A Technical Reference Manual For Plate Heat Exchangers in Refrigeration & Air Conditioning Applications
A Technical Reference Manual For Plate Heat Exchangers in Refrigeration & Air Conditioning Applications
A Technical Reference Manual For Plate Heat Exchangers in Refrigeration & Air Conditioning Applications
by Dr. Claes Stenhede/Alfa Laval AB Fifth, revised edition, November 17th. 2008
Alfa Laval AB
II
No part of this publication may be reproduced, stored in a retrieval system or transmitted, in any form or by any means, electronic, mechanical, recording, or otherwise, without the prior written permission of Alfa Laval AB. Permission is usually granted for a limited number of illustrations for non-commercial purposes provided proper acknowledgement of the original source is made. The information in this manual is furnished for information only. It is subject to change without notice and is not intended as a commitment by Alfa Laval, nor can Alfa Laval assume responsibility for errors and inaccuracies that might appear. This is especially valid for the various flow sheets and systems shown. These are intended purely as demonstrations of how plate heat exchangers can be used and installed and shall not be considered as examples of actual installations. Local pressure vessel codes, refrigeration codes, practice and the intended use and installation of the plant affect the choice of components, safety system, materials, control systems, etc. Alfa Laval is not in the business of selling plants and cannot take any responsibility for plant designs.
Copyright: Alfa Laval Lund AB, Sweden. This manual is written in Word 2000 and the illustrations are made in Designer 3.1. Word is a trademark of Microsoft Corporation and Designer of Micrografx Inc. Printed by Prinfo Paritas Kolding A/S, Kolding, Denmark
ISBN 91-630-5853-7
III
Content
Foreword. 1. Applications. The Basic Refrigeration Cycles and Applications.
1. The pressure-enthalpy diagram. 2. The basic cycle. 3. Intercoolers & subcoolers. 4. The components.
4.1. The expansion process and devices. 4.2. The compressor. 4.3. Evaporators, condensers and interchangers. 4.4. The interconnecting pipe work. 2 2 4 4 4 6 6 6 6 8 1 2
18
5. The flash economizer cycle. 6. The evaporator economizer cycle. 8. Definition of capacity and coefficient of performance. 9. Improvement of the cycles and the use of PHE in these. 10. Subcooler/superheater cycle. 11. The cascade cycle. 12. The absorption cycles.
12.1. Theory. 12.2. Applications. 12.3. The ammonia/water system. 12.4. The water/lithium bromide system.
20 14.1. Applications. 20 14.1.1. Cold stores. 20 14.1.2. Heat pumps. 20 14.1.3. Reversible A/C-Heat pumps units. 20 14.2. Design considerations. 20 14.2.1. Co- vs. counter current flow. 20 14.2.2. Freezing risk at the flow reverse. 22 14.2.3. Suction line separators. 22 14.2.4. Condensate subcoolers. 22 14.2.5. Oil separators, sight glasses and filter/driers. 22 14.2.6. Liquid receivers in reversible systems. 24 14.2.7. Installation of the liquid receiver. 24 14.2.8. One or two TEV systems. 24 14.2.9. Systems with a liquid receiver. 24 14.2.10. Systems without a liquid receiver. 27 15.1. What is a SECOOL system? 15.2. Direct vs. indirect systems. 15.3. Components. 28 28 28 28 32 32 33
34 40 42 44 44 44 44 44 45 46 46 47 47 47 47 48 48 48 49 49 50 52 52
3. SECOOL optimization.
3.1. What is SECOOL optimization. 3.2. Existing SW pumps. 3.3. Fixed pumping cost. 3.4. Summary.
IV
Content
54 60 60 60 60 62 62 62 62 64 64 64 64 64 66 67 68 68 68 68 68 68 70 70 70 70 72 72 72 72 72 74 74 74 74 76 76 76 76 76 76 76 76 76 76 78 78 79 79 79 79 79 79 79 81 82 82 82 82 84 84 5.4. The valve-evaporator system. 5.4.1. The bulb filling. 5.4.2. The valve characteristic curve. 5.4.3. The evaporator characteristic curve. 5.5. Matching the evaporator to the valve. 5.5.1. Operating point. 5.5.2. Capacity of the system valve-evaporator. 5.5.3. Stability of a valve-evaporator circuit. 5.5.4. Noise in an evaporator. 5.5.5. The evaporator stability curve. 5.5.6. Feedback oscillations in a valveevaporator loop. 5.5.7. Response of a valve-evaporator system to a sudden temperature change. 5.5.8. Factors affecting the valveevaporator behaviour. 5.5.9. Stability criterion of the valveevaporator system. 5.5.10. Choosing & installing an expansion valve. 5.5.11. Evaporators with distributors. 5.6. Troubleshooting the system valve-evaporator. 5.6.1. Liquid hammering. 5.6.2. Hunting. 5.6.3. Insufficient capacity. 5.6.4. Too low suction pressure. 5.7. Non-refrigerant BPHEs. 84 84 85 85 86 86 86 86 86 87 88 88 88 88 88 90 90 90 90 90 92 93 94 94 94 94 96 96 96 98 98
2. Properties.
2.1. Flow regime. 2.2. Control. 2.3. Compactness. 2.4. Liquid volume. 2.5. Pressure and temperature limits. 2.6. Thermal efficiency. 2.7. Fouling.
3. Flow arrangement.
3.1. The one-pass BPHE. 3.1.1. Evaporator. 3.1.2. Condenser. 3.1.3. Double nozzles. 3.1.4. Back end nozzles. 3.1.5. Vents and drains.
7. Thermal guarantees.
2. Evaporator classifications.
2.1. Pool boiling units. 2.2. Flow boiling units. 2.3. Refrigeration evaporators.
4. Vapour-liquid separators.
4.1. The purpose of this section. 4.2. The function of the separator. 4.3. Operation. 4.3. Equipment of a separator. 4.5. Separators for S&THEs and PHEs. 4.6. Separator types. 4.7. The horizontal & hybrid separators. 4.8. Vertical separators. 4.9. Summary of separator design.
6. Control.
6.1. Refrigerant expansion in a flooded evaporator. 6.2. The electronic expansion valve. 6.3. Matching the evaporator and compressor capacities. 6.4. Capacity control. 6.5. Pressure control. 6.6. Liquid temperature control. 6.7. Condenser/Evaporator.
Content
100 106 106 106 107 107 107 108 108 108 108 110 110 112 112 112 112
6. Control of condensers.
6.1. Condenser control requirements. 6.2. Change in a condensing parameter. 6.3. Control on the water side. 6.4. Control on the refrigerant side.
7. Desuperheating.
7.1. Draining of a desuperheater. 7.2. Desuperheating of a vapour. 7.3. Control for heat recovery.
4. Liquid receivers.
4.1. Types and purposes. 4.2. The pressure equalization line. 4.3. Operation.
8. Troubleshooting.
8.1. Insufficient capacity. 8.2. Instabilities. 8.3. Compressor HP cutout.
5. Venting.
5.1. Parameters affecting the venting.
114
122 122 122 122 122 124 124 124 124 124 124 125 125 125 125 125 126 126 126 126 126
6. Leakage in a BPHE.
6.1. Causes of leaks. 6.2. Leak seeking. 6.2.1. General. 6.2.2. Inspection of the system. 6.2.3. External inspection. 6.2.4. Locating the leak. 6.2.5. Cutting the BPHE. 6.2.6. Dissolving the BPHE in nitric acid. 6.3. Examining the result.
VI
Content
134 140 142 142 142 142 142 142 143 143 144 144 144 144 144 146 148 149
2. Oil cooling.
2.1. Temperature shocks at on-off cooling. 2.2. Temperature shocks at parallel connected compressors. 2.3. Control of oil coolers. 2.4. Cooling media at oil cooling. 2.5. Conclusion.
3. Refrigerants.
3.1. General.
9. Appendices.
Appendix I. Definitions.
1. Equipment. 2. Plates and channels. 3. Thermal and hydraulics.
152 152 152
152 152
152 153
154 158 158 159 159 160 160 160 160 160 160 160 160 160 160
5. The droplet settling velocity. 156 6. Correlations for the settling (terminal) velocity. 156 7. Example. 8. Pressure drops in a thermosiphon loop.
8.1. Separator - evaporator. One-phase pressure drops. 8.2. Evaporator. One- and two-phase pressure drops. 157 158 158 158
9. Pump circulation.
164
165
166 166
167
170
Foreword
The brazed plate heat exchanger was developed by Alfa Laval in the late seventies. It has increasingly proven useful in applications where compact, rugged heat exchangers with high heat transfer coefficients and high heat recovery are needed. One of the most important applications is refrigeration system, the topic of this manual. This is the fifth edition of the manual. Since the first edition, there has been an intense debate about refrigerants, both new, with and without glide and natural, such as propane and ammonia. This debate has influenced the development of equipment including plate heat exchangers. One of the effects on PHEs has been that the old boundary between the small BPHE for halogenated hydrocarbons in direct expansion systems and the large semiwelded PHE for ammonia in flooded systems has disappeared. Ammonia is increasingly used in direct expansion systems in nickel brazed PHEs and there are large brazed PHEs suitable for flooded systems. Hence, this new edition deals not only with brazed PHEs but also with the larger welded and semi-welded PHEs. After all,
the thermal and hydraulic properties are the same for all PHEs, from the smallest brazed to the largest gasketed. The purpose of the manual is to provide an understanding of the properties of the brazed and semi-welded plate heat exchanger in order to design, install, operate, and troubleshoot them correctly This Technical Reference Manual is divided into two sections. The first - the present booklet - contains general information about BPHEs in refrigeration systems. It comprises eight chapters and four appendices. Although all the chapters can be studied individually, some cross-references are made. The second part - the product catalogue - contains specific data about the product range. For the content of this manual, I have had the invaluable help of colleagues within Alfa Laval, in both Italy and abroad, but especially Bernard Pasquier, Loris Sartori, Mats Stromblad and Alvaro Zorzin to whom I express my thanks. Finally, I thank an understanding wife and son for their patience during all the working evenings. Especially Eric thinks now that a father comes with an integrated laptop.
1. Applications.
1. Applications.
BPHE condenser Capacity: 48.20 kW Cond. temperature: 40 C G. 40 C Cooling water F. 40.0 C Tap water
A. 35 C 100 % liquid 0 % vapour BPHE desuperheater. Cap.: 17.86 kW E. 114.6 C High pressure side
DX valve
-30 C, 66..1 % liquid 33..9 % vapour D. -25 C 100% vapour 1000 kg/hr R 22 COP.: 42.54/23.52 = 1.81 C. -30 C 100 % vapour B. - 30 C 33.9 % vapour Pressure, P (Saturation temperature)
Liquid A. 35 C Isotherm (-25 C) Two phases Vapour curve DX-valve Liquid curve C. -30 C G. 40 C BHE condenser The critical point
Isentrop
BHE desuperheater F. 40 C
E. 114.6 C
Compressor Vapour
D. -25 C
Enthalpy, H
1. Applications.
The actual performance of both the basic cycle and of any of the improved cycles obviously depends on the total required capacity, the efficiency of the compressors for the actual operating conditions, type of condensers and evaporators, availability of water, available space, etc.
4. The components.
4.1. The expansion process & devices.
In all types of the refrigeration cycle there is a step where liquid refrigerant is expanded from the condensing pressure to the evaporating pressure, from A to B in Fig.01. Refrigerant evaporates and simultaneously cools down to the evaporation temperature. The expansion device could be any type of restriction which exerts a pressure drop on the fluid which is equal to the difference between the condensation and evaporation pressure for the actual refrigerant flow rate. Fluid dynamic theory teaches us that the pressure drop increases with: Increasing flow length surface roughness direction changes de/accelerations hydraulic diameter. Cross-section area.
Decreasing
Theoretically, all these parameters could be used to control the pressure drop but it is difficult to imagine a device, which uses the surface roughness. Similarly, the number of directional changes and the flow length are closely interrelated as are the hydraulic diameter, the cross section area and the de/acceleration. There are three methods to control the pressure drop: A valve. A multitude of designs exists, but the basic parameter, which is changed, is the cross section and thus the velocity. Simultaneously the hydraulic diameter changes, the flow changes direction a little and the necessary acceleration energy is not fully recovered when the flow finally decelerates. Flow length and roughness have little importance. An orifice. The same principle as a valve but fixed. A capillary tube, which is a thin long tube usually spiral wound to save space. All the parameters above are important and have to be balanced with each other. Compared to an orifice, which is a hole in a wall, the cross section can be made larger as the flow length plays an important part for the pressure drop. This is important for small capacity devices, where an orifice would require such a small hole that it could easily be clogged by wear and tear products. It is usually a fixed design used for small systems such as refrigerators.
1. Applications.
85.6 C 35 C Air-cooled condenser Capacity: 59.34 kW Condensing temperature: 40 C 1000 kg/hr R 22 48.53 Liquid receiver C.O.P.: (9.53+10.67) = 2.40 18 C 2nd stage compressor Eff. : 80 % Cap.: 10.67 kW Pd/Ps: 3.08
S4
S1
27 C Well water
S3
S2
15 C
S4
S1
15 C Well water
-25 C
S3
S2
27 C
35.1 C 18 C
S4
BPHE evaporator Capacity: 48.53 kW Evap. temp.: -30 C
S1
-25 C
S3
S2
Compressor I -25 C
Enthalpy, H
1. Applications.
From the expansion device, a two-phase liquid-vapour mixture emerges. The liquid part has to evaporate in the evaporator, where it supplies the cooling effect. There are two ways to do this: A direct expansion evaporator. All the mixture enters the evaporator, the liquid part evaporates and a pure vapour leaves the evaporator and enters the compressor. To ensure that no liquid droplets enters the compressor, where they could cause liquid hammering and damage, the vapour has to be somewhat superheated. The superheat is also an excellent variable for the control of the expansion valve. Of special importance is the thermostatic expansion valve, which is controlled by the vapour superheat from the evaporator. As this is of prime importance for the function of a direct expansion evaporator, its function is treated in the evaporator chapter. The electronic expansion valve is increasingly common. The advantages, flexible control parameters and remote control, make it is very popular in supermarkets with numerous air conditioning units and refrigeration duties at different temperature levels. The most dependable valves are equipped with a heat motor, but pulse modulated and step motors are also used. Flooded flow. The vapour and liquid are separated, usually in a vessel large enough for the liquid droplets to settle by gravity or, sometimes, by centrifugal force. The liquid together with recirculated liquid enters the evaporator, partly evaporates, and the emerging vapour-liquid mixture re-enters the separator, where the vapour and liquid separate. The circulation can either be natural - thermosiphon or by a pump - forced circulation. The vapour joins the expansion vapour and leaves for the compressor. The liquid mixes with the expansion liquid and re-enters the evaporator. The vapour from the evaporator is neither dry or superheated. Thus, there is no superheat to be used as a control variable. Expansion is controlled by the level in the separator, instead. If the level decreases, the expansion valve opens and more refrigerant enters the separator-evaporator loop. See figure 02
1. Applications.
95.2 C
Air-cooled condenser Capacity: 61.62kW Condensing temperature: 40 C Liquid receiver Vapour, 210 kg/hr, 0 C 0 C
0 C, 21.0 % vapour Expansion valve A, controlled by the liquid level. Vapour-Liquid Separator Liquid, 790 kg/hr 0 C -25 C
The compressors can be two individual compressors or a screw compressor with an intermediate economizer inlet.
34.51 C
BPHE evaporator Capacity: 43.08 kW Evap. temp.: -30C 1st stage compressor Eff. : 80 % Cap.: 7.41 kW Pd/Ps: 3.04
40 C
95.2 C
Compressor II
DX-valve B
Separator -30 C
Mixing: 27.2 C
34.5 C
Compressor I -25 C
-30 C
BPHE evaporator
Enthalpy, H
1. Applications.
sometimes the electricity input to the motor) for given conditions. The given conditions include suction & discharge pressures, superheating, subcooling, compressor efficiency, cycle type. The semi-ideal COPs are used by designers to evaluate a new refrigerant, the effect on the performance by addition of a condensate subcooler, an economizer, etc. The cycles presented in this chapter use this COP. It is, thus, easy to see what the addition of an economizer or a condensate subcooler would mean. COP, plant: Refrigeration or heating effect of the plant divided with the total energy input to the plant, in consistent units for specific conditions. The plant COPs are used by the operators of a plant. A potential customer of a residential heat pump, comparing various machines, is probably only interested of how much heat a unit delivers at a given temperature, for a given heat source and electricity consumption. Whether a heat pump achieves a certain COP by addition of an economizer or a separate subcooler or a larger evaporator is of little importance to the final customer. Compression efficiency. The ideal adiabatic work (as found from the diagrams) of the compression of a volume of vapour to the actual work delivered by the piston or rotors of the compressor. Mechanical efficiency. The ratio of the work defined above to the actual shaft input work.
1. Applications.
71.7 C
5 C WPHE condenser Capacity: 56.08 kW Condensing temp. : 40 C 35 C 262 kg/hr, 35 C 0 C Cooling water
35.1 C
5 C
DX valve A
Unit cooler Capacity: 39.03 kW Evap. temp.: -30C DX valve B : -30 C, 17.5 % vapour
Compressor I -25 C
Enthalpy, H
10
1. Applications.
bcoolers/superheaters in low temperature systems, Economizers, Condensers/evaporators in cascade systems, Heat recovery in general, as well as numerous applications not directly connected to the refrigeration plant.
The evaporator operates with a lower inlet vapour fraction and a lower mass flow, which can decrease the K-value. It is possible to control the evaporator with the superheat after the interchanger. The evaporator can then operate with a much lower superheat or even wet without danger of liquid hammering. The MTD and the Kvalue increases as well. This should be done with an electronic expansion valve, though. Normal TEVs are apt to be unstable at this use. A BPHE, with its close temperature approach, is a good choice for the interchanger. If only a few degrees superheat is necessary, it will be too large. Common for both methods are: The circulating refrigerant flow decreases, important in cases of long refrigerant lines. The compressor operates at a considerably higher temperature than the evaporator. This eliminates the risk in a two-stage system that oil suitable for the high-pressure stage will become too sticky and lose its lubricating ability at the temperatures of the low-pressure stage. Moreover, at very low temperatures, the compressor materials could become brittle. Oil droplets are effectively stripped of its refrigerant content. Oil containing too much refrigerant could otherwise impair the compressor lubrification. This is mainly a problem for refrigerants with dissolves oil well, such as butane, propane, propene and the fully chlorinated hydrocarbons. The discharge temperature from the compressor increases. This could be an advantage in a heat pump, which heats tap water by desuperheating the discharge vapour but it could also lead to unwanted high temperatures. The vapour in the intermediate stage then has to be cooled, either by water, condensate injection or a combination, see figure 5. Example. A compressor, with a swept suction volume of 1 m3/s, compresses propene of a saturation temperature of 3.5 C and a superheat of 5 K. The condensation temperature is 43.5 C and the condensate leaves at 40.5 C. The vapour density is 13.291 kg/m3. The enthalpy difference between the evaporator inlet & exit is 283.08 kJ/kg giving a capacity of 13.291*283.08 = 3762.4 kW. The evaporator is changed and superheats the vapour to 23.5 C. The vapour density decreases to 12.322 kg/m3. If the swept volume is the same, the refrigerant flow decreases to 12.322/13.291 = 92.7 % of before. The refrigerant heats up 15 K more which means an increase of the enthalpy difference with 308.33 kJ/kg, i.e. with 8.9 %. Despite the lower vapour flow, the capacity increases from 3762.4 to 12.322 * 308.33 = 3799.2 kW. As an evaporator cannot normally manage 20 K superheat then try the superheater/ subcooler. Here the vapour superheats 8.5 to 23.5 C equal to 25.25 kJ/kg. The condensate enters the TEV containing 25.25 kJ/kg less enthalpy and thus has to expand less and enters the evaporator containing 0.0678 kg/kg more liquid refrigerant. The net result is thus the same increased cooling capacity, not by heating the vapour further, from 8.5 to 23.5 C but by evaporating 0.0678 kg/kg more refrigerant.
1. Applications.
11
1000 kg/hr R 22 46.91 C.O.P.: 22.86 Air-cooled condenser Capacity: 46.91 kW Condensing temperature: 40 C = 2.05
35 C
130 C
Brine
130.4 C
DX-valve
Compressor
Enthalpy, H
12
1. Applications.
There are other methods, though. In the absorption cycle, the vapour leaving the evaporator is absorbed (dissolved) in an auxiliary liquid fluid - the absorbent. The pressure of the liquid solution is then increased by a pump. Pressure increase of a liquid by a pump is a simpler and cheaper operation than compression of a vapour. Another advantage of the absorption cycle is the silence of operation. A pump makes considerably less noise than a compressor. At the higher pressure, the refrigerant and the absorbent are separated by distillation or a simple evaporation. The result is, just as in the vapour compression cycle, a high-pressure refrigerant vapour, which subsequently condenses and then expands, i.e. the normal cycle. At distillation, low-grade heat is added in a generator (reboiler). This heat replaces the mechanical energy at the vapour compression. Apart from a high-pressure refrigerant vapour, more or less regenerated absorbent is obtained, which then is recirculated to the absorber. There are mainly two types of absorption systems, ammonia/water and water/lithium bromide. Ammonia/water is a high-pressure system (~24 bar), suitable for compact installations. Ammonia is the refrigerant and water the absorbent. There are no particular corrosion problems except that copper and zinc may not be used. Mild/stainless steels are normally used. Water/lithium bromide is a low-pressure system (vacuum - 1 bar). Water is the refrigerant and LiBr the absorbent. The concentrated LiBr solution is potential corrosive against common construction material and special precautions have to be taken. Due to the large vapour volumes, the system tends to be bulky.
12.2. Applications.
The relative economy of the two cycles depends on the available energy. If low cost electricity is available, the vapour compression cycle might be the more economical. If low-grade waste heat is available at little or no cost, the absorption cycle might be the more economical solution. There are basically three types of application for absorption cycles, where plate heat exchangers can be used. Typically for most are that the heat for the generator has to be virtually free of charge and that the heat exchangers play a crucial part in the economy of the plant. Cogeneration. A dairy, a slaughterhouse, a fish processing plant, etc. have demands of electricity, steam/ heat and chilled water/refrigeration. The electricity is generated in diesel engine or a gas turbine and the resulting combustion gas produces steam. A part of the steam or possibly the combustion gas is then used for the absorption plant. Plate heat exchangers of the industrial types (all welded or semi welded) have been used for all positions in the absorption plant. Chemical industries at times produce large amount of excess heat, which can be used for production of chilled water. All- or semiwelded PHEs are used.
1. Applications.
13
1000 kg/hr R 22 54.52 C.O.P.: (10.28+15.24) 27 C Cooling tower water 32 C 35 C 2.5 C S&T condenser Capacity: 80.04 kW Condensing temperature: 40 C
73.8 C
= 2.29
The HP stage 1415 kg/hr R 22 tc/te: 40/-2.5 C 35 C Liquid receiver The HP stage HP DX valve: -2.5 C, 22.2 % 0 C
BPHE condenser/evaporator Capacity: 64.80 kW Condensing temperature: 2.5 C Evaporation " : -2.5 C
39.6 C
-25 C
The LP stage 1000 kg/hr R 22 tc/te: 2.5/-30 C 0 C Liquid receiver The LP stage
LP compressor Eff.: 80% Cap.: 10.28 kW Pd/Ps: 3.30 Unit cooler Capacity: 54.52 kW Evaporation temperature: -30 C
-30 C
Unit cooler
-25 C
Enthalpy, H
14
1. Applications.
Residential air conditioning/heat pumps. This is a new development. In many countries, especially in East Asia, there is a demand for residential heating during the winter and air conditioning during the summer. In contrast to more northern countries, which have an electricity consumption peak during the winter, the electricity demand is high throughout the year. The natural gas network has a consumption peak during the winter. The increase in air-conditioning units of the last few years has lead to an overload of the electricity grid during the summer while the natural gas net is underused. As in all residential A/C units, there is a need for silent units. Absorption A/C-heat pumps are thus an excellent alternative. However, units based on S&THEs will simply be too heavy (~ 250 kg) to be of practical use. These are sold as white wares, like refrigerators, stoves, etc. and have to be delivered on a pick-up truck by two men. The soltion could be heat pumps/air conditioners based on nickel (or other ammonia resistant materials) brazed heat exchangers. The weight can then be halved. More development of these systems is necessary, though.
All duties in an ammonia absorption plant can be carried out by PHEs, ranging from Ni-brazed units in small domestic air conditioning units to welded or semiwelded PHEs in large industrial or commercial plants. The thermal problems are much the same regardless of size. Most positions are uncritical and can be treated as normal one or two-phase heat exchangers but some merit special consideration. The rectifying condenser. In the figure, a rectifying condenser is installed. This condenses a part of the vapour and returns it as a reflux to the column. A reflux is necessary to provide the part above the feed inlet - the rectifier - with liquid. The vapour ascends in the condenser and meets the descending condensate flow, i.e. in counter current. This requires a condenser with a very low pressure drop. The plate heat exchanger can be used here but only if the velocity/pressure drop of the vapour is low, otherwise there is a danger of flooding the condenser. A low pressure drop is almost automatically ensured if the condenser is designed with high heat recovery, i.e. only a few degrees temperature difference. If the condenser is designed for cooling water, which normally means a larger temperature difference, the number of plates decreases and the pressure drop increases, creating conditions for flooding. It is difficult to give exact design condition. A calculated pressure drop of 0.05 - 0.1 kPa/(m flow length), corresponding to a flow of 20 kg/(m, hr) of 20 bar ammonia vapour in a 2 mm channel seems to be safe. It is also possible to operate the rectifier with downwards flow of both vapour and condensate. The position of it then becomes critical as a liquid column is necessary to overcome the pressure drop and force the condensate back into the column or an extra pump must be used. Thermodynamically this is less efficient than upwards flow as this separates ammonia and water more efficiently. The ammonia evaporator. It operates as a normal pure ammonia evaporator. However, in some cases, especially if a thermosiphon evaporator is used, the water content of the evaporated vapour is less than in the incoming condensate. Water then concentrates in the evaporatorseparator and the evaporation temperature increases. A special additional evaporator - similar to the oil evaporator in a freon thermosiphon, see 16 or 8. Oils & Refrigerants - is then necessary. A DX-evaporator is then a better choice, especially if it can be arranged so that all liquid droplets leaving the evaporator move directly to the absorber, without encountering any pockets, where they can collect. The best would be to have liquid injection at the top of the NBPHE but up to now all designs of evaporator with liquid injection at the top have shown a clear reduction in performance of the order of 25 to 30 %.
Finally, the strong aqua should be heated to the distillation temperature and the weak aqua cooled to facilitate the absorption. This is done in the feed preheater.
1. Applications.
15
Note! Only nickel brazed or welded units can be used in systems containing ammonia. Semiwelded PHE All welded PHE
Reflux 95 %, 41 C
2. Liquid receiver.
LP Ammonia, 99.95%, 2 C
Feed: Strong aqua, 50 %, 77 C Stripping section 44 C Return vapour Bottoms: Weak aqua 40 %, 96 C
8. Weak aqua strong aqua interchanger. 10. Generator. Strong aqua -7 C Weak aqua 4. DX-valve. 0 C -3 C Brine Cooling water or tap water for heating. Heat source, e.g. combustion gases, waste water, steam, etc.)
16
1. Applications.
Ammonia vapour
It is an unsolved question whether the inlet should be mounted from the top or from the bottom. Most installations are from the top. It is however easier to ensure a good distribution, especially of the liquid, from below, but the flow may be unstable at least at low capacities. The generator. The actual design of this depends on the type of heat source available. In industrial plants, where steam is available, a welded PHE is a good choice. From an operational point of view, it is uncritical. It performs as a normal steam generator. Domestic types of heat pump/air conditioner usually have natural gas as a heat source and the generator is integrated with the burner. The subcooler/superheater. In a compression cycle this is a questionable unit, see 10, p.10, due to the decrease of the circulating amount of ammonia. An absorber is not that much affected by the vapour density, the little more energy necessary to be cooled off can easily be accommodated in the absorber cooling part, possibly with a small increase of the cooling surface and/or amount of cooling water. Above all, the increased superheat of the ammonia helps to evaporate the last water in the liquid. Water leads to a substantial increase of the dew point, see 8. Oils & refrigerants, Table 3..
Weak Aqua A. Injection at several points into the pipe, before the absorber. Commercial nozzles, which give a cone shaped jet are used. B. Pipe with regularly spaced distribution holes, inserted into the port hole. Strong Aqua
1. Applications.
17
7. Separator
9. Condenser
8. LT generator
88 C 150 C
90 C
50 C
1. Evaporator
3. Solution pump
HT vapour HT condensate
LT vapour L T Condensate
1. The evaporator. Cold water evaporates under vacuum and cools the chilled water. At these temperatures, water vapour has a very low density and requires special types of heat exchangers. PHEs - brazed or gasketed - are usually not suitable for handling low-density vapours. After the evaporation, the water vapour moves to: 2. The absorber, which absorbs the water vapour in a strong solution of LiBr. 3. The pump, moves the resulting weak solution via: 4 & 5. The LT & HT regenerators, which heat the weak solution to the vicinity of the boiling point. A high regenerative efficiency is crucial to the economy of the process. The long temperature program make these duties excellent for PHEs - most likely in two-pass. After the regenerative, the weak solution enters
6. The HT generator. A part of the water boils off, usually in a gas fired boiler. The resulting strong solution releases its heat in the HT & LT regeneratives, more water evaporates in the intermediate placed LT generator. The HT vapour continues to 7.The separator, to be used as a heating medium in 8.The LT generator boils off water from the rich solution at a lower temperature/pressure than in the HT stage. The use of vapour from a HT stage to heat a LT stage is commonly used in evaporation plants to improve the economy. The conditions are similar here. 9. The condenser. Both the direct vapour (HT stage) and the LT vapour condenses here. Here too, the vapour volume is too large for PHEs. The resulting condensate, expands to the evaporation temperature in 10. The expansion valve and then to the evaporator.
18
1. Applications.
If the deep freeze store is remote from the machine room, the R404a lines might be too long. It might then be better to place the refrigeration unit close to the store and cool the condenser with brine from the main unit. Another alternative is a heat pipe with carbon dioxide as heat (refrigeration) carrier. A copper brazed PHE could then be used as the condenser. Compression from the evaporation temperature -12 C to the condensation temperature 45 C gives a low C.O.P. More importantly, ammonia gives very high discharge temperatures, in this case around 160 C. This is far too high, leading to oil break down and seizure of the compressor. The plant is therefore equipped with two compressors connected in series (or a two-stage compressor with intermediate refrigerant inlet) and an economizer (see Fig. 3); here a nickel brazed unit. This has the double function of increasing the C.O.P. and decreasing the discharge temperature to a more reasonable 96 C. A discharge temperature of 96 C is still high enough to recover the heat for hot water production. Water of a temperature of 80 to 85 C can easily be produced in a PHE. The condenser is hence equipped with a desuperheating section for production of hot water. 20 to 25 % of the cooling energy can be recovered as hot water. The separator is composed of two parts. The horizontal part mainly serves as a vapour liquid separator. The ammonia liquid level is maintained in the vertical vessel. Because of the comparatively small cross section of this as compared with the horizontal vessel, the ammonia content can be kept small. The liquid level in a separator is not quiet. To get a stable reading for the float, which controls the expansion valve, the float and valve is placed in a separate vessel, which communicates with main vessel. Such a valve-vessel combinations are sold as integrated units. Some are especially made for placement at the exit of a semiwelded plate condenser, Oil is insoluble in and heavier than ammonia. Oil from the compressors collects at the lowest point of the thermosiphon loop. From these points, the oil drains to the oil tank through the pipes with the valves A. There will inevitably be some ammonia entering the oil tank. The ammonia evaporates and the outside of the tank will be covered by ice. When the tank is full of oil ammonia cannot enter, the temperature increases and the ice melts. This serves as a signal, visual or by a thermometer, that the oil tank is full. The valves A are then closed and the valves B are opened. Through one of the valves B, highpressure ammonia pushes the oil out of the tank through the other valve B and to the secondary oil tank. This feeds oil to the compressors. If the ambient temperature is not high enough to evaporate the ammonia, an electric heater or a coil with condensate can be installed to improve the evaporation. A pump might be used instead of HP ammonia.
1. Applications.
+1 C
Liquid cooler
-30 C 12 C -20 C -12 C 12 C Level controlled Expansion valve Economizer, ammonia Cooling water Hot water 12 C Oil circuit 60 C 0 C Vent R404a 13 C
40 C
35 C
Desuperheater -4 C -8 C
96 C
Condenser
B
Flooded ammonia evaporation system with automatic oil return to the compressors. Brine tank
Brine, Tank - Evaporator An R404a DX-unit cooler connected in cascade to the ammonia system via an NB evporator/condensor unit
Economizer NB unit
19
20
1. Applications.
1. Applications.
21
Component
Best compressor Worst compressor Refrigerants TEV & Evaporator: Best evaporator Evaporation temp. Condenser Best condenser Condensate temp.
Heat pumps
Hermetic, indoors or outdoors Open, outdoors Water or brine better than air As large as possible Floor heating, direct would be the best but not feasible. Kept at sufficient level through a pressure controller, especially important in the winter. This level has to be as low as possible (large condenser) to obtain a high COP. Note that the various condensate subcooling methods are normally not justified. Important in Europe Not so important in US, APR
COP
10
80
R 22
50
R 40 7C
R 22
R 407C
0 Capacity, 100 % BPHE (evaporator) 20.6 kW Water 12 7 C R22 2C + 5 K 2 C Air coil (condenser) 25.9 kW R22 45 C + 35 K 42 C Air 35 C 30 C
40 Load, 100 % BPHE (evaporator) 23.1 kW R22 50 C + 39 K 47 C Water 45 40 C Air coil (condenser) 25.9 kW Air 10 C 7 C R22 0 C+5K 0 C
Fig. 11A. The temperature program in the BPHE at the cooling cycle.
Water Refrigerant
Fig. 11B. The temperature program in the BPHE at the heating cycle.
A.Condenser. The vapour enters from below and desuperheats somewhat. It then enters the last pass and condenses in downwards flow in counter current. B.Evaporator. The flow reverses and the evaporation is in cocurrent. The critical superheating in the last pass is now in counter current flow. The advantage of this is set-up is especially noticeable for refrigerants with a glide, notably R407c. Due to the glide, there might be a temperature pinch in cocurrent flow, which thus has to be avoided.
The refrigerant is connected in such a way that the last pass is always in counter current to the water.
The water is always in upwards flow, which secures the water filling even at low pressure drops.
1. Applications.
Long flow lines for liquid and two-phase refrigerant might mean a large body of refrigerant to evaporate. The four-way valve could be stuck by ice, mechanical damage or fouling and prolong the freezing time. One TEV systems. The TEV - see below - in such a system has a lower capacity in the reverse direction, which might prolong the freezing period, figure 13. Two TEV systems as in figure 14 on the other hand contain more pipe work and extra non return valves, i.e. more liquid refrigerant. TEV installation. This theme is discussed later in this chapter and in detail in the evaporator chapter. By-pass. One possibility would be to use a small solenoid valve parallel to the TEV. This opens for a certain time during the switch over and supplies the evaporatorcompressor with refrigerant while the pressure difference is low or the TEV is closed. There are TEVs, which gives a refrigerant flow at zero superheat. Basically they are TEVs with an internal bypass (bleed port). These could possibly be used.
1. Applications.
23
HP side Condenser.
Countercurrent.
LP side
Evaporator.
Cooling.
+-
Condenser.
Defrosting.
Fig. 13. A pilot operated, four-way valve used in a reversible cooling unit.
The design and function of the pilot and main valves is similar. The pilot valve is actuated by a solenoid, the main valve moves by pistons operated by the discharge and suction pressures and controlled by the pilot valve. When the solenoid valve is deactivated, the pilot valve connects the left cylinder in the main valve to the compressor discharge pipe and the right cylinder to the suction pipe.
The cylinder-slide assembly (green in the figure) then moves to the right. When the solenoid valve is activated, the pilot valve connects the right cylinder in the main valve to the compressor discharge pipe and the left cylinder to the suction pipe. The cylinder-slide assembly (green in the figure) then moves to the left and reverses the refrigerant flow.
1. Applications.
danger as a consequence or if not evaporated, it will enter the compressor resulting in liquid hammering and possibly damages. See the preceding chapter about freezing. The two condensate pipes in figure 14 & 15 can be joined before being connected to the LR. This simplifies the installation as standard LRs can be used. By carefully balancing the various components, pipe lengths and diameters, it might be possible to design a system, at least a small one, where the refrigerant difference between the cycle modes is small and can be stored in the port hole of the BPHE and interconnecting pipes and a LR is thus not necessary, see figure 16.
1. Applications.
25
A. A/C mode.
Active HP line Active LP line Inactive HP line Inactive LP line DX & one-way valves closed 45 C + ~ 35 K
Condenser
The set-up fulfils the most important requirements for a reliable installation, notably that there is a straight pipe between the TEV and the BPHE and that there is a direct uninterrupted passage from the condenser to the LR. Moreover, the pipe from the TEV to the air coil enters the condensate line from above. It is thus not filled with condensate when the air coil operates as condenser. The corresponding pipe from the valve to the BPHE is short and enters horizontally. The little condensate, which collects here, is unimportant.
L3
2 C + 5 K
Room chiller
12 C
L2
7 C
Evaporator
Filter Drier
The pipe parts L1 to L3 have constant B. Heat pump mode. flow direction and can be used to install equipment, which shall be used in both forward and reverse modes. A filter-drier is here installed at L1. In this case, it is designed for bidirectional flow as it is installed with upward flow. A sight glass could be installed here as well. An oil separator could be installed at L3.
40 C
Evaporator
6 C+5K 50 C + ~ 35 K
L3
Room heater
A suction separator at L2. L2 A condensate cooler using the flows at L1 and L2. Condenser
45 C
26
1. Applications.
Active HP line Active LP line Inactive HP line Inactive LP line One-way valves closed
A. A/C mode.
Condenser
45 C + ~ 35 K
Evaporator
2 C + 5 K
L3
Room chiller
This installation has one TEV and four extra one-way valves to route the evaporating mixture around the LR. Without these valves the receiver would have been positioned between the TEV and an evaporator in at least one mode. It also means that the TEV always operates in one flow direction thus maintaining full capacity in both forward and reverse mode. The set-up fulfils the most important conditions for a reliable installation, notably the straight pipe between the TEV and the BPHE and the direct unbroken passage from the condenser to the LR.
12 C
L2
7 C
L1 Filter Drier
Liquid Receiver
Evaporator
6 C+5K 50 C + ~ 35 K
Condenser
L3
Room heater
Here is installed condensate cooler/suction gas super heater and a filter/drier using the flows at L1 and L2.
L2
45 C
LR
1. Applications.
27
Filter drier
Evaporator
Condenser
Condensate
The relative condensate level in the two BPHEs can be equalized by the valve horizontal position.
28
1. Applications.
substation on each service floor. The service floors are placed at regular intervals, typically starting with the 13th (which doesnt show up in the lifts in e.g. a hotel). There are many ways to arrange such systems. The simplest and the most effective use of the available temperature difference between the room air and the sea water would be to have a SW cooled condenser in an A/C plants placed midway in the building. One FW circuit would feed the upper part of the building, another the lower. The drawbacks are that the corrosive and fouling SW enters the machine room and there is limit on the height of the building. In the system shown in figure 17, circulating FW is cooled by SW. The FW cools the condensers in the small, local A/C plants, placed in the service floors. The FW circuit has to stand the full pressure from the top of the building to the basement. There is also a loss of temperature difference as the condensers are fed with water some 3 to 7 K higher than the SW. An A/C unit for the whole building is placed in the basement. A FW circuit - from the basement to the top - delivers chilled water to the fan coils via a PHE (pressure breaker). A pressure breaker is a small installation and could be installed in a service room on each floor or grouped together for a number of floors. The advantage of the system is obvious, all the machinery is kept in one place in the basement and only chilled water is delivered to the building. The disadvantage is that the A/C unit has to deliver chilled water some 3 to 7 C colder than the previous in order for the fan coils to operate with the same temperature program but on the other hand, the condenser has a higher water temperature.
15.3. Components.
Water intake. In figure 17, the water is lead through a short canal to the basements where the sea water pumps are situated. The advantage of such a system is the low civil engineering cost. The disadvantage is the sometime excessive amount of debris found close to shore. Water can then be taken from an intake tower some distance from the shore and some meters from the bottom and piped to the pumping station. The water at such a point is usually much cleaner than either at the surface, the bottom or the shore. Moreover, it can be piped to a pumping station some distance from the shore and the system then doesn't occupy a possibly useful shore area. Strainer or grate. A grate or a self-cleaning strainer prevents large debris from entering the pumping station. The sea water pumps. In figure 17, the pumps are placed lower than the water surface, thus with a positive inlet pressure and no priming system is necessary. It is also possible to place them somewhat higher than the surface but then a priming system is necessary.
1. Applications.
29
Fan coils
Fan coils
Fan coils
A secondary cooling (SECOOL) system to cool circulating fresh water with sea water. The fresh water is used to cool condensers in a high rise building.
30
1. Applications.
SW pumps are usually made with bronze impeller, stainless steel shaft and bronze or cast iron casing. Large pumps should have double suction inlet, i.e. the SW enters the impeller from both sides, fig. 18 shows a single. Self-cleaning filters. See figure 17. Various types exist. Alfa Laval has developed a filter especially for use with PHEs and for sea or river water, see special documentation. The casing is rubber clad mild steel and the mesh of stainless steel. The smallest size has a connection of 100 mm. It is a stationary filter, where the cleaning action comes partly from flow reverse, partly from shearing forces when the water is made to flow along the sieve surface. The cleaning takes place automatically at predetermined intervals or controlled by the pressure drop. The absence of moving parts (except at the flow reverse) makes the filter very dependable. At the flow reverse the debris is flushed out in a small waste water pipe, which preferably connects to the main SW line after the SECOOL PHE. Rotary types exist which cleans continually, but some types have a tendency to be stuck by the debris. Back flushing. Small debris, seaweed, fish, etc., which have passed the water intake can, in the absence of a filter, tend to be stuck at the PHE inlet. A simple way of getting rid of the debris is to reverse the flow in the PHE and flush them out, see figure 18. This is done automatically, usually four to six times a day, 15 to 30 minutes each time. Note 1. At the flow reverse, the flow changes from normally counter current to cocurrent. There might thus be a temporary decrease in capacity. Note 2. Except in very special cases, it is not necessary to install both a filter and a back flushing system. Note 3. A back flushing system has no particular waste water pipe, which simplifies the installation, not the least from a legislative viewpoint. Note 4. A direct cooled condenser should have a better antifouling system than an indirect as an opening involves the refrigerant side, which is costly and time consuming. Chlorination. Sea and brackish water contain a multitude of marine life, algae, fungi, barnacles, bacteria, etc., fresh water somewhat less. The larger organisms are caught in the filter but smaller pass. Barnacles and mussel larvae can pass almost any type of sieve, settle on the walls, grow and finally detach and clog the condenser. HE of mild steel cooled by fresh water can harbour iron-eating bacteria. Especially in winter, the condenser is a warm cosy place to hibernate in. An effective way of controlling unwanted marine growth - where legislation permits - is chlorination, especially shock chlorination at irregular intervals as e.g. barnacles tend to memorize regular intervals and then close. The extent of chlorination is difficult to predict, marine activity
varies from site to site, season, time of the day, etc. 15 minutes, four to six times a day to a chlorine content of 1 2 ppm after the PHE is a suitable starting value. The injection point should be immediately before the PHE. If far away, all the microorganisms in the pipe between the injection point and the PHE will detach and clog the PHE. There are three principal chlorination systems in use: Direct injection of liquid chlorine, the preferred system in chemical industries, where knowledge exists on how to handle liquid chlorine, (but probably nowhere else). Injection of a water solution of calcium or sodium hypochlorite. It is an easy and safe but somewhat messy method to chlorinate, as the solution is highly corrosive. The most elegant method is electrolysis of sea water. Electrolysis of a NaCl solution yields chlorine and hydrogen, right where it should be, in the water. It is expensive, but the advantage is that there is no storage of a corrosive and/or dangerous liquid. Chemical cleaning. SW contains not only microorganisms but also pollutants, such as oil, smoke, and dead organic matter. All of these can attach to the surface and have to be removed. Cleaning must also be used if chlorination for one reason or other cannot be used. In a CIP (Cleaning In Place) system, the SW flow is shut off and cleaning solutions are circulated through the PHE. Alfa Laval deliver two types of cleaning solutions especially developed for PHE, see special documentation. One is a type of industrial detergent, which is effective against various types of organic fouling. The other is an acid solution, which dissolves calcium or magnesium carbonate scaling, the most common types in cooling water. Manual cleaning. Usually it is advisable to open a GPHE once a year for manual inspection and cleaning. The intervals can vary though, from several times a year to once every ten years. If an annual cleaning, it should be done just before the warmest season, i.e. when the load of the condenser is the largest, it is recently cleaned. The PHEs. Only GPHE & WPHE are used for SW. The material in a SW PHE, whether a condenser (WPHE) or a SECOOL PHE (GPHE), should be titanium as this is fully resistant to sea water and to most other chemicals. The exceptions are fluorine compounds either in solution or as compounds in the gaskets, e.g. fluoro rubber - Viton - that never should be used in titanium PHEs. There are some high quality stainless steels, e.g. Avesta 254 (and 654) SMO, which are resistant to brackish water but not in combination with chlorination. See the section on 6. Fouling & Corrosion. Note 1! A SWPHE for R22 cannot operate with a higher
1. Applications.
31
In the cooling plant below, the oil recovery system is not shown. It could be either an automatic oil drain from the bottom in case of an ammonia system or an oil return evaporator in case of a soluble oil.
Fan coils
Apart from a strainer and a selfcleaning filter or a backflushing unit as here, sea water might need chlorination or other chemical treatment and a CIP unit to clean the condenser while in operation.
Fan coils
Fan coils
Valves usually open. Sea water pumps. One Shut at backflushing. operating, one stand-by.
32
1. Applications.
refrigerant to almost pure oil. Especially the viscosity shows a very large variation. The evaporation should be divided into various zones from inlet to outlet, each designed with its proper physical properties. This is unfortunately a very tedious task. However, when 95 % of the liquid has evaporated, the oil concentration has increased to 10 %. Up to about this point, the increase in viscosity is reasonably small. The remaining part, when the oil concentration increases from 10 to almost 100 %, is the difficult one. Fortunately, the amount of liquid is small here, and the evaporation process is liquid droplets in a gas, i.e. the liquid physical properties have little influence. In practice, it is possible to design the unit with mean physical properties or as a pure refrigerant with a large margin (25 - 50%). The design of an oil return evaporator is not very critical. Suppose that the unit is underdesigned by 50 %. This would mean an oil concentration of 0.75 %, still a manageable value for the main evaporator. Finally, of the 5000 kg/hr mixture entering the evaporatorseparator loop, 1000 kg/hr is drawn of for the oil return evaporator. Will this decrease the plant capacity? No, as the heat needed to boil of the refrigerant is taken from the condensate, which then comes in cooler and contributes to the cooling load. Figure 19 shows the oil return evaporator. A flow is drawn off from the separator, evaporated, and an oil mist returns to the compressor. Thus, the main properties of the evaporator are: An as complete evaporation as possible of the oil-refrigerant solution. Return to the suction line of an oil mist only. The driving force for the flow can be the static head as in a normal thermosiphon. The static head might not be enough to force the oil mist into the suction line. If the line from the oil evaporator enters the suction line after a pressure drop producing valve, the same pressure drop has to be added to the oil evaporator as driving force. The entrance to the suction line could be as an ejector, producing some extra driving force. The inlet to the oil evaporator could be made from the discharge pipe of the pump in forced flow system. The flow is controlled by the superheat but without flashing of the refrigerant as in a normal DX system. The control valve simply regulates the flow so that no more enters than what can evaporate. The allowable pressure drop is low, i.e. distributor BPHEs can probably not be used. Similar components as in a DX system can be used, though.
1. Applications.
33
-3/-2 C -3/-1.2 C -2 C 41.5/101.3 C 41.5/101.3 C Oil filter 42/96.4 C 42/101.3 C Oil tank Vent EQL 37 C 37 C Oil cooler LR V1 Oil pump Condensate column V2 -2 C 3,4 C Oil evaporator
-2 C
Fig. 19. Oil cooler and oil return evaporator in flooded system with soluble oil. 16.2. The oil cooler.
The system is the same for both a soluble and an insoluble oil. The oil is cooled with evaporating warm condensate. The other possibilities - cold refrigerant or water give larger thermal shocks. After the evaporation, the vapour is condensed in the condenser again. It returns to condenser via either: A) The LR and the equalization line (EQL), figure 19. B) Directly to the condenser inlet. C) Back through the condensate pipe to the condenser exit and into the channels from below. In A) & B) this means that the refrigerant moves in a circle. For this to be possible in a condenser with a fairly high pressure drop, the following conditions must be met: The LR pressure has to be higher than the condenser inlet pressure, otherwise vapour cannot flow back. This is easy; the condensate is heated in the oil cooler until the vapour pressure is large enough to press the vapour back to the condenser inlet. The vapour must not enter the condensate pipe. The condensate pipe in the LR thus has to be drawn to below the lowest condensate level encountered. To over come the pressure drop in the condenser there must be a sufficient gap between the liquid level in the LR and the bottom of the condenser, to allow a liquid column to build up. What happens at the start of the oil cooler? Suppose that the EQL is closed. When the EQL opens, the condenser inlet pressure transfers to the LR. The pressure pushes the condensate into the condensate pipe. When the static height in the column equals the condenser pressure drop, equilibrium is reached and there is no flow in the EQL. The pressure is equal in the LR and at the condenser inlet. The refrigerant valve to the oil coolers opens and refrigerant evaporates. The pressure increases enough for the vapour to be pushed through the EQL to the inlet. The liquid column increases correspondingly. If the refrigerant contains inert, it has to be vented at the upper part of condensate exit pipe, not at the LR. In case C) no EQL is necessary and the condensate pipe discharges above the condensate level in the LR. Vapour moves in counter current to the condensate and enters the channels. The liquid column, necessary to overcome the
pressure drop in the channels, then forms in the channels. Parts of the channels are then submerged and are rendered useless as a condensing surface.
It is difficult or impossible to vent the conden-
34
2. Optimization.
1.2 . Conclusion.
Valves & other control equipment, pumps, and pipes & fittings do not change cost for minor changes in the operating parameters and are thus not optimized. The compressors come with fairly large steps between the nominal capacities. PHEs can be almost infinitely changed and should thus be adjusted to the compressor and not the contrary. The study is divided into three parts: The thermal and hydraulic properties of PHEs, which can said to be the optimization of the PHE without regard to other parts of the system including other PHEs. This forms the basis for the next steps. Secool optimization from the application for which this concept was originally developed. The interaction of PHEs with other parts of the system notably other heat exchangers, which can be said to be optimization proper.
2. Optimization.
35
P = 45 kPa 2. The pressure drop in the connections falls and the total number of channels decreases.
1. The pressure drop in the connections takes an increasingly larger part and a second unit is added. 3. The process repeats Flow, kg/s Flow, kg/s
Fig. 02. The area as a function of the water flow for parallel PHEs.
Area, m2 M=5%
3. The MTD approaches zero and the area approaches infinity. 1. The MTD departs from a constant value and the surfaces increases.
Flow, kg/s
36
2. Optimization.
The main reason for the large area is the increasingly small MTD. We cannot improve the MTD, for given flow rates and temperatures. On the contrary, a heat exchanger could impair the true counter current MTD, even though PHEs are good in this respect. However, the other reason for the increasing area is the low K-value, caused by the decreasing channel velocity. By splitting the area between two units and connecting them in series, the channel velocity is doubled, resulting in an increased K-value and thus smaller area. See Fig 4. For still lower flow rates the area can be split between three, four, ..... units in series. This reduces the rate of the increase of area somewhat but when the temperature difference approaches zero, the area approaches infinity.
At the left part of the curve - the red - the margin is the specified one, 5 % in this case. The pressure drop is smaller than the specified 45 kPa. Case b) occurs here. Heat transfer duties with very large heat recovery or limited amount of water, i.e. close temperature difference, fall on the left part of the curve. The heat transfer determines the necessary area. In one point - case c) - the pressure drop as well as the margin is equal to the specified values. The area has a minimum when both conditions are fulfilled. In the actual calculations a) to c), the values are close to the required 5% and 45 kPa, but not exactly. The addition of area is not continuous but occurs in discrete steps, e.g., when a plate is added, the margin might jump from less than 5% to more than 5%. The minimum area does not necessarily occur for equal flow rates on the warm and cold sides. It is a common misconception, that a PHE should have both flows equal, in order to be optimal.
A change of the pressure drop on warm side has no direct effect on the optimal flow, see Fig.7. In this example, we have permitted the warm side pressure drop to be free to study the effects of the cold side pressure drop.
2. Optimization.
37
Area, m2
5. MTD-> 0 Area->
M=5%
Area, m2
2. When the flow decreases, so does the MTD & K and the area increases. 4.The process repeats. 1. For an infinitely large flow, the MTD & K-value are constant, thus a constant area. 3. Two units in series. Channel velocity & Kvalue increase. The area drops.
M=5% The optimal heat exchanger flow rate occurs at the interP = 45 kPa section of the pressure drop and margin curves.
Flow, kg/s
Flow, kg/s
Fig. 04. The area vs. water flow for PHEs in series.
Area P < 45 kPa M=5% P = 45 kPa Case: b) M>5% P = 45 kPa M=5% Case: c) 0.625 0.77 1.0 G2/G1 Case: a)
Fig. 05. Both the thermal and hydraulic data are in accordance with the requirements.
Area
Case: b)
Case: c)
Case: a)
0.625
0.77
1.0
G2/G1
Area
P=30 kPa
M = 10 % 2. The area (# channels) then increases and the optimal flow decreases. M=5% P= 45 kPa
G2/G1
G2/G1
Fig. 08. The effect on the optimal flow, by a decrease of the cold side pressure drop.
Fig. 09. The effect on the optimal flow, by an increase of the margin.
The heat transfer coefficient decreases with increasing viscosity and decreasing thermal conductivity. The dependence on the specific heat and density is less clear and in any case less strong. The heat transfer coefficient increases with increases in both but then less fluid - less channel flow rate - might be necessary. Anyhow, if we switch to a fluid with higher viscosity and lower thermal conductivity, the area - the number of channels - increases. That would normally mean that the flow could be increased, increasing the MTD and overall heat transfer coefficient, again lowering the area and the number of channels until equilibrium is reached. If the increase in viscosity was very large, the pressure drop increases and this might force a reduction in flow. The area will increase but the optimal flow rate can occur for a higher or lower flow rate than for the first fluid.
38
2. Optimization.
a) We change the cold-water inlet temperature in our case study to 4 C, and try to find the optimal flow: Water, 10 kg/s 12 -> 7 C, Water 10.9 kg/s 9 <- 4 C, Two (CB76-70H, 6.8 m2), M = 5%, P1 = Free P2 45 kPa P1/2 = 35/43 kPa
As an example, we will see what happens when we switch to 20% propylene glycol in our case study. The area for the water optimal temperature program is (figure 10): Water 10 kg/s 12 -> 7 C, P1= Free 20 % PG 8 kg/s 8.5 <- 2 C, P1= 45 kPa Two (CB76-70H, 6.8 m2), M = 5 %, P1/2 = 35/26 kPa The area - the number of channels - has increased and the pressure drop has dropped to 26 kPa. The margin is the required. It is firmly on the left side on the glycol curve in Fig. 10. An optimization of the duty gives: Water 10 kg/s 12 -> 7 C, 20 % PG 8.82 kg/s 8.5 <- 2 C, Two (CB76-56H, 5.4 m2), M = 5 %, P1= P1= P1/2 = Free 45 kPa 52/45 kPa
For TTL = 5/3 = 1.67 the flow increases to G2/G1 = 1.09. b) In the second example, we decrease the TTL to 0.5 by changing the inlet water temperature to 9.5 C: Water, 10 kg/s 9.5 -> 7 C, P1 = Free Water 4.54 kg/s 7.5 <- 2 C, P2 45 kPa Two (CB76-30H, 2.8 m2), M = 10%, P1/2 = 164/42 kPa Comment: Side 1 pressure drop is very high, i.e. there will probably a line optimum as in figure 7. The optimal flow rate for TTL = 2.5/5 = 0.5, is G2 /G1 = 0.454. Thus, we can make a table:
This final heat exchanger is not very much larger than the original water heat exchanger but the optimum occurs for a much larger flow rate. As a conclusion, we can say that if the physical properties tend to make the PHE larger, then the optimum tends towards a larger flow rate and a larger unit.
1.0 0.77
1.67 1.09
Deduced
See also figure 12, which shows the optimal flow as a function of the TTL with the margin and the pressure drop as parameters. The direction of the parameters could be deduced from the previous results. The above results are in line with sound engineering. For a very close temperature approach and/or a large temperature change - a high TTL - case a), common sense says that a large temperature difference would be advantageous, thus a low exit temperature and a large flow. For a small temperature change and/or a large approach a low TTL, case b) - there is a comparatively small area for large flows and the number of channels is determined by the allowable pressure drop. The heat exchanger is probably thermally over dimensioned. Thus, we reduce the flow rate and the number of plates is reduced. As the heat exchanger is thermally over dimensioned this does not matter.
2. Optimization.
39
Area
Change to a fluid with higher viscosity means a larger area and a larger optimal flow. 20 % propylene glycol
Temperature, C tw t1 tw K t1 K
TTL = tw K
G2/G1
2000 1500
Increasing margin and pressure drop
12 - 8.14 = 3.86 12 - 7.86 = 4.14 12 - 7.48 = 4.52 12 - 6.86 = 5.14 12 - 6.30 = 5.70 12 - 6.14 = 5.86
203.5
Fig. 12. The optimal flow rate as a function of the Terminal Thermal Length.
a) We want to cool 16000 kg/hr water of 12 C to 8.2 C with an equal amount of 2 C water. 2000 kg/(hr,ch) gives 8.14 C, close enough. 16000/2000 = 8 channels (double) will do the duty. The pressure drop might be somewhat on the high side, 129 kPa, though. What if only 10 kPa? 500 kg/(hr, ch) gives 9.75 KPa and an exit temperature of 6.86 C, but we need 32 channels. We are firmly on the right side in the curves of figures 5 to 10 and pressure drop only determines the number of channels. b) Suppose that we want to cool 10000 kg/hr water of 16 C to 8.8 C with equal amount of 2 C water. Suppose that the K-value does not change with the change of the temperature from 12 to 16 C. P permitted is 10 kPa. The theta of this duty is (16 -8.8) / (8.8 -2) = 1.059. The table gives a theta of 1.056 for 500 kg/(hr, ch). That means an exit temperature of 8.81 C and a P = 9.75 kPa, which is OK. Thus we need 10000/500 = 20 channels. c) We want to cool 20 350 kg/hr water from 12 to 6.1 C with a maximum P of 70 kPa. The table shows that a flow of 203.5 kg/(hr, ch) gives a temperature of 6.14 C. Thus 20350/203.5 = 100 channels are needed. The pressure drop is only 2.14 kPa and we are on the left side of the curve in figures 5 to 10. d) There is another way. Suppose that we connect two channels in series and send 1500 kg/(hr, ch) through them. Each channel would have a theta value of 0.707 and we would get the following temperature program: 12 7.86 Channel 1 -> 9.07 <4.93 Channel 2 -> 6.14 <2 C C
This is close to what we want but now we only need: 2 * 20350 / 1500 = 27.1 ~ 28 channels. Unfortunately, the total pressure drop will be twice the channel pressure drop, i.e. 2 * 74.8 = 150 kPa. In order to get to the required pressure drop of 70 kPa, we can allow 35 kPa/channel, i.e. what 1000 kg/(hr,ch) give, thus: 2 * 20350/1000 = 40.7 ~ 41 channels, which is still better than the original 100 channels. As an extra bonus, the exit temperature will be about 5.8 C instead of the required 6.1 C. The operating point is now to the right.
40
2. Optimization.
f) Example b) would require about 57 L-channels (175.5 kg/(hr,ch)) or 35 H-channels (233 kg/(hr,ch)). g) Example c) would require about 151 L-channels (134.5 kg/(hr,ch)) or 29 H-channels (705 kg/(hr,ch)) with a pressure drop of 58 kPa, close to the required 75. Table 4 shows the result, which can be summarized as : a) is a low theta duty. Thus, a low theta channel fits this duty. Higher theta channels - M & H - need more plates to maintain the required pressure drop. b) is a medium theta duty and an M-channel is the best. Both L- & H-channels would require more plates to give the required thermal and hydraulic performance. c) is a high theta duty and the H-channels are the best. The next best are M-channels in series.
At a certain channel depth, the heat transfer would be the required for the correct pressure drop and with a reduced number of plates. Another possibility is shown in d). The heat exchanger, by being connected in two passes in series, is simply doubled in length. The result is a doubling of the theta value, i.e. the thermal length is connected to the physical length and we can simply change the plate length to obtain the required properties Thus, we can use the channel depth and the length to adjust the thermal and hydraulic properties of a PHE. Increased length Decreased depth => => Higher theta & & higher P
2. Optimization.
41
H-channel
L-channel
M-channel
12 - 7.35 = 4.65 12 - 7.00 = 5.00 12 - 6.51 = 5.49 12 - 5.77 = 6.23 12 - 5.14 = 6.86 12 - 4.97 = 7.03
64
1.059
57
20
35
1.439
151 100(41) 29
H-channels M-channels L-channels Flow
###### Channel # minimum. ###### Channel # determined by P. ###### Channel # determined by heat transfer.
Fig. 15. The general aspect of the area-flow curve for of three, different theta, channels.
In the case of a PHE with multiple channels, especially if these can be mixed, the optimal point (Fig. 10) is replaced by an area (figure 15). This permits a PHE to utilize the pressure drop for a range of thermal lengths. It is sometimes necessary to use a PHE outside its normal range of thermal lengths. The reasons could be material demands, pressure resistance; no better type is available etc. It might thus appear to be very large, thus "bad". It might not be the case; it is simply not optimal for the given duty but could be it for another. Example. A PHE optimized for high heat recovery, i.e. a high thermal length, will be very large if used to cool a large amount of water only a few degrees.
42
2. Optimization.
Conclusion: The higher the K-value is, the higher the margin becomes for a given Rf. The problem is, that published Rfs are almost exclusively developed for S&THE. If applied on PHE, which usually have much higher K-values, the margins become large, exceedingly large sometimes. Some points to consider: a) Control problems. A heat exchanger does not operate with a margin. It adjusts the temperatures until the margin (and the fouling factor) become zero. The resulting large margins, if S&THE Rfs are applied to PHEs, could result in freezing or overheating a fluid if it is throttled to decrease the MTD, and possibly instable operation. b) Decrease of the shear stresses. One possibility to increase the margin in a PHE is to add more plates in parallel. The pressure drop and thus the shear forces then decrease very rapidly. The result could be the contrary to the desired, increased fouling. c) Plates in series. If another PHE is added in series to the first, the p doubles. To keep the p, plates are added in parallel to both PHEs. The resulting area margin will probably be much larger than the required at the same time as the shear stresses have decreased and the effect will be a combination of a) and b). d) Temperature difference. Suppose that PHEs are designed for each duty below, both with a zero margin. 42 -> 37 40 <- 35 MTD = 2 K 49 -> 44 25 <20 MTD = 24 K
The nominal K-value decreases with 20 % due to fouling. How much must the warm side temperature change to keep the nominal duty? If the K-values decreases with 20 %, the MTD has to increase with 1/0.8 => 25 % and the new temperatures are: 42.5 -> 37.5 40 <- 35 MTD = 2.5 K 55 -> 50 25 <20 MTD = 30 K
A temperature increase of 0.5 K might be acceptable but 5 K is certainly not. Conclusion. It is important to consider the effect of the margin on the temperature program. e) Plant type. If an A/C plant cannot keep the design air temperature during an extraordinary heat wave, the result might be an inconvenience. If the same happens in a hospital or a pharmaceutical industry, the result could be a catastrophe. Obviously, the layout of the two plants, including the margins for the PHE, must be different. f) Cleaning should be done just before the warmest season starts. The PHEs then meet the most severe conditions with the highest K-value. g) Plant type. The margin could be less in e.g. an office, which could be shut down during off hours as compared with a continuous operating plant such as in a hospital.
Summary: PHEs should not be overdimensioned. Depending on the type of duty and other conditions, the margin should not be more than 0 - 15 %, rarely 15 to 25.
2. Optimization.
43
A. We want to cool water, the blue line, from 12 C to 6 C. The selected heat exchanger channel is too short, the water it delivers at the exit is only 8 C. What can we do?
8 C 12 C 6 C
6 C
B. One possibility is to add more parallel channels. This adds more surface and the channels cool better. The P and the K-value decrease though, making the added channels increasingly ineffective. The permissible P might not be fully utilized.
12 C
6 C
C. Another possibility is to rearrange the channels from all in parallel to two groups in series, each group containing. half the number of parallel channels. This doubles the channel velocity and thus increases the K-value. The DP increases about seven times though, which might necessitate more parallel channels. The HE will then be thermally overdimensioned.
12 C
6 C
4 C
5.5 C
D. Still another possibility is to use a physically longer channel. This cools better but the P increases. Channels might have to be added in order to keep a P limit.
12 C 6 C
6 C
E. A similar effect can be obtained with a thermally longer channel. In this case, the corrugations are tighter which creates an increased turbulence, thus a higher K-value but also a higher P.
12 C
5.5 C 6 C
12 C
F. Another method to obtain a thermally longer channel is to decrease the channel depth. In this case, the depth is less which creates increased turbulence thus higher K-value and but also higher P.
Fig. 16. Summary of methods on how to adapt the PHE to the duty.
44
2. Optimization.
pinch, which occurs when the vapour has been desuperheated and condensation starts. Unfortunately, the temperature pinch is more difficult to calculate than the end temperature difference. It might happen that a condenser is laid out with a pinch, i.e. the cold side is equal to or higher than the warm side, an impossible programme. Apart from the difficulty of determining the minimum flow, the area-flow looks about the same as for a liquid-liquid HE, e.g. as shown in Fig. 16.
3. SECOOL optimization.
3.1. What is SECOOL optimization.
A typical SECOOL system as described in 15 and figures 17 & 18 of 1. Applications and the simplified system shown in figure 19 in this chapter consists of pumps, pipes, valves, filters, fittings, etc. and the PHEs. We will here discuss how to design an optimal system, because the technique of this demonstrates the properties of PHEs and can be used not only for typical SECOOL systems but also for all system where a circulating liquid is cooled or heated in a PHE. The more spent on the system (PHE size, larger pipe & valve diameters, etc.) the lower and cheaper the pumping energy will be. A total optimization usually shows that the lowest total annual cost (investment and operating) occurs for high investment and low operating costs. The usual approach when making an investment is normally to keep the initial costs down, i.e. high running costs and low investment costs. SECOOL optimization circumvents the question of capital vs. operating costs and instead tries to minimize the investment cost, notably of the PHEs, for a given operating cost, whatever this operating cost might be. Consider the system in figure 19. Suppose that everything is decided - pipe length & diameter, filters, valves etc. except the sea water flow and thus the PHEs size. The larger the flow, the larger the MTD and thus probably the smaller the PHE will be. The primary side flow, whether it is a fresh water flow as in figure 19 here and figure 17 or a condensing vapour figure 18 both in 1. Applications, is usually considered to be fixed but sometimes a pressure drop limit is imposed. One more component has to be decided, the sea water pumps. We will consider two cases: The pumps already exist and we have the pump characteristic curve. Thus, we have to select a point on the pump curve. Only the pumping cost is decided. This depends on the flow times the total head, but these are not yet specified and we are free chose any combination of flow and total head as long as the pumping energy is constant. Moreover, we assume that the sea water is free of charge and that the only cost is the pumping cost. All other costs might be considered constant are thus not applicable.
2.18.1. Evaporators.
The refrigerant flow cannot be changed, only the warm side flow. It would be possible to change the water flow by keeping the inlet temperature constant and change the exit temperature. That poses some problems, though, as the exit temperature is an important value in a refrigeration system, which cannot readily be changed. Change of the entrance temperature while maintaining the exit temperature is meaningless. The duty cannot be changed in an air-conditioning system from 1 l/s water from 12 to 7 C to 0.1 l/s water 57 to 7 C. Thus, we have to take another approach. In most systems, the water leaving the evaporator cools air in some type of unit cooler. In a unit cooler, the flow is essentially in cross flow and the temperature difference between the air and the water is fairly large. In such systems, the most important value is the mean temperature of the water. The unit cooler will more or less perform in the same way, regardless of whether the water temperatures change 7 -> 12 C, 8 -> 11 C or 6 -> 13 C. Fig. 17 shows the temperature program. In the normal air conditioning temperature program, 12 to 7 C, the mean temperature is 9.5 C. The minimum flow occurs for a temperature program of 17 to 2 C, still with a mean temperature of 9.5 C, in an infinitely large evaporator. The maximum (infinite) flow, also gives an infinite large evaporator. The inlet and the exit temperatures are then equal to the mean temperature 9.5 C. The area-flow curve, when water changes from the minimum flow to infinity, looks very much the same as for a condenser or a liquid-liquid heat exchanger, see Fig. 16. Note! If the water is used to cool another liquid in a counter current HE, the exit temperature from the evaporator is the most important parameter. See later.
2.18.2. Condensers.
Condensers behave very much like one-phase PHEs, with some crucial differences, see Fig. 18. The figure shows the temperature programme in a condenser. Two cooling water temperatures are entered, one for an infinite flow and the other with a temperature pinch. It is only meaningful to study variation of the cooling fluid. The refrigerant flow cannot normally be changed. In a one-phase heat exchanger, there is a minimum flow, obtained when the exit temperature of the cold fluid approaches the inlet temperature of the warm. In the same way, there is a minimum water flow in a condenser. However, this minimum flow is not determined by the end temperature, but by the temperature
2. Optimization.
45
Fig. 17. Changing water flow around a constant mean temperature in an evaporator.
PPHE
PPipes etc.
PStatic
HPump
Fig. 19. The main elements in a SECOOL system. 3.2. Existing SW pumps.
The only equipment not yet decided is the PHE and the only operating parameters not yet decided are the SW flow and the PHE pressure drop. In order to obtain the best PHE we make a number of calculations for different SW flows. For each SW flow the pump gives a total lifting head, which can be found from figure 20. Here the various pressure drops in the system are also shown. Note in consistent units, either as height or pressure difference. The pump head. This is the driving force. The total head usually decreases with increasing flow rate. Somewhere in the middle of the curve is the point of maximum efficiency for the pump, i.e. lowest electricity consumption. Changing the impeller diameter and/or the rotational speed changes the characteristic curve. The static head. This is the total lift of the water from the pump inlet to the summit (p in Fig. 19). This does not change with the flow rate. Pipes, valves, filters, etc. The pressure drop for all parts of the system except the PHE. It normally varies as a little less than the square of the flow rate. The PHE. The PHE gets what remains when the static head and the pipe work pressure drop have been subtracted from the total pump head. Values for two SW flows are shown in the figure. They vary from a high value at zero flow to zero, when all the pump head is consumed by the pipe work and the static lift.
SW flow
Example 1:
Capacity 1000 kW, 10 % margin on the K-values R22 superheated to 73 C, condensing at 40 C Water 70 kg/s 32 to 35.4 C. From the pump curve (typical): Pump head at 0 kg/s: 40 kg/s: 200 kg/s: At nominal flow, 70 kg/s Head (Calculated) Static lift: Pipes Left for the PHE: 270 k pa 100 kPa 135 kPa 35 kPa 320 kPa 300 kPa 0 kPa
46
2. Optimization.
A design for this duty gives: A15BWFD 182 plates, 135 m2 DP1/DP2 = 0.73/34.6 kPa Margin: 10 + 20 % What are the minimum and maximum flows? The minimum flow occurs when the water curve just touches the vapour curve just where superheating is finished and condensation starts, see figure 17. From thermodynamic data, we note that the latent heat in this case is about 85 % of the total load. The same is then valid for the water side. At the temperature pinch at 40 C, 82 % of the water heat is transferred. Thus Total temperature change: (40 - 32) / 0.82 = 9.76 K Minimum flow rate: 70 * (3.4 / 9.76) = 24.4 kg/s When the flow rate increases, so does the pressure drop in the pipe work, roughly to the power of 1.8. The pump head decreases at the same time. At a flow rate of 77.3 kg/s we get: Pump head (interpolated) Pipe work: 135 * (77.3/70)1.8 = Static head: 261.5 kPa. 161.5 kPa 100.0 kPa
Example 2.
We calculate the various pressure drops as before. The result is shown in table 6. The operating point has thus moved from 70 kg/s & 270 kPa to 48 kg/s and 393.7 kPa. Considering all the pumps available, it should be possible to find one pump with efficiency as high at 48/393.7 as at 70/270.
3.4. Summary.
Note 1! The pressure drop on the condensing side increases when the number of plates is reduced from 182 to 102 (or 116 as in the previous example). In this case, it does not enter into the optimization as the exit pressure from the condenser has a very little effect on the function of the expansion valve. This will hardly notice that the inlet pressure has decreased from 15.4 to 15.2 bar. The decrease of the condensation temperature and the effect on the mean temperature difference is taken into consideration by the PHE design program. Note 2! In case of a liquid on side one, the pressure drop of should be taken into consideration when optimizing. This is best done by limiting the pressure drop, if necessary some optimizations with different levels. Note 3! The result above is astonishing. Few engineers would object to a design specification for heat exchangers, which says 70 kg/s and 35 kPa pressure drop. Most would object to the high pressure drop in a design specification 48 kg/s and 225 kPa even though the total pumping power is the same in the two cases. Note 4! If the same optimization were made for a S&THE, the result would be radically different, especially for liquid/liquid duties where close temperature approaches are important. In a S&THE of the most common type - 2/1 - it is difficult to heat the cold fluid to above the exit temperature of the warm fluid; the efficiency drops rapidly, with increasing temperatures. On the other hand a S&THE can be designed to accommodate large volumes at comparatively low pressure drop. Temperature programs like 1. FW 50 - 40 C SW 40 30 C 2. FW SW 50 - 35 C 45 - 30 C
Total static & pipe pressure drop 161.4 kPa = Pump head For these two flow rates the surface approaches infinity but for different reasons, at the lower because there is no temperature difference left, at the higher because there is no pressure drop left, see figure 21. A number of trials is needed to obtain the minimum area, see table 5. Conclusion: By changing the operating point we have managed to reduce the size of the PHE. The water flow has decreased and the allowable pressure drop increased, but the important result is that this is done using the same pump, therefore requiring no change in the pumping cost. Note that this is not the only possibility for optimization. At the nominal flow, the condenser is thermally overdimensioned. That might be used to decrease the condensing temperature and thus the compression cost. However, a change in the compressor involves the whole system and might thus be more difficult to carry out.
would be no problem for a PHE but a S&THE will get into difficulties, especially the second which would be close to impossible to obtain it in one S&THE; it has to be many connected in series. If we try to optimize the duties above according to the principles, we find that: A S&THE would benefit from as large a temperature difference as possible, especially important is to get the exit temperature of the cold media below the exit temperature of the warm, hence large flows but low pressure drops. A PHE has normally no problem with large heat recovery, but has problem keeping the pressure drop down for large volume flows, thus small flows and large pressure drops are optimal.
2. Optimization.
47
Head
Area
The minimum seems to be somewhere between 40 and 50 kg/s. A closer calculation gives the minimum. 43 100 56.2 297.6 141.4 116 Optimal
The minimum seems to be somewhere between 40 and 50 kg/s. A closer calculation gives the minimum. 48 100 68.4 393.7 225.3 102 Optimal
The first has less, expensive Ti surface than the second. This and the fact that the small water/water M10B is considerably easier to open than the large A15BW condenser, strongly favour an indirect condenser system.
Result:
1000 kW, MTD = 7.87 K, (Area/MTD) * 2.0 = 12.46 R22 76 -> 43 C A15BW 49.5 m2 SW 37.8 <- 32 C, 144 kPa Titanium Dir. B. Evaporator. 809 kW, MTD = 4.06 K, Area/MTD = 24.4 Glycol 9.2 -> 4.3 C, 38 kPa A15BW 99 m2 R22 5 <- 0 C Stainless steel Dir. C. Water chiller. 809 kW, MTD = 2.75 K, Area/MTD = 24.9 FW 12 -> 7 C, 44 kPa M10B, 68.64 m2 Glycol 9.2 < 4.3 C, 48 kPa Stainless steel The area/MTD are now close to each other for the two last designs. The first is now too small, indicating that a temperature increase of 3 K was too much. We get the following relative costs before and after the optimization. Before: After: 2.0 * 85.5 + 87 + 39.6 = 2.0 * 49.5 + 99 + 68.64 = 297.6 266.64
We could decrease the total cost further by increasing the temperature level a little less than 3 K. On the other hand, as it is such a large advantage to use pure water instead of glycol we accept the cost. In any case, the total cost is lower than before the change.
2. Optimization.
49
3) Select a liquid cooler. Alfa Laval has a model LCS 219, which gives 219 kW at standard conditions (15 K difference between the air and the liquid inlet temperatures and a temperature drop of the liquid of 5 K). In our case, that means 41140 kg/h 40 to 35 C (215 kW). This gives a pressure drop of about 42 kPa. This fits fine, thus we use that. 4) The actual pipe work (and valves, etc.) P = 55 kPa. 5) A calculation of a BPHE gives a: CB300-50M, 14. 44 m2, Margin = 0 %. R22, 78 -> 45 C Glycol 40 <35 C, P2 = 31 kPa 2. Comments: The design is acceptable. There is no margin - not necessary because of clean water - on the BPHE, some on the LC. The only way this design is not optimized is that the available pump head is not utilized. We have three ways of optimizing this circuit. 1) The pump head could be lowered, e.g. by changing the impeller, rotational speed, or another pump. 2) The pipe work could be changed, reduced diameter, smaller valves or the like. 3) We could play with the rate and temperature level of the circulating water and change either or both of the LC and the CB. See fig. 24 for an optimization curve. 4) The total pressure drop determines the flow rate we intend to spend on the circuit. The pressure drop has to be redistributed between the unit until it is fully utilized, which means some trials and errors. Note that the optimum usually follows the rule of thumb. 3. Recalculation: We will here see if we can get a smaller BPHE or a smaller LC. The fact that there are no large excessive margins and that the pressure drops are underutilised too by about the same degree, except in the pipe work, indicates that the design is close to optimal. There is no point in starting a series of designs from a water temperature level close to the LC to close to the CB. We will then study two cases 1) We keep the CB and use a smaller LC and see what it means. or 2) We keep the LC and use a smaller CB and see what it means. 4. Change of the LC. The next smaller model LCS 190 has a nominal capacity of 190 kW. 1) To squeeze out 215 kW we need to increase the MTD about two degrees. 2) That on the other hand means a much smaller MTD in the CB, possible but with the use of H-plates. 3) H-plates have a higher pressure drop than M-plates. We then reduce the flow rate by accepting 6 K temperature change instead of 5 K.
Example 3:
A 215 kW R22 condenser has to be cooled by air of 25 C via a closed loop liquid cooler. The circulated water contains 30 % ethylene glycol. The condensing temperature is 45 C and the entrance temperature is 78 C. The circulation pump gives a head of 19.5 m at a flow of 40 m3/h (E.glycol has a density of 1035 kg/m3). 1. The following steps are followed: 1) Recalculate the pump data to kg/h and kPa to get consistent data, i.e. 41 400 kg/h and 198 kPa. 2) This head is distributed as: LC: Condenser: Pipe work: Spare: 60 kPa 60 kPa 60 kPa 18 kPa
50
2. Optimization.
4) The lower flow rate reduces the pressure drop in the pipe work and LC and increases the head. 5) The LC has a correction table, which corrects the nominal capacity to other temperature programs. The correction factor should be F = 190/215 = 0.884. To get this correction factor for a temperature difference of 6 K, we need a temperature difference between the inlet air and inlet liquid of about 17.15 K (interpolated). 6) The circulation rate then decreases to 34283 kg/h. That means a recalculation of all the pressure drops. Pump head increases to: Pipe work decreases to: LC decreases to: Spare, 10 % of total: Left for the CB: 231 kPa (Pump curve) 55 * (5/6)1.8 = 41 kPa 37 kPa 23 kPa 130 kPa
Normally it is difficult to get hold of thermal data for the LC such as exact temperature difference and the application of the rule of thumb is then more tedious. Moreover, the rule of thumb works best when there is an almost continuous change in size as in a PHE. A LC comes in definite sizes and the difference could be considerably between a change upward and downward. Fortunately, that makes the choice easier. The next lower LC size is LCS 166, which requires almost 20 K temperature difference, i.e. there is no place for a water circuit. A decrease of the CB too much lower than 34 plates means a very low flow rate and/or a high pressure drop. The recalculated variants are both better than the original. Which is the better depends on the actual prices. As the LC optimization utilizes the total pump head better, at least from a purely technical point this is the best.
7) A calculation of a CB300 for the new duty gives: CB300-50H, 14. 44 m2, Margin = 0 %, MTD = 5.4 C R22, 78 -> 45 C Glycol 42.15 <- 36.15 C, P2 = 129 kPa 5. Comments. It fits more or less exactly to the duty and all the available head has been utilized. The pump operates at a somewhat changed point on the characteristic curve. In the next section, we will see if we can decrease the CB by keeping the LCS 219. 6. Change of the CB. To decrease the area of the CB300 we need to: 1) Change to H-plates to increase the K-value. 2) Reduce the flow otherwise; the pressure drop will be far too large. We need to reduce the flow more than in the previous recalculation. 3) The correction factor has to be 219/215 = 1.0186. Try with an 8 K temperature difference. We obtain this correction factor for a temperature difference of 16.13 4) The circulation rate then decreases to 25712 kg/h. That means a recalculation of all the pressure drops. Pump head increases to: Pipe work decreases to: LC decreases to: Spare, 10 % of total: Left for the CB: 258 kPa (Pump curve) 55 * (5/8)1.8 = 24 kPa 36 kPa 26 kPa 172 kPa
5) A calculation of a CB300 for the new duty gives: CB300-34H, 8.96 m2, Margin = 0 %, MTD = 6.7 C R22, 78 -> 45 C Glycol 41.27 <- 33.27 C, P2 = 159 kPa 7. Comments. Once more, the optimum for a PHE occurs for a fairly low flow rate and high pressure drop. A comparison of the results gives. LC Before LC Opt. CB Opt. LCS 219 LCS 190 LCS 219 BPHE CB300-50 CB300-50 CB300-34 Pump 41140 kg/h 198/128 kPa 34283 kg/h 231/207 kPa 25712 kg/h 258/219 kPa
2. Optimization.
Liquid cooler
B3. Evaporator - UC (1 C)
+1 C
+4 C
B2. Evaporator - UC (4 C)
B1. Evaporator - UC (12 C) -4 C -8 C 12 C Refrigerant, I Refrigerant, II Cooling water Brine, Tank - UC Brine, Tank - Evaporator
Brine tank
51
2. Optimization.
Note 1. There is no possibility to obtain a humidity of 80 % or more in c) as this would require a mean temperature of +4 C and a temperature program of -8 to +16 C. Note 2. The capacities of the UC differ slightly from the required but as the available capacity, 160 kW, is higher than the requested, 156 kW, it has no importance. Can we optimize anything? If the mean temperatures and the entrance temperature are fixed, the flow rates are also decided. It would be possible to change the entrance temperature, but a reduction would mean still more difficulties in maintaining the humidity in c) (if that is required) and increased frost formation. Increasing the temperature might be possible, but if the mean temperature of a) must be kept it means a very rapid increase of the circulation rate. The tank-UC loop is thus set to -8 C -> -1.9 C. What about the tank-evaporator loop -4 C -> -8 C? This temperature programme assumes that the circulating amount of glycol is larger than the tank-UC loop. A part of the cooled glycol then mixes with the returning glycol from the UC and form the entering -4 C glycol. A design of a CB gives: CB300 62M, 16.80 m2. Margin = 5 %, MTD = 5.2 K E. glycol 39215 kg/hr -4 -> -8 C. P = 25 kPa R22 -10 <- -15 C. A design for various glycol flow rates and corresponding inlet temperatures reveals that the area-flow curve is extremely flat, i.e. the present design is close to an optimum. A check for the loop temperature programme gives CB300 58M, 15.68 m2. Margin = 5 %, MTD = 6.1 K E. glycol 23762 kg/hr -1.4 -> -8 C. P = 12.3 kPa R22 -8 <- -15 C. There will be a saving in the pumping costs. However, it is advisable to keep the pressure drop fairly high, otherwise, there might be maldistribution of glycol.
Example 4:
Set-up. a) Meat & fish: +2 C/80 % Humi. 71 kW 47 kW 38 kW 160 kW
b) Fruits & vegetables: + 4 C/80 % Humi. c) Preparation: d) Totally available +12 C/70 % Humi.
The unit coolers for all three spaces are supplied with 30 % ethylene glycol of - 8 C. The return temperature to the tank is assumed to be -4 C. This is also the temperature program for the tank-evaporator loop, but as this has an individual pump, the flow rate and the temperatures can be adjusted. Frost formation. The glycol enters well below zero. Inevitably, frost will form on the fins. The preparation room coolers could conceivably be fed with glycol at above zero, the others not, but that would mean a double evaporator system, which is expensive. Thus, a system of defrosting has to be installed. Humidity. In a cooled, sealed off space the long-term humidity is set by ratio of water vapour pressure at the lowest temperature the air encounters and the vapour pressure at the bulk temperature. External factors have an influence, e.g. if there is a continuous influx of moist air, the relative humidity increases. In most cases, there is thus a limitation on the largest temperature difference possible in order to keep the humidity on a certain level. This looks something like: Temp. diff. C Humidity % 5 93 6 89 7 85 8 82 9 79 10 76 11 73
In the case of a liquid cooled UC, there is an extra complication that the temperature is not constant but varies from the inlet to the exit. In this example, we will assume that the mean temperature of the glycol can be assumed. In the above examples we have assumed at least 80 % humidity in a) and b) and 70 in c), corresponding to temperature differences of 8, 8 and 12 K. The temperature programs thus looks like: a) Glycol 3*5599 l/hr -8 to -4 C, b) Glycol 1*5520 l/hr -8 to 0 C, c) Glycol 2*1916 l/hr -8 to 8 C, d) Glycol 26149 l/hr -8 to -1.4 C 3* IRL 555, 3*24 1*IRL 657, 1*46 2*IRL 451, 2*19 156 kW kW kW kW
A decrease of the evaporation temperature to -16 C and a decrease of the condensation temperature to 43.7 C Cond: Evap: CB300-94H, CB300-56M, 4.4 K. 5.7K 21.4 9.82
There is thus a very sharp increase of the condensing surface with decreasing condensing temperature. The optimum probably lies very close to the first trial. Note! The rule of thumb has to be used with judgement as two-phase duties can behave somewhat erratic for high pressure drop, especially at evaporation and when approaching the temperature pinch (Fig.17) at condensation.
2. Optimization.
53
Cost
Sum
TR 22 Tmean TR 22 TAir
TR 22 Tmean
MTDCB TAir
MTDLC
Fig. 24. Optimization of two heat exchangers interconnected by a circulating liquid stream.
54
55
1A
1B
1C
A stainless steel sheet and a copper foil are pressed together in a hydraulic press (1A). A number of plates with their foils are then stacked with the corrugation angles in opposite directions (1B & C). The plate pack is
then placed in a vacuum oven and heated to the melting point of the copper. Because of the surface tension, the liquid copper collects at the edges and the contact points, effectively forming sealed off channels.
End channels Channel plate: The piece of corrugated sheet metal making up the heat transfer surface. Channel: The spaces between two channel plates. Thus: One channel can be made of equal or different channel plates.
Refrigerant Water
Last channel plate Back cover plate Adapter plates (not always) Mounting bolts Porthole First channel plate Four refrigerant (internal) channels surrounded by five water (external) channels. Seal plate, not corrugated Front cover plate Adapter plate (not always)
56
2.3. Compactness.
The BPHE weighs about 25 % less and occupies less than 25 % of the volume of its closest rival, the coaxial. To be fair to the coaxial, it has to be said that sometimes the space inside the helix wound shell can be used, e.g. as a liquid receiver.
The burst pressure of a BPHE is more than 150 bar, and most pressure vessel codes require a burst pressure of five times the design pressure. The above limits are general, for specific data see the product catalogue. The design pressure for nickel-brazed units is lower. Non-refrigerant BPHEs are sometimes used for high temperature, low-pressure service and are accordingly rated for a higher temperature, but a lower pressure.
2. Properties.
2.1. Flow regime.
The channels formed by the "herring bone" corrugation will force the liquid in a helix like flow path - a type of forced turbulence. Actually there is no sharp transformation between laminar and turbulent flow. This gives the PHE much higher heat transfer coefficients than normal S&THEs, especially for viscous flow.
2.2. Control.
The practical effect of the flow characteristics is that thermal performance does not exhibit the typical jumps or hysteresis effects which can occur when the flow changes between laminar and turbulent flow in a tubular HE, making control difficult or impossible.
2.7. Fouling.
This is a complex subject and is treated in detail in the section about fouling and corrosion. Suffice to say here that the absence of dead spots and high shear forces makes the BPHE very resistant to scaling. It is somewhat sensitive to clogging by fibre-like impurities in the water, though, but less so to sand like impurities.
57
Embossments
Washers
The channel plates have the edges folded towards the end cover plate.
Washers
Note! This method cannot be used with the special evaporator plates. Use connections at the back cover plate instead.
Embossments at positions S3,4 and T3,4 indicate a refrigeration unit with the refrigerant side at position 3 & 4. Embossments at S1,2 & T3,4 or S3,4 & T1,2, indicate a normal unit and equal number of channels on each side. 2. In most cases it is impossible to see the location of the washers from the outside. The only possibility then is a careful measurement of the porthole depth. 3. Some old models have closed end channels and thus an even number of plates. If the side with the shorter porthole has washers at the inlets, it is a refrigeration unit.
58
3. Flow arrangement.
3.1. The one-pass BPHE.
3.1.1. Evaporator.
The evaporating fluid is normally flows upwards and the heating media flows downwards. Downwards evaporation is not impossible but needs a comparably high channel and low nozzle pressure drop to distribute the liquid properly. A high nozzle pressure drop means that the fluid will have a maldistribution from the first to the last channel. A low channel pressure drop means that the liquid will not distribute properly over the channel width. Downwards evaporation in BPHE has been little studied and tests should be made before an installation is made. Expect a fairly large capacity reduction compared to upwards evaporation. A direct expansion evaporator with an entrance nozzle that is too large, resulting in low velocity, could cause the liquid and vapour to separate and make the evaporator difficult to operate. This is especially severe in downward flow. An evaporator with a restriction at the channel inlet distributes the refrigerant more evenly, both over the plate pack length and the channel width. A DX evaporator has the flow normally in counter current as this gives the highest MTD. A flooded evaporator normally flows in cocurrent. The reason for this is the resulting large temperature difference at the refrigerant inlet, which facilitates the start of the boiling. Note! In case of a large (> 10 K) temperature differences at the refrigerant inlet, there might be too much vaporization at the very first stage of the heating surface. Especially at the startup, when the evaporator and the exit pipe are filled with liquid, the pressure drop for this vapour is larger for the flow path through the evaporator to the separator than for the backwards path through the feed pipe to the separator. The flow might then reverse or oscillate. See also the section on evaporation.
loss. The lowest pressure drop, when the flow is on its lower control limit, should in this case be at least as great as the liquid static head. If the nozzle pressure drop is too large a part of the total pressure drop (> 25 %), the flow will distribute unevenly from the first to the last channel, causing a capacity loss. The solution is a BPHE model with larger nozzles or double inlet nozzles. If the pressure drop is a large part of the total pressure, there is a danger of sonic booms. This is, however, seldom the case in a refrigeration condenser. Counter current flow is normal.
59
No drain
No drain
No drain
60
and draining. As more than three passes are very uncommon in refrigeration, venting & draining is no problem. In normal operation, letting the liquid have a pressure drop per pass as least as large as the static height difference prevents formation of gas bubbles.
3.2.3. Applications.
The main reasons for using a multi-pass BPHE are: The nozzle positions. A two-pass arrangement permits an inline connection to the pipes; space in a chiller unit only permits certain nozzle positions, etc. Antifreeze or antifouling measures. If a BPHE changes - i.e. a new one - from one to two passes on the water side, the channel flow rate doubles. The number of passes on the refrigerant side remains constant. This has three effects: The wall temperature on the water side in an evaporator increases (for a more detailed study, see chapter 7. FREEZE), i.e. decreased freezing danger. (The wall temperature increases if the number of refrigerant passes increases, as refrigerant heat transfer is less velocity-dependent than that of water.) However, there could be problems of maldistribution of the evaporating refrigerant; see figure 9 B3. The wall shear stresses increase, i.e. fouling compounds have more difficulty adhering to the heating surface and the unit stays cleaner. The pressure drops increase about seven times. (The flow length and the channel velocity are now double.) Accommodate a long thermal duty. This is probably the major reason. An example will show why and how. Compare the two following duties: 1. Water Water 3 kg/s 50 -> 40 C or 3 kg/s 40 <- 30 C or 2. 1 kg/s 60 -> 30 C 1 kg/s 50 <- 20 C
Both the duties have the same capacity, the same temperature difference and the same fluids, thus the required area should be the same. Suppose we design a unit for the first duty. A-CB26-60H with an area 1.45 m2 and a pressure drop of 0.80 bar will do fine. However, when we try to use this for the second duty, problems arise. The flow rate is three times less, consequently a lower overall heat transfer coefficient and a larger heat transfer area is necessary. The only way we can achieve this is to increase the number of plates. This of course means a further reduction in channel velocity; a lower heat transfer coefficient and we end up with chasing our tails. The final result is two CB26 120H with 2 * 2.95 m2 area. Is there any possibility to reduce the area? There are two: We can switch over to a CB52. This has the same width as the CB26 but is twice as long. Thus, we can reduce the number of plates, which is fine for the lower flow rate. A CB52-50 with 2.4 m2 area is then required. A CB26 in two-pass. Effectively two plates are then connected in series and the result is similar to a CB50. A CB26-90, grouped as 23+22/2*22, i.e. two groups of 22 channels in parallel connected in series and with an area of 2.2 m2, is then necessary.
61
might become a little problematic. In A2 there is a temperature pinch in the turning point of the two-pass flow. This might become worse (or better) if the flow direction is changed on one of the sides as in A3. If the temperature difference increases, the pinch disappears, A4. Note that the flow direction is important.
R2 RL R1
RC
R R
C L L L1
L2
B
Fig. 08 B - C . Two circuits/One circuit.
B. Two refrigerant circuits and one water circuit is a common way of arranging a two circuit unit, but with some drawbacks, see 3.3.3. The arrangement is symmetrical i.e. both circuits have the same nominal capacity. It can be also used as a double circuit condenser. C. The condensate is subcooled in one circuit (RC/C) with part of the evaporating refrigerant, which superheats very much due to the large temperature difference. The subcooled condensate then expands, evaporates and cools the water in the other section (RL/L). The total superheat is controlled after the evaporator. Due to the high super heat in the (RC/C) section, the superheat in (RL/L) is low. This means a very efficient evaporator. Thus, where the heat transfer coefficient is low because of the high superheat (RC/C) the temperature diffe-
D
Fig. 08D. Two water/One 2-pass refrigerant circuit.
rence is large and contrary. Depending on the refrigerant type and the temperatures the arrangement can gain some 5 to 10 % in COP. As the leaving vapour is a mixture between high & low superheat vapours, a good mixing is necessary. An example of this is shown in 4. Evaporators and separators, figure 21, p. 95. See also figure 05, p. 11 in 1. Applications, regarding subcooling/superheating. D. This is a common arrangement for a desuperheater and a condenser. The refrigerants desuperheats in the first sections and heats tap water, then the condensing heat is dumped into cooling water. The disadvantage of two passes (8A & D) is the high pressure drop.. Compared to a single pass section with the same total number of plates this arrangement has about seven times higher pressure drop.
62
The results are similar, which is what could be expected. What happens if we try to use a CB50 or a CB26 in two passes for the first duty? We then have to increase the number of plates in order to reduce the pressure drop to an acceptable level, and we end up with either two CB26-69, 2 * 1.675 m2, connected as 2*17/2*17, or a CB50-70, 3.4 m2, connected as 35/34. Both are thermally overdimensioned, about 120 %. In the example above we have used water, but the result is similar if we use refrigerant, either condensing or evaporating, on one or both sides. By connecting the BPHE in two passes we obtain a HE which is twice as long as the one-pass HE. Better utilization of the available pressure drop. This is another way of expressing the previous paragraph.
exchanger, but it is entirely possible for the same number of passes on both sides. If, however, one flow has a small temperature change (approaching Fig. 8A1), it matters less if it is symmetrical or asymmetrical; e.g. for an evaporating or condensing liquid, asymmetrical flow could be interesting. Figure 9 B shows some possible ways of arranging the passes. Observe that it is not necessary to have the same number of plates per pass. The liquid side in multi-pass and the refrigerant in single pass are seldom used; see figure 9 B3. An unequal number of passes should only be used in two-phase flow, in order to accommodate a changing vapour fraction. An unequal number of channels on the liquid side are normally very uneconomical. Counter current vs. cocurrent flow. Thermal effectiveness prescribes counter current flow. The exceptions are: In flooded evaporators; see 3.1.1. In case of a very high pressure drop on the evaporating side, cocurrent flow could give a higher temperature difference. See figure 9 C. The practical consequences of the above are: All passes equal on both sides. One pass on the liquid side and unequal number of channels per pass on the refrigerant side. In some rare cases, e.g. refrigerant/refrigerant, there could be unequal number of channel per pass on both sides or one pass on the refrigerant and multi-pass on the water side (freezing danger).
63
64 3.3.2. Types.
There are some possibilities to make a PHE with two circuits on one side, three with any type of PHE and one with a special designed BPHE. 1) Normal, one-pass. Figure 8B shows a BPHE with two refrigerant circuits and a one-pass water circuit. 2) Normal, two-pass. Figure 10 shows a BPHE with two refrigerant circuits and a two-pass water circuit. Figure 8D is similarly arranged but reversed sides. 3) Dual circuit BPHE. Figure 11 shows a new design. 4) Mixed design, where there is refrigerant on both sides.
are then not equal, see figure 10. This is not necessarily a bad thing as it gives three capacity levels (Side 1 open, Side 2 open & both open) instead of only two. All the water always passes an active circuit, regardless of which is closed down, as opposed to the previous design. This is important to decrease the freezing danger. The design works well but it has one very important disadvantage; the pressure drop on the water side tends to be very high, thus only a few BPHE models can be used.
65
9 C 7 C 7 C
Circuit 2 49 kW
Circuit 2 12 C
Circuit 1 33 kW
7 C
tevap 2 C
This evaporator has two parallel water channels followed in series by another group of three channels, i.e. a total of five water channels. There are two refrigerant circuits, each with two channels.
Circuit 1
Flooded flow: If the liquid flow forms a U, the entering subcooled liquid meats the same liquid temperature in the two sections. This tends to equalize the nucleate boiling and the boiling starts simultaneously in the two sections. Thus: Liquid in/out at S1/T1 Condensers. The arrangement can be used for condensers as well, but with reversed flow directions on the refrigerant side. The water side connections can be either S1 & T1 or S2 & T2, depending on which gives the best tempera- ture difference. It is not very common, though. Refrigerant in at S3 (T3)
66
Beware of low p units placed on different heights. Flow Imbalance is all but inevitable.
Air purge
Air pocket To an open discharge Valve or a disk with a hole to increase the P can help balance the flows. Pflow
hfirst hlast
2. Loop
hfirst - Pflow = hlast In other words, the resistance to the water flow is so small that the water immediately enters and leaves from the first channels and doesnt fill the last channels. This usually happens only for very low pressure drops. Increasing the pressure dop would decrease hlast to a point lower than the channel exit, i.e. the air starts to leave the condenser
2. Drain by-pass
4. Free exit
The air pocket can be vented by the air purge. When the air has left and water enters the purge, the float rises and closes the vent. To facilitate the filling, there should be a restriction at the exit, symbolized by the valve. If the water is self draining, i.e. a free exit to the outside, the valve opens when the water leaves the purge, air enters and the draining is facilitated.
If the water does not fill the channels as described in figure 11 C, a loop as in 11 B2 might help. The BPHE is then not self-draining at off periods, but remains water filled. A small by pass line, from the exit connection to the pipe beyond the loop, allows for an automatic drain. There will be a continuous by-pass but that can be kept small. There might be an air cushion at the inlet port, which needs to be vented. Usually air exits with the water, but in the event of a very low flow rate, it might collect in the upper port. From this air can be vented either by: A purge valve at T1 (figure 11 C). This needs a certain pressure difference to the atmosphere to operate. A small by-pass line from T1 to the top of the loop. This does not actually vent the water side, it just moves the air further away, but the advantage is that it operates independently of the pressure difference to the atmosphere. The air can then be vented at a more suitable location or simply leave the system with the water. When the PHE has been vented, water bypasses the channels and enters the loop directly but this by-pass can be kept small. See figure 11 B1.
67
Brine
Ref. 4
Ref. 3
Ref. 1
Ref. 2
Ref. 1
Brine 1
Ref. 2
Brine
5. Instruments.
Thermometers. These can be placed in the extra connections on the water side. This has a drawback. The temperature of the water emerging from different channels can vary a little. Depending on the depth of the probe, the measured temperature varies somewhat. This location is acceptable for control purposes, where the error in the reading is taken into account when calibrating the system, but not for exact performance measurements of a plant. The thermometer should be located some distance from the BPHE, in insulated pipes. Refrigerant temperatures are often measured by the corresponding pressure. Refrigerant is often subcooled or su-
perheated, and if a performance measurement has to be made, the "true" temperatures have to be measured. Manometers. These are sensitive to turbulence. Manometers should if possible be placed downwards inside a straight pipe, at least 15 - 20 diameters long. This can be difficult to achieve in a refrigerant circuit, and errors of both pressure and temperature (see above) could then be introduced. It is important to be aware of this when evaluating the performance of a condenser or an evaporator. If possible, use differential manometers for pressure drop measurements as a small difference between two large absolute pressure readings could give large errors. Another possibility is to switch the placement of the manometers and take a second reading.
Brine 2
68
6. Installation.
6.1. Position.
A BPHE should generally be installed vertically. Figure 13. If a horizontal installation is necessary, there are some starting points to consider (figure 14): A horizontal installation should only be made after a thorough testing and evaluation, as the thermal and hydraulic properties of an inclined heat exchanger are less well documented and can vary considerably. Never install the unit on its back with the plates in a horizontal position (figure 14A). A horizontal installation as in Fig. 14B is preferable. Expect a capacity loss for an evaporator of about 25%, condenser unknown but probably less. A slightly inclined, 5 - 10 , unit (figures 14C-D) reduces the losses considerably. A condenser must be installed with the refrigerant nozzles on the lower part of the BPHE (figure 14C); otherwise, condensate will block the heating surface. This means that the water side cannot be drained. An evaporator could be installed with the refrigerant nozzles on the upper part. The water side can now be drained (figure 14C, but reversed connections).
an easily accessible place figure 13 C - for connection to an external water circuit. A possible leakage at the threads is then easy to detect and repair. Beware of liquid hammering. Note that valves, which close against the flow produce less hammering than the normally used valves, which close with the flow. Refrigerant side. These are normally made for soldering. In some cases, notably for nickel brazed units, welding might be necessary, figure 15C. Follow the local pressure vessel codes, safety regulations and refrigeration standards. Thoroughly clean and degrease the surfaces to be brazed with a solvent preferably a commercial solvent with instructions for disposal otherwise the flux does not wet the surface when brazing. Degreasing is normally not necessary when welding except for a highly corrosive atmosphere or medium. Degreasing prevents the increase of the carbon content in stainless steel. Carbon could then form carbide with the chromium in the steel during the welding, usually in a defined zone on each side of the weld. The chromiumdepleted steel is then no longer a stainless steel and is susceptible to corrosion. To prevent oxidation and to cool the unit, blow nitrogen through the pipe to be brazed or welded. Keep water running through the water side. The water should be running before the brazing or the welding starts and continues until the unit can be touched. Water-cooling is not strictly necessary when brazing, but it cools the unit faster and cheaper than nitrogen. Depending on the welding method, welding could release substantially more heat than brazing. TIG welding does not normally need any more protection than letting the water run. If more heat needs dissipating, the connection base should be protected by a wet rag or by flushing the body with water. It is easier to centre the connecting pipe and the brazing/welding is faster if the BPHE is place horizontally. Use a solder of at least 45 % Ag and braze at less than 650 C. Under no circumstance should the brazed part exceed 800 C. TIG welding and a protection gas is the least heat releasing welding method and should be used if possible.
6.2. Mounting.
See figure 13. Ensure that neither vibrations nor thermal expansions in the pipes are transferred to the BPHE. This could be prevented as follows: Use some rubber strip or Armaflex between the support and the BPHE. Use vibration dampener for the compressor. In case of long straight pipes, use bellows or some other expansion absorbing devices in the pipes.
6.3. Insulation.
A BPHE has a small exterior surface and can sometimes be installed without insulation. Should insulation be required, 5 - 15 mm Armaflex or similar is suitable. Pre-cut Armaflex sheets are available for most BPHE models. Units - both condensers and evaporators - installed outside in freezing ambient temperatures need insulation as well, see the chapter 7. Freeze. WPHE should be fitted with removable insulating sheets to allow easy dismantling for inspection and servicing of the plates and gaskets.
7. Thermal guarantees.
Alfa Laval can not take any responsibility for proving that a heat exchanger gives the capacity if there are no possibilities for proper measurements of temperatures, pressures, flows, etc. Thus in case of a claim, the customer has to prove that it does not give the capacity and not that Alfa Laval has to prove that it gives the capacity. The reason for this is simple. If the proof lies on the designer of the plant, he will equip it with sufficient measurement devices, a fairly cheap operation if foreseen at the design stage. Fitting measurement devices to an operating plant can be expensive, possible including shutdowns, emptying vessels, removing insulations, welding and subsequent controls, etc. Thus in the interest of the designer, operator and equipment vendor, a plant should be equipped with proper measurement devices at the beginning.
6.4. Connections.
Water side. These are normally threaded, figure 15 A. Keep the threads clean. Do not over tighten. (Recommended values are found in the product catalogue. A gasket between the pipe and nozzle seats and not at the threads should make the seal. Use Teflon or Molycote or similar products to protect and lubricate the threads. Threaded connections are acceptable if the connections are easily accessible from the outside. If the PHE is installed inside a package and maybe insulated, the PHE is better connected to the outside via brazed connection on the PHE and threaded connections on the pipes at
69
A bright line, pained diagonally along the plate pack, reveals a misplaced plate.
S4
S3
Flexible connection
S3 S2
A. Condenser
B. DX evaporator
C. Heat recovery
S4 S1
5 - 10
S3 S2
30
70
71
Fig. 01 B. Flow regimes for vertical, upward two-phase flow in a circular channel.
1. Subcooled boiling. The bubbles collapse when they enter the liquid bulk. 2. Bubble flow. A net vapour production is obtained. 3. Slug flow. The bubbles have merged into larger vapour slugs, which fill the channel cross section. 4. Churn flow. The liquid now flows along the walls, but the vapour has difficulty in carrying the liquid upward. The flow is unstable. 5. Annular flow. The vapour velocity is sufficient large to carry the liquid annulus upward. The heat transfer coefficient is here very large and the flow is stable. 6. Mist flow. The flow is droplets in a gas, not bubbles in a liquid, with the low heat transfer of a gas.
72
It can be mentioned that the refrigeration terminology is not very consequent. In general, industrial terminology the word evaporator is used, when the purpose is to separate a solute from a solvent by means of evaporation of the solvent. An example is sugar evaporators, found in sugar refineries. The word vaporiser is used, when the purpose is to vaporise a pure liquid, i.e. a condensed gas. Thus, the term vaporiser covers the function of a refrigeration evaporator better than the term evaporator.
2. Evaporator classifications.
2.1. Pool boiling units.
The heat transfer area - i.e. a number of channels, with the heating media flowing on the inside - is submerged in the liquid to be evaporated. The liquid is quiescent. The heat transfer is mainly nucleate boiling. The vapour separates from the liquid above the channels in an enlarged shell. Sometimes a vapour separator is fitted immediately above the evaporator. This type is mainly used in large industrial refrigeration plants. The advantage is stable operation for varying operating conditions. As PHE are not used as pool boiling units, this type is not discussed here.
73
Thermostatic expansion valve. Circulating refrigerant Liquid refrigerant from the condenser. Shut-off solenoid valve. Partly expanded refrigerant
Vaporized refrigerant to the compressor. Refrigerant from condenser Level controlled expansion valve
Circulating refrigerant
The pump is placed in a pit. This increases the absolute pressure and thus the boiling point. The risk of cavitation is then diminished.
74
the circulation, i.e. there is an interaction and interdependence between pressure drop, circulation rate and heat transfer in the circulating system. By definition thermosiphon means circulation owing to density differences between the fluids in two connecting legs, one hot and one cold. Assume that the unit in figure 03 B is not operating but filled with liquid refrigerant. Both the valves are open. The refrigerant level in the separator is the same as in the evaporator. When liquid enters the evaporator on the other side, the refrigerant heats up, slowly bubbles start to form and the channels are partly filled with ascending bubbles. Thus, the mean density in the leg formed by the evaporator is much lower than in the leg formed by the separator and the descending pipe. The two legs are thus not in balance and the refrigerant gradually enters the evaporator from the separator via the descending pipe. At the top of the evaporator, a two-phase mixture is pushed into the separator and the liquid and vapour separate. At the bottom, the entering refrigerant heats up and finally starts to boil. Thus, a two-phase mixture is always maintained in the channels. As the circulation rate increases, the various pressure drops increase and finally the driving force is balanced by the retarding forces. See figure 03. The system is now in balance and a constant refrigerant flow enters the evaporator while a certain fraction of the flow evaporates. When the liquid-vapour mixture enters the separator, it is saturated. The liquid separates and the refrigerant enters the evaporator again but it is now not saturated. The temperature is the same as in the separator, but the pressure is higher, increased with the static head from the liquid level to the inlet, i.e. the refrigerant is subcooled. This means that at the first part of the heat exchanger there will be no boiling, just a temperature increase. However, as the refrigerant rises, it will decrease in pressure, reducing the subcooling. These two effects-increasing temperature and decreasing pressure - mean that after a while boiling point is reached and the refrigerant starts to boil, albeit at a higher temperature than at the exit. The pressure continues to decrease because of the changing height and the pressure drop, and the refrigerant, now saturated, will continue ascending with decreasing temperature until it reaches the separator and the loop is closed. Figure 03 B shows how the temperature changes from the inlet to the exit of the evaporator. Observe that there is a small temperature drop of the refrigerant, because of the pressure drop in the exit piping. This temperature drop is not accompanied by a heat transfer as this is an adiabatic (i.e. no heat exchange with the surroundings) expansion of the refrigerant mixture. Because of this temperature drop, the temperature at the exit of the heat exchanger is slightly higher than at the inlet. At the lowest point in an ammonia system there is an oil drain. See 8. Oils & Refrigerants for further descriptions.
3.2. Applications.
The relatively small size of the flooded flow evaporator compared to the direct expansion evaporator favours this type at large capacities. Here the price of the evaporator is more important than the extra cost of a separator. The smaller BPHEs are thus normally not used as flooded flow evaporators. They are sometimes used as auxiliary coolers of oil, water or refrigerant in large plants. Larger BPHEs can be used as flooded flow evaporators for nonammonia refrigerants. Ammonia, for reasons later described, favours the use of flooded flow evaporators and the recent development of nickel-brazed PHEs could increase their use here.
3.3. Thermosiphons.
This is the most economical solution for a flooded flow evaporator as the cost for a pump and its operation is saved. The circulation rate is, however, very dependent on the heat transfer and pressure drops in the various parts of the system, which in their turn are dependent on
75
Shut-off valve
Vapour-Liquid separator
Churn flow
Nucleate boiling
Refrigerant enters the evaporator Oil return to the compressor Plate Heat Exchanger
Liquid
Liquid
76
If the ratio between the entrance and the exit temperature differences becomes too large though (more than double), it might be better to maintain counter current, as there might be a risk of instabilities, see 3.5.
77
Temperature
Temperature
3
Starting temperature difference at cocurrent flow.
2
Evaporation temperature
4 1
Unsymmetrical
78
This causes a decrease in vaporisation and pressure drop and the two-phase mixture can leave through the exit. Refrigerant may again enter the evaporator and the cycle is repeated. This problem is more serious than the first as it is oscillatory. Apart from causing control problems, the capacity decreases, as the evaporator is ineffective part of the time. The problem can be controlled by a valve at the refrigerant inlet. The problem is compounded if the inlet temperature difference is very large and the exit temperature difference small. This can happen in cocurrent flow (see figure 04). The vaporisation rate can then be larger at the first part of the channels than at the end and the vapour finds it easier to escape backward. In such a case, counter current could be better. Backward flow can be prevented by a control valve as above.
n) Does an ammonia system have efficient oil recovery? (1. Applications, fig. 10.) Otherwise, oil could cover the heating surface and decrease the heat transfer. o) Does a system with soluble oil have an oil evaporator? (1. Applications, fig. 19). If so oil will not cover the surface but the physical properties will be impaired (higher viscosity) and the heat transfer decreases. p) Measure the temperature (if possible the real temperature, and the corresponding pressure) at the evaporator exit, the separator exit and at the suction pressure. Too large differences indicate some flow restriction. q) Beware of refrigerant trapped in a downward loop. r) Is there water in an ammonia system? For a given pressure, addition of water to the ammonia increases the evaporation temperature. As the control system is mainly based on pressure, there could be a hidden decrease of the temperature difference. s) Inspect the refrigerant loop for flow obstructions. Are there excessive numbers of valves, valves with to high P, bends, etc.? Especially dangerous are obstructions at the vapour exit. How does the vapour inlet to the separator look? Obstructions? t) A superheated vapour could hint at: Water in an ammonia evaporator, see r) above. A restriction in the refrigerant flow to the evaporator, see s) above, means that the evaporation capacity is not fully utilised and the vapour superheats. Note that a superheated vapour means an underutilised evaporator, i.e. more refrigerant could vaporise if more enters. This is similar to a too cold condense from a condenser. In both cases, the problem is to be found outside the PHE. u) If the operation of the evaporator-separator system is very erratic, difficult to control or shows other irregularities, check the float valve. A float valve works best if moving with a certain velocity. A slow moving valve, especially after some time operation, can get stuck. An additional problem is that the float can implode, if the outside pressure becomes larger than the pressure inside the float.
79
4. Vapour-liquid separators.
4.1. The purpose of this section.
This section describes the function of the vapour liquid separator for a flooded - either pump or natural circulation - evaporator system. It is valid for: Refrigerants with insoluble oil, heavier than the refrigerant, i.e. in practice the ammonia-oil system. All refrigerants with soluble oils. The main stress is laid on the interaction of the separator with Plate Heat Exchangers (PHE). The description is by no mean complete. Security & control equipment, the mechanical design of the vessels, etc., are not covered. These depend on local refrigeration & pressure vessel codes, type, size & placement of the refrigeration plant, etc., and are thus beyond the scope of this manual. In (Appendix II. Design of vapour liquid separators.) is a review of some design methods for separators.
4. 3. Operation.
The separation of the vapour and the liquid is obtained by gravitational forces, sometimes assisted by centrifugal forces, which allow the heavier liquid droplets to settle. Thus, liquid droplets small enough to be kept in suspension by molecular movements (Brownian movements) do not separate. In practice, liquid droplets sometimes much larger than Brownian droplets do no separate either, but there are additional methods to separate these.
4. 4. Equipment of a separator.
A separator should be equipped with (see figures 06 - 08): Demister (for scroll & piston compressors). Sight glasses (one for viewing, one for a light) at the inlet and exit of the separator (from/to the evaporator). This is especially important for monitoring the operation of an injection recirculation evaporator. Connections for a thermometer and a manometer.
80
As long as possible
4 9 6 C 6 6 2
1 3
3 B
9 9 2
9 9 5
A. Utilize the total length as much as possible for the separation. B. Splash protecting plates (9) help keeping the liquid
body calm, especially at tangential inlets. C. Two sight glasses or even better three (6), for a light and for viewing, provides a cheap method to check the operation of the separator.
4 E 3 1 a D 5 2 h 3
G 1
D. An inclined vessel helps drain the oil in an ammonia system. However, note the danger of a vortex at the exit. E. The oil & evaporator return exits should be placed as far as possible from the vapour inlets. Fine oil droplets entering at 1 then have time to settle. See also H.
F. Avoid a high inlet as the static liquid column - h - and the pressure drop might be too high. G. Avoid injection of the flashing refrigerant into the liquid. It reduces mist and fine droplets but the agitated liquid gives an unsteady signal to the level controller.
7 4 K 1 2 2 2 5 J 8 1 L
K. A demister catches mist and fine droplets. L. Injection of the flashing liquid in the circulating loop is difficult. It could be made via an ejector pump (see figure 2 D) or via a distributor (a pipe with holes, one for each channel) into the PHE evaporator. In both cases, the expansion should be continuous (as in a TEV), not on-off as is normal in a flooded system.
H. Avoid vertical liquid refrigerant exits as the formation of vortexes is inevitable. If vapour is drawn into the evaporator the capacity might decrease and the pump in a forced circulation system might be destroyed. J. The upper and lower exits, 6 & 7, for the level controller (shown here as a pilot valve) should be spaced as far apart as possible. This facilitates changing of the operating liquid level should the need arise.
81
vapour exit A deflector plate turns the flow away from the vapour exit
C1 C2
Top view
A1
A2
B1
The return from the evaporator and the flashing vapour should be directed downward but prevented from hitting the surface by e.g. a horizontal exit. Horisontal exit to the evaporator
A1
A2
Oil drain
?
B2
Refrigerant to evaporator Heat from the ambient air, an electric heater or a heated coil. Compressor discharge vapour
B1 B3
B2
B4
The velocity is independent of the liquid level. The total refrigerant filling is less than the horizontal.
Fig. 07. The hybrid separator for ammonia with an oil recovery system.
Oil comes from the separator via the line & valve B1 and from the lowest point in the loop (see also Applications figures 03 & 10) and collects in an oil pot. The ammonia accompanying the oil into the pot evaporates and returns to the evaporator via the line & valve B2. The heat for this comes from the surroundings, an electric heat or discharge vapour. When the pot is full of oil, i.e. no more ammonia evaporates, the temperature increases. In case of an evaporation temperature below zero, this is clearly visible as the frost on the outside of the pot disappears. Valve B1 is then closed for a while and the last traces of ammonia evaporate. Finally, valve B2 closes and B3 & B4 open. High pressure gas then enters through B3 and forces the oil back to the compressor oil system.
The liquid level is less affected by the flashing refrigerant or the circulating vapour-liquid mixture.
The hybrid separator (Fig. 2B) has the properties: It is basically a horizontal separator with a vertical vessel attached to the bottom. The liquid level is maintained in this vessel.
82
The best type is probably the ball valve, which causes practically no extra loss, followed by the plug cock & gate valves, butterfly valves, and finally the ones with the largest pressure drop, glove & angle valves. The latter are extensively used as control valves. If control of the circulation is necessary, control valves should be installed in the liquid line (figure 2B). It is normally not necessary to control the flow, except possibly for very high placed separators. Heavy foaming can occur if the system has been cleaned with kerosene, white spirit or the like, some of which remains. Other likely foaming agents are oil decomposition products. A longitudinal standing wave can form in the horizontal separator. This can then give an incorrect level reading or expose a connection.
This works fine, but on one condition: a fixed evaporation temperature. The bulb alone only monitors the temperature, superheated or not. Conceivably, exchangeable sets of orifices/needles or springs, one set for each evaporation temperature, could be used. To extend the working range of the valve, the controlling variable is changed to the superheat. We already have the superheated temperature converted to a pressure. From this we have to subtract the evaporation pressure. This could be found where the temperature is monitored - external pressure equalisation - or immediately after the expansion - internal pressure equalisation. Port forces. Before the equalization methods are considered, another set of forces has to be discussed. When the liquid passes the port, it exerts a pressure, FF, on the spindle. The direction of the force is in the flow direction. In figure 09 it tends to close the valve. If the flow is reversed it tends to open it. In other words, the valve capacity depends on the flow direction. This might not be a drawback; a reversible A/C-HP have usually different capacities in the two modes. However, it is difficult to position the spindle exactly as the flow forces vary, sometimes irregularly. Precision of the spindle positioning is especially important at low capacities of a large valve. The spindle is then close to the seat and even small movements cause fluctuations. One way to solve the problem is a double spindle-seat assembly (port), mounted on the same axis, with liquid entrances in opposite directions, see figure 09E. The flow forces FFA and FFB now largely cancel each other and the spindle position depends only on the forces FA, FB & FC. The various manufacturers each have their own designs. External pressure equalisation. It is obviously most correct to monitor the pressure and temperature at the same point. However, an external pressure equalisation line is required. Downstream the bulb, a small pipe is attached to the exit. Through this pipe, the pressure enters compartment B and exerts a force on the underside of the diaphragm, in opposite direction to the bulb pressure. To see how it works, consider the following example for two evaporation temperatures 0 and -20 C, each with 5 K superheat. Evap. temp.: 0 C. Vapour pressure: Bulb ": 5 C. Bulb ": 5.81 4.95 bar bar bar bar bar bar
Force on needle: (FA-FB) ~ (5.81-4.95) = 0.86 Evap. temp.: -20 C. Vapour pressure: Bulb " : -15 C. Bulb ": 2.44 2.94
A valve without pressure equalisation thus faces forces ranging from 2.44 to 4.95 bar, i.e. a ratio of 2.03:1, while a valve with pressure equalisation faces forces ranging from 0.5 to 0.86, i.e. a ratio of 1.72:1. There is a decrease in the range of the forces, but not enough to justify the pressure equalisation. It now becomes clear why the bulb filling might not be the same as the refrigerant. If the bulb pressure at - 15 C remains the same 2.94 bar, the bulb pressure at 5C should be 5.45 bar and the force from the membrane in both cases should be proportional to 0.50 bar. If so, the settings of the superheat adjustment screw in both cases can be equal for a given superheat.
83
Water
Superheated vapour
Superheating
Vaporization
Refrigerant
Liquid
Liquid + Vapour
A. A direct expansion evaporator with a thermal expansion valve & external pressure equalization.
Seat Port
Spindle or pin.
Water
84
Internal pressure equalisation. A simpler design is the thermal expansion valve with internal pressure equalisation, figure 09 D. The wall between compartment C & D is simply removed and the pressure after the expansion is directly exerted on the underside of the membrane. The temperature is therefore monitored after, and the pressure before, the evaporator. The thermostatic expansion valve with internal pressure equalisation should thus be used for: A. Evaporators with a low pressure drop, i.e. exit pressure equal to the inlet pressure. B. Evaporators with constant pressure drop. As the BPHE has a relatively high pressure drop, thermostatic expansion valves with internal pressure equalisation should be avoided. Limitations. A thermostatic expansion valve is thus used to control the refrigerant flow and hence the evaporator capacity providing that the vapour leaves the evaporator superheated. It cannot be used to provide a pre-determined two-phase flow at the evaporator exit, nor can it be used to control the evaporation pressure.
This is one of the reasons why ammonia based plants normally use flooded evaporators, which do not superheat the vapour. Control of the thermostatic expansion valve. The valve lets through a determined amount of refrigerant controlled by the superheat after the evaporator. Why? To answer the question, imagine what would happen if the exit from the evaporator were to be saturated or even contain some residual liquid, say 1 %. What would then be the controlling variable to assure the 1% exit liquid? Unfortunately, it is difficult to find such a variable. Pressure and/or temperature could not be used, as these are the same for a mixture of 1 or 99 % liquid. The mean density could be used, but in practice, it would be extremely difficult or impossible to monitor this. The light dispersion or absorption could be used to detect liquid droplets, but then a certain amount of liquid refrigerants needs to be present. The sound of the liquid hammering monitored by a microphone in the compressor could also be used to control the expansion valve. When the vapour starts to superheat, the situation changes. The temperature and pressure then move independently, providing a set of two control variables. Rising superheat, i.e. the evaporator operates below capacity, the valve opens and more refrigerant passes. Sinking superheat, i.e. the evaporator operates above capacity, the valve closes and less refrigerant passes
85
Superheat K 5 10 5 10
Superheat K
86
When the R22 flow rate changes, the superheat changes according to the curve. For a flow of about 736 kg/hr, the resulting superheat is zero and when the flow approaches zero the resulting superheat approaches its maximum value, 10 K (12 K for a water temperature of 14 C). The evaporator curve is more complex than the valve curve. The amount, temperature and type of heating media will affect the superheat. It is also fairly difficult to calculate correctly the thermal characteristics of an evaporator because there are many parameters, e.g. surface finish, which are difficult to correlate, the flow behaviour is not well understood, etc.
At this low superheat - only 1 K operating superheat - the TEV is very vulnerable to disturbances. (See also 11, 1. Applications and 6.7, here). Conclusion. It is important that the valve and evaporator match each other, i.e. that the valve gives the required capacity for the superheat delivered by the evaporator at this capacity. The operating point should allow the valve a control range in both directions. Normally a valve can be used to ~ 20 % above its rated capacity, which should allow a sufficient range at the rated capacity.
87
CB52-50H Nominal operating point: 29.7 kW = 694 kg/hr at 5 K superheat. Actual operating points. Nominal operating point.
A. The operating point of a thermostatic expansion valve for two evaporator sizes.
The operating point is found by superimposing the two characteristics curves from the valve and the evapora- tor respectively. The crossing point between the curves ves gives the actual operating conditions, about 7.6 K at a flow of 654 kg/hr refrigerant for the small size and 7.9 K at a flow of 694 kg/hr for the large one.
B. The operating point of an evaporator and two sizes of a thermostatic expansion valve.
Assume that the required operating point of the system is 5 K and 696 kg/hr. Obviously, the valve in C is too small. The solution is a larger valve, one that gives 694 kg/hr for an opening superheat of just 1 K instead of for 3.9 K. The smaller valve shown operates fully open, i.e. far above its control area.
Unstable area
Stable area
Superheat
The superheat signal might be attenuated, e.g. by placing the bulb further away from the evaporator. Back mixing of the vapour due to the turbulence decreases the amplitude of the changes. A change of the bulb position means that a signal may no longer be in phase but signals of another frequency might then be in phase. The valve responds very rapidly if the superheat changes because of a pressure change. A pressure change transmits almost instantly to the valve, especially one with internal pressure equalisation. See figure 09. Conceivably, the response could be delayed in a valve with external pressure equalisation by placing an orifice or a small valve in the equalisation line. The valve reacts relatively slowly to a change in temperature as the wall of the connection has to be heated, heat has to be conducted through the wall and to the bulb and the bulb-filling has to change phase. The hunting risk could thus be different for a superheat change due to a pressure or a temperature change. If the static superheat decreases, i.e. the valve curve moves left and becomes relatively steeper, the hunting danger increases. See figure 11A. Two valves of different sizes see figure 11B. The smaller one might not give the capacity but gives less danger of hunting than the larger valve. The total mass of refrigerant and metal can vary from one system to the other and this affects the response. A valve can have the characteristic curve as in figure 15. This can be obtained by by-pass, either internal in the valve or external via a parallel-connected valve.
89
in. Signal to valve out amplified and in phase out attenuated and in phase
in
out
Time
Capacity
Capacity
4 1
3 2
The thick lines show an enlarged section around the operating point.
1 5
7 6
Superheat
Superheat
The new capacity corresponds to a new superheat at 4. The capacity then changes to 5 and so forth. It is easy to see that in case A, the operating point spirals outward and with a risk of hunting. In case B it spirals inward, back to the stable operating point.
Capacity
Superheat
Superheat
90
The evaporator should, if instabilities are anticipated, be designed with a larger superheat than the normal 5 K. Due to the normal overdesigning because of fouling, this is actually the case. This increases the inclination, but with some loss of capacity. The bulb of a MOP valve should never be installed on a vertically pipe with the capillary tube pointing down nor should it be warmer than the valve. This can facilitate the transfer of the filling from the bulb to the valve, especially if the bulb get warmer than the valve. Heat the valve to move back the filling to the bulb. Select a valve with a maximum capacity less than the capacity for which the evaporator gives zero superheat. Should hunting occur there is no danger that liquid refrigerant enters the compressor. Figure 17 shows some rules for installing a thermostatic expansion valve. The bulb is here uninsulated, but should usually be insulated. Refrigerant could be leaking through the packing box of the expansion valve and thus liquid refrigerant enters the vapour to the compressor. Normal design instructions say that the bulb should be installed upstream of the pressure gauge to prevent false bulb readings. However, this means that the actual superheat entering the compressor is not correct. However, if there is sufficient distance between the evaporator and the compressor to place the bulb some 400 -600 mm downstream of the pressure gauge, the bulb could be placed there. The liquid refrigerant from the pressure gauge then has time to evaporate and the bulb will monitor the correct superheat. See also figure 21. When the pressure and pressure difference decrease, so does the force the diaphragm exerts on the valve pin. A sufficient large force is important to obtain a stable operation. The low force can be compensated by a large diaphragm surface. Differently expressed, of otherwise equal valves chose the one with the largest diaphragm diameter, especially important at low temperature operation. Avoid valves with internal parts of plastic as the lifetime can be reduced. In case of suction line separators in a DX installation, consider installating the temperature bulb after the separator. See 6. Suction line separators here. Parallel connected evaporators should each have its own TEV as it is difficult to evenly distribute a two-phase mixture between the evaporators. See also fig. 11a, 3. Design & Installation.. A step motor valve does not close at a loss of power (nor does it change position). Even though it normally closes tight when the plant is shut down, it should therefore be preceeded by a solenoid valve. Avoid pulse width modulated valves, i.e. where the valve stays open for a certain period and then closes, see 6.2. The electronic expansion valve, as they can damage the evaporator.
Various solutions have been proposed; figure 18, but all have some drawbacks. Good distributors as in figure 18 A and B have fairly high pressure drops and can thus not be used in the condenser mode in a reversible cycle. The system shown in figure 18 C is composed of a premixer and the distributor pipe, which is pressed into the plates. It gives a good distribution over the plate pack without excessive pressure drop. It can be used as a condenser, but only total condensation. Partial condensation as well as onephase vapours and liquid cannot be accommodated. The pressure drop in the distributor is less than 0.1 bar at condensing, i.e. some tenths degree Kelvin. Refrigerant flashing is thus not a problem as the condensate is usually more than three degrees subcooled. The expansion valve has to be somewhat larger than normal to make up for the lower effective driving force.
5.6.2. Hunting.
If, despite all precautions, hunting occurs: Experiment with changing the static superheat. Depending on the shape of the valve curve this can affect the stability of the system. Move the bulb further away from the evaporator. This will attenuate oscillations. The bulb should not be too close to the compressor though, otherwise it has insufficient time to react to prevent unevaporated refrigerant entering the compressor. Try to open the valve completely and to keep it open. Are there still oscillations in the system? Try to find out from where they originate. Does hunting only occur at low capacities? At very low channel flow rate the operation is sometimes erratically. Is the flow from the condenser or liquid receiver constant and with constant properties? Is there a hot gas by-pass control or freeze protection (5, 7. FREEZE)? Is that the origin of the hunting? Try to increase the temperature difference between the media in the evaporator. This gives some leeway for moving the valve curve to an area where the inclination is less than the evaporator curve. (Figure 11A). Check the refrigerant filling of the system. If insufficient, the liquid receiver runs dry and the expansion valve receives an irregular flow of refrigerant. This affects both the stability and the capacity of the evaporator. MOP valves have normally less tendencies for hunting.
91
The figure shows a BPHE with the water connections at the back cover plate. 1. The bulb must be installed as the figure shows, never at the bottom of the tube, where oil can give errors. 2. The pressure probe must be installed downward from the bulb. See comments in 5.4.9. 3. The bulb and the pressure probe should preferably be placed after a bend, on a horizontal section of the pipe. The bend acts as a vapour-liquid separator, to prevent liquid refrigerant and oil from interfering with the measurements. 4. The pipe from the thermostatic expansion valve to the evaporator inlet should be straight and of the same diameter as the valve exit. 5. If the distance between the compressor and the bulb & the pressure probe is too short, liquid refrigerant could enter the compressor, as the valve does not have time to react to load changes. 6. The solenoid valve (not shown here) should be placed as close as possible before the thermostatic expansion valve. 7. Chose a valve with an as straight exit as possible. Be observant on possible turns and bends of the vapourliquid mixture inside the valve as the resulting centrifugal forces can lead to phase separation.
< 25 mm
> 30 mm
400 - 600 mm
150 - 200 mm
Holes f or liquid.
B. Sintered disk. D. Premixer and integrated distribution holes pressed into the plates.
Water in ammonia has the same effect. See the appendix for the rise in boiling point due water pollution. Detach the bulb from the exit pipe and let it warm up. The increased temperature in the bulb forces the valve to its maximum capacity. Does the capacity increase? Certain bulb fillings demand that the bulb is the coldest place in the valve-bulb system. If the valve is installed outdoors, the reverse might be true in winter. If there is a leak at the pushrod, refrigerant could cool down the membrane. The valve then stops working or works erratically. Is there vapour or air in the liquid line? These increase the pressure drop both in the TEV and in the lines and less refrigerant passes. Check the condenser pressure. If this is too low, there might not be enough pressure to force the required amount of refrigerant through the valve. Are there instabilities that can lower the capacity? Check the superheat from the evaporator. If the superheat is larger than the design value, this indicates that the evaporator should be able to evaporate more refrigerant than what it actually does, i.e. increase the capacity. The problem is then to be found outside the evaporator. In one way or other, the refrigerant flow to the evaporator is restricted. This could be a too small valve, an obstruction in the pipe, dirt in the filter, ice, the liquid receiver running dry, etc. More refrigerant cannot evaporate than what enters, and if this is insufficient, the capacity is insufficient as well. If the valve cannot provide the required capacity for the measured superheat and the static superheat is set to a minimum, there is no way the system could deliver more capacity. See figure 11. A small hand controlled valve in parallel could both increase the capacity and prevent hunting. See figure 15. Check the entrance temperature at the expansion valve. If it is substantially subcooled, e.g. by an economiser, less vapour has to evaporate than if it enters at the condensation temperature. The lower vapour content lowers the heat transfer coefficient, reducing the capacity. Check if liquid refrigerant leaves the evaporator. The cause of this could be a faulty distributor, a not correct cassette type, blockage of the brine side, etc. Liquid refrigerant then emerges from some channels, leaves the evaporator and causes the bulb to report a too low superheat. The valve then closes. The above is especially dangerous if the bulb is placed too close to the evaporator.
93
Induction coil
Set point
LP receiver (Separator)
HP receiver
A separate vessel is necessary because of an unstable liquid level - waves, bubbles, etc. Pressure motor (Adjustment system not shown)
Expansion valve
A dirty, damaged, frozen or too small TEV. Far too small condensing pressure. Vapour in the condensate line, before the TEV. A TEV with a leakage through the equalization line. Example. The LP guard cut out the compressor in a DX-evaporation plant after about 90 seconds operation. A thorough inspection did not reveal any anomalies. A thermal camera showed that the temperature decreased after the expansion valve and at the evaporator inlet the first seconds after startup. No temperature decrease was detected at the evaporator exit, but curious enough, in the suction line far away from the evaporator. Finally, a picture taken of the equalization line inlet to the suction line, revealed the cause. The inlet was vertical, to the top of the horizontal suction line. At startup, the image was a low contrast orange, indicating an even temperature. After a while, the equalization line inlet became blue, below zero, followed by the bottom of the suction line, just below the inlet. A study of the TEV showed that in this particular valve, the high pressure condensate was on the other side of the stuffing box for the push rod. (As it would be, with reversed flow, in compartment C in fig. 09, p. 83.) A leakage in the stuffing box would lead the high pressure to the underside of the diaphragm and close the valve. At the equalization inlet to the suction line there was a restriction, i.e. the pressure didn't decrease until there. A change of the valve solved the problem.
A too small, a too fouled or an evaporator with a too high pressure drop. Lack of brine. Obviously, if there is no brine entering the evaporator no vapour is produced.
94
6. Control.
6.1. Refrigerant expansion in a flooded evaporator.
The expansion of the refrigerant is seldom controlled by the superheat as in the direct expansion evaporator. There is simply no superheat or temperature difference between exit and inlet to use as control variable. The expansion valve is instead controlled by the liquid level in either the HP or LP liquid receiver. The HP controller should open the valve for increasing liquid level and the LP controller should close it. The level control could be mechanical, thermal, light absorption, etc. See figures 2B, 06K and 19 for some conceptual designs. A HP controller is normally used if there is only one evaporator and LP controllers if there are many parallel evaporators (each one with its own controller). In contrast to the direct expansion evaporator, a flooded evaporator could be installed with a pulse width modulated expansion valve. The separator acts as a giant dampening device, which will even out valve fluctuations. Flooded flow is often used with SWPHEs, which in any case are more insensitive to pulsation.
The electronic controller in figure 20 has the difference, 1 K, as the control variable. The TEV would use the true superheat, 5 K. The difference between the two is the evaporator pressure drop. Suppose that the load decreases to 25 % of the nominal capacity. The pressure drop then decreases to a little more than 0.5 K. The refrigerant thus enters at -3.5 C and leaves saturated at -4 C. The electronic controller will adjust the valve until a difference of 1 K between exit and inlet is obtained and the refrigerant leaves at -2.5 C, superheated to only 1.5 K instead of 5 K. This might be a too small difference to assure a vapour completely free of liquid refrigerant. In such a case, it is better to use the true superheat as the control parameter for the controller. Instead of the inlet temperature to the evaporator, the exit saturation pressure can be used and converted to a temperature. The use of the true temperatures, i.e. not converted from the pressure, to control the valve has one specific advantage: The valve will be refrigerant independent. The question is how to measure the saturation temperature at the same point as the superheat temperature, usually at the exit. Figure 21 gives one idea.
95
Set point ts t
te = -4 C ts = 1 K
ti = 0 C
ts Set point
Temperature
96
The compressor sucks a decreasing amount of refrigerant in kg/s, i.e. a decreasing capacity, due to the lower vapour density at the suction line. At a certain pressure, there is a match between the vapour volume leaving the evaporator and the space compressed by the compressor, i.e. a match between the compressor and evaporator capacities. See figure 22. If the compressor capacity increases relative to the evaporator capacity, temperature and pressure drop and vice versa. The equilibrium falls at the match between the compressor and evaporator capacities.
uid refrigerant from the HP receiver together with vapour can be used, etc. 8) & 9) are expensive, though. The compressor operates against a constant low pressure and no power saving is made. They are thus used mainly in small DX systems. 6) & 7) are more expensive to install but provide opportunities for power saving. These are the most common in large flooded systems. The most interesting method for good evaporator performance, is injection of vapour between the expansion valve and the evaporator, 10) and on-off 5). This keeps the vapour velocity more or less constant in the evaporator, which is important for a secure oil flow through the evaporator, figures 23.
97
To condenser
To compressor
Pressure adjustment
Note! The evaporation pressure can be higher but not lower than the nominal control pressure.
To compressor
Drive liquid Heat coil
Temperature
Fig. 25. The evaporation pressure controller with an electronically controlled set point.
This is basically the same as in figure 23 but instead of manual control of the pressure, the pressure is controlled by the exit temperature of the cooled fluid. Another option is to control the temperature difference between the inlet temperature of the cooled fluid and the evaporation temperature in order to avoid excessive thermal shocks, e.g. at oil cooling.
98
Control through a valve. The system is a hybrid between a flooded and a DX evaporator, see figure 28. A control valve is installed in the thermosiphon loop. Obviously, the refrigerant cannot be flashed to any great extent, as the pressure difference over the valve corresponds to the maximum static height. The valve can either be controlled by the superheat as in a normal direct expansion evaporator or as shown in the figure, by the exit temperature of the cooled liquid. The system is sometimes used for auxiliary evaporators such as oil coolers (8. Oils & Refrigerants, figure 8), oil evaporators (1. Applications, figure 19), etc. The advantage is that the vapour emerges dry, ready to enter the compressor, and there is still no need for the large pressure difference required for a normal thermostatic expansion valve. The simplest system is a solenoid (on-off) valve. A variant of this is a pulse width modulated valve. These should normally not be used together with BPHEs as expansion valves as the pressure shocks can damage the BPHE. In this case, there are almost no shocks. If the temperature difference between the refrigerant and the cooled fluid is not too large (< 30 C) the thermal shocks will be negligible.
hunting maybe starts, and at the exit, the superheat disappears and liquid refrigerant hits the relative hot walls causing thermal fatigue. See also 5.5.2 - 3. d) If a condensing pressure controller is used, it should be installed before the C/E, never after. Installation after could cause condensate to back-up in the C/E. See figure 10 A & B, p. 109 (5. Condensation and Liquid Receivers) and e) below. e) The condensing side should be drained properly, i.e. condensate is not allow to back-up in the E/C. See d) above and figures 05 & 06, p. 103ff (5. Condensation and Liquid Receivers). f) The above also means that condensate subcooling in the E/C is not allowed. g) When only one of many LT compressors operates, the E/C capacity will be only a part of full capacity. If only one TEV is used, this will then operate with a very low superheat and is thus susceptible to disturbances. See also 5.5.2, p. 86. The arrangement in figure 27 could be used to secure an inoffensive small flow of refrigerant at part load. (Compare with 8. Oils & Refrigerants, 8 D). h) Mufflers and oil separators tend to even out pressure variations from the LT-compressors. i) If possible, use slow moving solenoid valves, especially if the condensing side is started up first. j) A flooded evaporator, which remains filled with refrigerant during a shutdown, gives fewer temperature shocks than a DX evaporator.
6.7. Condenser/Evaporator.
In a BPHE with condensing refrigerant on one side and evaporating refrigerant on the other, there is a danger of rapid temperature changes because of the excellent heat transfer, due to the phase changes, on both sides. The danger is especially large if there are compressors in parallel on either side. To minimize the risk of damage due to thermal shocks, the control system should, especially if there are large temperature differences between the sides, be laid out as below and in figure 27: a) At startup, the condensing side should, if possible, be put into operation at the same time or after, but not before, the refrigeration side. See b) below. If the vapour line from the LT compressor is sufficient long there will automatically be a delay before the hot vapour enters the E/C. If the condensing refrigerant enters first, the BPHE heats up to the superheat temperature and the hot BPHE then receives the cold, evaporating refrigerant. Contrarily, a BPHE of the evaporation temperature will effectively receive refrigerant of only the condensation temperature, i.e. a much lower temperature difference. b) A suction line liquid separator (see 7.) could be installed to handle an initial liquid refrigerant surge from the evaporator if an arrangement as in a) is made. c) An evaporation pressure controller should be installed, especially in a multi-compressor installation. If the C/E (Condenser/Evaporator) is one of a number of parallel evaporators, e.g. unit coolers, the temperature setting should be somewhat higher than for the rest of the evaporators. The purpose is to insulate the C/E from rapid pressure changes, which can occur when a compressor cuts in or out. If the C/E then operates at a low capacity, the superheat is likely to be low. If a compressor on the evaporating side suddenly cuts in, the pressure drops. The lower pressure is immediately transmitted to the TEV while a possible temperature change takes longer time. The valve then senses a higher superheat and opens,
99
C2
C1
HT-Circuit
S3 S2
From HT Condensers V1 V2 V3
An alternative is one 1/3 TEV and one 2/3 capacity TEV. Depending on the number of compressors operating, one or both TEVs are active. In stead of a small TEV a fixed valve, which gives about 1/3 capacity could conceivably be used The pressure setting of the evaporation pressure controller PC2 should be higher than PC1 Or than the actual HT suction pressure). A possible condensing pressure controller should be installed before not after the E/C. Subcooling is not allowed in the E/C
Circular rizer
To ensure a liquid free vapour, a heating coil is at times installed in the separator. It is usually heated by condensate.
S4
S1
Circular descender
S3
S2
Oil Refrigerant
100
3. Design considerations.
3.1. Temperature profile.
In a condenser, the refrigerant enters superheated, with temperatures ranging from about 60 to 120 C. Higher temperatures than 120 C, though possible, are seldom encountered as oil in the compressor then starts to decompose and the system is, consequently, designed to keep the discharge temperature below this value. As in all heat exchangers, the cold side can only approach but never entirely reach the hot side temperature. When the exit water temperature increases by decreasing the water flow rate, the water temperature curve approaches the vapour curve, not at the end temperature but at the point where the vapour starts to condense. The water exit temperature is still far from the vapour entrance temperature, see figure 1, which shows the desuperheating and condensing of 1 kg/s R22 vapour entering at 87 C, condensing at 45 C and leaving subcooled at 40 C. 3.28 kg/s water is heated 25 to 40 C.
2. Condenser arrangement.
The selection of condenser type, cooling media, temperature level, control system, etc., depends very much on the purpose of the plant and is beyond the scope of this manual. Below are some indications: A) If the condenser is installed in an A/C or refrigeration plant without heat recovery, the only purpose of the condenser is to dump the heat in a suitable receptacle. The condenser type then depends on this receptacle and has to be as small as possible for the given data. For a water-cooled PHE, the techniques in the chapter 2. Optimization can be used to minimize the surface and the cooling water flow. B) The correct condensing temperature or better the pressure in A) depends on the type of evaporator: A direct expansion system needs a minimum pressure to provide the TEV with an ample driving force. The driving force in a flooded system is less critical and the condensing pressure might be left to vary with the cooling media temperature, thus saving energy. C) In a heat pump, condensation is the primary objective. Condensation might be split in a desuperheater for tap water heating and a condenser for space heating. The control system is probably arranged to maintain a specified minimum temperature level. D) An A/C can be used to recover heat from the condenser, usually the high level heat from the superheated
= 45 - 25 = 20 K (The temperature increase in the condensing section.) = 167 kW = 206 kW (The thermodynamic design.) (The thermodynamic design.)
thus, ttotal = 24.7 K and an exit temperature of 49.7 C, corresponding to a flow rate of 3.28 * 15 / 24.7 = 2.0 kg/s. Note that this is the limit for an infinitely large condenser. An even lower flow rate would mean a higher temperature of the water, which is clearly impossible. The nominal water flow, 3.28 kg/s, gives a temperature of 37.5 C at the pinch point, which is somewhat below the practical limit for a heat exchanger. In practice, a decrease of the water flow means that the condenser will not condense everything. That means moving the pinch point to the right to clear the water curve. This applies regardless of the size of heat exchanger.
101
Condensing 167kw
87 C % load
0.169 kg/s
80 C
~ 40 C
3.3. Inerts.
When a vapour containing an inert gas condenses in a condenser the following phenomena occurs, figure 03 a)The volume flow of the vapour-gas mixture decreases and the heat transfer coefficient decreases. b) The vapour layer close to the wall is progressively enriched with gas as condensation proceeds. This gas film acts as a barrier to the vapour. The vapour has no direct access to the cold surface to condense; it first has to diffuse through the inert gas layer. c) The partial pressure of the vapour decreases and the
102
In the absence of sightglasses it can be difficult to determine which of the two was the culprit. However, the suction and discharge pressure manometers, vibrated heavily. Especially the later was almost impossible to read. The third manometer, the oil pressure was completely stable. All were of the same type, had the same pressure range and were mounted side by side, close to the compressor. The stable oil manometer ruled out the cause as vibrations from the compressor. A vibrating discharge manometer hints of inerts. A check of the condensate filling would mean a lot of work with the piping while the simpler check with venteing could be done immediately without stoping the plant. The condenser was vented from the subsequent liquid receiver by simple connecting one end of a hose to the vent and the other into a bucket of water with cold water. Ammonia dissolves readily in water while the air bubbles rise to the surface. There was an initial very large air flow, which after about half an hour decreased to a stable value, which then continued for about five hours. After the initial surge of air, the manometers became stable and after the five hours both compressors could be connected. The subcooling was then normal, about 3 K. 4) Leakage into an evaporator operating at vacuum. 5) In once-through installations, see example below. Example II: A WPHE condensing propane was troubled by frequents cutouts of the HP pressostat. Inerts were suspected but an analyser revealed only 100 % propane. A closer investigation revealed that this was not a true refrigeration plant. The source of the propane vapour was the inevitable vaporization of propane stored in a commercial tank farm. This propane vapour is compressed, liquefied and returned to the tank storage. When the tank is emptied, the space above the propane is filled with an inert gas, normally nitrogen. The propane vapour to be compressed thus contains nitrogen. Furthermore, the analyser was actually for butane but the reading was converted to propane according to the formula X % butane = Y % propane. Unfortunately, it is impossible in this way to distinguish between 100 % or say 99.9 % propane. (A nitrogen analyser would have been better; the difference between 0 % and 0.1 % is large.) Installation of a venting system solved the problem. Inerts can have a dramatic effect on capacity. The problem is especially acute in a condenser with condensate level control (see below). Here, the inerts are effectively trapped in the condenser, see figure 04, p. 103.
cooling could mean a large change in the condensing part. It is thus impossible to control the temperature by a simple level control without affecting the capacity. The level could be kept constant and the condensate temperature controlled by the water flow but this too would affect the condensing part. It is impossible to control the capacity by changing the condensate level or the water flow without also affecting the temperature but this is usually less serious. The dynamic response when decreasing the condensate level is more rapid than when increasing the level. Decreasing, i.e. draining is almost immediate, while increasing could be slower, as sufficient refrigerant has to condense in order to build up the required condensate level. This could cause control problems. The condensate level might be different during summer and winter operation and the refrigerant difference has to be stored in the liquid receiver. This increases the total filling of refrigerant, which is expensive. The valve used to control the condensate level has to be installed after the condenser, but as close to it as possible. Otherwise, the volume in the pipe from the condenser to the valve unnecessarily increases the total refrigerant content while at the same time causing a delay when the condensate level is increased. Condensate cooling in the condenser should never be made for ammonia and water. The high latent and specific heats, together with the high thermal conductivity and low viscosity lead to very high heat transfer and rapid temperature changes, which could cause thermal fatigue. In general, ammonia and water condenser should be drained as completely and rapidly as possible. Note the similarities to oil cooling, see 8. Oil & Refrigerants, 2, p 138. To minimize the danger of thermal fatigue, do not use condensate subcooling (condensate should not be allowed to collect in the condenser) if the temperature difference between vapour and water inlet temperatures is more than 40 - 50 K for HCs, HFCs & HCFCs. Conclusion: Condensate temperature control should be avoided in the condenser. Capacity control, with the considerations and restrictions above, could be done if the exit temperature is accepted as it is.
103
Inerts (and light insoluble oil) can be trapped here Avoid a loop in an ammonia system as oil traps here.
Control valve
From compressor
A large pressure drop in a condenser is compensated for by a Avoid long horizontal liquid lines, large liquid column at the exit. they should be forward inclining.
To liquid receiver
104
4.3. Operation.
The main use of the EQ line is to direct vapour back to the condenser if the refrigerant in the receiver is heated and evaporates. In winter, the opposite could happen. A very cold condensate produces a low temperature, vapour flows from the inlet to the receiver and heats up the contents. As a warm condensate means loss of capacity, the EQ line should be closed. If the refrigerant evaporates in a through receiver, vapour flows back into the condensate line, enters the channels from below and condenses in the lower part of the condenser. Small amounts of vapour can be accommodated in this way but larger amounts could create a condensate plug in the lower part of the channels. If the condenser is properly designed, i.e. the extra vapour from the LR is added to the flow from the compressor, the condensation might work. The condensate plug is unstable, the more so if inerts are present even in very small amounts. It could selectively build up in some channels, then suddenly drain, rebuild in others, resulting in a fluctuating flow and loss of capacity. The design is recommended for S&THEs. A similar condition to the above could happen if the condensate subcools very much and subsequently reheats somewhat, by vapour or by the water if this is in cocurrent. The lowest pressure then occurs a bit inside the channels and a condensate column might be hanging in the channels. This too is unstable, though it should not give problems with insufficient capacity. Installation of the through LR above the condenser is not impossible but should be avoided as it can lead to irregular flow and control problems, notably in the case of parallel condensers. Care has then to be taken that the pressure in the condenser is higher than in the LR. Certain expansion valves for flooded operation are integrated with the LR. A small pipe from the vapour space of the LR connects to the line after the expansion. This creates the necessary lower pressure and these can be mounted above the condenser exit. A downward U-bend between the condenser and a through LR transforms this to a hybrid LR. It cannot be vented but it can be equipped with an EQ line as a condensate column now can form.
4. Liquid receivers.
4.1. Types and purposes.
The condensate leaving the condenser in larger systems is normally collected in a vessel, the liquid receiver (LR) See figures 06. It is sometimes called the HP (High Pressure) receiver to distinguish it from the vapour-liquid separator in a flooded system, which then is a LP (Low Pressure) receiver. There are mainly two types of LRs: A) The surge liquid receiver. The condensate normally bypasses this. It handles normal load variations in the plant. It is usually connected to the condenser inlet through a pressure equalization (EQ) line, see below. B) The through liquid receiver. The condensate always flows through this. It is better for dumping refrigerant contents from other parts of the plant. It should not normally be equipped with an EQ line. The liquid receiver serves the following functions: The LR has to absorb or release refrigerant during load variations as the refrigerant contents of the various components in the system then change. If a component is shut down, e.g. for service, its refrigerant content can be dumped in the receiver. Maintains a supply of liquid refrigerant at constant pressure to the expansion valve. In a flooded evaporator, the liquid level in the separator normally controls the expansion. The liquid receiver can be used as it has a more stable liquid surface. The drawback is that only one expansion valve (evaporator) can be controlled. Act as a vapour-liquid separator for cooling of the hot compressor oil in a small BPHE evaporator. The BPHE is subjected to less thermal shocks if using warm HP refrigerant rather than LP refrigerant.
105
Desuperheater
Vent, either to: The air or A refrigerant recovering unit EQualization line
Condenser
To expansion valve Dont connect an EQ line before a desuperheater. The higher liquid column can flood the condenser.
Sight glasses. Consider two: One for looking. One for a light.
Vent of an ammonia condenser. As ammonia is extremely soluble in water, air bubbles rise to the surface but ammonia vapour dissolves in the water. See text 5.3
Fig. 06. The liquid receiver and the pressure equalization line.
A. The surge liquid receiver.
1) The condensate enters and leaves from the bottom. As the condensate line is always liquid filled, a liquid column can be created in the condensate pipe. 2) Inerts cannot follow the condensate to the receiver and be vented from this. Use the position shown. 3) It should never be installed above the condenser (liquid level below condenser exit) as the condensate might flow back and flood the condenser. Troubleshooting a flooded condenser can be very difficult. 4) A pressure equalization line can be installed. The purpose of this is to carry vapour formed in the receiver back to the condenser inlet for recondensation. To overcome the higher pressure at the inlet, a condensate column - height h - forms in the condensate pipe. This is easily done as it is always liquid filled. 5) Sources of vapour in the receiver can be: Heat from the outside in a warm machine room. As the refrigerant content is more stationary here than in the through type there is a greater danger of heating from the outside. An oil cooler, cooled by evaporating condensate. The vapour could return to the LR and then through the EQ line to the condenser inlet or directly to the condenser inlet.
6) Note that the EQ line has to be open if the receiver is
106
5. Venting.
5.1. Parameters affecting the venting.
Inerts can be suspected and venting should be done if: The capacity is decreased and fouling is excluded. The exit temperature is very much lower than the condensing temperature of the vapour. The manometer needle vibrates rapidly. (Ex. I, 3.3) See 3.3 for the effect of inerts (noncondensable gases) on the condensation process. Note that a decreased capacity and a low exit temperature can depend on a condensate level in the condenser as well. The exit piping from the condensor to the expansion valve has to be checked for obstructions. See also 3.4 & 3.3.
Venting must not be done from the port itself. It can be done from the succeeding piping or the liquid receiver. Figure 06 shows one method of venting for both LR types. This method for the surge receiver can be used for the through receiver as well and should be used if there is any danger of condensate blocking the exit pipe.
107
TEV
To expansion valve
Expansion valve controlled by the liquid level. When the level and thus the float rise, the valve opens.
Fig. 07B. A level controlled expansion valve with a vent for an HP receiver.
The design is conceptual. Different manufacturers have various designs for manual control, pilot, venting, etc. Turn it upside down and it becomes an LP receiver, i.e. closing at rising level.
6. Control of condensers.
6.1. Condenser control requirements.
The control system depends on the type of system: a) Cooling plant, with flooded evaporators, either refrigeration or A/C. No control is necessary except that water could be saved at low cooling demands. b) Cooling plant, with DX evaporators. The condensing pressure has to be kept to a minimum, to give the TEVs sufficient driving force. c) Cooling plant, with heat recovery. The pressure has to be kept to a minimum to recover useful heat but the primary function of the plant, the cooling, must not be jeopardized. Thus, either hot water of a certain temperature or more water of a lower temperature might be produced but probably not both simultaneously. d) Heat pump. The heat output is the primary. If possible both requirements in c) should be complied with. In all systems, there are additional control requirements, which can be said to be security controls: e) The condensation pressure must not be so high that it jeopardizes the compressor. This a part of the compressor control and is not discussed here. f) The water must not freeze. This is not usually a problem in a condenser, but in a reversible HP/AC unit the condenser becomes evaporator and vice versa or a condenser could be installed outdoors in winter. As freeze control is very important, an entire chapter, 7. Freeze and a part of 1. Applications, is dedicated to this. g) The water temperature must not be so high that the cooling water scales. This is not only a control problem but also a water quality problem. h) The steady forward flow of the refrigerant has to be assured. There are thus a number of check-valves, solenoid valves, and level controllers etc. to assure that the refrigerant doesnt flow to nonoperating equipment. The design of a control system for a refrigeration plant is not a part of this manual, but some comments on the particularities of PHE will be given.
108
A B C
% Load
7. Desuperheating.
Desuperheating might occur in: The main condenser and is treated as a part of this. In a special desuperheater, usually for recovery of high temperature energy. Sometimes refrigerant condenses in the desuperheater and has to be drained from this. This desuperheater could be a separate heat exchanger figures 11A, 13 & 14 or integrated with the main condenser (11B). In pipe or a vessel in order to decrease the inlet temperature to the condenser, mainly SWPHE and with ammonia as refrigerant. The purpose is to decrease the thermal load on the ring gaskets.
109
From compressor. 1. Liquid receiver controller. Opens at falling exit pressure; not influenced by the inlet.
From compressor. 2. Pressure controller. Closes at falling inlet pressure; not influenced by the exit.
1. Liquid receiver controller. 2. Condenser controller. Closes at falling inlet pressure; not influenced by the exit. Opens at falling exit pressure; not influenced by the inlet.
One-way valve
To TEV
To TEV
The pressure in the air condenser is left free. The liquid column recovers the pressure drop in the air condenser and the controller. Tap water
Cooling water
Through LR
To TEV
110
A. If condensate for desuperheating is exrtracted far below the condenser, the pressure might be high enough to allow for injection of condensate into the vapour line. A large height difference might be necessary to allow for a good mixing.
From compressor Vent By injection of condensate the vapour is desuperheated. To expansion valve. B. A booster pump befor the injector allows for the use of nozzles to a ensure a good mixing of vapour and liquid. It is an expensive solution, though. C. An ejector pump sucks condensate into the vapour line. It also ensures a very good mixing. However, the effectiveness drops rapidly at part capacity.
To expansion valve
The liquid column makes the pressure in the LR larger than at the desuperheater inlet. The vapour can thus flow to desuperheater inlet By bubbling through the condensate, the vapour is desuperheated. The condensate heats up, though.
From compressor
111
Pilot valve
LR pressure controller
BPHE as tap water heater. Check valve Tap water valve Chck valve Air cooled (main) condenser
Solenoid valve
Diversion line
Pilot valve
Diversion valve
112
8. Troubleshooting.
8.1. Insufficient capacity.
Check all flow rates, temperatures, pressure drops, etc. Does the pressure drop suggest anything abnormal? An obstruction of the water flow, flooding from the LR, an unusual noise, etc. Interchange thermometers between various locations. Small temperature differences can easily be masked by incorrect thermometers. Inspect the condenser for temperature variations on the outside. Large differences suggest an inert gas pocket or a blocked water or refrigerant flow. Is a filter installed in the water inlet port? If this is partly blocked, the resulting unequal flow distribution decreases the capacity. Double-check the heat transfer, by using various combinations of temperatures and flows. Inspect the cooling fluid if it is not pure water. If glycol or similar, check for concentration and/or viscosity. A too high concentration will impair the heat transfer. Inspect the condenser for signs of fouling on the liquid side and oil on the refrigerant side. What are the signs of the insufficient capacity? If the condenser is used to recover heat and the exit water temperature is too low, the reason could be that the water temperature cannot pass the temperature pinch on the condensation curve, see figure 01. Are the evaporator and the compressor compatible? If so, the condenser capacity should be correct too. If not, there must be an irregularity somewhere. Check the compressor. Does it give sufficient refrigerant at the rated pressure? More refrigerant cannot condense than what enters. Is the compressor worn out, have internal leakages? Are the speed and electricity consumption compatible with the capacity? Are there instabilities? These can lower the capacity. A low capacity, but a very large condensate subcooling suggests a too high condensate level, blocking the heating surface for condensing. This could be due to a too large refrigerant filling in the system or too short condensate lines in a system of parallel condensers, thus causing one HE to flood. Another reason for low capacity/large subcooling could be inerts, which are not properly vented. In the absence of sight glasses it can be difficult to verify if this is due to inerts or condensate blocking. Inerts can manifest themselves by a heavily vibrating discharge pressure manometer, see example I, p. 101 Note that the capacity can be controlled by regulating the condensate level in the condenser if the condensate exit temperature is left free. The condensate temperature cannot be controlled by varying the condensate level as this would mean that the condensing capacity has to be left free, which normally is not possible. A too little capacity could thus be caused by a mistaken control of the condensate temperature. If condensate level control is used in the various condensers in a system, the necessary amount of refriger-
ant is larger during the cold season as a part of the refrigerant is used to block the heating surface. During the warm season, this extra refrigerant is dumped in the liquid receiver. If this is too small, it will block a heating surface somewhere. A leaking liquid receiver pressure controller bypasses refrigerant vapour around the condenser into the receiver and the appearance could be as if not all the refrigerant condenses. Check the EQ line. If this is open and the LR is of the through type, condensate could block the channels. Contrary, if the LR is of the surge type and is heated (warm machine room, oil cooler, etc.) and there is no EQ line, condensate could ascend to the condenser and flood the channels and impede venting. Check the installation according to the recom-
mendations in figures 05 & 06. After fouling on the water side, badly installed and operated EQ lines and vents are the principal sources of condenser failures.
8.2. Instabilities.
A refrigeration plant is a system of interconnected feedback loops, with the inherent risk of oscillation. See section Evaporation for a discussion on oscillations. Check the temperature difference. If the water temperature at some point is too close to the refrigerant temperature, a small change of the water flow rate could cause a large change of the temperature difference therefore with instabilities. Carefully choose the location of the control valve as in figures 10A 11. A large distance between valve and condenser means that condensate has to fill up the line before the condensate level increases, i.e. a long response time but a short evacuation time. Check the mechanical functioning of the various valves. Some valve types, notably expansion valves, have very small flow cross-sections. Debris such as metal chips from wear and tear are easily stuck there, impair the flow, and destroy the valve. If the flow decreases because of an obstruction, the capacity decreases as well. Such a valve could be revealed through an irregular control function and/or an unusual noise. If condensate blocks the channels e.g. because of a too high LR pressure, see figure 6, it could suddenly drain and cause instabilities.
113
Refrigerant valve
Desuperheater (& condenser) Heat recovery: On Heat recovery: Off Air cooled (main) condenser Liquid receiver pressure controller
LT pressure controller
HT pressure controller
LT
HT
HT: On
114
115
The net effect could be a fairly fouling and corrosive brew, which needs various treatments before being allowed into the more sensitive parts of a plant. Fortunately, as it is a closed loop, this is easy to do. A cooling tower is normally designed to cool water by about 5 C, thus specifying the available temperature increase in the condenser. D. Surface water from rivers and lakes. The concentration of scale-forming salts is usually fairly low. However, It can contain appreciable amounts of solids, ranging from sand & earth to leaves and fish. Its microbiological activity (algae, bacteria & fungi) can be high, especially in water from agricultural areas, which contains fertilizers. Usually some type of pre-treatment is necessary, especially to control the biological activity. The temperature is usually between that of well water and cooling tower water. Normally the temperature should not be allowed to increase more than 10 C for environmental reasons. E. Municipal wastewater. In certain locations, the only water available for large consumers of industrial cooling water such as steel works, refineries, chemical process industries, etc. is municipal wastewater, usually treated, which can be more or less crude. It contains large amounts of natural fertilizers, especially free ammonia. It is sometimes aerated to dissipate the ammonia. The suitability of this water is questionable as a cooling medium for a BPHE condenser, because of the risk of corrosion and fouling. F. Brackish & sea water should not be used in current BPHE condensers (stainless steel and copper construction) because of the corrosive action of the Cl ion. Note! Nickel brazed BPHE with plates of Avesta 254 SMO could be used with brackish water. Semiwelded heat exchangers with titanium plates are fully resistant to all types of sea water, even the most polluted.
If cooling tower water is not treated with an algaecide, the open structure of the tower provides an excellent breeding ground for algae. Algae are not found in closed circuits, as they need light to grow. Thus fouling from algae always enters from outside the condenser, making it at least theoretically open to interception and treatment. Bacteria, unlike algae, can grow anywhere and cannot be trapped by filters or strainers, making it impossible to prevent bacterial growth, if favourable conditions exist. Some bacteria can feed on sulphates in the water and convert them into sulphuric acid, which causes corrosion. Others feed on iron, either dissolved or oxides, forming a coat of a viscous mass. Still others feed on nitrogen or phosphor compounds forming slime on heat exchanger surfaces. As the condenser is the warmest place in a system, bacteria find its surface a warm, cosy place to settle on. Water can look innocent when inspected but still contain bacterial growths, making this type of fouling insidious and hard to detect. Fungi. Produce similar effects to bacteria but are usually harder to eliminate. The biggest problem with fungi is not the effect on the heat exchanger and other process equipment, but the way they attack the wooden structure of cooling towers. Deposits of sand, etc. Sand, dust, etc. can settle in the BPHE inlet, but usually only at very low water velocities. Growth of barnacles, mussels, larvae, etc. is normally only a problem in salt or brackish water.
117
5.0
pCa or pAlc
2.9 2.8 0 C 2.7 2.6 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.1
pS for total solids
5 10 15 20 25 30
4.0
pAlc. Alcalinity as CaCO3
3.0
10 mg/l CaO = 17.85 mg/l CaCO3 = 7.19 mg/l MgO = 15.04 mg/l MgCO3 = 1: (One German hardness degree)
2.0
35 40 45 50 55 60 65 70 75 80 85 90 95 100 10 100
1.0 1.0
1000
by titration of the water with an acid and with methyl orange as indicator. The result is expressed in various ways. Here we use the equivalent content of calcium carbonate, CaCO3. Get pAlc from the diagram.
Calcium content, pCa. This is normally expressed as either
the concentration of calcium as calcium carbo- nate, CaCO3 or as calcium ion Ca++. The relationship between the two is: 100 g CaCO3 <=> 40 g Ca++ Ex. An analysis gives the calcium content as 92 ppm Ca++, which corresponds to 230 ppm CaCO 3. With this value, get pCa from the diagram above.
Example. Water: pH = 10, 2300 ppm total solids, alca-
linity = 100 ppm CaCO3, Calcium = 500 ppm CaCO3. Is it scaling in contact with a surface of 55 C? pH = 10, pS = 1.72, pAlc = 2.6, pCa = 2.3 L.I. = 10 - 1.72 - 2.6 - 2.3 = 3.38 R.S.I.= 2*(1.72-2.6-2.3) - 10 = 3.24 Both show large scaling tendencies.
118
Warning 2!! Sodium hydroxide dissolves proteins. These not only include bacterial deposits in the heat exchanger but also your skin and especially your eyes. Warning 3!! Sodium hydroxide will corrode aluminium very rapidly. Apart from destroying the aluminium, hydrogen will develop, which means a danger of explosion. C. Scaling is best cleaned with commercial products, which usually contain additives to enhance the effect and/or to prevent corrosion. An example of a suitable product is P3 T288 from Henkel. In an emergency, weak acids such as formic, citric, acetic, oxalic or phosphoric acid can be used. A 5 % solution in water is normally sufficient. After use, do not forget to rinse the heat exchanger carefully with clean water. A solution of 1 - 2 % sodium hydroxide (NaOH) or sodium bicarbonate (NaHCO3) before the last rinse ensures that all acid is neutralised. Warning 4!! Concentrated acids are corrosive and should be handled carefully. Acetic acid, though innocent on the dinner table as vinegar, causes skin blisters. The author of this manual once was able to remove the skin of his left hand, almost like a glove, after only a few minutes contact with concentrated acetic acid. Warning 5!! Under no circumstances use strong inorganic acids such as hydrochloric, nitric or sulphuric acid. Under the right conditions, hydrochloric acid can corrode stainless steel in minutes, nitric acid corrodes copper and all, but especially concentrated sulphuric acid, are extremely hazardous. Warning 6!! When producing a diluted solution of an acid, lye, a detergent or whatever, always pour the compound to be diluted into the water, never the contrary. A little water poured onto a large amount of sulphuric acid will start to boil and splash acid. A little acid poured into water will sink to the bottom and cool off. D. Fouling from biological growths can be prevented through an appropriate treatment of the cooling water. This is easier in a closed circuit like cooling tower water than for once-through cooling water, but it is sometimes done for this as well. Biological water treatment should always be made in cooperation with a water treatment specialist. Below are some examples of chemicals that can be used for prevention of organic growths. Chlorophenols. Effective against a large spectrum of microorganisms, although certain microorganisms can develop resistance. Potassium permanganate. Being a strong oxidant it attacks and decomposes organic material in general, including microorganisms. Sodium dichromate or sodium chromate. These are used mainly as corrosion inhibitors, but being oxidants, they decompose microorganisms as well. Copper sulphate. This should not be used where there are components of mild steel, as the electrolytic deposit of copper from the copper sulphate could cause severe corrosion. See figure 04. Quaternary ammonia compounds. These are organic compounds where one or more hydrogen atoms in ammonia have been exchanged for an organic group. Being closely related to ammonia they could have similar corrosive action on copper. Use only proprietary brands with an assurance from the manufacturer that they are safe to use in equipment containing copper.
Chlorine. Chlorine is extremely effective against all type of microorganisms, and resistance cannot develop. It attacks organic compounds in general and some inorganic compounds too. As it is toxic, care should be taken (consult a specialist) when using it. Chlorine is a gas at normal pressure and temperature (BP -34.7 C). It is normally supplied as a liquid in pressurized vessels. As it is heavier than air it can collect on the floor in confined spaces, machine rooms, cellars, ships, etc. Being a strong irritant, its presence will be revealed long before toxic level is attained. The first effects of chlorine poisoning are pains in the chest and difficulties in breathing. It is useful to know that breathing some ammonia (e.g. from a bottle of ammonia solution in water) will relieve the pains and facilitate breathing. Due to the toxic properties, the use of liquid chlorine is usually limited to large chemical plants. Smaller plants use either calcium hypochlorite or sodium hypochlorite. When dissolved in water the active chlorine compound is the same as for free chlorine dissolved in water. The most elegant way to produce chlorine is with sea water, especially on board ships. Electrolysis of NaCl common salt - produces chlorine on one electrode and hydrogen and sodium hydroxide on the other. Chlorine is used not only in cooling water but also in drinking and swimming pool water. In cooling water systems chlorine (or chlorine salts) are used either for continuous or shock chlorination. Continuous chlorination. A concentration of 0.2 - 0.3 ppm is sufficient. This is consistent with stainless steel and copper used in CB heat exchangers. Shock chlorination. The period of chlorination and the frequency has to be established through experience, as the properties of the cooling water vary greatly from place to place and with the season. Usually a period of about half an hour for 4 - 6 times a day, at a chlorine concentration of 1 - 2 ppm after the heat exchanger, is a good starting value. The total chlorine consumption is less for shock than for continuous dosing, an important environmental factor. A disadvantage of shock chlorination is that certain types of marine organisms - e.g. barnacles - can learn to anticipate the chlorine shock and close their shells, rendering chlorination ineffective. In such cases, chlorination should be carried out at unequal time intervals. Warning 7!! Free chlorine in combination with the chlorine ion can be extremely corrosive on all types of stainless steel and some other high quality alloys, e.g., Hastelloy C. Recommendations for the use of chlorine in stainless steel heat exchangers can be found later in this chapter, in the section on corrosion. d copper used in CB heat exchangers. Ozone - chemical formula O3 - is a gas, which in higher concentrations is toxic, in lower irritating. It is the smell, which is felt close to large electric motors and during thundering. Like chlorine, it is highly reactive to organic material and thus kills all types of microorganisms. Resistance to ozone cannot be developed. The advantage over chlorine is that the decomposition products are more innocent than the organic chlorine compounds resulting from the use of chlorine. It is, however, more expensive to produce and difficult to transport. The efficiency in water and seawater seems to increase, the colder the water is.
119
II
I. Back flushing unit or II. strainer Both for flow: > ~ 40 m3/hr F. Drier/ filter
Evaporator
C. CIP (Cleaning in Place) unit. The hoses from the CIP unit are connected to the second pair of nozzles. The cleaning liquid should have reverse flow direction compared to the normal flow direction. D. Oil filter. E,. Filter for debris in the cooling circuit. F. Filter/Drier for the refrigerant.
4. Design recommendations.
The PHE heat exchanger provides good resistance to surface fouling due to: The plate corrugation causes a very high turbulence, even at low flow rates. Turbulent flow is important to keep suspended particles from settling which they do more readily with laminar flow. Ex. The shear forces are proportional to the pressure drop per meter flow length. A coaxial HE and a BPHE both have a pressure drop of 0.5 bar for given equal water flow rate. The coaxial, however, has a total pipe length of 6 m as compared to the brazed 0.6 m. Therefore; the BPHE generates shear forces ten times as great as the coaxial. No dead zones where fouling compounds can settle. A high surface finish prevents particles from attaching to the heat exchanger surface. The copper in a copper brazed unit creates an unfriendly environment for microorganisms. It is, however, susceptible to fouling by sticky products such as certain types of clay, slime, and mud and to fibrous substances. The PHE is susceptible to pressure drop fouling. It can, however, be easily controlled by the two methods previously mentioned; back flushing or filter/strainer.
B A C D
120
A PHE operating with potentially fouling fluids should be oversized to accommodate the capacity decrease due to the fouling. Unfortunately, it is difficult to say how much, as fouling tends to be either everything or nothing. If the fouling consists of debris, these should be prevented from entering the PHE. If so the PHE could operate infinitely, if not, the PHE could clog in some days. It is impossible to give a general recommendation of a suitable addition of surface to accommodate the fouling. Fouling due to scaling or due to growth of microorganisms such as slime can be kept in control by previous mentioned methods. Here too general recommendations of surface additions are impossible to give. See also 4 in chapter 2. Optimization, for the effect of surface margins under various conditions. Recommendation 1: Use as high a water pressure drop (i.e. shear stresses) as possible. Note! If the main source of fouling is debris, it might be questionable to increase the shear stresses by increasing the water flow rate. A larger amount of water, with potentially more debris then enters and the result could be more clogging. In case of scaling and slime formation though, large shear stresses are always beneficial. Recommendation 2: For a heat exchanger with a very high temperature (> 70 C) on the hot side and/or very hard cooling water (and danger of scaling), increase the pressure drop as much as possible on the cooling water side and reduce it on the hot side. This reduces the wall temperature of the cooling water side and increases the shear stresses, thus making it more difficult for the scaling compounds to adhere. Recommendation 3: Consider the use of cocurrent flow instead of counter current. The warmest part of the hot side, the entrance, will then face the coldest part of the cold side. This will normally lower the maximum wall temperature on the cooling water side and automatically limit the exit water temperature. See figure 03. Recommendation 4: The normal practice is to let the cooling water enter the lower nozzle. This arrangement should be used whenever possible since cooling water entering through the upper nozzle could encourage debris to enter the channels. Recommendation 5: In certain cases it might be advantageous to place a condenser horizontally or almost horizontally (with the individual plates vertically). There is little or no decrease of the condensing capacity. However, sand and other particles settle more easily in such an arrangement and clog the heat exchanger. Ideally, keep the normal vertical position for water containing particles that settle or increase the pressure drop.
5. Corrosion.
Corrosion is the nature's revenge on the engineer, who transforms metals from their chemically stable combined state as ore, into their elemental forms, which are usually less stable once they make contact with air or water. Corrosion in a refrigerant plant is largely confined to the liquid side of the heat exchangers, mainly to the condensers, which have the highest temperatures and the worst water quality. However, corrosion on the refrigerant side and in the evaporator can occur as well.
121
Pitting and crevice corrosion are basically the same phenomena. Conditions that favour this corrosion are: An oxidising environment, e.g. presence of oxygen, nitric acid or the Cu2+, or the Fe3+ ions. Presence of chlorine ions, even in low concentration. A weak spot on the passive surface film where the chlorine ion can penetrate. This could be a crevice between the gasket and the tightening surface of a flange, a badly executed weld or an overlap joint. . Pitting occurs on an open surface, below a spot covered by slag remnants, scaling, or a bubble. The penetrated spot has to be low in the oxidizing component. Low pH. At pH>12 the danger of localised corrosion is negligible (instead the danger of scaling will be high). Stagnant liquid. This is an important requirement. It would for instance be possible to use stainless steel in sea water, if a constant water stream always swept all surfaces in contact with the sea water. Normally there is always some stagnant spot, and when the heat exchanger is closed down all the surfaces would be susceptible to pitting. High temperature. In an open system (atmospheric pressure) and if oxygen is the only oxidising component, a high temperature (> 80 C) reduces the oxygen content, thus reducing the risk. If these conditions occur, then a local galvanic element forms with the penetrated area as an anode and the surrounding passive area as the cathode. As the cathode is large and the anode small the current density will be high in the anode, which then dissolves rapidly - several mm/day. As many different factors can contribute to crevice corrosion, it is very unpredictable and potentially dangerous. Since the right conditions are more likely to exist in a crevice than on an open surface. Stress corrosion is similar to pitting & crevice corrosion. A surface under tensile stress in a chloride solution forms the anode. The temperature has to be higher than that normally found in refrigeration plants or heat pumps.
6. Fouling & Corrosion. 5.2.6. Prevention of corrosion on mild steel and zinc by copper.
Galvanic corrosion can be prevented by electrically insulating the BPHE from surrounding equipment and/or galvanised iron or carbon steel should never be installed downstream of the BPHE unit. Micro galvanic corrosion cannot be prevented by electrically insulating the equipment. Here the concentration of corrosive compounds has to be within the limits specified in table 02. However, it does not happen on equipment upstream of the BPHE unit. Warning 8!! In a closed loop system, the concepts of upstream and downstream have of course no meaning, i.e. galvanic corrosion can occur on equipment placed immediately before the BPHE unit. However, there must be electrical contact for corrosion to occur.
123
100 C
Corrosion
pH > 7
50
The information here is for guidance only and not for warranties.
No corrosion
AISI 304 0 AISI 316
Fig. 04. The effect of chloride ion concentration and temperature on pitting and crevice corrosion tendencies on stainless steel.
Other factors which influence these corrosion types are: pH value. Water velocity. Fouling, both scaling and biological. Impurities, such as CO2, H2S, SO2, native iron etc. Surface condition. Redox potential of the solution, which depends on oxidizing compounds, i.e. oxygen, hypochlorous acids, sodium bichromate, etc.
Cl 2
Compounds such as hydrogen sulphide corrode copper and impairs the corrosion resistance of stainless steels. See also 5.2.4.
Cl 2
Table 02. Concentration limits of impurities in water to prevent microgalvanic corrosion of unalloyed or galvanized iron in the adjacent piping. See also 5.2.5.
pH SO42HCO3-/SO42Total hardness ClPO43NH3 Free Chlorine Fe3+ Mn2+ CO2 H2S Temperature Oxygen content 7.5 < > 4.0 < < < < < < < < < < 9.0 100 1.0 8.5 50 2.0 0.5 0.5 0.5 0.05 10 50 65 0.1 ppm dH ppm ppm ppm ppm ppm ppm ppm
ppb Note!
C ppm
125
Unfortunately, water is a living substance and the concentration of corrosive substances might change with season, weather (rain storms, dry spells), flooding etc. Several water analyses at different seasons and different meteorological conditions are necessary. The value of a water analysis made after a suspected corrosion leakage has occurred is also doubtful, especially if the unit has been in operation for many years. A single cleaning with a corrosive liquid could have corroded the unit but did not actually do it. Several years later, a further cleaning, leading to a temporarily lower pH and/or higher oxygen content or the like, could start the corrosion again but now the surface is penetrated. That the concentrations limits have been exceeded do not show up in a water analysis several years later.
Material defects. Most defects are discovered during pressure or helium tests of the BPHE, but some, e.g. slag inclusion in the plate, might turn up later. Manufacturing defects. As with material defects, most are detected at the testing of the BPHE. Sometimes a crack caused in the pressing might be covered with copper during the brazing and show up later, mostly in a corrosive environment. The brazing might be incomplete and, most likely in combination with some other factor, might be the cause of leakage. Exceeding the test pressure and/or the design temperature. This is rare though, as the BPHE will stand a pressure of at least five times the design pressure and the melting point of copper is 1083 C.
6. Leakage in a BPHE.
6.1. Causes of leaks.
Corrosion. This can occur on the entire surface, but tends to predominate on the lower part, close to the porthole. This is because some liquid can remain in the porthole of a closed-down unit and corrosive compounds then concentrate here. Corrosion in a BPHE is usually pitting/crevice corrosion on the stainless steel. Freezing. This occurs on the water side at a point where the wall temperature is lower than the freezing point. The point with the lowest temperature in a DX evaporator is normally some distance from the refrigerant inlet, but freeze damage normally occurs on the water side close to the refrigerant inlet, as the conditions for forming a closed-off space are greater here. Excessive torque and/or forces on the connections. The connections are kept in place by the double action of being expanded into the frame plate and brazed to it. However, excessive force can cause ruptures. Vibrations from the surrounding equipment. This can cause fatigue in the material, therefore with rupture of the plates and/or brazing. The problem will most likely occur in combination, e.g. with excessive forces on the connections. Water hammer or other pressure shocks. The amount of water flowing in a circuit represents a considerable kinetic energy. If the movement is suddenly stopped, e.g. by a fast-closing valve, the energy has to be captured somewhere. This may result in deformation and rupture of the confining surfaces. In a BPHE, the most likely damage is a deformation and, after repeated shocks, rupture of the end plate opposite the water inlet nozzle. Thus, a valve in the water pipe after the HE must never close rapidly. There is less danger on the refrigerant side since the vapour phase is compressible and absorbs the shocks. Thermal shocks. If cold water suddenly enters a very hot BPHE or vice versa, the thermal stresses between the first channel plate and the end plate might cause ruptures, as the thin channel plates expand/contract more rapidly than the end plate.
126
Does the water contain excessive amounts of chlorine ions or other corrosive compounds? Try to get a water sample. Be aware that water in alimentary installations is often chlorinated. As the chlorine is consumed, more is added, and soon there is an excessive amount of chlorine ions. This could lead to pitting.
127
Air
B
h h 2 C
B'
1 C'
B. Cutting a BPHE.
A'
Once the position of the leak has been determined, the BPHE has to be cut up for a closer examination. The first cut should be well away from the suspected place of the leakage, usually A-A' or the lowest B-B'. Finding a leakage by cutting is difficult. A too close cut could destroy the leak, a too wide (> ~10 mm) makes it impossible to see the leak. Check the cuts for damages. Freezing gives large de- formations, other types of damages maybe none.
Water
Water + acid
Liquid hammering
Normal channels
Do not dispose of nitric acid in the wastewater drain. Neutralize it first with a basic compound. In the absence of this, calcium carbonate stone, e.g. marble, is an excellent substitute. Ammonia solution is also suitable. See the next point, though. Before neutralizing the acid, dilute it with water. Pour the acid into the water, not the contrary. This makes the subsequent neutralization less violent. Neutralize with ammonia till the solution smells faintly ammoniac. Use protective gloves and glasses. Nitric acid leaves yellow stains on the skin and makes it leathery. After some days the affected skin will peel off.
Thermal shocks with displacement of the plates and as a consequence, ruptures of the brazing and the plates.
Frozen channels
128
7. Freeze.
2. Anti-freeze designs.
Non-refrigerant BPHEs. Because of the corrugations, there will be a space between the channel plate and the front cover. On non-refrigerant BPHEs, obsolete refrigeration BPHEs and some nickel BPHEs this space is open to the atmosphere. If the corrugation angle points downwards, water can collect in the space, freeze and possibly destroy the BPHE. See figures 01C and 3. Design & Installation , figure 02. Refrigerant BPHEs. A flat seal plate with a folded down edge is inserted between the front cover and the first plate to seal off the open space, see D&I, and figure 01. Note that a seal plate is not needed at the end plate; the folded down edge of the plate effectively seals off the space. Gasketed PHEs or the gasket side of SWPHEs, are resistant to damages due to freezing. The gaskets expand and the worst damage is some blown out gaskets. Tests have been made with a SWPHE where the evaporation temperature was reduced to -20 C with the water running. When the unit was entirely frozen, with no water flow, the evaporation temperature was increased until the full water flow was restored. No damage was detected. AWPHEs have basically the same freezing properties as BPHEs and the same advices apply.
3. Anti-freeze installations.
Figures 01 A - F contain some suggestions. A BPHE installed outdoors is subjected to the ambient winter temperatures, which might be below zero. If the BPHE is not used during this period or used only on & off, it should be drained during the off periods. The easiest way of doing this is to enter the water at the upper connection and let it leave at free flow from the lower connection, see figure 11 in 3. Design & Installation. Providing, of course that the thermal duty permits such an arrangement.
7. Freeze.
129
OK
A. Avoid positions for which water cannot be drained. B. Neither trap water in the BPHE between closed valves nor use downward loops. Except for the specially design oil traps in some refrigerant lines (see 1. O & R.), neither refrigerant nor liquid lines should have downwards pointing loops.
Note! Never install a valve with the drive mechanism at the bottom. Dirt and debris might collect there and impede the spindle movements.
C. BPHE without seal plate. For the following BPHE: CB14 Old models Non refrigeration BPHEs Some nickel BPHEs do either Corrugations of the first plate should point upward. or Seal the open space with silicon rubber. D. Low ambient temperatures. HEs installed outdoors, e.g. in heat pump should be protected from low ambient temperatures by: Insulation of the unit. Installation of a small electric heater. Draining of the unit when not in operation. Keeping the water always running.
F. Faulty insulation. E. Ice plugs. An ice plug and a valve can form a confined space. Note that an ice plug will be very solidly stuck in a BPHE. Damaged insulation or absence of insulation could provoke freezing. Especially dangerous is freezing on two different locations as the two ice plugs might create an enclosed water body. This will certainly rupture the unit as the freezing proceeds.
130
7. Freeze.
Moreover, there might be an air cushion at the entrance port. This air cushion causes maldistribution of the water and creates channels with low water flow or even channels with no running water at all. If the evaporation temperature is even slightly below zero, the water in these channels might freeze. This air pocket has to be vented. See figure 11 in 3. Design & Installation for some ideas. In evaporators operating close to or below freezing point, the water flow should be well balanced, both between the channels in a unit and between parallel units. See figure 11 in 3. Design & Installation.
4.4. Multi-pass.
Multi-pass on the water side is an effective way of increasing the water side wall temperatures. It is mainly used for a very long temperature drop of the water, which has to be cooled close to the evaporation temperature. A typical temperature program might look like: Water Refr. 28 -> -1 <1 -1 C C
Such a temperature programme is close to impossible in a single pass unit. The water pressure drop will be very low, too low, leading to unstable flow and oscillations. The low K-value and the low shear forces push the water wall temperature very close to the evaporation temperature. As the main use of this arrangement is to cool water close to zero with refrigerant which is sometimes far below zero it will be dealt with here even though it can be used for any temperature level. The water flow can be arranged as a U or an inverted U. U-arrangement, figure 03A. The temperature program in the passes would look something like: Pass Water Refr. 1, counter current 25 -> 6 C -1 <- -1 C 2, cocurrent. 6 -> 1 C -1 -> -1 C
The effect on the evaporator depends on if this is a direct expansion or a flooded flow evaporator. DX-evaporator. There is simply no room for any superheat in the second pass. The exit vapour is a mixture of much superheated vapour from the first pass and wet from the second, a not very suitable arrangement. Flooded flow. The entering refrigerant meets the same temperature difference in the two cases, which is an advantage as the boiling then starts simultaneously. The exit quality can vary very much but this has less importance in flooded flow. Inverted U-arrangement, figure 03B. The temperature program in the passes would look something like: Pass Water Refr. 1, cocurrent 25 -1 -> -> 5 4 C C 2, counter current. 5 4 -> <1 C -1 C
DX-evaporator. There is enough temperature difference in both passes at the end of the evaporation to secure a sufficient superheat. The middle water temperature, 5 C, defines the maximum superheat in both passes and there can be no large variations, e.g. at capacity changes and a DX evaporator is suitable. The entrance temperature difference is a little low at the second pass but as the refrigerant is already partly vaporized, it does not matter. Flooded evaporator. The low inlet temperature difference at the second pass might be too low to start the boiling, which makes this design unsuitable for a flooded evaporator.
7. Freeze.
131
Case study. A chicken slaughter plant in Portugal needed cool fresh water to wash and cool the slaughtered chickens. The entrance temperature of the water was a high 28 C in the summer, which had to be cooled to 1 C. As refrigerant, ammonia of -7 C was available from a LP receiver. This temperature could not be changed as many other evaporators were connected to the LR. A study showed that a flooded, single-pass WPHE would do the duty, but as the water pressure drop was very low, the water wall temperature was only slightly above -6 C. Though freezing would not damage the WPHE, it would act as a type of fouling with obstructed channels and therefore with a possible capacity decrease. The solution was a flooded WPHE in two passes on the water side, connected as in 4A. This is effectively two heat exchangers, in series on the water side, parallel on the ammonia side. The design criteria were: The pressure drop on the ammonia side should be the same in both sections. The K-value margins should be the same as well. The result was am M10BW with H-channels in the first section and L-channels in the second, a somewhat unusual combination. (The normal combinations are L, L & M, M & H and H.) It performed as expected with no operating problems except an occasional cleaning on the water side.
Note! The SWPHE is probably the only type of heat exchanger, which can be used for such an application. The low ammonia temperature means that freezing occurs if the water flow stops due to unforeseen circumstances.
132
7. Freeze.
The closing criterion. The compressor is normally shut down when a given pressure is reached. As seen from the above, this is no absolute proof that a liquid phase is not present. Far better would be to use the superheat at the lowest point or the point where liquid refrigerant is likely to remain the longest. As soon as there is a superheat there can be no liquid phase present and this should be the correct criterion for shutting down the compressor. Recommendations at pump-down. The water should always be running.
The compressors should run with as low a capacity as possible, i.e. the pump-down time is maximised. The solenoid & DX valves should be as close to the evaporator as possible, to minimize the liquid content. If possible, stop the compressors when a superheat is reached. This can be done automatically or manually. Record the superheat as function of the pump-down pressure and use the pressure at which there is a noticeable superheat as closing pressure. Alternatives to pump-down. Pump-down is not recommended or necessary at BPHEs. The refrigerant volume is small enough that the liquid will evaporate by the heat remaining in the BPHE for a temperature decrease of only 1 to 2 K. Install a bypass line between the compressor exit and the evaporator inlet. This is the same arrangement used to prevent too low temperatures at startup; see figures 04 L& M. The solenoid valve in the bypass opens when the refrigerant feed solenoid valve closes. The hot vapour from the compressor evaporates the refrigerant in the evaporator and maintains the temperature. The compressor runs for a determined number of seconds after the solenoid valve has closed, i.e. a sort of "semi pump-down".
The sequence at pump-down. 1. The solenoid valve before the expansion valve closes. The signal to close the valve is normally a decreased load, i.e. a reduced water temperature. See figure 04 F. 2. The compressor continues to operate and evacuates the refrigerant in the closed-off evaporator, thereby decreasing the pressure and the temperature. The relationship between the temperature and the liquid content in the evaporator is dependent on whether there is a heat transfer to the refrigerant during the pump down. No heat transfer. The pressure and temperature can theoretically reduce to absolute zero with liquid refrigerant still present. This means that the pump-down can be made to any given temperature and there might still be a risk that liquid refrigerant is present. Heat transfer. The temperature might decrease, remain constant or increase depending on the rate of the heat transfer to the suction power of the compressor. Thus, if the heat transfer is large enough, the pumpdown could be made with an increase of the temperature without a liquid phase being present. In praxis there is always a heat transfer to the refrigerant, and the liquid phase has disappeared at some point not too far below the starting temperature, but there is no temperature at which it is possible to be certain if liquid refrigerant is present or not. 3. When the pressure has reached the low set point, the LP pressostat stops the compressor, followed a little later by the water pump, see figure 04 G
7. Freeze.
133
Time relay
Note! An ice layer could mean that the exit temperature increases.
134
135
One phase
Refrigerant phase
Two phases
A. The oil is lighter than the refrigerant (R 502). Draining from the top, close to the interliquid surface.
Pure oil
Pure refrigerant
B. The oil is heavier than the refrigerant (ammonia). Draining from the bottom, preferably from a low point.
Oil-refrigerant solution
C. Completely dissolved oil. Draining is not possible (or necessary). Some liquid is sometimes taken out for evaporation of the refrigerant, thus concentration of the oil.
136
The behaviour of the film reflects the oil viscosity. A low viscosity oil flows at times as a mist in the channels, at times as a liquid layer on the wall. High viscosity oil might stick to the surface as a reasonably quiescent oil layer. An important but largely unstudied factor is the effect the surface tension has on oil fouling. Insoluble oil fouls the entire heating surface, in a DX or a flooded system. Soluble oil fouls only the superheating part of a DX evaporator. The only system that is not fouled by oil is a flooded system with soluble oil. If severe oil problems are expected, this might be an option, despite the need for an oil evaporator, see 1. Applications, 16. At low channel flow rates in a DX evaporator, the shear forces might not be large enough to carry the oil droplets upward. The oil - soluble or not - then collects in the evaporator. It might stick onto the surface or remain as an oil slug in the evaporator. On-off control maintains the vapour velocity. Consider hot gas by-pass control as a way of controlling the capacity in DX evaporators. 4. Evaporators & Separators, figure 24 shows one example. There are numerous variants and the valve manufacturers have specialized valves for this. Hot gas by-pass control has one advantage. When the refrigerant flow decreases because of decreased capacity, the missing refrigerant flow is replaced by compressor discharge gas. The refrigerant flow rate remains reasonably constant. A second effect of hot gas by-pass control is the increased temperature. This helps flush away the oil. A simpler variant of the above is an injection of hot discharge gas by a solenoid valve. This can be used as an alternative to a pump-down as well. See 7. Freeze, figure 04, L & M. Note this is not a capacity control, only a way to flush out oil and refrigerant. A third possibility is to heat up the oil with the warm side fluid, thus lowering the viscosity. The best way is not to let the oil enter the evaporator. Primarily a good oil separator is necessary if oil problems are expected. Secondly, a thermosiphon evaporator needs an oil drain system, see 1. Applications, figure 10 and 4. Evaporators & Separators, figure 07. A DX-evaporator normally does not need an oil drain. The exception is ammonia and insoluble oil otherwise, oil slugs could cause problems in the TEV. The drain should be in the HP receiver, i.e. before the oil enters the TEV. It could be arranged as an LP oil drain. The parameter most important to carry the oil through the evaporator is the wall shear stress or the more easily calculated pressure drop per meter flow length. The simple vapour velocity is misleading, as the hydraulic diameter and channel shape as well as the physical properties, specially the viscosity; of the oil and refrigerant have very significant effects.
A plate heat exchanger can carry oil with it for a velocity, which is about ten times lower than the velocity in the vertical connecting pipe. There is no absolute lower limit of the pressure drop per meter, but try to obtain a value larger than 5 kPa per meter channel length (port centre to port centre). Observe that a very large pressure drop means a high inlet temperature, and this decreases the effective temperature difference. In case of expected oil problems, try to make the temperature difference as large as possible, as this decreases the number of plates (channels) and thus increases the pressure drop on both sides. A high pressure drop on the refrigerant side means a high pressure drop on the liquid side as well. If this is not acceptable, a bypass on the liquid side could help. The BPHE with distributors improves the oil flow at DX evaporation, 4. Evaporators & Separators, figure 18. WPHE needs a distributor pipe with a number of holes with increasing diameter that is inserted into the port, figure 06. Heavy oil is drained from the inlet port of WPHE at DXevaporation, see figure 06.
137
Insoluble oil
General: There is practically only one system of this
Soluble oil
General: The oil impairs the physical properties of the
type, the soon to be phased out R 502/mineral oil. (R 22 at low temperatures has one phase oil in refrigerant and one phase refrigerant in oil.) An oil film could develop and fouling due to the oil might be a problem especially at very low temperatures. It could improve condensing heat transfer if dropwise condensation occurs. See 5. Condensers & Liquid Receivers, 1.2. The film could block the nucleation sites and impair the boiling heat transfer. See 4. Evaporators & Separators, 1.2.
refrigerant-oil solution, notably the viscosity and thus lowers the heat transfer coefficient. A little oil (< 1 %) could even improve the boiling heat transfer. An oil film cannot develop and fouling due to the oil is therefore no problem in the liquid phase.
DX-evaporator: At the end of the evaporation, when all
enters the channels and leaves with the refrigerant. Large amount of high viscosity oil (low temperature) could impede the proper function of the TEV. Then an oil separator and occasional flushing with hot gas might help (4. Evaporators & Separators, Fig. 23).
Flooded flow is not recommended as oil flows on the
refrigerant has evaporated, oil might get stuck on the heating surface: This is normally only a problem at low temperatures (< -20 C) and for high viscosity oil, especially for R22, as the oil is only partly soluble at low tempertures. Oil might then collect at the exit and flow back into the evaporator. The newer BPHEs with distributors improve the oil flow, see 4. Evaporators & Separators, Fig. 18., for WPHEs see also Fig. 06 here. Use a good oil separator and maybe hot gas by-pass control, see 4. Evaporators & Separators, Fig. 23 or 7. Freeze Fig. 04, L & M.
Flooded flow: As the liquid refrigerant washes all the
surface and is difficult to recover. In case of R22, an effective oil evaporator (1. Applications, 16) impedes formation of the light oil rich phase and the exit to the oil evaporator could then be from anywhere.
Condenser: No particular problems if the condensate
surfaces, oil fouling is no problem, but if no precaution is taken, oil concentrates in the system and drains the oil from the compressor. An oil evaporator is thus necessary, see 1. Applications, 16.
Condenser: No particular problems. The oil moves with
drains from the condenser. The oil then moves with the condensate. However, if there is a condensate level as in 5. Condensers & Liquid Receivers, figures 04, 05 & 11A, oil collects on the top of the condensate and impedes the heat transfer.
the condensate.
important one, ammonia/mineral based oils or synthetic polyalfa-olefin (PAO) oils. See above on the effect of an oil film.
The two commercial systems are ammonia / synthetic oils type polyalkylene glycol (PAG) and hydrocarbons (propane/propene/butane) / mineral oils. Note! The in ammonia soluble polyol ester oils (POE) react vigorously with ammonia.
DX-evaporator: The difference to HCF systems is that
inlet port and drain from there Fig. 06. The oil could also be drained from the HP receiver and only the remaining small amount from the inlet port. A good oil separator is necessary and maybe flushing by hot gas as described in 4. Evaporators & Separators, Fig. 23, Hot gas by-pass ....
Flooded flow: With a good oil recovery system, where
ammonia is normally used in larger evaporators, such as welded and semi-welded PHEs. The larger port holes in these can cause a maldistribution of the ammonia over the channels. A specially design distributor pipe is then installed in the porthole, see figure 06. Hydrocarbon/mineral oils are used in the same BPHE as the HFCs and have the same problems, see above.
Flooded flow: See above. In the case of ammonia it
oil drains from the lowest point of the evaporatorseparator loop, fouling due to oil will not be a problem. 1. Applications, Fig. 10 and 4. Evaporators & Separators, Fig. 03 & 07 and 3.4.7 & 3.6. Note! Some oils (polyalfaolefins) could be difficult to drain from the channels. Further experience is needed.
Condenser: No special problems. The oil moves with
the condensate. Condensate level should not pose any problems as the heavier oil drains from the bottom.
138
When the oil level is high enough, a float opens a valve and the discharge pressure forces the oil to return to the compressor inlet. Some oil separators can return more than 99 % of the oil to the compressor, but normally it is in the order of 80 %. An oil separator should be used whenever there are problems with oil retention in the evaporator. This will most likely be more severe the lower the temperature is. Apart from removing oil, an oil separator also acts as an efficient muffler, thus reducing noise. Take care if an oil separator is installed in an old system. The separated oil, which was previously distributed in the entire system, could flood the compressor.
Energy changes in the BPHE. Oil has about half the specific heat of, and a lower density than, water, i.e. the relative change of energy is lower if oil enters than if water enters. In case of an evaporating refrigerant, the relative energy change might be still larger as there is a phase change and thus a transfer of latent heat. Heat transfer. Oil normally has several times higher viscosity than the cooling medium, either refrigerant or water. Furthermore, the heat conductivity is much higher for water than for oil. Oil has about the same heat conductivity as refrigerants, but for the latter there is the additional heat transfer of the evaporation. Thus, there is not only a larger energy change for a sudden change of the cooling medium temperature, but it is also more easily transferred to the plates, i.e. with danger of rupture of the plates. The conclusion is thus that sudden temperature shocks such as from on-off control should be avoided, especially on the cooling fluid side.
2. Oil cooling.
2.1. Temperature shocks at on-off cooling.
If the compression ratio is too high, the exit temperature, including the oil temperature, might be too high as well. Some compressor types, notably screw compressors, heat the oil excessively. Certain refrigerants, e.g. ammonia, superheat more than others during the compression. A too high oil temperature could lead to a breakdown of the oil. Thus, oil cooling is necessary. This oil cooling can be done in a BPHE. However, if the heat exchanger is filled with hot oil and the cooling fluid water or refrigerant - suddenly enters, e.g. when an on -off controller is actuated, this could lead to thermal shocks and ruptures, usually between the end plate and the first heat transfer plate. In case of a S&THE, tubes could loosen at the tube plate. As BPHEs are effective oil coolers compared to other types of HEs, thus economically advantageous, some precautions have to be taken to avoid thermal shocks. Consider a BPHE working with hot oil on one side and cooling media on the other. The system has been brought gradually to the steady operating state, i.e. no thermal shocks have occurred. The viscosity of the oil entering the oil cooler is usually less than specified on the manufacturers data sheet as it contains dissolved oil. This can lead to an oversized oilcooler if the design is based on the nominal oil properties. An oversized oil cooler rapidly cools the oil and in case of an on-off control system it will probably be operating in many short cycles instead of a few longer ones. This lead to more temperature cycles and thermal fatigue. Beware of oversizing an oil cooler. What gives the largest shock? A sudden change of the oil temperature or a sudden change of the cooling media (Water or refrigerant) temperature? Study the two theoretical extreme cases: On-off control of the oil side (figure 08 B) or of the cooling medium side (figure 08 E). When the oil temperature has reached its upper limit, either oil or cooling fluid suddenly enters the BPHE. What is the effect on the BPHE when the oil or cooling fluid has just replaced the previous content of the channels? There are two effects to be considered, the change of energy and the effect on the heat transfer coefficients:
139
Oil pot
Heating
The distributor pipe should have a number of holes with increasing diameter from the inlet to the end. The holes should not point downward in order not to agitate the oil.
Entrance from above to a common main, in order to prevent back flow to Idle evaporators.
The condensate could flow back from the bottom inlet in this two-pass condenser. Loop to prevent liquid refrigerant and oil to flow back to the compressor from a higher placed condenser. Entrance from above to a common main, in order to prevent back flow to idle condensers.
Condensers Oil traps These should be the only downward pointing loops in a refrigeration plant.
140
Refrigerant temperature level. From where shall the liquid refrigerant be drawn; the HP or the LP receiver? A flooded evaporator as oil cooler could be added to the LP separator. However, the refrigerant could be very cold here, leading to large thermal shocks. The large temperature difference most likely makes the HE thermally overdimensioned and the oil will be cooled too much. Refrigerant from the HP receiver is usually of a temperature of 30 to 45 C, suitable for the oil, which normally is 70 to 90 C. If a direct expansion system is used (figure 09) and the evaporation temperature is close to the receiver temperature, there is no flashing of the refrigerant in the expansion valve; it simply ensures that no more refrigerant enters in the oil cooler than what can evaporate. A flooded evaporator could be connected to the HP receiver. The vapour produced is normally recondensed in the condenser. See the discussion in 5. Condensers and Liquid Receivers, figure 06, of the layout of the HP receivers. As a HP receiver is not a separator, the refrigerant should be evaporated as much as possible, e.g. controlled by the superheat through an expansion valve (but no flashing).
plant, PHEs benefit from a wealth of design and operating experience in refineries, chemical process industries, and in mechanical, food and marine application as well as for all types of duties such as condensation, evaporation, single-phase, viscous media, vapour with inert gases, etc.
3.2. The differences between the new and the old refrigerants.
All freon type refrigerants, are halogenated hydrocarbons. The halogenes are mainly chlorine and fluorine; for some special applications bromine. What then, are the differences between the new and the old refrigerants? The old refrigerants are mainly fully halogenated, i.e. the molecules contain only carbon and halogenes. They are denominated CFCs (chlorine, fluorine, carbon) and are in general: Good solvents, especially to mineral oils. Incombustible. Chemically stable, which is a bonus for the operating properties, a drawback for the environment. The chlorine gives good lubricating properties, especially to aluminium. Not hygroscopic. Inert to generally used construction materials. The chlorine reacts with atmospheric ozone, leading to the ozone depletion. Contribute to the greenhouse effect. The addition of a hydrogen atom converts the CFCs to HCFCs and changes the properties: Breaks downs easier, which decreases the ozone depletion potential considerably. Many hydrogen atoms in the molecule increase the flammability. Removal of the chlorine atom converts the HCFCs to HFCs and changes the properties: No more ozone depletion potential. Decreased greenhouse effect. Decreased solubility of mineral oils. Decreased lubricating properties. In many cases more hygroscopic. Special oils are necessary which brings about new problems, see 2 and 8. Fouling & Corrosion. Removal of the fluorine atom converts the HFCs to normal hydrocarbons, HCs, actually old refrigerants: No more greenhouse effect, nor ozone depletion. Flammable. Good solubility to mineral oils. Not hygroscopic. The system R134a/R12 could serve as an example what happens when a CFC is replaced by a HFC. R134A dissolves nitrogen 10 times better than R12. If R134A has been in contact with air, it might contain nitrogen, which could make trouble in the condenser.
2.5. Conclusion.
Rapid changes especially of the cooling medium temperature have to be avoided. Especially dangerous is on-off control, see figure 8 A & B. Try always to have continuous flows, however small, on both sides or attenuate the temperature changes. The range between the high and low temperature settings at on-off control should be as small as possible. This eliminates large temperature variations. If solenoid valves with slow opening and closing times are available, temperature shocks could be reduced; figure 08 C. Motor valves are excellent but expensive. Do not oversize the oil cooler. This is easily done if the pure oil viscosity is used instead of the mixture viscosity, which is considerably lower. Avoid valves which control a variable by changing the opening time, e.g. open 19 seconds, closed 1, changing this to open 1, closed 19 seconds as the load varies. This causes both thermal and mechanical shocks.
3. Refrigerants.
3.1. General.
Over recent years, the number of generally used refrigerants has increased from less than ten to multiples of ten. A number of different refrigerants together with the appropriate oils are thus under evaluation. Some of these combinations will prove to be unsuitable for various reasons and the large majority of these will never be commercially viable. In some years there will be a new set of environmentally friendly refrigerants, probably only between five and ten. Fortunately, plate heat exchangers are uncritical. In contrast to most other equipment in a cooling or refrigeration
141
Oil to compressor
The control system opens the solenoid valve, when the oil temperature exceeds the set point. It also maintains a constant temperature difference between the evaporation and oil inlet temperatures by controlling the evaporation pressure. Temperature shocks are thus minimized. In this way oil shocks are minimize at start-up of the individual compressors.
Control box
142
134A itself is more hygroscopic than R12. R134A in combination with water can lead to micro galvanic corrosion with copper, see 6. Fouling and Corrosion. Dry R134A is safe. As R134A is a smaller molecule than R12, desiccators (molecular sieves) for R12 might not work. If R 12 remains after a switch to R 134A, R 134A and R 12 form a high-pressure azeotrop, which has a high potential for damaging the plant.
C) Uncertain physical properties. There are sometimes deviations in the stated physical properties between different manufacturers (and from reality), especially for refrigerants with glide (see below). The deviations are usually small but in some cases, 2 C deviations in the saturation temperature have been observed. D) Composition changes. The most common example is probably water in ammonia. Water decreases the vapour pressure of ammonia. Examples that are more recent are composition changes in refrigerants with glides (see below) and composition changes because of an incorrectly executed drop-in replacement, resulting in a mixture of the old and new refrigerant. The deviation between the ideal and the real vapour pressure curves affects both the condensation and evaporation pressure and the superheat at an evaporator. E) Temperature glide. There is no constant boiling or condensation temperature, but the two-phase change occurs over a temperature range, the glide. This is treated extensively below.
3.3.1. Fouling.
As described in 1 and in the chapter 6. Fouling & Corrosion there might be a danger of fouling on the refrigerant side in the evaporator, most likely of tar like products and wear and tear products from the compressor. This type of fouling will most likely be a fairly slow process, probably over many years, which also means that there is little experience on maintenance and cleaning methods. Possible fouling will be still more detrimental to compressors and control equipment than to the heat exchangers, which are why, unsuitable combinations will be phased out.
3.3.2. Inerts.
If the performance of a condenser gradually diminishes, one reason could be blocking by inert gases. Thermal decomposition of synthetic oils could result in hydrogen and carbon dioxide. Some refrigerant and oils could also dissolve ambient air. Venting every now and then helps.
143
Fig. 10. Vapour pressures of the pairs R 12/R 134a and R 502/R 507.
D) An evaporator where R 502 is replaced with R 507 has a pressure of 6.10 bar (R 502 at 2 C). At this pressure, R 507 has a temperature of -0.52 C. To get 2 C (6.60 bar) for R 507, the R 502 has to be set to 5.2 C, which corresponds to 6.60 bar. Comments: 1) If the controller in the R 502/R 507 cases is not adjusted, the evaporation temperature decreases, the MTD and thus the capacity of the evaporator increase, and i.e. there is no incentive to change the setting. The evaporation temperature in this case is now below zero, i.e. freezing danger. 2) The settings change with temperatures and refrigerant system. R 502/R 507 is a well-behaved pair; the change in the setting is practically constant over the entire temperature range. 3) R 12/R 134a need no change around 18 C, below that the temperature setting is increasingly lower and above, in the condensing range, increasingly higher. 4) The old pressure controller might be used but the control range has to be sufficient large. 5) Adjustment of a TEV is more complicated than shown. In the example, the valve gets a superheat signal of 6.5 K, when expected was 5 K. It then tries to close the valve. However, the pressure difference has increased from 9.63 - 3.3 = 6.33 bar to 10.16 - 3.15 = 7.01 bar. That would force more refrigerant through the TEV and decrease the superheat. The compression ratio for R12 is about 2.92. Depending on the type of compressor, the compression ratio might have to remain. The condensing pressure for R 134a is then 3.145 * 2.92 = 9.18 bar => 37.8 C. This lowers the capacity of both the TEV and the condenser. The capacity of a TEV is not only dependent on the pressure difference and absolute pressure, but also on the thermodynamic and transport properties of the refrigerant. A TEV with R 134a seems to give about 20 % more capacity than the same valve with R12, while R507 gives basically the same capacity as R 502. The valve manufacturer's recommendation should be followed when adjusting valves, especially TEVs, at a retrofit of a plant.
144
Contrary to what is generally believed, the vaporization does not follow an exact defined path, from the bubble to the dew point. It does, under assumption that the liquid and vapour is intensely mixed and moves at the same velocity. In other words, a certain quantity of liquid has to be present and mixed during the entire process. The same is valid for condensation. The above condition might not be true, especially not for some condenser types. If the liquid is removed as soon as it condenses, it does not participate in the equilibrium at the condensation any more. When the very last vapour condenses, it is in equilibrium with the condensate just formed and not with the entire condensate amount. This just formed condensate contains relatively more of the LBCs than the total amount of condensate and the effect is that the mixture has to be cooled down further before everything is condensed. When a boiling zeotrop is in contact with the heating surface in an evaporator, mainly the LBC evaporates at the liquid close to the wall. The temperature in this layer then increases and as a result, the temperature difference is less than the theoretically calculated. The practical effect is that the K-value seems to be lower. The same happens in condensers but the effect is less. These three last effects above are less pronounced in a PHE with its intense turbulence and good mixing.
3.6.5. Leakages.
A leakage of a zeotrop from a vessel with similar amounts of both vapour and liquid, thus with different compositions, means that the original composition changes. The temperature programmes in the evaporators and condensers then change with possible malfunction. The majority of the tests, which have been made, seems to agree that the system continues to perform as designed if the refrigerant is topped-up with the original zeotrop.
3.6.2. The evaporation and condensation temperatures for refrigerants with glides.
If both the condensation and evaporation occur over a temperature interval, then the question arises of the definition of evaporation and condensation temperatures. The evaporation temperature is the mean value between the inlet temperature and the dew point at the exit pressure. As the refrigerant usually enters partly vaporized, the glide in the evaporator is less than the difference between the dew and the bubble points. The condensation temperature is the mean value between the dew and the bubble points. The evaporation superheat is the difference between the exit temperature and the exit pressure dew point. The condensate subcooling is the difference between the exit temperature and the exit pressure bubble point. There are some problems with these definitions (as are with alternate definitions). See also figure 13. The exit temperature is not the sum of the evaporation temperature and the superheat. In figure 13, it is 9.6 C for R407c, not the sum of 2 C and 5 K as for R 22. It is easy to make the mistake of using a too low water inlet temperature, especially for TEV with large superheats. An advantage with the glide is the possibility to cool the water a little more. If the approach of a certain HE is 3 K an R22 evaporator operating with 2 C evaporation temperature cools the water to about 5 C while an R407c could cool the water to 3 C. Note the above, though. The inlet temperature to an evaporator might vary as the temperature to the TEV varies. Despite a constant pressure, the mean temperature will then vary as well. This will make comparisons to other systems difficult. Despite a positive evaporation temperature, the inlet temperature might be negative, i.e. freezing danger. It is easy to forget, that at a retrofit of R22 to R407c the inlet temperature might be below zero for a 2 C.
145
C When the first bubbles start to form in the liquid, the bubble point is reached.
At the end of the evaporation all liquid has evaporated. When the first droplets, dew, start to form on a surface at a subsequent cooling, the dew point is reached.
% evaporated
Inlet evaporator
% evaporated
100 C
A. Zeotrop.
100 C
B. Maximum azeotrop.
Z 1 X YY
Constant pressure
2 Liquid or bubble point curve 100 % LBC 100 % HBC 100 % LBC
B. Minimum azeotrop.
100 % HBC
Fig. 12. The equilibrium curve for a zeotrop and the two types of azeotrops.
The concentration of the high boiling component in a zeotrop is always lower in the vapour than in the liquid. At 60 C (line 1), the vapour concentration is given by (X), 40 %, and the liquid concentration by (Y), 70 %. A liquid of 70 % has thus a bubble point of 60 C. The dew point of this mixture (line 2) is at Z), about 75 C. The components of a zeotrop can be separated by a series of evaporations and condensations, a disitillation or a fractionation. The bubble and dew points curves of an azeotrop touch each other at one point. In this point the vapour and liquid compositions are equal. The boiling point is either higher than for the individual components or lower. For an azeotrop we can talk of a boiling point as the bubble and dew points are the same. The components in an azeotrop can not be separate by an evaporation and a subsequent condensation as the vapour concentration is the same as the liquid concentration. An azeotrop thus behaves like a single component liquid.
146
Concentration, ppm Effect 5 20 50 100 300 400-700 1700 2000 - 5000 7000 Some people pick up the smell. Most Very distinct smell. Unpleasant, inexperienced persons get anxious. Experienced people leave inexperienced panic. Strong irritations in eyes, nose & lungs. Impossible to stay. Coughs, cramps. 0.5 hr exposure can lead to severe injuries. May be lethal in less than 0.5 hr. Lethal within minutes.
Hydro- Water gen group group H2=2 He=4 CH4=16 NH3=17 H2O=18 HF= 20 Ne = 20
Air group C2H2=26, HCN = 27 B2H6,N2,CO,C2H4=28 Air= 29 NO, C2H6=30 N2H4, O2=32
The gas density is proportional to the molecular weight. There are only three gases lighter than ammonia and seven lighter than air. Six more are lighter but so close that they hardly separate when the gas enters an air body. The lightest refrigerant, propene C3H6, is considerably heavier than air. The HFCs are between R32 and R125.
3.8. Ammonia.
Water and ammonia have heat transfer properties, both for single and two-phase heat transfer, far superior to all other fluids. The combination of low viscosities, high thermal conductivities and high specific & latent heats makes these fluids to the best possible heat transfer fluids. Ammonia could have a somewhat higher density to make it a perfect refrigerant, but this is a minor deficiency. There are some be tter fluids, but condensed hydrogen and helium are not liquids in the normal sense. Ammonia has optimal properties in the important temperature range of -50 to +10 C. Ammonia has one drawback, it is perceived as dangerous. No doubt, it is toxic and flammable but these properties alone do not make it dangerous, provided it is handled in a correct way. In certain respects, it is less dangerous than other refrigerants. General. Ammonia, NH3, is a gas at normal temperatures. It dissolves easily in water forming a complex compound simplified as NH4OH. This is a strong base, i.e. it neutralizes acids, usually under a strong heat release. It is a polar compound, i.e. the molecule has a positively and a negatively charged part, but is electrically neutral. It is a good solvent, especially for other polar compounds and salts. At temperatures over 450 C it disassociates and reverts to nitrogen and hydrogen. This is used when brazing, as the mixture is reducing and hence removes oxides. Occurrence. Ammonia occurs naturally with all living materia because of decay of nitrogen containing organic material. A cow barn smells strongly of ammonia. It is as intrinsically connected to the circulation of nitrogen as carbon oxide is to the circulation of carbon. Production. Ammonia was first isolated and produced from condensation of ammonia containing gases from decaying organic material. It is now mainly produced from
147
A. With superheat.
Water
B. Without superheat.
Water
C 10
R 407c
Water R 22 MTD = 7.21 K MTD = 6.91 K MTD = 6.45 K MTD =7.50 K R 407c
Load
Load
Note that modern pressure vessels, manufactured and controlled according to the valid codes are high quality equipments and are very unlikely to suddenly rupture. A liquid ammonia spill on a floor evaporates slowly, considerably slower than a freon spill. This is due to the large latent heat of ammonia. Note, never flush liquid ammonia with water, see also Warning 6, 3.2.4. 6. Fouling & Corrosion. Ammonia vapour can safely be scrubbed with water, though. Noncondensables. Despite precautions, a refrigeration system can and will contain noncondensables (inerts). The chief effect of these will be in a decreased condenser capacity. The presence of inerts are sometimes difficult to establish and the venting can be difficult. Ammonia is different from all other refrigerants. As it is readily soluble in water, it is enough to let the pipe/hose with the inerts discharge into a bucket with water. If bubbles emerges to the surface it is air, ammonia dissolves. See 5. Condensation and Liquid Receivers, 5. Venting, p. 106. The large danger of ammonia does not come from ammonia in refrigeration plants, which is about 1 % of the world production. Ammonia is used as fertilizer and as raw material for the chemical industry and is transported by road, railway and ships and pumped between various vessels in very large quantities. Despite these risky operations, there are relatively few severe accidents. Water in ammonia. The ammonia is now transformed to a refrigerant with glide, see table 3 and 3.6. Conclusion. Ammonia can be risky, but correctly handled it is in many respect safer than other refrigerants.
148
Hot air, 80 C Ammonia vapour has to be cooled to the vicinity of -100 C to be heavier than air of + 25 C.
Ammonia, 25 C
The strong smell of ammonia gives an early warning. It burns, but with difficulty Ammonia would escape upwards.
People and air have one thing in common: Both are heavier than ammonia. Air
Air
A new application is condensing of carbon dioxide with evaporating ammonia, see below. In this condenser/evaporator the carbon dioxide pressure is considerably higher than the ammonia pressure. In case of a leak on the heating surface carbon dioxide enters the ammonia side. Carbon dioxide and ammonia will then immediately form ammonium carbamate, which is corrosive and could destroy a compressor. It is readily dissolved in water. Check with the compressor manufacturer if the vapourliquid separator in a thermosiphon system should be equipped with a demister as extra protection. Note that leakage between the two side is a very rare phenomena in a PHE. Physiological effects. Low concentrations stimulates breathing - the refreshing effect from carbonated beverages but in higher concentrations - ~ 20 % - it causes death, mainly by suffocation. Symptoms of carbon dioxide poisoning are nausea, vertigo, dizziness, headache, sight & hearing difficulties, thus general and insidious. However. the allowable concentration is in the order of 5000 ppm, several times higher than most HCFCs. Environmental effects. Zero ozone depletion potential but a fairly low green house effect, see below though. Manufacturing. Carbon dioxide is, contrary to all other refrigerants except water, not manufactured, it is recovered, mainly from flue gases, fermentation processes and limekilns. The environmental effects are thus more realistically described as zero. Carbon dioxide escaping from a refrigeration plant is recovered elsewhere and thus the net effect to the atmosphere is zero. Refrigeration properties. Carbon oxide is cheap enough to be used in once-through refrigeration. It is liquefied (solidified), evaporates at the refrigeration and the vapour is discarded to the atmosphere. The liberated carbon dioxide has the important secondary effect that it protects food from attacks by bacteria and fungi. Heat exchangers. No particular problems are expected. Insoluble oil could pose difficulties at low temperatures. Water dissolved in the carbon dioxide can freeze and block components. A good filter/dryer is mandatory.
149
Saturation temp. C
Bubble point 100% condensed 10% condensed 1% condensed 0% evaporated 90% evaporated 99% evaporated 0.01 0.1 1.0 % H2O 0.01 0.1 1.0 % H2O 0.01 0.1 1.0 % H2O 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.4 0.4 0.4 0.4 0.4 0.5 0.5 0.5 0.5 0.5 0.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 0.1 0.1 0.1 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.4 0.4 0.4 0.4 0.4 0.5 0.5 0.5 0.5 0.5 0.6 2.0 2.1 2.2 2.3 2.5 2.7 2.8 3.0 3.2 3.4 3.6 3.7 3.9 4.0 4.2 4.3 4.4 4.5 4.6 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.4 0.4 0.4 0.4 0.4 0.4 0.5 0.5 0.5 0.5 0.5 0.5 2.0 2.1 2.2 2.3 2.5 2.6 2.8 2.9 3.0 3.1 3.1 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 43.1 42.3 41.6 40.7 39.8 38.9 38.0 37.0 35.9 34.8 33.7 32.4 31.1 29.6 28.1 26.5 25.0 23.4 22.1
Dew point 0% condensed 100% evaporated 0.01 0.1 1.0 % H2O 15.2 13.8 12.1 10.5 8.7 6.9 5.2 3.9 3.0 2.5 2.2 2.0 1.9 1.8 1.8 1.7 1.7 1.6 1.5 32.2 30.7 29.0 27.3 25.4 23.4 21.4 19.5 17.5 15.6 13.8 12.1 10.5 9.1 7.9 6.8 5.8 5.1 4.6 52.7 52.0 51.4 50.4 49.4 48.3 47.1 45.8 44.4 43.0 41.5 39.9 38.2 36.6 34.9 33.2 31.5 29.8 28.2
Ammonia evaporates at -15 C and condenses carbon dioxide at -10 C. Part of the liquid carbon dioxide is transported, by a pump or natural circulation, to the HT applications. It evaporates here at -10 C. This part of the system is actually a heat pipe. Because of the virtual isothermal operation, the total temperature difference between the evaporating ammonia and the cooled air is maximized. There is no loss due to a temperature increase as the case of a brine. The other part of the liquid is expanded to -35 C and used for the LT applications and finally compressed again to -10 C, i.e. a normal refrigeration cycle. In case of an emergency shutdown, e.g. a power failure, the carbon oxide can simply be vented to the atmosphere. The design pressure can then be limited to a more usual 26 bar. See Hydraulic properties, p. 148. There are many refrigerants, which can be used in a cascade system, but only carbon oxide can be vented without any concern for the cost or environment.
150
-30 -20 4200 3001 Ammonium hydroxide (40%): 4257 Calcium chloride: 3589 Sodium chloride: 3877 Ethanol-water: 3500 Methanol-water: 3707 Ethylene glycol: 3484 Propylene glycol: 3601 Potassium formate: 3542 Potassium acetate: 3830 Silicon oil (typical): 1512 Freezing point Data at Water at 0 C: Ammonia (l): C: C:
Thermal cond. Comments -30 -20 0.59 0.62 0.53 0.51 0.48 0.35 0.41 0.36 0.34 0.48 0.29 0.12 Bad ######## Alcohol solutions are good secondary refrigerants but the low boiling points means that the alcohol could evaporate in case of open systems at elevated temperatures. At elevated temperature, they become flammable as well. All commonly used metals and gaskets are resistant. Potassium acetate (Tyfoxit, GS4) and potassium formate (Freezium, Hycool) - the salts of acetic and formic acids - or mixtures (Temper, Pekasol) are recent compounds. Some properties: Excellent thermal properties. Potassium formate has ten times lower viscosity at low temperatures than propylene glycol. Acetate is somewhat less advantageous. Environmentally acceptable as both are naturally occurring compounds. Acetic acid comes from oxidation of alcohol and formic acid was first isolated by distillation of ants. At higher concentrations, both are toxic but the toxic level for e.g. potassium formate is higher than for sodium chloride, table salt. They are thus good replacements for propylene glycol. Both are corrosive, formates somewhat more, and have to be inhibited. Various manufacturers have different proprietary solutions, such as borax, metasilicates, carboxylic acids, etc. Long time experience is necessary, but zinc and high zinc brass (> 15%) are not allowed. l. Some formate/acetate mixtures have a high pH, which can be irritating for the skin, while other can be foaming. A thorough cleaning of the system is necessary before the brine is charged. The oxygen should be evacuated by e.g. a nitrogen pressure test. The manufacturers recommendations should be strictly followed, otherwise corrosion can occur on copper as well. Plastic pipes should be avoided. Air diffuse through these and the system can then be difficult to deareate. Silicon oils are advantageous at low temperatures due low viscosity. They are chemically inert, thus not corrosive, neither for metals nor gaskets. They do not have, apart from the low viscosity, very good thermal properties, but for temperatures lower than -50 C, they are sometimes the only choice. Thermal oils are mainly used at high temperatures but some are used at low temperatures as well. They are not aggresive to metals (some are towards gaskets). They have poor thermal properties but their advantage is the very large temperature span they cover.
W/m, C
-50 -40
-50 -40 0.69 0.52 0.48 0.30 0.35 0.30 0.29 0.48 0.27 0.12 Excellent, but limited to above 0 C. Irritant, toxic, high pressure, flammable. Irritant. Not flammable, nor high pressure. Corrosive. Corrosive, limited to -20 C Low boiling point, flammable. Low boiling point, flammable. Toxic Somewhat toxic. Less toxic but high viscosity Long time corrosive properties? Corrosive for zinc. Expensive & ineffective, used at low temp.
3046 4224 3465 2853 3370 3035 3265 3165 3715 1500
Excellent ########
Medium ########
151
HT Display
HT cold store Primary NH3 refrigeration system tcond = -12 C tevap = -17 C
LT Display
LT CO2: -35 C
HT CO2: -12 C
Figure 16. Ammonia as the primary and carbon dioxide as the secondary refrigerant.
Carbon dioxide condenses in the ammonia evaporator at -12 C. The liquid carbon dioxide is used either directly in the HT displays or expanded to -35 C for the LT displays. The flows to the displays are normally controlled by the display temperature through an on-off thermostat. Depending on the distance the carbon dioxide is pumped or flows by natural circulation. Natural circulation in an HT unit is essentially a heat pipe, i.e. condensation and evaporation at the same pressure. Refrigerant Volumetric latent Liq. kJ/m3 coolin capacity Vap. kJ/m3 COP Liquid ( *Cp) kJ/( K, m3) Viscosity cP Thermal conductivity W/m, K Temperature at 25 bar, C Common data: R717 839463 1030 6.78 3037 0.262 0.671 58.2 CO2 267590 8137 6.29 2197 0.184 0.142 -12.0 Heat exchangers for carbone dioxide should be: able to withstand the HT pressure. For carbon dioxide of -12 C it is 25 bar = 24 barg. A safety valve releases at higher pressure than 25 barg. A lot of equipment is designed for 25 bar, thus 25 bar is a suitable pressure. equipped with an oil filter, as oil at low temperatures can block part of the heating surface. designed to avoid thermal shock in case of S&THEs or CBs in especially DX systems. See 1. Applications, 11, p. 12. WPHEs are normally uncritical. R22 274065 1220 6.95 1528 0.323 0.111 61.4 ####### ####### ####### R404a 198828 1413 6.82 1605 0.280 0.093 53.5 R410a 278502 1926 6.87 1796 0.254 0.135 41.4 R507a 195548 1483 7.01 1605 0.277 0.081 52.5 R134a 262606 677 7.01 1773 0.438 0.108 77.6
Condensation temperature. C Condensation pressure, bar Compressor efficiency, % Discharge temperature C Condensation capacity kW Superheat capacity. % COP Heating Condensation temperature. C Condensation pressure, bar Compressor efficiency, % Discharge temperature C Condensation capacity kW Superheat capacity. % COP Heating Common data: Comments:
Evaporation temperature: 2 C, Superheat: 5 K, Subcooling: 3 K * Entrance temperature below zero. ** Far too high.
152
9. Appendix I. Definitions.
Parallel flow:
1. Equipment.
HE: S&THE: SHE: UC: LC: ACC: PHE: GPHE: SWPHE: A heat exchanger in general. A shell and tube heat exchanger. A spiral heat exchanger. A unit cooler. Used to cool air. A liquid cooler. Used to cool a liquid by air. An air cooled condenser. A plate heat exchanger in general. A gasketed PHE. A semiwelded PHE. A PHE where two plates are welded together, forming a cassette. The refrigerant flows inside the cassette, the water in the gasketed space between the cassettes. An all welded PHE. A SWPHE or a AWPHE. A brazed PHE in general. A copper brazed PHE. A nickel brazed PHE. A liquid receiver. A valve used to expand and vaporize the liquid in direct expansion system. A thermostatic expansion valve. An electronic expansion valve. Direct expansion.
The inlet and exit of the flows are on the same side of the PHE for each flow. The inlet and exit of the flows are diagonally placed on the PHE for each flow.
Diagonal flow:
AWPHE: WPHE: BPHE: CB: NB: LR: EV: TEV: EEV: DX:
The cold end temperature difference. (U-value). The overall heat transfer coC. efficient. Measured in W/m2, Ks , Ki Coefficients used in Appendix II. MTD, tm K: Mean Temperature Difference. Logarithmic Mean Temperature DifferLMTD, tlm,K: ence. See below for a closer definition of MTD, LMTD and K-value or Theta: Thermal length. t1 / MTD or t2 / MTD. gives the temperature change of a fluid pro degree temperature difference. TTL: Terminal Thermal Length. t1/tw, t2/tw, t1/tc or t2/tc. q, kW: The heat transferred up to a given point. Q, kW: Total heat transferred, duty or capacity. t, C: Temperature. T, K: Absolute temperature. T = t + 273.16. Tc,K: Critical temperature. tb, C, Tb, K: Normal boiling point. Tr: Reduced temperature = T / Tc p, bar: Gauge(over) pressure. P, bar: Absolute pressure. Pc, bara: Critical pressure. Pr: Reduced pressure. Note 4! In refrigeration terminology, temperatures are usually expressed in C, temperature differences in K. K-value:
Channels in series: A groups of channels arranged in parallel, a pass. The passes are arranged in series. A multi-pass unit. Note 3! The two sides in a PHE are independent, i.e. one side could have a parallel arrangement and the other two groups in series.
Note 5! Although a power unit, comparable to kW, horse power or kcal/hr, RT is customarily used only for cooling. A 400 RT condenser probably means a condenser serving in a system where the evaporator gives 400 RT and not a condensing capacity of 400 * 3.504 = 1401.6 kW. Note the danger of misunderstandings.
9. Appendix I. Definitions.
153
C or K
tw t MTD = t = tw = tc
C or K
t1 (tw - tc) MTD = LMTD = ln(tw / tc)
tw tc
tc
t2
Desuperheating Condensing. LMTD Subcooling. LMTD can be culated K-value varies. calculated
t Note1. t has to be a linear function of of the duty in order to use LMTD. t1 & t2 need not to be linear dependent on the heat transferred
t1 tw t t2 tc tc MTD LMTD
the quantities Q (duty) and A (Area) are fairly unambiguously defined. In case of a duty with no phase change, K (the overall heat transfer coefficient) and T (the temperature difference), are also easily defined. Especially in Fig.1 , where the temperature difference is constant from inlet to exit, T is simply equal to this temperature difference. If the temperature change varies from the cold side to the warm side, Fig. 2, the situation gets more complicated. The simple temperature difference T has to be replaced with some mean temperature difference, MTD. The question is how MTD shall be calculated. If the duty complies with the following requirements, then LMTD, as defined in Fig. 2, can be used as a suitable mean value in the calculations of HEs: The K-value has to be constant. t has to be a linear function of the heat transferred. Fig. 3 shows another valid case. The specific heat of the media varies with the temperatures. This makes the temperature vs. heat transferred curved, but in this case the T is still linear and LMTD can still be used. Figure 4 shows condensation of a superheated vapour as in the following example: R 22 vapour enters at 73, condenses at 40 and subcools to 37 C. Water enters at 27 and leaves at 37 C.
The K-value is thus a final result of, not a parameter used in, the calculation of the area.
154
1. Fg = ( l - g) * g * (Dp3 * / 6 ) 2. Ff = Cd * (D p2 * / 4) * g * vt2 / 2 3. At equilibrium: Fg = Ff 4. vt2 = 4 * g * D p * ( l - g) / (3 * Cd * g) 5. Cd = f(Re) from figure 02. 6. In order to calculate vt, Cd has to be known. However, Cd is a function of vt and a trial and error procedure is needed. Usually only a few trials are necessary. 7. The nominal design velocity is taken as a fraction of the settling velocity. 8. v = (0.75 to 0.9) * vt 9. D = ( 4 * V / ( * v ) ) 0.5
Vapour
Cd
100
1.2 D
vt
v
VapourLiquid
D
Min. liq. level Liquid
(2.5 - 3) * D
0.3 D
10
Re
0.1
1
Oil
10
100
1000
2. Vertical separators.
The principle of the vertical separator is simple. The vapour-liquid mixture enters the separator and the vapour flows upwards with a certain velocity v, see figure 01.
155
Sh St vt As high as possible Max. liq. level < D * 3/4 (3 - 5) * D The minimum liquid level can be lower than the separator. See text.
A: Liquid flow area P: Wetted perimeter vh: Vapor velocity Hydraulic diameter for liquid flow in a circular duct: dhl = 4 * A / P (General) (Lh = D or D/2) vl: Liquid velocity Lh D
dhl = D
vh
dhl 1.2D (Lh = 0.75 * D) dhl 0.6D (Lh = 0.25 * D) Liquid Reynold #.:
1. vt, the settling velocity, is obtained from figures 01 & 2. 2. Try a horizontal vapour velocity vh = (1 to 5) * vt Check with figure 4, if this is low enough. 3. Assume a maximum liquid level: Lh 0.75 * D. A suggested value is: Lh = 0.5 * D. 4. Calculate the diameter from D = ( 8 * V / ( * v h) ) 0.5 Note! The suggested value of Lh from 3 is used. 5. The settling height St is then: St = D - Lh = 0.5 * D. 6. Calculate the separation distance: Sh = Ma * St * vh / vt Ma is a margin which should be in the order of 1.10 to 1.25. Try to get a holding time, sec: 3 < Sh / vh > St / vt.
Eq. Rel N Re-entrainment velocity vre A. <160 >1.5 * (/l) * R * Rel-0.5 B. 160Rel1635 0.0667 >11.78 * (/l) * R * N0.8 * Rel-1/3 C. 160Rel1635 >0.0667 >1.35 * (/l) * R * Rel-1/3 D. >1635 0.0667 >(/l) * R * N0.8 E. >1635 >0.0667 >0.1146 * (/l) * R Maximal values of vh to prevent re-entrainment. Adapted from (3). Note that higher turbulence gives lower velocities, thus use equations D & E, which are not dependent on the geometry.
3. Horizontal separators.
Horizontal separators are more complicated than vertical. The basic operation principle is simple, even though it differs from vertical separators. As in vertical separators, droplets larger than a certain diameter Dp, with the corresponding terminal velocity vt, have to be separated. The vapour-liquid mixture enters in one end of the separator, travels horizontally along the separator and leaves at the other end. A droplet therefore moves in two directions: Horizontally with the vapour velocity vh the distance Sh. Vertically with the settling (terminal) velocity vt the maximum distance St.
156
The re-entrainment velocity depends also on the surface tension of the liquid. Values of the surface tension can sometimes be hard to find. There is however a surprisingly easy correlation for the surface tension , as a function of the absolute temperature T, using only the easily available critical properties Tc and Pc and the normal boiling point temperature Tb. Tbr = Tb / Tc Q = (0.12 * (1 + Tbr * ln(Pc) / (1 - Tbr)) - 0.28) / 1000 (Eq. 1) (Eq. 2) (Eq. 3)
placed by vapour, but on the other hand, the exit pipe now contains some liquid. The vertical cross section of the separator has to be tuned to the expansion valve, which is probably a float type. If this operates intermittently, i.e. opens at a low set point and closes at a high set point, the level variations shall be so designed that the valve opens & closes at a reasonable frequency. This should be checked with the valve manufacturer.
See Table 5 of values for the re-entrainment velocity vre for some common refrigerants. The separation efficiency also depends on the liquid height in the separator as this effects both the separation height and the vapour velocity, see figure 03.Suppose that the separator is designed for maximum liquid filling, which should not be more than about 80 % of the cross section. The separation distance is then about 25 % of the diameter. The limiting droplet moves as 1) in figure 03. If the filling is reduced to half the cross section, the vapour velocity is reduced 2.5 times (the cross section changes from 20 % to 50%). The separation distance only doubles and the limiting droplet now flows like 2). The maximum liquid level sets the design limit if the maximum capacity occurs simultaneously, but as this can change, it has to be checked from case to case. In a separator with two vapour inlets, e.g. as in figure 05, (4. Evaporators & Separators), the total vapour flow is split into two streams. The vapour velocity is therefore reduced to half of that in the one inlet separator. However, the separating distance is reduced to half as well, which means that the separating efficiency remains. The large advantage of splitting the vapour flow is that the reduced velocity means less re-entrainment and the shell can then be made longer and slender, i.e. normally reduced cost. In general can be said that the number of exits does not affect the separation efficiency, as the decreased velocity is balanced by reduced separation distances. The design of a multi inlet/exit separator should be based on the most loaded section of the separator. The injection points of the flashing refrigerant probably do not correspond to the vapour inlet/exits and the loads on the different sections are then unsymmetrical. Note that both the minimum and maximum liquid level can be lower than the separator if a vertical container as in figure 02B, (4. Evaporators & Separators) is added to the horizontal part. The separating distance is in this case the diameter of the horizontal part.
The Ks value is either taken as a constant or depends on one or more variables. In table 2 there is compilation of values from different sources. As can be seen, there is a large spread between the recommended values.
4. Common points.
The minimum liquid level shall be high enough that there is no danger of vapour entering the liquid line. This is especially important for pump circulation. The separator volume shall be large enough in order to at least contain the LP side refrigerant. In a system, where the PHE is placed close to the separator, a rough rule to calculate the operating refrigerant content is to use the PHE plus the inlet pipe volume. When the PHE starts, part of the PHE volume is dis-
157
Comments ft/s 0.1 0.263 0.1 0.227/0.034 0.13/0.26 Vertical Vertical Vertical With/without demister V/H with demister
Ref.
Ks m/s vt = Ks * ((l -g) / g)0.5 20 10 Surface tension, mN/m
4 5 6 7 8
5.0
Various F= Decontamination factor f(Ml/Mg)( g/l)0.5 = 0.02 to0.4 (British) f(Decontamination factor) = 0.014 to 0.11 (SI) The same as 9, but computerized Vapour velocity as a function of vertical distance The source of the above (12) but better Velocities between 0.5 and 1.0 m/s for ammonia
0.245 * (F-50)0.4
15 9 10 11 12 13 14
pp 60-64 Oil & Gas Journal pp 74-80 Chemical Engineering pp. 405 - 409 JPT (Society of Petroleum Engineers) pp. 201-202 Chemical Engineering pp. 155-156 Hydrocarbon Proc. and Petroleum. Ref. pp. 1.7 -1.9 ASHRAE Handbook. Refrigeration pp. 81-84 Chemical Engineering pp. 157-156 Chemical Engineering pp. 253-256 Hydrocarbon Proc. and Petroleum. Ref. pp. 511-515 CHEMICAL AGE OF INDIA pp. 55-62 Oil & Gas Journal pp. Proceedings I.I.R pp. 346-349 Unit Operations, John Wiley & Sons Inc. pp. 89-97 Kltetechnik - Klimatisierung 18 Jahrgang, pp. 11-3,36 Perrys Chemical Engineers Handbook pp. 104-107 Ammoniakverdichterklteanlagen, pp. 287-295 Pohlmann, Tachenbuch der Kltetechnik Design manual Heat Transfer Research Inc. Design of fractionating columns Ind. & Eng. Chemistry,
(Eq. 5)
7. Example.
Vapour from a 1000 kW ammonia evaporator, together with the flash vapour, about 0.92 kg/s at 0 C, shall be separated in a horizontal separator. Calculate vt from (eq. 5).
D) Use a demister. This can increase the allowable design velocity more than six times (7). Then follow the demister manufacturers recommendations. E) ASHRAE (6) gives two methods of calculating vt. One is (eq. 4) with Ks = 0.1 or 0.2 ft/s. (Note that the 1994 editions uses SI units, but have forgotten to convert Ks from British units, i.e. Ks is still reported as 0.1/0.2). The principal method refers to (12) which refers to (13) which in turn refers to (19). This is a method used to calculate the spacing in distillation columns. In a distillation column figure 07, (1. Applications), the vapour passes upward through the liquid with a velocity higher than the superficial velocity, i.e. probably more severe entrainment. The velocity depends also on the spacing. It probably gives a too low velocity but is then safe. The main problem is that the spacings used by (12), 12 & 24 & 36 (12 is incorrectly reported by ASHRAE as 10) are close to impossible to interpolate between. Thus, it is better to use the curves from (13), converted to SI units in figure 05. It is valid for vertical separators.
l = 638.6 kg/m3, g= 3.457 kg/m3, l = 1872 * 10-6 kg/m,s = 26.76 mN/m. The volume is = 0.266 m3/s. vt = 0.00272* ((l -g)*g)0.328* ( ( l - g)/g)0.5 = 0.46 m/s
The more stringent calculation in table 5 (A and Fig 1 & 2) gives 0.45 m/s. Try a horizontal velocity of 3 * 0.47 = 1.41 m/s and a half full separator. The diameter is then: D = ( 8 * 0.266/ * 1.41) 0.5 = 0.693 m, chose 700 mm. The min. length Sh (St = 0.7/2) is 0.35*1.41/0.47= 1.05 m. This is a too small volume; the residence time is also too small. Increase the length to 3.5 m. The residence time is now 3.5/1.41 = 2.5 sec. Small but on the other hand, the length is much longer than the calculated necessary. Will there be a danger of re-entrainment? Figure 04 gives N = 0.000995. Use equation D. Equation D gives Vmax. = 7.70 m/s, i.e. OK. See also table 5.
158
psl = l * g * Hl
(Eq. 6)
Friction pressure drop. Formulas for this can be found in most engineering handbooks. One is: m = M / ( * D2 / 4); Rel = m * D / l = vl * D / (l / l ) fl= 0.0055 / Rel0.2 (Eq. 7) (Eq. 8) (Eq. 9) (Eq. 10)
pfl = 4 * fl * (L / D) * m2/(2 * l)
The friction pressure drop has to be evaluated for each pipe diameter D, length L and the corresponding liquid mass flow, m or velocity, vl. The length shall include the passage through bends, valves, fittings, etc. Local pressure drops in valves, fittings, bends, etc. each for its corresponding liquid mass velocity, m:
pi = ki * m2 / (2 * l) = ki * l * vl2 / 2
(Eq. 11)
ki can be found in most engineering handbooks. Some examples are given in figure 06. Momentum change. Every time the velocity changes, by a change in the vapour fraction or in the cross section, there is a change in momentum, expressed as a change of the pressure. At a velocity increase, the pressure decreases, i.e. a pressure drop. At a velocity decrease, the pressure increases, i.e. a pressure recovery. In a single-phase flow when the velocity changes from state 1 to 2, the calculation is straightforward:
(Eq. 12)
Note that the pressure drop in e.g. a sudden contraction is of two types: Loss as given by Eq. 11. Momentum change as given by Eq. 12. Thus if a sudden contraction is followed by an equal large sudden expansion, there is a pressure loss due to the momentum change in the contraction but this is recovered in the expansion. The contraction and the expansion then have their unrecoverable losses as given by the coefficients k in figure 06 and table 5. In a closed single-phase loop, all contractions and expansions equalize each other and the net result is no change in momentum.
159
The homogeneous model is suitable to correlate local pressure drops in valves, bends, fittings, etc. and pipes for very large mass velocities and high vapour fractions. Homogenous flow is normally shear controlled. In this appendix, index h indicates homogeneous flow and tp a flow with slip, e.g. h and tp, the homogeneous and slip two-phase densities. All pressure drop correlations show large differences from each other, as it is very difficult to exactly correlate a twophase mixture. There is e.g. a large difference if the liquid and vapour enters a pipe well separated or intimately mixed. The surface tension, both liquid-liquid and liquid-wall, has a little understood effect. Entrance effects to a pipe play an important part. Correlations developed for a small pipe diameter do not necessarily scale up to a larger diameter. A standing sound wave can form and disturb the flow. Certain parameters are difficult to determine experimentally, which affects the quality of the correlations. Two-phase flow is correlated to single-phase flow data. There can be large deviations in these, depending on source, mechanical execution, methods, etc. Impurities such as oil, water, remains of detergents, etc. change the physical properties. After a few bends, valves, fittings, changes of directions etc, the flow appearance might have changed notably and the pressure drop in the last meter of a pipe might differ from the first despite the same flow parameters.
As in the single-phase design, the calculation is made for each pipe section with lengths L and the corresponding diameters D. Note that the two-phase correction factor is surprisingly constant for a given vapour fraction. Estimate of the two-phase pressure drop by multiplying the vapour pressure drop (note, as if vapour were flowing alone) with the factors below: All hydrocarbons: Ammonia:
g2 = 2.3 g2 = 2.0
Table 5B gives the multiplication factor g2 and the friction pressure drop for an one meter high, vertical twophase column, a vapour fraction of 85 %, a diameter of 0.1 m and for the corresponding minimum velocity Note that the static pressure drop is dependent on the velocity and diameter. It can easily be recalculated to other velocities and diameters as shown in table 5. Horizontal and downward flow give less friction pressure drop than predicted for vertical flow. Local pressure drops in valves, fittings, bends, etc. Experimental correlations give varied results. The homogenous two-phase density, h, used with the single phase pressure drop coefficients, usually gives a conservative result and is used in table 5B:
ki * m2 / (2 * h)
Momentum change in a two-phase flow is complicated, as the two-phase density has to be correlated. In general, the homogenous density correlates data in sudden changes better than the separated flow density and is probably conservative. There is a large change in momentum from the inlet of the evaporator to the exit as the fluid changes from liquid alone to two-phase with a large increase in volume. This change in momentum is included in the pressure drop for the evaporator and is not dealt with here. In the circulating loop, there is normally no change in the vapour fraction and change in momentum thus occurs only in contractions and expansions.
phase flows slower than the vapour phase - separated flow - and a given cross section thus contains proportionally more liquid, i.e. the slip two-phase density is higher than the homogenous density. Extensive work has been done to correlate the two-phase slip density tp with physical properties and operating parameters but the phenomena are still not fully understood and quantified. The static pressure drop is then correlated as:
(Eq. 19)
pstp = tp * g * Htp
(Eq. 13)
There is thus a large recovery when the vapour liquid mixture enters the separator (and a loss due to the k). However, this pressure recovery is normally not taken into account. There might be some reasons for this. The vapour probably turns a couple of times in the separator and this is normally not accounted for. There might also be guiding vanes, baffles, etc.
Pressure gain can occur in downwards flow, but is usually not taken into consideration. The reason is that in practice it can be difficult to obtain. Friction pressure drop. The most popular correlations correlates the two-phase pressure drop as the singlephase pressure drop, i.e. as if the vapour was flowing alone in the pipe. The vapour pressure drop, pg, is then multiplied with a correction factor g2.
pftp = pg * g2
(Eq. 14)
160
+ -
Eq. 6. Eqs. 10, 11 & 12. From the specification. Eqs.13-19 & table 5.
By definition, this balance is always valid for a thermosiphon loop in operation, but what happens if there is a difference between the nominal driving force and the sum of the nominal pressure drops?
(Eq. 20)
Note! Never put a control valve between the evaporator exit and the separator. If a valve is necessary, it should be placed before the evaporator.
9. Pump circulation.
The above described design methods could be used by forced circulation as well. The driving head is now mainly the pump head. Excessive head is released in a valve between the pump and the evaporator. Note that to avoid cavitations, a pump is usually placed appreciably lower than the separator and the heat exchanger, which increases the pressure and thus the liquid subcooling at the pump. For the same reasons, there should not be any larger local pressure drops, e.g. a valve, before the pump.
161
Sudden reduction d/D = 0.0 k = 0.5 0.2 0.48 0.4 0.42 0.6 0.32 0.8 0.18 1.0 0.00 Velocity based on d
Sudden expansion: d/D = 0.0 k = 1.00 0.2 0.92 0.4 0.71 0.6 0.41 0.8 0.13 1.0 0.00 Velocity based on d Open angle valve: Open ball valve: Open gate valve: Open globe valve: k=3 k=0 k = 0.2 k = 6 -9
Hl Long bend R / D = 1.5: k = 0.45 Short bend R / D = 1.0: k = 0.75 R: Centre curvature radius of a pipe D: Inside diameter of the pipe
The PHE pressure drop, PPHE, is defined between these two points.
0.9556 (Note the difference to the separator calculation) kPa 0.80 9.8054 * 2.5 = 15639 Pa = +15.64 2.5 psl = 638 * D = 0.05 vl = 0.7947 (pvhl = 638 * 0.79472 / 2 = 201 Pa = 0.201 kPa) (Rel = 0.795 * 0.05 / (0.000187 / 638 =135566) = 4* (0.055 / 1355660.2) * (2.5 / 0.05) * 0.201 = -0.21 2.5 pfl 4 pbl = 4 * 0.75 * 0.201 = -0.30 1 pvl = 1 * 6* 0.201 = -1.21 1 prl = 1* 0.45 * 0.201 = -0.09 1 pel = 1 * 1.1 * 0.201 = -0.22 0.15 13.6 1.96 4 3 1 0 1 1.04
m m/s m m
= = = =
1.96 * 4* 3* 1* 1*
petp = PPHE=
Min. vel. required m/s 5.98 (vhmin = 4.88 * (0.15 / 0.1)0.5 = 5.98) Back flow: (0.89 + 0.25 + 0.54 + 1.2 + 0.44 + 0.69) = 4.01 kPa which is 25.6 % of the total head Temperature increase in the evaporator: 4.01 * 6.23 /100 = 0.25 K. Check with the evaporator. Head margin: The driving head, 15.64 kPa is 24 % larger than the total p = 15.64 - 2.99 = 12.65 kPa 2) Calculated from table 5B.
1) Calculated by the use of the equations 6 to 11. The separator is connected to an MK15BW-35LG/ 36LW, 32.2 m2, a semiwelded unit. If the driving force becomes too small, consider the use of ball valves instead of angular and glove valves.
evaporator. One meter lower evaporator means a higher head by 6.25 kPa but only an increase of the static and friction pressure drop by 0.46 & 0.06 kPa. The evaporation temperature increases, though.
162
Temperature C
Pressure bar 1)
Slope
l
NH3 Mw = 17.03 Tc = 132.3 C Pc = 113.3 bar NBP = -33.3 C R134a Mw = 102.03 Tc = 101.2 C Pc = 40.67 bar NBP = -26.7 C R23 -40 -30 -20 -10 0 10 20 -30 -20 -10 0 10 20 0.717 1.194 1.901 2.907 4.294 6.150 8.575 0.844 1.327 2.006 2.928 4.146 5.717 0.624 1.137 1.937 3.119 4.789 7.065 10.070 0.810 1.310 2.022 3.002 4.308 6.003 8.158 0.864 1.387 2.128 3.144 4.493 6.240 8.453 0.705 1.110 1.677 2.445 3.453 4.746 6.368 8.368 0.911 1.419 2.123 3.066 4.295 5.859 7.809 10.200 0.722 1.081 1.566 2.203 3.020 6.823 10.040 14.280 19.700 26.490 34.81 1.76 1.14 0.77 0.54 0.39 0.29 0.22 3.40 2.32 1.64 1.20 0.89 0.68 3.63 2.20 1.41 0.95 0.66 0.48 0.35 3.17 2.12 1.48 1.06 0.78 0.59 0.45 3.02 2.03 1.42 1.03 0.76 0.57 0.44 1.72 1.18 0.84 0.62 0.46 0.36 0.28 0.22 0.71 0.98 0.70 0.52 0.39 0.31 0.24 0.19 1.94 1.39 1.02 0.77 0.60 0.41 0.30 0.22 0.17 0.13 0.10 1389 1360 1329 1297 1262 1226 1186 219.5 212.9 206.0 198.6 190.7 182.3 245.8 237.7 229.2 220.0 210.0 199.0 186.8 202.7 196.2 189.4 182.1 174.3 165.7 156.4 199.1 192.5 185.8 178.4 170.6 162.1 152.6 434.0 423.6 412.6 400.8 388.3 374.7 360.0 343.9 441.7 430.5 418.7 406.1 392.5 377.8 361.8 344.3 373.4 364.4 355.0 345.1 334.6 339.7 322.4 303.5 282.4 258.6 230.9 690.2 677.8 665.1 652.1 638.6 624.6 610.2 1388 1358 1327 1295 1261 1225 1476 1438 1399 1358 1315 1270 1221 1325 1293 1259 1225 1189 1151 1111 1327 1294 1261 1227 1191 1153 1112 590.5 579.0 567.2 554.9 542.1 528.7 514.7 499.9 611.9 599.4 586.5 573.1 559.2 544.6 529.3 513.0 603.2 592.0 580.5 568.6 556.2 1155 1116 1076 1032 982.9 928.1
v
0.6438 1.037 1.603 2.391 3.457 4.868 6.703 4.426 6.784 10.04 14.43 20.23 27.78 2.962 5.195 8.582 13.50 20.43 29.97 42.95 4.442 6.985 10.55 15.41 21.90 30.47 41.66 4.808 7.509 11.28 16.41 23.25 32.25 44.03 1.725 2.632 3.867 5.508 7.639 10.36 13.80 18.11 2.136 3.225 4.699 6.642 9.154 12.35 16.39 21.44 2.061 3.004 4.250 5.862 7.910 17.92 26.12 37.10 51.70 71.18 97.32
Hom.
0.8046 1.296 2.003 2.986 4.315 6.073 8.356 5.528 8.469 12.53 17.99 25.19 34.53 3.701 6.488 10.71 16.83 25.44 37.24 53.22 5.548 8.719 13.16 19.20 27.25 37.84 51.59 6.005 9.373 14.07 20.44 28.92 40.03 54.50 2.155 3.286 4.826 6.868 9.515 12.89 17.14 22.43 2.668 4.026 5.862 8.279 11.40 15.35 20.33 26.52 2.574 3.750 5.303 7.309 9.852 22.31 32.46 45.98 63.83 87.39 118.5 2) y = 0.80. 3) See eq.17
l
0.2812 0.2441 0.2144 0.1902 0.1701 0.1530 0.1383 0.4010 0.3476 0.3035 0.2665 0.2349 0.2074 0.3490 0.2953 0.2530 0.2187 0.1903 0.1660 0.1449 0.3881 0.3265 0.2778 0.2386 0.2066 0.1800 0.1574 0.3818 0.3213 0.2734 0.2349 0.2034 0.1772 0.1550 0.2172 0.1935 0.1731 0.1552 0.1396 0.1257 0.1133 0.1021 0.2011 0.1799 0.1617 0.1459 0.1321 0.1199 0.1090 0.0991 0.2526 0.2233 0.1984 0.1770 0.1585 0.2293 0.1938 0.1642 0.1393 0.1180 0.1054
v
0.00786 0.00815 0.00845 0.00875 0.00906 0.00936 0.00968 0.00964 0.01000 0.01036 0.01073 0.01110 0.01149 0.00903 0.00958 0.01012 0.01067 0.01122 0.01180 0.01242 0.01001 0.01050 0.01100 0.01152 0.01205 0.01262 0.01322 0.01004 0.01053 0.01104 0.01156 0.01211 0.01270 0.01333 0.00611 0.00636 0.00662 0.00689 0.00716 0.00745 0.00775 0.00809 0.00629 0.00657 0.00686 0.00717 0.00749 0.00784 0.00823 0.00867 0.00638 0.00662 0.00686 0.00711 0.00736 0.01131 0.01187 0.01246 0.01312 0.01386 0.01431
Separator velocities m/s Settling Reen152 traindrops ment vt vRe 0.718 0.633 0.559 0.493 0.435 0.385 0.341 0.693 0.589 0.502 0.431 0.371 0.320 0.838 0.677 0.553 0.456 0.379 0.316 0.264 0.660 0.554 0.467 0.397 0.338 0.289 0.247 0.644 0.541 0.457 0.388 0.331 0.283 0.242 0.596 0.515 0.446 0.389 0.340 0.298 0.261 0.229 0.570 0.491 0.425 0.369 0.322 0.281 0.246 0.214 0.566 0.495 0.435 0.383 0.339 0.364 0.306 0.257 0.217 0.181 0.151 21.2 16.5 13.1 10.5 8.53 6.99 5.76 6.01 4.77 3.83 3.10 2.52 2.07 8.31 6.13 4.64 3.57 2.76 2.16 1.68 5.59 4.42 3.54 2.86 2.32 1.88 1.52 5.32 4.22 3.39 2.74 2.23 1.81 1.46 8.61 6.85 5.53 4.51 3.72 3.08 2.55 2.12 7.98 6.36 5.14 4.19 3.45 2.85 2.35 1.94 7.41 6.06 5.00 4.16 3.49 3.11 2.46 1.95 1.53 1.19 0.878
-90 -80 Mw = 70.01 -70 Tc = 25.95 C -60 Pc = 49.0 bar -50 NBP = -82.03 C -40 -30 R404a Mw = 97.6 Tc = 73.0 C Pc = 37.8 bar NBP = -47.0 C R507 Mw = 98.9 Tc = 71.0 C Pc = 37.9 bar NBP = -47.0 C R290 -50 -40 -30 -20 -10 0 10 -50 -40 -30 -20 -10 0 10
-50 -40 Mw = 44.10 -30 Tc = 96.7 C -20 Pc = 42.48 bar -10 NBP = -42.09 C 0 10 20 R1270 Mw = 42.08 Tc = 92.4 C Pc = 46.65 bar NBP= -47.7 C -50 -40 -30 -20 -10 0 10 20 -20 -10 0 10 20 -50 -40 -30 -20 -10 0
R600a Mw = 58.12 Tc = 134.7 C Pc = 36.31 bar NBP = -11.61 bar CO2 Mw = 44.0 Tc = 31.06 C Pc = 73.84 bar NBP= -78.4 C
1) Reported properties show sizable variations depending on source, especially for refrigerants with glide. The data here are mainly based on NIST.
Table 5A. Physical properties, the settling velocity for 152 droplets and the reentrainment velocity in a horizontal separator for some refrigerants.
163
Refrigerant
Temperature
Minimum flow Mass Linear flow velocity 4), 6) Mmin kg/hr 170 220 280 355 447 560 700 693 901 1166 1501 1925 2461 559 776 1064 1448 1959 2635 3532 682 904 1188 1554 2022 2624 3394 717 948 1246 1630 2121 2751 3560 279 361 464 591 749 945 1186 1483 323 418 536 684 867 1094 1375 1721 314 397 498 623 774 1718 2300 3060 4055 5368 7112 4), 7) v hmin m/s 7.48 5.99 4.95 4.21 3.66 3.26 2.96 4.43 3.76 3.29 2.95 2.70 2.52 5.34 4.23 3.51 3.04 2.72 2.50 2.35 4.35 3.67 3.19 2.86 2.62 2.45 2.33 4.22 3.58 3.13 2.82 2.59 2.43 2.31 4.59 3.89 3.40 3.04 2.78 2.59 2.45 2.34 4.28 3.67 3.24 2.92 2.69 2.52 2.39 2.29 4.32 3.74 3.32 3.01 2.78 2.72 2.51 2.35 2.25 2.17 2.12
C NH3 -40 -30 -20 -10 0 10 20 -30 -20 -10 0 10 20 -90 -80 -70 -60 -50 -40 -30 -50 -40 -30 -20 -10 -0 10 -50 -40 -30 -20 -10 0 10 -50 -40 -30 -20 -10 0 10 20 -50 -40 -30 -20 -10 0 10 20 -20 -10 0 10 20 -50 -40 -30 -20 -10 0
Two-phase P Static, Vapour multi2), 5), plier 11) 2), 5) Pstp v2 kPa 0.070 0.125 0.210 0.321 0.454 0.606 0.772 0.600 0.908 1.268 1.667 2.092 2.530 0.381 0.703 1.118 1.602 2.130 2.680 3.235 0.603 0.926 1.304 1.718 2.155 2.602 3.048 0.653 0.988 1.375 1.797 2.239 2.690 3.140 0.232 0.354 0.496 0.655 0.823 0.997 1.172 1.347 0.289 0.427 0.585 0.758 0.938 1.123 1.308 1.490 0.280 0.401 0.540 0.691 0.851 1.873 2.347 2.823 3.288 3.735 4.156 1.80 1.86 1.92 1.97 2.03 2.08 2.13 2.12 2.19 2.26 2.32 2.37 2.42 1.98 2.07 2.15 2.22 2.28 2.34 2.38 2.12 2.18 2.24 2.29 2.34 2.38 2.41 2.13 2.20 2.25 2.30 2.35 2.38 2.41 2.05 2.13 2.19 2.25 2.30 2.35 2.39 2.42 2.07 2.14 2.20 2.26 2.31 2.36 2.39 2.42 2.13 2.19 2.25 2.30 2.34 2.35 2.39 2.42 2.44 2.43 2.44
Friction 4), 9), 10) Pftp 0.0079 0.0081 0.0085 0.0091 0.0100 0.0111 0.0126 0.0187 0.0207 0.0235 0.0271 0.0299 0.0339 0.0173 0.0191 0.0219 0.0253 0.0290 0.0339 0.0405 0.0181 0.0202 0.0231 0.0265 0.0298 0.0342 0.0399 0.0185 0.0207 0.0238 0.0271 0.0306 0.0353 0.0412 0.0080 0.0088 0.0099 0.0113 0.0131 0.0149 0.0169 0.0192 0.0086 0.0095 0.0108 0.0124 0.0144 0.0163 0.0185 0.0213 0.0087 0.0095 0.0106 0.0120 0.0137 0.0275 0.0317 0.0373 0.0444 0.0536 0.0656
Local P (homogenous) Velocity Angle Gate height valve valve k=1 k = 3.0 k = 0.2 2), 8), 9) pvhh pavh pgvh 0.0225 0.0233 0.0245 0.0264 0.0289 0.0323 0.0367 0.0543 0.0600 0.0678 0.0783 0.0920 0.1097 0.0527 0.0580 0.0661 0.0779 0.0943 0.1166 0.1466 0.0525 0.0586 0.0671 0.0786 0.0939 0.1138 0.1397 0.0536 0.0600 0.0691 0.0813 0.0973 0.1182 0.1454 0.0227 0.0249 0.0279 0.0318 0.0369 0.0433 0.0513 0.0613 0.0245 0.0271 0.0307 0.0353 0.0413 0.0488 0.0581 0.0698 0.0240 0.0263 0.0293 0.0332 0.0380 0.0828 0.1019 0.1274 0.1611 0.2062 0.2669 0.0674 0.0698 0.0736 0.0792 0.0868 0.0969 0.1100 0.1630 0.1800 0.2035 0.2350 0.2761 0.3291 0.1582 0.1740 0.1984 0.2338 0.2830 0.3497 0.4398 0.1574 0.1757 0.2013 0.2359 0.2816 0.3414 0.4190 0.1607 0.1801 0.2072 0.2438 0.2919 0.3547 0.4363 0.0680 0.0746 0.0836 0.0955 0.1107 0.1299 0.1540 0.1839 0.0734 0.0813 0.0920 0.1060 0.1238 0.1463 0.1744 0.2095 0.0720 0.0788 0.0879 0.0995 0.1141 0.2483 0.3058 0.3822 0.4834 0.6187 0.8007 0.0045 0.0047 0.0049 0.0053 0.0058 0.0065 0.0073 0.0109 0.0120 0.0136 0.0157 0.0184 0.0219 0.0105 0.0116 0.0132 0.0156 0.0189 0.0233 0.0293 0.0105 0.0117 0.0134 0.0157 0.0188 0.0228 0.0279 0.0107 0.0120 0.0138 0.0163 0.0195 0.0236 0.0291 0.0045 0.0050 0.0056 0.0064 0.0074 0.0087 0.0103 0.0123 0.0049 0.0054 0.0061 0.0071 0.0083 0.0098 0.0116 0.0140 0.0048 0.0053 0.0059 0.0066 0.0076 0.0166 0.0204 0.0255 0.0322 0.0412 0.0534
Long 90 Short 90 bend bend k = 0.45 k = 0.75 plbh psbh 0.0101 0.0105 0.0110 0.0119 0.0130 0.0145 0.0165 0.0245 0.0270 0.0305 0.0353 0.0414 0.0494 0.0237 0.0261 0.0298 0.0351 0.0424 0.0525 0.0660 0.0236 0.0264 0.0302 0.0354 0.0422 0.0512 0.0629 0.0241 0.0270 0.0311 0.0366 0.0438 0.0532 0.0655 0.0102 0.0112 0.0125 0.0143 0.0166 0.0195 0.0231 0.0276 0.0110 0.0122 0.0138 0.0159 0.0186 0.0219 0.0262 0.0314 0.0108 0.0118 0.0132 0.0149 0.0171 0.0372 0.0459 0.0573 0.0725 0.0928 0.1201 0.0169 0.0175 0.0184 0.0198 0.0217 0.0242 0.0275 0.0408 0.0450 0.0509 0.0588 0.0690 0.0823 0.0396 0.0435 0.0496 0.0584 0.0707 0.0874 0.1100 0.0394 0.0439 0.0503 0.0590 0.0704 0.0854 0.1048 0.0402 0.0450 0.0518 0.0609 0.0730 0.0887 0.1091 0.0170 0.0186 0.0209 0.0239 0.0277 0.0325 0.0385 0.0460 0.0184 0.0203 0.0230 0.0265 0.0310 0.0366 0.0436 0.0524 0.0180 0.0197 0.0220 0.0249 0.0285 0.0621 0.0764 0.0955 0.1209 0.1547 0.2002
Sudden reduc. k=0.45 psrh 0.0101 0.0105 0.0110 0.0119 0.0130 0.0145 0.0165 0.0245 0.0270 0.0305 0.0353 0.0414 0.0494 0.0237 0.0261 0.0298 0.0351 0.0424 0.0525 0.0660 0.0236 0.0264 0.0302 0.0354 0.0422 0.0512 0.0629 0.0241 0.0270 0.0311 0.0366 0.0438 0.0532 0.0655 0.0102 0.0112 0.0125 0.0143 0.0166 0.0195 0.0231 0.0276 0.0110 0.0122 0.0138 0.0159 0.0186 0.0219 0.0262 0.0314 0.0108 0.0118 0.0132 0.0149 0.0171 0.0372 0.0459 0.0573 0.0725 0.0928 0.1201
Sudden expan. k = 1.1 pseh 0.0247 0.0256 0.0270 0.0290 0.0318 0.0355 0.0403 0.0598 0.0660 0.0746 0.0862 0.1013 0.1207 0.0580 0.0638 0.0728 0.0857 0.1038 0.1282 0.1613 0.0577 0.0644 0.0738 0.0865 0.1033 0.1252 0.1536 0.0589 0.0660 0.0760 0.0894 0.1070 0.1301 0.1600 0.0249 0.0273 0.0306 0.0350 0.0406 0.0476 0.0565 0.0674 0.0269 0.0298 0.0337 0.0389 0.0454 0.0536 0.0640 0.0768 0.0264 0.0289 0.0322 0.0365 0.0418 0.0910 0.1121 0.1401 0.1773 0.2268 0.2936
R134a
R23
R404A
R507A
R290
R1270
R600a
CO2
Table 5B. A. Minimum flows in a vertical pipe, two-phase static, friction and local pressure drops.
164
Name R11 R12 R12b1 R13 R13b1 R14 R21 R22 R23 R30 R31 R32 R40 R41 R50 R113 R114 R114a R114B2 R115 R116 R123 R123a R124 R125 R132a R132b R133a R134 R134a R141b R142b R143 R143a R152a R160 R216 R218 R290 R227ea R236ea R236fa R245ca R245cb R245fa R600 R600a R717 R744 R1270 RC270 RC318 RE134 RE245 RE170 R400 R401a R401b R401c R402a R402b R403a R403b R404a R405a R406a R407a R407b R407c R407d R407e R408a R409a R409b R410a R410b R411a R411b R411c R412a R413a R414a R414b R415a R417a R500 R501 R502 R503 R504 R505 R506 R507a R508a R508b R509a R723 Formula CFCl3 CF2Cl2 CBrClF2 CF3Cl CF3Br CF4 CHFCl2 CHF2Cl CHF3 CH2Cl2 CH2ClF CH2F2 CH3Cl CH3F CH4 CFCl2-CClF2 CCl F2-CClF2 CCl2F-CF3 CBrF2-CBrF2 CCl F2-CF3 CF3 CF3 CHCl2-CF3 CCl F2-CHClF CHFCl-CF3 C2HF5 CHCl2-CHF2 CH2Cl-CClF2 CH2Cl-CF3 CHF2-CHF2 CF3-CH2F CFCl2-CH3 CF2Cl-CH3 CF2H-CFH2 CF3-CH3 CHF2-CH3 CH3-CH2Cl CClF2-CF2-CF2Cl C3F8 C3H8 CF3-CHF-CF3 CF3-CHF-CHF2 CF3-CH2-CF3 CHF2-CF2-CH2F CF3-CF2-CH3 CF3-CH2-CHF2 C4H10 (CH3)2-CH-CH3 NH3 CO2 CH3-CH-CH2 C3H6 C4F8 CHF2-O-CHF2 CF3-CH2-O-CHF2 CH3-O-CH3 R12/R114 R22/R152a/R124 R22/R152a/R124 R22/R152a/R124 R125/R290/R22 R125/R290/R22 R290/R22/R218 R290/R22/R218 R125/R143a/R134a Synonym NBP C Trichlorofluoromethane 23.76 Dichlorodifluoromethane -29.76 Bromochlorodifluoromethane -3.54 Chlorotrifluoromethane -81.48 Bromotrifluoromethane -57.83 Tetrafluoromethane -127.98 Dichlorofluoromethane 8.85 Chlorodifluoromethane -40.86 Trifluoromethane -82.03 Methylenchloride 39.64 Chlorofluoromethane -9.1 Difluoromethane -51.66 Methylchloride -23.38 Fluoromethane -78.12 Methane -161.5 1,1,2-Trichlorotrifluoroethane 47.65 1,2-Dichloro-1,1,2,2-tetrafluoroethane 3.59 1,1-Dichloro-1,2,2,2-tetrafluoroethane 3.0 1,2-Dibromo-1,1,2,2-tetrafluoroethane 47.1 Chloropentafluoroethane -38.94 Hexafluoroethane -78.20 1,1-Dichloro-2,2,2-Trifluoroethane 27.83 1,2-Dichloro-1,1,2-Trifluoroethane 30.01 1-Chloro-1,2,2,2-Tetrafluoroethane -11.94 Pentafluoroethane -48.14 Dichlorodifluoroethane Dichlorodifluoroethane(Est. values) 93.0 1-Chloro-3,3,3-Trifluoroethane 6.1 1,1,2,2-Tetrafluoroethane -19.81 1,1,1,2-Tetrafluoroethane -26.07 1-Fluoro-1,1-Dichloroethane 32.15 1-Chloro-1,1-Difluoroethane -9.14 1,1,2-Trifluoroethane 4.01 1,1,1-Trifluoroethane -47.23 1,1-Difluoroethane -24.02 Ethylchloride 14.0 1,1,2,2,3,3-Hexafluoro-1,3-Dichloropropane 35.7 Perfluoropropane -36.75 Propane -42.09 1,1,1,2,3,3,3-Heptafluoropropane -15.61 1,1,1,2,3,3-Hexafluoropropane 6.19 1,1,1,3,3,3-Hexafluoropropane -1.42 1,1,2,2,3-Pentafluoropropane 25.22 1,1,1,2,2-Pentafluoropropane -17.59 1,1,1,3,3-Pentafluoropropane 14.90 N-Butane -0.54 Isobutane -11.61 Ammonia -33.33 Carbondioxide (Subli.) -57.00 Propene(Propylene) -47.70 Cyclopropane -32.90 Perfluorocyclobutane -5.99 Bis(Difluoromethyl)Ether 6.17 2-Difluoromethoxy-1,1,1-Trifluoroethane 29.20 Dimethylether -24.84 60/40 % Zeotrop, Glide 6.4 K -23.20 53/13/34 % Zeotrop, Glide 6.4 K -33.05 61/11/28 % Zeotrop, Glide 6.0 K -35.70 33/15/52 % Zeotrop, Glide 6.7 K -30.5 60/2/38 % Zeotrop, Glide 2.0 K -49.00 38/2/60 % Zeotrop, Glide 2.3 K -47.00 5/75/20 % Zeotrop, Glide 2.4 K -50.00 5/56/39 % Zeotrop, Glide 1.2 K -51.00 44/52/4 % Zeotrop, Glide 0.7 K -47.00 RC318/142b/152a/22 42.5/5.5/7/45 Zeotrop, Glide 8.4 K -28.67 R600a/R142b/R22 4/41/55 %, Zeotrop, Glide 9.2 K -30.00 R32/R125/R134a 20/40/40% Zeotrop, Glide 6.6 K -42.30 R32/R125/R134a 10/70/20% Zeotrop, Glide 4.4 K -45.20 R32/R125/R134a 23/25/52% Zeotrop, Glide 7.1 K -40.00 R32/R125/R134a 15/15/70% Zeotrop, Glide 6.7 K -36.05 R32/R125/R134a 25/15/60% Zeotrop, Glide 7.3 K -39.15 R125/R134a/R22 7/46/47% Zeotrop, Glide 0.6 K -44.00 R22/R124/R142b 60/25/15% Zeotrop, Glide 8.1 K -34.00 R22/R124/R142b 65/25/10% Zeotrop Glide 7.2 K -35.00 R32/R125 50/50% Zeotrop, Glide 0.2 K -51.00 R32/R125 45/55% Zeotrop, Glide 0.2 K -51.47 R1270/R22/R152a 1.5/87.5/11% Zeotrop, Glide 2.4 K -38.46 R1270/R22/R152a 3/94/3% Zeotrop, Glide 1.3 K -40.52 R1270/R22/R152a 3/95.5/1.5% Zeotrop, Glide 1.0K -41.34 R22/R142b/R218 70/25/5% Zeotrop, Glide 8.1 K -36.06 R134a/R218/R600a 88/9/3% Zeotrop, Glide 6.9 K -35.00 R22/124/142b/600a 51/28.5/16.5/4 Zeotrop, Glide 8.3 K -29.9 R22/124/142b/600a 50/39/9.5/1.5 Zeotrop, Glide 8.3 K -30.27 R22/R23/R152a 80/5/15% Zeotrop, Glide 10.3 K -41.9 R125/R134a/R600 46.6/50/3.4% Zeotrop, Glide 5.5 K -41.80 R12/R152a 73.8/26.2% Azeotrop -33.5 R12/R22 25/75% Azeotrop -40.41 R22/R115 48.8/51.2% Azeotrop -45.50 R13/R23 59.9/40.1% Azeotrop -88.7 R32/R115 48.2/51.8 Azeotrop -57.2 R12/R31 78/22% Azeotrop -29.9 R31/R114 55.1/44.9 Azeotrop -12.3 R125/143a 50/50% Azeotrop -47.00 R23/R116 39/61% Azeotrop -87.40 R23/R116 46/54% Azeotrop -86.00 R22/R218 44/56% Azeotrop -47.50 NH3/Dimethylether 60/40% Azeotrop -36.6 T26bar C 162.7 86.4 126.3 11.3 46.3 -57.8 137.6 63.2 1.2 174.7 111.7 42.0 89.2 9.1 -101.9 196.1 132.5 131.5 197.0 70.6 13.6 162.6 166.3 104.6 51.1 111.7 127.3 89.4 79.4 172.3 111.3 118.8 55.4 85.0 141.6 176.7 71.6 70.3 96.1 125.1 114.3 150.9 97.1 135.7 129.3 115.0 59.8 -10.6 61.3 81.6 111.9 124.1 156.9 89.3 101.3 75.9 73.5 82.6 53.0 56.1 59.6 58.2 55.4 80.4 82.86 56.9 54.0 59.24 65.77 60.7 57.6 76.4 73.9 43.1 43.4 66.4 63.4 62.9 73.8 75.4 79.7 78.5 63.5 64.5 77.5 65.3 59.46 -1.6 38.32 103.6 54.3 -2.5 -2.9 57.0 58.7
165
1. Gasket properties
Gaskets can be of a wide variety of materials but in PHEs and SWPHE used in refrigeration applications, practically only elastomers - rubber - are used. Polar/Nonpolar. In the classification of gaskets, the polarity is an important property. A polar compound, despite being electrically neutral, has one end negatively charged and the other positive. A nonpolar compound has no such unequal distribution of electrical charges. Polar compounds have usually a group or atoms such as -OH, HCO, -COOH, -CL, or NH2- attached to the molecule. As a general rule, a polar solvent dissolves a polar compound but not a nonpolar compound and vice versa or differently expressed, the smaller the difference in polarity is, the more soluble is the compound Another condition for solubility is that the compounds are chemically similar. Components in a gaskets. An elastomer gasket is made of several components: The elastomer is a long chain of carbon molecules to which other atoms are attached. It is basically of three types: - Pure hydrocarbon - the chain is composed of only carbon and hydrogen - polymers, which are non-polar. Typical gaskets are EPDM & RCB. - Hydrocarbon with some hydrogen atoms substituted to other atoms than carbon/hydrogen or to a polar group. A polar polymer is then created. Typical gaskets are nitril rubber, chloroprene and fluorocarbons - The main chain is made of silicon atoms instead of carbon. Silicon rubber is an example. Curing agents. The raw polymer chains have to be cross-linked forming a three dimensional gitter, before becoming an elastomer proper. This is made by agents such as sulphur, resins, peroxides, etc. Fillers. Carbon black - the most common - increases hardness, tensile & shear strengths. Antidegradants, (such as anti-oxidants/ozonants) improve the resistant against various forms of attacks. Processing aids, e.g. softeners & mould release agents improve the manufacturing process. Gasket properties that characterize an elastomer are: Hardness. A hard gasket resists swelling better than a soft, but if too hard, the sealing properties might suffer. Tensile & tear strengths and elongation to brake. Compression set is the residual deformation after a pressure has been applied under specified conditions and then released. Relaxation is the rate which a compressed gasket looses its compression force. This is an important property to evaluate the sealing ability of a gasket. Gaskets for PHEs. The gasket in a PHE is subjected to much more severe conditions than e.g. an O-ring in a flange. The practical significance of this is that temperature and pressure limits issued by rubber manufactures are not always pertinent to PHEs.
Sealing. A PHE gasket is normally not self-sealing, contrary to a normal O-ring. Shape. The straight part of a field gasket - the gasket that surrounds the heating surface - have better sealing abilities than e.g. the corners. Length. The sheer length of field gaskets means that a relative better sealing ability is required than in a short flange O-ring. Nonglued gaskets should not be used: when frequent opening is necessary, for gaskets with large swelling, in oxidizing brines; when the gasket sticks to the matching plate.
2. Gasket corrosion
Attacks on the gasket are: Swelling, which is a similar phenomena to dissolution of a compound in a solvent and the same rules apply, e.g. a nonpolar solvent, oil, swells a nonpolar gasket, EPDM, but not a polar gasket, NBR. Swelling is a physical phenomena, the gasket is not destroyed by it, though properties can change if the gasket contains soluble additives. The gasket might then shrink. Oxygen, ozone, light and high temperatures attack and destroy the chain or the cross-links. The rubber becomes hard and brittle and looses its elasticity. Chemicals can have the same effects as above, but the resistance is very dependent on gasket and chemical. There are gaskets resistant to hot concentrate sulphuric acid but less to acetones. Chemicals can also attack the other components in the gasket and destroy or dissolve them. Fluoroelastomers (FPMs) have some or all of the hydrogen atoms in the chain exchanged to fluorine. The result is an elastomer, which is very resistant to all types of attacks but with one very important exception. It is readily disintegrated by ammonia, ammonia compounds, hydroxyles and amines, which then never should be used with FPM. This is especially important for by titanium, which is corroded by even minute amounts of fluorine. Diffusion of a fluid is closely related to swelling. If the gasket swells in a fluid, the fluid diffuses well through the gasket. It is temperature and pressure dependent. Cold leakage is a high temperature phenomena occurring when a PHE operating at high temperatures is shut down and started up a number of times. The cold gasket then obtains a permanent set causing leakage. Both the set and leakage disappears when the gasket heats up. EPDM is susceptible to cold leakage. Storage. Long-term storage of gaskets should be done in a dark place, preferably stored in e.g. plastic bag filled with an inert gas such as nitrogen.
166
Organic compounds Acids (R-COOH) Alcohols, Glycols, Polyols (R-OH) Aldehydes (R-CHO) Amins (R-NH2) Aromats (incl. alkylated) Esters (R-COO-R) Ethers (R-O-R) Halocarbons (R-F,Cl, Br, J) Hydroxyl amines Ketones (R-CO-R) Naphtenes Nitrates (R-NO2) Oximes (R=N-OH) Paraffins & Olefins
Gasket EPDM, FPM EPDM, FPM (EPDM) EPDM FPM EPDM (EPDM) FPM, Typ G EPDM, HNBR (EPDM) NBR, HNBR, FPM EPDM EPDM NBR, HNBR, FPM
Organic compounds Peroxides (R-O-O-R) Phenols (R-OH) Sulphides (R-S-R) Sulphonic acids (R-SO3H) Inorganic compounds Acids, oxidizing. (HNO3, HClO3, H2SO4(>75%), CrO3) non (H3PO4, HCl, HB3, H2SO4(<75%), H2S) Alkalies (NaOH, NH4OH, KOH, LiOH Hot pressurized air Salt solutions, oxidizing (K2CrO7, KMnO4, NaOCl) " non (NaCl,Na2SO4), K2CO3, CuSO4) Water, all types
4. Brine gaskets.
General. The field gaskets on the brine/liquid side are available in two commonly used materials and one on special request. The table below shows the general properties of these gaskets in different fluids. Nitril rubber (NBR) is the most common field gasket compatible with most brines and glycols. At very low temperatures LT-NBR to be used. Ethylene propylene dien monomer (EPDM) is used when polar compounds such as acids are used. It has good high temperature properties. Fluoroelastomers (FPM) have excellent resistance to a number of chemicals, e.g. hot concentrated sulphuric acid. It is however, a very expensive gasket and usually a minimum quantity is required. It should never be used with titanium se 3. It is thus seldom justified except in very special cases. Fluids, which can be treated in a PHE includes practically all common brines and glycols as well as most chemicals and alimentary fluids. There are however some types of fluid which should be avoided or discussed with Alfa Laval before being used together with gaskets. Some of them might be used with FPM gaskets but the price normally excludes this. Alimentary fluids, e.g. tap water, beer & soft drinks can be treated in a PHE. Contact Alfa Laval if formal FDA (the commonly accepted US standard) approval is needed. Heat transfer oils, except pure mineral and silicon oils, should in general be avoided Chlorinated (except R22) & many unsaturated (with double or triple bonds between the carbon atoms) compounds, aldehydes, ethers, ketones & aromatics (benzene rings) should in general be avoided together with gaskets. This is particularly true for compounds with low molecular weights. A large molecule with only one offensive group can be less aggressive. Contact Alfa Laval before use. Contact Alfa Laval before using strong oxidizing acids & salts, e.g. potassium bichromate. Note, that gaskets are in general resistant to nonoxidising acids and salts, e.g. hydrochloric acid, but it might be difficult to find a corrosion resistant plate material.
10. Index.
167
Absorption 12 ammonia-water 14 lithium bromide/water 16 Air, see: Inerts Ammonia, properties of 146ff and water, decrease of the pressure 78, 92, 142, 149 as secondary refrigerant 150 as refrigerant 146ff condensate subcooler 102 corrosion on copper 120, 124 in absorption systems 14ff in DX-valves 84, 92 in the subcooler/superheater cycle 10 insoluble oil 74, 75, 78, 80, 81, 120, 134ff separators properties 162, 163 venting 58, 106, 107 Azeotrops 143ff R12 & R134a 140 list of 164 see also: Glide Back flow: See reversed flow Balancing of water 66 Boiling, fundamental 70, 71 Brines, see also: Secondary refrigerants in a supermarket 18, 49 corrosion by CaCl2 and LiBr 122 Bulb, charges 84, 85 installation 91 Capacity, condenser control 107ff evaporator, matching compressors 94 surge in (pump circulation) 73 Carbon dioxide 148ff, 151 Cascade systems 12, 98 Characteristic curve, TEV & DX evaporator 85 Chlorination of cooling water 118 Chlorine, properties & use of 118 Circulation, forced or natural 6, 72ff back 76, 160 calculation of a thermosiphon loop 158ff Cleaning devices 119 Compressors, capacity in superheater/subcooler cycle 10 capacity matching 94 cut-out 112 liquid hammering in 6, 10, 22, 24, 32, 84 parallel 98, 138, 141 protection by the superheat 84 Condensate blocking of channels 102, 103, 105 level at subcooling 103 subcooling 102, 103 of HFC & HCs, control of 102 of steam or ammonia 102 Condensation process 100 Condensers general 100ff condensate subcooling 102 control 107ff desuperheating 100ff, 108ff parallel connected 102ff heat recovery 110ff inerts 101ff steam & ammonia 102 trouble shooting 112 venting 106 Connections soldering and welding 68, 69 reversed placements 56, 57
Control, capacity of evaporators & compressors 96 condensers 107 condenser/evaporator in cascade systems 98 condensate subcooling 102 desuperheaters 108 - 111 evaporation pressure 96 expansion in DX-system 4 expansion in flooded systems 93, 94 flooded evaporators 94 freeze control 132 heat recovery 110ff hot-gas bypass 97 in indirect systems 18 liquid receivers 109 liquid temperature control 96, 99 oil cooler 32, 33, 138ff liquid level & circulation in separators 81, 82 particularities for plate heat exchangers 56 pressure in evaporators 97 condensers 109 liquid receivers 109 superheat 6, 10, 82ff to zero margin 42 trouble shooting condensers 112 trouble shooting evaporators 76 - 78, 90, 93 water valves in condensers 107 water flow rate at condenser 107 COP, definitions of 8 Corrosion 120ff ammonia 120, 121, 124 by refrigerants, oils & decomposition products 120 corrosive tendencies of water 117 diagram for pitting 123 of copper on steel 122, 123 effect of inhibitors 124 gaskets 165 microgalvanic 122 on copper 122 on stainless steel 121ff Corrugation types of plates 40, 41 Counter current flow in condensation 14, 33 in DX evaporators 20, 21, 58 true or false 63 Counter vs. cocurrent flow flooded evaporators 58, 76 freeze control 130 instabilities in evap. 78 multi-circuit units 64 multi-pass units 60 reversible systems 20ff to limit scaling 120 Crevice corrosion 121ff Definitions 152 Design considerations condensers 100 draining 59, 66, 103, 105 DX-evaporators 88 flooded evaporators 76 flow arrangement 58 fouling 119 freezing 130 refrigerants with glide 144 retrofit with a new refrigerant 142 venting 59, 66, 105, 106ff Desuperheating, & condensers 101 control 107ff general 108, 109 bubble through or injection 110 Direct expansion evaporators 82 characteristic curve 85 hunting & stability 86
168
10. Index.
matching TEV to evaporator 86 stability criterion 88 trouble shooting valve-evaporator 90ff Distributors 90 Draining 59, 66 of condensers 103 -105 of a desuperheater 109 of a multi-pass unit 59 Dual-circuit evaporator 62 - 67 Dynamics of TEV - DX evaporator 86 - 89 Economizers, Evaporator 8 Flash 6 Ejector pumps 73, 79 Electronic expansion valves 94, 95 Equalization lines between condensers & liquid receivers 33, 104 pressure equalization at TEVs 83 Equilibrium, vapour-liquid of a multi-component mixture 145 Evaporators, classification 72, 73 pool boiling units 72 circulating 72 DX: see Direct expansion evaporators flooded: see Thermosiphons for soluble oil 32 once-through 72 Expansion valves,electronic 94 chosing & installation 88 thermosiphons in HP or LP liquid receiver 93 73, 107 thermostatic: see TEV various devices 4 Filter/drier for refrigerants 22, 114, 119 water 119 Flooded flow, see: Thermosiphons Flow boiling units 72 Fouling 114ff chemicals for cleaning and prevention 116, 117 oil 114 pressure drop 115 refrigerant 114 surface 116 water 115ff Freezing 128 control 132ff design to prevent 130 general 128 in reversible units 22 refrigerants with glide 132, 143, 144, 146 prevention in multicircuit units 64, 131 Gaskets 165, 166 Glide, definition 144 freeze control 132 in DX valve settings 143 in reversible units 21 refrigerants with glide 143ff, 164 Heat pumps, comparison HP & AC 21 refrigerant with glide in two-pass of HP|AC 20 comparison of refrigerants 151 Heat recovery 100ff control 110ff Hunting 86 - 90 Hydrazine 124 Identification of an unknown unit 57 Inclined units 69 Inerts 58, 101-103, 106, 107, 112, 142 purge of air pockets 66 see also: Vents Inhibitors, dilution of 124 Injectors, at hot gas by-pass 97 at flooded flow 73, 79 at desuperheaters 110 Instruments, Placement of manometers & thermometers 67, 79 Insulation 69, 129
Isenthalp 2, 3 Isentrop 2, 3 Isobar 2, 3 Isochores 2, 3 K-value, definition of 152, 153 Langliers Index 117 Leaks, causes 125 location of 125ff Liquid hammering Refrigerant side 6, 10, 22, 24, 32, 84, 90 Water side 68, 125, 126 Liquid receivers 104, 105 hybrid 104 installation in reversible systems 24 surge 104, 105 through 105 with, in reversible systems 24 without, 27 LMTD 151 Maldistribution 59 Mean Temperature Difference (MTD) 153 for refrigerants with glide 147 Minimum two-phase flow in thermosiphon pipes 160 MOP, Maximum Operating Pressure valves 85 Mounting positions of BPHE 69 Multi-circuits units 19, 62 - 67, 109 Multi-pass 59, 60ff, 109 reversible units 20 Noise in evaporators 86 Non-condensables, see: Inerts Oil control 138-141 cooling 32, 33, 138 - 141 drains of insoluble, flooded systems 18, 76, 78, 80-81 , DX 139 effect on evaporators and condensers 137 hot gas by-pass 97 properties 134 redistribution of refrigerants 135 separation 78 soluble oil, evaporation 32, 33 suction line separators 98 thermal shocks at cooling 141 Optimization heat exchangers alone 34ff heat exchangers in a system 47 optimization criterion for train of HEs 53 SECOOL 44 supermarket 49ff together with other HE 47ff Ozone 118 Pilot valves expansion at flooded 80 four way at reversible units 23 Piping, exit from flooded evaporators 77 loops 66 oil flow in 139 parallel 66 plastic 150 Pitting 121ff Plate channel types 40, 41 Plate heat exchangers balancing of water flow between parallel 66 brazed, non refrigerants 93 double circuits 64, 65, 67 general properties 54ff installation 68, 69 multi-circuit 62ff multi-pass 60ff one pass 58 pressure drop between parallel condensers 103 venting & draining 59, 66, 105 water filling 66
10. Index.
169
Pool boiling units 72 Pressure drop, two-phase 158ff Pressure-Enthalpy diagram 2 Pump-Down 132 Purging 66 Reentrainment at horizontal separators 155 Refrigerant circuits, double circuits 62ff, 67 multi-pass 60 one-pass 58 Refrigerant-oil behaviour 135 " " distribution 135 " " vapour pressures 143 Refrigerants 140ff ammonia 146 carbon dioxide 148 comparison of some common 151 corrosion by 120 difference between new and old 140 list of 164 retrofit with new 142 with glide 143 with glide, evaporating and condensing temperatures 144 Refrigeration ton RT, definition of 152 Refrigeration cycles, improvement of 8 the basic cycle 2 the cascade cycle 12, 13, 18, 19, 151 the evaporator economizer cycle 8 the intercooler & subcooler cycle 4 the reversible cyles 20 - 27 the flash economizer cycle 6 the subcooler/superheater cycle 10 the water/ammonia absorption cycle 14 the water/lithium bromide absorption cycle 16 Reversed flow at flooded evaporators 76ff, 160 Reversible units 20ff one TEV without LR 23, 27 one TEV with LR 26 two TEV 25 Ryznar Stability Index 117 Scaling tendencies of water 117 Secondary refrigerants, 18, 49, 150 corrosion 122, 124, 150 Secool systems 28 optimization of 44 Separators 79 - 82 design of 154 - 157 for thermosiphons 79 - 82 function 79 horizontal 81, 155 operation 79 suction line separators 98 types 79 vertical 82, 154 Settling, velocity of liquid droplets in vapour 154 of oil 76 Sight glasses 22, 79, 80, 105 Stability curve of an evaporator 87 Stability of a TEV/DX system 86 Stress corrosion 121 Subcooling 22, 102 Suction line separators 98 Subcoolers/superheaters 10, 99 in ammonia absorption units 16 Superheat, control purpose 84 Temperature control at thermosiphons 99 freeze control 133 measurements 55, 67, 95 pinch 21, 61, 101, 108 pinch at refrigerants with glide 147 reversible units, freezing 22 shocks at oil cooling 138ff shocks at parallel connected compressors 98, 138
Temperature profile, condenser control 107ff condensers 101 design to minimize scaling 119 DX evaporator 63, 83 flooded evaporator 75, 77 one pass/two pass units 61 vapour with inerts 103 TEV 82ff bulb filling 84ff by-pass valve to reduce hunting danger 88 characteristic curve of TEV 85 external pressure equalization 82, 83 hunting 11, 86ff, 90 installation of 90, 91 internal pressure equalization 82, 83 matching of a TEV to a DX evaporator 86ff MOP 85 multiple in parallel 98, 99 one or two TEVs at reversible units 24ff part capacity 86 Thermal properties of PHEs 34ff Thermosiphons 19, 31, 33, 73, 74ff design considerations 76 design of the evaporator - separator loop 158ff piping " " 77 instabilities 76 pressure drop balancing 160 temperature program 75, 77 trouble shooting 76ff, 144ff zeotrops 144 Thermostatic expansion valves: See TEV Theta values, adaptation of HE to the duty 43 Traps 129, 139 Trouble shooting, condensers 112 DX evaporators 90 refrigerants with glides 144 thermosiphons evaporators 76 Two-phase flow, design of 158 Valves two in parallel to reduce thermal shocks 33, 138ff pulse width modulated 94 Flow direction in liquid valves 68 thermosiphon 82 thermostatic expansion valves see: TEV Vapour-liquid equilibrium curves 145 Velocity minimum for two-phase upwards flow in a pipe 160, 163 reentrainment in horizontal separators 155, 162 settling in separators 156, 162 Venting 66, 106 and draining of a multi-pass unit 58, 59 ammonia 105, 107 condensate subcooling 102 in connection with liquid receivers 105 in oil coolers 33 location & methods 106 trouble shooting 112 Water analysis, components of 124 chlorinated 122 chlorination 118 demineralized, distilled or soft, surface, well, CT 124 fouling 115ff hot, production of 18-19, 61, 101, 109 -111 hydrazin, with 124 in swimming pools 122 scaling or corrosive? 117 sea water, treatment and use 30, 31 sea water, cooling with 28, 30 types of natural water 115 with chlorides, corrosion in 122 Wetting, of the surface with detergents 17 Zeotrops 144ff
170
2. Equipment:
A plastic or Al box and a plastic bag.
C > 10 cm
B cm
Sw. Eng. Ger. Fr. Ital. Sp. Bot.: Dill Dill Dill Aneth Aneto Eneldo Anethum Grave-olens
A - 2 cm
2. The fish.
The middle part of a salmon.
6. The marination.
B = 10 - 25 cm
Mix all the ingredients except the dill in a large flat bowl. (The sugar is necessary. It does not impart any sweet taste but prevents the salmon from picking up too much salt.) Grey -pink Orange -pink Put in the salmon halves and be sure that they are well covered with the mixture on the surface. Leave the mixture which remains on the salmon, but not more. Put the two salmon halves as shown in the plastic bag, intermixed with three layers of dill and put a weight on the top. Turn it around 1 - 3 times/day. It needs a curing time of at least two full days at a temperature of about 5 to 10 C Addition of some good cognac and/or tarragon is a variant worth trying even though it is not the traditional recipe. Lemon peels and/or crushed juniper berries (enbr, Wachholder, genivre, ginepro, enebro, Juniperus communis) are other excellent additions, both for the marination and the sauce.
Wild Farm
salmon: " :
A cm
The fish should be raw, fresh - neither smell nor red eyes - or frozen. If frozen, it should be as an entire fish, not cut-out parts. If unknown origin, freeze it at first, it kills most - microorganisms. The best salmon is the wild, as found in the rivers of Scandinavia and British isles. However, it has sometimes difficulty in competing with the farmed salmon due to the more vivid colour of the later. It is hard to find, though. Try to get a middle part. Do not wash it.
180
Touch with the middle finger to find spines. Pull out them with a pincher. Cut away the small spines
All ingredients should have the same temperature. Mix sugar, mustard, vinegar & egg yolk. Add oil while stirring vigorously. Add salt & pepper. Take the deepfrozen dill (dill freezes well), it is brittle. It is then easy to draw off the finer parts of the leaves between the thumb and the index finger, leaving the naked stem. If necessary, chop the leaves. Mix the leaves with the sauce.
5. The ingredients.
Prepare: Salt: Sugar: Black pepper: Dill: ~ 100 g ~ 70 - 100 g ~ 10 g (ground) ~ 100 g (Freeze 20 g for the sauce)
6. Serve with:
As entry: As main: Drink: Toast or avocado, lemon and the sauce above. Stewed or boiled potatoes & stewed spinach. Dry white wine or beer & absolut vodka.