Piping systems are the particular province of the mechanical engineer. They are prevalent
throughout his work. The analysis, design and costing of these systems should become as familiar to the
engineer as the multiplication tables.
Piping systems are the particular province of the mechanical engineer. They are prevalent
throughout his work. The analysis, design and costing of these systems should become as familiar to the
engineer as the multiplication tables.
Piping systems are the particular province of the mechanical engineer. They are prevalent
throughout his work. The analysis, design and costing of these systems should become as familiar to the
engineer as the multiplication tables.
Piping systems are the particular province of the mechanical engineer. They are prevalent
throughout his work. The analysis, design and costing of these systems should become as familiar to the
engineer as the multiplication tables.
INTRODUCTION Piping systems are the particular province of the mechanical engineer. They are prevalent throughout his work. The analysis, design and costing of these systems should become as familiar to the engineer as the multiplication tables. A piping system consists of the pipe itself, valves, fittings (elbows, tees, couplings, reducers, expanders, etc.), measuring instruments (temperature and pressure gages, flow meters, etc.) and pumps or turbines. The ultimate task of the engineer is to select appropriate materials and sizes of the various components to meet the objectives of the client. These objectives are generally met by a specification that minimizes the life cycle cost of the system.
THE ENERGY EQUATION SPECIALIZED FOR PIPE SYSTEM ANALYSIS
In general, we want to know how much work must be provided by a pump or produced by a turbine. To determine these work requirements we need to know changes in elevation, head loss due to friction and turbulence, and absolute pressures of the source of supply and the point of delivery. We start with the conservation of energy equation which, from your thermodynamics course, can be stated as follows:
Q W m (h ek e p ) m(h ek e p ) in
out
dE sys
dt
Lets modify this equation to put it in a form useful for analyzing piping systems. First, we are usually concerned with steady state operations; therefore, the last term on the right side becomes zero. Second, lets uncouple the flow work and internal energy terms from the enthalpy and divide by the mass rate to obtain terms based on unit mass. Finally, rearrange the equation to get:
p1
V12 p V2 gz1 w 2 2 gz2 (u2 u1 q) 2
where the subscripts 1 and 2 refer to the source and destination points, respectively. See Figure 1. This is one form of the energy equation. Engineers, however, generally prefer working with the individual terms expressed in feet (or meters) of head obtained by dividing each term by the acceleration of gravity; thus:
V12 p2 V22 hp
z2 hL 2g 1 2g p1
where: hL = head loss = (u2 - u1 - q)/g
hp = -dw = required pump head V1 and V2 are average values, constant over their cross sectional areas. We shall call this equation the pipe system equation.
See, e.g.,Black and Hartley, equation (4.29), p. 164.
We can measure pressure, velocity and elevation with appropriate instruments. Head loss, however, must be determined indirectly from theory, dimensional analysis and experimental work.
DETERMINATION OF HEAD LOSS
The first step in determining head loss is to calculate the Reynolds number. For flow in a circular pipe with a Reynolds number less than about 2100, the flow is classified as laminar, and a purely theoretical analysis results in an expression for head loss as:
LV2 hL f D 2g where:
(1)
64 Re
(2)
If the Reynolds number is greater than 4000, the flow is classified as turbulent and the expression for the friction factor, f, is based on dimensional analysis and experimental results and can be calculated from the Moody chart or the Colebrook formula 3:
/ D 1 2.51 2.0 log
f Re f 3.7
LV2 where, again: h L f D 2g
(3)
(4)
The above results are valid for flow in straight lengths of pipe. Where there are valves, fittings, bends and other components that cause significant changes in flow patterns, losses are accounted for by a loss coefficient, KL , for each class and size of a particular component:
V2 hL K L 2g
(5)
where values of KL can be found from the technical literature and manufacturers manuals. 4
See. e.g., Munson, p. 473.
See, Munson, p. 493,94. 4 See, e.g., Munson, p. 496-505. 3
NONCIRCULAR CONDUITS
For laminar flow, the friction factor can be calculated from:
C Re h where: Re h VDh / f
Dh =hydraulic diameter =4A/P
A = cross sectional area P = wetted perimeter C = constant based on theory/experiment5 For turbulent flow, use the hydraulic diameter for calculating the Reynolds number and the head loss from:
L V 2
hL f Dh 2 g Munson notes that applying circular pipe data to other cross sections will lead to results that are accurate to within about 15%. SYSTEM OPTIMIZATION The optimization problem entails minimizing the present value of total life cycle costs over a selected capitalization period for an assumed interest rate. The total life cycle cost is the sum of three cost categories: 1. Capital cost: material plus installation cost of piping components including the pump. 2. Annual maintenance cost which can be assumed to be a percentage of capital cost, say 10%, unless better data are available. 3. Operation costs: fuel, electric power, and operating personnel costs. The capital cost is a function of the diameter of the pipe and the size of the pump. We can gain an appreciation of these costs from the energy equation, which we rewrite as follows:
p p1 V22 V12
hp 2
z2 z1 hL 2g
(6)
Vj Li Vi 2 KL where: hl f i Di 2 g 2g i j
See Munson, p 508, 9.
(7)
The first term of equation (7) involves a summation if the system consists of lengths of pipe of different diameters and the second term accounts for component losses. If we assume a constant pipe diameter, which may be reasonable for cost comparisons, then the term in brackets is constant for the given system parameters. The variable part of the pump head requirement then depends on the head loss, which depends on the pipe diameter. We can therefore calculate head loss for available, nominal pipe diameters and thus determine the corresponding installed system cost. We then add the present value of the stream of annual maintenance and operation costs. The principal operation cost will be that for the pumps prime mover, either the cost of electricity for an electric motor, steam for a turbine, or fuel for an engine. The cost of electricity can be found by calculating the power input to the motor from known or estimated pump and motor efficiencies applied to the calculated pump head. Or the cost might be obtained directly from a manufacturers computer program such as that for Goulds Pumps (see Figure 2).
SYSTEM CURVE
Once we have established the geometry of the system, we can construct the system characteristic curve. This curve is needed to select a pump and to study the effect of changes in flowrate. Such changes typically result from fouling, which increases friction, or the clients need to operate at different flow rates. The system curve plots required pump head (ordinate) against flowrate (abscissa) (see Figure 2). Referring to equation (6), the first three terms, representing the pressure, velocity and elevation heads, hC, are constant for the given geometry; therefore we can write the required pump head as:
h p hC hL But Q=VA and hL is approximately proportional to V2 for turbulent flow (see equation 4) and V for laminar flow (see, e.g., Poiseuilles Law6). We can therefore write the system curve formula as: hp = hc + KQ2 (turbulent flow) and hp = hc + KQ (laminar flow)
PERFORMANCE CURVES AND PUMP SELECTION
Figure 2. is a set of performance curves from Goulds Pumps computer program. The intersection of the Pump Performance Curve with the System Curve determines the operating point. Where practical, this point should correspond to the best efficiency point on the Efficiency Curve.
NET POSITIVE SUCTION HEAD
After pumps are fabricated, they are tested to establish various useful parameters. One of these is the total head7 at the suction side of the impeller at which cavitation begins. Cavitation occurs when this head is reduced to the vapor pressure head of the fluid. The excess of this head over the vapor pressure 6 7
Equation (8.9), p.468, Munson.
Total head at the impeller eye, with zero elevation, includes both pressure and velocity head.
head is called Net Positive Suction Head Required, abbreviated NPSHR. The system designer must ensure that the excess of head calculated at the impeller eye over that of the vapor pressure head is at least as great as the NPSHR. This head is called the Actual Net Positive Suction Head, NPSHA. Referring to Figure 1., where the subscript s refers to conditions at the impeller eye, we can write the pump equation, corrected for vapor pressure, between points 1 and s as:
P1 V12 Pv Ps Vs2 Pv z1 hL
2g 2g
Or NPSHA = NPSHR Allowing a margin for design and manufacturing errors, NPSHA = NPSHR + 3 ft
