See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/261724835

Use of Cold Air Velocity Test (CAVT) to Locate Erosion Prone Zones in
Pulverized Coal Fired Utility Boiler

Conference Paper · July 2011

CITATION READS

1 1,354

6 authors, including:

Pradyumna Dhamangaonkar Milankumar Nandgaonkar

College of Engineering, Pune College of Engineering, Pune
19 PUBLICATIONS   23 CITATIONS    89 PUBLICATIONS   218 CITATIONS   

SEE PROFILE SEE PROFILE

Abhishek Y. Deshmukh
RWTH Aachen University
25 PUBLICATIONS   35 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Alternative fuel biodiesel View project

Vibration View project

All content following this page was uploaded by Abhishek Y. Deshmukh on 03 September 2015.

The user has requested enhancement of the downloaded file.

Proceedings of the World Congress on Engineering 2011 Vol III
WCE 2011, July 6 - 8, 2011, London, U.K.

Use of Cold Air Velocity Test (CAVT) to Locate

Erosion Prone Zones in Pulverized Coal Fired
Utility Boiler
P R Dhamangaonkar, S R Kajale, M R Nandgaonkar, Abhishek Deshmukh, Aditya Deshmukh,
Swaroop Thakur
 The sources of errors and their effect on the model in deviation
Abstract— Thermal power plants in India are facing the of the conditions from the actual ones are also discussed.
problem of forced outages due to unexpected boiler tube Index Terms— Boiler Tube Failure, CAVT, Fly Ash Erosion,
failures. Fly Ash Erosion (FAE) is one of the prime reasons for Porous Media approach. Utility Boiler,
these failures. The main cause of this FAE is the localized
increase in the velocity of flue gases which in turn increases the
velocity of fly ash particles. The rate and extent of FAE I. INTRODUCTION

T
depends on various parameters such as particle velocity, angle
of impact, particle composition, shape, size & sharpness factor, HERMAL Power Plants contribute about 75% to
temperature of particle, surface temperature of tube, all India installed capacity of electric power generating
population density of particles, coal quality, combustion stations. Coal continues to be the dominant fuel source
efficiency, erosive resistance of the tube surface including for fossil fuel steam generation in the Indian Electric Utility
compositional & temperature variations. Particle velocity is the
most important parameter as the rate of erosive loss is Industry accounting almost 55% of power generated. With
proportional to the velocity raised to an exponent that ranges ever increasing demand for electricity, it is very necessary
between two and four. Particle velocity is driven by the local for the power plants to generate electricity without forced
flow velocity at any particular boiler location. It is difficult to outages. Boiler Tube Failure is the prime reason of forced
physically measure velocities of flue gases inside the boiler outages at coal fired thermal power plants. The severe
when it is in operation. However, it is required to know the service condition in coal fired thermal power plants causes
velocity flow field in various zones so that its effect on the
various failure mechanisms can be predicted. The primary tool
failures as the effects of high temperature, erosion, stress,
to combat Fly Ash Erosion is flow modification in conjunction vibration and corrosion combined to yield failure of the
with a cold air velocity test before and after modification. The boiler tubes.
cold air velocity test is performed to predict the velocities in In pulverized coal-fired power stations, about 20% of the
the respective zones of the boiler. A comprehensive EPRI ash produced in the boilers is deposited on the boiler walls,
Guideline has been published which provides a step-by-step economizers, air-heaters and super-heater tubes. This
procedure of CAVT. In this paper authors have used cold air
velocity technique (CAVT) to determine local velocity profiles
deposited ash is subsequently discharged as slag and clinker
across various pressure parts. The use of CAVT to identify during the soot-blowing process. The rest of the ash is
regions of excessive velocity, followed by the installation of entrained in the stream of flue gas leaving the boiler. These
diffusion and distribution screens, may provide utilities with ash particles collide with the boiler steel components and
the most optimum solution to the problem. The authors have cause extensive surface erosion and may fail once they lose
tried to predict the Cold Air Velocity. The actual geometry of their structural integrity. Such erosion, together with the
the flue gas path of the 210 MW boiler is created using Pro-E
and meshed using GAMBIT is imported in FLUENT for
processes of blocking, fouling and corrosion, shortens the
analysis. The inlet and outlet are given pressure boundary service-life of boiler components. This leads to forced shut
conditions. The k-epsilon realizable is used as turbulence down of power plant in order to replace the damaged
model. The results obtained are then compared with the components. The resulting penalty is not only the cost of
experimental data of CAVT to validate the model. For replacing the components but also the cost of stoppage of
comparing the results, the points where the actual CAVT is power production. It is therefore, required to predict the rate
performed are replicated in the simulated model. Cold Air
Velocity Test is successfully simulated using FLUENT. The
of erosion of the coal-fired boiler components in order to
results of the simulation are in good agreement with the plan systematically for the maintenance or replacement of
experimental results of CAVT with error of order of ±23 %. these components and avoid forced outages.
Leading BTF mechanisms include Corrosion Fatigue, Fly
Manuscript received Mar 09, 2011; revised March 28, 2011. Ash Erosion, Under Deposit Mechanisms (Hydrogen
P R Dhamangaonkar is with the College of Engineering, Pune, Damage and Acid Phosphate Corrosion), Long Term
Maharashtra, PIN 411005, INDIA. (Corresponding author phone: Overheating/Creep, Short Term Overheating, Soot-blower
+919421670869 fax: +912025507299; e-mail: prd.mech@ coep.ac.in).
Dr. S R Kajale is with Shri Guru Gobind Singhji Institute of Engineering
Erosion, Fireside Corrosion (Waterwall, Superheater, and
and Technology, Vishnupuri, Nanded, 431606, INDIA, (e-mail: Reheater), etc. Fly Ash Erosion is one of the important
profsr.kajale@gmail.com) cause of BTF. The rate and extent of FAE depends on
Dr. M R Nandgaonkar is with the College of Engineering, Pune, various parameters viz. particle velocity, angle of impact,
Maharashtra, PIN 411005, INDIA. (e-mail: mrn.mech@coep.ac.in).
Abhishek Deshmukh, Aditya Deshmukh and Swaroop Thakur were with
particle composition, shape, size & sharpness factor,
the College of Engineering, Pune, Maharashtra, PIN 411005, INDIA. temperature of particle, surface temperature of tube,

ISBN: 978-988-19251-5-2 WCE 2011

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
Proceedings of the World Congress on Engineering 2011 Vol III
WCE 2011, July 6 - 8, 2011, London, U.K.

population density of particles, coal quality, combustion

efficiency, erosive resistance of the tube surface including
compositional & temperature variations. Particle velocity is
the most important parameter as the rate of erosive loss is
proportional to the velocity raised to an exponent that
ranges between two and four. Particle velocity is driven by
the local flow velocity at any particular boiler location.
Maximum design bulk velocities are on the order of 15 m/s
or less. It has been observed that local velocities in excess of
30 m/s are required to cause fly ash erosion failures in
10,000 to 50,000 hours.
It is clear that to predict the fly ash erosion; it is required
to know the velocity and temperature of flue gases in
various zones of the boiler. It is not practical to measure the
real time velocities of flue gases in all zones of the boiler
when it is in operation because of cost, clean environment Fig. 1. 3D view of the boiler model. Boiler geometry showing all pressure
parts.
needed for sensing instrument due to presence of flue gas
and fly ash, though the temperatures can be monitored.
Computational Fluid Dynamics was the obvious choice as
the analysis involves fluid flow. However, owing to
complex 3D geometry of the boiler, it was beyond the scope
of the project to develop indigenous code for solving fluid
flow equations. Hence, it was decided to use one of the
commercially available CFD codes like FLUENT, CFX etc.
FLUENT was chosen for the purpose.

II. GEOMETRIC MODELLING OF THE BOILER

A 210 MW pulverized coal fired boiler from one of the
thermal power stations in Maharashtra was considered for
investigation. The boiler components include boiler drum,
tangentially fired furnace, water walls, water wall platen Fig. 2. 2D view of the boiler model. Boiler geometry showing all pressure
parts. The fig shows the domain of interest.
(WWP), primary super-heater (PSH), cold re-heater (CRH),
hot re-heater (HRH), final super-heater (FSH), low The model was developed by replacing the tube bundles
temperature super-heater (LTSH), economizer (ECO) and with the porous block enveloping the tube bundles. This
air pre-heater. Detailed drawings of these components and was done for PSH, CRH-HRH, FSH, LTSH, ECO. LTSH
overall drawing were obtained from the power plant for was divided into two parts LTSH Strip which is like wall of
developing the model. A geometric model of the flue gas single tubes and LTSH Main which contains upper and
path containing the tube bundles of WWP, PSH, CRH- lower tube banks. WWP has only four platens, which means
HRH, FSH, LTSH, ECO was developed to the actual scale it is not as dense as other tube bundles. So, meshing was
using Pro-E Wildfire 4.0. Flue gas path from above of the possible without need to convert WWP to porous medium.
furnace and up to the outlet of economizer is considered. This model was then meshed with tetrahedral cells. (Fig. 3).
Air pre-heater is not considered for analysis. The efforts
were taken to ensure the correctness of the geometry. (Fig. 1 B. Porous Media Approach
& 2) The porous media model can be used for a wide variety
A. Meshing of problems In this case, a tube bank is replaced with a
Meshing is the intermediate step between the geometric porous block having equivalent overall dimensions (Fig. 4).
modeling and the flow analysis. Gambit 2.4.6 is used for When using this model, a cell zone is defined in which the
meshing of geometry. A dense arrangement with large porous media model is applied and the pressure loss in the
number of tubes in each bundle made the model complex flow is determined via user inputs.
and hence was difficult to be meshed with the limited Heat transfer through the medium can also be
computing resources. In order to overcome this situation, it represented, subject to the assumption of thermal
was decided to simplify the model. Literature survey equilibrium between the medium and the fluid flow. The
revealed the porous media concept which could be used to porous media model incorporates an empirically determined
solve such tube banks problems. In this approach, the tube flow resistance in a region of model.
bank is replaced with a porous block having equivalent
overall dimensions and properties like porosity, inertial
resistance, etc. which are calculated from original tube bank
model.

ISBN: 978-988-19251-5-2 WCE 2011

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
Proceedings of the World Congress on Engineering 2011 Vol III
WCE 2011, July 6 - 8, 2011, London, U.K.

The porosity of the porous medium is defined as

void volume in porous medium sample

total volume of porous medium sample

TABLE II
RESULTS OF POROSITY CALCULATIONS
Tube
Void Total
Volume Porosity
Zone Volume Volume
(Solid) 3 3 (φ)
(mm ) (mm )
(mm3)
PSH 8.005E+10 2.11E+11 2.19E+11 0.96352

RH 3.91E+10 4.58E+11 4.97E+11 0.92137

FSH 1.198E+10 6.849E+10 8.047E+10 0.85112

LTSH
5.92E+09 2.54E+10 3.13E+10 0.8112
Strip
Fig. 3. Meshed porous components and flue gas space.
LTSH
In essence, the porous media model is nothing more than 6.591E+10 4.42E+11 5.08E+11 0.87029
Main
an added momentum sink in the governing momentum ECO 4.032E+10 1.80E+11 2.20E+11 0.81682
equations. To balance the momentum loss, correct value of
inertial resistance is given as input in FLUENT.
III. COLD AIR VELOCITY TEST
It is difficult to physically measure velocities of flue
gases inside the boiler when it is in operation. However, it is
required to know the velocity flow field in various zones so
that its effect on the various failure mechanisms can be
predicted. The primary tool to combat Fly Ash Erosion is
flow modification in conjunction with a cold air velocity test
before and after modification. The cold air velocity test is
performed to predict the velocities in the respective zones of
the boiler. In this approach, units are evaluated by the cold
Fig. 4. Boiler Geometry with Porous blocks for pressure parts with air velocity technique (CAVT) to determine local velocity
equivalent inertia resistance. profiles, maximum local velocities of two or more times the
nominal velocity have typically found, and these peak
We cross checked the pressure drop obtained in module velocities usually correspond to the locations of known tube
with input value or else checked the mass flow rate over the erosion damage. The use of CAVT to identify regions of
tube bundles. The mass flow rate should match with the excessive velocity, followed by the installation of diffusion
expected mass flow rate of flue gases or air in case of Cold and distribution screens, may provide utilities with the most
Air Velocity Test. Following table gives the initial optimum solution to the problem. However, the technique
calculations of inertial resistances of various zones in case has not been adopted by sufficient utilities. The results of
of cold air velocity test simulation. The air at atmospheric this CAVT obtained from the concerned power station are
temperature is a flowing fluid over the tube bundles. No used for validating the model developed for predicting the
heat transfer is considered in CAVT simulation between the velocity of flue gases.
flowing fluid i.e. air and tube bundles. More information
A. SIMULATION OF CAVT USING FLUENT
about CAVT is given in subsequent discussion.
The actual geometry of the flue gas path of the boiler
created using Pro-E and meshed using GAMBIT is imported
TABLE I in FLUENT for analysis. The inlet and outlet are given
INITIAL CALCULATIONS OF INERTIAL RESISTANCE FOR COLD
AIR VELOCITY TEST SIMULATION
pressure boundary conditions. The k-epsilon realizable is
PD Velocity Thickness Density I.R.
used as turbulence model. The results obtained are then
Zone compared with the experimental data of CAVT to validate
mmWC m/s (m) (kg/m3) (1/m)
PSH 1 8 1.5577 1.225 0.161 the model. For comparing the results, the points where the
RH 8 9 2.97182 1.225 0.532 actual CAVT is performed are replicated in the simulated
FSH 6 12 0.84773 1.225 0.787 model. The planes considered for this are top of LTSH
LTSH upper bank, bottom of LTSH upper bank, top of LTSH
0.5 13 0.2445 1.225 0.194
STRIP lower bank, bottom of LTSH lower bank and top of
LTSH Economizer.
16.5 13 4.545 1.225 0.344
MAIN
ECO 17 14 2.0395 1.225 0.681

ISBN: 978-988-19251-5-2 WCE 2011

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
Proceedings of the World Congress on Engineering 2011 Vol III
WCE 2011, July 6 - 8, 2011, London, U.K.

the trend of predicted CAVT is more or less similar to that

of experimental results. There is an offset error between the
two.

Fig. 6. Trends of predicted and experimental velocities

Thus, from the above error analysis, though we are not

able to quantify the amount of error involved in the
prediction, we can conclude that the results can be fairly
predicted, on an average, within the range of ±25%. This
error range may reduce to ±10% on proper mathematical
Fig. 5. Positions of Reference Planes
error analysis of above mentioned sources of errors.

IV. RESULTS AND DISCUSSIONS

From the zone-wise comparison of the values of cold air
velocities obtained from onsite CAVT and predicted CAVT,
it can be observed that there is considerable difference.
Furthermore, on carefully observing the graphs it can be
seen that the predicted CAVT gives more or less uniform
distribution of the velocity over the plane under observation.

TABLE III
COMPARISON BETWEEN PREDICTED AND EXPERIMENTAL
Fig. 7. Velocity vectors, Ref. Plane- 0.1 m from left wall
VELOCITIES

Predicted Figures 7, 8 and 9 show the velocity vectors of the cold

Experimental % RMS
Plane Velocity air through the boiler flue gas path. Planes 0.1 m and 13.87
Velocity (m/s) ERROR
(m/s) m from the left wall of the boiler are near the boiler walls.
Top of LTSH Upper These planes pass through the gaps between the tube
1.89 1.49 21.75
Bank bundles and boiler walls. Hence, there is increase in the
Bottom of LTSH velocity of air due to throttling process occurring due to
1.74 1.47 17.17 change in the cross section. Increase in the velocity is more
Upper Bank
prominent in the gaps between the boiler walls and FSH or
Top of LTSH Lower
Bank
1.62 1.43 13.55 LTSH or Economizer. These figures also reveal that there is
increase in velocity near the walls. In the economizer zone
Bottom of LTSH there is increase in velocity near the bends which is
1.89 1.29 31.98
Lower Bank responsible for the erosion.

Top of ECO 1.82 1.34 26.63

On the reference plane “Bottom of LTSH Lower

Bank”, the error is the highest i.e. around 32%. The reasons
for these errors can be attributed to various assumptions and
simplification of theoretical analysis owing to lack of
computational resources. The exact physical phenomenon
has not been modeled while analyzing the flow of cold air
through the boiler over the tube bundles. Detailed error
analysis needs to be done . Few of the sources of error are Fig. 8. Velocity vectors, Ref. Plane- 7.60 m from left wall
(i) Porous Media Approach (ii) Coarse Meshing, (iii)
Geometry errors. However, it can be seen from fig. 6 that

ISBN: 978-988-19251-5-2 WCE 2011

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
Proceedings of the World Congress on Engineering 2011 Vol III
WCE 2011, July 6 - 8, 2011, London, U.K.

REFERENCES
[1] Boshu He, Laiyu Zhu , Jianmin Wang, Shumin Liu,
Baolin Liu, Yanting Cui, Lili Wang , Guoqiang Wei,
“Computational fluid dynamics based retrofits to
reheater panel overheating of No. 3 boiler of Dagang
Power Plant”, Computers and Fluids, 36 (2007) 435–
444, 22 March, 2006
[2] Masoud Rahimi, Abbas Khoshhal, Seyed Mehdi
Shariati, “CFD modeling of a boiler’s tubes rupture”,
Applied Thermal Engineering 26 (2006) 2192–2200, 8
Fig. 9. Velocity vectors, Ref. Plane- 13.87 m from left wall June, 2006
[3] Barry Dooley, Peter S. Chang, “The Current State of
Boiler tube failure in Fossil Plants”, Power Plant
V. CONCLUSIONS Chemistry
Cold Air Velocity Test is successfully simulated using [4] K. S. Bhambare , Sushanta K. Mitra, U. N. Gaitonde,
FLUENT. The results of the simulation are in good Journal of Energy Resources Technology, JUNE 2007,
agreement with the experimental results of CAVT with the Vol. 129, pp. 159-167
error of order of ±23 %. The sources of errors and their [5] E.H. Chui, A. Runstedtler, A. Majeski, “CFD Modeling
effect on the model in deviation of the conditions from the Study of a 500 MW Gas-Fired Utility Boiler”
actual ones are discussed. This model can be used to predict [6] Final Report, Demonstration of the Cold Air Velocity
Technique to Control Fly Ash Erosion at a National
the velocity of flue gases by changing the boundary
Thermal Power Corporation (NTPC) Plant, September
conditions and material properties of fluid flowing over the
2004
tube bundles. Heat transfer model is required to be enabled
[7] S K Das, K M Godiwalla, S P Mehrotra, K K M Sastry
for this. If the flue gas velocities are predicted successfully, and P K Dey, “Analytical model for erosion behaviour
the zones where the velocity of flue gases increases above of impacted fly-ash particles on coal-fired boiler
the critical velocity will be predicted. This critical velocity components”, Sadhana Vol. 31, Part 5, October 2006,
can be calculated for each zone from wear rate equation pp. 583–595.
which has been developed. Erosion rate can be calculated [8] R. P. Roy, M. Ratisher, V. K. Gokhale, “A
based on this equation. From this erosion rate, the life of the Computational Model of a Power Plant Steam
tube can be determined and in turn the failure of the tube Condenser with porous medium”, Journal of Energy
can be predicted. Resources Technology, Vol. 123, pp. 81-91, March
To reduce the erosion, the velocity flow-field may need to 2001
be redistributed. This is possible with the installation of [9] Hsin Chu, “Flame Temperature”, Dept. of
screens near the affected zone. Also the erosion zones which Environmental Engineering, National Cheng Kung
predominantly include economizer can be enclosed in the University.
box. In case of economizer, bends are more prone to erosion [10] Steam: Its Generation and Use, The Babcock and
as the velocity near bends increases due to throttling of the Wilcox Company, Edition 41
flue gases between the walls and the tubes. In case of finned
economizers, the surface area of the straight tubes is far
greater than that of the bends. Hence, enclosing the bends of
economizer will not affect its performance.

ACKNOWLEDGMENT
The authors are thankful to Chief General Manager and
his team, Nashik Thermal Power Station, Nashik,
Maharashtra, INDIA for the co-operation extended.

ISBN: 978-988-19251-5-2 WCE 2011

ISSN: 2078-0958 (Print); ISSN: 2078-0966 (Online)
View publication stats