March 13, 2012 REPORT ON INSPECTION AND PERFORMANCE OF AIR COOLED CONDENSER OF 35 MW UNIT AT JAI BALAJI, DURGAPUR By K.K.

Parthiban, Venus Energy Audit System The visit was made to study the Air cooled condenser unit of 35 MW. The thermograph study was done by our engineer Mr.Muthukumar on 5th March 2012. The temperatures readings were taken at drain pipes and at the surface of ACC outer tube bundles. The data are presented in annexure 3. The velocity measurement was done for all fans at 25 MW and 35 MW unit. The report on air flow calculated is enclosed in annexure 2. The observations and suggestions are furnished in annexure 1. The plant was stopped on 8th March 2012 to check the air cooled condenser inside. The inspection was made inside and the findings are covered in annexure 1. An action plan was given during the visit, which is necessary to improve the ACC performance. AGENDA The ACC of 35 MW unit was experiencing cold often. This was found out by plant staff. The ACC was supplied by Paharpur cooling tower. The problem was pending since long. The problem is to be resolved to make all the modules hot. ACTION PLAN There were many observations made by us during operation and shut down, which could drastically affect the overall performance of the ACC. A list was given immediately to plant engineers. These are time taking actions. It is learnt that some works are still to be completed in the ensuing shut down in May 2012 end. A. Free air & proper air circulation 1. ACC air inlet area to be cleared by 15 metres all around. 2. ACC bottom area to be cleared of all materials so that the air suction is available. 3. ACC bottom water sump is required for ensuring minimum dust infiltration in to the ACC. It also helps to drop the air temperature. It shall be preferable with flow of water. RO reject water can be used here before taking to sponge iron plant. A sump and pump system shall be arranged so that there is continuous flow of water so that there is no mud formation. 4. Floor grill at motor access area can be kept minimum so that air intake area is increased. An approach ladder can be readily fitted when required. This would help in allowing more air in to ACC. 5. Floor to ACC ribbed plate sealing is required. Ribbed plate to plate joint shall be fitted asbestoes felt sealing as done in roofing sheets. 6. Casing to ribbed sheet openings shall be filled with insulation wool and be plastered to ensure no air comes out from ACC inside. 7. Casing to casing joint shall be 100% seal welded. If required flats / filler rods can be added. 8. ACC bundles to supporting structure shall be sealed from inside using asbestoes felt. Shellac shall be liberally used to ensure the asbestoes felt stays inside. 9. The slot holes in the bundle frame are not used at the moment. They shall be plugged with a small plate. The ACC bundle shall not be welded to support as there will be relative thermal

expansion. 10. The side gap between the end casing and the end ACC blocks shall be fully closed using plate and asbestoes felt. 11. ACC block to block gap shall be closed with a plate at top ends where there is no support frame. Light could be seen at the top. See photograph in annexure 1. 12. All doors shall be made leak proof so that ACC fan air would not leak out. 13. The door shall be repaired where the hinge got broken. 14. Floor grills above the ACC fan shall be stacked above the hub area in each ACC section. This will enable free air flow in ACC. Cleaning of ACC tubes 1. Bottom headers of ACC blocks shall be provided with 40 nb drains with plugs at both ends. Each header shall be provided such plugs. This is required for flushing. 2. Water cleaning is suggested to remove loose iron oxide scales from inside of each ACC tube. The plugged tubes shall be made free. For inspection the drain header end shall be cut for inspection. Condensate pH maintenance 1. The condensate pH shall be maintained at 9.5 to 9.6 pH. This is to avoid inside corrosion of ACC bundles. Collecting header modification 1. The present collecting header is 150 nb per side of the ACC bundle. This has to be increased as the velocity works out to be 0.437 m/s. It is advised to split the present header in to two sections. The rear section shall be connected by another 150 nb pipe to the main pipe to condensate tank. Ejector line equalizer valve operation 1. It is seen that that the ejector lines are connected to CST through a common header. Valves are available for isolation. During plant start up these valves can be opened as the vacuum is created b y using hogger ejector which is connected to CST tank. When the hogger ejector is closed, these valves are to be closed to avoid problem in air removal in inner section of each bundle.

DETAILED REPORT About the aircooled condenser The air cooled condenser of 35 MW unit and 25 MW unit are compared below. There is a design discrepancy on the air temperature and exhaust steam temperature. Parameter Ambient air temperature Condenser vacuum Corresponding turbine Unit Deg C ata exhaust Deg C 25 MW unit 30 0.13 51 35 MW unit 36 0.22 62

steam

temperature Site elevation

Metres above MSL

30

26.4

The ACC sizing is based on the air & steam temperatures. Hence there will be undersizing of surface area if lower ambient temperature is specfied. Higher the ambient air temperature, more will be the surface area required. It is learnt that the plant was under performing due to vapor locks inside the ACC. The supplier had suggested to increase the fan capacity. It was learnt that some of the drain hoses were replaced earlier. The 35 MW unit was stabilised in April 2010. So far the unit had barely touched 25 MW. pH for ACC The air cooled condenser tubes are made of carbon steel tubes. The condensate pH is maintained by dosing ammonia / morpholine / cyclohexylamine. These chemicals are volatile in nature. The travel with the steam and combine with condensate. That is how, the pH of the return condensate is close to that of feed water. Each chemical has a volatility ratio. When the condensation begins at top, the first condensate has low pH since the amount of chemical would be less due its volatility. As the condensation process is complete at bottom, the condensate pH would be 8.5 or more as per the chemical dosage practice. Since the pH is less in the upper part of ACC, corrosion products are generated. These can plug the drain plugs. This can be reason for poor heat transfer as well. We need to raise the pH and wait for the inside of ACC to become normal. This can take a month too. In annexure 4, an article is enclosed on the pH requirement at ACC. In annexure 1 the debris removed from ACC bottom header can be seen. In addition, the choking occuring at CEP confirms that the condition of ACC is not good. Preservation of ACC during off-line period The ACC duct at top was found to have corroded and generated flakes of rust. This happens due to presence of water in CST when the ACC is off line. The rust flakes are large enough to block the ACC tubes during the operation. These flakes could block the small bore condensate drain pipes provided at the bottom headers. The square headers can also accumulate the rust inside. There is no way it can be cleaned well. Drains plugs are required, so that water flushing can be attempted. During off line period, the CST tank must be left dry with trays of limestone for absorbing moisture. Rightly a dehumdification system is required for preserving ACC tubes. Compressed air can be used for filling the ACC and keep it pressurised at pressure, say, 50 mmWC. Design parameters of ACC The design parameters selected for ACC is very conservative. The vacuum selected is only 0.78 kg/cm2 (a) as against 0.80 kg/cm2 (a). The steam temperature is chosen as 62.4 deg C and ambient temperature is chosen as 36 deg C. Lesser the difference between the steam temperature and the air temperature, more will be the heating surface requirement. Ideally the selection should have been made for a condensate temperature of 56 deg C & ambient temperature of 40 deg C. Table 2 gives a comparison of a successfully operating 7.5 MW & 20 ACC with 35 MW ACC. We can see the sizing

impact. Also the 25 MW and 35 MW are compared and found to have an undersized HTA. Hence vaccum will be a problem at full load. Bundle arrangement In this unit it is seen that the four bundles have been packed one over other. The first bundle has high T over the outer bundle. The effectiveness of the HTA in a upper bundle, comes down as the air gets heated in the previous bundle. The fourth bundle effectiveness will be the least. This can affect the draining of condensate as the driving force - T will not be efficient. In most of the units, three bundles are used. Four bundles was a wrong choice. This must have been doen to avoid two additional cells. Draining arrangement In other installations the collecting header to common condensate header is thorugh a large drain. Here it is seen that the drain pipes are smaller. It is necessary to add two more collecting hoses, preferable at the ends. In addition the common condensate header pipe has to be made in two sections to reduce the water side velocity. Small cross sections are prone to vapor locks. Additional drain would be of use on this aspect. Ejector piping arrangement In other installations the air removal pipe size is larger. Larger diameter pipes will have less pressure drop. The velocity and pressure drop are calculated and presented in annexure 5. It is worth drawing the air from both front and rear side of ACC. At present each bundle in each module is having two tubes devoted to air removal. In other designs one section in the middle is is devoted for air removal. The pressure drop in these units are lesser. While using smaller pipes, we would have blockages due to offline corrosion. The material of construction for air handling pipe must be SS 304. This is due to oxygen & wet steam handled in this pipe. It is not clear what is the material used. Air lock / dirt lock in ACC bundles Thermograpy camera was used by us to find out stagnation of water in bundles. It was clear that the drains were getting choked up. It was not like some tubes in the bundle were cold and some were hot. Since the drain was choked up, the entire bundle were found cold. There was no case that the drain was clear and part of the bundle was cold. It is not a vapor lock in some bundles. It is a flow lock for all the tubes of the bundle. Ejector inter connection to common hogger The design of ACC is with separate ejectors for inner bundle, bundle 2, bundle 3 and bundle 4. All inner bundles would condense more steam due higher T. Thus more non condesible gases would be handled by the 65 nb ejector than the remaining 3 x 50 nb ejectors. The valves connecting to common hogger should be in closed condition so that the vacuum is better. This is an operational

modification. Leakage of ACC casing The air from the ACC fan leaks directly at the casing around the fan. There are several locations of leakages. Many leakages were attended in our engineers presence. The locations of air leakages are brought out in the annexure 1. Free air & proper air circulation The space below ACC was to be cleared first to ensure the air circulation is available. The status has to be checked now. ACTIONS RECOMMENDED / TAKEN AND THE RESULTS The common collecting header end plate was cut to inspect the presence of choking material. The photograph in annexure 1 shows what was seen. The corrosion products were seen right below the drain connectiion at the common collecting header. ACC tubes were inspected by going inside the inspection door at the top header. Bottom headers are too small such an internal inspection. Hence its condition is not known. The ACC tubes were undergoing corrosion. The tubes and ducts are compared with another installation in annexure 1. Looking at the blockage and loose material inside the ACC tubes, water flushing was taken up immediately. The water connection was made from the CEP and used for cleaning the inside tube of each bundle. When the ACC was restarted after the shut down, the drains were getting choked again. Hence it was advised to use hogger ejector as often as required, to carryout dechoking on line. This had helped lot. It was learnt that the dechoking was almost complete after a month of this operation. CEP choking was indicating the corrosion product present in ACC.

CONCLUSION 1. The cause for cold ACC sections was identfied to be partly due to wrong pH regime. 2. The number of drains have to be increased at the drain header. As the drain header is small in size, only increasing the number of drains is possible. 3. The ACC design basis was not correct. The high condensate temperature and low ambient temperature has resulted in undersizing of the ACC. 4. The air circulation was checked with many materials inside the ACC space. Now it is learnt that the ACC base is cleaned up. A revised study is required to find out improvement in airflow. Or else the air flow can be increased as proposed by supplier. 5. It is necessary to add cells for the shortfall. There are units where additional cells have been added. This is to be done after the completion of revised study after incorporation of additional drains and after modification of air ejector piping.

K.K.Parthiban

ANNEXURE 1: PHOTOGRAPHS & COMMENTS

Photo 01 & 02: These are the photographs showing the space below ACC. These obstructions for free air circulation should be removed immediately. There should be 15 m free spacing all around for free air circulation.

Photo 03: These are the photos taken under the ACC fans. Due to this obstructions, the fan flow will not be uniform. This should be removed immediately.

Photo 04: There should have been 15 m free space all around. The ACC to TG building space is only 10 meters.

Photo 05: At present there is no water sump under the ACC fans. Dust ingress to ACC can be reduced if a water sump is added. We can see the dust pick marks in the floor.

Photo 06: The water trap arrangement provided elsewhere under ACC. Garden water / cooling water to kiln can be routed through the water trough here. 200 mm height of water is recommended.

Photo 07: The photograph showing the ribbed casing joint in ACC. Here the joint is not closed. Due to this air is leaking from this joint. This should be arrested by asbestos sealing as shown in the following photo. Also cement mortar sealing is required for ACC floor to ribbed sheet.

Photo 08: Hand sketch showing ribbed plate to plate joint by asbestoes felt with shellac.

Photo 09: The photograph showing the gap found between the ribbed sheet and the casing. Due to this gap air is leaking from inside. This shall be arrested to in order to reduce the passing of air outside.

Photo 10: Hand sketch showing the casing to ribbed sheet gap filling detail. After the insulation is filled, plastering should be done above to ensure no air leakages.

Photo 11: The photograph shows the incomplete sealing in the ACC casing. These shall be sealed as illustrated below.

Photo 12: Hand sketch showing casing to casing closing detail for ACC. Stitch welding is OK to arrest the leakage.

Photo 13: The photograph the air leakage mechanism at ACC at ribbed casing.

Photo 14: The ACC Inspection doors in between the ACC cells & outside should be in closed condition always. The leaky doors would over reduce the cooling capacity of ACC.

Photo 15: There is air leakage at the side casings to front & rear casing. These must be arrested.

Photo 16 & 17: Light seen between the ACC modules. These must be plugged. The bundle casing to casing can be closed from inside only with an overlap plate.

Photo 18 & 19: Photographs showing the debris generated by the ACC on wrong condensate chemistry. An article is enclosed informing the condensate pH monitoring.

Photo 20: Sealing required between the bundles. 75 mm wide x 4 mm thick overlapping plate shall be welded from inside. The light is visible from these joints when see from inside ACC.

Photo 21: ACC duct showing the corrosion products generated during offline storage. Dehumidified air is to be circulated during offline storage. Even compressed air is adequate as it is moisture free. Large amount of lime can be kept inside the ACC duct and at CST.

Photo 22 & 23: The Photograph on left showing the general pH corrosion. Photograph on right shows the oxygen corrosion.

Photo 24 & 25: The photographs show the corrosion of ACC tube. One tube is seen in plugged condition.

Photo 26: Photograph shows the four rows of condenser tubes. This is not good. As the air becomes hotter from one stage to another, the heat transfer reduces. Thus the effectiveness of last row will be less due to high inlet air temperature. Generally only three tubes are arranged by many manufacturers.

Photo 27: Photograph showing the best ACC internals with very grey surface at inlet of the ACC tubes. In this plant the pH was raised to 9.6 -9.8 as per our recommendations and the iron in condensate was the same as that of the feed water.

Photo 28: The Inside of ACC duct seen here. The rust flakes would travel to ACC tubes and end up plugging the 40 nb drain hoses provided.

Photo 29: Corrosion products collected at ACC duct. This is due to poor offline storage of ACC.

Photo 30: The photograph shows the low pH corrosion of ACC duct. The surface is being continually etched.

Photo 31: The surface of ACC duct is seen with grey surface at another plant after the pH was raised to 9.6 to 9.8. This plant uses all volatile treatment, in which the filming amines do this job. About 30% of pH booster is admitted at turbine exhaust duct..

Photo 33: The photograph showing awful condition of ACC tubes. The oxide scales reduce the heat transfer drastically. After adjusting the pH and on continuous running, we have to pull it out of the system. Water washing would help to a great extent. But high pressure water jetting is a must.

Photo 34: Debris collected at condensate pipe at ACC bottom. With this kind of continuous corrosion the life of the ACC would come down drastically.

Photo 35: The drain size in other ACC units. Any scale will drop in to the condensate piping below.

Photo 36: The hoses are too small for the long header. There would not be proper draining of water from the headers. Each header shall be provided with two more drains at the ends.

Photo 38: Air removal piping from bundles of modules. The air piping is now connected to front side. In order to gain some head the air piping can be taken from rear side also to the ejectors. It would reduce the pressure drop.

Photo 39: The photograph shows the additional valves added from each non condensable air piping. These valve should be in closed condition on regular operation.

ANNEXURE 2: AIR VELOCITY MEASUREMENT AT ACC FANS 35 MW & 25 MW

GENERAL METHOD

Air Velocity Measurement at 35MW & 25MW:

Input: Area Area1,m2 Area2,m2 Area3,m2 Area4,m2 35MW 4.48 2.68 2.42 9.34 25MW 4.36 2.59 2.34 6

Notes: 1.Thevelocityismeasuredfromthefanoutercasingring. 2.Thedistancefromwhichvelocityismeasuredis200,600, 1000&1400mm. 3.Fourareawillbecalculatedwithrespecttofourvelocity notedforeachquadrantoftotalfanarea. 4.Flowratewillbecalculatedforeachareainthequadrant dependinguponthevelocityobtainedatthearea area.Summing offlowinallfourquadrantsgivesthetotalflowofthefan. CurrentconsumptionofFans: Fanspeed: FanNo 35MW 25MW FanNo 35MW 25MW Fan#1,Amps 135 110 Fan#1 1480 1485 Rpm Fan#2,Amps 132 100 Fan#2 1490 1485 Rpm Fan#3,Amps 145 Fan#3 1430 1485 Rpm Fan#4,Amps 136 95 Fan#4 1490 1485 Rpm Fan#5,Amps 130 100 Fan#5 1485 1485 Rpm Fan#6,Amps 135 100 Fan#6 1485 1485 Rpm

MEASUREMENT AT 35 MW

AirvelocitymeasurementsinACCfans35MW Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat FanNo Quadrant measured,mm m/s measured,m2 m3/s eachQuadrant 200 4 4.48 17.92 4.5 2.68 12.06 600 m3/s I 135.752 1000 8.2 2.42 19.844 9.2 9.34 85.928 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
5.5 6.4 7.2 8.1 4.5 6.5 8.2 8.5 5.5 8.2 9.3 9.5

II
Fan # 1

4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34

24.64 17.152 17.424 75.654 20.16 17.42 19.844 79.39 24.64 21.976 22.506 88.73

134.87

m3/s

`III

136.814

m3/s

IV

157.852

m3/s

Total air flow from the Fan is

565.288 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat measured,mm m/s measured,m2 m3/s eachQuadrant 200 4 4.48 17.92 4 2.68 10.72 600 120.3 m3/s 7 2.42 16.94 1000 8 9.34 74.72 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
2.5 4.8 7.5 9.8 6.5 6.8 8.1 9.5 7.5 10 11.1 10.8

II
Fan # 2

4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34

11.2 12.864 18.15 91.532 29.12 18.224 19.602 88.73 33.6 26.8 26.862 100.872

133.746

m3/s

`III

155.676

m3/s

IV

188.134

m3/s

Total air flow from the Fan is

597.856 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat measured,mm m/s measured,m2 m3/s eachQuadrant 200 5 4.48 22.4 7.8 2.68 20.904 600 149.802 m3/s 8.5 2.42 20.57 1000 9.2 9.34 85.928 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
5 6.1 7.3 8.5 5.5 6.8 8 9 5.8 6.8 7.8 9.5

II
Fan # 3

4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34

22.4 16.348 17.666 79.39 24.64 18.224 19.36 84.06 25.984 18.224 18.876 88.73

135.804

m3/s

`III

146.284

m3/s

IV

151.814

m3/s

Total air flow from the Fan is

583.704 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat measured,mm m/s measured,m2 m3/s eachQuadrant 200 3.4 4.48 15.232 4.5 2.68 12.06 600 135.866 m3/s 8.2 2.42 19.844 1000 9.5 9.34 88.73 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
2.5 4.8 7.6 8.2 3.6 3.4 6 8.8 7.2 8.5 10.5 10.6

II
Fan # 4

4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34

11.2 12.864 18.392 76.588 16.128 9.112 14.52 82.192 32.256 22.78 25.41 99.004

119.044

m3/s

`III

121.952

m3/s

IV

179.45

m3/s

Total air flow from the Fan is

556.312 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat measured,mm m/s measured,m2 m3/s eachQuadrant 200 2.4 4.48 10.752 4.2 2.68 11.256 600 134.492 m3/s 7.5 2.42 18.15 1000 1400 10.1 9.34 94.334 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
6.1 9.3 8.3 8.6 6.5 7.3 8.5 9 4.2 6.5 8.5 9.5

II
Fan # 5

4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34

27.328 24.924 20.086 80.324 29.12 19.564 20.57 84.06 18.816 17.42 20.57 88.73

152.662

m3/s

`III

153.314

m3/s

IV

145.536

m3/s

Total air flow from the Fan is

586.004 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat measured,mm m/s measured,m2 m3/s eachQuadrant 200 2.5 4.48 11.2 3.5 2.68 9.38 600 128.186 m3/s 7.8 2.42 18.876 1000 9.5 9.34 88.73 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
3.8 6.8 7.8 8.5 4.5 6.4 8.5 9.8 6.5 10 10.5 10.6

II
Fan # 6

4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34 4.48 2.68 2.42 9.34

17.024 18.224 18.876 79.39 20.16 17.152 20.57 91.532 29.12 26.8 25.41 99.004

133.514

m3/s

`III

149.414

m3/s

IV

180.334

m3/s

Total air flow from the Fan is Total air flow from all the ACC Fans as per Design is Total air flow from all the ACC Fans at actual is Shortage of air in the ACC fans is

591.448 m3/s 3614 m3/s 3480.612 m3/s 133.388 m3/s

MEASUREMENT AT 25 MW

AirvelocitymeasurementsinACCfans25MW Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat FanNo Quadrant measured,mm m/s measured,m2 m3/s eachQuadrant 200 9 4.36 39.24 9.2 2.59 23.828 600 I 134.66 m3/s 1000 8.8 2.34 20.592 8.5 6 51 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
5 5.3 6 6.8 1.3 2.5 6.2 7.8 6.2 7.4 8.5 9.8

II
Fan # 1

4.36 2.59 2.34 6 4.36 2.59 2.34 6 4.36 2.59 2.34 6

21.8 13.727 14.04 40.8 5.668 6.475 14.508 46.8 27.032 19.166 19.89 58.8

90.367

m3/s

`III

73.451

m3/s

IV

124.888

m3/s

Total air flow from the Fan is

423.366 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat measured,mm m/s measured,m2 m3/s eachQuadrant 200 6.2 4.36 27.032 7.5 2.59 19.425 600 109.841 m3/s 7.6 2.34 17.784 1000 7.6 6 45.6 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
2.5 4.2 6.4 6.8 5.5 6.2 7.5 7.6 7.5 7.6 8 8.4

II
Fan # 2

4.36 2.59 2.34 6 4.36 2.59 2.34 6 4.36 2.59 2.34 6

10.9 10.878 14.976 40.8 23.98 16.058 17.55 45.6 32.7 19.684 18.72 50.4

77.554

m3/s

`III

103.188

m3/s

IV

121.504

m3/s

Total air flow from the Fan is

412.087 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat measured,mm m/s measured,m2 m3/s eachQuadrant 200 7.2 4.36 31.392 7.8 2.59 20.202 600 122.382 m3/s 8.2 2.34 19.188 1000 8.6 6 51.6 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
6.5 7.8 7.8 9 7.5 8.4 9.3 9.2 5.4 6.5 7.5 8.5

II
Fan # 3

4.36 2.59 2.34 6 4.36 2.59 2.34 6 4.36 2.59 2.34 6

28.34 20.202 18.252 54 32.7 21.756 21.762 55.2 23.544 16.835 17.55 51

120.794

m3/s

`III

131.418

m3/s

IV

108.929

m3/s

Total air flow from the Fan is

483.523 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat m3/s eachQuadrant measured,mm m/s measured,m2 200 3.8 4.36 16.568 5.5 2.59 14.245 600 97.797 m3/s 7.6 2.34 17.784 1000 8.2 6 49.2 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
5.5 6.3 7.4 7.6 4.5 6.5 8 8.2 6.5 7.2 8.5 7.9

II
Fan # 4

4.36 2.59 2.34 6 4.36 2.59 2.34 6 4.36 2.59 2.34 6

23.98 16.317 17.316 45.6 19.62 16.835 18.72 49.2 28.34 18.648 19.89 47.4

103.213

m3/s

`III

104.375

m3/s

IV

114.278

m3/s

Total air flow from the Fan is

419.663 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat measured,mm m/s measured,m2 m3/s eachQuadrant 200 4.5 4.36 19.62 6.5 2.59 16.835 600 101.945 m3/s 8.5 2.34 19.89 1000 1400 7.6 6 45.6 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
7 8.5 9 9.2 3.3 6.2 7.8 8 4.2 6 7.2 8.5

II
Fan # 5

4.36 2.59 2.34 6 4.36 2.59 2.34 6 4.36 2.59 2.34 6

30.52 22.015 21.06 55.2 14.388 16.058 18.252 48 18.312 15.54 16.848 51

128.795

m3/s

`III

96.698

m3/s

IV

101.7

m3/s

Total air flow from the Fan is

429.138 m3/s

FanNo Quadrant

Distanceatvelocity Velocity, Areaatvelocity Discharge, Dischargeofflowat eachQuadrant measured,mm m/s measured,m2 m3/s 200 2 4.36 8.72 2.5 2.59 6.475 600 72.027 m3/s 4.8 2.34 11.232 1000 7.6 6 45.6 1400 200 600 1000 1400 200 600 1000 1400 200 600 1000 1400
7.5 8.2 8 8.5 7.2 8.5 8.6 8.3 5.5 4.4 6.4 7.2

II
Fan # 6

4.36 2.59 2.34 6 4.36 2.59 2.34 6 4.36 2.59 2.34 6

32.7 21.238 18.72 51 31.392 22.015 20.124 49.8 23.98 11.396 14.976 43.2

123.658

m3/s

`III

123.331

m3/s

IV

93.552

m3/s

Total air flow from the Fan is Total air flow from all the ACC Fans as per Design is Total air flow from all the ACC Fans at actual is Shortage of air in the ACC fans is

412.568 m3/s 2818 m3/s 2580.345 m3/s 237.655 m3/s

` ANNEXURE 3: THERMOGRAPH STUDY ON AIR COOLED CONDENSER OUTER BUNDLE. Date of test: 5th March 2012

Photo 01: The hand sketch shows the arrangement of ACC of 35 MW as viewed from TG side. View AA shows at how many locations the thermograph reading is taken. Totally there are 9 bundles. In those 9 bundles 7 rows were divided equally and the temperature was measured. Therefore totally 63 locations the readings were taken and presented in other photographs.

Photo 02: This hand sketch shows the view-BB. There are 4 sections for each bundle. Here individual drain pipe / pipes are given from the drain box to connect drain header. Here also the temperature is measured. At some places it was cold proving there was no condensate flow. Like this for 9 bundles (column wise) the temperatures were checked. Therefore totally 36 readings were noted. The inner sections are supposed to condense more. But there was no flow of condensate.

Photo 03: This is the continuation of temperature values measured at individual drain hoses from drain boxes to common drain header. (Refer comment in photo: 02). From the readings we can observe that the ACC bundles on WHRB side alone are working completely. This may be due to better air availability for the 2-left side bundle.

Photo 04: These were the surface temperature readings at 1-Right side wall of ACC. Refer photo: 01 for legend.

Photo 05: These were the surface temperature readings at 1-Left side wall of ACC. Refer photo: 01 for sketch.

Photo 06: These were the surface temperature readings at 2-Right side wall of ACC. Refer photo: 01 for sketch.

Photo 07: These were the surface temperature readings at 2-Left side wall of ACC. Refer photo: 01 for sketch.

Photo 08: The hand sketch showing the plan view of fan with its orientation as view from TG building front. At the time of test, the wind direction was from south-west.

Photo 09: These were the photographs from IR camera. At some bottom rows of bundle it was cold in condition. The last two photos show the drain pipe stub to drain header. This means there is no flow of condensate downwards.

Photo 10 & 11: It was learnt that some of the pipes were changed due to air leakage. As the ACC is bottom supported, there is no need for flexible piping. We can add flexibility. The hose temperatures were measured by camera at the locations shown. As per the report, there was no flow in many drains. It is doubted that the stubs are choked due to ACC corrosion products. Had the line size was bigger as in other ACC units elsewhere, there would not be any choking. The inner most section of the bundle would collect more condensate. The outer most would collect less condensate due to hotter air.

ANNEXURE 4: ARTICLE ON AIR COOLED CONDENSER pH REQUIREMENT

PPChem

Assessing and Controlling Corrosion in Air-Cooled Condensers

Assessing and Controlling Corrosion in Air-Cooled Condensers

R. Barry Dooley, J. Denis Aspden, Andrew G. Howell, and Francois du Preez

ABSTRACT
An increasing number of air-cooled condensers (ACC) are being installed and operated on conventional and combined cycle plants worldwide. Unless understood and corrected, the corrosion associated with the ACC ducts and tube entries can become a major problem for operators of plant. Up to just a few years ago very little was known about the corrosion/ flow-accelerated corrosion (FAC) process. This paper starts to rectify the situation with a description of the corrosion/FAC process, a corrosion index and a relationship between the operating pH and the level of iron at the condensate pump discharge.

INTRODUCTION
Since the installation of the first small 1 MW dry cooling system in 1939, many hundreds of air-cooled condensers (ACC) have been erected worldwide. Corrosion of the carbon steel components in these large systems has been a concern because of the impact of high iron levels and air in-leakage [1]. The maximum corrosion has been observed at the entries to the A-frame ACC tubes. The mechanism of this corrosion is not fully understood and little work has been expended in trying to rectify this. As far as the authors know, only two ACC tubes have been removed for metallurgical examination. Based on these two analyses the actual corrosion appears to be a flow-accelerated corrosion (FAC) derivative where local indigenous magnetite is removed from the surface of the ACC tube leaving a very intergranular surface appearance. Adjacent to these areas where the local turbulence of the two-phase media is not as great, the magnetite deposits on the surface. There are clear boundaries between the regions where corrosion/ FAC takes place (white bare metal) and regions where deposition (black areas) occurs. Figure 1 shows one of the two ACC tubes which have been subjected to a metallurgical analysis. The low magnification photograph of the surface of the ACC tube near to the tube entry clearly shows the adjacent black and white areas. The other two parts of the figure show details of the black areas, which are essentially deposited magnetite (upper part, left), and the white bare metal area, which has an intergranular appearance (upper part, right). These metallurgical analyses have provided the first detailed investigations of the interfacial science taking place between the fluid and the ACC duct and tube surfaces. It is anticipated that more tubes will become available to assist in the development of a fuller understanding of the corrosion and deposition mechanisms. However, in the interim operators of ACC need guidance in operation so that the corrosion/FAC can be minimized. This paper describes two major steps which have been taken over the last few years to quantify ACC damage: a) an index to assess the severity of the corrosion/FAC process throughout the ACC system from the steam turbine exhaust to the collected condensate, and b) a relationship between the cycle chemistry (pH) and the level of corrosion/FAC. A parallel paper [2] outlines an interim cycle chemistry guidance for ACCs developed for EPRI.

ACC CORROSION INDEX

After inspecting a number of ACCs around the world, two of the authors developed an index for quantitatively defining the internal corrosion status of an ACC. This is known by the acronym DHACI (Dooley Howell ACC Corrosion Index) [3,4]. The index separately describes the lower and upper sections of the ACC, according to the following: Upper Section (upper duct/header, ACC A-frame tube entries). Index: 1, 2, 3, 4 or 5. Examples are shown in Figure 2. 1. Tube entries in relatively good shape; possibly some areas with dark deposits in first few inches of tube interior. No corrosion or FAC. 2. Various black/grey deposits on tube entries as well as flash rust areas, but no white bare metal areas. Minimal corrosion/FAC. 3. Few white bare metal areas on a number of tube entries. Some black areas of deposit. Mild corrosion/FAC. 4. Serious white (bare metal) areas on/at numerous tube entries. Extensive areas of black deposition adjacent to white areas within tubes. Serious corrosion/FAC. 5. Most serious. Holes in the tubing or welding. Obvious corrosion on many tube entries.

2009 by PowerPlantChemistry GmbH. All rights reserved.

264

PowerPlant Chemistry 2009, 11(5)

Assessing and Controlling Corrosion in Air-Cooled Condensers

PPChem

Figure 1: Details from the examination of an ACC tube section within about 38 cm (15 in) from the tube entry in the upper duct. The lower photograph shows the adjacent white and black areas on the tube surface. The two upper parts of the figure show surface metallographs of the black area after cleaning of the surface (left) and white bare metal (right).

Lower Section (turbine exhaust, lower distribution duct, risers); rate A, B or C. Examples are shown in Figure 3. A. Ducts have no indication of damage and are grey in color. B. Minor white (bare metal) areas on generally grey ducts. Some 'tiger striping' appearance with darker grey/ black areas demonstrating two-phase flow-accelerated damage. Overall assessment is 'mild damage'. C. Multiple, widespread areas of bare metal in the turbine exhaust and at abrupt changes in flow direction (e.g. where steam flow enters vertical riser from lower distribution duct). White (bare metal) areas are evident indicators of metal loss. Severe local damage exists. The index provides a number (from 1 to 5) and a letter (from A to C) to describe/rank an ACC following an inspection. For example, an Index of 3C would indicate mild corrosion at the tube entries, but extensive corrosion in the lower ducts. The DHACI can be used to describe the status of a particular ACC in terms of its corrosion history. A poor rating (e.g. 4B) indicates the need to consider options to reduce the corrosion rate, especially in the tube entry areas.

Additionally, the index provides a convenient tool for comparison between different units. This can aid in determining whether some factor in effect at one station, e.g. use of an amine rather than ammonia, is yielding better results. Finally, the DHACI is a very useful means of tracking changes that occur as a result of making changes in steam cycle chemistry. A plant that has a relatively poor rating for corrosion at a steam cycle pH of 8.58.8 (e.g. 4C) may increase the pH to 9.49.6 and determine whether this change improves its rating (e.g. 3B). Similar tracking may be done in association with mechanical or material changes, or with major shifts in operating pattern (cycling vs. baseload operation).

DEVELOPMENT OF RELATIONSHIP BETWEEN pH AND CORROSION IN ACC AND THE RAMIFICATIONS FOR CONDENSATE POLISHING SYSTEMS AND THE USE OF AMINES
The current understanding of the mechanism associated with the corrosion of ACCs has been outlined in the introduction. Regardless of the mechanism itself both chemistry data and visual evidence suggest that the pH

PowerPlant Chemistry 2009, 11(5)

265

PPChem

Assessing and Controlling Corrosion in Air-Cooled Condensers

Figure 2: Montage illustrating DHACI indices 15 for the upper ducts and tube entries.

environment plays a significant role in the corrosion process. Currently data is only available from organizations which employ ammonia as the feedwater conditioning chemical for pH control. It has also been discussed that the two-phase conditions present in the upper ducting of the ACC are likely to contribute to the corrosion process at the tube entries.

In two-phase conditions ammonia much prefers the vapour phase over the liquid phase. Therefore in the liquid droplets, associated with the first condensate, that come into contact with the ducts and condensing tube inlets there is far less ammonia present than that originally injected into the condensate/feedwater of the plant. Note also that some of the ammonia vapour will be removed in the dephlegmator of the ACC. As a result, where these

266

PowerPlant Chemistry 2009, 11(5)

Assessing and Controlling Corrosion in Air-Cooled Condensers

PPChem

Figure 3: Montage illustrating DHACI indices AC for the lower ducts from the steam turbine to the vertical risers to the upper duct.

liquid droplets come into contact with the ducts, in particular the tube entries, the solution pH can be as much as 1 pH unit lower than the condensate/feedwater pH at 25 C (77 F). This has been confirmed by measurement [5]. More importantly the differential between the solution pH and the neutral pH (at temperature) is significantly reduced. With this reduction in alkalinity and under the influence of locally increased turbulence, iron oxide (magnetite) dissolution is increased. Increasing the ammonia concentration in the condensate/ feedwater of the plant will increase the first condensate pH of the droplets and thus reduce the oxide solubility. However, because of the effect of the reducing basicity of ammonia with increasing temperature and the preferential vapour phase distribution of ammonia, a great deal more ammonia has to be added. Data gathered from several organizations around the world has resulted in the development of the Dooley/Aspden relationship for Fe in the condensate at the condensate pump and pH associated with units having ACC [3]. This is illustrated in Figure 4, which also shows the corresponding curve for magnetite solubility extracted from Sturla's mag-

netite solubility data [6] at 50 C (120 F). The steam/condensate temperature throughout the ACC is for all practical purposes constant since the ACC only rejects the latent heat in the exhaust steam. The steam temperature depends on the ambient temperature and the design of the ACC. The minimum and maximum design steam temperatures, and thus the steam temperature at the tube entries, are usually between 45 and 80 C. So the 50 C used in Figure 4 is near the lower end of the scale, but the solubility of magnetite varies very little up to 80 C. The band in Figure 4 is about 5 g kg1 wide to allow for uncertainty associated with the data from many organizations, but does have a strong track record around the world for units with ACCs using only ammonia conditioning for pH control. The relationship shows that these plants require a condensate/feedwater pH of at least 9.8 to achieve acceptably low iron levels at the condensate pump discharge. Alternatively, operating at the normal pH range associated with normal oxygenated treatment of between 8.0 to 9.0 would be absolutely disastrous for the ACC as has been illustrated [7,8]; high levels of corrosion will ensue at the tube entries and failures in the ACC can be expected. The high levels of iron will also completely over-

PowerPlant Chemistry 2009, 11(5)

267

PPChem
whelm the filtration capability of condensate polishers (where fitted) resulting in short resin runs due to differential pressure limits and significant deposition in boiler or HRSG tubes. Figure 4 also shows the solubility of magnetite extracted from Sturla [6] at the approximate temperature of the tube entries of an ACC (50 C (120 F)). The practical corrosion curve in the figure is elevated above this normally accepted magnetite solubility relationship by some considerable amount and shows for instance that at pH of 8.8 the ACC relationship is higher by 4050 g kg1 and at pH of 9.2 it is higher by about 30 g kg1. Only at pH of 9.8 and above do the curves become convergent. This suggests that the corrosion process at the lower pH range is faster or more aggressive than a normal classical two-phase FAC process [9] which may relate to the severe turbulence created by the tube entries. From the parallel field of nuclear plant twophase FAC, Table 1 illustrates the calculated effect of the different regimes on the liquid film pH under conditions approximating those of the condensate droplets in the ACC ducting.

Assessing and Controlling Corrosion in Air-Cooled Condensers

Figure 4: Iron levels measured at condensate pump discharge on units with ACC and only using ammonia for pH control. The iron solubility data which is extracted from Sturla [6] at 50 C is shown only as an example as actual operating temperatures at the ACC tube entries may be higher, as discussed in the text.

What is evident from the table is that when a plant is using conventional oxygenated treatment (OT) at pH 8.5 the delta pH, the pH at solution temperature neutral pH (last column), is only 0.26. It has been established by workers in the nuclear industry that a minimum delta pH of 1 unit (at solution temperature) is required to mitigate two-phase FAC. The table shows that to achieve this, the condensate/feedwater pH needs to be close to 9.8, which supports the Dooley/Aspden relationship shown in Figure 4.

are usually of the deep mixed-bed variety, which are designed for hydrogen form operation. This means that under these high pH conditions the operating cycles will be 2 to 3 days due to ammonium loading. The regeneration turnaround and effluent production make this mode of operation extremely difficult and costly. Ammonium form operation of mixed beds is an option [11] but presents significant challenges regarding polished water quality at these high pH values. At least one utility has approached this problem by equipping the ACC units with deep bed polishers using separate cation and anion vessels. This makes ammonium form operation quite simple and the required water quality is readily achieved. Capital costs are higher but the operaAmmonia concentration

CONDENSATE POLISHING
Where condensate polishing is fitted, these high levels of ammonia that are required to protect the ACC present a significant challenge. Where polishers are employed they

Treatment regime

Condensate

Condensate

Condensing steam Liquid phase Vapour phase g L1

8

Liquid phase

pH (25 C) OT High pH OT 8.5 9.8

pH* 1.5 2.8

g L 64

Molal 3.765 10
6

g L

Molal 1.804 10

Molal 3.747 10
6

pH (c) 6.18 7.05

pH** 0.26 1.13

0.31 7.02

63.7 5 012

5 018

2.952 104

4.129 107

2.948 104

Table 1: Calculations showing ammonia concentrations and delta pH for early condensate [10]. pH (25 C) = pH measured at the reference temperature of 25 C
pH* = difference between pH (25 C) and the neutral pH (25 C) pH** = difference between pH at temperature and neutral pH at temperature

268

PowerPlant Chemistry 2009, 11(5)

Assessing and Controlling Corrosion in Air-Cooled Condensers

PPChem
South Africa. The station (Figure 5) consists of 6 coal-fired units with a gross generating capacity of 665 MW each. The first unit at Matimba was commissioned in 1987 and the last unit went into commercial operation in 1991. The station was commissioned using an all-volatile treatment chemistry, AVT(R). The condensate pH ranged from 9.29.4 during the first few years of operation. A decision was made in the late 1990s to increase the steam/water cycle pH to 9.7 to help reduce the severe corrosion that was seen in the ACC ducts. Due to problems experienced in early 2000 with condensate polishing plant (CPP) resin cross contamination and early chloride leakages when the polisher resins were at the end of life, the steam cycle pH was controlled between 9.49.6 to help minimize the chloride leakage from the CPP [5]. The Matimba ACC exhaust ducts consist of rather complex arrangements of vertical and horizontal duct sections connected by a series of 90 bends, which in most cases are fitted with guide vanes to minimize the steam side pressure drop. Figures 6 and 7 illustrate the side and front views of the ACC duct arrangement to convey the exhaust steam from the LP turbine exhaust to the condensing elements. The two LP turbine exhausts are downward facing and the first vertical duct section (only one V1 is shown in Figure 6) is designed as a transition piece from a rectangular to a 5 m (16.4 ft) diameter round steam duct. The first guide vane bend connected to the transition piece diverts the steam in the horizontal direction (H1). After penetrating the turbine house wall another guide vane bend turns the steam direction in the vertical direction. Two vertical risers (only one V2 is shown in Figures 6 and 7) convey the steam upwards to an elevation of 30 m (98.4 ft) above the zero meter level, where the flow is distributed by two horizontal distribution pipes (only one H2 is shown in Figure 7) to 8 risers (4 risers per side) (V3) of 2.5 m (8.2 ft) diameter each,

tional costs are very attractive with cycles of 60 days or more being the norm. Powdered resin filtration systems also offer some benefits but the limited ion exchange capacity may deter some users, particularly if operation at high pH (ammonium form) is desired. The use of an amine injected into the crossover steam pipes (superheated steam) between the intermediate pressure (IP) and low pressure (LP) turbines is another avenue which could be considered. A relatively low concentration of an amine such as ethanolamine (ETA) or dimethylamine (DMA) would be required to achieve the required delta pH of > 1 unit (64 g kg1 ammonia for feedwater pH 8.5 and 100 g kg1 ETA). This would allow for reasonable and acceptable polishing cycles to be achieved with hydrogen form polishing. The injection of ammonia downstream of the polishers to achieve a feedwater pH of 8.5 would satisfy OT requirements. The removal of all the amine on the polishers would therefore retain the cation conductivity as a core measurement parameter. The authors are not aware of any organization using this technique and it would need to be evaluated to ensure that the correct mixing in superheated steam can be achieved and that the amine will not degrade in the short residence time at IP turbine outlet temperatures around 380 C (716 F). Also, although a number of combined cycle units with ACC are using amines and blends of amines including ETA, there is insufficient data at this time to superimpose another curve on the Dooley/Aspden relationship (Figure 4).

PRACTICAL CASE STUDY OF CORROSION IN AN ACC

Numerous internal inspections have been performed on the ACC ducts at Eskom's Matimba Power Station in

Figure 5: Aerial view of the Matimba Power Plant showing the ACC at the front of the power plant.

PowerPlant Chemistry 2009, 11(5)

269

PPChem

Assessing and Controlling Corrosion in Air-Cooled Condensers

Figure 6: Side view of the Matimba ACC ducts. Duct V1 takes the two-phase steam mixture from the steam turbine. V1, V2, V3, and H1 are considered to be the lower ducting of the ACC in the DHACI nomenclature. H3 and the ACC tube entries are considered to be the upper sections of the ACC.

which convey the steam to the inlet of the condensing elements (H3) at an elevation of 56 m (184 ft) above the local zero level. Depending on the ambient temperature, the design LP turbine exhaust pressure (absolute) varies from a minimum value of 11.5 kPa (1.66 psia) at minimum ambient temperature to a maximum value of 46.5 kPa (6.7 psia) in summer. Due to the corresponding variation in the specific volume of the exhaust steam, the average steam velocity in the ducts varies approximately between 35 m s1 and 120 m s1 (114 and 393 ft s1) at the maximum and minimum exhaust pressure respectively. At the weighted annual average exhaust pressure of 19.8 kPa (absolute) (2.87 psia), the average steam velocity is 76 m s1 (249 ft s1) with a moisture content of 5.4 %. Condensation of the exhaust steam occurs in bundles of finned heat exchanger tubes. The latter is characterized by rectangular fins fitted mechanically on an elliptical tube core as shown in Figure 8. A corrosion resistant bond between the fin root and tube is achieved by hot dipped galvanizing of the fin/tube. Heat exchanger bundles are constructed by arranging two tube rows in the air flow direction (A-frame). 48 axial flow fans per turbine unit of 9.1 m (29.8 ft) diameter each driven by a 270 kW electric motor force ambient air through the finned tubes as cooling medium. Internal inspections of the Matimba exhaust ducts and steam headers during outages revealed areas with corro-

Figure 7: Front view of the Matimba ACC ducts. H2, V2 and V3 are considered to be the lower ducting of the ACC in the DHACI nomenclature. H3 (above the A-frames) and the ACC tube entries are considered to be the upper sections of the ACC.

sion/FAC and a corresponding loss of material. The exact damage mechanism was not fully understood initially. At the time when the corrosion was first observed, a number of actions were identified to obtain more information for a

270

PowerPlant Chemistry 2009, 11(5)

Assessing and Controlling Corrosion in Air-Cooled Condensers

PPChem

Figure 9: Through-wall penetration of the tube wall immediately below the seal weld at the tube entry. Also note the loss in material of the seal weld. This is recorded as DHACI 5.

Figure 8: Matimba ACC heat exchanger tube. The tubes are 9.5 m (31.7 ft) long. The rectangular fin dimensions are 120 and 50 mm (4.7 and 1.96 in).

better understanding and to help mitigate the corrosion/ FAC mechanism. Among these options, cycle chemistry changes were made as discussed above as well as applying wear resistant protective coatings. Based on the various internal duct inspections conducted on the different Matimba units, it was evident that the liquid phase in the turbine exhaust steam had a significant effect on the corrosion pattern in the ducts. In areas were liquid accumulates significantly more corrosion was observed compared to areas exposed to dry saturated steam. As a result of the relatively high average steam velocity and the centrifugal forces exerted on the two-phase flow by the guided bends, the free water droplets in the exhaust steam accumulate and become separated from the steam [5]. The DHACI was used as a quantitative measure to define the internal corrosion status of the steam ducts and headers of the ACC. In the following examples the application of the Index is illustrated and used to quantify the benefit that the cycle chemistry changes had on the corrosion status. Of immediate concern during the initial Matimba inspections was the perforation of the relatively thin wall, 1.5 mm (0.059 in), of the elliptical fin tube at the tube entry. Any perforations would result in an air leak into the vacuum space with an adverse effect on condenser performance. Figure 9 shows a perforation in a tube entry surrounded by a large white corroded area. This white area is the same as the "bare metal areas" described in the introduction and is shown in more detail in Figure 1. Figure 10 shows patches of white corroded areas in a number of tube entries in a location of high liquid concentration caused by the two drain holes in the channel immediately above the tube entries. These white areas have

Figure 10: White bare metal areas at a number of tube entries which are immediately below a drain hole in a channel on the duct surface. This is recorded as DHACI 3.

been described in the introduction and are shown in Figure 1. The surrounding dark grey areas are the same as the black deposited areas also described in relation to Figure 1. Whilst the desired pH for operation at Matimba Power Station was 9.8 for the protection of the ACC, there was a period of operation when the pH had to be reduced to 9.4 because of problems related to the ability of the condensate polishing plant to operate in the ammonium form. This lasted for approximately two years and the resulting damage to the ACC is evident in Figures 9 and 10. Tube entries, rated at DHACI 1, with no visible white bare metal areas (FAC) are shown in Figure 11, which illustrates the improvement when a plant is operated at a pH of 9.8. Such significant improvements can occur in approximately two years. Corrosion of the exhaust (lower) ducts should generally not compromise the overall integrity of the ducts or vacuum

PowerPlant Chemistry 2009, 11(5)

271

PPChem

Assessing and Controlling Corrosion in Air-Cooled Condensers

Figure 11: Well-protected tube entries with no white bare metal areas and a well passivated red duct surface in an ACC operating with a pH between 9.6 and 10. This is recorded as DHACI 1.

Figure 12: Localized area of significant wall loss in the lower exhaust duct. DHACI C.

boundary of the ACC but will contribute to the iron transport in the condensate. However, in localized areas significant material loss can be experienced, generally in areas of highly disturbed flow or high liquid phase concentration. An example of such an area in the lower ducting, immediately downstream of the trailing edge of a guide vane in a bend, is shown in Figure 12. The figure shows a corroded area where material loss of about 3 mm (0.117 in) is experienced after a total exposure of 146 000 operating hours. The corrosion in Figure 12 is an exception, and generally minor wall loss is associated with the corroded areas. Such an example is given in Figure 13, which shows some 'tiger striping' appearance with darker grey/black areas demonstrating two-phase flow-accelerated damage. A detailed internal ACC duct inspection was performed on Matimba Unit 5 in September 2008. A definite improvement in duct appearance was observed in the steam flow direction, i.e. fewer white bare metal areas were observed with increasing distance from the LP turbine exhaust flange. In Table 2, a qualitative comparison is shown

Figure 13: Minor bare metal (white) areas on black background as observed in the duct section H1 (see Figure 6). DHACI B.

between the appearances of the different duct sections as identified in Figures 6 and 7. The improvement (lower corrosion/FAC areas) was due to the return to pH 9.8 and the associated "self repair" which took place.

Duct section (Figures 6 and 7) V1 H1 V2 H2 V3 H3

Table 2: Duct appearance of Unit 5.

% red area 0 40 60 70 75 85

% black area 95 50 40 30 25 15

% bare metal (white) 5 10* <1 <1 <1 <1

* The majority of the bare metal areas are caused by the LP bypass dump diffusers which are located inside duct section H1.

272

PowerPlant Chemistry 2009, 11(5)

Assessing and Controlling Corrosion in Air-Cooled Condensers

PPChem

Figure 14: Corrosion pattern in branch H2-V3 (Figure 7). DHACI C.

The effect of the moisture content in the exhaust steam is well illustrated in Figure 14, which shows the branch H2V3. On the downstream side of the branch a large bare metal area is visible with patches of flash rust, while the rest of the branch circumference is well protected with a stable oxide layer. Due to centrifugal action the free moisture in the exhaust steam impinges on the downstream side of the branch, resulting in the corrosion pattern shown in the figure.

REFERENCES
[1] [2] Howell, A. G., Pritekel, R. J., Dooley, R. B., PowerPlant Chemistry 2007, 9(5), 270. Dooley, R. B., Shields, K. J., Shulder, S. J, Aspden, D. A., DuPreez, F., Howell, A. G., Mathews, J., EPRI's Guideline on ACC, To be presented at the EPRI International Conference on Cycle Chemistry, Boston, MA, U.S.A., June 2009. Dooley, R. B., Aspden, D. A., Proc., Air-Cooled Condenser Workshop, 2008 (Sunshine Coast, Queensland, Australia). ACC Interest Group of the Power Plant and Environmental Chemistry Research (PPEC) Group of ASME, Interim Guidelines for Off-Line Inspection of AirCooled Condensers, To be published April 2009. Phala, S., Aspden, D., DuPreez, F., Goldschagg, H., Northcott, K., Proc., API Power Chemistry Conference, 2008 (Sunshine Coast, Queensland, Australia). Sturla, P., Proc., Fifth National Feedwater Conference, 1973 (Prague, Czech Rebublic). Prust, A., PowerPlant Chemistry 2008, 10(6), 393. Richardson, I., PowerPlant Chemistry 2008, 10(7), 393. Dooley, R. B., PowerPlant Chemistry 2008, 10(2).

[3]

CONCLUDING REMARKS
By examination of only two tubes worldwide the authors have started to understand the corrosion processes in ACCs. Based on this it has been possible to develop two tools for operators to use in assessing the condition of ACCs. The first, a corrosion index, DHACI, allows the operators to assess corrosion/FAC on a common basis and to monitor changes in corrosion as a result of chemistry or coating changes. The second, the Dooley/Aspden relationship between corrosion/FAC and pH in the cycle, clearly shows the importance of pH adjustment in optimizing the iron levels in the cycle.

[4]

[5]

[6] [7] [8] [9]

[10] Galt, K. J., Private communication to D. Aspden, 2009. [11] Sadler, M. A., PowerPlant Chemistry 2001, 3(10), 559.

PowerPlant Chemistry 2009, 11(5)

273

PPChem
THE AUTHORS
R. Barry Dooley (B.S. with first class honors, Metallurgy, Ph.D., Metallurgy, both from University of Liverpool, UK, D.Sc., Moscow Power Institute, Moscow, Russia) is a Senior Associate with Structural Integrity Associates. Before joining Structural Integrity Associates he was a Technical Executive, Materials and Chemistry, at the Electric Power Research Institute (EPRI) in Charlotte, North Carolina, where he managed the Cycle Chemistry, HRSG, and Materials Programs, as well as the Boiler Tube Failure Reduction/Cycle Chemistry Improvement and FAC Programs. Before joining EPRI in 1984, Barry Dooley spent nine years with Ontario Hydro in Toronto, Canada. From 1972 to 1975, he was a research officer in the Materials Division of the Central Electricity Research Laboratories (CERL) of the former Central Electricity Generating Board (CEGB) at Leatherhead, Surrey, England. Barry Dooley is the author or coauthor of over 260 papers and the editor of 14 international conference proceedings, primarily in the areas of metallurgy, power generation, boiler tube failures, HRSG tube failures, cycle chemistry, and life extension and assessment of fossil plants. He is the coauthor of a three-volume book on boiler tube failures, a two-volume book on steam turbine damage mechanisms, and a book on flow-accelerated corrosion in power plants. Barry Dooley is the executive secretary of the International Association for the Properties of Water and Steam (IAPWS). He was awarded an Honorary D.Sc. from the Moscow Power Institute in 2002. Denis Aspden (National Diploma, Chemistry, M.S., Engineering Management, University of Pretoria, South Africa) has more than 40 years of experience in power plant chemistry and is now an independent consultant.

Assessing and Controlling Corrosion in Air-Cooled Condensers

Andrew G. Howell (M.S., Chemistry, University of Missouri, Columbia, MO, U.S.A.) has worked for Xcel Energy and its predecessors since 1982. He has worked in the analytical laboratory, including laboratory and field investigation of power system component failures due to corrosion, and currently provides support for Xcel Energy's power stations as a 'hands-on' consultant. Abraham Francois du Preez (Ph.D., Mechanical Engineering, University of Stellenbosch, South Africa) joined Eskom, the electric power utility in South Africa, in 1994, where he has gained maintenance and operational experience on dry cooling systems, including air-cooled condensers. As the responsible engineer he provides design, operational and maintenance technical advisory and support service to power stations for condensers and cooling systems. His recent work includes the specification for the air-cooled condensers for the 6x800 MW Medupi power station, which is currently under construction.

CONTACT
Barry Dooley PhD., DSc. Senior Associate Structural Integrity Associates, Inc. 2616 Chelsea Drive Charlotte, NC 28209 USA E-mail: bdooley@structint.com

PowerPlant Chemistry Important Information Available on the Web

Subscription Rates 2009 http://www.ppchem.net/subscription/subscription.php Media Information 2009 http://www.ppchem.net/media-info/mi.php Ordering of Individual Articles http://www.ppchem.net/subscription/bi.php Ordering of Back Issues http://www.ppchem.net/subscription/bi.php Request for a Free Copy of the PowerPlant Chemistry Journal Send e-mail to freecopy@ppchem.net with your name, precise mailing address, and e-mail address and you will receive a complimentary copy of the latest issue of PowerPlant Chemistry.

274

PowerPlant Chemistry 2009, 11(5)

ANNEXURE 5: PRESSURE DROP IN EJECTOR AIR PIPING

PRESSUREDROPCALCULATIONFOREJECTORPIPING Specificvolumeofsteam Densityofairatcondenserpressure&temp Specificvolumeofairatcondenserpressure Suckedsteamflow Expected airflow p Air+Steamflow' Specificvolumeofmixture Mixtureflowrate Steamairflowrateinonetubeofabundle Steamairflowrateinonetubeofabundle Tubediameter Tubethickness Tubeflowarea Velocityinthetube Tubelength Lossnode12 Lossnode23 Lossnode34 Lossnode45 Lossnode56 Lossnode67 Lossnode78 Lossnode89 Lossnode910 Lossnode1011 Lossnode1112 Lossnode1213 Lossnode1314 Lossnode1415 Lossnode1516 Lossnode1617 Lossnode1718 Lossnode1819 Lossnode1920 Lossnode2021 Lossnode2122 T t lpressured Total dropi inejector j t piping i i m3/kg kg/m3 m3/kg kg/h kg/h g/ kg/h m3/kg m3/s m3/s kg/h mm mm m2 m/s m mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC mmWC WC 6.913 0.2273 4.399 20.4 9.2 29.6 6.132 0.050 0.0007003 0.4111111 32 2 0.0006158 1.1372285 10.5 0.14 0.001 0 004 0.004 0.005 0.015 0.013 0.013 0.026 0.058 0.043 0.091 0.064 0.131 0.089 0.178 0.118 0.233 0.152 0.525 2.818 21.809 26 26.526 526

ANNEXURE 6 DESIGN COMPARISON OF ACC WITH FEW OTHER PLANTS

TABLE2:COMPARISONOFACCDESIGNDATA&HEATINGSURFACE DESIGNCOMPARISONBETWEENBIRLAWHITE7.5MW&JAIBALAJI35MW Birlawhite Jaibalaji Jaibalaji Comment reqd Design 7.5MW 35MW 35MW Steamflow kg/h 31000 115000 115000 Condenserpressure kg/cm2a 0 18 0.18 0 18 0.18 0 22 0.22 Condensatetemperature degC 57.4 57.4 62.4 Designambienttemperature degC 40 40 36 Heatduty Finnedheatingsurfacearea Fandiameter Fanrpm Nooffans Fanairflow kcal/h m2 ft rpm no kg/h 16649100 76622.3 28 118 3 5463370 77695800 357571 58346000 134629 HTAislessby62.35% 32.8 114 6 14600000 flowislessby42.74%

25495727

DESIGNCOMPARISIONBETWEENTHEJAIBALAJI25MW&JAIBALAJI35MWUNIT Jaibalaji Jaibalaji Jaibalaji Comment Design reqd Design 25MW 35Mw 35MW Steamflow kg/h 77010 115000 115000 Condenserpressure kg/cm2a 0.22 0.22 0.22 Condensatetemperature degC 62 62 62.4 Designambienttemperature degC 40 40 36 Heatduty Finnedheatingsurfacearea Fandiameter Fanrpm Nooffans Fanairflow kcal/h m2 ft rpm no kg/h 39361120 113473 32.0 114 6 11222566 58346000 158862 32 114 6 15711592 58346000 134629 HTAislessby15.25% 32.8 114 6 14600000 flowislessby7.07%

DESIGNCOMPARISIONBETWEENSSIL20MW&JAIBALAJI35MWUNIT SSIL Jaibalaji j Jaibalaji j Comment Design reqd Design 20MW 35Mw 35MW Steamflow kg/h 115000 115000 Condenserpressure kg/cm2a 0.2 0.22 0.22 Condensatetemperature degC 62 62.4 Designambienttemperature degC 42 40 36 Heatduty Finnedheatingsurfacearea Fandiameter Fanrpm Nooffans F air Fan i fl flow kcal/h m2 ft rpm no k kg/h /h 39738000 151446 30.0 0 6 13280865 58346000 265031 32 114 6 18593211 58346000 134629 HTAislessby49.2% 32.8 114 6 14600000 flow fl i isl lessb by21 21.48% 48%

ANNEXURE 7 CONDENSATE CHEMISTRY CHANGE

Table1:Condensatechemistryreport 17/3/2012 Conduct UnitNo Description s/cm Unit#1 8.9 Condensatewater Unit#2 9.2 Unit#3 9.1 16/3/2012 Conduct UnitNo Description s/cm Unit#3 8.89 15/3/2012 Conduct UnitNo Description s/cm Unit#1 10.27 Condensatewater Unit#2 9.87 Unit#3 10.39

PH 9.5 9.4 9.4 PH 9.4 PH 9.6 9.5 9.6

Silica ppm 0.014 0.015 0.019 Silica ppm 0.018 Silica ppm 0.016 0.015 0.016

Iron ppb 30 30 80 Iron ppb 180 Iron ppb 30 30 320

Turbidity NTU 0.26 0.29 0.88 Turbidity NTU 2.11 Turbidity NTU 0.21 0.24 4.1

AftertheriseofpHatatACCcondensate,theironcontentrisedatcondensate.Chokingof strainerswasexperiencedinitially.LaterittaperedoffindicatingthecleaningoftheACC.

ANNEXURE 8 EFFECT OF VACUUM ON ACC CONDENSATE CHEMISTRY CHANGE