CASE STUDIES IN BOILER VIBRATIONS AND BFP CAVITATION By K.K.Parthiban, B.Tech ( IIT-M), M.E Thermal Engineering- RECT Trichy In this paper two case studies are presented, which are relevant to boiler operating and design engineers. One is a vibration problems experienced in CFBC boilers and other is about a repeated BFP failure in a power plant. VIBRATION OF SECOND PASS AT AIRPREHEATER There were three cases which had come to us in this subject. Two cases were about the vibration experienced in the APH casing of CFBC boilers. One case is the vibration of the steam cooled second pass housing the primary superheater & economiser. Vibration of boiler is not a rare phenomenon and many readers would have come across it. Vibrations of tube bundles / casing / panels are due to a scientific phenomenon known as flow induced vibration. Some of us may doubt whether fluid flow can cause vibration. Here is a famous example. The two photos below show the collapse of a bridge named Tacoma Narrows Bridge in USA in the year 1940. Within three months Photo 1 & 2 : Collapse of Tacoma narrows bridge - 1940 of construction, the bridge collapsed due to aero-elastic flutter due to wind at a speed of 68 kmph. This phenomenon is due to flow of wind around the bridge. The subject is unique and such situations occur in boilers too. Vibration and noise problems may be encountered as the air / gas / water / steam flow over tube bundles in a power plant. Fig 1: Von korman street- is a repeating pattern of swirling vortices caused by the unsteady separation of flow over bluff bodies Vortices are formed and shed beyond the wake of the tubes, resulting in harmonically varying forces on the tubes perpendicular to the flow direction. It is a self-excited vibration. If the frequency of vibration of the Von-Karman vortices, as they are called, coincide with the natural frequency of vibration of the tube bank, resonance occurs which leads to tube vibration. Another phenomenon that is relevant for discussion here is acoustic vibration, leading to noise. The duct or the waterwall / steam cooled panels vibrate when the acoustic frequency coincides with the natural frequency of tubes. The acoustic oscillation is normal to both
The casing vibration can lead to supporting structure and hand rails. It becomes Figure 2: Picture of standing wave in a duct / enclosure, due to air / gas necessary to column. identify the source of acoustics. In boiler and duct systems the source of acoustics can be from fans, improper ducts, and abrupt transitions, to list a few. Case study 1: vibration of APH casing in a CFBC boiler This is a 76 TPH CFBC boiler designed to fire 45% ash Indian coal. The boiler general arrangement can be seen in figure 3 below. The 2 nd pass APH casing was seen to vibrate heavily once the load crossed 65 TPH. The APH casing started cracking on the heavy vibration. The boiler was inspected during operation. Plant engineers had stiffened the duct casing, thinking that the duct stiffeners are inadequate. Even after the additional stiffening of the casing, the casing was cracking. See photo 3 & 4 below. On shut down inspection, it became clear that the cause for vibration was the acoustics created by the air leak from air side to gas side. Heavy air leak could be seen from the lifting holes of APH blocks. See photo 5 & 6. Figure 3: CFBC boiler Case study 1 Photo 3 & 4: APH casing cracks.
Photo 5 & 6: APH casing cracks can be seen due to acoustic vibration. The source of vibration was the high pressure air leak from air side to casing side through the lifting lugs. Photo 7 at top left shows the excessive gap between the casing and the end tube. Photos 8 shows the incomplete seal welding around the APH blocks. Figures 4 & 5 below show the erection welds. In shut down, more leakages were seen between the APH base and the structural supports. The vibration possibilities due to vortex shedding and standing waves were checked, by calculations. The procedure for checking the vortex induced vibration and the acoustic vibration is well written by Mr.V.Ganapathy, expert in heat transfer. There was no possibility of APH tube vibration. Readers could refer his article in internet. The spread sheet calculations is presented in this article for reader s interest. Another observation was that the tube to casing gap was much higher than the tube to tube clear gap. This can lead preferential gas flow along the casing and cause casing vibration. Perforated plates were installed at these locations, so that ash would not bridge and the gas flow would be retarded. The vibration stopped once all the actions were taken as per the recommendations.
Case study 2: vibration of second pass in a 270 TPH CFBC boiler The case is a CFBC boiler of 270 TPH with Indonesian coal as the fuel. See figure 6 showing the boiler layout. As the steam generation was increased around 80% MCR, the second pass began to vibrate. The boiler maker suggested dividing the gas path thinking that there is superheater coil vibration due to acoustics. When the visit was made, we found there could be other causes which needed to be attended first before implementing the gas side partitioning proposed by the boiler maker. Our observations included the following. Air was seen gushing in the soot blower stuffing box making a whistle noise. This was later attended by the plant engineers. The oxygen levels were also high. The CFBC operation was not stabilized with adequate dust in the upper furnace. There would be combustion pulsations which lead to vibration. Vibration was experienced at another plant, when O 2 level was at 12% in flue gas. We did notice that there was air ingress in the fabric joint flange, wherein sealing rope was used. The flange was not rigid too. We advised the flange be seal welded. Internal bracings were added at the flange frame. The boiler operation was corrected by proper loop seal parameters and adequate upper furnace dust. Vibration was gone after the modifications.
Figure 6: General arrangement of the 270 TPH CFBC boiler which had developed vibration in the second pass.
A clip from literature on vortex induced vibration and acoustic vibration modes. Gas flows / air flows in a heat exchanger bank can generate a loud noise called acoustic resonance. The noise can occur when the frequency of flow periodicity generated in the array resonates with one of the natural transverse acoustic standing waves of the duct. The relevant standing waves are oriented in a direction normal to both the tube axis and the flow direction. When resonance occurs an intense pure tone noise is usually produces that can cause damage to the HX internals by fatigue and will be harmful to plant personnel. The loudness of the noise can reach up to 175 db, depending upon the pressure drop of the tube array and the damping capacity of the tube bundle and the duct. The most common remedy is to install one or more anti-resonant baffles within the tube bundle to distort and suppress the resonance acoustic waves. Photo 9 & 10 show the position of the large fabric joint just below the economiser, which deflected inside and created air ingress. The internal tie pipes were added to prevent the caving of the flange. The flanges were seal welded. When the unit was put back on line, there was vibration. Acoustic disturbances can be in many forms. Air ingress also seem to create more problem in CFBC as the negative draft is high in the second pass as compared to other stoker / AFBC or even PF boilers. In fact the doors in the second pass had to be sealed well. Incidentally this boiler is provided with perforated baffles to prevent the preferential flow of the gases close to the casing / steam cooled wall.
CASE STUDY 3: FEED PUMP CAVITATION In this case, the BFP had been subject to cavitation damage during turbine trip / load throw occasions. The deaerator is given steam from the turbine extraction in order to maintain the feed water temperature at 160 deg C. This corresponds to operating pressure of 5.2 kg/cm2 g. When a load throw occurred in the turbine, the deaerator pressure comes down. During this period the NPSH is upset. The pump supplier had informed the plant engineers that the suction piping pressure drop is too high and hence the cavitation has been occurring. Plant engineers had made arrangements are to inject cold condensate from the CEP. But the basic problem remained. The various data for every second was available with plant engineers as this problem had been a seven year old problem for the plant. Why should there be a cavitation when the pressure drops at deaerator? During a transient condition the volume of water trapped in the suction piping is at a higher temperature as compared to the deaerator tank water temperature. The temperature of water at deaerator storage tank would be at the saturation temperature corresponding to current pressure above water level. In the event of turbine trip, the deaerator tank pressure will drop. The hot water at storage tank will produce flash steam in case of pressure drop and adjust its temperature. But the temperature sensors would indicate slow change. It is the trapped water in suction piping that can cause cavitation. It can be about 60 seconds for the water to be replaced, with the feed water flow rate present after the load reduction. It is within this period that the cavitation damage could occur. Any vibration in BFP should be in this period only. Once the water inside the suction piping is replaced, there should not be any continuity of cavitation or vibration. Our diagnosis points First thing was the pressure drop calculation for the suction piping. The pressure drop in suction piping with boiler MCR flow was calculated to be only 0.23 mwc. The truth was that the pump maker did not make a calculation or measured the pressure drop with calibrated pressure gauge. The pump suction strainer DP was seen to be 750 mmwc normally. This is considered to be high in a stabilized plant operation. The DP transmitter was calibrated to 5000 mmwc. The set pressure for alarm as per BFP supplier document was 0.35 kg/cm2. Trip was to be set at 0.5 kg/cm2. Plant engineers had set the trip at 1200 mmwc. Perhaps this was done after the failures were experienced in the BFP. The strainer DP was seen to increase at every load throw in the past. Whenever there was reduction in boiler feed water flow, there was an increase in DP. See photo 11 to 14. This indicates there is ARC passing. The DP rise had been instant, that this was not due to flashing. The ARC may be passing more flow than the stipulated minimum. ARC is designed to relieve 33 m3/h and act at a pressure of 142.3 kg/cm2 as the FCV starts closing. In the past instances the pressure had never touched the 142.3 kg/cm2 value, but there is sufficient rise in the strainer DP. This indicates there is passing of ARC at a lower pressure itself. More flow will call for higher NPSH. The available NPSH declines as the turbine load reduces, due to reduction of deaerator pressure. ARC was seen to be passing at the time of visit. The ARC outlet piping temperature at turbine hall was seen to be 138 deg C as measured by the IR camera. This was seen when the feed water
flow was 80% of MCR and the BFP discharge pressure was 125 kg/cm2. We suggested adding a pressure gauge near the TG floor in the ARC discharge line to know the passing of ARC. Power consumption is also an issue with passing ARC. We looked in to the details of strainers provided by the pump supplier as the normal DP itself was high. Strainer flow area was seen to be 1.5 times the cross sectional area of piping. We compared the area with strainer design by a well known pump supplier. Their BFPs are provided with 3 times the flow area of piping. We went in to further details of mesh selection. The mesh opening was 0.315 mm square. The usual strainer openings by another pump supplier used to be 0.5 mm sq. Small mesh openings can also be reason for cavitation. We looked in to the suction piping arrangement taken by plant engineers in the previous occasions. The feed pumps suction tapping at deaerator storage tank was found located near the ring stiffener. This could affect the flow. See photo 16 & 17. We advised to trim the part of stiffener which could obstruct the flow. We recommended that the wire mesh area at inlet shall also be equal to thrice the flow area of the piping. It was advised that the wire mesh openings should be made bigger as it was required only to trap the spray nozzle components. We made this suggestion by comparing with another deaerator in the same plant. Review of the transient curves given by the plant engineers With the latest DCS control system in place, nice transient curves were taken out by the plant engineers in the past. These curves are presented in this paper for readers interest. Such data could help the plant engineers to improve the analytical skills. The transient curve on TG load, deaerator pressure, BFP suction pressure, BFP vibration on 31/8/2013 5.43 AM was reviewed. The curve is attached in photo 11. The load throw (from 25 MW to 15 MW) occurred at about 5.46 AM. Both deaerator pressure and BFP suction pressure had come down and recovered after the pegging steam supply was commenced. In 30 seconds the difference between the suction pressure and deaerator pressure was re-established. The deaerator pressure change had been only 0.7 kg/cm2. But the vibration level had gone up and remained for a longer duration. It seemed to follow the frequency / rpm rise in the drive. The transient curve on TG load, deaerator pressure, BFP suction pressure, BFP vibration on 30/8/2013 16.50 PM was reviewed. The curve is attached in photo 12. The load throw (from 22 MW to 11 MW) occurred at about 16.56 PM. Both deaerator pressure and BFP suction pressure had come down and recovered after the pegging steam supply was commenced. In 90 seconds the difference between the suction pressure and deaerator pressure was re-established. The deaerator pressure change had been only 5.7-5.2 kg/cm2. But the vibration level had gone up and remained for a longer duration. It followed the frequency / rpm rise in the drive. In this case the suction pressure came below the deaerator pressure and lasted for 1.5 minutes. This is the case of cavitation. In both the above cases, there is a small change in deaerator pressure. The change is about 3 to 7 mwc in head. The NPSH available excluding the strainer DP is 11.47 mwc. If the pressure decay is too fast then there will be flashing of steam in suction piping. If we assume 1 kg/cm2 is the decay, the flash steam percentage is 1.4%.
Photo 11: Both BFP suction pressure and deaerator pressure declined during the load throw. The BFP suction pressure to deaerator pressure difference is maintained.
Photo 12: The BFP suction pressure had gone down below that of deaerator pressure. This can happen due to high DP across strainer / ARC acting or due to momentary frequency raise.
Photo 13: The DP had gone to 1325 mmwc instantly, even when the water flow is 90 TPH. There is a load throw of 24 to 20 MW. In one minute the deaerator pressure had come down by 0.7 kg/cm2.
Photo 14: The BFP strainer DP went up when the steam flow was dropped. Photo 15: The ARC back pressure considered is 12 bar. This may be checked by installing a pressure gauge downstream of ARC. ARC should operate only at 158.2 kg/cm2 as per the data sheet. The minimum flow for ARC is 33 m3/h.
Photo 16 & 17: The photo on the left shows the suction pipe location in the deaerator. The stiffener should be trimmed. The coarse mesh flow area shall be 3 times that of pipe c/s area. Photo on the right shows the coarse strainer in another deaerator of the same plant. Photo 18 & 19: The photos above show passing of ARC valve. The heat is seen in the piping. Photo 20 & 21: The photo on the left shows the strainer DP in another plant and the photo on the right shows the DP in the affected pumps. It implied the strainer flow area is less as compared to other installation.
Conclusions The ARC passing, small mesh in strainer, less flow area in strainer could reduce the NPSH available during transients. We recommended two important points for immediate implementations. Pressure decay can be reduced in the make up from CEP is stopped at the deaerator for one minute, the flashing of steam in the water entrapped in suction piping can be reduced. The deaerator vent pipes at this plant were 2 nos of 40 nb lines. We recommended adding 1 x 15 nb parallel vents to reduce the loss of steam and to reduce the rate of decay. M/S Venus energy audit system Trouble shooting of boiler failures and operational issues. Company carries out design audit, construction audit, shut down audit and operational audit. M/S Sri Devi engineering consultancy and agency engaged in non pressure parts spares supply for FBC boilers. K.K.Parthiban M/S Sri Devi boiler equipment and spares engaged in supply of pressure part spares for all type of boilers