Reducing Cavitation Potential At The Condensate Extraction .

International Journal of Mechanical & Mechatronics Engineering IJMME-IJENS Vol:19 No:06flow boundary surfaces however hard or tough the materialcan be [8].CEP is one of the multistage centrifugal pumps, the fluiddischarge directly into a volute casing which graduallyincreases in the area (bowl diffuser casing). The volutecasing which has a diffuser form is designed to decrease thefluid velocity when fluid exits the impeller, and thisreduction in kinetic energy is transformed to increase inpressure. The volute casing, with its increasing area in thedirection of flow, is used to produce an essentially uniformvelocity distribution as the fluid moves around the casinginto the outlet. Multi-stage pumps that are operated in series,i.e the liquid is coming out of the first stage flow to thesecond stage eye, the fluid exits from the second stage flowthe eye to the third stage, and so on. The same flow ratethrough each stage, but each stage provides an additionalpressure increase. Thus, a very high discharge pressure, orhead can be increased by a multistage pump [9].The results of the CEP inspection during the overhaulon 29 June - 5 August 2018 have indicated that there hasbeen found a form of pitting corrosion damage to the CEP2B bowl diffuser in the second stage as shown in Figure 1.One possibility of this damage is caused by erosion due tocavitation. In Figure 1, it appears that pitting corrosionpenetrates the outside of the diffuser bowl, where the outerpart is the line suction from CEP, so that local recirculationoccurs from the first stage discharge to the suction side ofthe CEP. When local circulation occurs, the CEPperformance decreases due to the fluid do not flow to thedischarge side, but returns to the suction side. Damagedbowl diffuser position and CEP suction flow path can beseen in Figure 1.At present the decline in CEP performance has becomean important issue at the Labuan Power Plant Indonesia,where the decrease in CEP performance is considered due tocavitation occurred in the CEP system. With the decliningperformance conditions, CEP cannot meet the needs of thepower plant loads and threatens the reliability of the LabuanPower Plant.Concerning the important role of CEP in a power plant,several studies on cavitation at the pump have been carriedout by many researchers, as follows:Authors [10] examined the effect of blockage on theinlet of a centrifugal pump against cavitation. Addingblockage to the suction side means increasing the headlosson the suction side. Cavitation events are detected by VectorSupport Machine (VSM) which detects high vibrationalvalues if cavitation occurs. The results showed that withgreater levels of blockage, there was an increase incavitation. And with higher flow rates, the formation ofcavitation increases larger too. Authors [11] conducted anexperimental test of the effect of temperature on cavitationon the centrifugal pump blade. Cavitation at the pump isindicated by the value of the Thoma cavitation number (σ).The authors have stated that the smaller the cavitationnumber, the easier cavitation occurs. The experimentalresults show that the cavitation number will be lower if thetemperature of the liquid rises. The authors have alsoconfirmed that cavitation numbers also affect the headcoefficient, which is a dimensionless number that states thepump's ability to convert mechanical energy into the pumphead. The lower the cavitation number will result in thelower the head coefficient. Authors [12] conducted an123experimental test and numerical simulation at a condensatepump. The research domain is divided into three parts: thesuction pipe zone of the pump impeller, the impeller zone,and the extension of the downstream impeller zone. Therotating coordinate system is used to adjust the impellerzone with measured rotation speed, while the other parts arein a stationary coordinate system. The simulation results ofcavitation flow adequately illustrate the development ofcavitation at the pump, and predictions of decreasedperformance due to cavitation. With a higher flow rate, headdrops are steeper than lower flow rates. At each flow rate,the occurrence of cavitation on the surface of the suctionblade and cavitation develops along the surface of theimpeller. Subsequent research requires studies of therelationship between the length of the relative cavity and thedecrease in performance. Due to cavitation is damage to thefluid flow area, indicated by increasing surface roughness.Authors [13] have studied the effects of roughness effect oncentrifugal pump performance. An increase in surfaceroughness will reduce the hydraulic efficiency of theimpeller, because roughness increases the flow resistance inturbulent flow whereas in laminar flow roughness has noeffect on the resistance due to no exchange of momentumacross the flow. Author [14] investigated the comparativenumerical study of open channel junction flows and lossesof energy. The numerical experiment set junction angle 30o,45o, 60o and 90o. The result of numerical simulation,secondary pattern and energy losses is dominated with abigger junction angle. The authors [15] focused onoptimizing the coefficient of local resistance in concaveshape tee junction. How to improve the tee junction byadded the concave contour. In the straight duct with a largerflow, the tee junction with concave contour can reduce headlosses between 20.45% and 248.2%.From several studies above, there is an important pointthat has not been widely discussed, namely how to reducethe potential of cavitation by increasing the pressure on thesuction side of the pump. This is the reason for this research,because cavitation can be prevented by increasing thepressure of the suction side of the pump or decreasing fluidtemperature. In this study, the method used to preventcavitation in CEP is by reducing the head losses on thesuction side of the pump, so that the pressure in this sectionbecomes above the saturated pressure of the flowing liquidin the working temperature. As is known, there are severalpossibilities to reduce head loss in the piping system, suchas; reduce the major losses, i.e. reducing the pipe length,enlarging the diameter of pipe and also smoothing the innersurface of the pipe, and/or reducing minor losses, i.e.avoiding connections with changes in pipe diameter,improving the connection system or junction so that itapproaches the rules of flow behavior, etc.Existing piping conditions can be seen in Figure 2. Itcan be seen that there is a connecting pipe with enlargementof the pipe diameter from 480 mm to 630 mm, so there is adiffuser effect that adds to the head losses. In Figure 2, thered arrows indicate the direction of the condensate waterflow, and the isometric of the existing suction pipe ismodeled in Figure 3.Based on the existing CEP piping system describedabove, there are potential improvements that can be made tothe suction pipe installation and proposed in this study, are:191606-8484-IJMME-IJENS December 2019 IJENSIJENS

International Journal of Mechanical & Mechatronics Engineering IJMME-IJENS Vol:19 No:06124Inlet 1Inlet 2Outlet 1Outlet 2Fig. 3. Isometric suction pipe installation of CEP Labuan Power Plant.a) Uniformity the diameter of the pipe on the suction side,becoming larger which is 680 mm,oob) Minimizing the tee junction angle, i.e from 90 to 60oand/or 45 .This study will propose three alternative suction pipemodels to reduce the potential cavitation in CEP. Inlet-2 inModel-1 (Figure 5), the pipe diameter same with the existingdiameter, 480 mm and inlet-1 is 630 mm. To remove thediffuser effect at the existing model, the uniformizeddiameter at the inlet-1 and the junction angle are 45o.Model-2 (Figure 6) was proposed with all uniform diametersoin 630 mm and 45 junction angles. And the last, Model-3(Figure 6) is proposed with all uniform diameters in 630 mmand 60o junction angles. The bigger diameter pipe andsmaller angle junction, predicted improve streamline andreduce secondary flow which will reduce head losses and itwill reduce potential cavitation in CEP. From the threemodels that were proposed, a numerical simulation wasperformed and which model was selected gave the smallesthead loss.II. RESEARCH METHODOLOGYIn this numerical simulation, Computational FluidDynamic software, Ansys Fluent version 19.2, is used whereactual data is needed as a reference for modeling andsimulation of the system to be examined. Experimental datais taken when Labuan power plant in operation. The actualdata used are two types, technical data and operating data.Technical data is used to do modeling and describe thegeometry in the field to be entered into the software. While,the operating data is used for modeling validation made.Modeling is declared valid if the difference between theactual data in the field and the model from the simulationresults has an error value of less than 5%. This means thatthe model system created using the software is the same asthe actual conditions in the field.A. SimulationSimulation methods have been widely used to predict thevelocity, pressure and temperature of fluid flow in severalcases. The working fluid used in CEP is condensate waterwhich is the steam condensation of the Power plant cyclewhich is accommodated in the reservoir tank as shown inFigure 4. Field operations data are taken when the pump isoperating and the unit is loaded. Operating data is obtainedfrom the parameters in the Distribute Control System (DCS)in the Central Control Room (CCR).In this study, the suction pipe CEP will be modeled in3D using commercial meshing software, Gambit version2.4.6. This software is used to create geometry, meshing,and to define domain modeling. Meanwhile, the pressuredrop from the inlet to the outlet is predicted using AnsysFluent 19.2 software. The operating data taken is as follows,(where the result can be seen in Table 1):a) Electric Current of CEP (in Ampere)b) Discharge Pressure (Pa)c) Mass Flow rate (kg/sec)d) Reservoir tank pressure (Pa)e) Reservoir tank level (m)f) Suction Pressure (Pa)og) Fluid temperature ( C)The geometry of suction pipeline CEP is shown inFigure 5. Where, the geometry has been modeled close tothe real condition, to get the same or close comparisonresults between numerical and experimental. Many researchstudies prove that non-matching geometry is the main causeof differences between experimental and numerical results.In this study, the geometry and modeling existing based onmanufacturing data and drawing in Labuan Power Plant.After making the geometry and the model is complete, thesimulation using Fluent-software continues to run on theexisting model by operating the data and parameters thathave been taken in Table 1.In addition to geometry, the specification of suitableboundary conditions is also important for the accuracy ofany numerical analysis. In this study, the inlet-1 and 2boundaries of the flow domain have been specified as massflow rate, while the outlet-1 boundary has been specified asa pressure outlet, which can be seen in Table 3. The massflow rate at the inlet boundary has been varied by circulatingvalve position. In this study, the total mass flow rate wasused in three conditions; (i) 180.56 kg/s (ii) 220.83 kg/s and(iii) 263.89 kg/sThe problem in this study was solved through numericalsimulations using ANSYS Fluent 19.2 software. Forvalidation, the initial results of the simulation of the existingmodel are compared with the data from the actualmeasurement results. After the existing model is valid, i.e.when the error is less than 5%, proceed with the simulationof three models made and compare the model which thelowest head loss in the CEP suction pipeline. The meshingmodel presented in Figure 7.Assuming that turbulent flow with kinetic energy coefficient( ) 1, the calculation of head losses (HL) follows thefollowing equation:𝑃1 𝑉12𝑃2 𝑉22𝐻𝐿 ( 𝑔𝑧1 ) ( 𝑔𝑧2 )𝜌2𝜌2where:HLP1ρV1gz1P2V2z2:::::head Loss (m)dynamic Pressure at point 1 (kPa)3density of water (kg/m )average velocity in point 1 (m/s)gravity (m/s2): elevation head in point 1 (m): dynamic Pressure at point 1 (kPa): average velocity in point 1 (m/s): elevation head in point 1 (m).191606-8484-IJMME-IJENS December 2019 IJENSIJENS

International Journal of Mechanical & Mechatronics Engineering IJMME-IJENS Vol:19 No:06125PressureStorage TankLevel tankTempSuctionPressureCEP A CEP BMassFlow RateAmpere,PressureFig. 4. The existing experimental test from DCS in Labuan Power Plant, Indonesia.Table IData experimental resultNo12345678ParameterLoadElectric currentDischarge pressureMass flow rateReservoir pressureSuction pressureReservoir tank 011012.40263.89-91.65-82.1016644.4Existing ModelModel-1Fig. 5. The geometry of the Existing Model and Model-1191606-8484-IJMME-IJENS December 2019 IJENSIJENS

International Journal of Mechanical & Mechatronics Engineering IJMME-IJENS Vol:19 No:06Model-2126Model-3Fig. 6. The geometry of Model-2 and Model-3Table IIBoundary condition settingMass flow rateTurbulent intensityHydraulic diameterModel-1122.78 kg/s3.20 %0.48 mValueModel-2150.16 kg/s3.20 %0.63 mModel-3122.78 kg/s3.20 %0.63mInlet-2Mass flow rateTurbulent intensityHydraulic diameter57.78 kg/s3.20 %0.48 m70.67 kg/s3.20 %0.63 m110.67 kg/s3.20 %0.63 m3Outlet-1Pressure outletTurbulent intensityHydraulic diameter-83,740 Pa3.20 %0.63 m-82,100 Pa3.20 %0.63 m-82,100 Pa3.20 %0.63 m4Outlet-2Wall---5FilterFilter strainer zonePorisity 0.174Porisity 0.174Porisity 0.174NoNameBoundary condition type1Inlet-12B. ValidationC. Simulation with Alternative DesignIn order to prove the accuracy of the model in this study,it is important to do the validation. The model is declaredvalid if the deviation between actual data and experiment isless than 5%. The result of the simulation of suction pipeCEP was validated by measuring the total head losses frominlet-1 to outlet-1 and the total head losses from inlet-2 tooutlet-2. This comparison is resumed in Table 3. From Table3, it can be seen that deviations are less than 3%. So, it cancontinue to run simulations for other models.There are three alternative designs of CEP suction pipeinstallation and meshing models are shown in Figure 7. Thecalculation result from simulations of the three models isshown in Table 4. Variation in this simulation is the changein total mass flow rate in three conditions in each model, sothat it can be known the change in the head loss at eachmodel. The mass flow rate settings are 180.55 kg/s; 222.22kg/s and 263.88 kg/s.191606-8484-IJMME-IJENS December 2019 IJENSIJENS

International Journal of Mechanical & Mechatronics Engineering IJMME-IJENS Vol:19 No:06127Table IIIValidation data head loss between experiment and simulation existing modelHead loss (m)Total mass flow rate (kg/s)Inlet-1Inlet-2Total mass flow rate (kg/s)Inlet-1Inlet-2Total mass flow rate Deviation1.08 %1.65 %1.75 %2.65 %1.01 %1.99 %Fig. 7. Meshing in existing model use for the validation processFig. 8. Meshing Model-1191606-8484-IJMME-IJENS December 2019 IJENSIJENS

International Journal of Mechanical & Mechatronics Engineering IJMME-IJENS Vol:19 No:06128Fig. 9. Meshing Model-2.Fig. 10. Meshing Model-3.Table IVHead Loss Calculation result in Existing Model, Model-1, Model-2 and Model-3ExistingParameter UnitTotal[kg/s] 180.55 220.83 263.86flow rateFlow[kg/s]122.78 150.16 [m]1.871.891.91head lossModel-1Model-2Model-3180.56222.22263.89 2.44 .861.881.861.871.90191606-8484-IJMME-IJENS December 2019 IJENSIJENS

International Journal of Mechanical & Mechatronics Engineering IJMME-IJENS Vol:19 No:06129Fig. 11. The comparison Velocity profile in total mass flow rate 263.88 kg/sec between; (a). Existing Model, (b). Model-1,(c). Model-2 and (d). Model-3.Fig.12. Comparison StreamLine profile in total mass flow rate 263.88 kg/sec between: (a). Existing Model, (b). Model -1 (c). Model-2 and (d). Model-3.191606-8484-IJMME-IJENS December 2019 IJENSIJENS

International Journal of Mechanical & Mechatronics Engineering IJMME-IJENS Vol:19 No:06130ACKNOWLEDGMENTSThis research did not receive any specific grant fromfunding agencies in the public, commercial, or not-for-profitsectors. Also, the authors have declared no conflict ofinterest.REFERENCES[1].[2].Fig. 13. Comparison Total Head Loss between Existing, Model 1, 2 and 3III. RESULTS AND DISCUSSIONThe simulation result of suction CEP has shown inFigures 11 and 12, shown that for all variations tested in thisstudy, head losses increase with increasing total mass flowrates. Weak secondary circulation even if the angle is largerand also even in smaller upstream pipe diameter. Model-2 isthe best streamline for all model was simulated.The calculation of total head loss has shown in Table 4and the comparison chart shown in Figure 13. For existingmodels, the mass flow rate of 180.55 kg/s has a total headloss of 1.87 meters; while for 220.83 kg/s it gives a total headloss of 1.89 meters; and for 263.86 kg/s produces a total headloss of 1.91 meters. On Model-1, that is for the mass flowrate of 180.56 kg/s has a total head loss of 1.85 meters; whilefor 222.22 kg/s it gives a total head loss of 1.88 meters; andfor 263.89 kg/s produces a total head loss of 1.90 meters. OnModel-2, namely for a mass flow rate of 180.55 kg/s has atotal head loss of 1.85 meters; while for 222.22 kg/s itgenerates a total head loss of 1.86 meters; and for 263.88kg/s produces a total head loss of 1.88 meters. On Model-3,for the mass flow rate of 180.56 kg/s it has a total head lossof 1.86 meters; while for 222.22 kg/s it gives a total head lossof 1.87 meters; and for 263.89 kg/s produces a total head lossof 1.90 meters.Existing suction pipe has the highest head losses at allflow rates operated. Model-1 and 3 practically have a greaterhead loss than Model-2, because with a larger angleconnection (junction), this will increases head loss. Model-2has the smallest total head loss among all the proposedmodels, so model-2 has the lowest cavitation potential inCED compared to other models.IV. CONCLUSIONThe simulation results show that for all variations testedin this study, head losses increase with increasing total massflow rates. By uniformizing the suction pipe diameter, whichis 630 mm, and changing the intersection angle from 90 to45 has been proven to have succeeded in giving the lowesthead loss in the suction pipe. In maximum mass flow rate, in263.88 kg/s, the total head loss Existing model is 1.91 m,Model-1 is 1.91 meters, Model-2 is 1.88 meters and Model-3is 1.91 meters. The alternative design of Model-2 has thesmallest head losses. This promises the smallest cavitationrisk in CEP and will certainly increase the reliability of theLabuan Power 13].[14].[15].Haque, M. E., Islam, M. S., Islam, M. R., Haniu, H., & Akhter, M. S.(2019). Energy efficiency improvement of submersible pumps usingin barind area of Bangladesh. Energy Procedia, 160, 2.127)Matlakala, M. E., Kallon, D. V. V., Simelane, S. P., & Mashinini, P.M. (2019). Impact of Design Parameters on the Performance ofCentrifugal Pumps. Procedia Manufacturing, 35, 5.027)Wang, C., Shi, W., Wang, X., Jiang, X., Yang, Y., Li, W., & Zhou,L. (2017). Optimal design of multistage centrifugal pump based onthe combined energy loss model and computational fluid dynamics.Applied Energy, 187, 1.046)Song-Sheng Deng, Guo-Dong Li, Jin-Fa Guan, Xao-Chen Chen, LuXing Liu (2019). Numerical Study of cavitation in centrifugal pumpconveying different liquid materials. Result in )Jean-Piere Franc, Jean-Marie Michel (2004). Fundamentals ofCavitation. France: Institute National Polytechnique de Grenoble(INPG).Donald P, Sloteman (1995). Avoiding cavitation in suction stage ofhigh energy pump. World Pumps(https:worldpumps.com)WORLD PUMPS (2018). Pump Cavitation and how to avoidit (https:worldpumps.com)Maxime Binama, Alex Muhirwa, Emmanuel Bisengimana (2016).Cavitation Effects in Centrifugal Pumps-A Review. Int. Journal ofEngineering Research and transfer.2019.05.008).Bruce R. Munson, Theodore H Okiishi, Wade W Huebsch (2009).Fundamentals of Fluid Mechanics. New York: John Wiley & Sons,Inc.Bordoloi, D.J. & Tiwari, R. (2017) Identification of suction flowblockages and casing cavitations in centrifugal pumps by optimalsupport vector machine. J Braz Soc. Mech Sci .092)Delly, J. (2009) Pengaruh Temperatur Terhadap Terjadinya Kavitasipada Sudu Pompa Sentrifugal. Jurnal Ilmiah Teknik Mesin Vol.1No.1 Tahun 2009.A Yu, W P Yu, Z B Pan, X W Luo, B Ji, X Y Xu (2014). CavitationPerformance Evaluation for a Condensate Pump. 6th InternationalConference on Pumps and Fans with Compressors and 4)Bellary, S. A. I., & Samad, A. (2013). Exit Blade Angle andRoughness Effect on Centrifugal Pump Performance. ASME 2013Gas Turbine India 31)Luo, H., Fytanidis, D. K., Schmidt, A. R., & García, M. H. (2018).Comparative 1D and 3D numerical investigation of open-channeljunction flows and energy losses. Advances in Water Resources, .2018.05.012)Gao, R., Zhang, H., Li, A., Liu, K., Yu, S., & Deng, B. (2018). Anovel low-resistance duct tee emulating a river course. Building andEnvironment, 144, .08.034)191606-8484-IJMME-IJENS December 2019 IJENSIJENS

A Pump is a turbomachine which increases the energy level of fluid when fluid flow through them [1]. In the power plant system, Condensate Extraction Pump (CEP) plays an important role. When the CEP does not work in design conditions, it can disrupt the reliability of the condensate