Simulation and testing of centrifugal pump cavitation.

Date
2024-02
Journal Title
Journal ISSN
Volume Title
Publisher
Stellenbosch : Stellenbosch University
Abstract
ENGLISH ABSTRACT: Cavitation, the formation of vapour filled bubbles in areas of low pressure, can be deleterious to centrifugal pump operation. Cavitating flow may cause efficiency loss, damage to impeller surfaces, and excessive vibration in pump systems. Some cavitation is unavoidable and even harmless, however, preventing cavitation will in general enhance pump reliability, and efficiency. In this research question, computational fluid dynamics was used to simulate cavitation inception using single‑phase, steady‑state flow, in a centrifugal pump. The pump was a low specific speed radial flow centrifugal pump that is used for the conveyance of water or other products. Simulations were performed using Cadence (NUMECA) Fine Turbo 16.1 and AutoGrid 5 meshing software. The simulated performance characteristics were compared to experimental performance measurements. The cavitation model determined the net positive suction head (NPSH) of cavitation inception (NPSHi) in the post‑processing stage of the simulation. The cavitation model is based solely on the inlet pressure and minimum static pressure in the impeller to determine cavitation inception. The simulated NPSHi values were 0.903 m, 1.53 m, and 2.33 m, at shaft speeds of 800 RPM, 1000 RPM, and 1200 RPM respectively. This was compared against scaled NPSH3 values of 0.281 m, 0.475 m, and 0.722 m respectively. As expected, the simulated NPSHi values were consistently higher than NPSH3 values. Vibration sensors were employed to measure cavitation inception in the pump, but due to stall occurring in the impeller, the point of cavitation inception could not be distinguished above the noise caused by backflow vibration. Broadband excitation was measured on the pump, which was consistent with other test setups where backflow was present. Backflow was also confirmed by the simulations. Experimental performance measurements showed that for accurate performance simulation, a model containing only the impeller and diffuser was insufficient. Adherence to the predicted NPSHi will prevent operating pumps under damaging cavitating conditions, however, the correspondence of these values with true cavitation inception is tenuous.
AFRIKAANSE OPSOMMING: Kavitasie, die vorming van dampgevulde borrels in areas van lae druk, kan skadelik wees vir sentrifugale pomp werking. Kaviterende vloei kan lei tot verlies van doeltreffendheid, skade aan stuweroppervlaktes, en oormatige vibrasie in pompstelsels. Sekere kavitasie is onvermydelik en selfs onskadelik, maar die voorkoming van kavitasie sal oor die algemeen pompbetroubaarheid en doeltreffendheid verbeter. In hierdie navorsingsvraag is gebruik gemaak van berekenings vloeimeganika om kavitasie‑aanvang te simuleer met behulp van enkelfase, bestendige vloei in ’n sentrifugale pomp. Die pomp was ’n lae spesi􀏐ieke spoed radiale vloei sentrifugale pomp wat gebruik word vir die vervoer van water of ander produkte. Simulasies is uitgevoer met behulp van Cadence (NUMECA) Fine Turbo 16.1 en AutoGrid 5 rooster sagteware. Die gesimuleerder werkverrigting‑karakteristieke is vergelyk met laboratorium werkverrigtingskarakteristieke metings. Die kavitasie‑model het die netto positiewe suighoogte (NPSH) van kavitasie‑aanvang (NPSHi) in die naverwerkingstadium van die simulasie bepaal. Die kavitasiemodel is uitsluitlik gebaseer op die inlaatdruk en minimum statiese druk in die stuwer om aanvangskavitasie te bepaal. Die gesimuleerde NPSHi‑waardes was 0.903 m, 1.53 m, en 2.33 m, teen snelhede van 800 OPM, 1000 OPM, en 1200 OPM onderskeidelik. Dit word vergelyk met geskaalde NPSH3‑waardes van 0.281 m, 0.475 m, en 0.722 m onderskeidelik. Soos verwag is die gesimuleerde NPSHi‑waardes konsekwent hoër as NPSH3‑waardes. Vibrasiesensors is gebruik om kavitasie‑aanvang in die pomp te meet, maar weens terugvloei in die stuwer, kon die punt van kavitasieaanvang nie onderskei word bo die vibrasie geraas wat veroorsaak is deur te rugvloei vibrasie nie. Breëbandopwekking is op die pomp gemeet en was konsekwent met ander toetsopstellings waar terugvloei teenwoordig was. Terugvloei is ook bevestig deur die simulasies. Die gesimuleerde kavitasie‑aanvang was egter konsekwent wanneer dit vergelyk word met gepubliseerde NPSH3‑ kurwes. Eksperimentele werkverrigtingsmetings het getoon dat vir akkurate werkverrigting simulasie, ’n model wat slegs die stuwer en diffusor bevat, onvoldoende was. Nakoming van die voorspelde NPSHi sal voorkom dat pompe onder skadelike kaviterende toestande bedryf word; egter, die ooreenstemming van hierdie waardes met ware kavitasie‑aanvang is broos.
Description
Thesis (MEng)--Stellenbosch University, 2024.
Keywords
Citation