A practical method for speeding up the cavitation prediction in an industrial double-suction centrifugal pump

Researches have over the past few years have been applying optimization algorithms to quickly find optimum parameter combinations during cavitation optimization design. This method, although better than the traditional trial-and-error design method, consumes lots of computational resources, since it involves several numerical simulations to determine the critical cavitation point for each test case. As such, the Traditional method for NPSHr prediction was compared to a novel and alternative approach in an axially-split double-suction centrifugal pump. The independent and dependent variables are interchanged at the inlet and outlet boundary conditions, and an algorithm adapted to estimate the static pressure at the pump outlet. Experiments were conducted on an original size pump, and the two numerical procedures agreed very well with the hydraulic and cavitation results. For every flow condition, the time used by the computational resource to calculate the NPSHr for each method was recorded and compared. The total number of hours used by the new and alternative approach to estimate the NPSHr was reduced by 54.55% at 0.6 Qd, 45.45% at 0.8 Qd, 50% at 1.0 Qd, and 44.44% at 1.2 Qd respectively. This new method was demonstrated to be very efficient and robust for real engineering applications and can, therefore, be applied to reduce the computation time during the application of intelligent cavitation optimization methods in pump design.