The Global Performance Prediction Of The Multiple Centrifugal Pump

The development of science and technology appeals to the improvement of pumps'performance, but traditional design and performance prediction method is not enough to satisfy the need. In order to predict the global performance of the multiple centrifugal pump, STAR-CD code is selected, which is popularly used in academic and industrial applications, as the research platform. A multiple centrifugal pump has been simulated for accurate computation and predicting its performance by using the STAR-CD software. Based on the ancient analysis of pump performance predictions, hydraulic loss, leakage loss and friction loss were considered in this paper. The object is to get a satisfied prediction result and conclude a method of global performance prediction of the multiple centrifugal pump. In this paper, the flow status in gaps was analyzed. The result showed that the leakage loss and the friction loss were related to the pressure difference between, the size of gap, the circumference velocity, etc. In this paper, to reduce workload, the multiple centrifugal pump was analyzed into several objects, and every object was just selected a part to do analysis. The three-dimensional models were built by using UG software. grids were Generated by using ICEM-CFD. The simulations of three-dimensional flows at design and off-design conditions are accomplished. The k ? εhigh Reynolds turbulent model was used to finish these jobs, and the SIMPLE based method for coupling the velocity and pressure fields is employed which facilities and the momentum equations. Some discussion about the effect of the gaps has been accomplished to give some suggest to the pump manufacturers. Calculating the applied power based on Torque, and actual power based on the head and flux. The computational performance values are compared with the tested values, and some error analysis was finished. This paper tried to accomplish the global performance prediction of the multiple centrifugal pump, and the results in this paper apply some data for the succeeding analysis.