Artificial intelligence approach for energy and entropy analyses of a double-suction centrifugal pump

As global energy and environmental concerns grow, reducing energy consumption and emissions has become crucial. Centrifugal pumps, widely used in various applications, are significant energy consumers, making energy loss reduction essential. This paper presents an optimization method for centrifugal pumps to minimize entropy generation losses while maintaining the same head. Using Computational Fluid Dynamics (CFD) and intelligent optimization algorithms, the study focuses on a double-suction centrifugal pump with a European specific speed of 43.8. The impeller blade profile is optimized to reduce entropy generation while maintaining head. Artificial Neural Network (ANN) and NSGA-II algorithm were combined to optimize the impeller blade Angle. The results show that the optimized design reduces the impeller's turbulent entropy generation by 19.3 %, and the volute's by 24.2 %. The optimized pump exhibits a more uniform pressure distribution, suppressing flow separation and vortices. The optimized design reduces entropy generation at the impeller outlet and shroud, while shrinking the low-pressure zone on the shroud. The total entropy generation loss decreases by 15.16 %, and the efficiency increases by 1.73 %, confirming the effectiveness of the optimization method in improving pump performance and energy efficiency.