Optimization design of centrifugal pump cavitation performance based on the improved BP neural network algorithm

• Published in: Measurement : journal of the International Measurement Confederation Vol. 245; p. 116553
• Main Authors: Wang, Yuqin, Shao, Jiale, Yang, Fan, Zhu, Qingzhuo, Zuo, Mengqiang
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 15.03.2025
• Subjects: Adaptive Genetic Algorithm; BP Neural Network; Cavitation Performance; Centrifugal Pump; Optimization Design; Centrifugal Pump; BP Neural Network; Cavitation Performance; Optimization Design; Adaptive Genetic Algorithm
• ISSN: 0263-2241
• DOI: 10.1016/j.measurement.2024.116553

•Identified key geometric parameters affecting cavitation via Plackett-Burman design.•Built a sample space using numerical simulation and Latin hypercube sampling.•Developed an adaptive genetic-BP algorithm for enhanced design efficiency.•Shown significant performance improvements in optimized pumps, reducing cavitation and noise. A novel approach was proposed to tackle the problem of cavitation in centrifugal pumps during operation, which resulted in reduced performance and equipment breakdown. By analyzing the cavitation mechanism and key impeller parameters, the sample space was successfully constructed using numerical simulations and Latin hypercube sampling. An adaptive genetic-Back Propagation (BP) neural network algorithm was designed, combining the global search advantage of the genetic algorithm with the local optimization characteristics of the BP neural network, thereby improving optimization efficiency and accuracy. The experiment verified the significant effect of this method in improving the cavitation resistance performance of centrifugal pumps and reducing noise, which has important theoretical and practical value for the research and design of pump equipment.