An improved ABC algorithm approach for shape optimization of the blood centrifugal pump impeller

Abstract

A high efficiency blood centrifugal pump has low blood damage and little energy loss, so it is important to improve hydraulic efficiency by optimizing the design. In this paper, an improved Artificial Bee Colony algorithm combined with the adjoint method (ABCA) is proposed to optimize the blood centrifugal pump impeller. The introduction of gradient information enhances the search performance and the convergence judgment of ABC algorithm. Besides, its random search feature can effectively avoid the adjoint optimization method from trapping in local optimum. The ABCA algorithm are applied to increase the hydraulic efficiency of a self-designed blood pump. After optimization, five design parameters of the impeller are optimized, and hydraulic efficiency of the impeller improves about 3.90%. Compared with ABC algorithm, the ABCA algorithm saves about 58% of calculation time, and its convergence performance is much better than ABC in the middle of the iteration.