Robust design Optimization of a blood pump by Reduced Order Models

The leading cause of mortality is cardiovascular diseases. In recent decades, significant effort has been put into designing ventricular assist devices (VADs) to treat heart conditions, as heart transplantation can benefit only a small number of patients [Wu, 2022]. These pumps are usually employed to assist the heart ventricle in pumping blood to the patient's body. However, the device reliability, as well as mechanical blood damage (hemolysis), still represent a serious challenge in clinical applications, limiting their potential benefits.

In this study we developed and validated a numerical methodology for the high-fidelity design of a centrifugal blood pump. The pump originally designed by the U.S. Food and Drug Administration (FDA) [Malinauskas et al, 2017] is subjected in this study to robust shape optimization. The effects of operating uncertainties on the hydrodynamic and hemocompatibility performance of the device are taken into account to obtain an optimal configuration whose characteristics are minimally affected by these variations.

To reduce simulation time, a limited series of CFD simulations was used to train a Reduced Order Model (ROM), which allows to approximate the flow fields inside the pump, as a function of a reduced number of modes (principal components). The entire process of simulation automation and robust optimization has been integrated into the VOLTA web-based collaboration platform, that includes the generation of ROM models through the romBOX tool and the analysis of results.

Although the approach is demonstrated on a benchmark case, i.e. the FDA blood pump, it can easily translate to more complex VAD geometries. More biocompatible and robust blood pumps can be designed using this approach, thereby minimizing the risk of postoperative complications for patients.