Multidisciplinary Optimization Design of pVADs Using Analytical Target Cascading-Guided Genetic Algorithm.

Area of Science:

• Biomedical Engineering
• Cardiovascular Devices
• Computational Fluid Dynamics

Background:

• Percutaneous ventricular assist devices (pVADs) face design challenges due to impeller-motor inconsistency, impacting efficiency in critical scenarios like cardiogenic shock.
• Current design inefficiencies lead to wasted computational and experimental resources, hindering optimal device performance.

Purpose of the Study:

• To address limitations in pVAD design by developing a novel, systematic optimization method.
• To resolve the trade-off between impeller performance and motor power requirements for diverse clinical applications.

Main Methods:

• A system-level optimization framework integrating artificial neural networks (ANN), analytical target cascading (ATC), and NSGA-II.
• ATC ensures coordinated parameter matching between motor and impeller; NSGA-II refines design parameters; ANN reduces computational load.

Main Results:

• The integrated method achieved balanced cardiac output and motor power, with a prototype pressure head >80 mmHg at 5 L/min and hemolysis index <0.02.
• Total pVAD efficiency increased to 33.65% from a baseline of 7.59%, demonstrating significant algorithmic improvement.
• The optimization successfully balanced hydraulic performance and minimized hemolysis risk.

Conclusions:

• The proposed framework enables efficient and feasible pVAD optimization by resolving blade performance and motor power trade-offs.
• Optimal blade design, including inlet/outlet angles and chord length, is crucial for hydraulic performance and reduced hemolysis.
• Comprehensive consideration of interdependent design aspects is vital for superior pVADs.