Sensitivity Analysis Powered Simplification of the Electric Vehicle Powertrain Design Optimization

The electric vehicle powertrain is a complex system containing multiple design parameters with mutual restriction, which brings difficulties to the design work. This paper proposes a data-driven simplification method aiming at quantifying the contribution of variables, which leads to identifying crucial parameters and mitigating the optimization burden. Global sensitivity analysis of powertrain design variables is adopted, and variables influencing significantly on energy consumption performance are definite. The result guides to simplify the powertrain design, achieving a 33% effect. It reduces the dimension of the optimization problem with almost the same performance in energy efficiency and stability in validation, promoting cost reduction in manufacturing.