A Simulation-Based Optimal Accelerated Degradation Test Plan for a Nonlinear Wiener Process

Purpose: This study proposes an optimal accelerated degradation test (ADT) plan based on a nonlinear Wiener process. The optimal ADT plan is devised using a simulation-based method that permits the effective determination of optimal ADT conditions and applied to nonlinear degradation data of lithium-ion battery pack.<br/>Methods: An optimal ADT design approach based on simulated ADT data is reported herein. First, maximum likelihood estimates (MLEs) of the nonlinear Wiener process degradation model with random effects were derived, then a Monte Carlo simulation was performed to generate lifetime data based on the MLEs. Using the general optimization procedure for accelerated life test (ALT), the optimal stress levels and sample allocation ratios were determined to minimize the asymptotic variance of the lifetime distribution.<br/>Results: The proposed approach was applied to real battery pack data for electric vehicles (EVs). Subsequently, the optimal sample allocation ratios at both maximum and minimum stresses were determined.<br/>Conclusion: The proposed simulation-based ADT plan is particularly effective when deriving the Fisher information matrix for the parameters, which is a challenging task given the complexity of the degradation model or when lifetime distributions are difficult to derive.