MODELING AND RELIABILITY ANALYSIS OF MULTI-STRESS COUPLED ACCELERATED LIFE MODEL BASED ON FUZZY INFORMATION FUSION
To address the challenges in identifying and accurately characterizing coupling effects in existing multi-stress acceleration models, this paper adopts fuzzy mathematics to characterize uncertainties in both empirical knowledge and data, and proposes a multi-stress accelerated life evaluation method that explicitly accounts for coupling effects. The method first adopts the Distance Correlation Coefficient (DCOR) technique to establish a preliminary screening framework for coupling effects. The objective statistical analysis results are then integrated with subjective empirical judgments via the Fuzzy Number–Entropy Weight–TOPSIS fusion method, yielding an importance ranking of the coupling effects. The selected and optimized acceleration model is subsequently embedded into the log-likelihood function of the Weibull distribution, and the Particle Swarm Optimization (PSO) algorithm is utilized to estimate model parameters and predict service life. Case study results demonstrate that, compared with the traditional Pearson correlation coefficient method, the proposed method exhibits a broader applicability under nonlinear and non-monotonic conditions and significantly improves the prediction accuracy of product service life under constant stress conditions, thereby validating the effectiveness of the model.