Lifetime prediction and reliability modeling of perovskite solar cells using the proportional Hazard Chen model

This study introduces a reliability-based framework that quantifies perovskite solar cell (PSC) degradation by estimating the T80 lifetime (time to 80% of initial power conversion efficiency) from experimental aging data using exponential regression. The resulting failure times, collected under an adaptive progressive Type-Ⅱ censoring scheme, were analyzed using the Proportional Hazard Chen (PHC) distribution. This model accommodates increasing hazard rates characteristic of PSC aging mechanisms. Model parameters were estimated using maximum likelihood estimation (MLE) and Bayesian inference, enabling comprehensive reliability metrics including survival probability, hazard rate, and mean time to failure (MTTF). The procedure was validated using degradation data from PSCs with titanium dioxide and nickel oxide transport layers, supplemented by Solar Cell Capacitance Simulator (SCAPS). The empirical results show that the survival probability decreases from approximately 0.98 at 0.5 hours to below 0.02 at 17 hours, while the hazard rate increases from approximately 0.04 to above 0.75, confirming an accelerating degradation pattern. The Monte Carlo simulation study further demonstrates that both MLE and Bayesian approaches can estimate the PHC model effectively, with improved accuracy when the effective number of observed failures and the termination time increase. Bayesian inference provides more conservative long-term reliability and MTTF predictions under heavy censoring, making it useful for risk-aware lifetime assessment. Overall, the proposed framework provides a practical statistical tool for predicting, comparing, and benchmarking PSC lifetimes in photovoltaic reliability studies.