Inverse Moment Estimation of Three-parameter Power Function Distribution with Extension to Accelerated Life Test

Power function distribution is used frequently in both reliability and life assessment. However, in some cases, the traditional method fails in parameter estimation of the three-parameter power function distribution. We propose a novel method to resolve this difficulty. This method utilizes the constant moment of the population while considering the bias-corrected estimator of the location parameter and scale parameter. Further, the proposed method is extended to accelerated life tests. The validity of life assessment is ensured by deriving the necessary condition for mechanism equivalence, and the effectiveness of the proposed method is proven using simulations. Finally, a real-world example is presented to illustrate the performance of the method.