Subsample Selection and Test Confidence Improvement in Equipment Accelerated Life Tests

Abstract

In this study, the evaluation of equipment technical status was enhanced by refining the secondary indexes of the exiting status evaluation system based on equipment quality monitoring standards. An objective assessment of equipment technical status was achieved through this refinement. Additionally, a scientifically grounded sampling scheme was developed, and a multi-objective nonlinear integer programming model was established for accelerated life tests which yields an approximate optimal solution. To address instances where the conclusions from small sample tests do not meet the confidence level in classical reliability sampling theory, an improved dynamic time warping algorithm was employed. This algorithm measures the similarity between the original part and the spare part, demonstrating that the inclusion of spare part samples increases the sample size without altering the distribution of the original sample. Furthermore, in addressing the challenges of inheriting information from both new and old equipment, the mutual information method was employed to enhance the calculation of inheritance factors. These factors were then fused with multi-source test data, improving test confidence and offering valuable insights for enhancing confidence in small sample tests.