Life Assessment for Motorized Spindle with Zero Traumatic Failure Data Based on Subdistribution Competing Risk Model

Considering the influence of performance degradation on a product's traumatic failure, under the condition that only degradation data are observed and no traumatic failure data are observed, this paper proposes a subdistribution competing risk model to achieve a motorized spindle life assessment. This paper assumes that the failure rate ratio of the tested products does not change with time under different stress levels. Basic reliability with zero traumatic failure data is modeled by a unilateral confidence limit method under a two-parameter Weibull distribution. Performance degradation data are taken as covariates. The regression coefficients of the covariates are calculated by SPSS software. Then, a subdistribution competing risk model is constructed, which reflects the dependency relationship between reliability and performance degradation, and the product's reliability life can be evaluated accordingly. The correctness and advantages of the model built in this paper are verified by a case analysis combined with the performance degradation information of a motorized spindle.

Automated summary

This paper proposes a subdistribution competing risk model to assess the life of a motorized spindle, considering the impact of performance degradation on traumatic failure. The model assumes that the failure rate ratio of tested products remains constant under different stress levels. Basic reliability, without traumatic failure data, is modeled using a unilateral confidence limit method under a two-parameter Weibull distribution. Performance degradation data are used as covariates, and the regression coefficients are calculated using SPSS software. The constructed subdistribution competing risk model reflects the dependency relationship between reliability and performance degradation, enabling the evaluation of the product's reliability life. The correctness and advantages of the proposed model are verified through a case analysis using the performance degradation information of a motorized spindle.