Additive hazard model based on equipment age and health index
| Autor: | Yan-Guang Hu, Jian-Ping Chen, Kun-Yun Wang, Zhi-Jun Xu, Xiang-Kun Liu |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
Trust region
Series (mathematics) Computer science Monte Carlo method Explained sum of squares Sampling (statistics) Failure rate 06 humanities and the arts 0603 philosophy ethics and religion 01 natural sciences 010104 statistics & probability 060302 philosophy Statistics 0101 mathematics Time series Variable (mathematics) |
| Zdroj: | 2016 Prognostics and System Health Management Conference (PHM-Chengdu). |
| DOI: | 10.1109/phm.2016.7819921 |
| Popis: | When describing the influence of equipment state to failure rate, traditional additive hazard model (AHM) has certain defects. Therefore, an improved AHM with equipment age and health index (HI) as variables is established, in which HI is used to describe the equipment state. In order to fit out the model's five parameters, a nonlinear constrained optimization problem was posed. A sampling algorithm based on Monte Carlo method was also proposed to obtain parameter ranges and initial values. When the sample data are complete, the parameters can be fitted out by Trust Region algorithm. Historical data of most equipment lack HI time series at present, so non-sequential Monte Carlo method is applied in the section of example analysis to obtain approximate solution of parameters. Fitting results of the example show that an AHM with practical significance can be acquired through the parameter fitting method. Compared with the results of two models with age as the only variable, sum of squares due to error (SSE) of the AHM decreases by at least 10%. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|