Analysis of importance and reliability in the hydraulic system based on failure data

In this paper, a novel technique for the reliability and importance analysis of hydraulic system is presented based on model-free sampling and state dependent parameter(SDP) method. The analysis results provide a reference guideline for the reliability assessment and design of the system. Instead of estimating a distribution function, this new method is applied to operate without a probabilistic model. According to the assignment criterion and distance criterion, the information contained in a small experiment sample is extracted directly to produce a second sample of considerably larger size, and the second sample completely reflects the properties of the original small sample. Due to the method performing without estimating a distribution, the analysis procedure reduces the error of the truncation and numerical fitting. Comparing to the traditional sampling, the model-free sampling specializes in sparse point data. To improve current probability safety assessment of the system, a new variance-based important measure is applied to measure the important effects of the basic components on the whole uncertainty of the lifetime of the system. Combining with advantages of the SDP method, a solution of the importance measures is provided, the importance rank is as follows: solenoid valve > pump > cylinder > oil filter > valve > tank, which illustrates the solenoid valve contributes to system lifetime most, while the tank contributes least. So, useful information can be obtained for system analysts and design engineers to trace the system weaknesses and to guide effective system improvement.