Reliability Modeling Of Parts Considering Correlation Between Failure Modes

Reliability, as one of the most important quality indices, has been embodied in the stages of development, manufacturing and maintenance of the part. Therefore, it is of importance to develop scientific and rational theory and method for the reliability analysis.The part usually has multiple failure modes. Sudden failure and degradation failure are the basic failure modes of the part. Considering that there exists correlation between failure modes, the reliability model obtained will have great error if the failure modes are assumed to be independent. In view of this, three circumstances were taken into consideration: sudden failure, degradation failure and both of them respectively, and the reliability models of the part were built up, in which the correlation between the failure modes was taken into consideration. The contents include:(1) First, in order to ensure that the part won’t fail at the initial stage, the reliability modeling method of the part considering the correlation between sudden failure modes was researched. Gear, as the most important mechanical transmission parts, has two main failure modes: bending fatigue and contact fatigue. The gear was assumed to suffer from tangential loads which follow normal distribution. Then, the Monte Carlo simulation was adopted to obtain the scatter points distribution of the failure modes. A bivariate normal distribution was proposed to establish the joint distribution of the failure modes by observing the distribution law of the scatter points, and the solution method of reliability of the parts was given. Finally, this method was applied to analyze the reliability of the gear in the wind turbine gearbox, which proved the validity of the model.(2) Then, considering that the performances degrade under the stochastic loads, the reliability modeling method of the parts considering the correlation between degradation failure modes was researched. As the strength degradation rate changes over time, a non-stationary Gamma process-based strength degradation model was proposed and the parameter estimation method in the model was given, which was applied to analyze degradation law of the gear, which validated the effectiveness of the model. Next, the Inverse Gaussian process was used to build the reliability model of degradation processes of each performance. Considering the correlation between the degradation increments of each degradation process, the reliability model of the parts was built based on the Copula function and the parameter estimation method was also given. The validity of the model was confirmed by numerical example.(3) Finally, considering that the part will face both sudden failure and degradation failure when it suffers from both shock loads and stochastic loads, the reliability modeling method of the part considering the correlation between sudden failure mode and degradation failure mode was researched. A reliability model under effective shock loads was proposed considering that the material has the ability to resist against the small shock loads. The reliability modeling method of the parts under one shock process and one degradation process, and one shock process and two degradation processes were studied respectively. The model was used to analyze the reliability of the pin shaft in the MEMS, which demonstrated the validity of the model.The reliability models of the parts by considering the correlation between sudden failure modes, degradation failure modes and the correlation between sudden failure and degradation failure were built respectively, which provided the accurate reliability analysis method for the part and was of great significance for guiding the reliability design of the part and for making maintenance plan.