Residual life prediction method for two-stage degraded product

Abstract

The invention discloses a residual life prediction method for a two-stage degraded product, which belongs to the field of residual life prediction in prediction and health management, and mainly comprises the steps of degradation process modeling, model parameter estimation and residual life prediction, wherein the degradation process modeling is to establish a two-stage degradation model by usinga nonlinear Wiener process; the model parameter estimation comprises the following steps: collecting historical degradation data; estimating model parameters based on a unit maximum likelihood estimation method; statistical analysis of random parameter distribution; the residual life prediction comprises the following steps: obtaining first arrival time distribution; deriving a state transition probability function; and predicting the residual life of the product based on the estimated parameters. The residual life of the two-stage degraded product can be effectively predicted, the operationreliability of the product is ensured, the maintenance cost is reduced, and safety accidents are avoided.

Description

technical field [0001] The invention belongs to the field of prediction and health management, and relates to a method for predicting the remaining life of a two-stage degraded product. Background technique [0002] Remaining life prediction is the core content of prediction and health management technology. It refers to the effective time interval for the product to lose the specified function at the current moment. It is an important indicator reflecting product reliability. important to economics. [0003] In the past ten years, residual life prediction has received extensive attention and in-depth research. Among them, the Wiener process has been greatly developed because it can describe non-monotonic degenerate trajectories and has good mathematical properties. But it is worth noting that in most life prediction methods based on the Wiener process, the degradation rate is usually fixed and does not change with time. However, in actual engineering, due to changes in e...

AI Technical Summary

Problems solved by technology

For such two-stage degradation products, traditional life prediction methods are not accurate enough

[0004] In order to improve the prediction accuracy, many researchers have proposed a two-stage degradation model, but the existing methods mainly focus on whether the change point is random or whether individual differences are considered, while ignoring the degradation nonlinearity and life analysis of each stage derivation of solution

Images