An integrated lifetime prediction method for double-nut ball screws subject to preload loss failure mode

Existing lifetime prediction methods for ball screws are mainly based on the failure mechanism modeling, while the accuracy is quite limited due to the need for a large number of failure samples and the ignorance of the uncertainty during the failure evolution. Therefore, we proposed an integrated lifetime prediction method for the double-nut ball screws utilizing both the degradation mechanism model and the uncertainty evaluation method. The Bayesian inference is used to update the posterior distribution of the wear coefficient on the basis of the condition monitoring data. The posterior distribution of wear coefficient gets narrower with more degraded preload data acquired, which indicates the uncertainty of the wear coefficient due to different material properties and working conditions is decreased. The experimental results showed that, without the need for a large number of failure samples, the proposed integrated lifetime prediction method for double-nut ball screws achieved much higher accuracy than existing methods.