An application-time-window based credibility assessment method for equipment digital twins in manufacturing

Abstract

To effectively assess the credibility of equipment digital twins (DT) for manufacturing, a new method fully considering the dynamic evolution process is urgently needed. In this paper, the influence of the evolution process on the applications is scrutinized, the connotation of equipment DT credibility assessment is expanded, and an assessment method to evaluate the general evolution process of equipment DTs in manufacturing is proposed. The method obtains the application-time-window (ATW) by fitting a joint probability distribution based on observation of actual application time points. Then the method uses the probability of application occurrence as a weight to quantitatively assess the model performance throughout the evolution process. The effectiveness of the method was validated by ablation experiments in manufacturing simulation scenes of workpieces transferring with robotic arms. Compared with existing assessment methods, the error of quantified DT credibility value is reduced by at least 65% and the assessment result can be obtained once the evolution is completed.