Ways to solve the problems of creating predictive measurement systems

The main problems of creating predictive measurement systems are considered. The role of predictive analytics, digital models and digital twins of measuring systems in predicting the drift of metrological characteristics, estimating the remaining resource and reducing the risk of metrological failure is shown. The key aspects of the development of intelligent measuring systems are considered: requirements for metrological support and features of measuring instruments based on artifi cial intelligence, the infl uence of their multicomponence and opacity of algorithms on the procedures of verifi cation, calibration and metrological self-control. Based on current standards and scientifi c publications, the terminology used in this fi eld is evaluated, and a defi nition of a predictive measuring system is proposed, including functions for predicting both process parameters and the metrological dependability of measuring instruments. Recommendations are formulated to improve methods for assessing the risk of metrological failure and the need to detail the requirements for intelligent measuring systems through the development (revision) of standards is noted.