Technologies of digital modeling and artificial intelligence in Tribo-Fatigue
Abstract:
Artificial intelligence technologies are considered in comparison with mathematical modeling using the examples of a number of critical purpose technical systems studied in tribo-fatigue. An applied definition of artificial intelligence as a technical automated system rather than a set of specific technologies is proposed. Tribo-fatigue methodology is presented for the consistent formulation and solution of problems of interaction of a system of many bodies with previously unknown contact surfaces, determination and prediction of their three-dimensional stress-strain state, volumetric damage state and multi-criteria limiting states taking into account its simultaneous complex thermal-force loading by contact and non-contact forces. This made it possible to create multi-element digital twins of a number of technical systems of critical purpose, used to optimize these systems for damage and support management decisions Unified approach to application of artificial neural networks and mathematical modeling in intelligent models of tribo-fatigue and mechanothermodynamic systems is proposed. The positive direct effect of artificial neural networks on the approximation of the results of mathematical modeling and the inverse effect of high-quality data obtained as a result of mathematical modeling on artificial neural networks are shown.
