Digital Twin Technology As A Basis Of The Industry In Future

Abstract

The transition to the next technological order requires robotics and automation in the industry. Management of such process of production is impossible without modelling of equipment, forecasting of possible emergencies based on information from various sensors. Such models are called twin models. The purpose of the study is to identify approaches to creating a platform for implementing dynamically changing models - twins of physical and technical objects that are capable of performing ongoing training and adapting to changing external conditions based on the processing of a large volume of measurement information.It is shown that the use of RBF ANN for models that allow us to obtain solutions of systems of partial differential equations to perform constant additional training and dynamic change of its structure in the process of application. The results of the development of the landing system of the module, descending to the surface of the planet using the NARX ANN, are presented. The model has the property of additional training to adapt to the composition of the underlying soil.A model of an early warning fire system for fire using the NARX ANN is developed. Model includes a 3D model of premises, a genetic algorithm for the optimal placement of sensors, and modeling fires of different type of ignition. The proposed methods and models form the basis for creating digital twins of the considered classes of physical and technical objects.The stages of the unified process of constructing similar models are formulated.