INTELLIGENT EQUIPMENT COMPLEX MANAGEMENT SYSTEM WITH FAILURE PREDICTION

Abstract

The article considers the methodological and applied aspects of the development of an intelligent technical equipment management system with the function of failure prediction. The relevance of the work is determined by the need to ensure high reliability, continuity and economic efficiency of the functioning of modern aggregated technological complexes in the conditions of industrial digitalization. The task of transition from traditional maintenance according to the regulations to adaptive preventive management of the technical condition based on the analysis of real operational parameters is set. Within the framework of the study, an architectural solution of the system was proposed, including modules for data collection, pre-processing of signals, analytics, forecasting and decision-making. To implement the predictive function, an artificial neural network model of the LSTM type, trained on time series of vibration and temperature characteristics, was used. The mathematical model of the technical state is described by the degradation function σ(t), which reflects the current level of wear and tear of the equipment, and the forecast of the residual resource (RUL) is defined as an estimate of the time until the critical state is reached. A combination of spectral analysis (STFT) and multidimensional normalized feature space was used to improve the accuracy of the results. Experimental testing of the system was carried out on a laboratory stand with a model of a mechanical drive. Confirmation of high accuracy of forecasting (up to 91.3 %) and practical effectiveness of algorithms in real time mode was obtained. It is shown that the implementation of intelligent management allows to reduce the number of unplanned downtimes by more than 25 %, to optimize maintenance schedules, to integrate data processing into SCADA/MES systems of the enterprise. The results of the study prove the feasibility of implementing intelligent monitoring for critical equipment, as well as create prerequisites for the formation of digital doubles and multi-agent control systems within the framework of the concept of Industry 4.0.