МЕТОДЫ ПРОГНОЗНОГО МОНИТОРИНГА ТЕХНИЧЕСКОГО СОСТОЯНИЯ ЭЛЕКТРИЧЕСКИХ МАШИН

Abstract

Relevance During operation of electromechanical machines, their specifications may change, which can lead to machine failures. Like all electromechanical equipment, rotating machinery is subject to many different adverse effects, such as thermal and environmental stresses and mechanical damage, which require the utmost attention. In industry, any machine or system failure or unplanned downtime can degrade or interrupt a company's core business, potentially resulting in significant fines and immeasurable loss of reputation. Existing traditional approaches to maintenance (maintenance by failure or regulation) suffer from some assumptions and limitations, such as high prevention or repair costs, inadequate or inaccurate mathematical degradation processes. Due to the trend toward smart manufacturing, data mining, and artificial intelligence, predictive maintenance is proposed as a new type of maintenance only after analytical models predict certain failures or degradations. Modern systems for assessing the technical condition of electromechanical equipment are decision support systems based on machine learning. Aim of research The aim of this research is to provide a general overview of maintenance goals and objectives, which mainly include cost minimization, availability/ reliability maximization, and multicriteria optimization. In addition, an overview of existing approaches for fault diagnosis and prediction in predictive maintenance systems is proposed, which include two main subcategories: knowledge-based approaches and traditional machine learning methods. Research methods Currently, many methods based on processing information from measurement transducers, which use acoustic and vibration sensors, current and voltage sensors, temperature sensors, and electromagnetic transducers, using machine learning techniques, have been developed for fault diagnosis and failure time prediction of electromechanical units. Results As a result, the paper presents a brief overview of the aims and approaches of a predictive maintenance system for electromechanical systems based on machine learning techniques used in various electromechanical equipment. An overview of various predictive maintenance methods is presented. An overview of existing approaches based on machine learning is given.