Fuzzy logic based predictive maintenance application in an asynchronous motor

In industrial facilities, systems are expected to operate with high performance and the principle of continuity. A malfunction in any part of the system may cause production losses. The use of various systems in the production and service sectors necessitates maintenance. To avoid production and labor losses due to failure, it is necessary to plan maintenance in a way that does not cause malfunctions in the systems. In this project, a predictive maintenance method, also known as stimulant maintenance, which is increasingly used in planned maintenance processes, was designed. In this context, it has been studied on a system that determines the parts that will need maintenance with the fuzzy logic method before the failure occurs and prepares the maintenance plan accordingly. In this thesis, the vibration, temperature, and current values obtained during the operating conditions of an asynchronous motor were collected from the field with the help of PLC, then written to the database with timestamps, and used both to create test data and to determine membership function values. A fuzzy logic-based inference system was designed using the "MATLAB Fuzzy Toolbox" application based on the data collected from the engine in the field, engine and instrument label values, and expert opinions. The output variable determined as a result of the evaluation of the test data given as the input variable to the inference system with the rules defined in line with the expert opinions and catalog data determines the "motor health" of the asynchronous motor operating in the system. This study proposes an inference system related to the detection rather than the diagnosis of the fault. Enterprises engaged in industrial production intend to increase total equipment efficiency and prevent unplanned stoppages in the first place. Therefore, the probability of a malfunction is more important than the content of the malfunction in the engine. Only motor stator currents, vibration, and thermal-based analysis are available in the literature. In this study, the data obtained by monitoring the current, vibration, and thermal values are evaluated using fuzzy logic from many perspectives to determine motor health and maintenance needs.