Abstract
The paper examines the relevant issue of using neural network diagnostics as an auxiliary element of a set of measures to detect failures of the asynchronous electric drive of a vibratory jaw crusher used for crushing mineral fertilizers. A distinctive feature of this equipment is its ability to generate powerful vibration loads necessary for the effective destruction of the material, while maintaining the reliability and simplicity of the design. According to statistical data, up to 95 % of asynchronous motors installed in vibratory jaw crushers for crushing mineral fertilizers are machines with a power output ranging from 75 to 250 kW. These motors are most susceptible to the effects of aggressive environments, humidity, temperature fluctuations and increased vibration loads, which primarily leads to the failure of the mechanical parts of the equipment and the stator winding bars. The authors demonstrate the main problems of applying the method of neural network diagnostics of equipment and develop an algorithm for predicting failures of the actuator, namely, rod breakage in the phases of the stator of an asynchronous motor that operates as a drive for a vibrating jaw crusher of mineral fertilizers. Moreover, the research presents the assessment methodology with an expanded algorithm that allows eliminating a number of typical failures for this equipment, thereby increasing the efficiency of using an asynchronous electric drive in this industry. The research produced guidelines for operating and configuring the neural network algorithm designed to detect broken stator bars in an asynchronous motor.