Intelligent system for monitoring and controlling the technical condition of mechatronic process facilities

Abstract

Introduction . Digital systems that control the maintenance of separate mechatronic process facilities (MPF) and sets of production machines are mainly considered. Numerous issues on maintaining the reliability of the condition and emerging malfunctions, as well as the multifactorial nature of the using the existing monitoring and diagnostic systems, are noted. In this regard, the relevance of the tasks of developing methods of processing equipment maintenance to make decisions under the data veracity and limitation is specified. Materials and Methods . To analyze the criticality of the technical condition, an assessment of the efficiency of the autonomous control of the device state is formed. The method of the neuro-fuzzy system is used to determine the aggregate criterion of criticality. It is proposed to apply this approach to develop recommendations on equipping a production facility with the necessary means of maintaining overall performance and reliability. Results . The solution provides predicting the development of the state of mechatronic process equipment, alerting personnel in case of emergency and other dangerous conditions, and, if necessary, updating or adjusting control programs. Provision is made for performing of some of the technical state maintenance functions by the mechatronic facility itself, i.e., equipment self-service. The concept of “autonomous management of the technical condition” is formulated. The system structure and control functions are considered. It is noted that the implementation of the systems under consideration will significantly increase the efficiency of the equipment use. The performance of the autonomous control of the device or MPF in general is evaluated in accordance with ISO 13381-1: 2004. Based on this standard and the data presented earlier, a neural network structure is built to assess the autonomy of state management. The system training efficiency is considered taking into account the standard deviation of the network outputs from the target values of the training sample. Discussion and Conclusion . A list of the basic control functions at different levels of maintenance autonomy is presented: from alarm for failure prediction to complete maintenance autonomy without the direct involvement of an operator.