The development of sensor technology provides massive data for data-driven fault diagnosis. In recent years, more and more scholars are studying artificial intelligence technology to solve the bottleneck in fault diagnosis. Compared with other classification and prediction problems, fault diagnosis often faces the problem of data scarcity. To overcome the lack of fault data, the transfer learning based on different working condition is gradually introduced into fault diagnosis by scholars. This paper discusses the current mainstream AI-based fault diagnosis methods, and analyzes the advantage of transfer learning for fault diagnosis problem. Then, a transfer component analysis (TCA) based method is proposed to transfer data features between different working conditions. Through the TCA-based method, the fault diagnosis model under the working condition can be established with the help of historical working condition. It effectively alleviates the problem of data scarcity under the condition to be predicted. Different from other fault diagnosis studies, this paper considers the online maintenance process based on TCA. A fault diagnosis framework including online maintenance process is proposed. Finally, a case study of bearing diagnosis from Case Western Reserve University proves the feasibility and effectiveness of the proposed TCA-based method and our fault diagnosis framework.

/pdf/transfer-learning-based-data-feature-transfer-for-fault-j1sxx9nhm6.pdf

Transfer Learning Based Data Feature Transfer for Fault Diagnosis

This article is devoted to the problems of development and implementation of digital production technologies at industrial enterprises. The structure of the technological processes and principal approaches to automation of production enterprises with various technological processes is considered. An analysis of the conditions of the change in the enterprises to multiproduct small-scale manufacturing is carried out. The problems of changing modern production enterprises to the Industry 4.0 concept are specified. The urgent problems are stated, solving of which is required to develop digital machine-building production in the Industry 4.0 concept. A definition of the digital production is proposed.

Development of Digital Machine-Building Production in the Industry 4.0 Concept

A set of methods for estimating the parameters of the production process is proposed. This set includes expert assessments, Pareto analysis and elements of neural networks based artificial intelligence. The set forms up the methodology for assessing the state of the production process and forecasting its implementation. The sequence of application of individual methods is substantiated. Algorithms for constructing informal decision are developed. Decisions are based on expert assessments and Pareto analysis for the formation of a training sample of a neural network. Approbation of the proposed methodology performed by means of simulation.

Methodology for assessing the implementation of the production process

An imitating model for monitoring the work of a workshop that executes several technological processes was developed. Model was developed with usage of probability theory methods, mathematical statistics, and the theory of random processes. The model visualizes a certain set of states of the executed operations belonging to the technological process and allows user to track the percentage of failure for each machine and of the entire system as a whole at each stage of the technological process. Approbation of the model was carried out by calculating the estimated probability of technological process completion for the blanking mechanical engineering production of FSUE NPO Tekhnomash.

Simulation model for monitoring the execution of technological processes

The digital twin (DT) of a process is a prospective tool for designing modern digital productions based on the Industry 4.0 concept. The development of technological process (TP) DT provides a large amount of data acquisition connected with product processing into a single information structure. The structure may contain the values of different technological parameters and the workshop's utilized resources data. From this point of view, the monitoring of machine-building TP can be implemented and based on the DT process. The development of this approach was the purpose of the research. The scientific publication analysis showed the approach under consideration is relevant and insufficiently developed from a scientific point of view. The authors propose the structure of TP DT for high variety low volume machine-building production in this paper. The possibility of its application for the majority of these types of production is proved. An algorithm for solving the TP monitoring task based on the DT process is described.

Application of the Digital Twin Concept to Solve the Monitoring Task of Machine-Building Technological Process

The article examines the development of digital manufacturing production, in particular the system for monitoring the work of CNC equipment, CNC equipment circuits and other automated manufacturing equipment, as well as systems of monitoring and resource suppy for the implementation of production tasks and production plan of the digital manufacturing production. The developed system has the following features: monitoring of the main parameters of the equipment and its components; assessment of the current situation and the consequences of equipment failure; assessment of resource supply for production orders and the production plan, and tracing their completion; reporting on the formation of diagnosed objects. The solution is based on using the equipment condition index to ensure objective assessment of the equipment condition. The authors developed software and circuit solutions to be implemented in real digital production.

A. Kushnir

Papers

Development of digital production engineering monitoring system based on equipment state index