Data-driven machinery faults detection techniques using Artificial Intelligence in Industry 4.0 concept

Abstract

The research is devoted to the development of an intelligent technology for diagnosing industrial equipment of oil and gas facilities based on an improved FMEA methodology (Analysis of Modes, Failures of their Influence, Degree of Criticality) in combination with a unified artificial immune system (UIIS) and the principles of immunological homeostasis. The main trends in the development of bioinspired artificial intelligence technologies are considered. A unified artificial immune system is built on the basis of modified algorithms of the artificial immune system (AIS) in order to identify the most effective ones (in data processing and forecasting) for a certain set of production data. The application of the principles of immunological homeostasis to assess modified algorithms allows identifying the «homeostasis area» in which the algorithms have the best predictive properties and can form an adequate immune response. The extension of the FMEA methodology with an intelligent block based on UAIS allows to automate the information processing previously carried out manually by experts, reduce time and resources when diagnosing equipment, and eliminate errors associated with the «human factor». The technology has been approbated on real data on equipment failures at TengizChevroil company (oil and gas industry) and on experimental data on equipment from Schneider Electric (Industrial Automation Lab).