At present, modern trends in the development of Smart production based on bioinspired approaches of artificial intelligence are relevant and promising for modern high-tech science-intensive industrial complexes equipped with complex and expensive equipment. Particularly important is the development of intelligent technologies for predictive diagnostics of the state of industrial equipment, which forecast malfunctions before emergency situations occur. Prompt correction of problems that are brewing in production due to possible malfunctions, and timely preventive work with industrial equipment allows preventing many situations leading to technical and financial costs associated with stopping the production line, repairing equipment, eliminating accidents, etc. These studies are devoted to the development of an intelligent system for predictive diagnostics of industrial equipment based on the use of a modified endocrine-immune algorithm (MEIA) for oil and gas facilities. The endocrine component of the algorithm solves the problem of identifying informative features. The algorithm of artificial immune systems acts as a classifier. Two sets of industrial data were used for modeling: the «Motor» database on various operating modes of a DC motor and the «TSHO» database from sensors on the production line of the TengizChevroil oil refinery. The results of modeling and experiments are presented, which showed the advantage of the MEIA algorithm based on the analysis of various metrics on the «TSHO» database.
Samigulina et al. (Thu,) studied this question.