Abstract
The research is devoted to the current problems of introducing artificial intelligence into industrial production in order to reduce emergency emissions into the atmosphere and create an environmentally friendly production environment within the framework of the implementation of the Smart manufacturing concept. Bioinspired approaches to industrial artificial intelligence and the features of their implementation on modern industrial equipment are considered. The developed intelligent multi-agent system MASBio is proposed based on neuroimmune-endocrine interaction for the analysis of multidimensional production data, production process control, equipment diagnostics and emergency prediction in order to prevent negative impact on the environment. The architecture of a multi-agent system is presented taking into account the real technological process of gas purification from acidic components of the Tengizchevroil oil refinery, as well as the specification of agents and the functions they perform. The MASBio system was tested at the Industrial Automation Lab (Schneider Electric) of the Kazakh-British Technical University. The intelligent MASBio system can work with industrial controllers of the PAC and EPAC class, as well as SCADA systems. Modeling and experiments on production data showed better results compared to other machine learning methods.