Technological Aspects of Control Adaptation Based on Neural Network Modeling

Abstract

An innovative technology for adapting control signals of the technological production cycle control system to external control actions and random disturbances is presented. A new solution for reducing the impact of destabilizing environmental factors in real time is proposed. A method for adapting the automated technological cycle control has been developed, based on constructing control feedback algorithms to reduce the sensitivity of process operation parameters to changes in the operating conditions of process equipment and the environment. A description of the tools for implementing the means of adapting the technological cycle control based on constructing control feedback algorithms and synthesizing neuroregulators using neural network algorithms is provided. A procedure for generating a knowledge base in accordance with the "technological processes with probabilistic characteristics" ontology is proposed, providing a formal description of the component classes of the process cycle, their properties and relationships between classes, coding in the SC code format and integration with logical inference mechanisms.