AUTOMATIC CONTROL SYSTEM WITH NEURAL NETWORK CONTROLLERS FOR IMPROVING THE QUALITY OF CCCM

Abstract

The article discusses quality assurance issues of carbon-carbon composite materials during their production stage. It emphasizes that achieving the desired quality levels of the manufacturing process depends significantly on the stages of СССМ production technology, starting from the selection and preparation of raw materials and ending with graphitization and surface treatment to improve the necessary special properties. The main characteristics of composite materials, by which СССМ can be evaluated, and key factors ensuring quality in their production are provided. The main organizational aspects playing a key role in ensuring the quality of carbon-carbon composites production are considered. Within the scope of the study, factors of uncertainty typical for the technological process of manufacturing blanks from СССМ were identified and classified. These factors include changes in the properties of initial materials, fluctuations in the surrounding environment, as well as variations in processing and manufacturing processes. The issues regarding the quality assurance of products made from carbon composite material are discussed, which can be improved through the application of the proposed intelligent management system for the technological process of obtaining blanks from СССМ using the thermogradient method with the use of a radially movable pyrolysis zone. The control system has been improved by introducing an additional block of a neural network adaptive controller based on a PID controller, in which the integral part is implemented as a tunable multilayer neural network, with the ability to connect and reconfigure for specific control channel inputs through parameter identification that may affect the behavior of each component. Adding intelligent elements will help capture and formulate deterministic quality indicators for each component or for the entire system, reflecting the necessary quality characteristics. The proposed intelligent system for automatic control and monitoring of parameters in the technological process of manufacturing products from СССМ using neural network algorithms will improve control quality by increasing the system's adaptive capabilities based on the use of macro-information about the dynamic state of the process, aimed at guaranteed quality formation of the СССМ.