APPLICATION OF NEURAL NETWORKS IN MODELING AND OPTIMIZATION OF THERMAL PROCESSES IN MECHATRONIC SYSTEMS

Abstract

The paper discusses the application of artificial neural networks in the modeling and optimization of thermal processes within mechatronic systems, with a particular focus on industrial applications. In modern systems, precise temperature control is crucial for maintaining functionality and extending the lifespan of devices, and neural networks represent an efficient alternative to traditional modeling methods. Through experimental measurements and data analysis under real-world conditions, a numerical model based on a multilayer neural network was developed, which accurately predicts temperature profiles in various parts of the system. Additionally, by optimizing control parameters, energy losses were reduced and system efficiency was improved. The results show that neural networks provide high accuracy in predicting temperature changes even in dynamic operating regimes, with an average prediction error of less than 2°C. Moreover, using this model, the system was optimized to reduce energy consumption, contributing to sustainability and efficiency in industrial applications. This research opens new possibilities for the application of artificial intelligence in the modernization and automation of thermal systems across various engineering disciplines.