Development of a neural network model for an automated hvac system based on collected data

Abstract

The object of research is ventilation and air conditioning systems, which act as the object of data collection for the development of a neural network model based on them. The main attention is paid to the choice of algorithm, data collection for training a neural network model based on the MATLAB software package, to simplify the model development process. The main problem that was considered in the study is the complexity of building mathematical models for ventilation and air con¬ditioning systems. Traditional approaches require significant computing resources and in-depth analysis of physical processes, which complicates their development and practical use. The research results show one of the approaches to creating a model of ventilation and air conditioning systems using neural net-works. The proposed approach provides fast training of the model based on real data, which in further studies will allow adapting the system to changing operating conditions and increasing its efficiency. The obtained results are explained by the fact that, unlike classical mathematical models that require precise formulation of all dependencies and parameters. Neural networks are able to approximate complex nonlinear functions without the need for a complete understanding of physical processes. The proposed approach can be used for ventilation and air conditioning systems provided that there is a sufficient amount of data for training the neural network. Also important is the integration of such a system with controllers and SCADA systems that provide operational collection of parameters from the environment. The use of neural network models is especially effective in smart buildings, industrial facilities and energy-saving systems, where it is important to optimize energy consumption and provide comfortable condi¬tions for users. In addition, such models can be implemented in cloud platforms for centralized management of climatic parameters in various buildings or production complexes.