Method of Creating an Automatic Heat Supply Control System Using Simulation Models

Abstract

Ncreasing the intellectualization of heat supply systems is a key task in the development of urban infrastructure. Existing automation approaches often fail to account for the complex interdependence between thermodynamic and hydraulic processes in pipelines, as well as the human factor leading to data noise. The aim of this study is to develop a method for creating an automatic control system for heat supply that utilizes simulation modeling for testing and parameter tuning, considering structural changes and environmental impacts. Materials and Methods. The method is based on representing the heat supply system as a complex structure described using a metric graph. Fuzzy neural networks are proposed for processing hard-to-formalize conditions and noisy data. The physics of processes (single-phase and multi-phase flows) is described by a system of partial differential equations, while the interaction between system elements is modeled based on a modified FitzHugh-Nagumo model. The developed simulation model is implemented in MATLAB using SSPRK and GFORCE numerical methods for optimizing and visualizing system dynamics. To assess the information content of each element, an additive indicator is introduced, calculated using safety factors and Shannon's formula. Results. The authors propose a method for the continuous description of an automatic heat supply control system through state functions, which constitutes the scientific novelty. A system of differential equations is derived, describing the dynamics of control vectors and their interrelations in the state space. A simulation model has been developed, including blocks for physical flow modeling, environmental interaction, and a neural network module. The reliability of the results is ensured by the use of proven mathematical methods (graph theory, partial differential equations) and numerical schemes. The model allows not only for predicting system behavior but also for identifying data anomalies by comparing different information flows.