Application of agent-based approach for heat conduction processes simulation

Abstract

Heat transfer processes are the basis of the most technological processes of the energy sector. Thus, the development of modern approaches for computer simulation and visualization of the phenomena of thermal energy transfer in various objects are of great importance and is relevant. Classical models in the form of differential equations of various types describe processes in continuous space and time. So, it is difficult to apply classical models to study nonlinear phenomena, and processes in inhomogeneous media in the presence of discontinuous solutions at the boundaries. In these cases, simplifying assumptions are used, thus, the adequacy of the models is reduced. It is of great interest to apply fundamentally different approaches to describe transfer processes, which include discrete dynamic models. The purpose of this project is to study the possibilities to apply discrete approaches to simulate nonlinear heat transfer processes under conditions of material inhomogeneity and the presence of volume sources of variable power. The paper studies the possibilities to apply the agent-based approach to simulate models of complex systems. This approach allows us to consider a continuum as a set of interacting elements (agents). The behavior of the elements is completely described by local dependencies. At the same time, the laws of functioning of individual elements are accepted as deterministic and they correspond to the fundamental principles of the theory of heat transfer. The possibility to apply a discrete approach for simulating the process of heat transfer by the molecular mechanism has been studied. The general methodology to develop an agent-based deterministic model is described. Its applicability to describe quasi-linear and nonlinear heat conduction processes is considered. The examples of simulation of combustion processes complicated by exothermic and endothermic effects are considered. The advantages and disadvantages of the proposed method are indicated. The results of the study have shown that discrete agent models are a good alternative to classical continuum approaches to study heat transfer processes in inhomogeneous media. The results obtained do not contradict modern approaches to the description of thermal processes. It has also been found that the simulation algorithms used in the agent-based approach are quite universal and easily adapt to changes under the conditions of problem setting. The analysis of the results makes it possible to recommend a discrete agent-based approach to develop simulation models of complex technological processes and systems.