Optimization of Tubular Gas Heaters on Pellets with Recirculation of Gas-Air Mixture Using a Quasi-Two-Dimensional Model

The article is devoted to decision-making in the control and design of autonomous heat supply systems with tubular gas heaters. The results of mathematical modelling and optimization of tubular gas heaters (TGN) are known. Tubular gas heaters are an extension of the term “infrared tubular gas heaters”. The main elements are: a gas burner, a tubular heater inside which gas combustion products with air move, and a mechanical fan (supply or exhaust), which ensures the movement of the coolant inside the tubular part and its removal outside. There are a number of new technical solutions that expand the scope of tubular gas heaters, for example, tubular gas heaters on pellets. Mathematical models of tubular heaters on pellets and solutions to the problems of optimal design of tubular heaters of linear structure are known. Another possible structure of tubular gas heaters is heaters with recirculation of the heating gas-air medium. Optimisation of such pellet heaters has not been performed before, which determined the subject of this paper. The article is devoted to the presentation of the method of optimization of the design solution for tubular heaters taking into account recirculation under the existing constraints. The novelty of the optimization lies in the use of a quasi-two-dimensional mathematical model for the hydraulic circuit of the heater. An evolutionary search algorithm with binary choice functions is used for numerical search of solutions, for which convergence with probability 1 to the optimal solution is shown. The algorithm contains two consecutive functions: the function of solution generation and the function of solution selection. The function of solution generation is built largely independently of the content of the problem to be solved, while the function of selection is built in such a way that the resulting binary selection relation is completely determined by the requirement of finding the necessary solution. The resulting binary selection relation includes both the selection components of the available constraints and the basic optimiztion requirement.