The Research on Temperature Control and Microstructure Optimization of Regenerative Heat Treatment Furnace Based on Deep Reinforcement Learning

This research explores temperature regulation and microstructure design optimization in regenerative heat treatment furnaces by deep reinforcement learning (DRL). The research aims to enhance the heat treatment process efficiency and accuracy and retain the necessary material properties. By integrating DRL with the simulation of material properties, heat treatment technology optimization, and high-temperature furnace behavior modeling, this paper presents a novel real-time temperature control and process optimization algorithm. The synergy between microstructure and temperature control achieved by DRL is advantageous from both product quality and process efficiency perspectives. With this approach, operation performance is not only optimized but a window is opened for adaptive, automatic, and intelligent heat treatment systems. Sub-sequent work will focus on the optimization of the DRL algorithm, with the incorporation of high-sensitivity sensing technology and deeper investigation of the temperature control-microstructure relationship to achieve even more precise material design. The findings are significant in their theoretical support for heat treatment technologies, and thus offer great instruction for technological innovation and industrial transformation in the subject industries.