Deep Reinforcement Learning for the Heat Transfer Control of Pulsating Impinging Jets

TL;DR

This study applies Deep Reinforcement Learning, specifically variants of Deep Q-Networks, to optimize thermal control of pulsating impinging jets, demonstrating high efficiency and superior temperature regulation performance.

Contribution

It evaluates and compares different DRL variants for thermal control, highlighting the effectiveness of soft Double and Duel DQN in maintaining desired temperature thresholds.

Findings

Soft Double DQN outperforms hard Double DQN in thermal control.

Both soft Double and Duel DQN maintain temperature within thresholds over 98% of the cycle.

DRL shows promising potential for thermal management systems.

Abstract

This research study explores the applicability of Deep Reinforcement Learning (DRL) for thermal control based on Computational Fluid Dynamics. To accomplish that, the forced convection on a hot plate prone to a pulsating cooling jet with variable velocity has been investigated. We begin with evaluating the efficiency and viability of a vanilla Deep Q-Network (DQN) method for thermal control. Subsequently, a comprehensive comparison between different variants of DRL is conducted. Soft Double and Duel DQN achieved better thermal control performance among all the variants due to their efficient learning and action prioritization capabilities. Results demonstrate that the soft Double DQN outperforms the hard Double DQN. Moreover, soft Double and Duel can maintain the temperature in the desired threshold for more than 98% of the control cycle. These findings demonstrate the promising potential of DRL in effectively addressing thermal control systems.