| Keywords |
Deep Q-learning; Thermal comfort; PMV; Space cooling; System integrated control; Building energy |
| Abstract |
The objective of this study is to develop an algorithm that integrates the PMV prediction model and deep Q-learning for minimized energy consumption while maintaining thermal comfort in the space. In this study, the proposed performance-based thermal comfort control (PTCC) based on a deep Q-learning algorithm for cooling in a space (or room) developed using a co-simulation was proposed. By comparing the thermal comfort (PMV) and energy consumption resulting from using fixed set-point (rule-based) control and PTCC for cooling, the efficiency of the proposed performance-based thermal comfort control was evaluated. As a results, it was found that PTCC yielded the optimal control action value that minimized the energy consumption while satisfying the thermal comfort conditions.
PTCC reduced the total VRF energy consumption by 32.2% and 12.4% compared with that required by 22°C and 24°C, i.e., the fixed set-point control, respectively. |