header-logo

강화학습 기반냉방 자율제어 모델의 기축 공동주택 상 적용성 평가

주 저자김세헌
공동 저자김태연, 남영도, 정재원
소속-
AbstractThis study proposes a reinforcement learning(RL) model for residential cooling control. The proposed model derives a preferred temperature range from thermostat control data of the occupant. A neural network estimates the preference range, which is then used to define the reward function. The model was evaluated in a chamber experiment simulating real residential conditions by comparing measured thermal sensation and preference votes. Results show that proposed model achieved similar thermal preference level to occupant-controlled case, while reducing energy consumption by 36.15 Wh. These findings highlight the model's ability to optimize the trade-off between comfort and energy efficiency. Future work will be conducted with extended control variables and validation duration in actual apartments.
KeywordReinforcement learning (강화학습), Cooling control (냉방제어), Residential buildings (주거 건물), Thermal preference (열적 선호도), Autonomous control (자율제어)
페이지-
논문 파일 없음
게재일시 2025-06
DOI-
학회/저널명2025 대한설비공학회 하계학술발표대회
년도2025
추가 문구-
등록 일시2025-12-24 15:44:43