基于人工智能技术预测热感觉的室内热环境控制

INDOOR THERMAL ENVIRONMENT CONTROL BASE ON THERMAL SENSATION PREDICTED BY ARTIFICIAL INTELLIGENCE

提出一种基于人工智能技术预测用户的热感觉、根据用户热感觉进行室内热环境控制的方法。该控制系统中使用计算机视觉技术识别用户皮肤温度,采用机器学习方法来确定用户的皮肤温度舒适范围,并且使用模糊控制逻辑来预测用户的热状态并调整室温设定值。为了验证该控制方法的有效性,搭建了试验台并进行了一系列试验,试验结果表明用户满意率较高,满意度评价达到97%,随着机器学习的进展,用户的舒适度得到提高,解决了传统控制存在的既不舒适,又不节能现象的问题。

This article proposes a new thermal environment control method using the room occupants’ thermal sensations predicted by artificial intelligence. The control method uses computer vision technology to measure occupants’ skin temperature and uses machine learning to identify the comfort range of skin temperature. And fuzzy logic is used to predict the thermal sensation states of room occupants and the determine the optimal room temperature settings. In order to verify the effectiveness of the proposed control system, test bed was built at an ordinary office room and experiments were conducted. The experiment results show that the room occupants are very satisfied with the thermal environment with a satisfaction ratio of 97%. The hot/cold complaint times dramatically decreased accompanying to the progress of machine learning, which shows the proposed control system can provide comfortable indoor thermal environment and effectively solve the problem of uncomfortable thermal environment and energy waste caused by the present set-point based control.