个人空间中结合人员偏好的空调自学习控制方法

An Air-conditioning System Self-learning Control Method Integrated with Occupant Preference in Personal Space

室内环境影响着人们的身心健康和工作效率,而热舒适度是影响室内环境质量的最重要因素,但个体间热舒适存在很大差异,有明显的人员偏好。随着智能时代的到来,人们对空调的个性化和智能化需求也不断增长。因此本文设计并实现了一种适用于个人空间的结合人员偏好的空调自学习控制方法,基于人员历史空调使用行为对人员偏好分类,基于朴素贝叶斯分类器和神经网络模型的人员偏好预测模型来刻画和学习人员的温度偏好和降温/升温速率偏好,将人员偏好分类和预测模型集成到空调控制系统中,结合多阶跃输入的空调控制方法,提供不同人员所需的室内环境条件。在实际运行中,随着人员空调使用行为反馈和模型更新,将持续优化学习结果并且实现更优的控制效果。同时,该方法在现有空调系统的基础上无需增加额外的传感器,利于推广和应用。通过模拟和实验的方式验证了结合人员偏好的空调自学习控制方法的有效性和准确性;通过能耗模拟验证了结合人员偏好的空调自学习控制方法在提高人员舒适度的前提下不会增加许多能耗,甚至可以节省能耗。研究表明,结合人员偏好的空调自学习控制方法能够准确预测人员偏好,提供个性化舒适的室内环境,且具有一定的节能潜力。

The indoor environment affects occupants’ physical and mental health and work efficiency.The thermal comfort is the most important factor affecting the quality of indoor environment,but there are large differences in thermal comfort among individuals.In other words,occupants have significantly different thermal preferences.With the advent of intelligence times,occupants’ demand for personalized and intelligent air conditioners continues to grow.Therefore,an air-conditioning system self-learning control method is designed and implemented to integrate the occupant preference and fit the needs for personal space.A method is first set up to classify occupant preferences based on their historical air-conditioning usage behavior.An occupant preference prediction model based on naive Bayes classifier and neural network model is used to characterize and learn people ’s temperature preferences and cooling/heating rate preferences.The self-learning model of occupant preference is integrated into the existing air-conditioning control system.The multi-step input air-conditioning control method is combined to meet the control needs of different occupant preferences.In actual operation,the learning results will be continually optimized to achieve better control by virtue of the feedback of occupant air-conditioning usage behavior and model updating.At the same time,the method does not require adding extra sensors to the existing air conditioning system,which facilitates its dissemination and application.The effectiveness and accuracy of the air-conditioning system self-learning control method integrated with occupant preference are verified through simulation and experiments.Energy consumption simulation confirms that the air-conditioning system self-learning control method integrated with occupant preference will not increase much energy consumption and even save energy consumption on the premise of improving occupant thermal comfort.Preliminary research results show that the air-conditioning system self-learning control method integrated with occupant preference in personal space can accurately predict occupant preferences,provide a personalized and comfortable indoor environment,and has energy saving potential.