基于室内外环境参数的人员开窗行为识别模型  

Recognition Model of Personnel Window Opening Behavior Based on Indoor and Outdoor Environmental Parameters

作　　者：刘咏爽 于丹[1] 崔治国 张亮 曹勇 LIU Yongshuang;YU Dan;CUI Zhiguo;ZHANG Liang;CAO Yong

机构地区：[1]北京建筑大学,北京100044 [2]中国建筑科学研究院有限公司,北京100013

出　　处：《煤气与热力》2023年第4期32-37,共6页Gas & Heat

摘　　要：以寒冷地区某住户为研究对象,采用C4.5决策树算法,提出一种基于室内外环境参数的供暖期人员开窗行为识别模型,选取识别模型的特征变量。结合真实开窗情况,以正确率、准确率作为指标,评价识别模型的识别效果。识别模型的特征变量为室内温度、室内相对湿度、15 min室内温度变化量、60 min室内相对湿度变化量、室外温度、室外相对湿度、壁面逐时太阳入射角。各测试房间识别模型的识别正确率均比较高,范围为89.6%~94.8%。识别模型在准确率方面也有较好表现,范围为84.9%~91.6%。各房间识别模型均具有一定的正确性、准确性。识别结果与真实开窗情况变化趋势基本一致,且吻合程度比较高。C4.5决策树算法在寒冷地区供暖期人员开窗行为识别方面表现良好。Taking a household in a cold area as the research object,using C4.5 decision tree algorithm,a personnel window opening behavior recognition model based on indoor and outdoor environmental parameters during the heating period is proposed,and the characteristic variables of the recognition model are selected.Combined with the real window opening situation,the accuracy rate and accuracy rate are used as indicators to evaluate the recognition effect of the recognition model.The characteristic variables of the recognition model are indoor temperature,indoor relative humidity,indoor temperature change in 15 minutes,indoor relative humidity change in 60 minutes,outdoor temperature,outdoor relative humidity and hourly solar incidence angle on the wall.The recognition accuracy rate of each test room recognition model is relatively high,ranging from 89.6%to 94.8%.The recognition model also has good performance in terms of accuracy,ranging from 84.9%to 91.6%.Each room recognition model has a certain degree of correctness and accuracy.The recognition results are basically consistent with the change trend of the real window opening situation,and the degree of coincidence is relatively high.The C4.5 decision tree algorithm performs well in the recognition of personnel window opening behavior during the heating period in cold areas.

关 键 词：室内外环境参数 开窗行为 识别模型 C4.5决策树算法 

分 类 号：TU832[建筑科学—供热、供燃气、通风及空调工程]