大型商业建筑空调负荷预测方法的研究

发布时间：2018-01-28 01:48

本文关键词： 商业建筑负荷预测室内扰量正交试验多元线性回归季节性指数平滑　出处：《天津大学》2014年硕士论文　论文类型：学位论文

【摘要】：针对商业建筑暖通空调系统能耗高的特点,为优化空调系统运行策略,把大型商业建筑的空调负荷预测作为研究对象,分析了大型商业建筑空调负荷显著性影响因素,建立基于气象预报和室内人员密度的空调负荷逐时预测模型,用于预测下一日空调逐时负荷,作为优化空调系统运行策略和确定蓄能空调蓄能量的依据。以天津市为例,首先调研大量规范文献和30座实体商业建筑,建立大型商业建筑特征模型。基于特征模型,采用正交实验,对空调负荷的显著性影响因素进行了研究,确定了室内人员密度为商业建筑空调负荷的显著性影响因素。其次,分析了室外气象参数与室内人员密度对空调负荷的延迟影响效果,初步确定了合适的室外气象参数以及室内人员密度作为预测模型的输入变量。之后,利用逐步回归法,排除空调负荷不显著影响因素,减少预测模型的输入变量,简化预测模型。最后,对几种预测方法的适用性及预测精度进行了计算分析,结果表明:带反馈的多元线性回归模型预测空调逐时冷负荷取得较高的预测精度;改善的温特季节指数平滑模型对于波动较大的空调逐时热负荷取得较高的预测精度。通过本文研究,发现人员密度为商业建筑空调冷负荷和热负荷的最显著性影响因素,新风量及室内照明和设备功率密度对空调冷热负荷的影响作用也比较大,为显著或接近显著性影响。把室内人员密度作为自变量,这样做提高了商场建筑空调负荷预测精度。将本文提出的负荷预测方法应用于天津市某商业建筑A的负荷预测,预测结果表明:利用带反馈的多元线性回归模型预测逐时空调冷负荷取得了较高的预测精度,其平均相对误差MRE为7.90%,误差平方和RMSE为316.8;最大逐日相对误差maxRE为5.14%,minRE为0.108%。带输入的温特季节性指数平滑模型预测逐时空调热负荷,其MRE为10.05%,RMSE为262.0;逐日热负荷相对误差最大值maxRE为14.5%,最小为minRE为0.506%,冷热负荷均取得较好的预测效果。综上证明:本文构建的基于气象预报和室内人员密度的负荷预测模型的方法可行。
[Abstract]:Aiming at the characteristics of high energy consumption of HVAC system in commercial buildings, in order to optimize the operation strategy of HVAC system, the air conditioning load forecasting of large commercial buildings is taken as the research object. This paper analyzes the influential factors of air conditioning load in large commercial buildings, and establishes an hourly forecasting model of air conditioning load based on meteorological forecast and indoor personnel density, which can be used to forecast the hourly load of air conditioning in 1st. As a basis for optimizing the operation strategy of air conditioning system and determining the storage energy of energy storage air conditioning system. Taking Tianjin as an example, a large number of normative documents and 30 commercial buildings were investigated. Based on the feature model of large commercial buildings, orthogonal experiment was used to study the significant influence factors of air conditioning load. The indoor personnel density is the significant influence factor of air conditioning load in commercial buildings. Secondly, the effect of outdoor meteorological parameters and indoor personnel density on the delay of air conditioning load is analyzed. The suitable outdoor meteorological parameters and indoor personnel density were preliminarily determined as the input variables of the prediction model. After that, the air conditioning load was excluded by stepwise regression method. Reduce the input variables of the prediction model, simplify the prediction model. Finally, the applicability and accuracy of several forecasting methods are calculated and analyzed. The results show that the multivariate linear regression model with feedback can predict the hourly cooling load of air conditioning with high accuracy. The improved Winter seasonal Index smoothing model can achieve high prediction accuracy for hourly heat load of air conditioning with large fluctuation. It is found that the density of personnel is the most significant factor affecting the cooling load and heat load of air conditioning in commercial buildings, and the influence of fresh air volume, indoor lighting and equipment power density on the cooling and heat load of air conditioning is also relatively large. The density of indoor personnel was regarded as independent variable for significant or near significant effect. In this way, the precision of air conditioning load forecasting in shopping mall is improved. The load forecasting method proposed in this paper is applied to A load forecasting of a commercial building in Tianjin. The prediction results show that the prediction accuracy of hourly air conditioning cooling load is higher with the multivariate linear regression model with feedback, and the average relative error MRE is 7.90%. The sum of square error (RMSE) is 316.8; The maximum daily relative error (maxRE) is 5.14 and RE is 0.108. The temperature seasonal index smoothing model with input is used to predict the hourly air conditioning heat load, and its MRE is 10.05%. RMSE was 262.0; The maximum relative error of daily heat load is 14. 5% (maxRE) and the minimum is 0.506% (minRE). It is proved that the method based on meteorological forecast and indoor density forecasting model is feasible.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TU831.2

【相似文献】