生成同时发生设计日的参数降维分区方法研究

Research on parameter dimension reduction and partition method of generating coincident design days

为了减少同时发生气象参数的数据量和便于工程应用,提出了一种生成同时发生设计日的参数降维分区方法,该方法基于辐射时间序列法,确定各房间特征参数的近极端组合.以香港长年逐时气象数据作为数据源,在近极端组合下对各房间特征参数进行降维分区,降维分区后代表组合共计960种.对生成同时发生设计日的参数降维分区方法进行验证,结果表明:利用挑选出的同时发生设计日所确定的设计冷负荷与长年逐时负荷计算后确定的设计冷负荷相比,99%以上的相对误差在±5%内,提高了设计冷负荷计算的准确性.将同时发生设计日气象参数与传统方法得到的设计日气象参数进行了对比,结果表明:传统方法得到的设计日参数被高估,导致空调设计冷负荷偏大.

To reduce the amount of data of coincident meteorological parameters and facilitate the engineering application,a method of parameter dimension reduction and partition was presented to generate coincident design days.Based on the radiation time series method,the near extreme combination of each room’s characteristic parameters was determined.Hong Kong’s longterm hourly meteorological data was as the data source to reduce the dimension and partition for each room’s characteristic parameters under the near-extreme combination,and there were 960 representative combinations after dimension reduction and partition.The method of parameter dimension reduction and partition for generating coincident design days was validated,and results show that the relative error between the design cooling load determined by the selected coincident design days and determined by the long-term hourly load calculation is within±5%in more than 99%cases,improving the accuracy of the design cooling load calculation.The meteorological parameters of the coincident design day were compared with those generated by traditional method,and results show that the meteorological parameters obtained by the traditional method are overestimated,leading to a large design cooling load of air conditioning.