| Meteorological parameter is one of the main factors affecting indoor thermal environment and energy consumption of air conditioning system. Several issues are discussed based on meteorological parameters, which are related with the design, operation in HVAC area.Researched and developed the interpolation method based on the existing meteorological data. Analyzed the intrinsic link between the design meteorological data and the geographic information (longitude, latitude, altitude, continentality, etc.) using neural network, and developed a new method of getting design meteorological data using geographic information and neural network. The results show that the new method has unique advantages over the traditional mathematical ones. At the current condition that there aren't enough meteorological data in China, it is a useful addition to standard meteorological data and has a certain application value, providing the basis for the HVAC design with unknown meteorological data,. As the summary and the application of the method, a database program about the HVAC design meteorological data of small towns in China is developedAnalyzed the important of online amendment of the load prediction, and pointed that the ultra-short-time predicting of the solar radiation, ambient temperature and ambient humidity (moisture content) is the basis of online amendment. Then, developed an ultra-short-time predicting model about them with Bayesian BP neural network method, and used it in the research of online amendment of the load prediction.Proposed that correlation analysis and principal component analysis should be used to select of the model predictors and to transform predictors.The results show that both methods effectively improve the generalization ability of established neural network models about the prediction of outdoor meteorological parameters and the prediction of air-conditioning load.In the finality, the problems requiring further studies are discussed.
|
PDF Full Text Request