Research On The Ice-Storage Air-Conditioning Load Forecasting Based On Wavelet Analysis And Neural Network

The usage of ice-storage air-conditioning system can help to save operating cost. At present, chiller priority is the most common control strategy for existing ice-storage systems, which can not achieve the goal of minimizing the running cost. But only optimal control can significantly reduce the operating cost to the minimum. The accuracy of the load prediction is a key for optimizing the system control. Accurate cooling load forecasting can make the system operated in the way of programming more economic and more scientific, optimal management can be achieved accordingly.This thesis takes the ice-storage system building in Hang Zhou as research example, mainly deals with the BP neural network model and wavelet-BP neural network for cooling load prediction.The prediction results of meteorological parameters can directly influenced the effect of load forecasting firstly. This thesis chooses appropriate methods to predict the outdoor temperature, the humidity and the sun radiation, and these parameters are used as the input of neural network.The neural network can solve some problems which can not be settled by traditional method. In this thesis, the author forecasted the ice-storage air-conditioning load by the BP neural network, satisfied results are obtained.The orthonormal bases can be constructed by translation and dilation of a mother wavelet. Wavelets have local property in time domain and frequency domain; we can extract information from signals using wavelets, so it is called as"math microscope". In this thesis, decomposed cooling load series to low-frequency and high-frequency by wavelets, forecast with BP neural network respectively, synthesize finally. Contrast with the BP network method, the forecasting results by wavelet-BP network is much better than BP network.