| The dissertation deals with the study of load forecasting based on time series analysis and its application in heat supply system. During winter heating period, heating supply has some randomicity. The running of traditional heating system is only based on the weather situation. The yardman adjusts the parameter of district heat supply network usually according to the weather situation and their working experience, the accuracy of district heat supply network is inferior. Depending of the method of time series analysis and using the collected real time data of heating system to forecast the load of the next sampling period, and serving for optimizing heating system operation will take important effect on the saving operation and real-time control of heating system.The dissertation adopts the original data collected in heating supply station and makes data pre-processing. The certainty part of load time series can be estimated by OLS method, and the random signal is separated. After making the signal stable, the signal can be analyzed by time series methods.By using Yule-Walker method and Dynamic Data System modeling method, this dissertation forecasts the heating load transversely and vertically, and optimizes forecasting results by crossover algorithm. Theoretical and simulation study indicate the applying of crossover algorithm is significant for improving the forecasting precision, and the result of using DDS is better than Yule-Walker.Load forecasting method based on time series analysis to the monitoring and controlling is applied in software of heating supply system. By using the database of original data and forecasting data, depending on the display interface, the adjustor of heating supply station can easily use the software to monitor and control the heating supply system, and instruct the energy saving operation of heating supply system. It is significant for improving management level, control level of heating supply process, and saving energy.
|
PDF Full Text Request