The Research On Short-term Load Forecasting In Fujian Power Grid Under Special Weather

Power system short-term load forecasting is the base of power generation planning and grid security check analysis,the forecast level will directly related to the economy,stability and security for power system operation.The study on short-term load forecasting method has matured through the long-term efforts of domestic and foreign experts.However,when the weather conditions change,the load level will follow the fluctuations occur,and the load forecasting accuracy rate still can not meet the requirements.This thesis started with weather factors and grasped the impact of load change mechanism,focused on the research of load forecasting under special weather conditions such as high temperature,ranfall,typhoon,and aimed to increase the load forecasting accuracy when weather conditions change.The main research outcomes are as follows.The significance of short-term load forecasting and its research status are generalized,and the difficulties for short-term load forecasting of Fujian in practice are also summarized in the thesis.Then it deeply analyzed the load characteristics of Fujian power system,and briefly described the factors that lead to chang load characteristics including weather and non-weather factors.Additionally,something about data pre-processing and error analysis are presented.Temperature-lowering load of Fujian power grid is analyzed.Firstly,the temperature-lowering load is separated from the total load by reference load comparison method.Then,the thought of holistic analysis and subnet-accumulation are both applied to temperature-lowering load analysis for the grate difference of economic level and meteorological conditions in various cities in Fujian Province.It deeply studied in the aspects of the characteristics of temperature-lowering load,temperature sensitivity and annual maximum temperature-lowering load.Lastly,it showed that subnet-accumulation is better than holistic analysis.A short-term load forecasting method considering the impact of small hydropower is proposed.Firstly,the characteristics of small hydropower load and its influencing factors are analyzed,and a small hydropower load forecasting method based on decomposition and reduction of load curve is introduced.Nextly,it put forward a way to construct large hydropower virtual power plant with a unit of watershed.And then the predictions of each virtual power plant can be obtained by aforementioned small hydropower load forecasting method.So,the load curve of small hydropower generation can be obtained by accumulating.Finally,it proposed a short-term load forecasting method considering the influence of small hydropower:the load curve of network supply can be restored by the load curve of small hydropower and the load curve of whole province.A short-term load forecasting method during typhoon is proposed.This paper analyzed the typhoon activity patterns that affect Fujian power grid,and described the features of a typical load characteristics change during typhoon.In the face of available similar day during typhoon,a forecasting method based on load modification of similar day is developed.Without available similar day,this paper proposed a forecasting method based on load correction of reference day.Examples show that the results of this paper can improve the accuracy of short-term load forecast in Fujian power grid.