Research On Short-term Load Forecasting Of County Area

Load forecasting of county area is one of the most important task of themanagement department on the power grid planning, power utilization, scheduling, and itis also the basic work, so it plays a pivotal role to the safe, reliable and economicoperation of power grid. Power load forecasting of country area is related to thereliability of power system planning, as well as the economic interests of the electricpower enterprise. It is an important task of country area’s power system operationsdepartment, and also is the main basis to formulate the power generation planning andtransmission scheme. So, research on the power load forecasting model applied to thecountry area has been the important content in the power load forecasting.This thesis puts forward the necessity of county area’s research on power loadforecasting through the introduction of the research status of power load forecasting athome and abroad. Thus, it puts forward a related theory’s research on power loadforecasting of country area, and takes Shi Jiazhuang Wuji country as an example, then itanalyses respectively from the power load characteristics of country area, loadclassification of country area and the influencing factors of power load in country area,and for the load characteristics of Wuji County, it selects the appropriate power loadforecasting model applied to Wuji County. As Shi Jiazhuang Wuji country is a largetraditional agricultural county, the change of load is mainly reflected on the irrigationload during the irrigation period, and the fluctuation of load is relatively largely. To thelargely fluctuant load, a single model is difficult to meet the accuracy requirements forload forecasting. First, use the fuzzy clustering and distance coefficient load forecastingmodel, and make the maximum temperature, minimum temperature, humidity,precipitation, holidays and other influence factors of the irrigation load as the model’sinput to do the irrigation period’ load forecasting. Moreover, considering the whole trendof the load, it uses the GM (1,1) model. But, as the load fluctuation is largely in theirrigation period, so Markov state transition matrix is used to revise the GM(1，1)model. Fuzzy clustering and distance coefficient model and Grey Markov model arecombined to forecast. Combination forecasting gives full play to the advantages of eachsingle forecasting model and avoids the forecasting risk of each single forecasting model,and it reflects the system change rule comprehensively. The empirical results show thatthe combination of forecasting accuracy has been greatly improved, the combinationforecasting model of county area put forward by this thesis has a certain theoretical andpractical guiding significance to the short-term power load forecasting work of thecountry area.