Research On Spatial Load Forecasting Considering Urban Electricity Consumption Structure And Land Division

Spatial load forecasting is one of the basic tasks of power system planning.It is the premise of automatic power generation control and economic dispatch control.As the continuous construction of Global Energy Interconnection,our country has a strong demand for electricity.Power system management mode needs to be changed from the extensive type to fine type.The traditional load forecasting methods and technology is gradually out of step with the latest requirements of power system planning,so the spatial load forecasting has becoming a current research hot spot.This study analyzed the spatio-temporal distribution characteristics of the spatial load historical data from Shanghai Power Grid in 2004-2014,and the industry data from Shanghai Statistical Yearbook 2004-2017.According to the characteristics of different areas,the paper respectively established a rapid estimation model on Shanghai spatial load forecasting,which can predict the 2015-2017 load data of each area.Therefore,it provides a scientific basis for urban spatial load forecasting of power system research.First of all,this study carefully combed and summarized the characteristics of the intelligent methods at home and abroad,and selected the intelligent method for spatial load forecasting,including regression analysis,feature selection and extraction technology,clustering analysis,artificial neural network.Then introduced the principle of the algorithm and implementation steps in detail,and compared the advantages of these analyses methods,which used in power system and spatial load forecasting.Secondly,this study analyzed and summarized the research results,which domestic and foreign scholars proposed on power load influencing factors,and proposed the index system of power load forecasting.According to the analyzing of Shanghai electricity consumption structure in 2007-2017 and "standard of urban land classification and planning construction land",this study selected 19 indexes as the indicators factors of spatial load forecasting.Therefore,the rules and characteristics of Shanghai power grid can be obtained.The proportion of primary industry is very small.Shanghai’s secondary and tertiary industries have a greater impact on power load.Industry,urban life,finance,real estate,business and resident services,public utilities and management organizations,business,accommodation industry and catering industry accounted for 96.77% of the total load in Shanghai,so these kinds should be seriously considered in load forecasting.The three largest districts that consumed electricity in Shanghai are Pudong 1,Jinshan 8 and Qingpu 7.Pudong 1 not only ranked first in the total load,but also maintained a high growth rate.Songjiang 5 area shows a trend of continuous growth,and the future load also has a large growth space.The load of the old districts of Shanghai,including Shiqu 2,Shibei 3 and Shinan 4,has tended to be stable since 2009,with a small change range.This study divided the historical load data of Shanghai power grid into five types.The first type mainly contained the outlying suburbs of Shanghai,so its regional urbanization level and economic lever is relatively low.Thirdly,this study optimized the impact factor(X1 ~ X19)by dimension reduction and noise reduction.Then the load data were divided into five types by the principal component analysis,system cluster,and the rule and analyzing of Shanghai power grid.The first type mainly contained the outlying suburbs of Shanghai,so its regional urbanization level amd economic lever is relatively low.The second type contained Shibei,which includes Baogang and Shibei industrial park with high industrialization level.The third type is the traditional old city of Shanghai,where the economy is very developed and the industry is less.The fourth and fifth type are the two stages of Pudong.Influenced by the development and opening up of Pudong and Shanghai Free Trade Zone,this area enjoyed rapid growth in population,economy,industry and other aspects.Its load accounts for the largest proportion of all the districts in Shanghai and shows a rapid growth trend.Fourthly,five types of load data were modeled separately.According to the characteristics of the data,the first and second type used artificial neural network methods.The third,fourth and fifth type used multiple regression analysis methods to verify the results.The models forecasting the load data of each district from 2015 to 2017.To sum up,this study analyzed the electricityc omposition of Shanghai in 2007-2017,mined the several main factors that affect Shanghai load,and analyzed its spatio-temporal distribution characteristis.Then this study proposed a new method and model on spatial load forecasting,which can provide reliable support for Shanghai power grid planning.