Based On The Big Data Of Power Distribution Industry Classification And Power Consumption Demand Forecast Modeling Analysis

The core technology and application of smart grid is the deep combination of power energy flow and information communication technology.Along with the high-speed development of information technology,the construction cost is greatly reduced,which caused a power base in grid database data explosive growth,the data contains the huge value,at present,both at home and abroad through the technology of data mining was carried out by different levels of intelligence with the electric field data research,expect to build economic and reliable support the smart grid.Based on the data of the electricity industry with TV university classification and analysis of electricity demand forecasting model,mainly by the power of the mining and analysis of regional electric power customer model control group composition and its characteristics of electricity utilization,identify the key factors influencing the power consumption,to forecast the electricity consumption of different industries,so as to realize the fine management of customers,provide quality service to the electricity.The accurate classification of the electricity industry is an important basis for realizing the customer's electricity pattern recognition.But because of various factors influencing the sensitivity of the different,lead to the business scope of the same enterprise electricity model may also be different,so the traditional industry classification method cannot effectively distinguish between different power mode.By big data mining method in the second chapter to the dongguan industry power consumption data from 2008 to 2015 were analyzed,and the selected power consumption accounts for 30 higher electricity industry,and then take the k-means clustering analysis method,based on the two standards to select key electricity industry: first,the industry of electricity consumption in the whole society;the proportion in the total power consumption Second,the fluctuation of electricity consumption in this kind of industry has a great influence on the fluctuation of total electricity consumption in the whole society.Finally,the 10 major industries with the highest electricity consumption ratio were obtained,and the electricity consumption of these industries reached about 80% of the electricity consumption of the whole society.In the third chapter,the over-classification scale is used to find the prediction model with high precision and good fitting.In contrast prediction model based on the results of the fitting and prediction of screening,found that the predicted effect of different models for different time periods,such as affected by seasonal fluctuations of data not fitting on the grey system model;And sample amount is insufficient and poor quality data will lead to the neural network model can't achieve satisfactory accuracy,when the budget for the industry power consumption in data so the Box-Jenkins model(ARIMA)is more appropriate,for the annual power consumption data or the type of error and fluctuation point usually is not caused by the characteristics of power industry itself,so using the grey system model for the forecast of annual data error is smaller.Finally,relevant economic variables are tried to be added.By substituting economic variables into the model,it can be seen that with the gradual addition of economic variables related to electricity consumption,the prediction accuracy can be significantly improved.