Research On The Forecasting Of The Monthly Electricity Sales Based On The Combination Method

Electricity is the main product of the electric power grid enterprise. And the monthly electricity sales, which is the basic data to calculate the electricity sales price, sales profits, loss management and a series of indicators, is an important economic indicator of the operation and management in electric power grid enterprise. Changes and development in electricity sales are influenced by many uncertain factors, such as the development of national economy, weather, electricity price policy, the characteristics of grid equipment, the habit of using energy, etc. So it’s very difficult to forecasting the electricity sales accurately.Therefore, forecasting the monthly electricity sales correctly plays an important role in making business decision, and providing an important marketing decision support within the jurisdiction. Forecasting the monthly electricity sales, which can make power plants and transmission and distribution grid running reasonably, has great significance to promote the development and construction of the electricity market.Based on the load characteristics analysis of the historical dada of monthly electricity sales in Dalian, analyze the rules and characteristics of electricity sales growth in Dalian. Using a combination forecasting method to predict monthly electricity sales in Dalian for a month, find a new method to improve forecasting accuracy. Before forecasting, the monthly electricity sales is divided into four categories, the first industrial electricity sales, the second industrial electricity sales, the third industrial electricity sales and residential electricity sales. The sum of the four categories of electricity sales predictive value is namely the electricity sales in that month. For forecasting the electricity sales at a month, firstly, according to the electricity sales history data of the corresponding month of the previous five years, by using gray theory method, linear regression methods and cubic fitting method, we can obtain three monthly electricity sales predictive value. Secondly, analyzing the degree of the error between the true value and the three predictive value, proper weight distribution can result in the further calculating a new data as the final monthly electricity sales. Taking into account the effect of the Chinese New Year on electricity sales, we forecast electricity sales in first quarter as a whole, using combination forecasting method to predict the ratio between the electricity sales in January and February and the quarter electricity sales. The monthly electricity sales can be forecasted if we can calculate the electricity sales in first quarter and the ratio. Forecast the monthly electricity sales in2013in Dalian based on the electricity sales history data from2008to2012. By using combination forecasting method, the minimum relative error is0.42%, and1.93%for maximum. Compared with other three methods, the prediction accuracy has been greatly improved. So it has a certain forecasting accuracy and reliability using combination forecasting method to predict the monthly electricity sales generated from currently existing load characteristics in Dalian.