Multi-factor And Multi-dimensional Power Demand Forecasting Method And Its Application Research

A safe and stable power grid environment is an important guarantee for the rapid development of the social economy.Scientific and effective power demand forecasting is an important basis for power marketing,power system planning,power market analysis,and power company decision-making.There are many researchers at home and abroad dedicated to the study of power demand.After years of research,there have been many methods and theoretical foundations for power demand forecasting.In the process of power demand forecasting,it has always been to find and determine the factors and mechanisms that affect power demand.It is a hot issue that needs to be studied,and with economic development and technological progress,the influencing factors also change accordingly.Many external environments and internal factors can affect power demand,making it more difficult to predict power demand in complex environments.Based on this,based on the power demand forecasting under the background of economic transformation,energy transformation and electricity market-oriented reform,this paper has carried out the following research work:(1)based on the analysis of short-term and long-term power demand influencing factors.this paper uses econometric tools to explore the correlation between economic development factors and power consumption of some provinces and cities in China,and constructs the power demand forecasting for the following paper The selection of influencing factors in the model provides the basis;(2)from the perspective of power load related characteristics,considering many influencing factors such as economic operation index and monthly meteorological index,a short-term(monthly)power demand prediction model based on SSA-LSSVM is constructed.(3)Combined with the idea of per capita electricity consumption level method and output value unit consumption method,using the analysis method of Kaya identity and scenario analysis method,a long-term electricity demand prediction model considering multiple factors based on Kaya identity is constructed.(4)Taking a province as a research sample,by collecting relevant historical data and future development scenarios,using the multi factor and multi-dimensional power demand forecasting model constructed in this paper,the short-term and long-term power demand forecasting are carried out respectively,and the scientificity and effectiveness of the model and method are verified.The results show that:(1)economic and social development factors:GDP,industrial structure,urbanization rate and power consumption have a strong correlation,which is the key factor affecting power demand;(2)according to the prediction results of a province’s long-term power demand,the first generation power intensity develops according to the historical trend,the second generation power intensity declines rapidly,and the third generation power intensity declines Under the situation that the elasticity of per capita income per capita GDP and the elasticity of per capita electricity consumption per capita income of urban and rural residents are accelerating,the power demand of a province can reach the future planning of the province.To sum up,the research results of this paper have certain reference value for the formulation of relevant power development policies,and can provide more scientific and reasonable suggestions for energy power demand prediction and power planning.The main innovations of this paper are as follows:(1)Combined with the actual situation in China,this paper analyzes the main factors that affect the power demand forecast from the short-term and medium-term dimensions respectively,clarifies the internal relationship between the power demand and the influencing factors,and lays the foundation for building a multi factor and multi-dimensional power demand forecast model;(2)from the perspective of economic development and power load related characteristics In this paper,the X-12-ARIMA method is used to decompose the data into trend term,cycle term and fluctuation term,and the grey correlation analysis method and SSA-LSSVM method are used to build the monthly load forecasting model based on the main operation indexes of macro economy and meteorological factors.It provides an effective tool for short-term power demand forecasting.(3)Scenario analysis is introduced into the established Kaya long-term power demand forecasting model,and different development scenarios of each influencing factor are set up,and then the forecast is carried out in different scenarios,which improves the reliability and scientificity of the forecast results.