Power Load Forecast Under Intelligent Grid Environment

Smart grid is the development direction of future grid to self-healing, security, power generation resources compatible, power user interaction, electric power market coordination and resource optimize efficiency and power quality quality, information system integration as the main characteristics of the realization of intelligent power grid, without accurate load forecasting technology support. The development of the electric power industry on the one hand it directly restricts the national economic and social development, and the correct power load forecasting that the development of the national economy can provide enough power for power system, also can help the development of itself, especially for the power system planning is concerned, precise load forecast is the whole planning the basis and premise for work. This paper in smart grid environment to load forecast of load signals at the same time, increase the reliability of the forecast results.Introduced the wavelet transform and the neural network, the basic principle of clustering theory and the basic concept of intelligent power grid. The discrete wavelet transform smooth wavelet transform, stationary wavelet transform the redundancy and panning invariability of the time-frequency transform, in the process, to avoid the sampling processing signal distortion. Using wavelet transformation noise reduction processing of load signals, is able to on signal processing while, not classify the trend. Change signal On load signal wavelet decomposition, used statistical methods after using probability theory, the method to weed out poor data. In the load forecast this step, use wavelet clustering of data load classification, then use Elman neural network algorithm forecast.Due to the complexity of the communication network, may introduce the interference of bad data, this paper deal with the problem of load forecast, the influence of bad data adopts the automatic identification method, i.e. used statistical methods to solve the mathematics paper, this is the first innovations.For intelligent power grid, AMI's main function is to provide a intelligence platform, through this platform accurately grasp the characteristics of load node, the main method is to use wavelet clustering algorithms for load classification. Because the purpose of doing that for industrial enterprises and civil point speaking, its power load forecasting is different, the result of power load node to classify and separately predict, will greatly enhance the load forecasting results accuracy and dependability. This is the second innovation points. This method is far superior to expedite the load forecast intelligent master load forecast, and the whole of actual cannot consider different demands of electricity load.After a great deal of simulation results prove the validity and reliability of this algorithm, and the results show that, using WNN to load signals of multilayer signal decomposition, with periodic components to carry on the forecast, the result is accurate and reliable, and the load forecasting process based on intelligent power grid, the senior measure system provides massive data support, the use of wavelet clustering method to load, classification precision load forecasting technology to provide support. This method is AMI load forecasting methods and development direction of one of the feasibility.