Short Term Load Forecasting Based On Improved Neural Network Algorithm For Smart Grid

The realization of smart grid has brought profound changes and development to the traditional power grid. Smart grid provides a direction for the establishment of the self healing power grid structure with saving, no pollution,safety and reliability. The research and utilization of smart grid has become the development strategy of European and American countries and China, and many countries have made some achievements in the field of smart grid research.Power system short term load forecasting is the basis of power system dispatching department to make power generation plan, and it is the basis of scheduling scheduling, power supply plan and trading plan in market environment. Short term load forecasting refers to the load forecast of the week,day, hour and so on. According to the change of the load curve, the daily power generation plan of each power plant can be put forward. So accurate load forecasting is beneficial to improve the economy and reliability of power system operation.Wind energy, solar energy these renewable energy grid connected power generation, is an important part of the structure of smart grid. In the case of smart grid, the user’s consumption pattern will be significantly changed, this significant change is that the user can adjust their consumption patterns based on real-time electricity price according to the demand of electric energy. In order to improve the accuracy of short-term load forecasting in smart grid environment, this paper proposes a real-time pricing based on principal component points(principal component analysis) genetic algorithm optimizing neural network short-term predictions. First of all, in the study of load forecasting, the main model isestablished, this paper refers to many factors such as holidays, temperature and so on to build a model. Secondly, the principal component analysis is used to reduce the dimension of the principle, when the dimension of the matrix is reduced so that it can reduce the amount of computation, and reduce the dimension of the matrix also contains the information of the original matrix.Again, because the neural network algorithm is easy to fall into the local minimum in the process of operation, so that the use of genetic algorithm to optimize it, remove the shortcomings of its. At last, using the high nonlinearity of neural network and genetic algorithm, the optimization of neural network and the principle of PCA reduction dimension are used to get the final prediction result through Matlab simulation. Experimental results show that the proposed method has high accuracy and good adaptability to load forecasting.