Research on the Daily Gas Load Forecasting Method Based on Support Vector Machine

A daily gas load forecasting method based on Support Vector Machine theory is developed in this paper. Some aspects of data preprocessing are discussed, such as normalization method, data grouping method and the period of history data using as input vector. Proper normalization method, which is to map gas load data from the small and narrow range to the big and wide, will improve the forecasting accuracy. The feature of grouped gas load plays an important role to the model effectiveness, because of the different consumer compositions corresponding to different data groups. The principle to define a proper period of history data which are used as input vector is relevant to the number of training samples and the characteristic of the nonlinear regression problem. The period of 5 days is better than 7 days for this research, although the latter one is corresponding to a week. As the average error is about 2% in heating period, the accuracy of engineering practice is satisfied by this model. Moreover, the research about the proper data preprocessing principle is helpful to solve the similar problems.

[1]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[2]  Ming-Wei Chang,et al.  Load Forecasting Using Support Vector Machines: A Study on EUNITE Competition 2001 , 2004, IEEE Transactions on Power Systems.

[3]  F. Girosi,et al.  Nonlinear prediction of chaotic time series using support vector machines , 1997, Neural Networks for Signal Processing VII. Proceedings of the 1997 IEEE Signal Processing Society Workshop.

[4]  Vladimir Cherkassky,et al.  The Nature Of Statistical Learning Theory , 1997, IEEE Trans. Neural Networks.

[5]  Peng Shi-ni.  Methods of City Gas Load Forecasting and Their Analysis , 2005 .

[6]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[7]  Carlos E. Pedreira,et al.  Neural networks for short-term load forecasting: a review and evaluation , 2001 .