Load Modeling Based on Support Vector Machine Based on Bayesian Evidence Framework
Wang Zhenshu1, 2, Li Linchuan1, Niu Li3
1. Key Laboratory of Power System Simulation and Control of Ministry of Education Tianjin University Tianjin 300072 China 2. Shandong University Jinan 250061 China 3. State Nuclear Electric Power Design & Research Institute Beijing 100032 China
Abstract:Load modeling is still one of the difficult problems in power system. Accurate load model plays a very important role in power system digital simulation. This paper presents a support vector machine (SVM) load modeling method which bases on Bayesian evidence framework. According to the load characteristic data acquired from wide area measurement system(WAMS), the load model based on SVM method is founded, and it chooses Gaussian radial basis function (RBF) to optimize the structure of the model. Among three levels of Bayesian evidence framework inference, the level 1inference is used to explain SVM training, both levels 2 and 3 can also be applied to SVM. Load model parameters are identified and optimized by using three inference levels of Bayesian evidence framework. Simulation tests for SVM load model verify the validity of this method. SVM load model based on Bayesian evidence framework has good generalization ability, flexible structure, and rapid calculation speed, so it can describe the actual load characteristics more accurately.
王振树, 李林川, 牛丽. 基于贝叶斯证据框架的支持向量机负荷建模[J]. 电工技术学报, 2009, 24(8): 127-134.
Wang Zhenshu, Li Linchuan, Niu Li. Load Modeling Based on Support Vector Machine Based on Bayesian Evidence Framework. Transactions of China Electrotechnical Society, 2009, 24(8): 127-134.