AVC系统电压预测及控制研究

Research on Prediction and Control of Voltage in AVC System

摘要: 通过建立支持向量机回归模型，对某一段母线电压进行预测。并通过几次模型参数的优化，将误差逐渐减小，取得令人满意的结果。利用模糊信息粒化方法，获取电压曲线的特征。结合优化的支持向量机回归方法，提出模式匹配方法对模型进行优化。实验结果表明，该方法能取得进一步的改进结果。

Abstract: The voltage of a bus bar is predicted by the Support Vector Machine Regression Model. And through several times of parameters optimization, the error is reduced gradually and the result is satisfying. With the method of Fuzzy Information Granulation, we can get the feature of voltage curve. By combining this method with the previous one, we propose the pattern matching method in order to optimize our model. The result shows that we can achieve further improvement with the combined method.