【摘要】 系统边际电价是电力工业改革的关键因素之一,是电力市场的杠杆和核心内容。为克服神经网络预测法易陷入局部极小,隐层数不易确定,训练速度慢等问题,提出一种基于相似搜索和最小二乘支持向量机的系统边际电价预测方法,该方法对相似搜索得到的相似日的负荷—电价数据用最小二乘支持向量机建立电价预测模型,同时利用网格搜索和交叉验证自动选取最小二乘支持向量机相关参数。用美国加州电力市场的真实数据做实例验证结果表明该方法可有效提高预测精度。更多还原

【Abstract】 System marginal price is the key point of the electricity industry innovation and one of the key problems in electricity market.In order to exactly forecast the day-ahead system marginal price of electricity market,a novel system marginal price forecast model based on similarity search and least square support vector machine is put forward.The load-price data obtained by similarity search is used by least square support vector machine to establish the price forecast model.And grid search and cross validation are adopted to search the best parameters of least square support vector machine.Theoretical analysis indicates that this method is effective for selecting the optimized parameters of least square support vector machine.The historical data from California electricity market is used in the case study to forecast the day-ahead system marginal price,the result shows that the price forecast model based on similarity search and least square support vector machine has effectively increased the forecasting precision compared with the forecast method based on BP neural network.更多还原