电力市场短期边际电价的分时重构混沌相空间预测

ELECTRICITY MARKET SHORT-TERM MARGINAL PRICE FORECASTING BASED ON PERIOD CLUSTERING RESTRUCTURING CHAOTIC PHASE SPACE

为了实现高精度的电力市场短期边际电价预测,该文对市场边际电价时间序列数据分时段聚类进行了相空间重构,并分别计算分形维数和提取最大Lyapunov指数,经分析得出了边际电价分时序列数据的演化具有混沌特征,由此提出了短期边际电价的分时重构混沌相空间预测算法,相比目前通常采用的单一时间序列混沌预测算法,该算法具有相空间嵌入维数少和模型参数配置灵活的特点,通过电力市场短期边际电价预测实例验证,结果表明该算法比单一时序混沌预测算法在预测精度上有显著提高。

This paper addresses the forecasting algorithm of short-term marginal price in electricity market. Phase space of time series marginal price data in electricity market was restructured by period clustering, their fractional correlation dimension and maximum Lyapunov exponent were calculated, and the conclusion that period clustering data of marginal price has chaotic property was deduced. A new model of short-term marginal price forecasting algorithm based on period clustering restructuring chaotic phase space was presented. The algorithm has advantages of less phase space dimension and flexible parameters to compare with the single-time series chaotic forecasting algorithm. Test results for using this algorithm to forecast the short-term marginal price data in actual electricity markets are reported, and show that there is great improvement in forecasting precision.