基于优化权重的卡尔曼滤波与无偏灰色组合模型的短期负荷预测  

Short Term Load Forecasting based on Kalman Filter and Unbiased Grey Combination Model

作　　者：林天祥 张宁[1] 胡军 

机构地区：[1]福州大学电气工程与自动化学院,福州350116 [2]山西漳泽电力股份有限公司,太原030006

出　　处：《电气技术》2017年第9期19-23,共5页Electrical Engineering

摘　　要：本文针对短期负荷预测动态、随机的特点,提出了一种基于优化权重的卡尔曼滤波与无偏灰色组合预测模型。该模型充分发挥了卡尔曼滤波准确估计动态系统的优势,并合理利用无偏灰色模型挖掘随机数据潜在规律的特点。首先根据卡尔曼滤波预测中出现特殊日收敛不足的缺陷,利用趋势稳定,规律性强,消除固有偏差的无偏灰色理论加以弥补。根据无偏灰色理论趋势稳定向上,在短期负荷预测中某些下降趋势数据点误差较大的缺陷,利用卡尔曼滤波依据大量数据最优估计的平均思想加以弥补。并且采用线性组合法进行结合进一步规避了预测风险。算例结果表明,该预测模型精度较高,具有实用性。According to the dynamic and stochastic characteristics of short-term load forecasting, a Kalman filtering and unbiased grey combination forecasting model is proposed. The model gives full play to the advantage of the Kalman filter to estimate the dynamic system, and makes use of the unbiased grey model to explore the characteristics of the random data. According to the defects of the special day convergence in the Kalman filter prediction, this paper makes use of the unbiased grey theory which is stable, regular and eliminates the inherent deviation. According to the trend that the trend of the unbiased grey theory is stable and the error of some data points in the short-term load forecasting is large, the Kalman filter is used to compensate the defect. Linear combination method is used to further avoid the prediction risk. The results show that the model has high precision and practicability.

关 键 词：卡尔曼滤波模型 无偏灰色模型 短期负荷预测 组合预测 

分 类 号：TM715[电气工程—电力系统及自动化]