基于联合卡尔曼滤波的配电网多源异构数据融合

由于传统方法没有对配电网多源异构数据采集的时间进行配准，导致配电网多源异构数据融合时误差大、效率低。针对该问题，提出了基于联合卡尔曼滤波的配电网多源异构数据融合方法。构建了数据纠偏机制，采用最小二乘法对数据采集的时间配准，并采用拉格朗日插值方法对时序数据填充，计算数据关联性。在此基础上，采用联合卡尔曼滤波算法将相同数据融合到同一个类中，以此实现配电网多源异构数据融合。实验结果表明，所研究的数据融合方法不仅能够根据需求一直追踪需求的融合误差，还能够降低节点电压与功率估计的相对误差、提高配电网多源异构数据融合的效率。实验结果不仅证明了所研究融合方法的有效性，还证明了联合卡尔曼滤波算法在数据融合中的适用性。

Multi-source heterogeneous data fusion of a distribution network based on a joint Kalman filter

The traditional method does not register the time of multi-source heterogeneous data collection in the distribution network. This results in large errors and low efficiency in the fusion of multi-source heterogeneous data. Thus a distribution network multi-source heterogeneous data fusion method based on joint Kalman filtering is proposed. First, a data correction mechanism is constructed, the least squares method is used to register the time of data collection, and the Lagrangian interpolation method is used to fill the time series data to calculate the data relevance. Then, the joint Kalman filtering algorithm is used to fuse the same data into the same class, so as to realize the multi-source heterogeneous data fusion of the distribution network. Experimental results show that the method not only can track the fusion error of the demand according to the demand, but also reduce the relative error of node voltage and power estimation, and improve the efficiency of multi-source heterogeneous data fusion, demonstrating the effectiveness of the proposed method and applicability of the joint Kalman filter algorithm in data fusion. This work is supported by the Science and Technology Project of China Southern Power Grid Co., Ltd. (No. 670000KK58200011).