从工频磁场数据到架空输电线路三相电流的反演逆推方法

Inversion Method for Calculating Three-phase Currents of Overhead Transmission Lines by Using Power-frequency Magnetic Field Data

随着智能电网的发展,沿架空输电线路对其电量参数进行实时监测的需求日益增长。然而,在已有线路中加装互感器存在诸多困难,甚至根本不具备实现条件。鉴于此,提出了一种全新的三相架空输电线路电流检测方法,即利用线路下方空间的工频磁场测量数据,通过逆推寻优算法反演线路电流值,然后进行曲线拟合并提取特征参数。该方法可实现线路电流的非接触测量,在实际工程应用中具有更安全、更方便灵活的优点。为了使逆推所寻到的解尽可能接近物理真实值,在寻优算法中设置了3个目标函数,采用带精英策略的快速非支配排序遗传算法(NSGA-Ⅱ)进行多目标寻优。此外,为弥补多目标寻优计算量大、求解时间长的缺点,选择特定测量点,推导特定测量点磁场分量与三相电流之间的约束关系,将寻优算法中的3个决策变量缩减为单个决策变量,极大地降低了逆问题求解的复杂度。最后通过算例验证了所提出的方法的可行性和准确性。

With the development of smart grid, the demand for real-time monitoring of the electric parameters （e.g. current） along the overhead transmission lines is increasing. However, there are many difficulties in the installation of transformers in the existing lines, or even there are no the achievable conditions for installation. In order to solve the problem, a new detection method for three-phase currents of overhead transmission line was proposed, in which the measured data of power-frequency magnetic field under transmission lines were used to back-calculate the current values in the conductors by inversion and optimization algorithm. Then the calculated current data were fitted as curve and the characteristic parameters were identified. Non-contact measurement of the currents of overhead TLs can come true by the use of the proposed novel inversion method, which is more secure and more flexible for the practical engineering application. In order to obtain the solution as close as possible to the physical reality, three object functions were set up in the optimization algorithm and the fast and elitist non-dominated sorting genetic algorithm （NSGA-Ⅱ） for multi-objective optimization was applied. Moreover, to remedy the defect of multi-objective optimization being featured in large computation and long time-consuming, some specific measurement points were singled out and the constraint conditions between magnetic field components of the given measurement points and the currents of transmission lines were deduced. The three identification variables were pared to only one in the inverse calculation process, which greatly reduce the computation complexity of the inverse problem solving. Finally, the simulation analyses verify the feasibility and accuracy of the proposed inversion method.