基于多预测校正内点法的WLAV抗差状态估计

A Robust WLAV State Estimation Based on Multiple Predictor-Corrector Interior Point Method

针对预测-校正内点法(predictor-corrector primal-dualinterior point method,PCPDIPM)加权最小绝对值状态估计(weighted least absolute squares,WLAV)可能发生校正方向指向错误方向的不足,提出一种基于多预测-校正内点法(multiple PCPDIPM,MPCPDIPM)的WLAV抗差状态估计算法。该算法在PCPDIPM的基础上,通过多次校正,对中心参数动态估计,并采用2阶段线性搜索法确定校正方向在总的牛顿方向中的最优比重,从而保证迭代点向中心轨迹靠拢。最后,通过IEEE算例仿真和我国某省网的测试结果验证了所提方法的有效性。与含不良数据辨识功能的加权最小二乘状态估计相比较,所提方法的收敛速度及抗差能力具有明显的优势。

In allusion to the defect that during the weighted least absolute squares（WLAV） state estimation based on predictor-corrector primal-dual interior point method（PCPDIPM） it is possible the corrector direction possibly points to wrong direction,a multiple PCPDIPM based robust WLAV state estimation algorithm is proposed.On the basis of PCPDIPM,through multiple correction the proposed algorithm dynamically estimates the centrality parameter and by use of two-stage linear searching method the optimal proportion of the corrector direction in the Newton direction is determined to ensure that the iteration points draw close to centrality parameter.Finally,the effectiveness of the proposed method is verified by simulation results of IEEE 14-bus system and test results of a certain provincial power grid in China.The convergence speed and robustness of the proposed method are much better than the weighted least square state estimation with the function of bad data identification.