Iteratively re-weighted least absolute value method for state estimation

A modified version of the weighted least absolute value (WLAV) method for the solution of the power system-state-estimation problem is presented. The WLAV method can be seen as minimising a linear objective function subject to a set of nonlinear constraints. The modification is aimed at producing a WLAV estimator that remains insensitive to bad data, even if they are associated with leverage-point measurements. This is achieved by bending the linear objective function, so that the residuals of bad measurements are allowed to grow without incurring much additional cost in the objective function. Consequently, the optimisation procedure would find an optimal solution where the residuals of all bad measurements are nonzero. The optimisation is carried out via sequential linear programming. It is shown that each linear program corresponds to a linearised WLAV problem with weights adjusted automatically during the iterations. To ensure fast execution time, the linear program is solved using a homogeneous interior-point method. Computational results show that the proposed method can identify bad data in leverage points.