Robust State Estimation Of Power System Based On Uncertainty Measure

With the rapid development of the power information physical system,accurate prediction,exact decision-making and precise control of power system become necessary,which depend on accurate data.However,as the basis and core of the energy management system,the traditional weighted least squares estimation method and the bad data identification method based on residual have limited ability to identify strong correlated bad data,which has been unable to fully meet the growing production and security requirements.The robust performance and computational efficiency of power system state estimation need to be improved urgently.For this reason,this paper makes a deep research on the state estimation under the uncertainty theory.The main work of this paper is as follows:(1)The normal measuring point,abnormal measuring point and normal rate are redefined based on the uncertainty measure,and a new state estimation result evaluation is given to take account of the normal rate and the deviation from the truth value,which can be applied weather the true value is known or unknown.(2)Based on the evaluation indexes of both the normal rate and the deviation of the measuring points,the normal point participation function and the state estimation criterion function are defined,and the maximum normal least deviation estimation method(MNLD)is proposed.MNLD is based on multi-objective programming model with the highest normal rate and the smallest deviation of the normal measuring points,which guarantees the rationality of the estimation results in theory.(3)In order to apply the MNLD model to the practical solution,3 equivalent transformations are made.(a)Multi-objective programming is transformed into a single objective programming problem by objective programming method.(b)A hyperbolic tangential rectangular pulse function(RPF)is used to replace the normal point participation function.(c)The improved cohesive function is used to approximate the function like max(7)(8)(infinite norm function)and solved by interior point method.(4)Simulation among MNLD and other methods is done including robust performance and computational efficiency.A large number of examples show that MNLD has good performance in robust performance and computational efficiency.