Research On Power System Harmonic State Estimation Based On Robust Least Squares

The harmonic state estimation based on Robust least-squares,deals with the harmonic synchronous measurement data according to certain criteria for robust estimation, so it can get the objective function’s optimal state value. After studying the status of harmonic state estimation in domestic and foreign, we can see that wide area measurement system which based on GPS provided a technical foundation for harmonic state estimation, measurement data is Synchronized in time. It selects the bus harmonic voltage phasor as state variables, the bus harmonic voltage phasor and branch harmonic current phasor and harmonic current of injected into the bus as measurement phasor.That makes the harmonic state estimation equations as linear equations, which greatly simplifies the difficulty of solving the equation. Robust least squares combines of robust estimation principle and least squares by equivalent weight.Therefore, its subject is least squares,the core is the choice of weight. The weight is nonlinear function of residual. In the iterative process, the weight is constantly changing. In recent years, robust least squares is gradually applied to power system state estimation, and achieved some results. This paper proposes that it uses the robust least squares for harmonic state estimation, which is proposed for shortcoming of least squares having no ability to resist gross errors; It used MATLAB simulation software to build a simulation model of three-phase power system with5nodes network and IEEE14classic simulation model:It wrote out the harmonic state estimation equations according to the harmonic synchronization measurement data and wrote out procedures of the least squares and robust least squares; Used the procedures to solve the equations. The simulation results can be obtained that: least squares does not have the ability to resist gross error;When measurement data contain gross error, robust least squares well resists gross error, the estimate error is smaller; When the content of gross error is more,the advantage of robust least squares is more obvious.