Parameter optimisation of robust power system stabilisers by probabilistic approach

An approach for the parameter optimisation of power system stabilisers (PSSs) is presented, in which system multioperating conditions are taken into account by probabilistic eigenvalue analysis. With the assumption of normal distribution, the statistical nature of the eigenvalues is described by their expectations and variances. The objective function used for PSS parameter optimisation is constructed from the expectations and variances of eigenvalues, and minimised by the quasi-Newton technique of nonlinear programming. The proposed algorithm is employed on an eight-machine test system. In which the PSS site selection and the convergence property are discussed.