Weighting coefficients based probabilistic PSS parameter optimization

The objective function of the nonlinear programming method, used to optimize power system stabilizer (PSS) parameters under multiple operating conditions in a multi-machine system, is modified to improve the convergence property. The improved objective function, with enlarged eigenvalue range, is composed of the real part and damping ratios of all unstable and some stable eigenvalues These considered eigenvalues are multiplied by weighting coefficients determined by their corresponding stabilization degree. Based on an eight-machine system, convergence properties of the quasi-Newton method used to solve the optimization problem are discussed in details, with different eigenvalue range values and several weighting coefficient expressions, including the inverse function, linear function and so on. From these testing computations, both the weighting coefficient selecting rules and the preferable eigenvalue range are concluded. The proof analysis shows that the convergence property can be improved efficiently by the proposed approach.