Research On Algorithms For Parameter Optimization Of Power System Damping Controllers

Interconnections of regional power systems and operation of Three Gorges power plant indicate that China nationwide power system has been basically established. Due to weak connections between regional systems, low frequency oscillations have become one of constraints on the power transfer limits among regional power systems. Currently, Power System Stabilizers (PSS) and FACTS damping controllers are widely used to suppress the low frequency oscillations and to improve small signal stability. Coordination design of PSS and FACTS damping controllers in multi-machine power systems is a challenging task and further research efforts are required to obtain more insights in this topic. This thesis mainly focuses on optimization design of damping controller parameters. Some contributions are summarized as follows The thesis proposes an approach called Bidirectional Module Reduction (BMR) for power system transient stability simulation and trajectory sensitivity computation. The BMR technique comprises processes of forward reduction and backward evaluation. The forward reduction process deals with increment equations of generators and their controllers to obtain solvable network equations. And the backward evaluation process evaluates generator variables using the solution of the network equations. By means of the BMR technique, the simulation approach based on Newton iteration will become more modular,expandable,easier for programming while remaining Newton method's quick convergence with high precision and good numerical stability. A method based on time-domain simulation and trajectory sensitivity analysis is proposed for setting PSS parameters in multi-machine power systems. The method considers the nonlinear characteristics of power systems by means of time domain simulation and coordinates the influences of PSS parameters on system eigenvalue and nonlinear interaction coefficient by optimization approaches. One of advantages of the method is that PSS designed by this method not only can improve small signal stability of systems, but also can significantly damp out great oscillations brought about by a severe disturbance. The thesis first adopts a simplex-simulated annealing (SSA) approach for optimal design of PSS parameters. This approach formulates the PSS parameter design as a nonlinear programming problem with inequality constraints. Then a novel nonlinear transformation is employed to transform the constrained nonlinear programming into an unconstrained one. This unconstrained optimization problem is solved for optimal damping controller parameters by the SSA approach. The SSA approach, an improvement of Simulated Annealing method (SA), hybridizes the downhill simplex method with the SA method such that SSA possesses the advantages of fast convergence and ability for a global minimum.