On Power System Stabilizers: Genetic Algorithm Based Tuning and Economic Worth as Ancillery Services
The problem of power system stabilizer (PSS) parameter tuning is a complex exercise because of the presence of poorly damped modes of oscillation, and continuous variation in power system operating conditions. In recent years, as a result of the newly deregulated market environment, power systems tend to be operated at reduced security margins, thus making the system more vulnerable to disturbances. The role of PSS, among other power system control devices, becomes then even more critical, and subsequently, in light of the new operating paradigm, new methods of financial compensation for the generators providing this service to system would be needed.
The work presented in this thesis focuses on aspects related to PSSs tuning, on one hand, and evaluation of their contribution to system stability and security from an economic perspective in the context of ancillary services, on the other. Thus, a genetic algorithm (GA) based method to simultaneously tune PSSs is developed in the first part of this thesis, while the second is dedicated to developing a game theory based method to financially compensate the PSSs for the control effort they provide for the power system.
The simultaneous approach for tuning PSSs, as opposed to sequential approaches, usually involves exhaustive computational efforts, but, in turn, ensures the parameter setting optimality. The classical Lyapunov's parameter optimization method employing an Integral of Squared Error (ISE) criterion has been integrated within a GA framework to simultaneously tune PSSs. Within the genetic process, a potential solution the PSS parameter setting is coded as an individual, which is part of a population of such potential solutions randomly generated, and by applying the survival of the fitness principle based on each individual's fitness with respect to the objective, a sound basis to finding the best individual, i.e. global optimum solution, is created. The method thus emerged has been used for tuning of lead-lag and derivative PSS. A similar GA based optimization process is implemented for tuning the proportional-integral-derivative (PID) PSS. Since the PID PSS acts in discrete mode, the power system model has been accordingly developed in discrete-time domain. An optimal sampling period has been determined considering the conflicting requirements of computation time vis-à-vis accuracy of information on system dynamics due to discretization. Tests for transient events, such as three-phase short circuits and transmission line outages, have also been performed with satisfactory results.
In the second part of this thesis, an attempt has been made to examine the role and performance of PSS in the context of deregulated power markets. It is proposed that the PSS control effort to enhance power system stability and security be regarded as an ancillary service PSS-control service and subsequently, the allocation of system savings/benefit, as accrued from a PSS, becomes an important issue. A game theory based approach, namely the Shapley value criterion, is used to develop a scheme for allocation of payments to generators equipped with PSS and providing this service. The PSSs contribution is evaluated in different ways: by assessing the transfer capability of the system due to PSS, or by employing various performance indices based on system dynamic behavior. A contingency analysis is also performed and the N-1 security criterion is taken into consideration as well, for evaluating the PSS-control payment.