The Simultaneous Design of Multiple Power System Stabilization Controllers Using the System Identification and the Non-linear Optimization Technique

Abstract

We have developed a simultaneous control design method for multiple power system stabilization controllers such as PSSs using the system identification and the non-linear numerical optimization technique. The developed controller design can take into consideration of controller structure and several design constrains in the controller design. For multiple controllers design, a reduced order MIMO linear system, which show similar dynamic properties to some original full order power systems, is identified from time domain simulation results by N4SID algorithm. The nonlinear optimization technique is applied to minimize the Modal Performance Measure, which is directly related with damping of the power system oscillations. As an example of the design, PSSs design for IEEJ EAST30 system model is described.

System identification of power network for coordinated setting of PSSs

“…2. The objective of this study was to obtain an identified model to be used in the MPM method [1,2,5]. For the model output, we used the active power P and the frequency ω of every generator.…”

“…Power network models on simulator that is, the active power P and frequency ω of every generator, are input to the target PSSs, compensated for phase lead-lag and gain, and then input to the model AVRs. The objective of this study was to obtain an identified model to be used in the MPM method [1,2,5]. The MPM method has the distinctive characteristic that the lead-lag constants and gains are set in coordinated fashion so as to optimize the damping of the closed loop composed of the PSS and the identified model; therefore, pole placement must be identified precisely for proper evaluation of the damping variation.…”

“…There have been attempts to use identified models with multiple oscillation modes existing in the power system for coordinated setting of multiple ∆P + ∆ω-PSSs by the MPM (Modal Performance Measure) method [1,2,7,8]. There have been attempts to use identified models with multiple oscillation modes existing in the power system for coordinated setting of multiple ∆P + ∆ω-PSSs by the MPM (Modal Performance Measure) method [1,2,7,8].…”

“…Such ∆P + ∆ω-PSSs installed on multiple generators to suppress common long-period oscillations are usually set individually using the one-machine infinite-bus model, which does not necessarily produce optimal settings in terms of the entire power system. There have been attempts to use identified models with multiple oscillation modes existing in the power system for coordinated setting of multiple ∆P + ∆ω-PSSs by the MPM (Modal Performance Measure) method [1,2,7,8]. In addition, it has been shown that coordinated setting using a large-scale system model becomes possible when the damping torque coefficient K d and other long-known performance indices are incorporated as constraints for optimization [5].…”