Multivariable adaptive control algorithms and applications to multi-machine power system stabilizer (PSS) design.

摘要

Modern power systems consist of numerous generating units (G.U.) dispersed all over networks at different geographical locations. In case of insufficient damping in both electrical and mechanical modes, low frequency oscillations (LFO) in generating unit speed, output power and terminal voltage may be triggered by various disturbances. Without timely and properly handling and control, these oscillations can sustain, continue to grow and eventually cause systems to collapse. Conventionally, Power System Stabilizers (PSS) either with fixed parameter structures or designed on single-machine infinite-bus (SMIB) system base is employed to damp out the oscillations. However, due to lack of adaptivity to system changes and/or coordination to other PSS in the system, its performance is limited.; This dissertation proposes four multivariable adaptive control algorithms and applies them to multi-machine power system (MMPS) PSS design. These algorithms and the corresponding PSS design schemes are all self-tuning controllers in the nature; therefore, they can adapt to system operating conditions and parameter changes to continuously provide desired damping effect to the system oscillations. They are also computationally efficient thus suitable for on-line implementation. In addition, the fourth algorithm, the Generalized Integral Feedback Minimum Variance (GIFMV) control, and the associated PSS design scheme are directly derived from synchronous machine equations in a multi-machine environment. This type of PSS is decentralized, with all input variables locally defined and physically measurable. External system dynamics are coupled into the PSS through carefully selected interfacing variables. These features assure that all PSS in the system are properly coordinated for overall system stability improvement.; All four proposed PSS have been tested on sample SMIB and MMPS respectively, under different generating unit and system operating conditions with various disturbances. Results presented show that the proposed PSS meet design requirements and provide consistent as well as effective damping to the generating unit low frequency oscillations in multi-machine systems.