Commande non lineaire adaptative des reseaux electriques multi-machines (French text)

A decentralized adaptive nonlinear multivariable speed and voltage controller is proposed to improve the stability and to dampen local and inter-area oscillations in multimachine power systems.; The power system is first modeled using a novel approach that unifies the classic single machine infinite bus and multimachine modeling approaches. We obtain a decentralized nonlinear time-varying fifth order multivariable system. The time-varying parameters represent the influence of the rest of the power system on each generator. The excitation and the hydraulic turbine servomotor reference signal are the input variables. The terminal voltage d-q frame components, the rotor speed, the turbine flow rate and the gate opening are the new model's state variables.; A straightforward application of the input output linearization design technique leads to an unstable closed loop system because of the non-minimum phase property of the system. A novel design method for nonlinear non-minimum phase systems is then proposed. The technique consists of applying the input-output linearization procedure first and stabilizing the closed loop system using a nonlinear auxiliary control law of internal and external state variables. The novel technique is used to design a robust decentralized voltage and speed regulator. The gains of the controller's linear part are obtained by solving an algebraic Riccati equation. It considerably attenuates inter-area and local oscillations. Integrators are then introduced in the control law to improve the robustness of the controller. The effectiveness of our design procedure is tested on a four machines two-area power system. The simulation results assess the performance of the new controller compared to the standard power system regulator with PSS (power system stabilizer). An adaptive version of the controller is proposed to improve the robustness property. The parameters of the controller are updated after a load variation. The design of the new regulator is based on the adaptive version of our nonlinear control method for non-minimum phase systems. The adaptive voltage and speed regulator is also tested on the four machines power system. The results prove the effectiveness of the novel modeling approach and design procedure.