A latencies tolerant model predictive control approach to damp Inter-area oscillations in delayed power systems

摘要

Automatic voltage regulators and power system stabilizers have been employed successfully to deal with fast dynamics associated to local oscillations phenomena. Whereas, inter-area oscillations damping in large power systems requires remote feedback controllers fed by Wide Area Monitoring Systems (WAMS). Although these improvements in the inter-area oscillations behavior have been effective, a new challenge emerges: reaching stability with a closed loop control despite latencies due to measurements taken far away from the control centre. The research was motivated by the need for modernization of power systems capable of dealing with control difficulties in centralised WAMS for damping inter-area oscillations in power systems caused by delays in the communication system. Herein, time delay and control problems are addressed separately. The time delay problem is solved by a database based time compensation solution relying on the most updated available state of the system. The control problem is solved by a Model Predictive Control (MPC) with terminal cost and constraint set to handle complexities due to nonlinearities of the power system, the large scale of the problem and the parametric variations. Both solutions work in a coordinated way with local controllers to implement a decentralised coordinated strategy that manages slow global dynamics and fast local dynamics as well. The integrated proposed approach is called Time-Delay-Tolerant Model Predictive Control (TDT-MPC). Coordinated and coherent performance of the two TDT-MPC components (Kalman compensator and MPC) is achieved thanks to unifying power system reference models for both strategies. The approach has been tested on the IEEE 39 system and validated with time domain nonlinear simulations, obtaining post-fault damped oscillations and a good tracking of new power references when tripping tie lines.