Trajectory-Based Supplementary Damping Control for Power System Electromechanical Oscillations

Abstract

Abstract-This paper considers a trajectory-based approach to determine control signals superimposed to those of existing controllers so as to enhance the damping of electromechanical oscillations. This approach is framed as a discrete-time, multi-step optimization problem which can be solved by model-based and/or by learning-based methods. This paper proposes to apply a model-free tree-based batch mode Reinforcement Learning (RL) algorithm to perform such a supplementary damping control based only on information collected from observed trajectories of the power system. This RL-based supplementary damping control scheme is first implemented on a single generator and then several possibilities are investigated for extending it to multiple generators. Simulations are carried out on a 16-generators medium size power system model, where also possible benefits of combining this RL-based control with Model Predictive Control (MPC) are assessed.

(Deep) Reinforcement learning for electric power system control and related problems: A short review and perspectives

“…Oscillatory angle instability relates to the problem of low-frequency oscillations in the system (local modes in the range [0.7 − 2.0] Hz and inter-area modes in the range [0.1 − 0.8] Hz). This type of instability was considered in the context of controlling individual system components [17,[58][59][60][65][66][67][68][69] and wide-area controls [70][71][72]. Some of the works [17,[58][59][60] are already discussed in the context of transient angle instability and some comments are valid for oscillatory angle instability consideration.…”

“…In [68] a set of quadrature boosters devices, controlled by fuzzy controllers, are coordinated by Q-learning for oscillations damping. Fitted Q-iteration RL method was considered in [69] where a trajectory-based approach was designed as supplementary to existing controllers. In [70] a wide-area decentralized power system stabilizer was designed using Q-learning for damping both local and inter-area low frequency oscillations.…”

“…• In general, there is a lack of efficient fusion of (D)RL models with control theory and practice in electric power systems. Few exceptions exist [59,62,68,69] where RL was fused with the concept of control Lyapunov functions [62], model predictive control [59,69] and fuzzy logic based control [68].…”

“…• Fusion with advanced methods coming from control theory and engineering (some example exists, see [55,59,62,68,69], for some future on control see [2][3][4]). Recent work presented in [100] is another example of combining RL and a model predictive control and is worth of considerations in electric power systems.…”

Design of Nonlinear Robust Damping Controller for Power Oscillations Suppressing Based on Backstepping-Fractional Order Sliding Mode

“…However, these schemes require additional equipment. Meanwhile, the second category involves techniques that modify conventional control schemes [6][7][8]. In power systems, damping capability can be improved by way of the first category; however, these approaches increase in cost and complexity.…”