An Adaptive Load Shedding Technique for Controlled Islanding

The main contribution of this paper is the development of a novel load shedding methodology so as to regulate unacceptable frequency deviations, with least disruption of service, in the case of controlled islanding. To achieve this, an abstracted model of each islanded area, consisting of one lumped generator and turbine model, is used. Since in the controlled islanding the bounds of the areas are determined, a good estimation of the generation-load mismatch can be obtained. Taking advantage of the latter an amount of load to be shed in the load-rich areas, at the moment of the area separation is proposed. This amount is optimally selected so as to minimize the total load shed, taking into account the primary frequency control dynamics and the conventional Underfrequency Load Shedding schemes (UFLS). Markov Chain Monte Carlo (MCMC) methods and Model Predictive Control (MPC) are the two optimization tools that are used to solve the resulting optimization problem. The effectiveness of the proposed technique is evaluated via simulation on the IEEE 118 bus network.