Development of Optimized Under-Frequency Load-Shedding Scheme With Renewable-Energy Integration.

Abstract

Under-frequency load shedding maximizes the performance of a power system at times of severe power imbalance. It involves trading off the total amount of load shed versus required frequency stability level, in order to prevent frequency collapse of the grid. For reliable power supply to consumers, renewable energy powered off-grid systems complement the main grid supply. Renewable energy micro-grids experience adverse frequency instabilities due to intermittent nature of the generation sources, and thus requiring load shedding. This comes at a loss of load and reduces operator’s revenue, in pursuit of power system security and stability levels that comply to grid codes. Given the unpredictable consumer behavior, and the fact that shedding loads leads to economic losses for the utility, and a blackout to consumers, an optimal process of load shedding that disconnects the least load from the grid was developed in this research. The developed load shedding attains the grid code frequency stability levels and only applies as the option of last resort, towards securing the network from frequency collapse. A hybrid particle swarm – genetic algorithm, coded in MATLAB, is utilized to optimize the load shedding process given the conditions and constraints surrounding the modeled micro-grid, which is powered purely by renewable energy sources. Results from this study ascertain the optimization of load shedding process for the studied micro-grid. Though load shedding restores the micro-grid back to its stable operation, 100% optimization was not achieved. For instance, a maximum power imbalance of 269MW, has an optimum load combination constituted of 12 loads and falls in the second stage/ second priority of load shedding i.e. combines loads in domestic/ agricultural and commercial categories and spares industrial category of loads as per the prioritization. The total load disconnected for this imbalance is 295MW, consisting of 5MW, 6MW, 8MW, 10MW, 14MW, 16MW, 18MW, 23MW, 25MW, 31MW, 42MW and 97MW feeders. Hence a difference of 26MW (295-269) is disconnected unnecessarily since there is no single combination that adds up to 269MW. In conclusion, this research improves the design and practices of load shedding processes, lowers the Incidence of blackouts in off-grid areas and offers a solution for recovering the frequency stability of renewable energy islanded micro-grid networks, at times of severe power deficiency.