An effective Load shedding technique for micro-grids using artificial neural network and adaptive neuro-fuzzy inference system

In recent years, the use of renewable energy sources in micro-grids has become an effective means of power decentralization especially in remote areas where the extension of the main power grid is an impediment. Despite the huge deposit of natural resources in Africa, the continent still remains in energy poverty. Majority of the African countries could not meet the electricity demand of their people. Therefore, the power system is prone to frequent black out as a result of either excess load to the system or generation failure. The imbalance of power generation and load demand has been a major factor in maintaining the stability of the power systems and is usually responsible for the under frequency and under voltage in power systems. Currently, load shedding is the most widely used method to balance between load and demand in order to prevent the system from collapsing. But the conventional method of under frequency or under voltage load shedding faces many challenges and may not perform as expected. This may lead to over shedding or under shedding, causing system blackout or equipment damage. To prevent system cascade or equipment damage, appropriate amount of load must be intentionally and automatically curtailed during instability. In this paper, an effective load shedding technique for micro-grids using artificial neural network and adaptive neuro-fuzzy inference system is proposed. The combined techniques take into account the actual system state and the exact amount of load needs to be curtailed at a faster rate as compared to the conventional method. Also, this method is able to carry out optimal load shedding for any input range other than the trained data. Simulation results obtained from this work, corroborate the merit of this algorithm.