Load Shedding Optimization Considering Consumer Appliance Prioritization Using Genetic Algorithm for Real-time Application

In the event of energy supply-demand imbalance caused by deficiency in energy generation, Distribution Utilities (DUs) implement load shedding methods to avoid system damages. In the current distribution set-up, consumers experience unscheduled or scheduled total blackouts with no control over which appliances to spare. This paper introduces a novel method to implement automated load shedding, which considers appliance activities and priority levels as predefined by the consumers, in a smart distribution system. The proposed method utilizes the information from the distributed appliance controllers which are assumed to have power monitoring and direct load control capabilities with bidirectional communication. Since consumer appliance switching is binary in nature, Genetic Algorithm (GA) is used to perform the optimization which is to allocate the available power supply to as many appliances as possible considering the consumer-defined appliance priority levels. With the limited power supply, consumer power allocation is determined by executing two GA processes, in each appliance controller and in the central station, respectively. The GA process in each appliance controller allocates the available supply capacity to the enrolled appliances to determine their switching on or off considering their priority levels. In order to avoid repeated switching of particular appliances, ‘fairness’ of switching implementations is judged by a proposed criterion. The remaining unallocated supply capacity is collected and optimally redistributed by GA in the central station. The case study results showed that the proposed method ensures optimum power utilization to avoid total blackouts with fast convergence signifying a promising capability for real-time applications. Furthermore, the proposed method is able to involve consumers in deciding which appliances to deload through their priority level inputs.