Electric power supply grids are vital to social and economic activities as well as to public safety and wellbeing and are ranked as the highest critical infrastructure. There are substantial adverse impacts on society when power grids fail such as disruption to traffic and shut down in the operation of other critical infrastructure elements. This paper presents a novel method to assist in forecasting the probability of power outage based on weather condition in four Canadian provinces-Quebec, Ontario, New Brunswick, and Nova Scotia. System disturbances reports, provided by the North American Electric Reliability Corporation (NERC) from 1992 to 2009, have been scrutinized to determine the conditions that lead to power outage. Based on the reports above, weather condition is found to be a major cause behind power outage that justifies the necessity of a comprehensive study in this area. As a result, a forecasting model for power failure based on weather conditions is developed by artificial neural network (ANN). Once the prototype model is trained, it is able to predict the probability of power outage occurrences by utilizing forecasted weather data for a specific location. Finally, a case study is presented to illustrate the applicability and accuracy of the developed method.
展开▼
