Belgian gas transmission company Fluxys reports successful deployment of Energy Solutions International’s (ESI) GasLoadForecaster (GLF). GLF uses neural net technology to forecast volume and energy targets at supply points along the pipeline network. This neural network finds patterns in complex pipeline operations resulting from weather fluctuations, calendar information and other variables. The simulator is ‘trained’ regularly to fine tune prediction parameters.
Fluxys needed a forecasting tool to anticipate gas demands help its operators react to demand fluctuations and transport constraints as well as to additional shipper capacity requests.
Fluxys project leader Sophie Jehaes said, ‘We chose GLF because of its accuracy and its comprehensive administrative functions and overall ease of use. As an IT team member, I am interested in a solution that can be easily integrated with our other business applications but also can work standalone.’
ESI CTO John Sherman said, ‘GLF is a fast, flexible and easy-to-use tool. Users see immediate improvements in their forecasts after implementing this product. The multivariate, non-linear nature of gas market forecasting makes this a perfect application for neural networks.’
© Oil IT Journal - all rights reserved.