85th AMS Annual Meeting

Monday, 10 January 2005: 5:00 PM
Electric load forecast sensitivity to weather inputs
Daniel O'Connor, ISO New England, Holyoke, MA
Accurate forecasts of next-day electrical demand are critical components of efficient energy markets. Control Area operators, like ISO New England, rely on hourly demand forecasts to create generation dispatch schedules that meet reliability and economic standards. Market participants rely on load forecasts to construct bidding strategies that enhance revenue and avoid losses. Weather concepts are an integral part of these daily forecasts.

Mr. O'Connor will present an overview of how an Independent System Operator (ISO) produces its day-ahead load forecasts and the importance of weather variables in producing workable estimates of consumer demand for electricity at the regional level. The talk will address inputs to neural network and other demand model constructs, and the sensitivity of load forecasts to temperature as a primary driver of variations in seasonal electrical consumption. As an end-user of weather forecast services, Mr. O'Connor will challenge the meteorological community to develop enhanced weather products that can quantify uncertainty and help improve the efficiency of the national market for electricity.

Supplementary URL: