Date created: | Last Updated:

Category: Project

Description: Electricity flows on high-voltage alternating current transmission systems follow the laws of physics, not the laws of supply and demand. Thus there can be large differences between scheduled and actual electricity flows when the two sets of laws are not aligned. This is a significant operational challenge for electric power system operators, as increases or decreases in generation may be required to ensure the reliability of the electric power system. System operators have attempted without success to accurately model high-voltage electricity flows. In this paper, we present evidence that the electricity flows between Ontario and New York can be modeled using econometric methods. The modeling framework makes use of weather data, electricity load data, measures of transmission “network effects,” proxies for geomagnetically induced currents (GICs) as well as the expected level of power grid conditions. The model is estimated using hourly data over the period of 1 May 2002 through 31 October 2003. The model is evaluated using out-of-sample hourly data over the period of 1 November to 9 December 2003. The out-of-sample predictions are more accurate when the forecasting equation reflects the estimated contribution of geomagnetic activity. The structural modeling results for the sample period indicate that peak predicted effect of geomagnetic activity on the electricity flow was about 1,604 MWh in absolute value when the GIC proxy achieved a value of 363.2 nT/min. The findings also indicate that geomagnetic activity is an important driver of the volatility in the electricity flows.