Document Type

Master's Culminating Experience

Publication Date



The changing natural gas industry has increased the importance for utility companies to develop an accurate peak day forecast. The model developed for this midwest utility estimates a firm natural gas sendout econometrically, using the Ordinary Least Squares regression analysis. The peak is primarily weather driven, thus the model made use of wind-chill variables from the current day, the previous day, and the current day squared. Also included is a disposable income variable to reflect the level of economic activity.

The peak day forecast depends on using extreme weather conditions for the design day parameters. The parameters used represented the second worst wind-chill for the utility's service area to occur in the last thirty years of history. The forecasted peak for the 1993/1994 winter season is calculated from the model to be 515,654 MCF, or 530,092 DTH. The actual peak occurrence for this season fell on January 18, 1994 where firm sendout reached an all-time high of 513,876 MCF, or 528,261 DTH.