Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Universities have adopted many programs to conserve electricity and counteract its high costs. Electricity can be conserved by adopting a class scheduling policy which schedules a reduced number of classes when the weather is more inclement. A regression model is postulated to measure the impact of the actual class scheduling policy on the 3-hourly electricity consumption at Louisiana State University. This model explicitly considers the effects of students' load, temperature, and humidity, and implicitly considers variables such as electricity prices and implemented electricity conservation programs through dummy variables. Also, two- and three-way interaction effects are considered in this model. Since weather-sensitive electrical devices are utilized with different degrees of intensity throughout the year, twelve models, one for each month, are estimated using OLS. It is found that the estimated regression coefficients for students' load, temperature, and humidity are not stable through time. The models are validated using intra (1977 to 1979) and extra (1980) sample observations. It is found that the models meet the OLS assumptions. Subsequently, the models are used to predict electricity consumption during the year 1980 under each of three alternative policies: Policy 1, constant student load; Policy 2, modified constant student load, and Policy 3, inverted student load. These predictions are compared to the consumption under the actual class scheduling policy. It is found that under all three alternative policies electricity dissavings result. However, Policies two and three allow the implementation of a daily building shut-down policy from 12 to 3 p.m. Electricity savings result when these policies and a building shut-down policy are assumed. Temperature and relative humidity are not the only weather factors affecting electricity consumption. Other factors such as precipitation, cloud cover, and wind speed affect electricity consumption. These factors are imbedded in the eight weather types and three weather indices developed by Muller {24, 25}. It is found that weather types or indices associated with warm air (cool air) correspond to higher (lower) electricity consumption levels.