Forecasting electricity demand using regression and Monte Carlo simulation under conditions of insufficient data

The problem studied is that of a summer peak residential energy demand model for Appalachian Power Company's service area in West Virginia. By restricting the forecast to a region smaller than the state, serious data problems result due to insufficient data to obtain reliable forecasts.;Regression analysis and Monte Carlo Simulation are the two methods used to forecast energy demand. Both methods incorporate risk into the analysis in different ways. Regression analysis yields a measure of the reliability of the coefficients of the variables and of the reliability of the forecast. The resulting forecast and confidence limits of the forecast values give an indication of the risk using regression analysis and Monte Carlo Simulation. Monte Carlo Simulation uses a probabilistic range of input values rather than a single discrete value, which accounts for future uncertainty to determine the probabilistic future summer peak.

Name thesis:degree_name

MS

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Economics

Year dc:date.available

1999

Author dc:creator

• Cullen, Kathleen Ann

Contributors dc:contributor

• Thomas F. Torries.

• Economics
• Energy
• Operations research

Identifier

https://researchrepository.wvu.edu/etd/974

OAI identifier oai:identifier

oai:researchrepository.wvu.edu:etd-1977