Date of Award

Summer 2012

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Electrical and Computer Engineering

First Advisor

Brown, Ronald H.

Second Advisor

Corliss, George F.

Third Advisor

Povinelli, Richard J

Comments

Natural gas utilities need to estimate their customers’ gas demand accurately. This thesis develops a number of daily forecasting models for test the possibility to extend the weather inputs in the current method for three different operating areas. Our goal is to improve the accuracy of our forecast by extending the number of inputs used by the existing GasDay model. We present a detailed explanation of the identification of the significance for each of the new weather input candidates. The significance of the new weather inputs was tested by statistical hypothesis testing, by forecasting performance testing, and by unusual day evaluation. We show that with some combinations of additional weather instruments, the accuracy of the forecast is improved. For most gas utilities, the primary use of natural gas is for space heating, so temperature is a critical factor when we build forecast models. In this thesis, we develop a method to split the Heating Degree Day (HDD) term into smaller pieces and generate the forecast based on these small factors. We name the method that developed as Multiple Weather Station (MWS) model in Chapter 4. We show that the MWS model yields better results compared to the existing method.

Download

Included in

Engineering Commons

COinS