Date of Award
Spring 2007
Document Type
Dissertation - Restricted
Degree Name
Doctor of Philosophy (PhD)
Department
Electrical and Computer Engineering
First Advisor
Brown, Ronald H.
Second Advisor
Corliss, George F.
Third Advisor
Hock, Jeffrek
Abstract
We introduce two algorithms to disaggregate time series data to a desired frequency given a training set of data at the desired frequency and factors at the desired frequency. These algorithms use parametric models and can be used to generate estimates of the times series data. This research can be applied to a variety of fields. Here we apply it to hourly natural gas flow recordings used by natural gas utility companies. An algorithm is introduced to identify and remove outlying hourly flow recordings either in an existing set of hourly flow recordings or in near real time. Then this research is applied to generate hourly flow forecasts for natural gas utility companies, and the accuracy of the forecasts is compared to the forecasting accuracy of an existing hourly flow forecasting software, GasHour. We concluded that the proposed algorithm can generate more accurate hourly forecasts in each of three cases tested.