Date of Award

Summer 2017

Document Type


Degree Name

Master of Science (MS)


Electrical and Computer Engineering

First Advisor

Povinelli, Richard J.

Second Advisor

Brown, Ronald H.

Third Advisor

Ababei, Cristinel


Short-term load forecasting is important for the day-to-day operation of natural gas utilities. Traditionally, short-term load forecasting of natural gas is done using linear regression, autoregressive integrated moving average models, and artificial neural networks. Many purchasing and operating decisions are made using these forecasts, and there can be high cost to both natural gas utilities and their customers if the short-term load forecast is inaccurate. Therefore, the GasDay lab continues to explore new ways to make better forecasts. Recently, deep neural networks (DNNs) have emerged as a powerful tool in machine learning problems. DNNs have been shown to greatly outperform traditional methods in many applications, and they have completely revolutionized some fields. Given their success in other machine learning problems, DNNs are evaluated in energy forecasting. This thesis examines many DNN parameters in the context of the short-term load forecasting problem including architecture, input features, and use of synthetic data. The performance of the model is compared against several traditional forecast strategies, including artificial neural networks and linear regression short-term load forecasting strategies. Additionally, the DNN forecaster is evaluated as part of the GasDay ensemble. The DNN forecaster proposed in this thesis offers an average 6.98% improvement in terms of weighted mean absolute percent error (WMAPE) when included as part of the GasDay ensemble. Finally, ideas for future work are discussed.