Date of Award


Document Type


Degree Name

Master of Science (MS)


Mechanical and Civil Engineering

First Advisor

Hamidreza Najafi

Second Advisor

Aldo Fabregas Ariza

Third Advisor

Troy Nguyen

Fourth Advisor

Ashok Pandit


Energy use worldwide has been increasing at an exponential rate for the past several decades. As a result, the demand for energy has put a heavy strain on the world’s climate and on existing infrastructure. To combat these issues, many researchers have been looking toward changing how energy is collected and improving how energy is used in different sectors. To achieve this, there are a variety of programs and incentives to help promote energy saving measures in residential, commercial, and industrial sectors. The goal of this thesis is to develop the building blocks for a tool that uses publicly available data from one of these government-sponsored programs, the Industrial Assessment Center (IAC), to help show the possible energy savings that may be available for an industrial plant, specifically plastic manufacturing, and water treatment plants. The Department of Energy (DOE) created the Industrial Assessment Center (IAC) program to help small and medium enterprises (SMEs) lower their energy usage and improve the efficiency of their energy usage [1]. The IAC program has maintained a publicly available database of the assessments data covering information regarding energy usage and energy efficiency measures (EEMs) to help decrease the overall energy usage. The data available under the IAC database are analyzed. Nationally, the two most common industry types in this database according to their NAICS and SIC code combinations were miscellaneous plastic manufacturing plants and water treatment plants. Using this information, a series of regression analysis and artificial neural networks (ANNs) models were developed to predict energy usage and savings given the area and annual production of each plant. One of the networks was able to achieve a relative error of up to 0.37% for energy usage, while another reached 10.9% for savings. The networks were not without error, with the worse estimate having a relative error of 1393.45%. These networks can be further improved by including weather data, technologies/techniques used to create the products, age of the plants, and by including more data points (e.g., more assessments being included in the database). The present work provides insights into the publicly available data related to industrial energy assessments in the U.S. The presented approach can be used to develop tools that can facilitate pre-audit energy analysis for the auditors and/or inform industrial facility managers about potential saving opportunities at their plants and encourage them to pursue a detailed audit.

Available for download on Sunday, December 17, 2023