Date of Graduation
Statler College of Engineering and Mineral Resources
Civil and Environmental Engineering
H.L. Roger Chen
The necessary materials for most human activities are water and energy. Integrated analysis to accurately forecast water and energy consumption enables the implementation of efficient short and long-term resource management planning as well as expanding policy and research possibilities for the supportive infrastructure. However, the integral relationship between water and energy (water-energy nexus) poses a difficult problem for modeling. The accessibility and physical overlay of data sets related to water-energy nexus is another main issue for a reliable water-energy consumption forecast. The framework of urban metabolism (UM) uses several types of data to build a global view and highlight issues of inefficiency within the network. Failure to view the whole system contributes to the inability to comprehend the complexity and interconnectivity of the issues within the system. This complexity is found in most systems, especially with systems that must be able to support and react to vacillating human interaction and behavior. One approach to address the limitations of data accessibility and model inflexibility is through the application of univariate time-series with heterogeneous hybrid modeling addresses. Time-series forecasting uses past observations of the same variable(s) to analyze and separate the pattern from white noise to define underlying relationships to predict future behavior. There are various linear and non-linear models utilized to forecast time-series data sets; however, ground truth data sets with extreme seasonal variation are neither pure linear nor pure non-linear. This truth has propelled model building into hybrid model frameworks to combine linear and non-linear methodologies to reduce the fallacies of both model frameworks with the other's strengths. This problem report works to illustrate the limitations of complex WEN studies, build a timeline of hybrid modeling analysis using univariate time-series data, and develop a parallel hybrid SARIMA-ANN model framework to increase univariate time-series analysis capabilities in order to address previously discussed WEN study limitations. The parallel Hybrid SARIMA – ANN model performs better in comparison to SARIMA, ANN, and Series hybrid SARIMA-ANN; and shows promise for research expansion with structure flexibility to expand with additional variables.
Yarrington, Cory Sumner, "Review of Forecasting Univariate Time-series Data with Application to Water-Energy Nexus Studies & Proposal of Parallel Hybrid SARIMA-ANN Model" (2021). Graduate Theses, Dissertations, and Problem Reports. 8078.