Predictive modeling of solid waste generation for aggregate building and material types across geographical contexts

Solid waste generation is increasing at alarming rates, globally. Challenges to decreasing solid waste generation and landfill disposal are widespread, and U.S. Army installations are a unique basis for providing qualitative and quantitative data from a breadth of geographical locations. The purpose of this study was to assess the modeling prediction capability of solid waste streams from data of 12 Army installations at the aggregate material and building type level. Solid waste generation data was collected by quantifying materials found that have potential for diversion (e.g., source reduction, ruse, recycling, composting, etc.) and are currently being sent to landfill. In coordination with key personnel, buildings were selected that were representative of the main activities conducted at each of the installations. These buildings represent 28 different building categories as defined by the System Master Planning classification tool. Over the period of one week, 100-pound random samples from dumpsters at selected buildings were obtained for each installation studied. Materials were manually separated into 22 categories, weighed, and recorded. Results from the study identified considerable amounts of materials with value and diversion potential in the solid waste stream. A total of three building types and five material types were down selected for model construction and validation based on robustness of data available and applicability outside military contexts. Models were constructed for each material and building type combination to avoid error with multiplication factors of coefficients for each independent variable. Results showed statistical significance (p-value = 0.05) for 12 of 15 modeling combination predictions, indicating that these 12 models for each material and building type are uniquely capable of predicting solid waste generation. P-values for the 12 significant models ranged from 6.94e-07 to 0.033. Each of the 12 statistically significant models differed in R-squared and adjusted R-squared values, ranging from 0.823 to 0.997 and 0.764 to 0.996, respectively. This study provides a unique data source demonstrating the ability to use predictive modeling to forecast solid waste generation at the aggregate building and material type level. Using Army installations as a case study may increase data available across the continental U.S. to focus targeted source reduction efforts.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Agricultural & Biological Engr

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2023

Author dc:creator

• Rice, Abigail Marie

Contributors dc:contributor

• Davidson, Paul C

• Diversion
• Waste
• Predict
• Model
• Resilience
• Generation
• Reduction

Statement dc:rights

• Copyright 2023 Abigail Rice

Language dc:language

en, eng

Handle dc:identifier

https://hdl.handle.net/2142/120154

OAI record oai:identifier

https://www.ideals.illinois.edu/oai-pmh?verb=GetRecord&metadataPrefix=etdms&identifier=oai%3Awww.ideals.illinois.edu%3A2142%2F120154