Graduate student David Agu and Associate Professor Matt Campbell with the award, GraphSynth Imaging Software, and leaf blowers.
University of Texas ME graduate student David Ikechukwu Agu was awarded the Shell Research Prize for the best paper and presentation in the area of Energy/Environment/Transport Processes at the Graduate And Industry Networking (GAIN) 2009 conference. His project was entitled "Automated Analysis of Product Disassembly to Determine Environmental Impact." It was selected as one of twelve papers presented at the conference on February 17, 2009 out of fifty submissions. The paper and accompanying poster were co-authored by Dr. Matthew Campbell, an associate professor in the area of Manufacturing & Design. The GAIN conference is an annual event held at The University of Texas which allows graduate students to present their research to their peers within the Cockrell School of Engineering as well as to visitors from industry. Visit the GAIN website for more information.
The Project
Manufacturers are increasingly being held responsible for the fate of their products during the end-of-life phase. In this research, a product's end-of-life environmental impact is calculated as a function of cost and recyclability. Cost was determined to be the time taken to disassemble a product and the value of the materials in the product assembly. Recyclability is a measure of the amount of material separated from the assembly and ready for recycling and is either referred to as the recovered weight or inversely as wasted weight. The paper presents a computational approach which will analyze any product given information concerning individual components and how they are connected within the overall assembly. The result of this analysis is a set of Pareto optimal candidates representing various stages in the disassembly process and evaluated using associated costs and recyclability. This Pareto set can be used to judge a product's end-of-life suitability against the manufacturer's or industry standards, or against the suitability of a similar product. The real innovation in this project lies in the computer's ability to discern how best to disassemble a product given only its exploded view.
Motor from the Troy-Bilt TB190BV
Click image to enlarge
The Product Disassembly Problem
Depending on the complexity of a product's assembly there are a number of different orders in which it may be disassembled and the total disassembly time varies for these differing orders. It would be advantageous to know the particular order which simultaneously minimizes the total disassembly time and the wasted weight however given the potentially large number of disassembly sequence alternatives, determining the optimal sequence without a suitable evaluative method is difficult.
Product Assembly Representation
Dr. Matthew Campbell and the Automated Design Laboratory had previously developed a program called GraphSynth for use in automated design projects. GraphSynth provides a suitable tool for representing product assemblies and analyzing their disassembly.
Within GraphSynth, graph-based representations of product assemblies are created. Nodes represent components within products while arcs, which connect nodes to each other, represent the connections between components. In order to accurately model product assemblies, it was necessary to investigate each of the ways components can be connected. To date the connection types which have been evaluated fall into one of the following categories: rectangular constraints, radial constraints, press fits, threaded constraints, adhesives, modules and welds.
One of the main ideas behind the research is that if a product's environmental impact can be determined, it can be compared with that of a similar product. In this case two similar leaf blower models were chosen: the Toro Power Sweep #51586 and the Troy-Bilt TB190BV.
Author Information
David Ikechukwu Agu is currently pursuing a Master's degree in mechanical engineering with an anticipated graduation date of May 2009. He works in the Automated Design Laboratory under Dr. Matthew Campbell.
David Agu's winning poster at the GAIN 2009 conference
Click image to enlarge
