ADVERTISEMENTS
Learn about diversity at ASEE
ASEE would like to acknowledge the generous support of our premier corporate partners.

JEE SELECTS - Research in Practice

JEE Selects

Design in the Wild

A case study reveals how engineers integrate mental models and simulations to tackle complex problems.


By Joshua Aurigemma, Sanjay Chandrasekharan, Nancy J. Nersessian, Wendy C. Newstetter


Researchers have long had an interest in the processes of engineering design, particularly how a design engineer makes progress under conditions of uncertainty and constraint to arrive at a solution. Over the years, laboratory studies have yielded valuable insights into the problem-solving steps of both expert and novice designers. However, nothing can replace direct observation of an engineer working to solve a significant design problem, or “design in the wild.”

In our longitudinal case study, we observed and chronicled the reasoning and problem-solving processes employed by an engineering master’s student as she designed a micro-fluidic device to measure the response of blood T-cells to a chemical stimulant over time. Such “labs on a chip” bring together many complex experimental processes and include such constraints as cell damage and other factors unique to biological material.

For this study, we undertook field observations, wrote field notes of the design work unfolding on the lab bench, and conducted informal, unstructured interviews with the engineer at work. At weekly lab meetings, we took notes and audiotaped her presentations and the accompanying discussions. We analyzed the engineer’s PowerPoint lab presentations, two posters she created for conferences, two of her publications, and her master’s thesis. In our analysis, we traced her iterative actions and activities as she progressed from a partial design to a fully working lab on a chip.

The engineer designed the device in an iterative and parallel fashion, revising different facets repeatedly. When design impasses loomed, she generated representations that served multiple cognitive functions. One design challenge involved determining the geometries of a herringbone sluiceway to mix the cells and stimulant. The engineer built a computational fluid dynamics model that allowed her to simulate flow patterns for different channel widths and liquids. She compared the simulation results against actual flow patterns in prototype channels, using water and sucrose to mimic viscosity. Adding a fluorescent tag to the liquids allowed her to observe and track the level of mixing. She then compared confocal microscope images of the flow against patterns generated by the model’s visualization. Once a good correspondence between simulated and actual output was established, the validated model was used to generate and test different possible geometries. The model geometries that produced the desired mixing the fastest were built and tested.

In this single design phase, we observed how the engineer utilized external models to simulate possible scenarios, visualize the device output, tag device elements to track the level of mixing, and interrogate to gain greater understanding of a local problem, making it possible for her to efficiently design her device.

This study in the wild establishes that there was an interactive process between the internal conceptual models of flow and viscosity that the engineer had gained from extensive prior electrical engineering experience and the locally generated external representations in the computer simulations. Consistent with distributed cognition theory, this integration of mental models, drawings, and prototypes formed a distributed cognitive system that constituted the reasoning process, leading to a novel object with features that supported new experiments.

This case study also clarifies the need to revise our students’ understanding of learning and problem solving to one that appreciates the bootstrapping value of distributing complex cognitive tasks across internal and external representations. Engineering educators should help students understand that sketched or built models that support simulation, visualization, and interrogation are essential to successful design. Additionally, the classroom is a place where students can begin to develop the representational fluency and flexibility exhibited in the design process by the engineer in this study.


Joshua Aurigemma is a recent graduate at the Georgia Institute of Technology; Sanjay Chandrasekharan is on the faculty of India’s Tata Institute of Fundamental Research; Nancy J. Nersessian is a professor in the School of Interactive Computing at Georgia Tech, where Wendy Newstetter is the College of Engineering’s director of educational research and innovation. The work was supported by NSF grant DRL097394084.


ASEE
© Copyright 2013
American Society for Engineering Education
1818 N Street, N.W., Suite 600
Washington, DC 20036-2479
Web: www.asee.org
Telephone: (202) 331-3500