Validating computer simulations

Imagine this: an engineer develops a new product for the automotive industry using state-of-the-art computer simulation programs. After conducting an exhaustive series of modeling exercises on screens that show the product operating correctly, the company begins manufacturing the product. But after the first few hundred are shipped to customers, a problem arises. Customer calls suggest that the product does not work as indicated by the hundreds of computer simulations. The result? Hundreds of returned products, lost revenue and a plethora of potential legal issues.

Fortunately, this elementary example does not happen often in U.S. industry. Although technology (such as computer simulation programs) allows engineers to make important decisions about the operation of a product before it's produced, reducing production time significantly, companies "must also validate that the product operates as expected," explained Henry Kowalski, or Doc K as students affectionately call him.

"Computer modeling is a great tool for engineers, but until we produce a working example of the product and test it, we cannot determine if the actual product will behave according to results obtained from computer simulations," added the professor of ME and director of Kettering's Experimental Mechanics Laboratory.

This is why testing of a computer model to validate the simulation and/or the function of the product is essential. And although technological improvements in testing and validation help in this endeavor, Kowalski said that time-tested methods of experimental stress analysis and validation are important steps engineers can never forgo. More importantly, the hands-on approach of testing and validation is in his view a necessary part of engineering education.

But this has not always been the case in past years. In the early 1950s, all Kettering/GMI students took a course in testing and validation, also called experimental stress analysis (ESA). At that time, Kowalski said conventional wisdom "required students to develop knowledge and the ability to apply practical skills in the science and art of ESA."

However, in the 1980s, the engineering field shifted toward increased use of analytical tools like computer-aided calculations, which reduced the dependency on prototype testing. Thus, ESA went from a mandatory course to an elective as changes to the curriculum evolved according to what was happening in the industry. The result was the incorporation of what Kowalski defined as "virtual engineering that replaced the reality of hands-on experience."

In the midst of these changes at Kettering and in the industry, computer simulation packages began replacing basic engineering topics. "Over time, we realized that our graduates were deprived of the knowledge and skills necessary to perform rudimentary analysis of the products they created in a virtual environment," Kowalski said. Kettering responded to this issue by reintroducing basic instrumentation and technologies to provide students experience with the physical testing of products.

Today, students take an extensive course titled MECH 514: Experimental Mechanics. This course and related work in the Experimental Mechanics lab introduce students to the utilization of strain measurement techniques and the importance of identifying and measuring physical quantities such as displacement, weight, load and temperature during product testing and validation.

Today, the lab, located in 1-231 of the Academic Building, continues to serve as an integral resource where student teams undertake term projects to test products designed using computer simulation programs. The goal is to determine whether or not these products will achieve the desired operation levels depicted by the simulation software.

Companies are beginning to understand why such a lab at Kettering is necessary. Several years ago, for instance, ArvinMeritor Technologies pledged $100,000 over a five-year period toward the development of this lab following a persuasive presentation by Kowalski regarding why students should engage in hand-on testing and validation processes. Since this initial contribution, Kettering has purchased testing equipment for the EM lab and leveraged this funding to obtain additional investments.

Image removed.

The success of the lab and reintroduction of Experimental Mechanics is substantial, Kowalski noted. In 2004, for example, a student group designed, constructed and successfully tested a device for indirectly measuring the force exerted on the spine of an athlete when incorrectly performing a bench press exercise.

The results of this project could aid a fitness equipment company when developing a new bench press. Other projects undertaken by student groups include stress analysis of a prosthetic hip implant. In this project, students examined the affect of bone degradation on the strain in the femoral component by varying the level of fixation of the prosthesis.

This team concluded that the strain in the prosthesis was minimal in the areas that were fixed, which shows that as long as the prosthesis is completely fixed in bone cement, strains are minimal. Thus, the life of the prosthetic life is greatly improved.

"No company has ever refused participating in a student research project," Kowalski said. "Head Ski Company, for example, sent us different types of K2 skies for vibration testing when they were made aware of a student research project. These companies want us to help for several reasons: first, the research is free or of a nominal charge and second they receive important results to help them produce products of better quality in the future."

He has taught at Kettering for more than 40 years and is one of the longest tenured professors at the institution. Although his enthusiasm for the lab and what the students can accomplish in the facility is contagious, he recognizes a hard truth: the need for future funding to help students continue their testing and validation work. The EM lab requires new state-of-the-art technology among other things. "As I get closer to retirement, I would like to attract supporters to the school, who could possibly help start an endowment of $2 million to support student projects," he noted.

In recent years, Kowalski has parted with approximately $200 of his own salary for each student team to obtain the materials they need to conduct their research for his course. He recognizes that there is clearly a desire on the part of students to take the course and conduct the research, as well as a desire from companies that seek assistance testing and validating their own products.

To learn more about the Experimental Mechanics Laboratory, or to inquire about supporting the lab, please contact Dr. Henry Kowalski at (810) 762-9926, or via e-mail at hkowalsk@kettering.edu.

Written by Gary J. Erwin
(810) 762-9538
gerwin@kettering.edu