It takes more than an expensive software package and top-of-the-line workstations to get the most out of FEA.
|An engineer with limited FEA experience and no library of best practices to guide him tried to mesh a thin-walled bike frame with solid elements rather than using more appropriate beam elements.|
|Most design costs are committed during preliminary design, even though little is typically known about a design at this stage, according to a study by Boeing during the 1960s. As design moves from concept towards production, redesign costs skyrocket while flexibility plummets.|
|Switching to a concurrent-engineering process and implementing FEA gives designers greater freedom and control over costs and the final product than do more traditional serial engineering processes.|
John P. Leahey
Pro Specialists Inc.
Why do some companies succeed and others fail when making FEA part of their design process? Surveys reveal several obstacles that make it difficult to get FEA up and running, including serial rather than concurrent product-development processes, lack of management commitment, and inadequate user skills. Even organizational size contributes.
These hurdles and their solutions sort themselves into three major elements required for a successful implementation of FEA: adoption, adaptation, and retention. Whether planning a FEA implementation or analyzing where one went wrong, these elements provide a useful organizational framework.
Why adopt FEA?
The idea of concurrent engineering (CE) is to front-load the concept and preliminary design process with as much information as possible, such as functional characteristics, manufacturability, and serviceability. A 1960s study by Boeing determined that most design costs are committed during preliminary design. As designs progress from concept to production, redesign costs skyrocket and flexibility for change plummets. Evaluating several aspects early in the process lets users consider their effects when changes would have the greatest impact at the least cost.
In many companies using the more traditional serial-engineering (SE) process, analysts cannot evaluate or optimize every product component due to time constraints. This often forces engineers to overdesign parts in an attempt to account for functional unknowns, resulting in suboptimal products.
To realize the full benefits of CE, it is essential to incorporate FEA into the preliminary design stages. Yet these early stages are largely the domain of nonanalyst design engineers -- staff that often does not have the background to successfully run FEA tools.
Design engineers inexperienced in FEA risk building unnecessarily complex models that will not run, or making incorrect modeling assumptions that yield invalid results. These common mistakes often lead to failed designs and abandoned CAE tools. How can companies avoid these pitfalls? By carefully planning the implementation of FEA based on the adoption, adaptation, and retention framework.
Adoption refers to an organization's commitment to implementing the technology, including end-user commitment and management buy-in with a realistic view of the costs. Purchasing CAE software and investing in training is the first step. Unfortunately, CAE vendors often neglect to mention the additional assistance necessary for long-term success. One vendor for example, has a ROI calculator on their Web site. Presumably, it is to help prospective customers see how quickly they can recoup their investment in FEA tools. Costs of training and implementation assistance are not included in the calculations, giving customers false expectations and setting the stage for failure.
Compounding the problem, most FEA training focuses on navigating user interfaces rather than analysis. Training alone does not ensure the proper use of FEA. Subsequent implementation consulting, including assessment of onsite engineering expertise and developing analytical best practices, is generally necessary for successful adoption and retention of FEA.
After getting the right technology and training, companies must adapt product-development cycles to incorporate the technology. This includes changing design practices from SE to CE. Far too often, having made the commitment to adopt the technology, companies will continue with the same design processes and simply replace a few preproduction prototypes with digital versions. While this may save money, the company will not obtain the full benefits of FEA. In fact, implementing FEA in an SE environment can negate ROI assumptions, potentially leading to abandonment of FEA.
Once a company has cleared the adoption and adaptation hurdles, they must ensure FEA knowledge is actively developed and retained. This calls for building and documenting a knowledgebase of FEA models and techniques correlated to physically tested products, keeping practitioners up to date through continuing education and seminars.
Problems at large companies
Companies with designers and analysts are not likely to abandon FEA, but they can end up using it inefficiently. The challenge for these organizations is to put FEA into the hands of the design engineers during the conceptual design phase, fostering effective concurrent engineering. Typical impediments to this shift include turf wars, minimal management buy-in, lack of designer analysis skills, and job security fears on the part of full-time analysts.
An inefficient design process stalled one large company even before it adopted CAE tools. In 2001, International Truck & Engine Corp. launched its latest medium-duty truck 20 years after its previous version. Their sequential engineering process was largely to blame for this long development cycle.
To streamline product development, they made the plunge into CE and prepared to buy several CAE tools, including FEA. They wanted FEA done as early as possible to get the most from CE, but they were aware nonanalysts might have a hard time solving complicated FEA problems. To resolve this, they devised a two-tiered approach: engineers would perform component-level analyses (with analyst supervision), while analysts handled the system-level studies (such as overall load calculations, vibration, and fluid studies).
To put FEA safely into the hands of part-time users, International Truck chose a focused FEA tool. It allowed programming the analysis process for design engineers so they can only make a few selections but still get accurate and useful results. As designers become more proficient at FEA, the narrowed pathways in the software can be expanded. This let designers test ideas early in the product-development cycle while limiting the potential for missteps.
Finally, International Truck employed retention strategies to preserve on-site know-how about the technology. Analyst oversight and mentoring of designers' simulation work helped develop a best-practices database. As company analysts correlated results of digital and physical prototypes, they built confidence in CAE tools and developed a valuable knowledgebase of analytical methods for future projects.
Smaller companies have problems too
Smaller companies are often "design engineer only" environments, where engineers are responsible for everything from concept through production. A typical project engineer might require FEA early in a project then spend months focusing on other development issues.
Due to the lack of analysts, smaller companies must somehow gain the analytical knowledge required to run CAE tools, as well as determine and document analytical best practices. Avenues open to smaller companies include hiring an engineer versed in FEA, learning and documenting FEA techniques on the job, or outsourcing to a consultant who can derive and document the best analysis practices for the organization.
Schwinn Cycling and Fitness tried two of these three methods. For its first 90 years, the company was an innovative market leader. Then in a miscalculated move, it delayed entering the mountain bicycle market thinking it was a fad. Schwinn quickly lost market share to its competitors and declared bankruptcy.
Scott Sports Group brought Schwinn out of bankruptcy and was determined to return to innovation. They invested in a state-of-the-art 3D CAD system, mechanism-analysis software for suspension engineering, and structural FEA software for durability and weight optimization analyses. While the mechanism and FEA tools were similar, their implementations were not equally successful.
With design innovation a priority, management made a firm commitment to adopt the mechanical analysis. Schwinn engineers adapted their suspension design practices to be concurrent, and all new suspension designs started with mechanism analyses. Detailed design started only after optimizing the suspension layout.
To retain experience in-house, Schwinn built a knowledgebase of best modeling practices for Schwinn-specific designs. Computer results were correlated to laboratory and rider testing results. After a few product iterations, it became easier to predict how new suspension designs would ride or feel based on computer data alone.
Things didn't go as smoothly with FEA. Schwinn continued to rely on SE methods -- testing physical prototypes prior to production. This slowed their development process and often caused additional problems. For example, in a rush to introduce a new mountain bike, they cut short its testing. The bike was released prematurely which lead to a costly recall to replace fatigue-prone components.
Prodded by the ineffectiveness of physical testing, motivated by the recall, and reassured by their success with the mechanism-analysis software, Schwinn renewed its commitment to adopting structural FEA. Knowing they lacked FEA experience, they hired a design engineer familiar with their software. The FEA implementation began by building a retainable analytical knowledgebase prior to adapting to a concurrent-design process. The engineer determined best modeling and analysis practices for the thin-walled components that dominate the bicycle industry, and put in place an analysis documentation procedure. Unfortunately, the engineer left Schwinn before the library was formerly documented. Products still had to get out the door, but without an engineer with analysis skills, FEA again took a back seat to prototype testing methods.
A year later, at the beginning of a new design cycle, Schwinn again renewed its commitment to FEA and sent their entire design staff to training. Eighteen months later, Schwinn released the Fastback.
After the bike's release, the company looked at their design processes and the bike to see what impact FEA had made. A survey of athletes riding the new design found that while the bicycle frame was light, it was too stiff. In fact, it was so stiff many cyclists replaced the stock metal seatpost with a carbon-fiber alternative to reduce road shock and vibrations.
It turned out that the engineer responsible for the design attempted to mesh the thin-walled bicycle tube with solid elements, even though more appropriate techniques were available. Meshing with solid elements generated a huge model that took so long to solve, the engineer abandoned FEA completely and relied on make-and-break techniques to get the bike out the door. Thus, a less than optimal bike was Schwinn's first foray into a competitive market segment.
While management committed to adopt the technology when they invested in hardware and training, Schwinn's implementation failed during the adaptation and retention phases. To be successful, Schwinn needed to continue building and documenting an analytical best-practices knowledgebase so engineers would know which analytical method to use when evaluating radically different designs. Because the company was unable to implement FEA internally, they should have outsourced development and documentation of best practices. Soon after release of the Fastback, Schwinn was forced to declare bankruptcy again. Pacific Cycles has since bought the Schwinn brand, and most engineering is done overseas.
Innovation often drives companies to seek competitive advantages promised by FEA. Unfortunately, a pioneering spirit alone does not guarantee a successful implementation and the competitive advantages that come with it. Switching to a concurrent-engineering design process and placing FEA in the hands of nonspecialists requires strategies to ensure long-term success. The adoption, adaptation, and retention approach provides a framework to ensure effective integration of FEA and its positive effects on product design and development.
Pro Specialists Inc., (303) 443-2745, www.prospecialists.com