Transitioning from a traditional build-and-test process using physical prototypes to a new approach that tests designs virtually will greatly reduce costs and speed time to market.

It is often the case that the simpler the product, the greater the complexity surrounding its design. That is certainly true at Dana Holding Corp.’s Sealing Products Group, which manufactures top-to-bottom engine sealing systems for the automotive sector, including cylinder head covers, all kinds of gaskets, and valve stem seals. The division’s products are highly engineered and have to perform consistently over a protracted period in a dizzying array of unique environmental and system conditions. Because of that scenario, the process calls for extensive testing and verification of designs, yet the group could hardly be cost-effective and time-efficient if it relied exclusively on a build-and-test process involving physical prototypes.

With that in mind, the Sealing Products Group set out several years ago to transform its traditional processes and pursue an analysis-led design approach. Analysis-led design entailed standardizing on a well-developed, integrated computer-aided engineering (CAE) software portfolio and a set of engineering processes so that the team could design products properly up front, test them virtually to optimize performance, and cut back on the number of physical prototypes it had to build. Over time, the new design approach would help Sealing Products Group vastly reduce development costs, not to mention pushing product out the door faster.

The importance of an analysis-led design strategy becomes clear when one considers the competitive landscape the Sealing Products Group faces. Regulations for emissions and fuel and noise control are evolving continually and becoming more stringent, forcing a new set of requirements for the exhaust system along with the associated sealing components. Despite the increasing complexity of product requirements, the number of global players vying in this territory is on the rise. Thus, the analysis-led design transformation was viewed as a key survival strategy for differentiation, a path that would help the Sealing Products Group meet its goals of accelerating time to market and achieving consistent high-quality performance across its broad spectrum of products.

Despite the technical limitations of both the hardware and the CAE software when the journey began nearly a decade ago, the team was confident that it had embarked on the right course. Because its products are all about forming, contact, and assembly, they are highly nonlinear, making it difficult to predict stress and strain performance by using traditional prototyping techniques or even to track the results with manual tools such as spreadsheets. In terms of durability, one also is looking at extremely long testing setups and cycles. It was not uncommon to put a design through 1,000 hours of dyno testing only to find a problem when the tests were complete; as a result, the design had to be scrapped and the process restarted. On simpler designs, this kind of iterative test and verification approach might have been workable, but with the increasing complexity of emissions requirements and the need to reduce fuel consumption, the Sealing Products Group had no choice but to turn things around much faster while vastly reducing the amount of time and effort dedicated to testing.

SIMULATION IN THE EARLY DESIGN STAGE

Reorienting an engineering organization to support analysis-led design does not happen overnight. It is a gradual process that has been under way at the Sealing Products Group for years. Although early simulation applications and computer hardware put limitations on analysis-led design, the original goal was always to achieve 100% simulation before prototyping. The group is not quite there, but advances in technology are allowing it to get close. The group is working toward a system simulation approach coupled with individual component simulation, and today multidisciplinary CAE groups are leveraging simulation tools from a variety of companies to analyze and test designs for everything related to structural and thermal stress, durability, tribology, and computational fluid dynamics (CFD). In addition, the use of CFD and finite element analysis (FEA) has expanded over the years and now is applied to a wide range of applications in the base engine area as well as in other powertrain components.

The first leg of the journey was to identify the main outputs that could be simulated without taking into account any specific process flow. At this early stage it is necessary to determine whether it is preferable to simulate and test for durability or instead to evaluate the stress and strain of a product to avoid overdesigning from a materials standpoint. It helps to have a highly-trained CAE analyst involved in this step as well as test engineers who can develop the material properties required for the simulation.

Mapping out process flow is the next logical step. As manager of advanced engineering, I took on that key role, and the process required a couple of months. Instead of hunkering down with consultants behind closed doors, the start-up team involved engineers early on to get their input on defining the processes and identifying areas in which simulation might fit. Taking a bottom-up approach instead of mandating process changes from top management goes a long way toward ensuring buy-in and getting the broader engineering community to adopt the new practices.

Once the group became accustomed to the changes, no one questioned the CAE tools or the analysis-led design approach. In fact, it was at that point that engineers began looking for additional areas in which to apply simulation. Rather than let simulation expand organically, the team realized that it needed to organize and manage its growth. Simulation islands had begun to spring up, with different product teams using different software and different approaches that did not allow them to collaborate and reuse the results easily. An incredible amount of data was being created from the ongoing simulation efforts, and there needed to be a cohesive way to store the information and make it widely accessible.

At the same time, other engineering groups within Dana were facing similar challenges. At that point the global CAE teams within the engineering group decided to create a CAE Council, a formal body that would establish a standardized set of software and simultaneously create best practices so that the company as a whole would approach analysis-led design in a consistent manner. Established around seven years ago, the council meets at least once a quarter either virtually or face to face. The CAE Council defined a core portfolio of six CAE packages, including Abaqus from Dassault’s SIMULIA brand and Altair Hyperworks, and has nurtured partnerships with those vendors that give Dana input into how the tools will evolve. The CAE Council also is steering future directions around analysis-led design, including a pilot Simulation Lifecycle Management (SLM) project spearheaded by the Sealing Products Group that is intended to create a one-version-of-the-truth repository for all the simulation data.

Years into the journey, the biggest lesson learned is patience. Transitioning from many CAE tools to a standardized suite takes time; in fact, it took the Dana group nearly three years to make the switch, and during that time it had to incur the cost of some overlap in licensing fees. There was also a time lag before all the divisions were on the same renewal cycle for hardware upgrades, which was necessary for the company to maintain a consistent hardware platform to run the power-hungry simulation software.

Although the Dana Sealing Products Group is very close to its goal of 100% simulation for testing, that goal remains a moving target. The more simulation is implemented, the more possibilities come in sight. There may be no official end to this journey, and the Dana team has never lost touch with its well-defined strategy and the mission it envisioned for analysis-led design.