By Bruce Jenkins, President, Ora Research
Moving from wind tunnel testing of physical prototypes to simulation-driven design processes can offer typical automotive OEMs more than 500% ROI (return on investment), according to research from Tufts University’s Gordon Institute for engineering management.
In a project sponsored by CFD software developer Exa Corp., the Tufts team analyzed the cost-benefit of deploying digital prototyping and simulation software to replace the physical prototypes and test procedures conventionally used in design, development and validation of vehicle aerodynamics, thermal management and aeroacoustics.
Through surveys of automotive engineering executives, wind tunnel experts and digital simulation technologists, the Tufts researchers quantified ROI of simulation-driven design for three different categories of automotive OEM:
• Most conservative—146% ROI (1.5X gain)
• Most likely—531% ROI (5.3X gain)
• Most inclusive—1209% ROI (12.1X gain)
Most conservative—Automotive engineering organizations in this category already use simulation software extensively. They are not heavily invested in physical test infrastructure and do not use physical prototypes or tests in any instance where this is not mandatory. Thus, for these companies, ROI available from increasing the use of simulation and eliminating the few remaining prototypes and tests is comparatively low—even though an almost 150% ROI remains noteworthy and worthwhile.
Most likely—The ROI calculation for this category is based on typical or average industry costs for prototyping, testing and simulation. While significant variation exists across the major automakers, this ROI measure approximates the industry average.
Most inclusive—Automakers in this category will see the greatest benefit from moving to simulation-driven design. They can avoid the investment costs for a new wind tunnel and its upkeep, use simulation software for design optimization to reduce part costs in high-volume models, and avoid costly late-stage changes that are likely in the absence of a robust simulation-based development process.
ROI model
ROI measures the amount of return on an investment relative to the investment’s cost. To calculate ROI, the net benefit of an investment (gain minus cost) is divided by the cost of the investment, and the result is expressed as a percentage.
In the Tufts project, the gain from investment was defined as the physical prototyping and testing costs that will no longer be incurred, plus the additional gains (or losses) that result from using simulation. Thus, the ROI formula used in the study was:
ROI = (Cost of prototypes and tests + Additional gains or losses – Cost of simulation) / Cost of simulation
Cost of prototypes and tests consists of:
• Cost of prototypes—Cost of all the required prototypes built for aerodynamics, thermal and acoustics tests in the design and development process that will no longer be incurred upon transition to a fully digital design process.
• Cost of tests—Cost of all the required tests for aerodynamics/thermal/acoustics in the D&D process that will no longer be incurred upon transition to a fully digital design process. These tests can be either done in-house or outsourced.
• Test facility investments—Investments in in-house aerodynamics/thermal/acoustics test facilities, plus the costs to maintain and upgrade them.
Additional gains or losses consist of:
• Design optimization—With advanced simulation software, products can be optimized to reduce cost and improve performance. There have been many successful cases, such as using aeroacoustics simulation to reach a low noise level without the costly laminated glass originally required for this.
• Late-stage changes—These arise from issues with the design that are discovered during testing and must be corrected after the design is completed or nearing completion. Compared with simulation, the iteration cycle of building a prototype and sending it for testing is much longer; thus, late-stage changes are more likely to occur with physical testing. To preserve program schedule, late-stage changes usually come with high retooling costs and/or an increase in the cost of production parts.
• Test deviation—Factors such as manufacturing deviation, transportation damage, test repeatability and reproducibility all influence test outcomes. As a result, the actual number of prototypes built to verify product design is typically more than necessary.
• Warranty costs (unquantifiable)—If a problem is not found through testing, it may eventually lead to quality or safety problems after consumers have purchased the automobile. This leads to warranty problems resulting in repair or recall.
• Styling feasibility (unquantifiable)—Early-stage simulation enables product designers (stylists) to create styling themes more flexibly, balancing tradeoffs between design attributes necessary to meet aerodynamic, thermal and acoustic parameters, and attributes that the styling team considers most attractive and appealing to customers.
• Performance and perceived quality (unquantifiable)—Better aerodynamic, thermal and acoustic performance will yield increased customer satisfaction and potentially attract more customers.
• Effectiveness and efficiency in communication (unquantifiable)—Simulation results are much more easily processed into clear, informative visualizations than are physical test results. Such visualizations can reduce misunderstandings and improve communication among different functional teams, as well as reduce rework and design cycles.
Cost of simulation consists of:
• Cost of licensing—Cost based on the use of simulation software for aerodynamics/thermal/acoustics, measured in CPU hours.
• Cost of computing power—Accompanying costs for implementing the software, including the investment in IT infrastructure necessary to run the software.
• Cost of training—Training courses to teach engineers how to use the software, and to keep users familiar with new features and new releases.
Qualitative benefits outside ROI model
In addition to the quantifiable gains captured in their ROI model, the researchers found important qualitative benefits in moving from wind tunnel testing to simulation-driven design:
• Wind tunnels do a poor job of reproducing real-world road and environmental conditions, and thus fall short of simulation in accurately predicting product performance in use.
• Physical prototypes give feedback on performance, but do not provide the insights that simulation does into how to improve a design.
• Studio designers (stylists) and engineers need to collaborate early in the design process to evaluate and refine the performance of their proposed designs. Simulation supports this because it can begin much earlier in design than physical testing.
• Working with wind tunnels and clay models is a rigidly sequential process, thus much less fluid and flexible for iterative design investigation and optimization than simulation-based workflows.
Source: Aly K., Costa A., Garreffi M., Yu H.; advisor Liggero S. 2015. ROI Analysis of Simulation-Driven Design. Medford, MA: Tufts Gordon Institute.