Multidisciplinary optimization of ground vehicle dynamics

Pratt & Miller Engineering evolved from a small business focused on designing and building race cars into a full-service engineering and low-volume manufacturing company serving a global customer base in the defense, automotive and powersports industries. After evaluating multiple optimization tools, Pratt & Miller selected Red Cedar Technology’s HEEDS MDO and its SHERPA algorithm as the only optimization technology that could solve its highly constrained models.

Pratt & Miller’s typical automotive optimization studies are set up as nearly overconstrained problems. The firm usually needs to minimize or maximize only one or two objectives, but often needs to satisfy up to 50 constraints. Prior to using HEEDS MDO, Pratt & Miller would plan for four to 12 man-weeks to find a solution to its typical problem. This was in addition to the time required to generate the baseline model using a combination of engineering intelligence and DOE methods. Using HEEDS MDO and parallel processing, this time is typically reduced to one to two weeks.

In the past, Pratt & Miller structured simulation investigations primarily around DOE studies performed using MSC Adams/Insight. However, an assumption of the model order had to be set before the simulations were started, and the response surface did not always conform to that order, resulting in a poor fit. When designing new vehicles from the ground up, the parameter variations can be very large, and even a cubic model does not fit very well to the actual response surface. HEEDS MDO removes this problem by operating on the actual response values and determining how the simulation should progress, explains Pratt & Miller, and it is very easy to exclude failed simulations from the optimization process in HEEDS.

Pratt & Miller’s other problem with DOE simulations was the large number of evaluations. The firm often ended up doing 21,000 evaluations for a cubic model and 1500 evaluations for a quadratic model. “Using HEEDS MDO with SHERPA, we can find a good solution in 400 evaluations, and often less,” says CAE Manager Jesper Slättengren. “HEEDS allows us to study the progress of the optimization while it is running. We can stop the simulation at any point when the results are close enough to the desired target. We can also extend the run beyond the initial number of runs without any loss of optimization progress.”

Finally, Pratt & Miller struggled to manage multiple simulations. In a typical scenario, it would conduct six to 14 different simulations for each model, and the optimal design was dependent on all of the simulations. This required six to 14 different DOEs with the same factors and different responses. It also required a polynomial optimization method that could work over the combined response surfaces from all the DOEs. Incorporating HEEDS into the design process has completely eliminated this problem, the company reports.

Optimizing chassis components for military vehicles

pratt1
Source: Pratt & Miller Engineering

HEEDS MDO is used daily at Pratt & Miller to optimize chassis component parameters to meet specified military requirements for heavy vehicles. These include spring and damper parameters, anti-roll bar dimensions, and suspension geometry points. Among the constraints are understeer gradient, ride-quality measures, and spring and damper balance. Typical problems include up to 50 factors and as many constraints. Using MSC Adams with HEEDS MDO parallel processing, Pratt & Miller is able to produce reliable results overnight.

pratt2
Source: Pratt & Miller Engineering

The development process for military vehicles differs significantly from that of passenger and commercial vehicles, including different ride and handling requirements that must be satisfied. For example, the ride specification includes passes over half-rounds at various heights during which no occupant can be exposed to more than 2.5G vertical acceleration. For stochastic roads, the absorbed power must be less than 6W. For each of these events, a threshold and an objective speed are specified. Handling requirements include a range for the understeer gradient and a maximum roll gradient. Besides these explicit specifications, a good ride and handling balance is desired.

pratt3
Source: Pratt & Miller Engineering

With tougher requirements each year, it gets progressively harder to tune the vehicle suspension. With decreased development times and longer lead times for components such as springs and complex damper systems, the need for multi-event or multidisciplinary optimization is increasing. By coupling MSC Adams and Adams/Car with HEEDS MDO to perform the multidisciplinary optimizations, Pratt & Miller designed, developed and built a new vehicle design in 12 weeks, with the suspension element taking less than six weeks.

pratt4
Source: Pratt & Miller Engineering

“We believe in HEEDS and have seen tremendous benefit as we’ve applied it on several internal programs with great success,” says Pratt & Miller’s Mark Palmieri. “HEEDS integrates well with our internal tools, as well as with the MSC products we represent and use. Because of our confidence in this technology, we have committed to reselling it and working closely with Red Cedar Technology to share it with both our customers and prospects.”