Many cars are designed today as digital mockups long before physical prototypes hit the test track. That way engineers can analyze virtual parts before stamping metal and building the full assembly.
Ford uses software tools like EDS' I-DEAS for CAD design and Mathworks' SimuLink for analysis to create vehicles like the Escape SUV.
But when they began to design the Escape HEV (hybrid electric vehicle) a gasoline and electric version of the original, they hit some challenges—they couldn't predict how all the virtual parts would work together in the full assembly.
When it hits production at the end of 2003, the new Escape will be the first hybrid SUV, Ford says. But they didn't want to rush the product to market. Their goal for the Hybrid Escape was to create a "no compromise" hybrid-electric vehicle, so drivers wouldn't notice the difference between the standard and hybrid electric vehicles, says Michael Tiller, a technical specialist in Ford's Powertrain and Vehicle Research Laboratory.
The new car uses regenerative braking to store energy normally lost during deceleration (Tiller won't say whether it's stored as electric energy or pressurized hydraulics). It will automatically turn off its internal combustion engine at a stoplight, and use the stored energy to accelerate.
Those techniques work great for saving, but one of the biggest drawbacks of a hybrid vehicle is that drivers get annoyed when they notice the engine stopping and starting, Tiller says.
Ford had software tools written in Fortran and C++ to aid in analyzing various aspects of the powertrain individually, such as combustion in the cylinders, wave dynamics in the manifold, and mechanical systems in the transmission. But analyzing the system as a whole is a much more difficult problem. "We needed a combined engine, transmission, and simple vehicle model, to see what vibration feels like when it gets to the driver," he says.
So Tiller uses Dymola from Dynasim (Lund, Sweden), a multi-do-main modeler designed to simulate several types of physical stress at once. Once it has a model to solve, Dymola streamlines the process by generating efficient simulation code.
Another hindrance to Ford's overall productivity was trying to pull results together, especially during optimization studies, which involves exhaustively varying parameters within a range, sequentially. So another way Tiller speeds the process with Dymola is to use a standard library of components; electrical, mechanical, hydraulic, and Ford's own thermodynamics library. "So our models resemble the schematics of the actual system," he says.
Dymola can do that because it reads-in Modelica models from many sources: third-party libraries, companies' internal databases, and the Modelica Association, the industry consortium that created the language. "We can incorporate models that other people have developed, quite easily, because the Modelica modeling language is non-proprietary. When we write Modelica code, it's done in an open way," he says.
For more information about Dymola from Dynasim: Enter 534