At Ford Motor Co., Jack works the manufacturing line, testing out production cells to identify possible safety hazards or ergonomic issues like awkward postures or hard-to-fit assemblies. Yet, Jack is no ordinary plant floor employee. He’s a digital human and now a core part of Ford’s virtual 3D tool set, allowing the car giant to evaluate manufacturing feasibility and address safety concerns long before physical prototypes are constructed.
Jack, from Siemens PLM Solutions, is a member of a growing class of digital human modeling applications that help companies simulate human behavior in relation to their product designs or manufacturing operations in the hopes of catching problems early on in the design cycle when it is less costly to make changes. Companies in such industries as automotive, aerospace and capital equipment are leveraging the tools, initially to design assembly operations to reduce worker-related fatigue and stress injuries associated with repetitive tasks. The software, available commercially from such companies as Siemens PLM Software, Dassault Systèmes and NexGen Ergonomics Inc., as well as proprietary functionality being designed by large OEMs and in university research labs, is also being tapped to evaluate human ergonomic and safety factors in how products are designed — for example, a tractor cabin or car control panel. As the digital human models evolve to be more anatomically correct, companies, including most of the major car manufacturers, are looking to go a step further and leverage the tools to do more realistic crash testing in lieu of physical dummies.
The reasons for simulating human factors around ergonomics and safety are much the same as doing digital mock-ups to test a design for aerodynamics or durability. It takes cost out of the equation by identifying and correcting possible problem areas upstream in the design process, as well as reduces the number of design iterations. Also, digital human modeling ensures safer, more useable products and work spaces, particularly if looked at as an extension of the 3D model and computer-aided engineering (CAE) practices already in use today.
“Just like the whole notion of CAE is to save on physical prototypes to see how a crane or tractor operates, now we have operators in the cab working in the vehicle six to eight hours a day and we’re simulating the motion of those people,” says Alain Iung, vice president of Tecnomatix marketing for Siemens PLM Software. “It just extends the fidelity of that simulation to include the actual person in the loop, to see how they will operate the machinery or survive in the event of an accidental mishap.”
The Virtual Build
Ford, for one, believes its virtual ergonomic work has eliminated a significant number of on-the-job injuries. The initiative, which has become a mainstream component of the Ford engineering toolset since 2003, has also boosted the automaker’s quality performance by as much as 11 percent last year in the U.S. compared to 2 percent for the industry average, according to a Global Quality Research System study conducted in 2007 by the FDA Group for Ford.
In the past, Ford would build a prototype vehicle and its ergonomic specialists would spend hours on the physical plant floor, evaluating how the workers went about the assembly process and identifying bottlenecks like awkward postures or difficult insertion efforts. From those observations of real human labors, the specialists would run the calculations to draw conclusions about safety concerns related to back stress or fatigue injuries. “It was quite late when you think of the design of a vehicle to go back to a design or process engineer and tell them we need a design change,” says Allison Stephens, technical specialist, assembly ergonomics at Ford. “Now, as part of their engineering process, engineers have a human component to base all of their decisions instead of addressing the human factors at the end.”
Using Jack, along with other human modeling software, Ford typically conducts a virtual build long before any physical parts are produced, a prototype vehicle is built or any tooling or workstations have been configured. The digital “Jack” or “Jill” (the female counterpart) avatars assemble the vehicles on a wall-sized screen, while the program team evaluates manufacturing feasibility, as well as assembly sequences. From that evaluation, vehicle design changes are sometimes made. For instance, a design specification to place a satellite radio antenna in the center of a Ford model’s roof was modified when the virtual build showed smaller plant floor operators would have trouble with the installation. Since the human component has become part of the virtual model, Ford has seen substantial safety improvements, Stephens says, including a drop off of 80 percent in ergonomic-related injuries in one plant alone.
Caterpillar Inc. is also employing Jack, using the software to evaluate human factors when designing its new tractor cab systems, according to Bernie Meegan, program manager for ergonomics machine research. A 3D model of the system, designed in PTC's ProEngineer CAD tool, is brought into the Jack environment and the digital avatar is used to check ergonomic factors like visibility or whether operators of different sizes can comfortably reach controls. “The benefit of doing this in a virtual environment is that you do the evaluation prior to the cost of developing a piece of iron,” Meegan says. “In that way, you root out a lot of issues around control placement and adjustability before the first part of the product is physically developed. It reduces the cost and the number of iterations in the development cycle and allows you to develop ergonomically sound and safer products.”
Eventually, the Caterpillar team would like to model and assess systems completely in the virtual world. “Whereas in days past, we’d get that first tractor built, run all the tests and then go through iterations of adjusting for safety and making changes to the equipment. This will allow us to make all the changes in the virtual environment and only use the physical tractor to validate the model,” Meegan says.
Caterpillar, along with Ford and other companies out in front with digital human modeling, are collaborating with universities to push the technology even further. The University of Iowa’s Santos project, part of The Virtual Solider Research Program funded by the U.S. military, is a fifth-generation digital human that is an accurate, biomechanical model of a person, right down to the physics of every bone and muscle. Unlike many of the existing digital human modeling programs, which are static, Santos employs predicative dynamics methods so it will accurately calculate walking, running or other movements on a digital human when presented with such variables as body size, strength and weight, according to Karim Abdel-Malek, director, Virtual Soldier Research program and a professor in the departments of Biomedical and Medical Engineering at the University of Iowa.
These capabilities will help Caterpillar consider ergonomic issues like fatigue patterns and impact on muscles in a much more realistic way, according to Meegan. “It’s another level of fidelity, which will allow us to make some better design decisions,” he says of Santos, which is still under development. Ford’s Stephens agrees. With models like Santos taking into account real human factors such as skin and joint movement as opposed to robotic functions, Ford can make better decisions when it comes to safety and ergonomics. “When you’re studying to see if someone can put lug nuts on a vehicle, it’s important that they have fingernails and that the fingers are the same size as humans to give direction to engineers that the lug nuts need to be recessed,” she says. “It continues us on the path to making our models more humanly accurate.”
Beyond the university research, software vendors are doing their part to evolve digital human modeling capabilities. Dassault, for example, which acquired the DELMIA Human technology in 1999 as a result of its acquisition of SafeWorks, has prioritized ease of use in its last few releases. For example, the current version of the software has libraries of hand and body positions such as kneeling or climbing so engineers can easily manipulate the mannequins without having to program them. The software has also been enhanced to work in a virtual reality CAVE in addition to on a workstation, giving users a full solution.
“In 1999, it was pretty difficult for an engineering ergonomics person to program a virtual mannequin,” says Julie Charland, Dassault’s industry solution and consulting manager for virtual ergonomics. “In 2008, we’ve made it easy.”
Siemens PLM Software has similar ease-of-use efforts for Jack under way along with development of additional tools to close the gap between the virtual and physical worlds and to make the model more accurate, Iung says.
The major automotive OEMs are also committed to taking digital human modeling to the next level. Nine automakers, including DaimlerChrysler, Ford, General Motors Corp., Toyota Motor Corp. and Honda R&D Co., along with two suppliers, have formed the Global Human Body Models Consortium to pool resources around human body modeling into a single effort to advance crash safety technology. The consortium plans to develop the first six human body models (large, medium and small males and females) by March 2011 at an estimated cost of up to $18 million, officials say. A second phase will consist of adult models of any age and size followed by a third phase focused on developing child models.
Pooling resources and focusing on a common standard will ensure the creation of a standard, state-of-the-art virtual model, according to Lou Carlin, director of vehicle safety and structure integration at GM. The higher fidelity human body model will help car manufacturers get more specific detail on injuries resulting from crashes and the virtual simulations also allow GM to perform far more crash tests a year. Since 1991, GM performs somewhere in the neighborhood of 400 barrier crashes annually; since it started with digital modeling in crash testing, that number has skyrocketed to more than 8,000 crash simulations each month, Carlin says.
The digital human modeling technology under development will also give GM and other carmakers the ability to measure the impact of crashes on a broader spectrum of users compared with the standard set of physical crash dummies. “An 18-year-old young man in good shape is going to have a different injury response than an 80-year-old woman,” he says. “Armed with this information, we’ll be able to not only design for the robust human models, but adopt designs for an aging population. Digital human modeling is really what will drive the safety of vehicles going forward.”