Simulation – Virtual Reality
Visit to a Virtual World
Global competition is forcing manufacturers to develop and make better products—and to do so faster and at lower cost. Simulation technologies and virtual reality have a key role to play here, as they make it possible to virtually create products and entire production lines long before any real machines or parts exist.
Observers in the VR lab embark on a fascinating virtual flight through a power plant. Locomotive cabs and entire electric motor production plants can be simulated
A train being driven by locomotive engineer Armin Richter has left the station at the Virtual Engineering lab, and is now traveling through the Tauern region of Austria. The train is not really moving, however, as Richter is actually sitting in a life-sized model of the locomotive cab in the Virtual Reality lab operated by Siemens Corporate Technology (CT) in Munich, Germany. Overhead projectors display the route as a realistic stereo image on a giant curved screen, and Richter thus sees tracks, stations, and tunnels racing toward him as he continues his trip. It’s not as easy as it all sounds, however, since Richter has never been in this locomotive cab; nor have any of the other 38 test drivers from 13 countries, who are here to put a new model for a European-wide standardized locomotive cab through its paces at the Siemens facility as part of the "European Driver’s Desk" project.
"The train drivers and the participating train manufacturers were thrilled by the simulation," says Bernd Friedrich, head of Virtual Product Development at Siemens. Rail systems specialist Uwe Mades, who is working on the project on behalf of Siemens Transportation Systems in Erlangen, Germany, is also excited, as European Driver’s Desk "enables us to test various conditions and situations. Each locomotive engineer drives his or her train under the same conditions, which makes it very easy to compare results. This wouldn’t be possible if we used real trains." The standardized cab, which was tested for the first time two years ago, doesn’t really exist yet. However, its new control unit has already been integrated into the MODTRAIN project that will design a modular train concept for the European Union by 2007.
Friedrich’s guests in the VR lab aren’t always train drivers, of course. Some 90 % of the facility’s customers are from Siemens groups such as Medical Solutions, Automation and Drives (A&D), and Siemens VDO Automotive. "They come to us because we can support both their product development and production planning activities with the help of digital, numerically based techniques," says Friedrich. "Here they can test prototypes in the early development stage before actually building them as hardware." Even entire power plants can be digitized—complete with machinery, pipe systems, wires, cables, process technology, major components and production processes. Experts refer to such a system as a "digital factory."
Friedrich’s colleague Carsten Selke built a complete digital factory for electric motors together with experts from A&D in Nuremberg. The assembly line, with everything from material flows to plant control systems, was planned in 3D and tested with a process simulation system. The plant is now being built in China, and although only the structural works have been completed to date, Siemens planning engineers can already predict how the materials will flow during production operations, and what has to be done to ensure that the processing steps are coordinated.
"We’re very satisfied with the support we’ve received with our planning activities for the new plant in China," says Dr. Georg Nerowski, head of the Global Motor project at A&D. "The structured approach to process analysis was particularly important, as it enabled us to obtain a very clear overview of the plant’s resource arrangement. This, in turn, speeded up the layout work considerably, and the layout visualization and subsequent simulation provided us with planning security regarding the necessary investment levels."
Simulation helps engineers avoid errors from the start—be it in an air-mass sensor or an automotive head-up display
Selke still isn’t completely satisfied with the simulation tools, however. "The digital factory could actually be further refined to include functional control of individual machines," he says. A&D experts are already simulating the behavior of numerically controlled machine tools, but have not yet linked them with other elements of the production process. "There are still too many different data models from the various manufacturers of software tools," says Selke. That means that methods and interfaces must be standardized across all industrial sectors, which is why Siemens participates in standardization bodies such as ProSTEP, an association consisting of 200 leading companies in the automotive, aerospace, and plant construction industries, alongside representatives of other industrial sectors.
Flying Through a Sensor. Friedrich has prepared a new simulation in the VR lab. "Get ready for a flight through an air-mass sensor no bigger than a penny," he says, and then hits a few keys on a computer that controls six others located in an adjoining room. These, in turn, provide the six projectors on the ceiling of the Virtual Reality lab with image data. The projectors use vertically and horizontally polarized light to display the calculated stereo images on three screens. When the engineers or visitors put on polarization glasses, they see everything in 3D. In this manner, they can watch as air—depicted as flow lines—moves through a sensor in the intake port of a passenger car engine.
A click of the mouse makes observers feel as if they themselves are being carried along with the air flow into the sensor, which measures air mass. Tiny black particles can also be seen—diesel exhaust particulates that are also sucked in, and that can collect as dirt on the sensor. Friedrich clicks again to turn the sensor and alter the air-flow direction. "This visualization is based on sophisticated fluid mechanics calculations whose results we convert here in real time into stereo images that run in the simulation," he explains. "Our colleagues at Siemens VDO want to know, for example, what the flow field will look like, how much air will actually flow through the sensor, and how many dirt particles will hit it. In other words, they want to examine and test complete system functionality." And they want to do so long before the sensor is built.
Mirror, Mirror on the Screen. Another simulation example involved the testing of design and mounting tolerances in the VR lab for a new type of head-up display for vehicle cockpits. Here, various mirrors for the display were simulated to determine their effectiveness in projecting high-contrast speed and navigation data onto the windshield. Friedrich’s colleagues at Siemens VDO in Babenhausen were interested in finding out which tolerances with regard to the shape and position of the mirrors would be permissible to ensure an optimal image on the windshield. If, for example, the mirror were placed in a slanted position, the result might be a distorted image in the display.
The projectors are turned on and the 3D image of a head-up display mirror appears on the screen. Forces depicted as vector arrows impact the component and deform its surface. "This animation enables our engineers to intuitively determine where the weak points are and where they need to optimize the component’s design," Friedrich explains, and then mentions one of the most important benefits the VR lab offers: "Our simulations reduce development times by up to 30 %." This not only means that new products reach the market sooner than was previously the case; such products are also more mature and reliable because the simulations help to uncover errors at an earlier stage. And since the required manufacturing resources and materials have already been precisely simulated beforehand, production launches can also be implemented more rapidly in the factory.
Friedrich’s team requires a huge amount of data from designers to carry out their simulations. Along with the information on the geometric shape of the CAD models, they also need to know which materials the product is to be made of and under what conditions it will operate—for example, details on the electrical and mechanical properties of the surrounding environment. Using this data, as well as off-the-shelf software tools, specialists draw up a numeric simulation model. "We don’t develop the software ourselves," says Friedrich. "Our job is to make very good use of existing complex tools."
The original CAD model is usually first simplified by the team and then freed of details not pertinent to the simulation. This minimizes computing times. After this simplification process, the model is broken down into a network frequently consisting of several million tiny triangular bodies known as finite elements. Numeric procedures are then used to compute physical conditions such as temperature, pressure and force for the finite elements. Because each value can be changed with just the push of a button, it is very easy to run through various scenarios. The simulation becomes more complex, however, when values from different physical realms interact with one another. For example, simulation specialists have already succeeded in linking procedures for fluid mechanics and acoustics, thus enabling them to study the sound propagation of air vortices created by the current collectors on high-speed ICE trains.
Holistic Simulations. "The highest form of linkage that we’re trying to achieve is a mechatronic simulation in which a product’s mechanics, electronics and software interact," says Friedrich in describing his vision of a virtual prototype. Up to now, the mechanical aspects of a component have generally been developed first, and then the suitable software system. "We could speed up product development if we could develop the software in parallel with the design and then test both together," Friedrich explains. The pioneers in this field are Friedrich’s colleagues at A&D in Erlangen, who have already taken a step in this direction with their Sinumerik Machine Simulator (see box). "We will need another two years in order to achieve complete integration, however," says Friedrich, "We’re currently working on the development of the required methods and calculation procedures."
Rolf Sterbak
A personal computer as a virtual machine tool? Thomas Menzel, product manager at Siemens Automation and Drives in Erlangen, doesn’t see this as a problem. His Sinumerik Machine Simulator program demonstrates how this amazing transformation can be achieved. In this system, a computer is hooked up to a numeric control unit—a common feature for machine tools. Menzel then starts his program on his PC, which begins acting out all the functions performed by a real machine connected to such a control unit. Conversely, the automatic control unit reacts as if it were communicating with an actual machine. "Our customers, who are mechanical engineers, can use the system to test and improve new machines in the design phase, and such a procedure offers huge benefits," says Menzel. With the Sinumerik Machine Simulator, Siemens can offer designers a type of kit that already contains many machine components in virtual form, for example delay elements. These are typically used to depict machine tool feed processes.
Designers can also flexibly generate additional components and store them in a model library, from which they can be accessed at any time. All these elements can be used by designers to create their machine drafts in a drag & drop process, whereby a virtual depiction of the machine is gradually generated and this depiction simulates the actual behavior of the planned machine tool or production machine before anything is even built. Questions can thus be answered at an early stage regarding whether the mechanical systems and software will interact properly, or whether the future machine will actually do what it has been designed to do. Designers can find out how fast machines will drill or mill and also determine the quality of the products they will manufacture. Thresholds can also be tested without any danger. For example, designers can find out what will happen if a sensor fails. In the worst case, even if a virtual machine breaks down, the "damage" can be repaired with just the push of a button. "The investment in the simulator pays off after less than a year," says Menzel. "The quality of the customer’s machine tools and production machines increases, and they can also be delivered and put into operation much more rapidly." One of Menzel’s main customers confirms this. Bernd Zapf is head of development at Heller Maschinenfabrik in Nürtingen, Germany, one of the leading manufacturers of machine tools and production equipment in the automotive industry. He says his company’s experience with the machine simulator has been so good that "we will use it to simulate all of our new machines."