Go to content

SIEMENS

Research & Development
Technology Press and Innovation Communications

Dr. Ulrich Eberl
Herr Dr. Ulrich Eberl
  • Wittelsbacherplatz 2
  • 80333 Munich
  • Germany
Dr. Ulrich Eberl
Herr Florian Martini
  • Wittelsbacherplatz 2
  • 80333 Munich
  • Germany
pictures video

In August 2012, Curiosity landed on Mars and began to explore the planet’s surface. The entire system was developed and tested using Siemens simulation software.

The entire unit consisted of around 90,000 parts, many of which were allowed to deviate from design by only 100 micrometers — about the width of a human hair.

Meeting on Mars

NASA's Mars Science Laboratory mission - which includes the latest rover, popularly known as Curiosity - is the most technologically complex project in the space agency's history. Developed and tested using Siemens design and simulation software, it is an example of the seamless communication that is developing among machines - the meeting of the real and virtual worlds.

Image Image Seven Minutes of Terror. Lowering the rover to the Martian surface on cables was a particularly delicate maneuver. The landing phase was simulated 8,000 times.
Image Image The entire unit consisted of around 90,000 parts, many of which were allowed to deviate from design by only 100 micrometers — about the width of a human hair.

Uncertainty Quantification: The Next Step in the Science of Simulation

Image open

Whether you’re sending a laboratory to Mars or crossing Europe in a high-speed train, there are innumerable phenomena that are virtually impossible to predict, and they all add up to one overwhelming factor: uncertainty. So why not try to quantify uncertainty itself? That’s what Dr. Utz Wever and Dr. Meinhard Paffrath, both specialists in simulation and mathematics at Siemens Corporate Technology (CT), are working on. Their focus: railway wheelsets. “Wheelsets — including axles — are the top safety item on a train,” says Wever. “They also account for a significant proportion of operators’ maintenance costs.” With this in mind, the researchers are developing a new way of looking at these crucial parts — a method that amounts to a virtual inspection. Normally, inspections rely on ultrasound to determine whether cracks are present. If a crack is found, the part at fault is scrapped. But a lot of the time inspections don’t find anything because the probability of finding a crack with at least the minimum depth during an inspection is itself a random number. After all, the factors that cause cracks — things like loads, speeds, and temperatures — constantly fluctuate. “Our approach is radically different,” says Wever. “What we want to do is to predict crack size so that inspections take place only when the probability of discovering a crack is high enough to warrant the cost of an inspection.” Adds Paffrath: “As we move toward that goal, we are currently computing failure probabilities for fixed inspection intervals.” To do that, Wever and Paffrath started out with a program from Germany’s Fraunhofer Institute that computes crack growth without consideration of uncertainties and inspections. They then “added sophisticated algorithms to the program in order to compute failure probabilities for different inspection scenarios.” And the result? The software is now being evaluated on a pilot basis by Siemens Rail Systems in Graz, Austria. “We are now at the point where we can develop predictions of crack sizes, thus reducing the uncertainty of expensive but unnecessary real inspections,” says Wever. “We will then compare the results of the real inspections with our predictions. The initial results look promising.”

close

Thirty-six weeks in the unspeakable cold of outer space and a 567-million-kilometer voyage at speeds of over 76,000 km per hour were just the beginning. As the Mars Science Laboratory (MSL), a $2.5 billion, 900 kg rover the size of a small car streaked into the Martian atmosphere on August 5, 2012 it was still traveling at 21,000 km per hour. At that moment it had seven minutes to reach a landing speed of less than two kilometers per hour — or crash.
In order to touch down gently enough to avoid damaging instruments designed to search for the chemical ingredients of life, hundreds of complex steps had to be executed flawlessly and without human intervention. How did engineers prepare for such a challenge — one that could not be tested on Earth because our atmosphere is 100 times thicker than that of Mars? “Well,” says Chuck Grindstaff, President and CEO of Siemens PLM Software (product lifecycle management), a business unit of the Siemens Industry Automation Division, “NASA’s Jet Propulsion Laboratory (JPL) designed the whole thing using our simulation software — everything from thermal analysis to the multi-physical interactions the vehicle would encounter as it entered the Mars atmosphere. Our software was in the center of solving all of it.”
Daren Rhoades, a Senior Product Development Manager at PLM Software’s Cypress, California development center who, until recently, was a member of the NASA team that developed the MSL (Curiosity) mission, adds that, “There were subsystems that I worked on where simulation allowed us to go from concept to detailed parts, to assembly and testing entirely within the virtual world.” MSL’s crucial landing sequence — what NASA called “7 Minutes of Terror” — for instance, was optimized in the course of 8,000 simulated landings. “The significance of being able to go from simulations to real-world deployment,” says Rhoades, “is huge.”
It really is. Consider, for instance, the Sky Crane, a never-before-used system designed to brake the Rover’s final descent and gently lower it to the Martian surface. PLM software simulated the dispersion of the plumes of fire from the Crane’s rocket engines to ensure that they would not damage the rover or the harness it was attached to. “Not only did the cables have to let the rover down flawlessly without interference from the flames,” explains Joel Rooks, PLM Account Executive for NASA, “but there was a kind of umbilical cord that kept the rover tethered to the Crane until the split second before the crane took off. Everything had to separate at once. To do this, there were little guillotines that had to simultaneously sever all those lines. All of that was simulated using our software.”
Complexity was also a major issue. The entire assembly consisted of approximately 90,000 custom-made parts, many of which were allowed to deviate from design by only 100 micrometers — about the width of a human hair. What’s more, to minimize wasted space, those parts had to be folded into an extraordinarily dense package, while nevertheless allowing just enough wiggle room for the extreme shaking of launch and reentry, as well as the expansion and contraction of different materials under temperature variations of as much as 1,648 degrees Celsius. “To design things that are that densely packed requires new capabilities and new techniques,” says NASA Director of the Mars Exploration Program Doug McCuistion. “We used a lot of systems from Siemens to help us do this.”
“It was important to design all system parts so that they would not touch each other and potentially cause damage,” adds PLM Solutions Consultant Kent Rash. “The only way to ensure that, given the different materials involved, was to have a finite element model — FEM — of each, which is a method for dividing complex surfaces into small elements that can be calculated in relation to each other.”

From Models to Manufacturing. And that’s where one of PLM’s primary software tools — NX — comes in. NX brings products to life by not only producing CAD (computer-aided design) models, but also through the use of computer-aided engineering (CAE), a process that imbues a design with related functional and physics-based data, such as how much stress or heat a part can safely withstand given the material it is made of. And because many parts in machines are designed to move and interact with each other, NX can perform so-called “kinematic simulations,” allowing engineers to “animate and test even very large assemblies in the context of events that might happen in the real world — such as a rover landing on a rock, and figuring out the stresses that would impose on the entire structure,” says Rhoades. NX is also a computer-aided manufacturing (CAM) tool. “NX programs were used to generate the code that drove the machines that manufactured the parts for the MSL mission,” says Rhoades. “You start out with the original NX CAD model of a part in the virtual world, define which tools will produce it, and then actually run a machine using the NX CAM software based on the original CAD model that you used for your finite element analysis.” This process offers many advantages, the most meaningful of which for a multi-billion dollar space mission is accuracy. “The accuracy you can achieve from using a digital model combined with computer numerical control (CNC) in parts production is tremendous,” adds Rhoades. “By the time the MSL had been built, many of its machined parts were off by no more than the width of a human hair.”

Connecting Desks…and Industries. Not only do complex projects require exceptional simulation software, they also demand systems that allow large teams of engineers to collaborate securely on the same project. And when it comes to “connecting the desks,” nothing beats PLM’s Teamcenter software. “Teamcenter is the critical supporting system that permeates the entire engineering design process,” says PLM CEO Grindstaff. “By providing a set of applications for things such as requirements management, project management, regulatory compliance, and design data management, it is the backbone of the design process.” Indeed, as is the case with NX, Teamcenter is used not only for all of NASA’s Jet Propulsion Laboratory flight missions, but by SpaceX, which, in May 2012, made history when its Dragon spacecraft became the first private commercial vehicle to successfully dock with the International Space Station.
Why do organizations like NASA and SpaceX turn to Teamcenter? “Simple,” says Rhoades, “because no part in a space vehicle is unimportant. Not only does Teamcenter minimize the possibility of human error by ensuring that each authorized participant sees only the latest version of the project data he or she is working on, but it makes it possible to trace every single part back to its original design, analysis, and manufacturing data. And when a design is approved for manufacturing, it is documented as such by Teamcenter. All of this can add up to huge potential savings.”
Not surprisingly, high-flying industries are not the only ones to have taken notice of the competitive opportunities offered by virtual prototyping and collaborative development. “We have 70,000 customers,” says Grindstaff, who points out that the total market Siemens PLM Software addresses is around $20 billion per year and is growing at an annual clip of five to seven percent. “We are the leaders in collaboration and data management. We are the leaders in digital manufacturing, and we are a strong number two in mechanical CAD/CAM,” he says.
What’s more, the company is on an impressive run with major new contracts. It recently signed a ten-year agreement with Boeing to expand the use of PLM technology, as well as a worldwide agreement with Daimler to develop all its products and factories in a Teamcenter and NX environment. “The introduction of parallel processes in development, design, production planning, and production will further optimize Daimler’s entire value chain, allowing it to produce better products faster,” says Grindstaff. The company has also recently signed major contracts with Chrysler and Johnson Controls, a top-ten, tier-one automotive supplier and the automotive industry’s largest seating supplier.

Out-of-this-World Advantages. Among the many reasons for such successes is the fact that Siemens PLM Software helps its customers to cut costs and become more efficient. For instance, as it has integrated PLM technology with its manufacturing operations, Samsung Electronics has been able to reduce its use of physical prototypes by 30 percent, thereby cutting errors by 50 percent in first production runs, and accelerating development time by 30 percent, says Grindstaff.
So how much productivity did NASA gain on its latest ticket to Mars? The amount, says Rhoades, is impossible to quantify because, unlike virtually every earth-bound product, MSL is one-of-a-kind. “It isn’t as though they started out with a two-ton version and cut it to one,” he says. What is clear is that the Curiosity / MSL mission has set a new standard for integrating everything from concept to production and testing. Says Siemens Industry CEO Prof. Siegfried Russwurm, a member of Siemens’ Managing Board, “In the past, processes were sequential. What NASA was able to do with the Mars rover was a paradigm shift, an integrated database, an integrated approach from product design to production design — a seamless transition from the virtual world to real production in one consistent database for hundreds of engineers working on one consistent model.”
Adds PLM Account Executive Rooks, “MSL was the most technologically-complex project NASA has ever had. What’s important is that our tools helped them to simulate and optimize everything, and when they tested it and flew it to Mars, it worked.”

Arthur F. Pease