Main content of this page

Anchor links to the different areas of information in this page:

You are here: MEDICA Portal. Magazine & More. MEDICA Magazine. Interviews. Interviews Concerning Innovations. Prosthesis.

“It Is Really Exciting“

Robust Design: “It Is Really Exciting“


Photo: Tanja Clees

Doctor Tanja Clees;
© beta-web

Computer simulations are indispensable in many industry sectors. The advantage is obvious: Before a part goes into production, it can already be thoroughly “put through its paces”. Defects are noticed much faster and production can be adapted at any time.

Robust Design is headed for medical technology: talked to Doctor Tanja Clees from the Fraunhofer Institute for Algorithms and Scientific Computing about the emerging field of Robust Design and what methods from the automotive industry could be interesting for medical technology. Doctor Clees, what exactly is Robust Design?

Clees: It is a relatively new field and includes very different areas and functions. The aim of Robust Design is to design production processes and products with robustness. Robust is not meant in the sense of something for instance falling on the ground and ending up minimally damaged, but rather in the sense of products becoming more robust towards production variations or external forces. In reality you often discover that a product is not optimal due to miscellaneous influences, if variations were not incorporated in the optimization process. In practice, an optimization cannot be achieved or very poorly if it’s determined this way. You also deal with simulations of flow mechanics and structural mechanisms in production processes. How should we envision this?

Clees: You can take a car as an example. Flow occurs if you drive the car into a wind tunnel – in the simulation this only happens virtually on the computer – and you can see how the air flows around the car. Flow does not always have to pertain to air, but could for instance also pertain to mixing machines that are supposed to blend different components. Generally you want to optimize the flow, so for example as much blending as possible occurs or in the case of the car, so as little aerodynamic resistance as possible takes place. Structural simulations serve other purposes: You might want to know how stress resistant the material is, and hence produced structures that are ideally suited for your products. In the medical field for example, man-made material to make bones is tested via simulation. Mathematical methods aid in calculating the synthetic bone structure in a way to make it light on the one hand and on the other hand – for instance for walking purposes– strong enough to withstand the same loads as a normal human bone would. Our body also conducts optimizations. Bones grow in a way where they ideally withstand walking processes and its load requirements, respectively – and you try to replicate this. During optimization you basically copy what the body tries to do on its own. If you constantly strain your hip in the wrong direction for instance, you will alter your bone structure because if you are only able to put pressure on one side and have to support the other side, the body will react to it and try to better absorb the new load case. Can you give examples of such simulations and their computational models?

Clees: We look at what happens during a car crash for example. If in the simulation the car hits the person, you can see what happens to the bones and the tissue among other things. Methods that we developed for the automotive industry are transferred to a human being in a crash situation. For these simulations you first need to mold the “human material“, because the so-called soft tissue, meaning soft tissue like flesh for example, acts differently than metal or aluminum. If this was successful you can use the same simulators as in the automotive industry. To do this you install material models and adapt them a little. Then you check how the stresses and strains affect the human body.

Photo: Doctor Clees working on her computer

Different conditions can be clearly represented in the computer simulation; © beta-web How do you proceed when you analyze or optimize a production process, related for instance to manufacturing a hip implant?

Clees: At first we parameterize the process that is, we specify parameters which can be varied, just like in the production process. However, there are other parameters dictated by nature which we cannot control. Overall we try to grasp the effect of a process through parameter variations. In production this usually refers to parameters like size, length, temperature or pressure. Clearly this also applies to a hip implant, which can be built quite differently. Next you consider, how these separate parameters vary, because you cannot set them to a fixed value. If a metal for example needs to have a certain size, in practice this is called plus/minus tolerance. That is why it makes the most sense to include experimental data or physical assumptions. Next you closely look at the whole range of possibilities. We do experiments, but in simulation. Here you can vary parameters or set up an experimental design, let it run and do a subsequent assessment. We virtually run through many different configurations. By doing tests we learn something about the performance – for example we learn about the load case of a hip implant, like for example whether it can bear the load that is impacted on it. This can be quantified in data. You simply evaluate different criteria: load level, mass and cost. Afterward, we interpolate the performance and thanks to this interpolated data we can then optimize. As a mathematician you try to find the so-called Pareto optimum, because there is never just one optimum here, but always multiple compromises of equal value. As a mathematician you never know whether the weight for instance bears more importance than cost or load. This is something the customer needs to determine himself. What innovations are you contributing to the area of medical technology?

Clees: We contribute to research in the area of kidney stone destruction and also tumor treatments. The concept is this: Take for instance an instrument that applies radiation to the human body. We try to do similar testing for such a device as we do in crash situations of the automotive industry. We check what happens to the human tissue as a result of the radiation and try to assess it. Just now we started a project on this topic where we transfer these quantifications for the first time ever. It is really exciting. So the organ virtually replaces the car in this computation?

Clees: Indeed it does. In place of the car, we act on the assumption of a human being or his/her organs, respectively. We look at the physical effects of waves on the body just like we do with a car crashing into a wall, because we want to find out how the variations proceed “internally”. One example is the destruction of a kidney stone through impulse waves. Needless to say you have to reasonably model the kidney, by examining kidneys of animals among other things. Afterward we can analyze the effects to the tissue on the computer. The whole purpose is then for instance to determine an optimal dosage or the optimal length of treatment. However, this can only be achieved if the models were cleanly validated beforehand – this is one of the most difficult aspects of this project, since we cannot conduct the experiments on actual human beings.

The interview was conducted by Simone Ernst and translated by Elena O’Meara


More informations and functions

Further Interviews!

Photo: Microphone

© Shumskiy