How Simulation Ensures Better Vacuum Cleaners

2023 Author: Hannah Pearcy | [email protected]. Last modified: 2023-06-02 20:50

The calculation engineer Stefan Koch, who grew up in Romanshorn, Switzerland, was enthusiastic about the Dyson company from the start. Already at the 2016 launch date, he felt the fascinating atmosphere in the company. "All people were very motivated and committed to meet the high demands on the newly developed Dyson products with their know-how," reports Stefan Koch. “They took me through the labs and explained their approach to basic research and product development. I was thrilled with the Dyson campus in Malmesbury, England, and it was immediately clear to me: I want to work here!

After completing his mechanical engineering bachelor's degree in Winterthur, the Swiss completed a master's degree in aerodynamics in Southampton. Today he is a senior aerodynamics engineer in Dyson's research and primarily responsible for floor care. There he heads a five-person simulation team that works with the Ansys software.

Picture gallery

Picture gallery with 7 pictures

How Dyson's ideas come about

Dyson has always been very experimental. The experiment and try things was and is an important component in the creative process of brainstorming and evaluation. As a result, many prototypes were built and tested, then improved and tested again. This approach was coined by company founder James Dyson, who built 5,127 prototypes for the first bagless vacuum cleaner.

"We do not only want to carry out optimizations again and again, but also to fundamentally address the existing challenges in order to provide our customers with the best possible products for the respective application area," emphasizes Stefan Koch. For example, the latest cordless vacuum cleaners from Dyson, with their high suction power, can completely replace conventional vacuum cleaners. The suction power of the Dyson Cyclone V10 has been increased by around 20% compared to the predecessor V8. The suction power of the V11 launched in early 2019 was increased by a further 15%.

Product development

Rethinking product development

Simulations at the center of development

"Tinkering and trying out is in our company DNA," explains the calculation engineer. “Although simulations are the focus of our developments today, prototypes still play a very important role. Because we have to validate and test our designs again and again to get a real feel for how our products are handled.”

At Dyson, numerical simulations are mainly used in two ways:

• On the one hand, simple, fast simulations are used to evaluate new or modified concepts.
• On the other hand, it makes it possible to understand complex physical processes that, for example, run so quickly that measurements can only be carried out with great effort or not at all.

Here Stefan Koch names the floor nozzle of the vacuum cleaner, which works at 3000 revolutions per minute. For him, this is one of the most complex multiphase systems, which includes brushes and carpets as well as particles in the nanometer range and also large amounts of dust. To build a realistic test system to carry out appropriate measurements seems hardly feasible for him.

Counter increasing demands with simulation

Simulation applications ultimately enable efficient new development and optimization of products and components in order to be able to present improved products on the market again and again. In addition, Stefan Koch emphasizes: “The frequency with which we launch new products has increased enormously. In my opinion, this has only been possible through extensive use of simulation. Because simulations are the best tool for fast and efficient product development. They not only provide knowledge about the physical behavior of the product, but also make it possible to look inside to understand what is happening there.”

Dossier simulation

How simulation improves product development

What does a good calculation engineer have to do?

Stefan Koch answers with a basic statement on the question of the necessary computing power, which can be increased by the parallel use of processors:

“Calculation engineers have a responsibility to think intensively about how a problem can best be solved with simulations. This means that not just as many processors as possible are used, but that consideration is given to how a sensible simulation model can be constructed. To do this, I have to think about exactly what I want to know and what model or software tool I need for it. It should also be clarified how detailed the results should be and which aspects should be taken into account and how. Do I need a 3D calculation and which turbulence model is best suited? My experience is: The easier I can design my simulation model, the easier it will be later to interpret the results and draw conclusions from them. Otherwise it could happen that the data flood floods me. All of these criteria distinguish a good calculation engineer from a simple software user.”

Smart engineering day

The digitization of production requires a rethink in product development. The Smart Engineering Day offers decision support for the selection of the most suitable methods for the conception, design and development of smart products and machines.

More info

How Dyson uses simulation in product development

Flow optimization: New software functions available:

• The purchase decision when choosing a vacuum cleaner mainly depends on its suction power, battery life and filter system. That is why the calculation engineers at Dyson are concentrating on optimizing these product properties, for example on increasing the suction power through better motors and compressors.
• Due to new software functions, completely new approaches to flow optimization arise. So far, people have always designed the paddle wheel and modified the degrees of freedom and parameters, but now part of this work can be done by the software. This is an advantage because people have to think problems non-linearly and find solutions that they cannot intuitively recognize. With the Ansys Fluent Adjoint Solver, on the other hand, geometries can be optimized very efficiently.
• After the flow field has been calculated, the Adjoint Solver uses sensitivity studies to analyze how the geometry has to be modified to improve the flow. “With this modified geometry, the cycle starts again until we are satisfied with the result of the flow simulation,” reports Stefan Koch. “After 10, 20, 30 or even 40 iterations, we get an optimized blade wheel geometry that gives us the desired flow results. How could we try to achieve such a result?"

Minimize unwanted sound sources:

• A similar procedure is used for minimizing noises caused by the flow around the impeller. To better understand the processes at the rear edge of the paddle wheel, the Detached Eddy Simulation (DES) is used locally. It can be used to generate very high-resolution results in which the acoustically relevant turbulence is effectively calculated, which takes a lot of time and computing power.
• The strength and frequencies of the vertebral detachments could be analyzed in detail. The result was that the intensity of the sound sources with corrugated impeller edges is not as high as with straight trailing edges and the noise was reduced due to the simulations. And the tests also showed that the modifications to the real paddlewheel provided the sound reductions that the simulation promised.

Coupled two-way particle simulation:

• But it is also important to look at the individual cyclones of the vacuum cleaner, which serve as centrifugal separators, in order to increase the suction power. The flow in the cyclones - sometimes more than a dozen concentrically arranged - are always unsteady due to the global instability, so that an abstract concept of different flow modes has to be analyzed. Here, too, extensive detailed calculations are necessary in order to obtain high-resolution results for the individual turbulence structures so that the deposition efficiency can be further improved.
• Individual, sporadic vortices that reduce the deposition efficiency could be prevented by modifications that were carried out on the basis of corresponding simulation results. "To do this, we had to go into great detail, for example, to be able to take the dust particles in their different sizes - from the micrometer range to the size of so-called woolen mice - into account in the calculations," explains the calculation engineer.
• To do this, he uses a coupled two-way particle simulation, because the air flow influences the particles and, conversely, the particle concentration in the cyclones slows down the flow. It should be borne in mind that unimaginable centrifugal forces, which are a hundred thousand times gravitational force, are generated in the cyclones. This can result in pressure fluctuations and vibrations that cause the cyclone to vibrate, which results in pressure losses and noise, and of course reduced separation efficiency.

Multi-scale problems with filter considerations:

• With regard to filter simulations, Stefan Koch explains: "For me as an engineer, they are fascinating and at the same time a nightmare." Because of the filters, a multi-scale problem has to be considered, on the one hand the filter system, which is several centimeters in size, and on the other hand the micro level with nanometer-sized fibers in the filters and the smallest particles that should get caught there.
• In addition to the different orders of magnitude, different time scales have to be taken into account. This is hardly possible in a simulation. That is why the Dyson engineers, in collaboration with the University of Oxford [Printsypar, Galina & Bruna, Maria & Griffiths, Ian. (2018). The influence of porous media microstructure on filtration] developed a homogenized multiscale model, with a simulation for the macro problem and one for the micro problem. The flow simulation through the filter material is coupled with the calculation of the particle separation in the filter, so that changes in the porosity of the filter are returned to the macro simulation.

Focus on system simulation and democratization

When asked what future focal points will be considered in the simulation, Stefan Koch sees two important subject areas: “We are already looking more and more at the systems and not just at the individual components. An excellent system needs optimal components, but they also have to work together efficiently. We improve this with the help of system simulations. The second important area I see in the so-called democratization of simulation. In the future, this should enable every engineer in development to use the simulation so that everyone can benefit from the advantages.”