Skip to main content
To KTH's start page To KTH's start page

Mathematics for Life

Computer simulation can decrease suffering while helping protect the environment.

Research in Depth

Published Sep 14, 2012

KTH researchers are developing advanced computer simulations that may help decrease the number of experiments on animals and at the same time improve environmental protection.
“Simulation can already reduce the need to produce industrial prototypes,” says Johan Hoffman, professor of numerical analysis at KTH. “And computer simulation may soon be used as decision support for certain medical procedures.”

n his research, Johan Hoffman and his team of scientists develop computer programmes that can model turbulent flow in the environment, and show how this turbulence will affect us.

Fluid mechanics is hardly a new science — the mathematical equations describing flows have existed for 250 years. But with the progress of computer development in recent decades, the development of technology has taken off considerably.

Johan Hoffman and his team of scientists at KTH’s School of Computer Science and Communication (CSC) uses the most powerful computer system in the Nordic region: a 306 TF Cray XE6, also known as Lindgren.

“The main focus is to develop methods for large-scale simulation of water, air and blood flows,. This allows us to see how flow affects the functionality of airplanes, cars and boats, and even human speech and heart functions,” Hoffman says.

His team of scientists has already contributed to a range of industrial technologies, including production of silencers for trucks.

“Our simulation technology can lead to considerable environmental savings, because it allows the automotive industry to optimize designs without actually producing the physical prototypes. Simulation makes it be possible to test product functionality and properties on the drawing board and computer screen through virtual prototyping,” says Hoffman.

In 2011, his team managed to predict air resistance for an entire car with mathematical modelling and computer simulation, and the results were presented at the Super Computing conference SC11 in Seattle.

In the workshop BANC-II at the American Institute of Aeronautics and Astronautics conference in Colorado Springs this year, a similar methodology was used to calculate sources of acoustic noise from airplane landing gear. Here it became apparent that the KTH scientists are at the international forefront of computational fluid dynamics.

The physical tests conducted on the landing gear by Boeing and Nasa illustrated that the computational method was reliable. Results from the digital simulations tallied with reality.

“The fact is that our research is at the frontier and we don’t have any other research group to turn to for answers to our questions,” says Hoffman.

“As a researcher, I want to increase understanding of the opportunities and limitations of simulation models. We will be able to make the development process more effective by avoiding the use of expensive prototype production. In the future we can also reduce the number of animal experiments. But our simulation models in medicine will be used as decision support for professionally active surgeons – they will not become an app that everyone can download to their telephones,” says Johan Hoffman.

New findings and industry news are shared in trade shows and conferences, and the Swedish scientists have implemented a new computational technology in the open source software Unicorn as part of the FEniCS project .  

Recently, Johan Hoffman has directed more of his research toward medicine. He has contributed to the creation of a computer model of blood flow in the human heart in collaboration with KTH School of Technology and Health, Umeå university and Linköping University.

Mathematical equations describe blood flow, based on ultrasound measurements of heart wall movements.

“The simulation gives us a better understanding of how the heart works. The computer model can provide a more comprehensive picture of blood flow than can be observed with pure imaging techniques such as magnetic resonance imaging (MRI) and ultrasound,” Hoffman says.

“In addition, we are looking for ways to identify the best treatments for given individuals, for example by testing the functionality of several different types of cardiac valves, virtually, prior to an operation.

Success for a project of this sort requires investments in reliable computational tools, accurate quality control and quantifying errors appearing as a result of approximations in the models.

“One strong argument in favour of continued investment in advanced computer simulations in medicine is that they increasingly will be able to replace animal experiments. Instead of testing products and medical treatments on animals, the tests could be conducted on virtual, human models in a computer,” Hoffman explains.

He points out that the field of computer simulation is undergoing rapid change with gains in computer performance.

“Only 20 years ago, scientists worked with two-dimensional models in these types of projects. But now, when we work in 3D, we can make all the tests more realistic. My goal is also for models to become interactive, so that users feel that they enter into a virtual laboratory, like in a computer game.”

Johan Hoffman is deputy head at a new department, High Performance Computing and Visualization (HPCViz), at the KTH School of Computer Science and Communication. The aim is for HPCViz to become the leading Swedish research environment for high performance computational mathematical modelling and advanced computer visualisation.

“What we see in e-science generally and globally is the gathering of maximal amounts of data on all fronts, and that we now have to learn how to manage, sort and use all that data,” he says.

“My vision is that we will be able to combine all data with advanced simulation models in the future, for example to be able to create models of the ocean and the atmosphere. In another decade, we are likely to have achieved that vision.

For more information: Johan Hoffman at +46-8-790 77 83, or jhoffman@kth.se.

By Katarina Ahlfort

Three application areas for fluid mechanics

1. Medicine: Blood flow in the heart can be calculated with a computer model. Technology makes it possible to predict how the heart will be affected by, for example, new cardiac valves.
2. Aerodynamics: Calculations of air resistance, fuel consumption and noise can be conducted without the need to produce and test physical prototypes.
3. Geophysics: It is possible to predict how ocean currents can be altered following the construction of wind power stations at sea. Simulation could also help to organise crucial rescue operations during oil spills at sea.

Hoffman’s unique computational methodology

The team of scientists uses mathematics to optimise the use of computer power by adapting and delimiting the method of computation. With the help of “adaptive methods” (a type of self-instructing algorithms), scientists find correct answers with minimal effort.

Here’s how it is done: The flow is computed with the help of a grid that is fine-meshed only in the areas important to the study, for example concerning air resistance or blood flow. If the correct areas are selected for fine-meshing, the method can predict with a high degree of reliability the turbulent flow, and enable studies that were previously infeasible.

Read more about the Computational Technology Laboratory

Johan Hoffman's personal web page