Electromagnetic simulation on a grand scale
01 May 2010
Electromagnetic simulation software has successfully supported the design of a groundbreaking particle beam-steering magnet for cancer treatment. The 70 tonne, 1.81 Tesla unit is believed to be the largest ever constructed for a medical application
The Opera electromagnetic simulator from Cobham Technical Services has played an important role in the design of a particle beam steering magnet, currently being commissioned at Italy's new hadron therapy centre in Milan - the Centro Nazionale di Adroterapia Oncologica (CNAO). Weighing in at 70 tonnes, the 1.81 Tesla dipole magnet is believed to be the largest ever produced for medical applications. Positioned at the end of a particle accelerator, it turns the particle beam through 90 degrees and directs it towards a patient treatment table.
The magnet was manufactured by Sigmaphi, a company specialising in custom magnetic systems and beam transport lines for particle accelerators. As with almost every magnet design that Sigmaphi creates, the CNAO specification was unique, as well as being extremely challenging. The specification called for a very large magnetic field region of 20 x 20 cm, combined with exceptional field homogeneity. Sigmaphi had an established track record in this application, having already produced a similar bending magnet for the Heidelberg Ion Therapy Centre. However, CNAO's specifications called for even higher performance - with field homogeneity improved by a factor of two.
The sheer scale of the CNAO magnet meant that the design had to be right first time, as any post-design modifications would have had a dramatic effect on project construction time and costs. To help ensure positive outcomes for its projects, Sigmaphi makes extensive use of simulation using the Opera finite element analysis tool. For CNAO, the company created a very detailed three-dimensional model of the magnet concept and performed dozens of simulations of design iterations before settling on the final optimised parameters for manufacture.
Fast simulation is very important in complex applications like this; in the case of the CNAO magnet, the speed of simulation was aided by Opera's ability to use Biot-Savart calculations for computing coil fields, greatly reducing the need for complex finite element meshing of the model, which would have had a big impact on computation time.
The Opera simulator is the world standard for this scientific application because of its accuracy and execution speed. It is also commonly used by both designers and end user organisations, which helps to simplify the specification, design and test cycle of projects. Sigmaphi engineering director, Frederick Forest takes up the story.
"We spent around six months designing and optimising this magnet before we put it into production. Even though the design model was complex, the 3D Opera simulations took only seven or eight hours on a standard PC, a speed that helped us to investigate a large number of design variations. I know from my own investigations into other simulators that Opera provides more accurate results in this magnetics application, and it's a very important tool in our work."
Over the company's 29-year history Sigmaphi’s engineers have built up a globally-respected knowledge of techniques that can be employed to exert precision control over magnetic fields. Several of these techniques were used in this application, and the Opera simulation exercises helped to optimise the function of a number of the features. These included optimisation as to the size and shape of field clamps on the magnet, which improve field homogeneity, and similar modifications to an iron collar that helps enhance the fidelity of beam steering.
The Opera simulator used in this application - known as Tosca - employs a discrete finite element model in order to solve the partial differential equations governing the behaviour of static electromagnetic fields. It computes the total magnetic scalar potential in the magnetic material and the reduced magnetic scalar potential in the regions where source currents in coils are specified.
The reduced potential represents only that portion of the field produced by magnetisation, the remainder of the field being computed directly from source currents. This avoids the drawbacks of other methods which can produce cancellation errors or require complex meshing that conforms to the geometry of the coils. As a result, the accuracy of the computation is far higher than alternative methods and is proven by nearly 30 years of comparison with measured results.
Contact Details and Archive...