Mosaic’ finite element meshing and auto-optimiser will shave weeks from development.
Advanced electromagnetic simulation software is helping one of the world’s leading producers of cryogenic equipment to speed the development of application-specific superconducting magnets for research. The software – the Opera 3D simulator – has been provided to Cryogenic Ltd. The package includes a suite of 3D electromagnetic design, simulation and analysis tools, plus a unique optimiser. This latter tool automatically employs multiple goal-seeking algorithms to eliminate the need for manual intervention when evaluating the best solution for a particular design.
Cryogenic designs and manufactures a wide range of superconducting magnets and associated measurement systems for laboratory research and industrial uses worldwide. In addition to producing liquid helium cooled systems capable of generating magnetic fields as high as 22 Tesla, the company is renowned as a leader in cryogen-free magnets with flux densities up to 20 Tesla.
Generally, Cryogenic uses its own in-house software to design the basic mechanical layout and coil structures of a magnet, fine tuning its field profile with magnetic material to meet the customer’s specific needs. The shape and placement of the magnetic pieces are critical to the magnet’s performance, and are determined through extensive electromagnetic field simulation. Following the initial design phase, the company employs Opera’s 3D Modeller to create a very detailed geometric model of the proposed design, from which is generated a mesh of finite elements for numerical solution using the static electromagnetic field simulator. Until the advent of the latest version of the Opera-3D Modeller, the finite element shapes were limited to tetrahedra, but can now include other shapes such as hexahedral, prism and pyramid elements as well. Opera uses the term ‘mosaic’ to define a mesh which can include this mixture of element types.
This advanced meshing feature helps Cryogenic to achieve the very high accuracy of simulation it needs much more quickly. According to Jonathan Cole, Magnet Consultant at Cryogenic, “the ability of Opera-3D Modeller to now handle hexahedra in conjunction with automatic tetrahedral meshing is extremely interesting. In the past we have handled this aspect of finite element analysis manually, using as many hexahedra as possible to secure the best accuracy. This is an extremely time-consuming process, with each iteration typically demanding overnight simulation. We now expect to produce even more accurate results within hours. Given that design optimisation involves a very large number of iterations, an optimal design often took weeks to achieve. By shortening simulation times and homing in on the best solution more quickly, we expect to shave weeks off our development schedules.”
Opera-3D – which is part of the Vector Fields Software product line – has been used by Cryogenic for several years to simulate, analyse and fine tune the field profiles of many of its high power magnet systems. A typical example is the 5 Tesla muon detector magnet installed at the ISIS pulsed neutron and muon source at the Rutherford Appleton Laboratory. This split pair magnet is used on the muon beam line for a high-field muon spectrometer known as Hifi, and creates a highly homogenous and stable longitudinal field over the sample volume. Optimising the parameters of the magnet, especially with regard to fine tuning of the magnetic field profile, was a highly iterative process involving many lengthy simulations and weeks of development effort.
Recent versions of Opera-3D include a powerful additional tool known as Optimizer, which is designed specifically to accelerate the design process using finite element analysis by automatically selecting and managing a large family of goal-seeking algorithms. The tool can support multiple design goals – even when they compete with each other – and, by eliminating the need for manual intervention, gives the designer freedom to concentrate on other aspects of the design project while minimizing the number of numerical simulations needed.
Until now, Cryogenic has devoted considerable time to repeating simulations until a satisfactory solution was found, often having to devote weeks of effort to solving a particular problem manually. As Jonathan Cole points out, “Optimising the shape and placement of magnetic material to achieve the desired magnetic field profile necessarily involves a series of design compromises. The Optimizer software is one of the key reasons we chose this package. It enables us to reduce the number of simulation runs, while at the same time giving the option of considering a wider set of potential solutions, resulting in a much faster and more cost-effective approach to magnet design.”