By Colin E. Brench
F
or those new to the idea of applying computational electromagnetics (CEM) to solving EMC problems, two things become immediately apparent. First, there are a variety of numerical techniques available, each with unique strengths and weaknesses that need to be understood. Secondly, there is the question of how these techniques can be applied to a particular problem. These two issues create a tough barrier that must be overcome before further investigations of CEM are made. However, once these have been addressed it becomes easier to determine if CEM modeling will be a beneficial tool for any given situation.The numerical techniques used in CEM are the means by which a set of equations is solved. For all EMC problems, it is Maxwell’s equations that are being solved. However, in many cases the full set of equations is not necessary and constraints or simplifications can be applied. It is the addition of such constraints that result in the variety of techniques in use today for EMC modeling.
The three most common numerical techniques used in EMC modeling are the Finite Difference Time Domain (FDTD), the Finite Element Method (FEM), and the Boundary Element Method (BEM), more commonly though less accurately known as the Method of Moments (MoM). These techniques are all used for solving a variety of field problems. Other techniques include the Partial Element Equivalent Circuit (PEEC) method and the Transmission Line Method (TLM). These two techniques were especially developed to provide a link between physical structures and circuit models.
The most general solution to an electromagnetic problem might well be obtained using the FDTD technique, as there are very few constraints required in formulating a practical EMC tool. The price to pay for the flexibility afforded by FDTD is the magnitude of computational resources required for a given problem. A more specific tool such as the Numerical Electromagnetics Code (NEC) uses the MoM. This tool was written to address a specific class of problems, namely antenna modeling. NEC has considerably more constraints within its code but because of this it runs much more efficiently. Different formulations, also using MoM, are written to address other specific tasks such as solving coupling problems.
The other numerical techniques mentioned have also been developed with the appropriate constraints to efficiently address a wide range of EMC related problems. This area is rapidly developing, with new customizations of these techniques being added frequently. Therefore, it is often equally important to consider what kinds of problems a tool is intended to solve rather than simply which technique should be used.
The application of numerical techniques to solving an EMC problem is the second hurdle to face. Some problems match so well to an easily implemented formulation of a particular numerical technique that the electromagnetic problem and numerical technique have almost become synonymous. The antenna modeling tool, NEC, is a good example of this, where MoM has become so closely associated with NEC that many don’t realize the wider range of applications that the MoM can be used to solve.
The creation of EMC models has much the same procedure as the preparation of a problem to be solved by any other means. The first three steps are to determine what is wanted from the model, what is known as a starting point, and what pieces of information might be missing. This information is then incorporated into the framework of the available modeling tools and the best method of solving the problem can then be determined. While it is the clear goal of TC-9 to help develop all aspects of EMC modeling, the primary goal is to help EMC engineers understand the subject so that they can decide on its relevance to their particular problems. Not everyone will need EMC modeling, and EMC modeling is not needed to address every problem.
While there is still much that can be done without the need to apply modeling, this is slowly changing. Design and analysis in the related discipline of signal integrity (SI) were once also performed by following rules and through the use of basic equations. However, as data rates increased and tools improved, SI modeling grew from occasional use to a mainstay in the field. It is to be expected that EMC design and analysis work will develop in the same manner. A more in-depth overview of several common numerical electromagnetic methods, listing of available commercial and public-domain codes, and a brief bibliography of CEM books is available at the web site https://www.emclab.umr.edu/numer.html.