A fast and easy way to come up with the best design is to check it out while the design is still in CAD format. For this reason, computer aided concurrent analysis became an integral part of the design process. After all, it's the design engineers who create designs--it's only natural they, rather than dedicated analysts, also analyze and modify those designs.

Why do design engineers find it difficult to analyze their own designs and what analysis tools can make concurrent analysis easier?

Not exactly birds of a feather

There are fundamental differences between analysts and design engineers running analysis. Design engineers prefer simple types of analysis, but want to use their CAD models directly for it, with no need for modifications like defeaturing, idealizations and clean up. While avoiding Finite Element Method (FEM)-specific issues, design engineers need to produce accurate results necessary to make design decisions. At the same time, analysts need to deal with complex custom simulations, which call for extremely flexible analysis software with features like application programming interface, extensive library of finite elements, etc. Due to those differences, the same analysis software can't serve both analysts and design engineers well.

Responding to market demands for analysis software suitable for design engineers, the industry developed scaled down versions of their mainstream FEM-based programs. Those user friendly programs come with an interface similar to a CAD interface and meshing is done in the background, with no users input required. Idealizations like beams, shells or 2-D elements are discouraged or not available at all. At the first sight, that is exactly what design engineers have long asked for--a simple to use software that would not overburden users with non-value added, FEM specific issues. Unfortunately, this customization of mainstream FEM based products is often quite superficial and just hides the FEM specific tasks. That makes problems worse because it takes analysis control away from users. As long as analysis software is based on the FEM, it will always be subjected to all of the limitations of mesh-based analysis.

So what are the prospects of providing design engineers with easy to use and, at the same time, reliable analysis tools?

Don't fence me in

The Precise Solid Method (PSM) doesn't use any mesh at all and, therefore, it's free from all mesh related limitations. Defeaturing, idealization and cleanup, required to make geometry meshable for FEM are not even applicable in PSM. The PSM will analyze the geometry of any complexity with no modifications required. Moreover, unlike in FEA, the complexity of geometry does not drive the number of degrees of freedom (DOF). The number of DOF necessary to solve the problem depends on the complexity of the geometry. Complex geometry is more likely to cause higher stress gradients, which in turn requires more DOF to approximate those gradients.

In PSM, discretization involves splitting geometry into subparts to keep approximation functions simple. However, because stress patterns--not geometric details--drive the number of DOF, those regions of the structure which are idle in the stress sense, need no splitting and very few DOF to model their behavior. The discretization process is governed by factors other than the smallest edges and surfaces, as it is now in FEM.

Clear communication

Design engineers know how parts interact with each other, how parts are constrained and how they are loaded. However, that knowledge must be translated into the language of FEM analysis. Boundary conditions are often just a consequence of geometry--pressure is represented by a set of concentrated forces, a cantilever I-beam is modelled as a line-by-beam element supported in one point, etc. Several decades of experience indicates that modelling boundary conditions is where severe mistakes are most often made. In fact, the ability to model realistic boundary conditions is a derivative of the ability to provide a realistic geometry representation. The ease of use and functionality are matters of implementation and not of the particular analysis method.

So, at present, there is no perfect design analysis tool for designs engineers. Both FEM and the PSM offer attractive features and the user's choice between those two can be based on complexity and meshability of the analyzed geometry. It is certainly tempting to speculate about future developments. FEM based technologies, after several decades of intensive growth, have now reached a plateau. The PSM seems to have the best potential.

Edited from a longer paper by Yuri Apanovitch of Procision Analysis Inc. in Mississauga and Paul Kurowski of Design Generator Inc.