Other services

Contents:

For more than three decades, I have thrived by the motto "No simulation is an island, entire of itself." That is, one must always prepare for loopbacks and later reuse (with variations) of any small part of the sequence of events leading to any full-blown simulation.

The standard simulation workflow of "preprocessing-simulation-postprocessing" should only be seen as a teaching tool. Newbies need teaching tools, so the world is full of clickety-click software applications that ease the start of the learning curve.

Further up the learning curve, software applications for beginners start to annoy you. Creative minds should never need to waste their time repeating the same sequence of mouse clicks, spending man-hours on what could have been computer-minutes. Instead, nerds like me start dreaming of addressing simulation software with programming code instead. This is called scripting. And scripting is what it is all about.

When you have scripting, you almost have automation

If you don't want automation, your manager probably does. Most development departments repeat the same processes over and over again with slightly different geometries and different parameters.

As stated elsewhere on this website: already on the first delivery from Simxon, you are close to having the entire procedure automated.

Finding the right approach to a tailor-made software application is an iterative process where Simxon should be paid by the hour. Simxon will probably leave the application at a development stage where it works as intended but is unappealing to the end users. You can then (possibly with a more attractive cost-benefit ratio) pay somebody else to code exactly the user interface that you want.

Such a tailor-made automation solution from Simxon will never pull a license while operating!

Templates to start from

If the concept of "automation" seems a little too highbrow for you, we could call it "templates" instead. When you have realized which kind of simulation you want to repeat over and over again, Simxon may help you establish a recipe for the procedure. Also here, it may come in handy that you need not pay for the everyday use of your simulation tool.

Extraction of key values

At the end of the day, an engineering manager wants to make decisions: "Should alternative ZQ514 be preferred to alternative WX287 or not?" Pretty pictures rarely help when dealing with such choices. But a number which indicates some kind of a (possibly inverse) score may be very easy to handle. In my years as an employee, I have frequently entered the office of a colleague, informed the colleague of one number and left the office again.

For that kind of communication to work, the score number must be clearly defined and it must be considered highly useful by all relevant stakeholders. Simxon can assist in defining such score numbers and in calculating them automatically.

History tracking

On the "Decision-making support" page, I mention the fact that mechanical engineering lives in a different world than most other modern pursuits. That goes for the attitude to the concept of volatility, too. Today, almost everyone thinks that volatility is an inevitable fact of life and therefore a virtue. However, in order to be able to step back in the decision-making trees of mechanical engineering, all tree nodes should be clearly defined and available at any time. Simxon applies and recommends a way of working with simulations which fullfills requirements with respect to history tracking that you may not even know that you had. As always: ask some questions.

FEA and CFD are just TLAs...

The IT and numeric analysis toolboxes which I use for simulating solids and fluids may be used for other purposes, too. I cannot know what you need until you tell me. But I can tell you this story from my LinkedIn profile:

Balancing an asymmetric, partly submerged rotor during the design phase: I got an idea slightly outside my ordinary subject area (outside anybody's subject area). I then single-handedly carried out all project phases. I coded the final software product in the Siemens NX API with Python 3 as the programming language. In an intermediate phase, I coded the same thing in Open CASCADE with Python 2 as the programming language. I consider the whole project my masterpiece when it comes to providing value for money.

See also the paragraph on "Dimensional reduction".

Development projects with academia

Academia is driven by enthusiasm. If you want to establish a simulation software development project with both commercial and academic stakeholders, the option of relying on open-source software may be preferred solely because of the inherent enthusiasm of the open-source world. But the possibility of sending code snippets between otherwise identical virtual computers also speaks in favor of using open-source software for development projects between commercial companies and academia.

Smart devices

This may sound far-fetched, especially because I expect never to program such a thing. On the other hand: with open-source software, you have a possibility of cutting off unnecessary program modules and make something work with a small footprint on a small device. With license-bound software, you never have a chance.

Poor man's optimization

You can buy optimization of simulation solutions as a licensed software package. One reason why I have never participated in buying such a package is that each time I presented a solution which was optimal in some way, I was told: "We don't want it to look like this."

Once I presented a home-spun Pareto curve created by ad-hoc tools. The mere thought of deliberately balancing two parameters against each other seemed so repulsive that the decision became to change some other parameters which I had been told were fixed parameters...

If your organization makes fundamental decisions in a somewhat opaque way, extensive optimization studies may be of little value. But small, ad-hoc loops may be worthwhile. Ask Simxon for assistance.

Dimensional reduction

The computer tools of today can model our common three dimensional reality conspicuously as such. However attractive this may seem, there are also elements of deceit to it. If the effort to crunch a lot of numbers exhausts all our resources, a more general view may get lost in that process.

Enter dimensional reduction. If a component fails at a load of 93.5 N or if it has a stiffness of 837 N/m, why not focus on these numbers, then, and de-emphasize a comprehensive 3D finite element analysis as a noisy representation of those two figures? That would be a reduction to 0D. Reduction to 2D (plates or shells) or 1D (beams, rods or rotors) is possible, too.

Thoughts as these lie behind 0D standards like Modelica and FMI which enable very powerful model assemblies to be created and analyzed. Sometimes, a single component in such an assembly is called a surrogate model. See here for an alternative description of these issues.

You need scripting and a mature overview to create and maintain simulation based 0D components. A task well-suited for Simxon as demonstrated with the introduction of SimxonCARE.

You will find my latest contributions to the art of dimensional reduction here and here.

Other kinds of integration

During my years as an employee, I often heard a colleague lament that this IT system or device unfortunately had no connection with that IT system or device. Sometimes, I created the connection then. On a recent occasion, I even did it in public.

Teaching

I believe that the way I teach resembles the way I communicate in writing. The latter you should know about by now.

A pretty picture. For more pretty pictures, see here.

Creative Commons License
Website for the company Simxon by Kim Ravn-Jensen is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Permissions beyond the scope of this license may be available at info@simxon.com.