_{original text was writen 2011.06.21}

Today we will talk about computer calculations.
And our talks would be not about “Start/Programs/Accessories/Calculator”. And even not about more complex calculation programs: Maple, MathCAD, athLAB and other “Math”something. We will talk about CAE – software for engineering calculations. ANSYS, ABAQUS, NASTRAN and other from the list.
The main question of this publication will be about whether you can trust in the results of the calculations in CAE. And why not?

Foreword

I’ll want to start with an abstract example. Imagine that Uou had decided to “translate” an important text from “Your Native Language” into “Another”. You don’t know this “Another Language”. Even if you had studied this Language many and many years ago. You can do this translation in some ways:

• order a translation to a professional linguist-translator;
• order a translation for a non-professional. This way will help to save some moneys;
• translate by yourself. It’s free because You can use any online translator.

All ways have they own benefits and flaws:

1. Professional translation has only one benefit – quality. The translator-linguist has a special education. He values ​​his reputation. In that reason he always tries to translate well. Hi will takes into account all the nuances of the text. Also he uses professional dictionaries and verification services. The problem is the high price.
2. Non-professional translation can be much cheaper. You don’t have to waste your time. The problem are on the detail. If You doesnt have good knowledge of both laguages – You can not always distinguish such a translation from a quality one. You cant compare how good or bad this translations results. But result can be bad. Realy bad.
3. You can translate the text by yourself using a software for translations. If a free web-translator is used, you will spend no money, only your time. A little time (less then with 1 or second way). And the only one problem – is a quality. The quality is usually poor. If you don’t belive, you can check. Translate one text for several times into another languages and back to original. Than vice versa results. In result You will get text with no sense (nonsense). If you don’t want to do re-re-re-translation – search examples for “Funniest Examples of Translations Gone Wrong”.

Why am I described this situation? Because it’s an example of situation that is very simple for understanding, and it is absolutely the same with engineering calculations:

• There are people with a diploma/PhD in field of Math, Computational Mechanics, CFD, and so on.
• There are people who had learned some CAE software during their work. But CAE are not their specialty, only additional tool.
• And there are some CAE software. CAE as a tool wih can help in solving engineering tasks
  • Some Software are similar to “dictionaries”. To “translate” the entire text, you need to “translate” each word one by one (and than do lot of other routine).
  • Others are like automatic translators: past the text + press the button = got the answer.

And now, We need to return to the question at the very beginning of this text.

Most people believe in the results of calculations with CAE. This is especially true for Young people and Big-Bosses.

Hovewer! Do You can understand that fully-automatic translations or translation from the hand of unprofessional translators is worse than translation which was done by profesional translator-linguist? And if answer are “Yes” – you understanf differences, why do you think the situations with CAE calculations can differ? Wy people belive in result of calculations, only because thay was obtained with a COMPUTER?
We will not recieve high-quality text after automatic translation. We cannot press some buttons and automatically recieve the correct result after CAE calculations.
You can’t trust in Software or Computers. Even if they don’t lie. But, in that reason who can be trusted? Only people. It doesn’t matter does this people use some software or not.

Why it’s no mater? Because not a Computer, not a Software can give you anything without People. Computer made by People. Software created by People. CAE calculations are also maded BY People. And even if we can trust for people that create those Softwares and Computers, we need to trust a one another Person. Person who push the buttons for obtaining Results with this Software and Computers.

So “Distrust” for CAE software – is mean a distrust of the Persons. The person who uses it.
If You are this Person who use CAE Software, you MUST distrust Yourself either. Why? Because You don’t know the “languages” you want to “translate” your “text” into.

Two opposing opinions

Here are two type of CAE professionals.

1. CAE “Engineers”: A persons that has been doing engineering calculations for many years. They knows and understands different disciplines – numerical analysis, strength, dynamics, theory of elasticity and plasticity. They knows how to provide calculations and analysis using the finite element method. They know CAE Software – ANSYS, ABAQUS and others are based on FEM.
2. CAE “Sellers”: A persons that has developed and been selling CAE Software for many years.

Lets ask those specialists for opinions about calculations in the “CosmosWorks” or similar CAE Software for Designers and Express Calculations. The question is: “can we trust in those calculations”:

1. CAE “Engineers”: This person will say that it is impossible to trust.
2. CAE “Sellers”: This person will say that the calculations can be trusted.

Both Opinions are fully opposite. And both Opinions are fully true.

How it can be so? Two different opinions, and both correct. It doesn’t happen! And this is also true. 😉

Let’s consider the issue in detail.

If the calculation results in CAE programs would be incorrect, the CAE Software developers cannot be successfull in sold they software for such many years. SolidWorks Simulation has been sold under various names for over 20 years. Its basis on the Cosmos/M solver (GEOSTAR). Cosmos/M are well known to specialists who are engaged in FEA. Especially for university scientists. Cosmos/M has feature to use of different solvers for different tasks. thats why Cosmos/M has possibility to solve large and complex problems on low-level computers.
But we don’t talk about Cosmos/M. We talk about SolidWorks Simulation!

Yep! Hovewer, we need to understand that SolidWorks Simulation is just an “interface” (Front-end). It automates the process of setting initial data. And then you use a slover from Cosmos/M that relies on proven theory. Lot of people have used this solver to do Billions of FE Calculations. And when we says “Billions” – this is not an exaggeration.

Another example. There are relatively new CAE programs on the market: NX NASTRAN, appeared in 2003 and Autodesk NASTRAN in 2014. Why new? Because ANSYS has been around for over fifty years. But, for examole, NX is a “fork” of MSC NASTRAN version released in 2003. MSC NASTRAN for now have more than 60 years. After 2003, several clones began to develop independently. Autodesk Nastran it’s a NEi NASTRAN that exist from a 1991 and in 2014 was bought by Autodesk. So all these programs are not newbie to the market.

Or we can talk also about Autodesk Simulation. This name was used for “prepost module in Autodesk Inventor. The solver was ANSYS. Later Autodesk start to sell another product with this name. Those product was based on Algor. This is also a well-known brand in the CAE market (from 1984). Unfortunatelly Autodesk kill this profuct. Hovewer we are talking about other things, so…

I hope you believe: there are many such examples about good old “newbies”. And all these CAE Software have been on a market around for a long time. They have been tested many times in real “combats”. They were certified by serious organizations. So, they are of very good quality. And you must know that, if you submit the correct data, you will get the correct results.

And here is the place the difficulties begin. This is why “calculators” (expiriensed FE Engineer) don’t believe in “quick” analysis. I will repeat again:

IF you submitted the correct data, you will get a good result. IF.

Do not trust express calculations

Big, Scary and Complex CAE Software (ANSYS APDL=Classic, Abaqus, Nastran/Patran, Femap, Cosmos/M) require a lot of data. You must know all data BEFORE you will start. You must know certain place to enter those data. You must do it in certain order. Otherwise, you won’t get ANY results.

In “express analysis” CAE, most type of data is set automatically. You want change them? Firstly You must find them. In some “fastest” programs, those data cannot be changed at all.

As result of such “simulations” and “CalcuLIARsion” is some kind of “colored pictures”. Hovewer if You simulate something with input data that is not fully clear (for you), You must know that you will received something that is not clear, too.

CAE Software has no knowledge , experience or understanding. CAE has only the “dumb” algorithm of work.

An inexperienced user does not know what data was used for the calculation. He does not know if additional data is needed. He doesn’t understand why they should be changed. Doesn’t know how to change. And the sadness moment of truth – he also cannot understand the result of the simulation.

It’s like driving a car with your eyes closed. Imagine. You got into the car blindfolded. You are driving fast on a mountain road. On one side there is a wall. On the other hand, a fall. And computer tells you where to turn.

Are you ready to drive like this? No? So learn to ride. In reason lern basics of simulation your engineering problems. And understand why experts do not believe in your “express calculations”. Why they ask you about calculation parameters, elements types, loading conditions, solver parameters, and more. They don’t want to make you angry. They don’t want to seem smarter. They want you not to fall into the abyss or hit the wall.

The text is partly based on the questions from the participants of the last seminar (“Reliability of strength calculations in the CosmosWorks environment”). And partialy based onreview of bachelor’s degrees thesis of our students.

P.S. Gentlemens and comrades! Learn theory! And may accuracy will be with you!

P.P.S. This text is result of translating in second way $)