Monday, October 29, 2007

Why Engineering In The Future Will bring More Disasters.

Cheaper, faster and better.

That is what engineer's are hearing from their management. Management is of course made up of MBA's and accountant's, who have no idea what it takes to engineer a safe, reliable product. All their interested in raising their stock price. At the cost of civilian safety. Don't believe me? Just keep your eyes open on the number of product failures that arise these days. If its not a design fault, its usually manufacturing where its easiest to cut cost.

Come on, after all, who will know whether you've added in 2 or 3 cups of water to your cement mix. Will you? You wont. But your house/building/bridge will pay the price down the road, usually taking lives with it. Who do you blame? The engineer or the constructor?

CAD, CAE and CAM companies are all chanting the mantra, "Faster, Cheaper and Better". Their all rushing to create products that work like a black box. The vendors all say that you don't have to know what goes inside the black box, all you have to do is to come up with your CAD model, apply your loads and constraints, then believe the results! Yeah, right, like that was what we were taught in school.

But the irony is that management is buying all this bull shit. In their blind haste to bring products to the market faster so that profits go up, their stock price rises, and they get a huge Christmas bonus every year, management is taking risks and closing an eye to the complains and constraints raised by their engineering team.

Speaking from a point of view of an ex-CAE vendor (technical manager), I have seen engineers design products without knowing why they chose one particular type of element over the other. They don't know why their using a 8-noded quadratic element over a 4-noded element and they've been using the software for some time now. Okay, most of them as not as ignorant as this person but I can assure you that majority of they don't know when to use the Arc Length method in a non-linear analysis or do they have a rough idea on how the Netwon Raphson method works in a nonlinear analysis?

The worst is during results interpretation. Everyone feels good when the stress plot shows more blue than red (in FEA, red color indicates regions of high stress while blue color indicates region of low stress). Some of these engineers dont even think whether their results are plausible. They just note the highest stress value, put it into their report and sends them across. The senior engineer verifying their work looks over the loads and constraints and the corresponding results and OKAY's them. This report then goes through the channels to other vendors who based their design on this report. The young engineer who just did this report doesnt know the cascading effect his report might cause if any of the numbers of wrong.

With the use of tools like CAE, engineers are not thinking anymore! We're so used to sitting in front of the computer and looking at this pretty stress plots that we dont spend them time to evaluate whether our results are reasonable. How many of us look at our stress results and think about the inputs that went in. Did we make any mistake when defining the loads - was force inserted in the positive x direction, as per the design specification. Did we make a mistake with gravity? In ANSYS, gravity, surprisingly, always acts in the positive direction of an axis, while in MSC.Software, Abaqus and other FE softwares, gravity is just a force and its direction depends on whether its acting upwards or downwards. When gravity is coupled with other forces, and is applied wrongly (or in the wrong direction), it lowers the maximum stress in the model - in other words, a wrongly applied gravity could make an otherwise failed design pass! This is what I meant by checking the inputs thoroughly!

When approached by CAE vendors especially, ask them how you can check and verify your work. ANSYS is stupid in the sense that there's not output listing of your loads. You just have to ensure that you have inserted them correctly in the GUI. Software's like LSDYNA (the best output I have seen), Abaqus and MSC.Software all list the inputs in a text file that you can verify that you've inserted your loads and constraints properly. Always ask for this. make sure you know what is going into the black box you're using so that you can be reasonable sure of the output.

You're an engineer. Be proud of the fact that the safety of someone elses life is in your hand. Even if your design didn't for the public consumer, think of the person installing your design. He might be doing it alone. What if a bracket breaks and injures him - he will have that injury for the rest of his life while you still sit comfortably in your chair going on designing other stuff, oblivious of what harm your design creates.

So, don't be pressured by the CAE vendor to purchase their so called easy to use software. The risks are just too great. And explain the consequences to your management on the perils of buying something that you cant verify the inputs and outputs. As an engineer, you must know all inputs and outputs. And the CAE software that you're using must be able to provide you with such information! Don't let it leave you blinded.

No comments: