Robustness Evaluation – Why Bother?

This article will explain how ANSYS optiSLang can be used for robustness evaluation in virtual product development.

A successful product. Isn’t that the goal for every product company? It begins right from the step where engineers come up with world class product innovations to getting the right marketing mix that brings commercial success. Is every product successful? No. Is every product with a great design successful? Maybe.

The Symptom
Robustness Evaluation - Why Bother?
Courtesy: Android Authority

More often than not, we find market leaders stumble with product failures. The infamous Samsung’s Note 7 will come to your mind instantly. Hundreds of users were at the forefront of dangerous incidents where phones caught fire due to short-circuiting. Samsung conducted severe internal testing and several independent investigations. They found that, in certain extreme situations, electrodes inside each battery crimped, weakened the separator between the electrodes, and caused short circuiting. In some other cases, batteries had thin separators in general, which increased the risks of separator damage and short circuiting. Economics-wise, the incident caused Samsung to recall 2.5 million devices, lose over $5 billion and damaged its reputation.

Faulty Takata airbags’ inflators contained a defect that cause some of them to explode and project shrapnel into drivers and passengers. 50+ people worldwide lost their lives due to this design failure. 70 million Takata airbag inflaters were to be recalled at a cost of $9 billion to its automaker customers. For a Tier-I supplier, this liability was so huge that they filed for bankruptcy.

Such glaring errors after product launch, with severe economic implications, aren’t limited to Samsung and Takata alone. Honda, Michelin and many more companies have been involved in product recalls due to design failures.

Obviously, such design flaws need to be mitigated. Isn’t it?

The Probable Solution

To preempt design failures, today’s engineers use state-of-the-art engineering technology. Traditionally, product development teams used extensive prototyping and testing to validate design variants during the design life cycle. Of course, this is cumbersome, expensive and time-consuming.

Over the past few decades, engineering simulations have opened up a whole new range of possibilities for the design engineers. ANSYS, Inc., the market leader for engineering simulations, provides state-of-the-art technology to simulate systems involving mechanical, fluid, electrical, electronic and semiconductor components. With added insight, design engineers are able to test a lot more design variants on a virtual platform using this technology.

Consequently, the benefits – innovation, lowered cost of product development, higher product profitability and faster time-to-market. The staggering economic benefits and tremendous value on the offer have prompted several product companies to introduce simulations upfront using a Simulation-Driven Product Development approach.

Companies like Samsung and Takata were power users of engineering simulations. They used technology extensively in their design phase and perform virtual tests to validate designs. Only validated designs were put through production, QA and then sent off to the market. Despite simulating and validating designs, these companies witnessed monumental product failure in the market that caused loss of life, led to economic losses and damage to their reputation.

If they used simulation-driven product development, what went wrong?

The Cause

While the probable solution can mitigate and even eliminate design failures, there are other forces at play that you will need evaluate carefully. Hence it is imperative to understand the root cause for occurrence of design failures despite conducting extensive state-of-the-art simulations.

Many design engineers often undermine or do not consider one important aspect due to lack of proper understanding. Variability. Just as design parameters such as thickness or physical loads can be varied to test different design variants, some parameters display inherent variability.

Let me explain it with a material parameter: Young’s Modulus. If you’re an engineer by qualification, you would’ve come across the Universal Testing Machine (UTM) in your freshman or sophomore year of college. To test the Young’s Modulus of any given material (say steel), the UTM pulls a material specimen at extreme ends to create tension. Using mathematical calculations, you’ll arrive at a number close to 210 MPa as the Young’s Modulus of mild steel. Let’s say you repeat this test for 99 other specimens of the same material. Each test result will be different and it will never be the same. Other than the odd case of a faulty UTM apparatus, there’s only one reason for that. Natural Scatter.

The Hero: Robustness Evaluation

Such variability (statistical) will lead to variability in the performance parameters of the product. Obviously this is quite important and engineers need to assess designs for variability well ahead of product launch. For variability, you have only one way to assess designs for product failure or risks: Robustness Evaluation.

Robustness Evaluation with ANSYS optiSlang

The preferred choice of tool for robustness evaluation is ANSYS optiSLang. For better understanding, there is a lot of material available in more detail. Instead of reading, you may also want to consider watching these webinars here and here.

Can you attribute lack of design robustness to any other product failures that you have witnessed? Do you have alternate views? Please let me know in the comments section.

Share this on:

Leave a Reply

Your email address will not be published. Required fields are marked *