One of the most important things that scientists have come to understand in the last few hundred years is that the nature of scientific truth is always going to be something that we can pursue and approach but never attain or possess. In his very interesting lecture series, available through “The Great Courses,” Professor Steven Goldman of Lehigh University tracks the history of this dialogue from the time of the ancient Greeks until today.
For most working engineers, objectivity is something that “we should try to achieve.” I’d like to suggest a new goal for our technical work, which is clarity of thought. The concept of objectivity has been dealt a death blow by modern understanding of how the world works. There is simply no way to have a perfectly objective observation of anything else. The very fact that we ask a question a certain way makes it more or less difficult to see certain aspects of the relevant situation, as David Levy puts it in his famous book To Observe is to Disturb.
For very simple situations, this may not matter much. But for a failure analysis, we are obligated to do everything in our power to try and understand what went wrong. We can’t do this reliably well if we can’t step out of the artificial distinction between objective and subjective. We can learn to think more clearly about many situations that confront us.
A good way to start thinking more clearly is to practice seeing things more clearly. This is particularly important since failure analysis of engineering components and assemblies/structures always starts with a visual examination of the physical evidence. Sometimes we don’t have the physical evidence to start, so we perform background documentation review first. Especially in these cases, we must have a way to compare what we think we are going to see with what is really present.
In order to learn to see more clearly, we need to distinguish between looking at something and looking for something. If you don’t know what potential things might have caused the problem, then you aren’t a failure-analysis practitioner. You’re just someone looking at a failure. If you are a failure-analysis practitioner who is familiar with the type of failure in question, then you have somewhere to start to try to figure out what might have happened. You’ve probably seen other similar situations. You know what to look for. This is good. This is necessary. This is what eventually gives the expert a big reason for being called an expert. But this is only part of the story. This is “looking for.” We have a sort of list in our minds, or on paper somewhere, that may include the types of common failures for the type of component in question.
Let’s say it’s an externally threaded fastener (a bolt) that is not a single, intact object. The most common type of failure for a bolt is probably a fatigue crack due to improper clamp load. So we can look to see if there is evidence of a fatigue crack. We can look to see if there is evidence of inadequate clamp load, such as rubbing on the threads or damage to a hole that the fastener was inserted through. Excessive clamp load is harder to detect, unless it is extremely severe. We can look for evidence of “the opposite” of a fatigue crack, which is a single load cycle that caused the crack. There might be some obvious deformation of the fastener.
But we also need to look at the fastener. We look at the fastener to see its appearance. To do this properly, we need to “turn off” the technical dictionary in our heads. We need to resist using the easy vocabulary that is the jargon we hear among people who do failure-analysis work. We need to take the time to try and describe the part using regular words that a high-school kid would understand. Learning to use words such as round and trying to being more precise (such as specifying cylindrical or circular or spherical or spheroidal) will increase not only how other people understand what you say, but will make the images in your own mind more clear. Words such as sparkly, shiny or dull, and phrases such as uniform color or splotchy or stained, are useful descriptions. Color words are important too. But we are not writing a fashion catalogue. “Greenish” is better than “teal,” even at the expense of exact description. “Reddish” is better than “rusty.” I have seen a lot of reddish stains on steel parts over the years that turned out to be dirty, degraded oil and not iron oxide at all.
It is during the visual inspection where you are looking at the part that you have the greatest chance to notice something that is not expected. When you are looking for certain features, even prominent features can remain invisible if they are not on your list. How many times have you missed something on a grocery-store shelf because they changed the package color or design? What if they changed where they put it on the shelf? Our minds are hard-wired to take many shortcuts to help us get a grasp of our environment and surroundings, and taking the time to look around to see what is present is step one in learning to think more clearly.
Try this exercise:Look around you to try and notice as many things as you can in 20 seconds. To do this, you will need to look up, down, in front and behind (yes, stand up and make sure you turn around). As soon as your eyes focus, move your gaze to another location. If what you just focused on was near, look at something far away and vice versa. If you found something light in color, look for something dark. If you noticed a detail, try to focus on a large object.
When you are done, write down as many things as you remember seeing in that time. Don’t look around for new things. You might want to try this two or three times and see if you notice new things. Ask one of your friends to do the same thing from the same space. Compare your lists. I would be very happy to have some people post comments on their experience with this.
I have been doing this exercise with various failure-analysis training classes and thinking-skill seminar attendees for about 15 years. It is always very interesting to find out what people saw and did not see!