A common source of confusion in doing a failure analysis is figuring out how strong the material was supposed to be. There are a lot of sources of material properties. MatWeb.com, for example, has thousands of different materials. The problem is that they give such a large range that it is almost useless for either design or failure analysis.

For example, I’ve been working on a finite-element analysis for some 430 stainless steel. MatWeb actually has quite a bit of mechanical property data for 400-series stainless steels. They list 165-1,900 MPa or 23,900-276,000 psi for the yield strength. This is a factor of 11.5!

I was thrilled to see that they even have some fatigue-strength data. This characteristic is listed as 85.0-772 Mpa or 12,300-112,000 psi. Again, a huge factor. In order for the data to be useful, you have to be able to correlate multiple properties with each other. The overview data tells you how many materials they averaged to get the range and average shown. In general (for ferrous materials, for example), as the yield strength goes up, the elongation goes down. This is not the case, however, for many wrought aluminum alloys. The detailed material may be viewed by clicking on the individual materials.

Great! I found some 430 that was annealed and then cold worked to HRB 89 with a tensile strength of 78 ksi, close to the 76.4 ksi value found in my material when tested at my lab. However, their yield strength was 64 ksi, and mine was only 52 ksi. My total stretch was 28%, and theirs was 32%.

What form was the material? Mine was a wire, and I suspect that theirs was a sheet since it is pretty tough to get a yield strength on a wire. (We used a nonstandard procedure, and it was a PROCEDURE!) I kept on looking and did find some wire, but it was dead-soft annealed or soft temper. Anyway, the diameter was way bigger than my material. Neither of the references listed what I was really looking for: fatigue-strength estimates.

I suspect that many designers don’t go to the trouble of looking up the individual materials, which is understandable. I actually consider it a miracle that I found something as close to my results as I did. I suspect that most designers will just use the average. They don’t necessarily understand why this may not be a good idea. On the other hand, there are probably as many designers who know beforehand what material they are going to use. They may only consult the data sources when the prototypes fail.

As pressure for cost reduction pushes us to design with less margin for error, maybe the ratio of people who check out the properties prior to specifying a material compared to those who don’t is increasing. It’s hard to say. In the absence of codes requiring strength calculations prior to building, it probably depends on the industry and recent history of pain within the company.

Another point I really wish to emphasize here is that the published data is for standard test coupons. They rarely truly represent the characteristics of the manufactured component. Whether we are talking about metal or plastic parts, the shaping procedures that create the final object have an effect on the strength.