Donnerstag, 17. August 2023

Fact check: No, this work does not falsify my thesis.

On the contrary. Its evidence that my thesis is right.

The speaker of an organisation which is financed mainly by manufacturers, sent me following mail:
  
"Rope life is not unlimited. ... For me ... Bernstein etal’s Polymer Degradation and Stability justifies a 10-year lifespan. I find figures 3 and 13 particularly compelling. This Berstein article is one of a series published by Berstein & colleagues at a US national lab regarding the long-term degradation of nylon under a variety of conditions."
 
The speaker refers to a paper, the full quotation is:
 
R. Bernstein and K.T. Gillen: "Nylon 6.6 accelerating aging studies: II. Long-term thermal-oxidative and hydrolysis results."

 The article is published in

https://www.sciencedirect.com/science/article/abs/pii/S0141391010002636?via%3Dihub

After reading the whole article  (which is not free available but can be requested) I got the feeling that it actually states the opposite. Confused, I contacted the authors and asked some questions.

I received an email from Mr. Bernstein.

Here are some of the questions to the author and his answers:

Regarding Fig. 3


Question: I am missing the "stars", marking the 37 degree results. 

Answer: Yes, there are more, but not a lot, at the -very early- part of the curve, and buried. If you look carefully at figure 1 you can see them, but they are buried in this plot under other data.

 I’d also argue that Figure 3 suggests very little change at room temperature storage after 10 years. It is a log graph, so after ~20 years it may lose on the order of ~10% according to that graph. That means that whatever application has less than a 10% safety margin if you have to stop using them at 10 years? That seems strange to me."

Regarding Fig. 13:

Question: Do you have an explanation for the scatter?

Answer: Scatter happens. Also because these were parts stitched together….unstitching was a messy process, and could have caused weak spots. But the scatter seems normal and fine to me for a set of parts that were manufactured over decades and not uniform. Pinch points could be possible while testing causing unequal loads; this is a network, not a mono-fiber. Some were used/deployed before testing, so that could certainly account for the scatter as well.

 Question: Were the parachutes unused?

 Answer: Some were used/damaged (not visually, but assume some could have been) some were not if I remember correctly. So again, the scatter does not bother me, that is -real life- data. Not ideal lab data. Real life is what we care about when making decisions about field performance.

 Unused climbing gear, properly stored, did not show significant degradation over 40 years. The curve in fig. 14 would mirror my results for unused material.

 

"With regard to 10 year lifetime….I’d argue that the data from figure 13 is scatter. I bet that if we did testing from virgin to 10 years, we would see the same scatter. So at first glance, I would not use that to justify a technical argument for 10 years.

My conclusion

 In their work Mr. Berbnstein and Mr. Gillen compared values from a temperature and humidity chamber with reality. The "70 years" are NOT evidence, they are the result of extrapolisation. They exposed Nylon to different temperatures up to 137 deg C and extrapolised the results to long term aging at 21 degrees.

The authors underscore the dangers of long extrapolations of high-temperature aging data.

"It is interesting to note that in the absence of the low-temperature results, an extrapolation using the high-temperature activation energy of 96 kJ/mol would predict a 50% drop in tensile strength at 21 C after 17,000 years."

The only relevant evidence I found in this work are the breaking loads of parachutes which have been partly used and stored packed under different conditions. They were deconstructed and the ribbons, reefing lines, suspension lines, vent bands and skirt bands were tested and compared with new products of the same kind.

Here is the chart # 13 mentioned above:

The scatter reminds me to my own work where I found similar patterns..

In chart 14 they took the mean of the values.

The residual breaking strength is higher than 80% of a new item after 40 years.

By the way, in permanent topropes in a climbing gym I identified a 50% reduction of breaking strength after a couple of days!

Sorry,  this work confirms my thesis that the reduction in breaking load is not dangerous over a period of 40 years (we do not have enough evidence for a longer period).