To the editor:
I enjoy reading the GMA Techline in Geosynthetics magazine and the snappy to-the-point answers provided by my friend and colleague, Dr. Koerner.
However, I have some concerns with the recommendation to accept 4-out-of-5 passing results for the peel separation parameter for seam peel testing.
I do not think a 20% failure rate, displayed by four passing peel separation specimens and one that peels 100%, is acceptable, particularly these days when a welder worth his or her salt can regularly get five passing shear strength/elongation specimens and five passing peel strength/separation test results.
Ian D. Peggs
Ocean Ridge, Fla.
[Dr. Peggs], As you know almost all specifications for peel and shear strengths of thermally welded geomembranes in the field are written around 4-of-5 passing the specified strength with the fifth one achieving 80% of the specified strength. This follows U.S. EPA Technical Guidance Documents EPA/530/SW-91/051 and EPA/600/R-93/182. This is also the situation as written in the GRI-GM19 Specification for HDPE, LLDPE and fPP geomembrane seams since it was first adopted in 2002.
While welding advances have certainly been made over time, this criterion persists and, even with it, failed samples are not uncommon. When ASTM subsequently revised its D6392 seam test method some years ago, it introduced the additional measurements of maximum incursion for peel testing and minimum elongation for shear testing. At GSI we added these requirements by updating our GRI-GM19 specification to include a maximum peel separation of 25% and a minimum shear elongation of 50%. We were (and still are) silent on the issue of having some flexibility of these values within the five peel and five shear test values for both observational values. In other words, “Are outliers allowed?”
This question posed in the GMA Techline was the first time the issue of allowing outliers was asked of us. In the context of consistency, I suggested in my answer to the questioner to allow 4-of-5 test specimens to meet the incursion and elongation criteria test specimens, but failed to mention any limitation insofar as the amount of leniency in the fifth specimen as opposed to the required specification values. For example, should we allow the one outlier to be limited to a maximum 30% incursion (i.e., an increase of 20% in this maximum value) and a minimum 40% elongation (i.e., a decrease of 20% in this minimum value)? If so, these two numeric criteria would then be consistent with the 80% strength value mentioned above.
I do think that this is reasonable in the context of being consistent with the strength criteria and we will bring up the issue at the next GSI Geomembrane Focus Group meeting.