Title: So, How Did it Fail?
By: Ian Peggs, I-CORP INTERNATIONAL

When a liner fails by sliding down a steep slope or as a result of whales getting entangled in aerator blades, there is often a great pile of mangled liner to remove and replace as quickly as possible. Similarly, failed pipe fusion welds need to be repaired as soon as possible. Often, the old material is discarded altogether. Then there is an investigation to determine what happened, often with insurance companies and attorneys involved, to determine who is at fault. A little forethought can save much wasted time and money later by removing key samples before tearing everything out. A tangled mess of liner does not destroy the evidence of the cause of the failure, but if all the material is removed by bulldozer or excavator, it just makes the liner features of interest that much more difficult to find. And in the worst case, one can lose a case as a result of prematurely disposing of the evidence.

A 20 mil (0.5 mm) HDPE liner for a decorative pond turned out to be a 16 mil slit film HDPE woven with 2 mil of LLDPE on each surface. It leaked like a sieve, literally. To ensure there was sufficient water in the pond for the development's opening (and major sales weekend), the contractor ripped the liner out, disposed of it, and replaced it with the intended homogenous HDPE liner. Despite retaining excess unused material that also leaked like the same sieve, the case brought by the contractor against the liner supplier/installer was dismissed because the installed material was no longer available; the evidence had been destroyed.

With appropriate materials science knowledge, it is quite easy to trace the initiating site(s) of the failure and to recover appropriate samples for detailed examination if required. The fracture edges and fracture faces, like the rings of a tree, tell the complete history of the failure - exactly where it started, at what feature, on which surface or internally, which way it propagated, how it propagated (fast or slow, steadily or in steps), then where it transitioned into the large displacement event. Even if a unique initiation location cannot be identified on site, a few key samples can be selected that together will tell the complete story. With these critical samples preserved, the balance of the unusable material can then be removed and repairs made.

In one case many years ago, the liner on the whole side slope (200 ft by 40 ft) of a surface water run-off pond on a hazardous waste site had shattered by rapid crack propagation (stress cracking), but the initiation site was quickly traced to a fusion seam that had been "repaired" by extrusion welding. A sample of the seam was removed for detailed microscope examination in the laboratory. The precise initiation site was found on the underside of the weld at a crimp made by overheating - clearly a combination of a poor material and poor welding.

Slow crack growth stress cracking started at overheating crimp on underside of extrusion-weld-repaired fusion seam (large arrow), propagated downslope, turned into geomembrane to left of seam, continued upwards where (small arrows) it bifurcated, trifurcated, etc., before "shattering" the liner by rapid crack propagation in a cascading firework pattern along the slope, into the anchor trench, and under the ice on top of the water (where I am standing).

In another case where stress cracking occurred at extrusion and butt welds in a thick HDPE embedment liner in a mine facility, the installer was faulted by the owner and engineer for improper installation. However, the welds were well made and the material was good. In fact, the liner manufacturer suggested that a chemical resistance test be performed but the engineer claimed it was unnecessary. So, a saving of $10,000 resulted in a multi-million dollar failure. Fortunately, in-situ samples were taken before the liners were ripped out.

In this case, the design engineer was ultimately faulted for specifying that HDPE be used to contain a solution with which it was incompatible - the HDPE never stood a chance.

In two sewer pipe liner failure projects, one liner was removed before I was able to get proper samples for the optimum materials science examination. In the other case, I was able to enter the sewer pipe and take samples before the repairs were made. The second case was much more effectively presented.

In a very large evaporation pond containing precipitated salts, damage to the liner was claimed to be slit-type punctures made by large chisels done as an act of vandalism in the final stages of construction. Under the microscope it was apparent that the slits were not punctures but were shear cuts made, in fact, by the precipitate fractured as a result of the chisel impacts during removal of the precipitate. The contractor was absolved of responsibility, the damage becoming an operations concern.

If a lined slope fails and tears the liner, examination of the liner's fracture faces will determine if, perhaps, the liner failed first and caused the soil to start sliding. Thus, if the geomembrane, geogrid, geocomposite, high strength geotextile, or any other geosynthetic fails because it is overloaded, the cause of failure is not the geosynthetic. However, if the geosynthetic fails when it is not overloaded; i.e. at stresses below the yield or break strength of the material, then the geosynthetic may be the cause of the failure, or may contribute to it.

It is more than likely the breaks in the geogrid would identify the location of initial failure of the wall and how it progressed, whatever the cause of the failure...

Break in liner before waste slipped. If edges of break are ductile, the soil-imposed stresses have simply overloaded the liner; i.e. induced stresses have been higher than the yield and break stresses of the liner. If edges of break are brittle, stress cracking has occurred at a stress lower than the yield stress of the liner. The initiation point and therefore the cause of the break could easily be identified. If stress cracking had not occurred, movement of the geomembrane would not have occurred and the waste may not have slipped catastrophically. This is a very unique and instructive photograph. The development of this break was noted and monitored by an in-situ system of leak location electrodes.

Therefore, as soon as the failure has occurred, call in a materials performance expert to identify samples that should be taken and retained before the bulk of the geosynthetic is removed and stored or replaced, or repaired. Even if a few defects require patches, it could be helpful to remove the flaw before the patch is placed rather than putting the patch on the flaw and hiding it.

Broken geonet strands across full width of geocomposite roll.

Initial failure of this reinforced floating cover occurred on the upper surface immediately above the warp reinforcing yarns under the cyclic stresses caused by wind uplift.