We read with interest the recent correspondence on accelerated permeability testing (GE July). From 1993 to 1996 one of us, Robert Silver, now of Babtie Geotechnical, carried out intensive research into the compaction and permeability performance of minerals for landfill liners, in collaboration with the University of Paisley and Shanks & McEwan.
It included laboratory and field-based investigations, and led to recommendations on the suitability and use of compaction and permeability test methods. One of the objectives was to determine the effect of different methods of sample preparation and testing on observed values of permeability. A full technical paper is in preparation.
Triaxial cells have been used for many years to determine permeability for landfill liners. The strengths and weakness of the method have been investigated by many different researchers worldwide, mostly in the USA. There are, of course, at least as many different opinions about the best procedure as there are methods and variants.
The principal problem with all permeability testing is the structural changes caused in the material. Extrapolation of results from the laboratory to the field is fraught with difficulties. If laboratory-based permeability values are to be used as a basis for landfill liner design, the effects of the test method on those values must be considered. This is particularly true of the relationship between sample stress and observed permeability. Permeability tests should always be scheduled on the basis of the final use to which the data will be put.
One of the soils used in the Paisley research was a glacial till from central Scotland. The permeability range observed varied from 8 x 10-8 to 1 x 10-10 m/s solely as a result of increasing the effective stress from 25kPa to 200kPa. It was also found that increasing the stress above the preconsolidation pressure for any cohesive materials leads to a decrease in permeability. A subsequent reduction in stress on the same sample will cause the permeability to rise slightly again. It does not, however, return to its initial value because elevated stresses cause irreversible structural change within the sample. The pressure differential used to control the gradient through a sample also produces effective stresses and this, in turn, leads to changes in permeability. There is no significant change, however, if the induced stresses are less than the preconsolidation stress level.
Such effects are significant because permeability tests are accelerated by adjusting the stress state and hydraulic gradient through the sample. During the research, samples were tested for periods from just a few days to up to six months. A protocol was developed which enabled determination of the stress state of the sample and allowed the permeability to be assessed fully. The work on each sample took about six weeks. The minimum was three weeks, when permeability had only to be determined at low stress. Thus, while the protocol was suitable for research it would be impractical in a commercial context. This is primarily because long test times would delay the site CQA process in addition to tying up laboratory equipment. CQA, in particular, depends on fast turnround of results.
No permeability test should be carried out unless it has been designed for purpose. For CQA purposes it may be sufficient to define the full stress/permeability relationship only for a sub-set of representative samples. The majority of the samples could then be tested under relatively high stresses and gradients, to allow a fast turn-round of results, and related back to the low stress sub-set. It is commonly accepted nowadays that permeability cannot be relied upon as the sole means of validation for a clay liner. It may, therefore, be perfectly appropriate to use accelerated permeability tests as part of the overall CQA strategy when engineering a landfill liner.
A huge range of options is open to laboratories in preparing samples for permeability tests. There is urgent need for guidance and standardised approaches for determining the permeability of soils used for landfill liners. No single approach can be appropriate in every circumstance. The guidance, therefore, should include an explanation of the conditions for which a particular procedure is relevant as well as the protocols for sample preparation, etc. It must be remembered that permeability is not a definitive property of soil. It is an empirical coefficient determined from a series of parametric observations which will themselves vary depending on the laboratory equipment and test procedures used.
Dr RK Silver
Babtie Geotechnical Glasgow
Professor JB Joseph
visiting professor of environmental engineering
Department of Civil Engineering
University of Paisley