Percolation tests on rock samples are more difficult than corresponding tests on soil, mainly because of the slower rate of percolation. It is difficult to measure rates lower than 0-01 cm3/h. Though it is desirable to use higher percolation rates, alteration of the rock samples may occur, and they should be watched closely. Gradients up to 1 in 1000 are used in the laboratory to obtain the desired rate of percolation, whereas gradients of only 1 in 10 are usual in nature.
Terzaghi (19626) distinguishes between 'primary percolation' dependent on microfissures and some of the microfractures, and 'secondary percolation'
Permeability tests 61 dependent mainly on microfractures and macrofractures. The latter, although frequently greater than 'primary percolation', cannot be observed under laboratory conditions and can only be estimated accurately from tests in situ.
In laboratory examination of primary percolation it is important that the samples be as large as possible. The Paris laboratory of the ficole Poly- technique started testing samples of 36 mm diameter and 72 mm height on a standard soil test rig but found it necessary to develop a technique better adapted to rock samples. Samples of 60 mm diameter and 150 mm length, using gradients as high as 1 in 1000 were used for the more impervious rock types.
Water percolation is measured as a function of rising pressure gradients and time. Rock samples may, however, deteriorate with the increasing pressure gradient or with time, and it is essential to keep both effects separated. This is determined by observing the percolation rate on a time curve, which shows some horizontal sections of the curve, corresponding to constant percolation flow.
Air enclosed in the pores of the rock samples must be eliminated before starting the tests. It can be done conveniently by saturating the sample with water under vacuum and allowing the water to percolate through it until the effluent water collected from the sample is absolutely free of air bubbles.
Complete saturation of dense rock samples may take as long as one week.
Water used for percolation tests should be free of dissolved gases. For high- pressure applications it is suggested that the percolation water should be protected with a layer of oil.
The formula to be used for calculating the permeability factor K is:
K = QL/pA,
where Q is the discharge of water percolating through the sample, L the length of the sample, A the cross-sectional area of the sample and p the pressure differential between the two faces of the sample.
Some research laboratories use apparatus (fig. 4.19) derived from conven- tional soil-testing equipment, where the rock sample is encapsulated in an epoxy resin, with the object of preventing leakage along the external cylin- drical face of the sample. Under practical conditions this is not always possible, and in addition it is difficult to remove the epoxy coating when the sample is required for further tests.
The Paris Laboratory developed an apparatus (fig. 4.20) where the sample is protected with a plastic coating and plunged into pressurized water. The radial component of the water pressure is always superior to the pressure in the sample itself and no water can seep along the cylindrical face.
The upstream and the downstream faces of the cylinder are normal to the axis, and the distance between them is equal to L.
It is possible to classify the rocks in two main categories according to the percolation results in a direction parallel to the axis of the sample. When the
seeping water water under
I pressure
rock sample
Fig. 4.19 Longitudinal percolation test.
plastic coating
de-compressed water rock sample
water under pressure
water under pressure Fig. 4.20 Longitudinal percolation test
(Paris Laboratory).
percolation factor is independent of the pressure gradient, it can be assumed that the voids inside the sample are more or less spherical or ellipsoidal. When the rock is microfissured, with fissures much longer in one direction than in the other, the factor K will decrease with the pressure (fig. 4.21).
(a) (b)
Fig. 4.21 Rock samples: (a) with spherical voids; (b) fissured.
Among the rocks tested by Habib & Vouille (1966) were: limestone with 15 to 25% voids (K=7-5 x 10"8), and hard sandstone with 15 to 21%
voids (K = 2-4 x 10"8). In both cases the voids were more or less spherical, and the K value remained constant with varying pressures.
With a microfractured quartz (parallel fractures, 0-1 mm wide) the K value varied from 1*3 X 10"8 to 1*5 X 10"9, when the pressure rose from 0 to 45 kg/cm2. The microfractures were then parallel to the axis of the cylindrical sample. When they were at right angles to the axis, percolation through the very compact rock material was practically nil.
Similarly, a stratified hard schist showed a K value that varied from l*2x 10" 7 to 1-9 X 10" 8, as the pressure rose from 0 to 45 kg/cm2. This schist was formed from layers of quartz grains alternating with layers of mica. The thick- ness of the layers was approximately 0-5 mm with many of the micro- and macrofractures parallel to the direction of foliation. Porosity was not more than 5%.
It is probable that increasing pressures tend to close the thin fissures and fractures, and to reduce the free passage to percolating water.