Variable-head Tests Research Articles

A note on falling-head and rising-tail permeability tests on saturated soils in centrifuge

The falling-head and rising-tail method for permeability test allows for both inflow and outflow rates to be conveniently measured. This helps to detect any possible swelling or consolidation of specimen during testing and to identify the continuity of flow of water through the specimen. It has been widely adopted to measure the hydraulic conductivity of saturated porous media of extremely low permeability, such as hydraulic barrier soil materials. Centrifuge permeameter test with this method has recently been devised to accelerate water flow and reduce the time needed for testing. The accompanying equation for calculating hydraulic conductivity is, however, found to be simplified because it was formulated by idealizing such a variable-head test as multiple segmental constant-head tests. In this paper, a new generalized equation is established based strictly on the variable head. Moreover, based on curve matching between the rates of flow into and out of the soil specimen, an approach is also proposed to identify the continuity state of flow. The data relevant to this state are extracted for performing the calculation of hydraulic conductivity. The practical application of the proposed equation and its associated approach is illustrated through a reanalysis of recently published experimental data in the literature.

Field Permeability Tests: Importance of Calibration and Synchronous Monitoring for Barometric Pressure Sensors

Pressure transducers (PTs) and an atmospheric pressure transducer (APT) were used to register test data during two types of permeability tests, which were performed in 14 wells monitoring a confined aquifer installed in the lab, and a field rising-head test in clay. The constant-head tests were performed using a peristaltic pump and thus functioned as constant flow rate tests until stabilization of the water level in the well riser pipe. The rising-head tests were started by the sudden removal of a slug of water. This article presents, first, the method used to calibrate the transducers to assess their systematic calibration error (offset) values. Then, it quantifies the influence of synchronized monitoring for the (PT-APT) pair on short- and long-term test data, which had never been done before. The results indicate that the pair calibration cannot be neglected and that the synchronized monitoring is important for all tests, except maybe for a short-duration variable-head test. For most tests, however, the barometric fluctuation with time plays a significant role.

Forming condition of transient saturated zone and its distribution in residual slope under rainfall conditions

Rainfall, as one of the most significant factors triggering the residual soil slope failure, leads to not only the reduction of soil shear strength, but also the increase of soil weight and the decrease of matric suction as well. All these modifications in soil properties have important influence on the slope stability. The water infiltration and redistribution inside the slope are the preconditions of the slope stability under rainfall conditions. Based on the numerical simulation via finite element method, the water infiltration process under rainfall conditions was studied in the present work. The emphases are the formation, distribution and dissipation of transient saturated zone. As for the calculation parameters, the SWCC and the saturated permeability have been determined by pressure plate test and variable head test respectively. The entire process (formation, development, dissipation) of the transient saturated zone was studied in detail. The variations of volumetric water content, matric suction and hydraulic gradient inside the slope, and the eventually raise of groundwater table were characterized and discussed, too. The results show that the major cause of the formation of transient saturated zone is ascribed to the fact that the exudation velocity of rainwater on the wetting front is less than the infiltration velocity of rainfall; as a result, the water content of the soil increases. On the other hand, the formation and extension of transient saturated zone have a close relationship with rainfall intensity and duration. The results can help the geotechnical engineers for the deeper understanding of the failure of residual slope under rainfall condition. It is also suggested that the proper drainage system in the slope may be the cost-effective slope failure mitigation method.

Parasitic Head Losses During Laboratory Permeability Tests

Parasitic head losses incurred outside the tested soil specimen may alter the results of laboratory permeability tests. They can be measured by running a blank test in an empty permeameter, and determining the flow resistance of the testing equipment. They can also be assessed by using the difference between a constant-head test, where the hydraulic heads are measured within the tested specimen, and a variable-head test, where the hydraulic heads are measured outside the specimen and include the parasitic head losses of the testing equipment. When not accounted for, they can lead to permeability values that are underestimated. An electric analog provides an analytical relationship between variable- and constant-head test results. The parameters in the analytical solution can be obtained via blank permeability tests. This article emphasizes the need to measure hydraulic heads inside the soil specimen during laboratory permeability tests.

Using a leaky swimming pool for a huge falling-head permeability test

Despite recent repairs, a heated municipal swimming pool had major leaks. To determine the possible position of the leak or leaks, a full scale falling-head permeability test was performed. All the valves for circulating hot and cold water were closed for monitoring the variation with time of the water level in the pool. The velocity graph, which represents the conservation equation in variable-head tests, helped to determine the positions of leaks. Subsequent excavations found broken pipes and connections, their failure being related to high differential settlement caused by poor compaction of pipe foundations.

Field variable-head test in low-permeability materials: Assessing the effects of trapped gas pocket and cavity expansion

Trapped gas and cavity expansion may affect the data of a variable-head test in a monitoring well. In a low-permeability material, some gas may be trapped in the top of the filter pack. Presence of a gas pocket may be alleged when the initial change in water level in the pipe is less than the theoretical value due to the slug volume. A variable-head test with trapped gas is described by three equations to be solved simultaneously. The gas pocket contraction leads to an underestimation of the hydraulic conductivity, k, especially for small pipes used to test natural or compacted clay. It must be considered, however, that the difference between the theoretical and measured changes in initial water level is due to the sum of trapped gas contraction plus cavity expansion (falling-head) or trapped gas expansion plus cavity contraction (rising head). This difference cannot be used to assess the volume of trapped gas unless an independent mechanical test has established the material deformability. For example, pressuremeter tests can be done in cavities that are prepared with the same care as those used for the installation of piezometers. A simple test that consists of adding or removing known volumes of water in the pipe of the piezometer is described. Field data are provided along with the method to handle them and extract the correct k value. To avoid gas entrapment in the filter pack, a small diameter plastic tube can be installed and used as a gas vent.

Falling-Head Permeability Tests in an Unconfined Sand Aquifer

Abstract This paper examines the reliability of hydraulic conductivity estimates, k, obtained using equations of Hvorslev (1951) or Bouwer and Rice (1976) for falling-head tests in monitoring wells (MWs) having short screens at the bottom of an unconfined sand aquifer. Two issues are examined. First, the equations come from the theory of steady-state flow whereas a variable-head test means transient flow with a changing water table position. This first problem was investigated using a finite element analysis, taking into account the sand capillary retention curve and its saturated-unsaturated permeability. The main result was that the equations can still be used for variable-head tests. The effects of specific storage (elastic deformation of the solid matrix in saturated conditions) and delayed gravity drainage (frequently schematized by a specific yield) can be neglected for these tests. The second issue is how partial clogging or fine particle washing against the screen can influence the k value. This practical problem was investigated by using a surge block to develop MWs and then performing successive permeability tests to assess the effects of development. Before development, the tests provided k values in the range 1 to 12×10-2 cm/s, the average being equal to 2/3 of the large-scale k value obtained using a pumping test under steady-state condition. After development, the tests provided k values that were increased by 50% on average, and thus their mean value became almost equal to the pumping test k value.

Interpreting Variable-head Tests Performed in Open Boreholes or Monitoring Wells with Several Screens

AbstractVariable-head permeability tests can be performed in open boreholes or monitoring wells having either a long screen or several screened zones. In all cases, a test involves several layers composed of fractured rocks or pervious soils. In the hole or the pipe, the static water level represents some mean value of the hydraulic heads within the aquifer layers, a complex pattern of water circulation having been established within the open hole or the screened aquifers. This paper provides first the equation relating the static water level to the hydraulic and geometric parameters of the aquifers, and then the extended theory for variable-head tests under such conditions. According to theory and numerical analyses, such tests should provide straight semi-log head responses and velocity graphs. In real conditions, however, as shown by a series of field tests performed in a monitoring well equipped with three screens, broken line velocity graphs may be obtained. Such piecewise linear lines indicate that the product “hydraulic conductivity × shape factor” varies during the test within one or several aquifer layers. These variations can be related to internal erosion, which happens close to a screen when the gradient and seepage forces change direction during testing. This is confirmed by numerical modeling.

Closure to “Measuring the Hydraulic Conductivity of Soil–Bentonite Backfill” by Jeremy P. Britton, George M. Filz, and Wayne E. Herring

A discussion of an article with the aforementioned title, published in this journal in December 2004 is presented. The author discusses testing details and scale effects in relation to previously published concepts, laboratory test results and field data. Specific topics discussed include: pilot scale walls, scale effects, in situ variable-head tests, pore pressure dissipation, laboratory results and the presence or absence of scale effects. The original authors address the issues raised by the discusser, especially in the areas of the measurement of the flow rate through pilot-scale wall W3, the reduction of piezometer test data, piezocone soundings and the laboratory tests on undisturbed samples.

Evaluating the hydraulic conductivity at three different scales within an unconfined sand aquifer at Lachenaie, Quebec

This paper presents the evaluation of scale effects for the hydraulic conductivity, k, of a sand. Data were obtained using several methods at a site that was densely instrumented with piezometers within an area of about 100 m × 100 m. The 3.1–3.6 m thick sand deposit rests on a thick, nonfissured Champlain Sea clay deposit. At the small scale, local k values were obtained from the grain-size curves of sand samples and also from ground-penetrating radar surveys. At the intermediate scale, k values were obtained from field variable-head tests using piezometers. At the large scale, k values were obtained from pumping-test drawdowns in steady-state conditions. In this sand aquifer, the careful evaluations at three scales gave similar narrow distributions for the hydraulic conductivity, and therefore no scale effect is evidenced.Key words: hydraulic conductivity, sand, scale effect, gradation, monitoring well, pumping.

Variable-Head Field Permeability Tests in Driven Flush-Joint Casings: Physical and Numerical Modeling

Abstract Two types of variable-head permeability tests in driven flush-joint casings, the end-of-casing test and the lateral injection test (Lefranc test), were performed in two sand tanks. Hydraulic conditions in the sand tanks included a constant vertical hydraulic gradient that was null, positive, or negative. The gradients were monitored by lateral piezometers, which confirmed that the large-scale hydraulic conductivity (k) value of the sand layer remained constant throughout the test program. Any interpretation method for these variable-head tests implicitly requires an assumed piezometric level (APL) for analysis. According to a common practice, the elevation of the ponded water surface above the sand deposit was taken as the APL to plot the test results as the logarithm of the difference in total head versus time. After each casing installation, permeability tests were performed for the three gradient conditions. When the vertical gradient was null in the tank, the test data provided straight lines as assumed in theory. When the gradient was positive or negative, the test data provided curved graphs that may be interpreted as giving k values that vary with time. The velocity graph method, however, provided the same constant k value for the three gradient conditions. It also detected the error in the APL and provided in all cases a local piezometric level (PL) equal to that given by lateral piezometer monitoring. Thus, the velocity graph method, a graphical representation of the conservation equation, eliminated the potential for misanalysis of variable-head tests. The experimental k values, as determined by the two types of tests, usually ranged between 0.5 and 2 times the larger-scale values determined independently by the monitored flow rates and gradients in the sand tanks. The variability in experimental k values was partly due to the local variations of void ratio and effective diameter of the sand, and partly due to the influence of casing installation. Numerical simulations of the tests confirmed the validity of the velocity graph method for providing k and the local piezometric level for a test.

Overdamped slug test in monitoring wells: review of interpretation methods with mathematical, physical, and numerical analysis of storativity influence

Several methods are available to interpret slug tests; however, when applied to the same test data, they usually yield very different results. The methods are classified into three categories depending on their assumptions about the solid matrix deformability during the test. This paper deals with overdamped tests for elastic solids that deform instantaneously. It provides a unified interpretation of transmissivity T and storativity S based on the velocity graph for variable-head tests in monitoring wells or cased boreholes. If S has little influence, the velocity graph is a straight line. If S has some influence, the graph should give a smooth curve. However, smooth curves are exceptions in practice, thereby leading to a reexamination of the influence of S during a slug test. Three independent approaches are used. (1) A mathematical review shows that the overdamped solution, as adapted from a heat conduction problem, did not correctly treat storativity terms and the type of problem: it corresponds to a special pulse test, not a slug test. (2) A physical investigation of deformability shows that the influence of S does not exceed 1% of the initial slug for most compressible materials. Thus, it is almost impossible to detect its influence in test results. (3) Numerical analyses confirm that S has a negligible influence: test results provide straight lines, not curves. The numerical analysis of the special pulse test provides exactly the classical solution, and the correct values of T and S after eliminating the confusion about storativity terms. It is concluded that (1) S has a negligible influence in slug tests, (2) the existing classical solution giving T and S must be abandoned, and (3) the velocity-graph equation and its integral equation (Hvorslev or Bouwer and Rice) which correctly describe the process must be used.Key words: slug test, hydraulic conductivity, storativity, numerical modeling.