Single Well Pumping Test
The essentials required for a single well pumping test (Figure 12.17) are:
- (i) A central water abstraction point, usually a single deep well, which forms the ‘test well’. Ideally, this should penetrate the full thickness of the aquifer, as this simplifies analysis of test results.
- (ii) A series of monitoring wells (standpipes, standpipe piezometers or VWPs (vibrating wire piezometers) depending on the aquifer type) installed in the aquifer at various radial distances from the test well. These allow the depth to groundwater level to be measured (manually by dipmeter or using datalogging equipment) to determine the drawdown at varying times after commencement of pumping and during recovery after cessation of pumping.
- (iii) A means to determine the rate of pumping, such as a weir tank or flowmeter (see Section 22.7).
There is no single template for the design and execution of a pumping test. However, a common plan for single-well pumping tests includes the following phases:
- 1. Pre-pumping monitoring
- 2. Equipment test
- 3. Step-drawdown test (sometimes omitted)
- 4. Constant rate pumping phase
- 5. Recovery phase
The term ‘constant rate’ is a misnomer. In many cases, during a pumping test, the water level in the test well will fall slowly during the later period of the test. If the pump is not adjusted, the pumped flow rate will tend to reduce slightly as the increased drawdown in the
Figure 12.17 Well pumping tests, (a) Components of a pumping test. For clarity, only two monitoring wells are shown, but several are normally required, (b) Well head arrangement for test well. Test equipment visible includes well head with sampling point and control valve, with the discharge pipe leading to a flowmeter. A dipmeter is being used to record water levels in the well. (Panel (b) courtesy of Stuart Wells Limited.)
well changes the discharge head at the pump. In some hydrogeological settings, for example where the aquifer is bounded and large drawdowns are generated, the pumped flow rate may reduce significantly. Where the test is for the design of groundwater control systems, it is normal practice to allow the flow rate to reduce and not to attempt to adjust the pump to keep the flow rate constant.
To obtain the most relevant data for groundwater control schemes, the ‘constant rate’ phases of pumping test are typically carried out at the maximum achievable flow. The test well is commonly ‘over pumped’ with drawdown in the well lowered to close above the pump intake. As a result, there may be a significant seepage face, so the water level in the pumped well often does not provide a useful guide to conditions in the aquifer outside the well.
In addition to monitoring of groundwater levels and the discharge flow rate, a test is also a useful opportunity to obtain groundwater samples from chemical testing. Further details on the execution of pumping tests is given in Appendix 4. Test methods are defined in BS ISO 14686: 2003 and BS EN ISO 22282-4:2012. The drillers’ borehole log of the test well itself may provide additional information about ground conditions. Geophysical logging of the well may also provide useful information (see Sections 188.8.131.52 and 12.8.7). Such methods can be especially useful for wells installed into rock aquifers to allow identification of fractured zones where groundwater is entering the well.
A pumping test is a relatively expensive way of determining permeability. The cost of a pumping test is seldom justified for a small project or for routine shallow excavations. However, for any large project or deep excavation, or where groundwater lowering is likely to have a major impact on the construction cost and/or schedule, one or more pumping tests should be carried out. It is essential that a pumping test is planned to provide suitable data for dewatering design. Issues to be considered include:
- • The drawdown that will be observed in nearby monitoring wells (which at the end of constant rate phase should be at least 10 per cent of that in the proposed dewatering system). The drawdown should also be significantly greater than any natural variations in groundwater level due to tidal or other effects.
- • The duration of the constant rate phase. For pumping tests used for the design of groundwater lowering systems, typical durations of pumping are between 2 and 7 days. Longer tests (with durations up to 60 days) are occasionally carried out, especially where the environmental impact of drawdown on distant groundwater-depen- dent features (see Chapter 21) is a concern.
- • The number and position of monitoring wells (which should allow the drawdown pattern around the test well to be fully identified).
- • The design of the test well (which, ideally, should be of similar design to, and be installed by the same methods as, the proposed dewatering wells).
- • The requirement (or benefits) of additional monitoring over and above monitoring of pumped flow rates and groundwater levels in monitoring wells. Additional monitoring may include more detailed water quality monitoring or observations of groundwater levels in more distant groundwater-dependent features (e.g. water supply wells, wetlands, springs or streams).
To meet these aims, a pumping test is generally carried out in the second or subsequent phases of ground investigation, when ground conditions have been determined to some degree. This allows the depth of the test well screen and monitoring well response zones to be selected on the basis of the data already gathered.
A fundamental requirement for a pumping test is the need to dispose of the pumped water. For most pumping tests, the volume of water generated is too large to store on site. The water is typically discharged to surface watercourses or to sewers (pumping and reinjection tests are discussed in Section 184.108.40.206). When a pumping test is planned, checks must be made that the disposal route has adequate capacity, and the necessary permissions must be obtained from the regulatory bodies (such as Environment Agency in England) or from the utility companies responsible for the sewer network. Regulatory permissions associated with pumping of groundwater are discussed in Section 25.4.
Analysis of the test results can provide information that is useful in a number of ways:
- a) Data from the step-drawdown test can be used to analyse the hydraulic performance of the well in order to determine well losses and efficiency. This approach is widely used in the testing of water supply wells (see Clark, 1977, for methods of analysis), but is less widely relevant to temporary works wells for groundwater lowering purposes. It can still be useful when trying to optimize the performance of deep well systems and may allow comparison of the performance of wells drilled and developed by different methods.
- b) Data from the constant rate pumping phase can be used to estimate the permeability and storage coefficient of the volume of aquifer influenced by the test. The permeability values estimated from the test results are used as an input parameter in the design methods described in Chapter 13. Permeability is estimated by conventional hydrogeological analyses (described later), which can also provide some information about the aquifer boundary conditions.
- c) Observations of the way drawdown reduces with distance from the test well can be used to construct a distance-drawdown plot, which can then be used to design ground- water lowering systems by the cumulative drawdown method (see Section 13.8.2). This method is interesting because the design does not need a permeability value, since that is implicit in the distance-drawdown plot.
Most pumping tests are analysed by ‘non-steady-state’ techniques, which are relatively flexible methods and can be applied to data even as the test is continuing. This allows data to be analysed almost in ‘real time’. ‘Steady-state’ methods of analysis can be used but may require much longer periods of pumping than are necessary with non-steady-state methods. In general, analysis by non-steady-state techniques is to be preferred.
There are a wide variety of methods of analysis that can be used to analyse the results of the constant rate pumping phase, many of which are usefully summarized in Kruseman and De Ridder (1990). Each of the methods is based on a particular set of assumptions about the aquifer system (unconfined, confined or leaky), the well (fully penetrating or partially penetrating) and the discharge flow rate (generally assumed to be constant). Methods suitable for analysis of data from the recovery phase are also available.
These methods should be viewed as a ‘tool kit’ providing a range of possible analysis methods. Provided that the basic details of the aquifer and well are known, it is normally straightforward to select one (or more than one, if uncertainty exists over aquifer conditions) method appropriate to the case in hand. Commonly used methods of analysis fall into two main types:
(i) Curve fitting methods: These typically involve displaying on a log-log graph, for each monitoring well, drawdown against elapsed time. The data will generally form a characteristic shape. The data curve is then overlain with a theoretical ‘type curve’ and the relative positions of the two curves adjusted until the best match of the shape of the two curves is obtained. Once a match is achieved, the permeability and storage coefficient can be determined by comparing values from each curve. These methods were developed from the work of Theis on simple confined aquifers, but variations are available for various other cases (see Kruseman and De Ridder, 1990). The curve fitting process can be done manually but can be rather tedious. However, in recent years, commercial software packages are being increasingly used for pumping test analysis, which speeds up the process.
(ii) Straight line methods: This approach involves plotting sets of data so that characteristic straight lines are produced, allowing the permeability and storage coefficient to be determined from the slope and position of the line. These methods are a special case of the Theis solution based on the work of Cooper and Jacob (1946) and are often called the Cooper-Jacob methods. Two approaches are possible and can be used on the same drawdown dataset from a test:
a. Time-drawdown diagrams involve plotting the drawdown data from one monitoring well against elapsed time since pumping began (Figure 12.18); this process is repeated for all monitoring wells being analysed.
b. Distance-drawdown diagrams plot the drawdown recorded (at a specific elapsed time) in all monitoring wells against the distance of each monitoring well from the test well (Figure 12.19).
The Cooper-Jacob straight line method is a widely used method of analysis on groundwater control projects, mainly due to its relative simplicity. The original Cooper-Jacob method was based on horizontal flow to fully penetrating wells in confined aquifers, but it can also be used in unconfined aquifers where the drawdown is a small proportion (less than 20 per cent) of the original aquifer saturated thickness.
For the time-drawdown data from a single monitoring well, aquifer permeability k and storage coefficient S are determined as follows. From the semi-log graph (Figure 12.18), draw a straight line through the main portion of the data (the data will then deviate from the straight line at early times and possibly at later times). From the graph, obtain the slope of the straight line, expressed as As, which is the change in drawdown s per log cycle of time. Also determine tQ, the time at which the straight line intercepts the zero drawdown line - k and S are then obtained from
Figure 12.18 Analysis of pumping test data: Cooper-Jacob method for time-drawdown data.
Figure 12.19 Analysis of pumping test data: Cooper-Jacob method for distance-drawdown data.
q is the constant flow rate from the test well
D is the aquifer thickness
r is the distance from the centre of the test well to the monitoring well This process can be repeated for each monitoring well.
For the distance-drawdown approach, data from all monitoring wells, at elapsed time t after pumping started, are plotted on the semi-log graph (Figure 12.19). A straight line is drawn through the monitoring well data. If the test well has a very much larger drawdown than the monitoring wells, this may be the result of well losses. In such cases, the straight line should be based on the monitoring wells only and should not include the test well drawdown. From the graph, obtain the slope of the straight line, expressed as As, which is the change in drawdown s per log cycle of distance. Also determine R,„ the distance at which the straight line intercepts the zero drawdown line - k and S are then obtained from
The analysis can be repeated for various elapsed times. This approach also allows the distance of influence R„ to be estimated, which can be a useful check on values used in later dewatering designs.
Care must be taken to ensure that consistent units are used in these equations. To obtain permeability k in metres per second (its usual form in dewatering calculations), the following conventions are used: As, D, R(> and r must be in metres; q must be in cubic metres per second (not litres per second - this is a common cause of numerical errors); and t and t„ must be in seconds (not in minutes, which is often the most convenient way to plot drawdown-time data).
Because the Cooper-Jacob method is only a special case of the more generic Theis solution, a check must be made to ensure that the method is valid. The Cooper-Jacob method can be used without significant error provided that (r S)/(4kDt) < 0.05. This means that the approach is valid provided that t is sufficiently large and r is sufficiently small.
Comprehensive analysis of a pumping test may produce several values of permeability. Time-drawdown analysis will produce one result for each monitoring well, and distance- drawdown analysis can produce several values depending on how many times are analysed. These permeability values are all likely to be slightly different either due to uncertainties in analysis or in response to changes in aquifer conditions across the zone affected by the test. This highlights that pumping test results may still need detailed interpretation; in complex cases, it is prudent to obtain specialist advice.
Variations on the Cooper-Jacob method for certain other aquifer conditions are given in Kruseman and De Ridder (1990).
Alternative Forms of Pumping Test
Although the single well pumping test is the most commonly applied test, other types can be relevant in some circumstances. Different test types are summarized in Table 12.2, which also includes the phases commonly included in conventional single-well tests.
Where artificial recharge of groundwater (see Section 17.13) is planned, special pumping test programmes involving re-injection of water may be required. Typically, the test is started with a conventional pumping phase, whereby water is pumped from well(s) and discharged to surface water or a sewer. In later stages, the water from the pumped well(s) is diverted to recharge well(s), usually located at a significant distance from the pumping wells. This allows data to be gathered on both the capacity of the recharge wells and the interaction between the pumping and recharge wells. A test using two pumping wells and three recharge wells is described by Roberts and Holmes (2011).