A grout curtain is not defined by a row of boreholes on a drawing. It is a controlled intervention in a variable rock mass, intended to reduce hydraulic conductivity, manage uplift or inflow, and preserve the behaviour assumed in the wider design. Good grout curtain design steps therefore begin with the required performance, not with a standard hole spacing.
This distinction matters at dams, underground works, shafts and cut-offs where ground conditions can change materially over a few metres. A curtain that looks complete geometrically may still contain connected water-bearing features. Conversely, an overly dense curtain may add cost, induce unnecessary fracturing and provide little additional benefit.
Grout curtain design steps: start with performance
The first design task is to state what the curtain must achieve and how that achievement will be judged. The objective may be limiting reservoir seepage, reducing uplift pressure beneath a structure, controlling groundwater inflow to an excavation, or isolating a permeable zone around a tunnel. These objectives are related, but they do not necessarily lead to the same acceptance criteria or layout.
Performance criteria should be measurable. Depending on the project, they can include a target residual water loss from pressure testing, a permissible piezometric level, a maximum inflow rate, or a specified reduction in hydraulic conductivity. A water-loss criterion alone is rarely sufficient. It describes the response around a test section at a particular pressure and duration, whereas the structure must perform under its own hydraulic boundary conditions over time.
The design basis should also distinguish between the primary curtain and associated measures. A foundation treatment scheme may include contact grouting, consolidation grouting, drainage holes, local treatment of faults and a deeper grout curtain. Treating these as one operation makes it harder to assign responsibilities and interpret results.
Build a ground model that explains flow
A geological model is useful only when it helps predict where water travels. For grout curtain design, this means moving beyond a lithological description towards a hydrogeological and structural interpretation. The relevant questions are whether flow is concentrated in open joints, sheared zones, bedding planes, karstic features, contacts or weathered rock, and whether those features are connected at the scale of the structure.
Core logging, borehole imaging, packer testing, groundwater levels, surface mapping and excavation observations should be brought together. Particular attention is needed at changes in rock type, valley shoulders, abutments, intrusive contacts and known deformation zones. A low average water loss can conceal a narrow but highly conductive feature if the test intervals are too long.
The likely response to grouting also needs consideration. Fresh, tightly jointed crystalline rock may require relatively limited treatment once the principal conductive features are addressed. A weathered or stress-relieved zone can consume large volumes with uncertain continuity. In soluble or highly heterogeneous ground, a conventional curtain may need to be combined with different cut-off measures. The appropriate design is therefore conditional on the ground model, not merely on depth below foundation level.
Define the curtain geometry
Curtain depth, alignment, inclination and number of rows should follow the anticipated seepage paths. A single line of holes may be appropriate where the objective is to intercept sub-vertical fractures beneath a dam axis. Multiple rows or inclined holes can be justified where the dominant structures dip unfavourably, where a fault zone has width, or where the curtain must connect reliably with less permeable rock.
The depth criterion should be stated explicitly. It may be a fixed elevation, a penetration below a weathered zone, an intercept of a geological unit, or a depth derived from hydraulic analysis. Fixed depth is convenient but can be misleading when rock quality varies significantly along the alignment. A deeper local treatment is often more effective than extending the entire curtain to the same conservative depth.
Hole spacing should be viewed as an initial hypothesis. Spacing is governed by fracture persistence, aperture, grout penetrability, borehole deviation and the intended overlap between treated volumes. Primary, secondary and tertiary holes give a practical way to refine the curtain, but they are not a substitute for evidence. If secondary holes reveal persistent high takes or high water losses, the response may be altered orientation, local deepening or targeted fault treatment rather than simply adding another regular split.
Select materials and pressures together
Grout selection and injection pressure must be developed as a pair. Cement grout is effective for many fractured rock masses, but its ability to enter fine apertures depends on particle size, water-to-cement ratio, admixtures, stability and filtration behaviour. Microfine cement or chemical grout may be required in fine fissures, but higher material cost and compatibility with the project environment must be considered.
The pressure regime requires equal care. Insufficient pressure may fail to drive grout beyond the immediate borehole vicinity. Excessive pressure may hydrofracture the rock, lift near-surface ground, connect previously isolated features or produce a misleadingly high take. Pressure limits should reflect depth, in-situ stress, overburden, rock mass condition and proximity to structures or excavations.
A simple rule based only on metres of cover is not enough for every ground condition. The pressure recorded at the collar also differs from the effective pressure at the injection stage because of hydrostatic head and line losses. Stage length, upward or downward sequencing, refusal criteria and the maximum volume allowed per stage should be defined in the method statement and reviewed against field observations.
Some projects use grout intensity number methods to control the combined effect of pressure and volume. Such approaches can support consistent execution, particularly where a substantial dataset is available, but they should not be applied mechanically. A limiting pressure-volume relationship does not establish that a fault has been sealed, nor does a high grout take automatically demonstrate a beneficial treatment.
Prove the design through a trial section
A trial curtain is where the conceptual design becomes an evidence-based construction method. It should be located in ground representative of the principal risk, not merely where access is easiest. The trial can test hole spacing, stage length, grout mixes, pressure limits and the value of tertiary holes before those decisions are committed across the full works.
The trial programme should include independent verification holes. Their position needs to be selected so that they test the untreated spaces and suspected conductive features, rather than repeating the geometry of production holes. Water pressure tests conducted before and after grouting can provide a useful comparison when test pressures, stage lengths and procedures are consistent.
Acceptance should use several observations together: water loss, pressure response, grout takes, piezometer trends, drilling returns and geological records. Each measure has limitations. A low residual water loss may result from local plugging close to the test borehole. A declining grout take can indicate successful filling, but it can also reflect poor communication with the target fractures. The engineering interpretation is as important as the numerical threshold.
Treat execution data as part of the design
Grout curtain construction produces a large quantity of information, often at a much finer scale than the site investigation. Borehole coordinates, measured depth, inclination, stage intervals, water-loss results, grout mix, pressure, flow, volume and refusal condition should be recorded consistently. Without reliable records, it is difficult to identify a connected weak zone or defend the adequacy of completed works.
Visualising these records along the curtain alignment and in section is particularly valuable. Zones of high take or persistent water loss may indicate a structural feature that was not recognised in the original model. Conversely, uniformly low takes in competent rock can support reducing treatment density where the verification evidence permits. Simple, transparent calculation and plotting tools are useful here because they allow the engineer to review the evolving ground model without separating design work from field data.
Borehole deviation deserves explicit control. At depth, small collar positioning errors and deviations can create larger gaps between intended treatment paths than the nominal spacing suggests. Surveying or deviation measurements may be necessary for critical deep curtains, closely spaced rows and inclined drilling. The planned geometry must be compared with the geometry actually achieved.
Verify the finished curtain under relevant conditions
Final verification should address the original performance objective. If the key concern is uplift, piezometric monitoring beneath the structure may be more meaningful than an isolated residual water-loss value. If the concern is tunnel inflow, the observed response during staged excavation and groundwater monitoring may provide the decisive evidence.
Verification holes should remain genuinely independent where possible. Testing only production holes can give an optimistic view because they have already been directly treated. The number and distribution of checks should reflect the consequences of failure and the variability of the geology. A short curtain through uniform competent rock calls for a different level of verification from a deep foundation curtain crossing several faulted zones.
A well-designed curtain also allows for uncertainty. Define hold points for reviewing results, criteria for additional holes, and clear authority for changing the programme. This makes adaptation controlled rather than improvised, while retaining the traceability needed for design records and future operation.
The most useful final deliverable is not a statement that every planned hole was drilled. It is a defensible account of how the curtain was adjusted to the ground encountered, what evidence supports its performance, and where monitoring should continue to confirm that behaviour in service.
