Site investigation and soil exploration

In short

Site investigation is the planned collection of information about the soil, rock and groundwater at a site so that foundations, pavements, embankments and excavations can be designed safely and economically. It covers desk study and reconnaissance, boring and sampling, in-situ and laboratory testing, and finally a soil profile with design parameters. In India it is governed mainly by IS 1892 (Code of practice for subsurface investigation for foundations) along with IS 2131, IS 2132, IS 4968, IS 4434, IS 1888 and NBC 2016 Part 6; typical investigation cost is well under 1% of project cost but it decides the success of everything above. It is examined in GATE (CE) almost every year through SPT corrections, sampler area ratio, depth of exploration and plate load test scaling, and it is the first topic in interviews at L&T, AFCONS, Tata Projects, NHAI/NHIDCL consultants, geotechnical firms and metro/railway design cells.

Explained

Purpose, stages and planning of the programme

• Objectives: determine the sequence, thickness and extent of strata; depth and fluctuation of groundwater; strength, compressibility and permeability parameters; chemical aggressiveness (sulphates, chlorides); rock depth and quality; and suitability of borrow material.
• Stage 1 - Reconnaissance / desk study: topo sheets, Geological Survey of India maps, seismic zone from IS 1893 Part 1, existing records, aerial imagery, performance of nearby structures, site walkover (seepage, cracks, fills, old wells).
• Stage 2 - Preliminary exploration: a few widely spaced boreholes, soundings (SPT/CPT/DCPT) and test pits to get the broad profile and plan the detailed stage.
• Stage 3 - Detailed exploration: boreholes at designed spacing with sampling and in-situ tests, laboratory testing, groundwater monitoring.
• Stage 4 - Construction-stage checks: foundation-level inspection, additional probes where the profile differs, pile integrity and load tests, field density checks on fills.
• Number and spacing of boreholes (indicative, as per IS 1892 and NBC): compact building up to about 0.1 ha - one borehole at centre plus 2-4 at corners; multi-storeyed building grid - roughly 10-30 m; large area/industrial layouts - 30-60 m; highways and runways - 150-300 m along alignment, closer at structures; earth dams and embankments - 30-60 m. Spacing is tightened where strata vary and relaxed where they are uniform; a minimum of three boreholes is needed to define a plane/dip of a stratum.
• Depth of exploration: go to the significant depth - the depth at which the net stress increase from the foundation falls to about 10% of the net applied pressure, or to about 10-20% of the effective overburden pressure, whichever is shallower. Minimum guidance: 1.5 B below the base of an isolated footing, 1.5 × least dimension below a raft, 1.5 × group width below the pile tips for a pile group, 1.5-2 times the embankment height, and at least 3 m (or one base width) into sound rock after proving rock by coring.
• Footings closer than about 2 B interact; their pressure bulbs overlap, so the exploration depth must be based on the equivalent larger loaded area, not the single footing.

Methods of boring and exploration

• Open test pits and trenches (IS 4453): cheapest and best for visual examination, block/chunk sampling and in-situ density; practical only above the water table and to about 3-6 m with shoring.
• Hand/post-hole and helical augers: 3-6 m in cohesive and partly saturated soils; simple, cheap, disturbed samples only; collapses in clean sand below water table.
• Power-driven continuous flight auger: fast profiling to 15-20 m; hollow-stem augers allow sampling and SPT through the stem without casing.
• Wash boring: water or drilling fluid jetted through a chopping bit washes cuttings up the annulus; cheap and widely used in India for SPT holes, but cuttings give a poor record of strata - identification must rely on samples and SPT.
• Percussion boring: heavy chisel/bailer repeatedly dropped; the only practical method in boulders, gravel and hard strata, but it badly disturbs the soil below the hole (IS 10042 covers gravel-boulder deposits).
• Rotary drilling and core drilling: a rotating bit with circulating mud (bentonite) or air; gives the best record, is essential in rock with a double/triple-tube core barrel and diamond bit, and is standard for deep investigations and bridge foundations.
• Soundings (no hole, continuous record): static cone penetration (CPT) and dynamic cone penetration (DCPT) - fast, cheap profiling between boreholes.
• Boreholes are usually 150 mm diameter so that 100 mm undisturbed samples can be taken; casing or mud is used to keep the hole open, and the water level inside the casing must be kept at or above the groundwater level to prevent the base from "boiling" and loosening the sand.

Soil samples and samplers

• Disturbed samples: structure destroyed. *Non-representative* (auger cuttings, wash returns) - only for identification; *representative* (split-spoon, bagged) - for index tests, grain size, Atterberg limits, compaction, specific gravity.
• Undisturbed samples: needed for shear strength, consolidation and permeability. True "undisturbed" is an ideal; disturbance is minimised by thin-walled tubes pushed (not driven) continuously, sealed ends, careful transport and quick testing.
• Sampler geometry controls disturbance - area ratio, inside clearance, outside clearance and recovery ratio (formulas below). A thin-walled Shelby tube has an area ratio under 10%; a split-spoon sampler (50.8 mm OD, 35 mm ID) has an area ratio of about 110%, so it can never give an undisturbed sample.
• Thin-walled tube sampling (IS 2132): tube pushed hydraulically at least 10-15 tube diameters, with inside clearance to prevent friction and a short overdrive avoided.
• Piston samplers (stationary piston, Osterberg hydraulic piston) hold a vacuum above the sample and retrieve soft, sensitive clays; Denison and Mazier core barrels sample stiff clay and weathered rock with an inner tube and outer rotating cutter; sand bailers / core-catcher samplers retain cohesionless soil; Sherbrooke/ block samples are the highest quality.
• Soil-sample classes: Class 1 (undisturbed - strength and consolidation), Class 2 (slightly disturbed - density, some strength), Classes 3-5 (progressively only for classification and water content).

In-situ (field) tests

• Standard Penetration Test, SPT (IS 2131): a 50.8 mm OD split-spoon is driven by a 63.5 kg hammer falling freely through 750 mm; the sampler is driven 450 mm in three 150 mm increments. The first 150 mm is the seating drive (discarded, disturbed zone); N = blows for the last 300 mm. Refusal is recorded at 50 blows for any 150 mm increment or 100 blows total. Tests are normally at 1.5 m intervals and at every change of stratum.
  • Corrections: overburden pressure (CN), dilatancy correction for saturated fine sand and silt (Terzaghi-Peck), hammer energy ratio to 60% (CE), rod length (CR), borehole diameter (CB) and sampler type (CS). IS 2131 requires overburden correction first, then dilatancy.
  • Uses: relative density and φ of sands, liquefaction assessment (as (N1)60 in IS 1893), allowable bearing pressure charts, site class. Reliability in clays is poor; in gravel the N value is inflated by particles jamming the shoe.
• Static cone penetration test, SCPT/CPT (IS 4968 Part 3): a 60° cone of 10 cm² base area pushed at about 20 mm/s records cone resistance qc and sleeve friction fs; friction ratio Rf = fs/qc identifies soil type (low in sand, high in clay). CPTu (piezocone) adds pore pressure u2 for dissipation and drainage. Continuous, repeatable, no borehole - but gives no sample and cannot penetrate gravel or rock.
• Dynamic cone penetration test, DCPT (IS 4968 Parts 1 and 2): a 50 mm (or 62.5 mm) cone driven by the SPT hammer; blows per 300 mm (Ncd) are recorded continuously, with or without bentonite slurry to kill rod friction. Very fast and cheap for checking uniformity between boreholes; correlations with N need site calibration.
• In-situ vane shear test (IS 4434): a four-bladed vane (h/d usually 2) rotated at about 0.1°/s gives the undrained shear strength of soft to firm saturated clay (cu up to roughly 100 kPa); repeating after 10 rapid rotations gives the remoulded strength and hence sensitivity. Unreliable in fissured, sandy or stiff clay; results need Bjerrum's correction for plasticity in stability analyses.
• Plate load test (IS 1888): a rigid steel plate (300-750 mm, square or circular, minimum 300 mm) is loaded in increments at foundation level in a test pit at least 5 Bp wide; the load-settlement curve gives ultimate bearing capacity and the modulus of subgrade reaction k. Its influence depth is only about 2 Bp, so it misses deeper weak layers and must not be used alone for large footings or for settlement of clay (consolidation is time-dependent).
• Others: pressuremeter (Menard) for in-situ modulus in bores, flat dilatometer, field permeability tests (IS 5529), field density by sand replacement/core cutter (IS 2720 Parts 28 and 29), and pile load tests (IS 2911 Part 4) at the construction stage.

Geophysical methods, groundwater and rock

• Seismic refraction (IS 15681): travel times of a hammer/explosive-generated P-wave to a line of geophones give layer velocities and depth to rock; needs velocity to increase with depth. MASW/cross-hole and down-hole tests give shear-wave velocity Vs for dynamic design and site classification.
• Electrical resistivity (Wenner/Schlumberger, IS 15736): apparent resistivity distinguishes saturated clay (low) from dry sand, gravel and rock (high); used for groundwater, bedrock depth and soil corrosivity for buried pipes and earthing design.
• Ground penetrating radar (GPR) locates utilities, voids and shallow interfaces; magnetic/gravity methods are occasional. Geophysics is fast, non-destructive and covers large areas, but it is indirect and must always be calibrated against boreholes.
• Groundwater: record the water level on striking and after 24 h stabilisation (IS 6935); install standpipe or Casagrande piezometers for long-term and perched/artesian conditions; note seasonal fluctuation - the design water table is the highest likely level. Collect water samples for sulphate and chloride to choose cement type and cover.
• Rock coring: report core recovery and RQD (Deere); RQD classes - 0-25% very poor, 25-50% poor, 50-75% fair, 75-90% good, 90-100% excellent. Also log weathering grade, joint spacing and orientation, and test unconfined compressive strength and point load index; IS 12070 covers shallow foundations (bearing pressure) on rock and IS 13365 rock-mass classification (RMR).

The bore log, report and Indian codes

• Bore log contents: project and borehole number, coordinates and reduced level, date, method and diameter of boring, casing/mud details, depth-wise strata description with IS 1498 symbols, sample type and depth, N values, water levels, core recovery and RQD, termination depth and reason.
• Report structure (IS 1892): scope and work done; site and geology; borehole location plan; bore logs and soil profile sections; laboratory results; interpreted design parameters (γ, c, φ, cu, mv/Cc, k, Vs); recommendations on foundation type, founding depth, safe bearing capacity and expected settlement; special issues - liquefaction, expansive soil, fill, collapsible soil, sulphate attack; limitations of the investigation.
• Key Indian codes: IS 1892 (subsurface investigation), IS 4453 (pits, trenches, drifts, shafts), IS 2131 (SPT), IS 2132 (thin-walled tube sampling), IS 4968 Parts 1-3 (DCPT and SCPT), IS 4434 (vane shear), IS 1888 (plate load test), IS 5529 (field permeability), IS 6935 (water level in a borehole), IS 15681 and IS 15736 (geophysical exploration by seismic refraction and electrical resistivity), IS 2720 series (laboratory tests), IS 6403 and IS 8009 (bearing capacity and settlement), IS 1893 Part 1 (site classification, liquefaction), NBC 2016 Part 6 Section 2, and IRC:75/IRC:78 with MoRTH specifications for highway and bridge investigations. Code revision years change - confirm the current edition on the BIS portal before quoting one.

Key points

• Site investigation - planned subsurface fact-finding; its cost is a small fraction of project cost but controls safety and economy.
• Stages - reconnaissance, preliminary, detailed, construction-stage verification.
• Significant depth - where stress increase falls to about 10% of net applied pressure; minimum 1.5 B below a footing, 1.5 × least dimension below a raft.
• Spacing - 10-30 m for buildings, 150-300 m along highways; minimum three boreholes to define a stratum's dip.
• Wash boring - cheap but cuttings misrepresent strata; rotary coring gives the best record.
• Percussion boring - the method for boulders, the worst for sample quality.
• Disturbed vs undisturbed - index tests vs strength/consolidation tests.
• Area ratio - <10% for undisturbed sampling; split-spoon ≈ 110%, hence disturbed.
• Inside clearance 0.5-3%, outside clearance 0-2% - reduce side friction and driving force.
• SPT - 63.5 kg hammer, 750 mm free fall, blows for the last 300 mm of a 450 mm drive.
• SPT corrections - overburden, then dilatancy for saturated fine sand/silt; energy, rod length, borehole and sampler factors give N60 and (N1)60.
• CPT - continuous qc and fs, friction ratio identifies soil, but no sample is recovered.
• Vane shear - in-situ cu of soft clay; strength before and after remoulding gives sensitivity.
• Plate load test - influence depth only about 2 Bp; cannot detect deep soft layers or predict consolidation settlement.
• RQD - percentage of core in pieces ≥100 mm; 50-75% is "fair" rock.
• Geophysics - fast and areal but indirect; always tie it to boreholes.
• Groundwater - record level on striking and after 24 h; design for the highest probable level.
• Deliverable - bore logs, soil profile and a report giving design parameters and their limitations.

Formulas and examples

Example 1 - Depth of boring for a footing. A square footing 2.5 m × 2.5 m rests 2.0 m below ground and carries a net pressure of 150 kPa. Find the required borehole depth.

• Significant-depth criterion: Δσz = 0.10 × 150 = 15 kPa. Using 2:1 spread, 150 × 6.25 / (2.5 + z)² = 15 → (2.5 + z)² = 62.5 → 2.5 + z = 7.91 → z = 5.41 m below the base.
• Code minimum 1.5 B = 1.5 × 2.5 = 3.75 m - smaller, so the stress criterion governs.
• Borehole depth from ground level = 2.0 + 5.41 = 7.41 m → adopt 8 m (round up, and continue if a soft layer is still being penetrated).

Example 2 - Sampler suitability. A thin-walled tube has cutting-edge ID 72 mm and OD 74 mm; the tube ID is 73 mm and tube OD 73.5 mm. A 450 mm push recovers 430 mm of soil.

• Ar = (74² − 72²)/72² × 100 = (5476 − 5184)/5184 × 100 = 292/5184 × 100 = 5.6% (< 10%, acceptable for undisturbed sampling).
• Ci = (73 − 72)/72 × 100 = 1.39% (within 0.5-3%). Co = (74 − 73.5)/73.5 × 100 = 0.68% (within 0-2%).
• Rr = 430/450 = 0.96 - slight compression, sample acceptable.
• For comparison, a split-spoon with D1 = 35 mm, D2 = 50.8 mm: Ar = (2580.6 − 1225)/1225 × 100 = 110.7% → disturbed sample only.

Example 3 - SPT correction and φ. At 9.0 m depth in fine saturated sand, field N = 25. γ = 18 kN/m³ throughout, water table at 3.0 m. Hammer energy ratio 60%, 65 mm borehole, standard sampler, rod length 9 m (CR ≈ 0.95).

• σ'v = 3 × 18 + 6 × (18 − 9.81) = 54 + 49.14 = 103.1 kPa.
• N60 = 25 × (60/60) × 1.0 × 1.0 × 0.95 = 23.8 ≈ 24.
• CN = 0.77 log10(1915/103.1) = 0.77 × log10(18.57) = 0.77 × 1.269 = 0.977 → (N1)60 = 0.977 × 23.8 = 23.3 ≈ 23.
• Dilatancy (fine saturated sand, N > 15): N' = 15 + 0.5(23 − 15) = 19 - this is the value carried into design.
• φ ≈ 27.1 + 0.30 × 19 − 0.00054 × 19² = 27.1 + 5.7 − 0.19 = 32.6°; N' = 19 → medium dense sand.

Example 4 - Vane shear and sensitivity. Vane d = 60 mm, h = 120 mm. Peak torque 30 N·m; after 10 rapid rotations, 10 N·m.

• d²h/2 = (0.06)² × 0.12/2 = 0.0036 × 0.06 = 2.16 × 10⁻⁴ m³; d³/6 = 2.16 × 10⁻⁴/6 = 0.36 × 10⁻⁴ m³; sum = 2.52 × 10⁻⁴ m³.
• π × 2.52 × 10⁻⁴ = 7.917 × 10⁻⁴ m³.
• cu = 30 / (7.917 × 10⁻⁴) = 37,890 Pa ≈ 37.9 kPa (firm clay); remoulded cu = 10 / (7.917 × 10⁻⁴) = 12.6 kPa.
• Sensitivity St = 37.9/12.6 = 3.0 → medium sensitive clay; strength is largely lost if the clay is disturbed during piling or excavation.

Example 5 - Plate load test scaling. A 600 mm square plate on sand settles 12 mm at 200 kPa; the plate failed at 450 kPa. A 2.0 m square footing is proposed; permissible settlement 25 mm.

• Settlement: Sf = 12 × [2.0 × (0.6 + 0.3) / {0.6 × (2.0 + 0.3)}]² = 12 × [1.80/1.38]² = 12 × (1.3043)² = 12 × 1.701 = 20.4 mm < 25 mm → safe at 200 kPa.
• Ultimate capacity in sand scales with width: qu,f = 450 × (2.0/0.6) = 1,500 kPa; with FS = 2.5, qsafe ≈ 600 kPa, so settlement, not shear failure, governs the design pressure of 200 kPa.
• Caution: the plate stresses only about 1.2 m of soil, while the footing stresses about 4 m - a soft layer at 2-3 m would make this answer unsafe, which is why the borehole must reach the significant depth.

Example 6 - RQD and core recovery. A 1.5 m NX core run recovers 1.25 m of core; the intact pieces longer than 100 mm measure 120, 200, 150, 110 and 250 mm.

• Core recovery = 1.25/1.50 × 100 = 83.3%.
• Σ pieces ≥ 100 mm = 120 + 200 + 150 + 110 + 250 = 830 mm → RQD = 830/1500 × 100 = 55.3% → "fair" rock, jointed; design must account for joints rather than intact rock strength.

Where it is used

• Building and infrastructure consultancy: geotechnical investigation reports for high-rise foundations, metro stations and depots (DMRC, MMRDA, BMRCL projects), industrial plants and warehouses - deciding raft versus pile and the founding level.
• Highways and bridges: NHAI/MoRTH-specified boreholes along alignment and at every pier and abutment, with SPT/CPT, water sampling, CBR for pavement design, and borrow-area exploration.
• Specialist investigation and drilling firms and equipment suppliers: rotary rigs, CPT trucks, piezometers and instrumentation; roles such as site/field engineer, geotechnical engineer, drilling supervisor, soil laboratory in-charge.
• Dams, canals, ports and power projects: CWC/NHPC-type investigations with drifts, shafts, permeability (Lugeon) tests, grout curtains; jetty and breakwater borings over water from barges.
• Seismic and liquefaction studies: (N1)60 and Vs profiles for IS 1893 site classification and liquefaction screening in Indo-Gangetic plain, Gujarat and North-East sites.
• Forensic and remedial work: investigating distress in existing buildings, settlement of fills, failed slopes and sinkholes; mining, metro tunnelling and oil-and-gas onshore pads also rely on the same toolkit.

Interview questions

• Q: How deep should a borehole go for a 20 m × 30 m raft? A: To the significant depth - about 1.5 times the least dimension (≈ 30 m) below the raft base, or where the stress increase drops to roughly 10% of the net pressure, and deeper if soft strata or rock proving requires it.
• Q: Why is the first 150 mm of SPT penetration ignored? A: It is the seating drive through soil disturbed and loosened by boring; only the next 300 mm represents the in-situ stratum.
• Q: What is area ratio and what value is acceptable for an undisturbed sample? A: Ar = (D2² − D1²)/D1² × 100, the ratio of soil displaced by the sampler wall to the sample volume; it should be below about 10% (ideally under 10% for soft clay).
• Q: Why can a split-spoon sample never be used for a consolidation test? A: Its area ratio is about 110% and it is driven dynamically, so the structure is destroyed - it is only a representative disturbed sample.
• Q: Which corrections are applied to a field N value and in what order? A: Energy, rod length, borehole diameter and sampler corrections to get N60; then overburden correction to get (N1)60; then the dilatancy correction for saturated fine sand and silt (IS 2131 applies overburden before dilatancy).
• Q: SPT or CPT - which would you choose for a soft clay site and why? A: CPT (preferably piezocone), because it gives a continuous, operator-independent profile and reliable cu correlations, whereas SPT is insensitive in soft clay; a few boreholes are still needed for samples.
• Q: Why is the water level inside the casing kept above the groundwater table? A: Otherwise the upward seepage gradient loosens or "boils" the sand at the hole base, giving falsely low N values and poor samples.
• Q: What are the limitations of a plate load test? A: Its influence depth is only about twice the plate width, so deeper weak layers are missed; it is a short-duration test and therefore cannot predict consolidation settlement of clay; scaling to a real footing is empirical.
• Q: What does sensitivity from a vane shear test tell you about construction? A: It shows how much undrained strength the clay loses on remoulding; highly sensitive clays can lose most of their strength from pile driving, vibration or trafficking, so disturbance must be minimised.
• Q: How do you report rock quality from cores? A: As core recovery and RQD (percentage of core in pieces of 100 mm or more), supported by weathering grade, joint spacing and orientation, and UCS or point load strength.
• Q: Can geophysical surveys replace boreholes? A: No - they are indirect, give layer velocities or resistivities rather than strength, and must be calibrated against boreholes; they are excellent for interpolating between them over a large area.
• Q: Which Indian codes govern SPT, thin-walled sampling and plate load tests? A: IS 2131, IS 2132 and IS 1888 respectively, within the overall framework of IS 1892 and NBC 2016 Part 6 - editions are revised, so the current BIS version should be checked.
• Q: What would you add to the investigation if the site has black cotton soil? A: Swelling tests - free swell index, swelling pressure and the depth of the active zone - plus water-content profiles in both seasons, because design is governed by heave, not bearing capacity.

Common mistakes

• Taking exploration depth as a fixed figure (say 10 m) instead of computing the significant depth and checking the 1.5 B / 1.5 × raft dimension minima, and forgetting that nearby footings interact.
• Using N for the full 450 mm of SPT drive, or forgetting to apply the overburden and dilatancy corrections - or applying the dilatancy correction to clay or coarse gravel, where it does not belong.
• Writing the area ratio with the outside diameter in the denominator; the denominator is D1², the cutting-edge inside diameter squared.
• Treating split-spoon or auger samples as undisturbed, or testing shear strength on them and reporting the result as a design parameter.
• Extrapolating a plate load test on a 300 mm plate directly to a large footing without the width scaling, and ignoring the limited depth of influence.
• Mixing units in SPT and vane formulas - σ'v in kPa against t/ft² in Peck's expression, or torque in kg·m against N·m - and quoting cu in Pa instead of kPa.
• Logging strata from wash-boring cuttings, failing to record the 24-hour stabilised water level, or omitting the report's limitations and the number/locations of boreholes - all of which invite disputes later.

Related topics

• Soil properties and classification · Compaction, consolidation and shear strength · Foundations: shallow and deep · Pile foundation design · Earth pressure and slope stability · Ground improvement techniques