Geotechnical investigation

What an investigation is for

An investigation answers three questions: what the ground is made of, how strong it is, and where the water sits. Everything downstream depends on those answers. Foundation type, floor slab design, how much fill a site needs, whether a slope will hold, whether the soil will liquefy in an earthquake.

The methods below are not alternatives to one another so much as different resolutions. A cone gives you a continuous profile at one point. A borehole gives you samples you can hold and test. A test pit shows you the ground in section, at the scale a digger works at. Choosing well is mostly about matching the method to the question and the site.

Cone penetration testing (CPT)

A CPT pushes an instrumented cone into the ground at a steady rate and records tip resistance, sleeve friction and pore pressure continuously as it goes. The output is a profile every few centimetres rather than a reading every half metre, which is why it is the method of choice for soft and layered ground, and for liquefaction assessment in particular.

LDE runs its own CPT rig under SOP009, with its own pre-start, inspection and maintenance regime. Having the rig in-house is the difference between booking a subcontractor around their programme and getting on site in the window the project actually has.

• Continuous profile, so thin soft layers are not missed between samples
• No spoil and a small footprint, which suits occupied and urban sites
• Feeds liquefaction and settlement assessment directly
• Seismic CPT adds shear wave velocity for stiffness and seismic site class work

A cone cannot tell you what the soil is, only how it behaves. Gravel, cobbles and shallow rock will stop it. So a CPT programme is usually paired with something that brings material up.

Boreholes and drilling

Drilling recovers material. That gives you soil and rock you can describe, classify and send for testing, and it gets you through ground a cone will not enter. It is also the way to install a piezometer and get a real groundwater record rather than a single reading on the day.

An engineering geologist logs the hole as it is drilled, which is the part that matters: the log is the record everything else is built on, and it cannot be reconstructed afterwards from a sample bag. LDE works to its own machine borehole logging and core logging supervision procedures for exactly that reason.

Drilling costs more per metre than a cone and needs access for a rig. On a tight or sensitive site, a survey first is often cheaper: see geophysics for the non-intrusive options and where they earn their place.

Test pits and hand augers

A machine-excavated test pit is the cheapest way to see the ground in section. Fill depth, buried topsoil, an old rubbish layer, the base of a slip: all of it is obvious in a pit face and ambiguous in a borehole log. Pits are limited to roughly digger reach and they make a mess, so they suit early-stage work, shallow foundations and earthworks.

A hand auger reaches a few metres with no plant at all. On steep ground, inside a building, or behind a house with no access, it is often the only option. LDE covers both under SOP002 and SOP010.

Sampling and laboratory testing

Field description tells you what the soil looks like. Laboratory testing tells you what it will do. Atterberg limits and particle size distribution classify it, compaction and shear strength testing feed design, and consolidation testing is the only way to put a number on long-term settlement under fill or a footing.

Samples are sent to an accredited testing laboratory. The scheduling matters more than the volume: a handful of well-chosen tests on the layer that governs the design is worth more than a broad suite on everything recovered.

Putting a programme together

A workable programme starts from the decision it has to support, not from a number of holes. A three-lot subdivision on flat sand and a retaining wall below a house on a slope need different work, and neither is served by a standard grid.

• What is being built, and what would go wrong if the ground is worse than assumed
• What is already known: neighbouring reports, council records, published geology
• Access, services, and what the site can tolerate being disturbed
• Whether the answer needs strength, stiffness, groundwater, or all three

The output is a geotechnical report, and the full range of work it feeds into is on the geotechnical engineering page.