Soil mechanics and soil classification

Every decision about protecting an excavation starts with the soil. The slope you may cut, the shoring you may use and the depth a trench box is rated for all depend on the soil type. Classifying soil is therefore one of the competent person's most important jobs, and one of the easiest to get wrong. This lesson explains how soil behaves, how the regulations classify it, and the field tests used to decide which class applies.

What soil is made of

Soil is a mixture of mineral particles, water, air and sometimes organic matter. The particles are grouped by size:

• Gravel: particles larger than about 2 mm (roughly the size of a pea or bigger).
• Sand: particles from about 0.075 mm to 2 mm, gritty to the touch and visible individually.
• Silt: fine particles that feel smooth or floury when dry and are too small to see individually.
• Clay: the finest particles, which are sticky and plastic when wet and very hard when dry.

Real soils are mixtures, described by their main components: silty sand, sandy clay, clayey silt, gravelly loam and so on. Glacial till, common across much of Canada and the northern United States, is a dense, poorly sorted mix of everything from clay to boulders. Fill, placed by people, can contain anything, including old foundations, construction debris and voids.

Cohesive and granular soils

Soils fall into two broad behaviour groups:

• Cohesive soils, such as clay and clayey silt, hold together because of the attraction between their fine particles. They can stand in a vertical face for a time, which is exactly what makes them treacherous. A vertical clay face that stands for an hour may fail without warning as it dries and cracks, absorbs water, or relaxes under its own weight.
• Granular soils, such as sand and gravel, have little or no cohesion. Their strength comes from friction between grains. Dry granular soil runs until it reaches its natural angle of repose, typically around 30 to 35 degrees from horizontal. A thin film of moisture can make sand appear to stand vertically, but the apparent cohesion vanishes when the soil dries out or becomes saturated.

Water changes everything. Below the water table, granular soils can flow into a trench as fast as they are dug out, and pressure from groundwater pushes on supports. Clay softens as it absorbs water and loses strength. Freezing creates temporary strength that disappears on thawing, and frost can mask fissures and weak layers that will fail in the spring.

How the regulations classify soil

Ontario's O. Reg. 213/91, section 226, sorts soil into four types according to descriptive properties rather than laboratory numbers:

• Type 1: hard, very dense soil that can be penetrated only with difficulty by a small sharp object, with low natural moisture content, a high degree of internal strength and no signs of water seepage, and which can be excavated only by mechanical equipment.
• Type 2: very stiff, dense soil that can be penetrated with moderate difficulty by a small sharp object, with low to medium natural moisture content, a damp appearance after excavation and a high degree of internal strength.
• Type 3: stiff to firm, compact to loose soil, or previously excavated soil, that exhibits signs of surface cracking or water seepage, may run easily into a well-defined conical pile when dry, and has a low degree of internal strength.
• Type 4: soft to very soft, very loose soil that may run easily or flow unless completely supported before excavating, has almost no internal strength, is wet or muddy, and exerts substantial fluid pressure on its supporting system.

The soil must be classified by visual and other inspection of the excavation walls and of the ground adjacent to the excavation, within a horizontal distance equal to the depth of the excavation. Where an excavation contains more than one type of soil, the whole excavation is treated as the least stable type present. Alberta's OHS Code, section 442, uses three categories with similar descriptions: hard and compact; likely to crack or crumble; and soft, sandy or loose. Mixed soils are again treated as the least stable type.

Visual tests

Visual assessment starts before digging and continues as the excavation deepens:

• Look at the spoil as it comes out. Does it come out in large clumps that hold together, suggesting cohesive soil, or does it run and crumble, suggesting granular soil?
• Look at the walls for layers, especially sloping layers and layers of sand or gravel between clays, which can act as slip planes or water paths.
• Look for cracks parallel to the excavation, fissures in the face and chunks of soil falling off.
• Look for water: seepage from the face, pooling on the floor, a sheen on the walls, a nearby water table or a recent storm.
• Look for signs of previous excavation: old trenches, backfill of different colour or texture, buried utilities, foundations.
• Look at the surroundings: traffic, railways, pile driving or other sources of vibration, nearby buildings and their foundations, and surcharge loads.

Manual tests

Manual tests give a more objective reading of soil strength and plasticity. They are done on a fresh sample taken from the excavation wall, not on dried spoil.

• Thumb penetration test: press a thumb firmly into a fresh face. Soil that can be indented by the thumb only with very great effort is consistent with the strongest cohesive class. Soil that the thumb penetrates several centimetres easily is weak.
• Plasticity or thread test: roll a moist sample between the palms into a thread about 3 mm thick and 50 mm long. If it can be held at one end without breaking, it is cohesive.
• Dry strength test: a dry clump that is very hard to crumble, or that breaks into clumps that cannot be crushed further by hand, is cohesive. A dry sample that crumbles easily into individual grains is granular.
• Pocket penetrometer: a small spring-loaded instrument pushed into the soil face to a marked depth. It reads unconfined compressive strength directly. Several readings should be taken at different points and depths, and the lowest values used.
• Torvane or shear vane: a small vane device pressed into the soil and rotated to measure shear strength, which can be converted to compressive strength.
• Wet shaking or drying tests: shaking a wet sample to see whether water rises to the surface, which indicates silt or fine sand, or drying a sample to see whether it cracks and how it breaks apart.

Factors that override the tests

A strong reading is not the end of the analysis. Several conditions automatically downgrade soil regardless of strength:

• Previous excavation. Backfill is never as strong as undisturbed soil, even if it has been in place for decades. Utility corridors are full of it.
• Fissures and cracks. They form planes along which blocks of soil slide.
• Vibration. Heavy traffic, railways, compaction equipment and pile driving break down soil structure.
• Water. Seepage, a high water table, a nearby watercourse or a leaking main weakens soil and can turn a granular soil into a flowing one.
• Layers dipping into the excavation. A sloping bed of sand or silt between clays can act as a slide plane, especially when wet.
• Surcharge. Spoil piles, buildings, roads and equipment near the edge add load.

When to call a geotechnical engineer

Field classification is adequate for most routine trenches, but some situations need a geotechnical engineer's assessment: deep excavations, excavations next to buildings or roads, high groundwater, organic soils and peat, sensitive clays such as the Leda clay found in the Ottawa and St. Lawrence valleys, permafrost, contaminated ground, or any time the competent person is unsure. A geotechnical report from the design stage, if one exists, should be on site and read before excavation starts. It may describe soil layers, groundwater levels and recommended slopes.

Classifying from the geotechnical report

Where a geotechnical report exists, it should be the starting point, not the final answer. Boreholes are typically tens of metres apart and describe the soil only at those points; the trench will cross ground between them that may be quite different. Use the report to anticipate what the excavation should encounter, then confirm with visual and manual tests as the excavation proceeds, and treat any difference between what the report predicted and what the walls show as a reason to slow down and reassess.