Ground bearing pressure: five ways engineers get it wrong

A crane on outriggers is a four-legged table carrying a load that moves. The counterweight sits behind the slew centre, the boom and the hoisted load sit in front of it, and as the superstructure slews, that whole arrangement rotates over a fixed rectangle of support.

At no point does each leg carry a quarter.

Take the worked machine: 1400.0 kN gross, on outriggers at a 3.0 m front span, 3.0 m rear span and 2.5 m half-width, with 0.6 m square floats on 2.0 m timber mats over ground declared good for 250 kPa.

The manufacturer's chart for the same machine gives 420.0 kN on the front left leg with the boom over the front, which is 20.0 percent more than the average. That is before the crane turns.

Use the chart. The manufacturer's outrigger load chart is the primary source for these reactions and any estimate is a fallback for when it is not available. What the chart will not do is tell you which of its rows governs, and it will not tell you anything about the mat.

Rotate the same machine and the corner loads redistribute. Over the front, two legs share the boom side. Over a corner, one leg takes most of it, because the resultant of the counterweight, the boom and the load passes closest to that leg. On the worked machine, slewing from the parked position to 90 degrees puts 45 percent more load on the worst leg.

The worked machine's chart, swept:

Two habits follow, and both are cheap:

• Sweep the chart, do not sample it. If the lift plan permits the crane to slew through an arc, every angle in that arc is a load case.
• Name the governing leg and angle in the output. "1694 kPa" is not a result. "1694 kPa under the FL float at 90 degrees of slew" is a result somebody can argue with, which is what makes it useful.

This is the mistake that produces the most confidently wrong numbers, because the arithmetic looks so reasonable. The mat is 2.0 m by 2.0 m, which is 4 square metres, so a 610 kN leg gives 153 kPa, which is comfortably inside 250 kPa, so the mat is fine.

It is not fine. The float is 0.6 m square. Unless the mat is stiff enough to distribute load out to its edges, the pressure under it is concentrated beneath the float, and the area doing the work is nearer 0.36 square metres than 4.

Declaring a spread angle is how you say what the mat is capable of, and it is a claim about the mat that has to be defensible: its thickness, its material, its construction, and the load it is spreading. Thirty degrees through a solid timber mat is a common and defensible starting point. Thirty degrees through 0.2 m of timber buys 0.115 m of width on each side, and no more.

Notice a pair of numbers in the table above that a single-check calculation would never separate:

• The float puts 1694 kPa into the mat.
• The mat puts 883 kPa into the ground.

Those are both real, both are checks, and they have completely different allowables. Timber in compression perpendicular to the grain will take a few megapascals; the ground under it will take a few hundred kilopascals. Roughly a factor of ten separates them, and that factor is the whole reason a mat is worth putting down.

There is a third check, and it is the one that catches out large thin mats: the mat is a structure. Under the float it is a plate spanning outwards, so it has a bending check and a punching shear check of its own. A mat that fails in punching under the float does not spread anything; the float goes through it and the crane is standing on the float again.

The two-level verdict from the worked case says it plainly:

A calculation that reports one pressure and one allowable has silently chosen which of these three questions it is answering, and it has usually chosen the easy one.

Presumed allowable bearing values published for preliminary design span roughly an order of magnitude between soft clay and dense gravel. They are useful, they are in every foundation textbook, and they come with assumptions that get dropped when they are copied:

• Drainage. Values are quoted for a stated groundwater condition, and a saturated fine-grained soil is a different material.
• Depth. Bearing capacity increases with embedment, and a crane mat is on the surface, which is the least favourable case there is.
• Settlement. Presumed values are usually settlement-governed, not strength-governed, and a crane cares about differential settlement between legs far more than about the absolute value.
• Proximity. A value is meaningless within the influence zone of an excavation, a slope, a buried service or a basement wall.

Two honest ways forward, and the choice between them is a project decision, not a calculation:

Have the value confirmed for the actual setup position. A geotechnical engineer confirming that this ground, at this position, in this condition, may be taken at a stated value is the shortest defensible route.

Calculate a capacity from tested parameters. With a friction angle, a cohesion, a unit weight and a groundwater depth you can compute a bearing capacity from classical theory, apply your own factor, and show your working. The US Army Corps of Engineers publishes a complete manual on doing exactly that, free, which makes this the rare geotechnical calculation a reader can start the same afternoon.

And where the ground is not good enough at any credible value, the answer is neither a bigger mat nor a more generous table.

Both instruments state a duty and leave the arithmetic to you. That is not a gap: it is the correct division. A regulation cannot know your machine, your chart, your mats or your soil. What it can do is make it somebody's job to have found out, and both of these do.