Barton Roof Support Pressure (simplified)

This is the form of Barton's roof support pressure that most people meet first, and it is the one to use when the rock mass has three or more joint sets — which, in most tunnels, it does. With $J_n$ at 9 or above the block size term has stopped discriminating, and two ratings carry the whole estimate: the quality index $Q$ and the joint roughness number $J_r$.

The cube root is the most important thing on the page. $Q$ covers six orders of magnitude, from 0.001 in squeezing ground to 1000 in massive rock. The support pressure covers two. That compression is entirely due to the exponent of $-1/3$: ground a hundred times worse needs only about 4.6 times the support pressure. And that is what makes a rating system with four subjective terms in it usable at all. An argument between two engineering geologists about the stress reduction factor that moves $Q$ by a factor of five moves this answer by about 1.7 — which is comfortably inside the scatter of the case records the line was drawn through in the first place.

The same compression is also a limit. The equation cannot discriminate finely, and quoting its answer to three significant figures claims a precision the underlying cloud of tunnels never had. Two significant figures is generous. Run it at the high and the low end of your $Q$ range and quote the pair; a single number here is a fiction dressed as a result.

Read the cube root the other way and it turns nasty. Solving backwards for $Q$ from a support pressure raises everything to the third power, so a 20% error in the pressure becomes a 70% error in the quality index it implies. The direction that forgives uncertainty going forward amplifies it coming back. There is a legitimate use for the reverse — checking whether an installed support is consistent with the ground being logged is a real audit — but it is not a way to arrive at a rock mass rating, and a $Q$ reverse-engineered from the support you would like to install is a circle with no ground in it.

What $J_r$ is really saying. It runs from 0.5 for a slickensided planar joint to 4 for a discontinuous one, and it behaves roughly like the tangent of a friction angle: it is a measure of whether a block will lock or slip. Notice that it appears in the numerator of $Q$ and in the denominator here, and both are the same physical statement made twice. A rough, undulating, clean joint raises the quality index and lowers the pressure the support has to carry, because the rock is doing some of the work of holding itself up. A clay-coated slickenside does neither.

Two closing cautions that belong on every page in this shard. The leading 2 is 2 kg/cm², a pressure and not a coefficient. And this number is a pressure, not a support system: turning it into bolts and shotcrete is a design decision made against a standard by a person who signs it.