Barton Q-System Rock Mass Quality

Nick Barton, Reidar Lien and Johan Lunde published the Q-system in 1974 out of the Norwegian Geotechnical Institute, fitted to around two hundred tunnelling case records. It multiplies six ratings in three pairs, and the pairing is not decoration — each pair answers a different question about the rock, and reading them separately tells you far more than the product does.

$RQD/J_n$ is block size. How big are the pieces the mass has been divided into? A high RQD divided by a low joint set number means large blocks; a low RQD over four joint sets and random ones means gravel. $J_r/J_a$ is inter-block shear strength. Whether a rough, clean, undulating joint will lock and dilate, or a slickensided one with a clay coating will slip; the ratio behaves roughly like the tangent of a friction angle, and that is exactly how Barton intended it. $J_w/SRF$ is active stress. What the water and the stress field are doing to the mass right now — and it is the pair most likely to be different on the day of excavation from what it was during the investigation.

Q spans six orders of magnitude, from about 0.001 for squeezing ground to 1000 for massive unjointed rock, which is why it is plotted logarithmically and why it works well as a support-selection tool. Barton's chart pairs Q with the equivalent dimension of the opening and returns a support category, and that chart — rather than the number — is what the system was built to deliver.

Four of the six terms are ratings, not measurements. $J_n$, $J_r$, $J_a$ and $SRF$ are read off descriptive tables by a person standing at a face, and two competent engineering geologists will not always pick the same row. Because Q is a PRODUCT of ratios rather than a sum, those disagreements multiply instead of averaging out: three terms each off by one row in the same direction can move Q by a factor of five. Quote a range and not a point.

SRF is the worst of them and deserves naming. It is the term that has been revised most since 1974, it does duty for loosening, squeezing, swelling and high stress all at once, and moving it from 1 to 5 divides Q by five without anything about the rock having changed. It is also the term most easily argued down when a project needs a better number. If you find yourself adjusting SRF to reach a Q you had in mind, you have stopped assessing and started negotiating.

Note finally what Q does not contain. Joint orientation relative to your excavation appears nowhere in it. Barton's argument was that $J_n$ and $J_r$ already carry most of the effect and that the case records supported leaving it out. For a tunnel that is defensible. For a slope, where the relationship between a joint set and the face is usually the whole of the problem, it is not, and Q is simply the wrong instrument.