RMR Calculator — Bieniawski (1989) Rock Mass Rating, Free

Stepwise, not interpolated — and why that is a choice

Bieniawski published continuous charts for three of the five parameters — intact strength, RQD and spacing — and some implementations read a rating off those charts instead of the table. This one does not. It applies the published intervals as a step function, and the figure says so.

The reason is that the class boundaries, the stand-up-time guidance and every RMR-to-Q and RMR-to-GSI correlation in the literature were calibrated against ratings read from the TABLE. An interpolated 13.6 is not a more accurate 13; it is a different quantity with the same name, and it also invites a precision the method does not have.

There is a concrete consequence, and it is the source of a lot of confusion in secondary sources: the tabulated floors give a minimum RMR of 8, while the continuous charts run down to 0. Bieniawski only spelled that out in 2011. If your spreadsheet and this page disagree about the very worst ground you have logged, that is probably why.

What happens exactly on a boundary

Every step function has a question the printed table appears not to answer: what happens at exactly 250 MPa, or exactly 90 % RQD, or exactly 0.6 m spacing. It answers more of it than it looks. The open bands carry strict inequalities and are printed that way on purpose — "> 250" excludes 250, so 250 MPa belongs to the 100–250 band and scores 12; "< 25" excludes 25, so an RQD of exactly 25 scores 8, not 3; "< 60 mm" excludes 60 mm, so a 60 mm spacing scores 8.

That leaves only the boundaries shared by two written ranges — UCS 100, 50, 25 and 5; RQD 90, 75 and 50; spacing 0.6 m and 0.2 m. Nothing in the source decides those, so this page decides, and says which way: the higher band wins. It is worth stating plainly, because two RMR tools that pick differently disagree by a whole band at exactly those inputs and agree everywhere else — the kind of discrepancy that ends up being blamed on the rock.

The orientation adjustment is the biggest number on the page

The five ratings describe the rock. The adjustment describes the relationship between the rock and your excavation, and it is by far the largest single swing available. The same unfavourably oriented joint set costs a tunnel 10 points, a foundation 15, and a slope 50. Very unfavourable costs a slope 60.

A spreadsheet written for tunnels and then re-used for a cutting is not conservative. It is wrong by four rock-mass classes. So this page asks what you are building before it applies anything, and prints the application on the figure. You can also switch the adjustment off entirely, which gives basic RMR — the five parameters only — and that is the right thing to do when you are characterising ground rather than designing an excavation in it.

Two honest notes on this table. Its slope row is where reputable reproductions actually diverge: some print −60 for very unfavourable and Hoek's own reproduction leaves the cell blank. −60 is used here. And RMR was superseded for slope work by Romana's Slope Mass Rating, which is very probably why that cell was dropped — a slope reading from this page is a starting point, not a slope classification.

Reading the figure

Each scenario is one stacked bar, built bottom-up in the order the published table lists the parameters, so the height of each segment is that parameter's contribution. The orientation adjustment hangs back off the top as a hatched deduction rather than as a negative bar below the axis, because that is what it is: a removal from a total you had already earned. The final RMR89 is ruled across in the scenario's own colour.

Behind everything are the five class bands. That is the point of drawing this rather than tabulating it — you can see at a glance not just which class each scenario is in, but how close it is to the next one, and which single parameter would move it. A bar sitting two points above a boundary with a 4-rated groundwater segment is telling you something a table of totals cannot.

Below the figure, one row per scenario: all five ratings, the basic total, the adjustment, the final RMR89, the class and its description, and the two correlated values.

The correlations, with their conditions attached

GSI ≈ RMR89′ − 5 (Hoek, Kaiser & Bawden 1995). The prime is not decoration: the relation applies to your rating RE-SCORED with groundwater at 15 and the orientation adjustment at zero, because neither is a property of the rock mass. This page does that re-scoring and prints the primed number. The stated limit of 23 is the arithmetic floor of that primed rating, not an empirical threshold. And the caveat carried alongside it every time: Hoek's current position is that this correlation is unreliable, particularly in poor ground, and that GSI should be read off its own chart instead — which is a page on this site.

RMR ≈ 9 ln Q + 44 (Bieniawski 1976, from 111 case histories), inverted here to give a Q from your RMR. The scatter is wide — ±50 % or more — and reviewers of both systems are blunt that there is no scientific basis for assuming a universally valid regression between them: they characterise and weight the rock mass differently, and later authors who fit their own constants to their own regions disagree with each other. Treat it as an order-of-magnitude cross-check. Nothing on this page feeds it into another calculation.

Who uses this

Tunnel engineers producing the classification that a contract, a support class drawing or a payment schedule is written against. Engineering geologists characterising structural domains along a drive or across a site. Mining geotechnicians, who often need RMR and Q side by side because different parts of the same organisation ask for different ones. Students, who need to see the tables rather than a black box that returns 62.

It is a classification tool, not a support design. It will not tell you what to install. What it gives you is the correctly scored, correctly cited rating those decisions start from, with the breakdown visible so somebody can check it.

Common questions

Is this free? What is the catch?

Free, with no account and no gated options. Enter your own scenarios, change every setting, export the figure. Free exports carry a small watermark, as they do everywhere on this site. There is no catch — RMR is five published tables and an addition.

Can I type a rating instead of a description?

Yes, for the two descriptive columns. Joint condition takes either a phrase from the published table or a bare 0–30 rating, and groundwater takes either a condition or a 0–15 rating. The two ways people arrive here are opposite: one has a logging sheet with words on it, the other has a spreadsheet column of numbers from a colleague and only wants the total, the class and the correlations. The results table prints which route each row took.

What about the joint condition sub-ratings?

The 1989 guidelines decompose joint condition into five sub-parameters — persistence, aperture, roughness, infilling and weathering — scored 0 to 6 each, summing to the same 0–30. Score them yourself and type the total into the condition column. They are not six separate grid columns on purpose: six columns to score one parameter is a worse table than one column that accepts a number. One caution from the original, worth repeating: where infilling is present it overshadows roughness, and you should use the descriptive table directly rather than summing.

Why does the spacing rating look conservative?

Because it assumes three joint sets, and Bieniawski notes that with fewer than three the rating should be increased by about 30 %. This page does not apply that automatically — it has no way of knowing how many sets you logged, and silently inflating a parameter is exactly the kind of hidden step the page exists to avoid. It is printed as a note on the figure instead.

How reliable is the stand-up time?

Treat it as guidance, and read it with the span it belongs to — "one week for a 5 m span", never "one week". It is also the one part of the published table this page will not present as settled: classes III to V are unanimous across every reproduction, but classes I and II appear in two printed generations (20 years at 15 m and 1 year at 10 m, or 10 years at 15 m and 6 months at 8 m) and no reproduction pins either pair to the 1989 book. Both are shown, with the disagreement named. Picking one would be inventing a citation.

Which RMR is this, exactly?

RMR89, from Bieniawski (1989), Engineering Rock Mass Classifications, Wiley. Ratings are read from the tables, not interpolated from the charts. One warning about secondary sources: at least one widely circulated summary sheet prints 5 rather than 3 for the RQD < 25 band and reproduces the 1976 stand-up times under a 1989 heading. The peer-reviewed comparisons of the three editions give the RMR89 RQD range as 3 to 20, which is what this page uses.

Does anything leave my browser?

No. The whole calculation runs in this tab, there is no account and there is no upload.