84 lnfad rock: deformabi/i% strength and failure
Hoek-Brown criterion
We can either simply take appropriate values for the parameters in the
criterion from a table or attempt to statistically fit the criterion to the
data. Using the former method, we have s = 1 and m = 15, and with
O, = 200 MPa we obtain:
03 -13.3 0 5 10 20 50 100
-3.3 200.0 239.5 274.6 336.2 485.9 683.1
Using the latter method, we rearrange the criterion to
(2 - z)’ = m, a3 +s
which corresponds to the linear form y = mx + c, and obtain values for
m and s by linear regression. In this case, we know that s = 1, and so we
perform a linear regression on
while constraining the solution to pass through the origin. The result is a
parameter value of m = 20.5, with corresponding stress values of
03 -13.3 0 5 10 20 50 100
01 -10.4 200.0 251.0 294.6 369.3 545.1 770.9
-20
-100
20 40 60 80 100
Assessment
As the plot above shows, the statistical fit of the Hoek-Brown criterion is
the best of the three criteria. The curvilinear nature of the results means
that the Mohr-Coulomb criterion will never be a good fit over the entire
stress range, and the fundamental behaviour of rock over large stress
ranges means that this is a general conclusion.