Correlated UCS values from PLI

It is a general practice to carryout Point load test for 2-3 specimens retreived from Pile bores and convert to UCS and evaluate the socketing depths with other criterias, such as CR/RQD of adjacent boreholes, sample sizes, angularity/shape of samples, weathering conditions, rock condition and depths of adjacent piles etc. However as per the ASTM D 5731 to obtain accurate PLI values for lump samples at least 20 specimens shall be tested for a sample, and which is not possible/practicle with samples from bored piles. However it is understood testing 2-3 specimens are carried internationally and an accepted criteria and since it is not the sole parameter which determines the socketing the accurancy of the test results are not required to a manner it is requested to carry out as per ASTM D 5731. 

However I would like to know from any expert who dealt in this area, the reliability of UCS recieved from PLI test if we test only 1 specimen, 2 specimen or 3 specimen.



Rock Mechanics Construction Deep Foundations

1 Answer

Votes: 1

David Nutakor


Please note that the purpose of testing at least 20 samples is to be able to estimate the strength value using any available statistical tool. You cannot get a meaningful result when you try to determine the mean value or the standard deviation using results from only three samples. If you have a very limited samples to test, the best practice is to perform both PLI and UCS on a few samples and then develop a graph for the two tests (i.e., UCS vs PLI graph). Once this graph is plotted (hopefully, with a reasonable coefficient of correlation), you can easily determine corresponding UCS values of any PLI tests from the equation of the graph going forward. I have attached a plot of UCS versus PLI on a project that l was actively involved during 2004-2005 timeframe.