DCPT Test

[Geotechnical76]

How accurate is the DCPT test in estimating the CBR and k for pavement design? I understand that it is site dependent since the CBR will be estimated based on the DCPT test results so it will need to be calibrated/correlated based on the practice in that area/site.

 

[Ron]

If you use only the DCPT with no correlation to an actual CBR or plate load test, then it's just a rough approximation.

You can correlate the DCPT to site specific lab data (or field plate load tests), then use the DCPT to check lots of areas on the site without running a bunch of CBR's or plate load tests.

 

[BigH]

. . . and I'd like to again point out - as has been discussed in numerous threads on the site - that DCPT means different things to different people. You are probably thinking of the TRL miniature DCPT; I'm used to a 2 inch 60deg apex cone driven with a 65 kg hammer dropping 760. Others think of the MacIntosh Probe - others to Sowers . . . try to site google previous threads on the subject - there are quite a few.

 

[Ron]

BigH...good point....I was thinking Sowers (after 18 years at the company that George Sowers helped to start, what else would I think!!)....but you're right...others could apply as well. In any case, a site-specific correlation should be done.

 

[iandig]

Follow this link and it will take you to a free copy of IAN 73/06 (Draft DMRB 25) which is the UK's foundation design for road pavements.

http://www.standardsforhighways.co.uk/ians/index.htm

This includes the use of the TRL DCP, plate load tests (static and dynamic), FWD and DPLT etc... If you need more info, it is a case of following the references. The real question may be is how accurate is a CBR, not the other way around. The IAN now relies upon stiffness and refers to CBR as a means of assessing E not the other way around. Plate tests to determine CBR use k, but they MUST be done on repeated cycles to ensure the ground is behaving eleastically. Trying to convert a single cycle plate test to k then to a CBR is full of errors.
The equations in this document are different to other relationships used to compare CBR and k (many of which I have taken from Eng-tips) so be aware and reference which one you have used. For CBR's less than 5%, the IAN will return a CBR lower than other published relationships, so be consistent with what you are using and how the data is presented. On a recent project, using an older widely used relationship (in the UK) the equivalent CBR came out at 5.8%, using the procedure in the IAN it was 5.1%, even though the k value was very similar. (there was also a plate size correction included as the test was done on a 450mm plate).