A New Approach to the Integrity Testing of Ground Anchorages

Ground anchorage systems in the form of rock bolts are used extensively throughout the world as supporting devices for civil engineering structures such as bridges and tunnels. It is widely recognized that non-destructive testing methods for ground anchorages need to be developed as a high priority [1 ], with only a small proportion of anchorages currently being monitored in service [2]. The condition monitoring of anchorages is a relatively new area of research, with the objective being a wholly automated or semi-automated condition monitoring system capable of repeatable and accurate diagnosis of faults and anchorage post tension levels. The GRANIT tm system developed by the University of Aberdeen is capable of measuring the integrity of ground anchorages and has won two major UK Awards - the National John Logic Baird Award for Innovation in 1997 and a Design Council Millennium Product Award in 1999. The normal procedure with the GRANIT tm system [3,4] is to train a neural network on dynamic response data that has been sampled from an anchorage over a range of post-tension levels. Further data is needed in order to test the neural network. The paper presents neural networks that have been trained on data taken from a number of anchorages on different sites, and the results of the diagnosis of these neural networks on new test data is presented. The results show that the cross-diagnosis of anchorages is possible, where a neural network trained on one anchorage is capable of successfully diagnosing the post-tension level for an anchorage of similar construction, and the high level of accuracy that can be attained with the GRANIT tm technique is shown. The paper also introduces a mathematical model of the GRANIT tm process.