Mehdi Izadpanah

A significant part of monument conservation due to in situ weathering processes requires an assessment of the unconfined compressive strength (UCS) of granite to assess the monument’s structural integrity. Based on the minimum intervention principles of modern monument conservation, the retrieval of intact granite samples from monuments to determine their unconfined compressive strength is generally not possible. A viable alternative is to perform non-destructive tests on the in-situ monument material and correlate it with the unconfined compressive strength of granite. To this end, a site and data independent database comprising three non-destructive test indexes and unconfined compressive strength data of very soft granite corresponding to high weathering degrees likely to be encountered during monument conservation was compiled and used to train artificial neural networks. The results show that the developed ANN significantly outperforms the prediction accuracy of the unconfined compressive strength of granite currently reported in the literature.