International Journal of Computational
Intelligence Research (IJCIR)
Volume 2, Number 1 (2006)
Application of an artificial neural network on depth to bedrock prediction
Department of Civil Engineering, MingHsin University of Science & Technology, No.1, Hsin-Hsing Road, Hsin-Fong, Hsin-Chu, 304, Taiwan
Department of Civil Engineering, National Central University, ChungLi, Taoyuan, 320, Taiwan
Material and Pavements Section, Texas Department of Transportation, 4203 Bull Creek Road, Austin, TX USA
Despite the importance of bedrock depth on Falling Weight Deflectometer (FWD) backcalculation, there is only limited literature and field verification to address this issue. It would be easiest and least expensive if the bedrock depth could be accurately determined from the FWD data itself. Efforts were made in this study to determine the bedrock depth by using augers. Prior to augering, FWD tests were conducted with the expectation of establishing this correlation at 10 different sites and 84 locations with the wide variety of pavement structures. The correlation of bedrock depth with FWD data is developed by Artificial Neural Networks (ANNs). Ten percent of the data was used solely for testing purposes. After the ANN was developed, 5 % of the data was employed for verification. Based on the results from the training, testing and verification set, the developed ANN represents a significant improvement in accuracy as compared to previous analysis model.
artificial neural network, depth to bedrock, falling weight deflectometer, backcalculation, prediction.