Damage Identification of Rock Mass with Artificial Neural Networks

Zi Chang Shangguan; Shou Ju Li; Mao Tian Luan

doi:10.4028/www.scientific.net/KEM.353-358.2325

Paper Titles

Application of Nondestructive Detecting Technology for Geological Radar in a Large-Scale Highway Lanslide Treatment
p.2309

Infrared Thermal Image Study on Failure of Granite with Hole
p.2313

Damage Detection Based on the Cross Correlation Function Amplitude Vector and its Application to the ASCE Benchmark Structure
p.2317

Influence of Heterogeneity on Direct Tensile Failure Process of Rocks and Associated Fractal Characteristic of AE
p.2321

Damage Identification of Rock Mass with Artificial Neural Networks
p.2325

Experimental Study on Crack Propagation by AE Location under Uniaxial Compression Condition
p.2329

Experimental Verification and Theoretical Analysis on Felicity Effect of Acoustic Emission in Concrete
p.2333

Statistical Damage Detection of Laminated CFRP Beam Using Electrical Resistance Change Method
p.2337

Compaction Quality of Bitumen Mixtures Based on Ultrasonic Methods
p.2341

HomeKey Engineering MaterialsKey Engineering Materials Vols. 353-358Damage Identification of Rock Mass with Artificial...

Damage Identification of Rock Mass with Artificial Neural Networks

Abstract:

The inverse problem of rock damage detection is formulated as an optimization problem, which is then solved by using artificial neural networks. Convergence measurements of displacements at a few of positions are used to determine the location and magnitude of the damaged rock in the excavation disturbed zones. Unlike the classical optimum methods, ANN is able to globally converge. However, the most frequently used Back-Propagation neural networks have a set of problems: dependence on initial parameters, long training time, lack of problemindependent way to choose appropriate network topology and incomprehensive nature of ANNs. To identify the location and magnitude of the damaged rock using an artificial neural network is feasible and a well trained artificial neural network by Levenberg-Marquardt algorithm reveals an extremely fast convergence and a high degree of accuracy.