Neural Networks for Backcalculation of Moduli from SASW Test. Williams, T. P abstract bibtex The spectral-analysis-of-surface-waves (SASW) method is a seismic technique for the in-situ evaluation of elastic moduli and layer thicknesses for pavement and soil systems. The inversion process currently used to analyze the experimental dispersion curves from SASW tests is complex, and requires experienced test operators. Prototype neural-network models have been developed to perform the inversion of SASW test results. Three-, four-, and five-layer back-propagation models are employed. The use of a general-regression neural-network model has also been studied. Ninety-eight cases of synthetic dispersion-curve data were developed to train and test the neural networks. All of the neural-network models produced results that were reasonably close to the actual output. Best results were found with back-propagation neural networks using multiple hidden layers and jump connections. The results indicate that back-propagation neural networks are useful for performing the inversion procedure of SASW tests.
@article{williams_neural_nodate,
title = {Neural {Networks} for {Backcalculation} of {Moduli} from {SASW} {Test}},
abstract = {The spectral-analysis-of-surface-waves (SASW) method is a seismic technique for the in-situ evaluation of elastic moduli and layer thicknesses for pavement and soil systems. The inversion process currently used to analyze the experimental dispersion curves from SASW tests is complex, and requires experienced test operators. Prototype neural-network models have been developed to perform the inversion of SASW test results. Three-, four-, and five-layer back-propagation models are employed. The use of a general-regression neural-network model has also been studied. Ninety-eight cases of synthetic dispersion-curve data were developed to train and test the neural networks. All of the neural-network models produced results that were reasonably close to the actual output. Best results were found with back-propagation neural networks using multiple hidden layers and jump connections. The results indicate that back-propagation neural networks are useful for performing the inversion procedure of SASW tests.},
language = {en},
author = {Williams, Trefor P},
pages = {8},
}
Downloads: 0
{"_id":"3LXT297zWSwtu9qCj","bibbaseid":"williams-neuralnetworksforbackcalculationofmodulifromsaswtest","author_short":["Williams, T. P"],"bibdata":{"bibtype":"article","type":"article","title":"Neural Networks for Backcalculation of Moduli from SASW Test","abstract":"The spectral-analysis-of-surface-waves (SASW) method is a seismic technique for the in-situ evaluation of elastic moduli and layer thicknesses for pavement and soil systems. The inversion process currently used to analyze the experimental dispersion curves from SASW tests is complex, and requires experienced test operators. Prototype neural-network models have been developed to perform the inversion of SASW test results. Three-, four-, and five-layer back-propagation models are employed. The use of a general-regression neural-network model has also been studied. Ninety-eight cases of synthetic dispersion-curve data were developed to train and test the neural networks. All of the neural-network models produced results that were reasonably close to the actual output. Best results were found with back-propagation neural networks using multiple hidden layers and jump connections. The results indicate that back-propagation neural networks are useful for performing the inversion procedure of SASW tests.","language":"en","author":[{"propositions":[],"lastnames":["Williams"],"firstnames":["Trefor","P"],"suffixes":[]}],"pages":"8","bibtex":"@article{williams_neural_nodate,\n\ttitle = {Neural {Networks} for {Backcalculation} of {Moduli} from {SASW} {Test}},\n\tabstract = {The spectral-analysis-of-surface-waves (SASW) method is a seismic technique for the in-situ evaluation of elastic moduli and layer thicknesses for pavement and soil systems. The inversion process currently used to analyze the experimental dispersion curves from SASW tests is complex, and requires experienced test operators. Prototype neural-network models have been developed to perform the inversion of SASW test results. Three-, four-, and five-layer back-propagation models are employed. The use of a general-regression neural-network model has also been studied. Ninety-eight cases of synthetic dispersion-curve data were developed to train and test the neural networks. All of the neural-network models produced results that were reasonably close to the actual output. Best results were found with back-propagation neural networks using multiple hidden layers and jump connections. The results indicate that back-propagation neural networks are useful for performing the inversion procedure of SASW tests.},\n\tlanguage = {en},\n\tauthor = {Williams, Trefor P},\n\tpages = {8},\n}\n\n","author_short":["Williams, T. P"],"key":"williams_neural_nodate","id":"williams_neural_nodate","bibbaseid":"williams-neuralnetworksforbackcalculationofmodulifromsaswtest","role":"author","urls":{},"metadata":{"authorlinks":{}}},"bibtype":"article","biburl":"https://bibbase.org/zotero/kongzh95","dataSources":["tAdPv9cJdcbuFicdv"],"keywords":[],"search_terms":["neural","networks","backcalculation","moduli","sasw","test","williams"],"title":"Neural Networks for Backcalculation of Moduli from SASW Test","year":null}