Application of BP Neural Network in Identifying Soil Strata by CPTU

摘要

Piezocone penetration test (CPTU) is a new technique of in-site testing for soils, and it is used widely in the engineering investigation and soil testing today. Soil strata can be identified according as classifying diagram from CPTU. But, because the relation between CPTU parameters and depth is obviously nonlinear, the method to identify soil strata not only takes time and also brings forth biggish error. In this paper the relation between CPTU parameters and soil types and strata is analyzed, the structure of BP Neural network is designed and the application program is programmed with MATLAB language. The application results of two test segments show that BP neutral network can identify the mud and muddy soil, sand and clay and the identify result is trusty. This research accomplishes identifying soil strata automation, supplies a new way to deal with the CPTU datum and has actual significance in promoting efficiency and precision of CPTU data processing. It will take important function in CPTU data processing automation in the future.