Data fusion technique for predicting shear strength and stress history from piezocone penetration tests

摘要

Existing methods used to infer soil properties from piezocone penetration test (PCPT) data are not always reliable due to the complexity of cone penetration. This study examines the feasibility of training an artificial neural network (ANN)-based data fusion model to estimate soil properties, including overconsolidation ratio (OCR), coefficient of lateral earth pressure at rest (K_o), and undrained shear strength (s_u), directly from multiple piezocone penetrometer sensor measurements. Additional features were created by mathematically combining the PCPT measurements in a manner consistent with the work of previous researchers in an attempt to improve the performance of the trained data fusion model. Overall, the values of OCR, K_o, and s_u predicted by the data fusion models were found to compare very well with the reference values and to be generally more reliable than the results of the current interpretation methods.