Dimensionless and unbiased CPT interpretation in sand

The cone penetration test (CPT) is widely used, and although initially developed as a stratigraphic logging tool its excellent repeatability and accuracy offers a benchmark quantitative test for sand in particular. A continuing difficulty, however, is that the CPT does not measure any soil property directly, so that parameters of interest must be recovered from solution of an inverse boundary value problem, which is difficult. To date most CPT interpretations in sand have been based on very limited calibration testing carried out in large chambers on a few sands from which mappings are developed. But there are differences in the CPT response from one sand to another leaving the interpretation imprecise (and arguably even speculative) because these differences remain poorly understood. In this paper we use the familiar spherical cavity expansion analogy to the CPT including large strains and a good, critical‐state‐based, soil model to develop a pattern of behaviour which we then compare to some of the reference chamber test data. We find that one of the issues of dispute in the empirical interpretation methods, the so‐called stress‐level effect, is caused by neglect of elasticity and that there are several additional parameters of first‐order significance to cavity expansion in sands. More generally, we show that the difference in CPT response between various chamber sands in predicted. Our results are cast in dimensionless form and the inversion illustrates that extreme care is required in interpreting CPT data if the in situ sand state is to be determined with precision approaching that suggested as achievable by the repeatability of the CPT data itself. Aspects requiring particular care in interpreting CPT data in sand are discussed. © 1998 John Wiley & Sons, Ltd.