Method and system for dividing submarine soil strata by pore water pressure static cone penetration

Abstract

The invention discloses a method and a system for dividing submarine soil layer by pore water pressure static cone penetration, which relates to the field of soil layer dividing. The current stratification method in the CPTU index selection has a greater subjectivity, stratification results are not very accurate. The invention comprises the following steps: acquiring original index data, data processing, dimension reduction processing and clustering; based on the K-means clustering method, the self-encoder is used to reduce the dimension of the submarine pore pressure cone penetration test indexes, remove the redundant features, optimize the weights of the features, and cluster the subsets of the features by K-means. It is found that the outline coefficients and other clustering evaluationindexes of the cluster stratification results are greatly improved. The technical scheme utilizes a self-encoder combined with K-means clustering, the accuracy of the interface of the seabed soil layer is high, and the number of soil types, and the result display can be intuitively displayed.

Description

technical field [0001] The invention relates to the field of soil layer division, in particular to a method and system for subsea soil layer division by pore pressure static penetration penetration. Background technique [0002] Due to its rapidity, convenient operation, continuous data, complete functions, and good scalability, the pore pressure static penetration test has gained more and more attention in land and marine engineering geological surveys and geotechnical engineering surveys in recent years. Its data indicators have gradually become an important basis for the evaluation of marine engineering geological conditions and geological hazards. Whether it is engineering geological survey or geotechnical engineering survey, the correct division of strata and identification of soil types are the most basic and important content, which affects the quality of follow-up work. [0003] Compared with the method of drilling combined with laboratory tests, piezostatic penetra...

AI Technical Summary

Problems solved by technology

[0005] At present, the existing stratigraphic soil classification methods mainly rely on manual selection of features. Although there are not many CPTU indicators, there are still redundancy, and some have low correlation with soil types. Selecting such features will cause interference. Selecting indicators Too little will lead to missing information

Moreover, the manually selected features depend on human experience, and different people choose differently, so the reliability is not high.

Images