Abstract: Accurately obtaining the geological information of soft foundation is an essential basis for determining the construction technique and controlling the pile quality of vibro-replacement stone columns. The existing geological exploration technology used to determine stratum information is considerably random and discrete, which makes it impossible to comprehensively understand the geological conditions of the reinforced areas. To overcome these limitations, this study relies on a large amount of data related to stratum classification attributes collected by the real-time monitoring system during the construction process of vibro-replacement stone columns at Lawa Hydropower Station. By cleaning big data, features such as penetration depth, speed, and current related to stratum classification attributes were selected for fuzzy C-means clustering algorithm-based study of stratum identification of the soft foundation. The results indicate that compared to the traditional K-means algorithm, the method proposed in this paper exhibits higher accuracy and superiority in identifying strata and enables real-time research and judgment of geological conditions. The research findings presented in this paper are of great significance in the rational evaluation of vibro-replacement stone column construction quality and the intelligent construction of the pile formation process.

Key words: vibro-replacement stone column, stratum identification, fuzzy C-means clustering algorithm, real-time monitoring system, construction process parameters