Rock mass quality classification based on fuzzy comprehensive evaluation and neural network

Abstract

Aiming at the problems of subjectivity and fuzziness in the traditional rock mass quality classification methods, a basic model of rock mass quality classification was established based on the fuzzy theory. According to the engineering geological characteristics of underground water sealed storage, the rock mass quality classification was determined by the method in combination of fuzzy comprehensive evaluation and neural network. A case study was carried out to an underground water-sealed storage in China, specifically, a neural network model of rock mass quality classification with good stability was built with the representative sample data measured on site, and the rock mass quality classification of Borehole ZK1 was also verified. The results show that, compared with the traditional rock mass quality classification methods, the result of rock mass quality classification determined by the fuzzy comprehensive evaluation method is more reasonable. The test results of rock mass quality classification in Borehole ZK1 are in good accuracy, which can meet the actual engineering requirement. The newly constructed rock mass quality classification method of underground water sealed storage is a quantitative analysis method, which provides a theoretical basis for the stability analysis and optimization design of underground water sealed storage.

Keywords

• underground water sealed storage
• fuzzy comprehensive evaluation method
• neural network
• rock mass quality classification