Quality Evaluation of Rock Mass Using RMR14 Based on Multi-Source Data Fusion

Area of Science:

• Geotechnical Engineering
• Data Science
• Geology

Background:

• Rock mass quality evaluation faces uncertainties due to multi-source testing data.
• Existing methods require robust data-driven approaches for improved accuracy.

Purpose of the Study:

• To develop a data-driven rock mass quality evaluation method using multi-source data fusion.
• To enhance the Dempster-Shafer (D-S) evidence theory for correlated rock mass indices.

Main Methods:

• Applied Dempster-Shafer (D-S) evidence theory for multi-source data fusion.
• Incorporated belief reinforcement and Murphy's average belief theory to handle correlated rock mass data.
• Developed the method based on the Rock Mass Rating 14 (RMR14) system.

Main Results:

• The proposed method significantly amplified the distance between possibilities for different ratings (0.0666 to 0.5882).
• This amplification leads to more accurate and distinct quality evaluations compared to conventional D-S theory.
• A case study in Daxiagu tunnel confirmed the method's practical value, yielding a Level III rock mass rating with high confidence (0.9838).

Conclusions:

• The enhanced D-S evidence theory provides a more precise rock mass quality evaluation framework.
• The method effectively mitigates uncertainties from multi-source data in geological assessments.
• Validated through simulation and a real-world tunnel case study, demonstrating practical applicability.