Research on factor analysis and method for evaluating grouting effects using machine learning

Area of Science:

• Geotechnical Engineering
• Hydrogeology
• Machine Learning Applications

Background:

• Grouting engineering effectiveness evaluation is crucial but complex.
• Traditional methods often rely on empirical data, limiting precision.
• Directional drilling and grouting are used for limestone aquifer reformation.

Purpose of the Study:

• To analyze factors influencing grouting effects in limestone aquifers.
• To develop and validate a machine learning model for assessing directional grouting effectiveness.
• To compare machine learning model performance against traditional evaluation methods.

Main Methods:

• Established a "dual-process, multi-parameter, and multi-factor" system for analysis.
• Employed correlation analysis to validate selected engineering indices.
• Utilized eight machine learning models and three optimization algorithms, including Adaboost and genetic algorithms.

Main Results:

• The developed system clarifies relationships between influencing factors and engineering parameters.
• Eight selected indicators showed low correlation, validating their use as independent evaluation metrics.
• The Adaboost model, optimized by a genetic algorithm, achieved superior accuracy in evaluating grouting effects.

Conclusions:

• The "dual-process, multi-parameter, and multi-factor" system effectively models grouting effect intricacies.
• Machine learning, particularly the optimized Adaboost model, offers a more precise and realistic assessment of directional grouting effectiveness.
• The findings provide a robust framework for improving grouting engineering practices.