|
چکیده
|
The Geological Strength Index (GSI) is an important metric in rock mechanics and geotechnical engineering that indicates the structural quality and mechanical properties of rock masses. Its precise assessment is essential for the design, safety, and cost-effectiveness of underground structures, mining operations, and slope stability evaluations. Nonetheless, conventional techniques for determining GSI largely depend on field observations and qualitative assessments, which frequently are subjective and challenging to replicate. These limitations often lead to inconsistencies in classification and may affect the reliability of engineering design decisions. To overcome these shortcomings, the present research employs artificial intelligence (AI) and metaheuristic optimization techniques to establish robust, data-driven predictive models for GSI estimation. A comprehensive dataset was compiled from the Beheshtabad water transfer tunnel in southwestern Iran, comprising paired measurements of compressional (Vp) and shear (Vs) wave velocities, together with field-evaluated GSI values. Because seismic velocities are relatively inexpensive and easy to measure, they serve as practical predictor variables to replace or complement traditional classification systems such as rock mass rating (RMR) and Q-system, which require extensive field and laboratory investigations. Three modeling strategies were explored: the trust region reflective (TRR) algorithm, the genetic algorithm (GA), and a hybrid TRR–GA approach. To model the nonlinear dependency between GSI and the seismic parameters, five mathematical formulations, including exponential, logarithmic, power, quadratic polynomial, and linear functions, were systematically tested to determine the most suitable representation of the relationship. Among these, the quadratic polynomial model consistently achieved the highest predictive accuracy across all algorithms. Model performance was evaluated using a suite of statistical metrics, including error and accuracy indices, along with additional indicators such as index of scatter (IOS), index of agreement (IOA), and the A20 accuracy index. Comparative analyses revealed that the hybrid TRR–GA model produced the most accurate and stable predictions, outperforming the standalone TRR and GA models. Although both models based on the Vp and Vs yielded satisfactory results, the Vp-based formulations demonstrated slightly superior predictive capability. In all analyses, the order of the models was stable: TRR-GA was the top performer, succeeded by TRR and GA. These results emphasize the benefits of combining global optimization through GA with local refinement using TRR, which improves model stability and decreases prediction inaccuracies. Significantly, the research shows that GSI can be accurately forecasted through inexpensive seismic measurements, providing a feasible and dependable substitute for subjective evaluations conducted in the field. The suggested hybrid TRR-GA framework offers significant insights for geotechnical engineers, enhancing decision-making in tunneling, excavation, and rock stability evaluation.
|