Answers are grounded in this paper and cite their sources.
Through systematic field investigation in the Helan Petroglyph Area, this study documented the developmental characteristics of 146 spalling disease cases alongside fundamental carrier rock properties. We established a comprehensive hierarchical model integrating three critical dimensions: rock physical-mechanical parameters, fracture development characteristics, and petroglyph spatial distribution patterns. The fuzzy analytic hierarchy process (FAHP) was implemented to compute precise indicator weights and generate quantitative spalling development scores (K), enabling systematic classification into three distinct risk tiers: moderate (100 cases), low (11 cases), and high (35 cases). Observed discrepancies between evaluated classifications and field conditions highlighted both methodological subjectivity in FAHP implementation and the complex etiology of spalling pathologies.Advanced machine learning approaches were subsequently employed, with C1-C11 indicators serving as input variables and K scores as targets. Two optimized neural network architectures GA-BP and PSO-BP demonstrated exceptional predictive capability, achieving determination coefficients (R2) exceeding 0.9 with minimal error metrics (MAE, MSE, RMSE approaching 0). Comparative analysis revealed the PSO-BP model’s superior accuracy in spalling prediction. Comprehensive correlation analysis through Pearson coefficients identified four predominant influencing factors: rebound strength (C1), particle density (C5), crack proximity (C7), and crack width (C9), providing valuable references for the conservation of petroglyph cultural heritage.
No takes yet. Share an insight, caveat, or question.
Yu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: