ToolFusion
📋 Case Study

Limestone Mine Digital Twin for Karst-Related Grade Uncertainty

Solution cavities caused unpredictable grade drops (CaCO₃ purity <85%) in otherwise uniform deposits, resulting in 11% product rejection

🏗️ Project Overview

Karst-hosted limestone quarry, USA

🎯 Challenge

Solution cavities caused unpredictable grade drops (CaCO₃ purity <85%) in otherwise uniform deposits, resulting in 11% product rejection

🔧 Design Approach

Ground-penetrating radar (GPR) + micro-gravity time-series fed into physics-informed neural network; digital twin updated hourly with cavity growth simulation

📐 Design Diagram

GPRMicro-GravityPhysics-Informed NN(Hourly Digital Twin Update)GPR data streamMicro-gravity time-seriesGrade DropCaCO₃ <85%Alert output• Min void: 0.8 m³ @ 12m depth• Forecast horizon: 4.2 days• Rejection rate: 11%

AI-generated project design illustration

📐 Key Calculations

Cavity Detection Sensitivity

Min detectable void volume
Result: 0.8 m³ @ 12m depth
Enables pre-blast cavity avoidance

Grade Purity Forecast Horizon

Time until predicted purity drop
Result: 4.2 days
Triggers proactive stockpile management

📊 Results

Product rejection reduced from 11% to 1.9%, quarried tonnage increased 13% without expanding footprint, 98% compliance with ASTM C150 cement specs

💡 Lessons Learned

  • •Physics constraints prevented unrealistic cavity morphologies
  • •Micro-gravity drift compensation required daily base station recalibration
  • •GPR antenna frequency selection was site-specific (400 MHz optimal)

✅ Key Takeaways

  • 1Physics constraints prevented unrealistic cavity morphologies
  • 2Micro-gravity drift compensation required daily base station recalibration
  • 3GPR antenna frequency selection was site-specific (400 MHz optimal)