🎓 Lesson 24 D5

Equation of State–Based Digital Twins for Real-Time Optimization

A digital twin for blasting is a real-time computer model of a blast site that uses thermodynamic equations to predict how explosives and rock will behave—like a live simulation that helps engineers adjust plans before detonation.

🎯 Learning Objectives

  • Calculate detonation pressure and specific energy output using the JWL equation of state for common ANFO and emulsion explosives
  • Design a real-time feedback loop that adjusts burden and spacing based on EOS-predicted fragment size distribution
  • Analyze EOS model fidelity by comparing simulated peak particle velocity (PPV) against field accelerometer measurements
  • Explain the trade-offs between computational speed and thermodynamic accuracy when selecting an EOS for embedded digital twin applications

📖 Why This Matters

In modern open-pit mines, unplanned overbreak, flyrock, or poor fragmentation cost millions annually in rework, downtime, and safety incidents. Traditional blast design relies on static empirical rules (e.g., Kuz-Ram), but rock properties change with depth, moisture, and jointing—and explosives behave differently under confinement. EOS-based digital twins close this gap: they translate physics-first thermodynamic models into actionable, real-time blast optimization—enabling adaptive drilling, dynamic stemming adjustments, and AI-driven charge tuning *minutes* before detonation.

📘 Core Principles

At its foundation, an EOS describes the relationship among pressure (P), specific volume (V), internal energy (E), and temperature (T) for energetic materials under extreme conditions (10–50 GPa, >3000 K). For blasting, the JWL EOS is industry-standard because it accurately captures the non-ideal detonation products of ammonium nitrate–fuel oil (ANFO) and water-gel explosives across varying densities and confinements. Digital twins embed this EOS within a coupled multi-physics framework: linking explosive reaction kinetics → shockwave propagation → rock fracture mechanics (via Mohr-Coulomb or damage models) → vibration and fragmentation outputs. Real-time updates come from IoT sensors (strain gauges, borehole pressure transducers, microseismic arrays) that feed boundary conditions back into the model—transforming it from a pre-blast simulator into a closed-loop control system.

📐 JWL Equation of State

The JWL EOS expresses pressure as a function of relative volume and internal energy, enabling prediction of detonation product behavior critical for blast modeling accuracy.

Jones-Wilkins-Lee (JWL) EOS

P = A·exp(−R₁·V) + B·exp(−R₂·V) + ω·E/V

Predicts pressure of detonation products as a function of relative volume and internal energy; used to initialize shockwave propagation in blast simulation.

Variables:
SymbolNameUnitDescription
P Pressure Pa Detonation product pressure
A, B JWL coefficients Pa Material-specific energy constants derived from cylinder tests
R₁, R₂ Volume exponents dimensionless Control decay rates of exponential pressure terms
ω Grüneisen coefficient dimensionless Relates internal energy to pressure work in condensed explosives
V Relative specific volume dimensionless V = v/v₀, where v₀ is initial specific volume
E Internal energy per unit mass J/kg Energy released by chemical reaction, including kinetic and thermal components
Typical Ranges:
ANFO (ρ = 0.8–0.9 g/cm³): A: 5.0–6.5×10¹¹ Pa; B: 6.0–9.0×10⁹ Pa; ω: 0.25–0.35
Emulsion explosives (ρ = 1.1–1.3 g/cm³): A: 8.5–11.2×10¹¹ Pa; B: 1.2–2.1×10¹⁰ Pa; ω: 0.20–0.30

💡 Worked Example

Problem: Given: ANFO density = 0.85 g/cm³, detonation velocity = 4800 m/s, measured detonation pressure = 12.4 GPa. Using standard JWL coefficients for ANFO (A=5.96×10¹¹ Pa, B=7.52×10⁹ Pa, R₁=4.15, R₂=0.95, ω=0.30), calculate predicted P at V = 1.2 (i.e., 20% expansion from initial specific volume) and E = 2.1 MJ/kg.
1. Step 1: Convert specific volume ratio V = V/V₀ = 1.2; compute exponential terms: exp(−R₁·V) = exp(−4.15×1.2) ≈ 0.0067; exp(−R₂·V) = exp(−0.95×1.2) ≈ 0.316
2. Step 2: Plug into JWL: P = A·exp(−R₁·V) + B·exp(−R₂·V) + ω·E/V = (5.96e11)(0.0067) + (7.52e9)(0.316) + (0.30)(2.1e6)/1.2
3. Step 3: Compute: 3.99e9 + 2.38e9 + 5.25e5 ≈ 6.37 GPa — compare to measured 12.4 GPa; discrepancy indicates need for density-corrected calibration or confinement adjustment.
Answer: The predicted detonation pressure is 6.37 GPa, which falls below the measured 12.4 GPa—highlighting the importance of confinement correction and site-specific JWL parameter calibration per ASTM D7603.

🏗️ Real-World Application

At Newmont’s Boddington Mine (Western Australia), an EOS-based digital twin integrated JWL-calibrated ANFO models with real-time LiDAR-derived rock mass rating (RMR) maps and down-the-hole pressure sensors. During a 12-row production blast, the twin detected reduced confinement in a weathered shear zone (identified via resistivity logs) and autonomously recomputed optimal burden (reduced from 4.2 m to 3.6 m) and delay timing to suppress wave superposition. Post-blast image analysis confirmed 92% compliance with target fragment size (x₅₀ < 0.8 m), versus 74% in prior conventional designs—reducing secondary crushing energy by 18% (Boddington Annual Blasting Report, 2023).

📋 Case Connection

📋 Liquefied Natural Gas (LNG) Train Optimization

Excessive compressor power consumption and suboptimal refrigerant blend performance

📋 Ammonia Synthesis Loop Purge Optimization

Excessive purge gas loss (3.2% of fresh feed) causing hydrogen waste and emissions

📚 References