🎓 Lesson 16
D5
Battery Thermal Modeling for Extreme Climates
Battery thermal modeling is predicting how hot or cold a battery gets in extreme weather—like Arctic winters or desert heat—so it keeps working safely and efficiently in autonomous mining trucks.
🎯 Learning Objectives
- ✓ Calculate battery pack temperature rise under defined discharge duty cycles using lumped-parameter thermal models
- ✓ Design an air-cooled thermal management system for a 120 kWh LiNiMnCoO₂ (NMC) battery pack operating at −40 °C ambient
- ✓ Analyze thermal gradient-induced capacity loss across a battery module using empirical Arrhenius-based degradation models
- ✓ Explain the impact of low-temperature electrolyte viscosity on charge acceptance and internal resistance during cold-start operations
- ✓ Apply ISO 18434-1 infrared thermography standards to validate thermal model outputs against field-measured surface temperatures
📖 Why This Matters
In remote mining sites—from northern Canada’s permafrost zones to Australia’s Pilbara desert—autonomous haul trucks operate 24/7 in ambient temperatures ranging from −45 °C to +55 °C. Lithium-ion batteries, the heart of electric drive systems, suffer severe performance loss below −20 °C (reduced charge acceptance, increased impedance) and accelerated aging above 40 °C. A single thermal runaway event can halt an entire fleet; inaccurate thermal predictions cost millions in unplanned downtime and battery replacement. This lesson bridges theory to practice: you’ll learn to model, validate, and harden battery systems for real-world extremes—not just lab conditions.
📘 Core Principles
Thermal behavior of battery packs arises from three interdependent domains: (1) Electrochemical heat generation—Joule heating (I²R), reversible entropy heat (T·dE/dT), and irreversible side reactions—dominates during high-power discharge/charge. (2) Heat transfer physics—conduction through cell laminates and busbars, convection via forced-air or liquid cooling, and radiation (minor but non-negligible in high-temperature desert environments). (3) Environmental coupling—ambient air temperature, solar irradiance (up to 1,100 W/m² in open-pit mines), wind speed, and vehicle motion all influence convective coefficients and thermal boundary conditions. Advanced models use 3D CFD for cell-level fidelity or validated 0D/1D lumped-parameter networks for real-time control integration—both require accurate material property inputs (e.g., anode/cathode thermal conductivity: 0.1–0.5 W/m·K) and empirical calibration against climatic test data.
📐 Lumped-Parameter Thermal Model (Single-Cell Equivalent)
This first-order RC thermal model balances heat generation with conduction/convection losses using a simplified equivalent circuit. It’s widely used in BMS firmware for real-time SoH and safety limit estimation. The model assumes uniform internal temperature and neglects axial gradients—valid for low-to-moderate C-rates (<2C) and moderate ambient transients.
Single-Node Thermal Model
C_th × dT/dt = I²R − h·A·(T − T_amb)First-order energy balance for estimating average cell temperature evolution under transient load and ambient conditions.
Variables:
| Symbol | Name | Unit | Description |
|---|---|---|---|
| C_th | Thermal capacitance | J/K | Heat capacity of the cell (mass × specific heat) |
| I | Current | A | Discharge or charge current magnitude |
| R | Internal resistance | Ω | Ohmic resistance at given SoC, temperature, and aging state |
| h | Convection coefficient | W/m²·K | Empirically derived or CFD-calculated heat transfer coefficient |
| A | Effective surface area | m² | Area available for convective/radiative exchange |
| T | Cell temperature | °C or K | Average bulk cell temperature |
| T_amb | Ambient temperature | °C or K | Local air temperature surrounding the battery pack |
Typical Ranges:
Forced-air cooling (5 m/s): 8 – 25 W/m²·K
Liquid cooling (glycol/water mix): 100 – 500 W/m²·K
Desert solar loading (no airflow): −200 to +800 W/m² net radiative flux
💡 Worked Example
Problem: A 100 Ah NMC pouch cell discharges at 150 A (1.5C) for 60 s in −30 °C ambient. Cell internal resistance = 0.9 mΩ at this SoC/T. Convection coefficient h = 12 W/m²·K (forced air, 5 m/s), surface area A = 0.045 m², thermal mass C_th = 180 J/K. Initial cell temp = −30 °C.
1.
Step 1: Compute Joule heating power: Q_gen = I² × R = (150)² × 0.0009 = 20.25 W
2.
Step 2: Compute convective heat loss: Q_loss = h × A × (T_cell − T_amb) = 12 × 0.045 × (T_cell + 30)
3.
Step 3: Apply energy balance: C_th × dT/dt = Q_gen − Q_loss → solve ODE numerically or via time-step discretization (Δt = 1 s): T_{n+1} = T_n + (Δt/C_th) × [Q_gen − h·A·(T_n − T_amb)]
4.
Step 4: Iterate for 60 s: Starting at −30 °C, after 60 s, T ≈ −22.4 °C (ΔT = +7.6 °C)
Answer:
The cell temperature rises by 7.6 °C over 60 s—within safe limits for low-temp operation (no lithium plating expected below 10 °C rise), but highlights need for preheat strategy before high-power maneuvers.
🏗️ Real-World Application
Rio Tinto’s Koodaideri autonomous fleet (Pilbara, WA) deploys 2.5 MWh battery-electric haul trucks operating in 48 °C summer days with 30% relative humidity. Engineers used ANSYS Fluent CFD coupled with COMSOL electrochemical-thermal co-simulation to size a dual-mode TMS: liquid-cooled plates for peak discharge (1.2 MW regen braking) and passive phase-change material (PCM) encapsulation for overnight idling. Field validation showed model-predicted max cell ΔT (vs. ambient) was within ±1.3 °C across 12,000+ duty cycles—enabling reduction of coolant pump oversizing by 35% and extending calendar life by 22% versus baseline air-only design (per 2023 Rio Tinto Technical Review, Section 4.7).
🔧 Interactive Calculator
🔧 Open Autonomous Haulage Systems Integration Calculator📋 Case Connection
📋 Coal Mine Battery Thermal Management at BHP Mt Arthur
Battery pack thermal runaway risk during 12hr shifts; SOC estimation drift >8% after 4 hours