Spatial Discretization Principles for Large Pouch Battery State Estimation
Spatial discretization maps planar potential drops and localized overpotentials in large pouch cells, preventing undetected lithium plating during high-rate charging.
Foil
Large-format lithium-ion pouch cells exceeding 50 ampere-hours introduce severe in-plane electric potential drops across their current collectors. Thin copper and aluminum sheets, typically ten to fifteen micrometers thick, carry total pack currents across lengths exceeding 300 millimeters. At discharge rates above 2C, localized ohmic losses along these metal layers generate uneven current density distributions across the active electrode surface.
A lumped equivalent circuit treats the terminal tabs as direct taps into a homogenous electrochemical volume. That simplification collapses on continuous fast charging, where local potential drops create spatial current variations exceeding forty percent across the sheet area.
Subtab ultrasonic welding configurations fix the boundary conditions for all in-plane potential fields. Pouch cells with tabs on opposing narrow ends force current along the longitudinal axis, creating a predictable linear potential drop. Cells with adjacent tabs on the same end face severe current crowding near the collector base, concentrating charge exchange into the upper quarter of the cell casing.
The electrical resistance of an aluminum positive collector foil at twenty-five degrees Celsius measures approximately 0.028 ohms per square meter, while the negative copper foil exhibits roughly 0.017 ohms per square meter. These values appear negligible until continuous currents of 150 amperes pass through a cross-sectional area of less than five square millimeters at the tab-foil interface.
Where Does Planar Current Density Diverge?
Current density peaks at the collector-tab transition zone. Current crowds where the tab tabulates into the multi-layer foil stack. Active material directly beneath the tab carries five to seven times the current per unit area compared to material at the bottom seam.
This disparity distorts local state of charge across the planar coordinates during aggressive drive cycles or high-rate DC charging pulses.
Single-tab pouch cells operating at 3C discharge exhibit local current density variations exceeding sixfold between the tab root and bottom edge.
When current density diverges across the plane, electrochemical utilization becomes non-uniform. Active material near the tabs hits voltage cutoffs prematurely, while distal material remains underutilized. A single terminal voltage measurement records only the mixed potential of these parallel, non-uniform zones.
Standard cell monitoring algorithms interpret this composite reading as an average state of charge, concealing the fact that local regions operate beyond safe electrochemical boundaries.
Cell vendors frequently justify lumped parameter models by claiming that planar gradients equalize during rest intervals. That argument ignores the irreversible degradation occurring during continuous current throughput before equilibrium is reached.
Node
Mathematical representation of large spatial dimensions demands discrete meshing of the electrode area into coupled electrical and electrochemical units. Discretizing a 500-millimeter by 120-millimeter cell sheet into a network of interconnected finite volumes resolves the lateral potential drop on both current collectors. Each discrete segment contains a one-dimensional electrochemical model representing solid-state lithium diffusion and charge transfer through the separator.
The resulting formulation couples planar two-dimensional ohmic conduction with through-thickness one-dimensional mass transport, establishing a 2D plus 1D model architecture.
Selecting grid density involves balancing spatial resolution against onboard processor execution limits. A coarse mesh of four planar nodes captures basic longitudinal gradients but misses localized current crowding at the tabs. A dense mesh of one hundred nodes predicts current distribution accurately yet exceeds the computing throughput of automotive microcontroller units.
Evaluating state estimation accuracy across discrete grid densities identifies practical thresholds for real-time safety files.
| Mesh Configuration | Total State Variables | Execution Time per Step | Peak Current Density Error | Local SOC Estimation Error |
|---|---|---|---|---|
| 1D Lumped Single-Node | 14 | 0.12 ms | 48.2 percent | 11.4 percent |
| 2×2 Planar Grid (4 Nodes) | 56 | 0.68 ms | 22.1 percent | 5.8 percent |
| 4×4 Planar Grid (16 Nodes) | 224 | 3.15 ms | 6.4 percent | 1.9 percent |
| 8×8 Planar Grid (64 Nodes) | 896 | 14.80 ms | 1.8 percent | 0.6 percent |
| 12×12 Planar Grid (144 Nodes) | 2016 | 42.60 ms | 0.5 percent | 0.2 percent |
Memory footprints expand linearly with node counts. Microcontroller units running safety-critical state estimators allocate limited random-access memory to battery management tasks. A sixteen-node model strikes a documented balance, containing peak current errors below seven percent while executing well within the typical ten-millisecond task scheduling window.
Underestimating the required nodal density produces blind spots in state of charge tracking. The controller allows local overcharging near the current tabs, triggering premature cell isolation during warranty operation and inflating warranty reserve costs across the pack deployment.
Gradient
Spatial non-uniformity across large cell formats involves linked electrical, chemical, and thermal fields. Local current crowding generates non-uniform Joule heating across the foil surfaces. In addition, the entropic heat of reaction varies with local state of charge, meaning that different regions of the cell liberate heat at different rates.
Because heat rejection occurs primarily through conduction to cooling plates positioned either against the pouch faces or clamped to the terminal tabs, steep temperature differentials develop across the active area.
Temperature differences directly govern local reaction kinetics and ionic conductivity. The Arrhenius relationship controls the exchange current density and solid-state diffusion coefficients. Warmer zones near the tabs display lower charge-transfer resistance, drawing a higher proportion of the total current.
This dynamic creates a self-reinforcing feedback loop. Increased local current produces elevated ohmic and reaction heating, which further lowers local impedance and attracts even more current.
- Tab Region Overheating accelerates electrolyte breakdown, elevates solid electrolyte interphase growth rates, and reduces local state of health faster than the cell average.
- Distal Edge Underheating suppresses liquid-phase ionic diffusivity, heightens concentration polarization, and increases the local overpotential during cold charging regimes.
- In Plane State Of Charge Drift forces localized active material to cycle across wider operational windows than commanded by the pack supervisory controller.
Surface cooling configurations produce in-plane thermal gradients exceeding twelve degrees Celsius during 2C continuous fast charging.
Will Lumped Boundary Resistance Mask Plating?
Lithium deposition on graphite occurs when the local potential of the solid phase drops below the potential of the electrolyte. In cold conditions or at high charge rates, current crowding drives the local overpotential negative in the active material immediately adjacent to the negative tab. A lumped cell model calculates an average anode potential based on total cell current and bulk temperature, indicating a positive safety margin.
Localized electrochemistry under the tab operates simultaneously in active plating conditions.
Thermal gradients cannot be decoupled from spatial state of charge estimation. Liquid-phase diffusion limitations at the coldest boundary restrict lithium intercalation rates. When an estimation algorithm lacks thermal-electrochemical spatial coupling, it fails to predict the onset of localized metallic lithium deposition during fast charging protocols.
Thermal fields and current pathways settle at their own distinct rates during high-power operation.
Observer
Real-time tracking of distributed internal variables requires state estimation architectures that map boundary measurements to interior nodal states. Pouch cells present minimal external sensing interfaces: two terminal posts, one overall current sensor, and surface thermocouples. Internal states like local overpotentials, regional state of charge, and concentration profiles remain unmeasured directly.
The observer synthesizes boundary current, terminal voltage, and surface temperatures through discretized state equations to correct internal nodal trajectories.
Kalman filtering formulations adapt naturally to discretized systems. The extended Kalman filter linearizes nonlinear electrochemical equations around the current operating point, updating the covariance matrix of state estimation errors at each time increment. In a sixteen-node spatial model, the state vector contains nodal lithium concentrations, solid-phase potentials, and local temperatures.
Matrix multiplication costs scale quadratically with the state vector dimension, demanding algorithmic reductions for production embedded microcontrollers.
- Reduced Order Modeling projects high-dimensional state vectors onto low-order orthogonal subspaces using proper orthogonal decomposition, cutting arithmetic operations by eighty percent without sacrificing boundary fidelity.
- Dual Extended Kalman Filters decouple fast electrochemical dynamics from slow thermal and parameter updates, running the spatial voltage observer at ten hertz and the spatial temperature estimator at one hertz.
- Observability Gramian Verification ensures that surface voltage and temperature readings provide mathematical observability of interior states across the full operating state of charge window.
State observability diminishes significantly when open-circuit voltage curves exhibit flat plateaus, as seen in lithium iron phosphate chemistry. In these plateau zones, terminal voltage shifts negligibly despite substantial movements in local state of charge. The estimator relies heavily on current integration, accumulating spatial drift unless persistent dynamic excitation perturbs the cell.
Whether boundary surface temperature arrays can uniquely reconstruct interior through-thickness lithium concentration gradients during prolonged dynamic discharge regimes remains an open technical dispute among battery estimation teams.
Clause
Procurement dossiers and supply contracts for large pouch cells enforce strict boundaries on allowable spatial gradients. Factory acceptance tests under standard specifications evaluate cell capacity and direct-current internal resistance as lumped metrics. Those lumped tests mask cell-to-cell variations in foil coating uniformity and tab alignment that dictate real-world spatial behavior.
A sourcing practice protects its commercial position by establishing spatial discretization validation tests within the technical agreement annex.
Regulatory certification under UN 38.3 and IEC 62619 demands rigorous thermal abuse and overcharge compliance. Test T.5 of the UN Manual of Tests and Criteria imposes an external short circuit at fifty-five degrees Celsius. In large-format pouches, the initial current surge exceeds 1,200 amperes.
Internal potential and current distribution models prove compliance with safety margins before destructive physical testing begins, preventing expensive redesign cycles on packed shipping crates.
| Parameter | Test Condition | Contractual Threshold | Verification Standard |
|---|---|---|---|
| In-Plane Temperature Variance | 3C Continuous Discharge, 25°C Ambient | Maximum 4.5°C Spread | Multi-Point Thermocouple Array |
| Collector Tab Potential Loss | 150A Pulse for 10s at 50% SOC | Maximum 18 mV Drop | Four-Point Tab Kelvin Sensing |
| Localized Degradation Spread | 1000 Cycles, 1C/1C, Tab Cooled | Maximum 3.0% Capacity Delta | Post-Mortem Sectioning Protocol |
| Fast Charge Plating Margin | 4C Pulse to 80% SOC at 10°C | Local Anode Potential Above 5 mV | Three-Electrode Pilot Batch Audit |
Technical annexes mandate that cell suppliers deliver validated multi-node state estimation parameters alongside physical cell lots. The buyer embeds clauses establishing that inaccurate spatial model parameters invalidating pack thermal safety margins constitute a latent non-conformity. Responsibility for safety-critical functional failures under ISO 26262 traces directly to defective cell parameter dossiers, reallocating warranty and recall liabilities to the component manufacturer.
Technical supply contracts that define battery quality exclusively through lumped capacity metrics leave warranty liability on the pack integrator when localized degradation accelerates.
Incoming inspection procedures verify parameter accuracy through pulse-power testing combined with high-resolution infrared thermal tracking. Cells exhibiting anomalous planar temperature distributions fail incoming quality gates and face immediate rejection at the container dock.