Battery discharge analysis

The restored tool-battery-suite is an explicit scenario model, not the former generic chemistry/aging-curve suite. It supports primary and rechargeable battery references, but models discharge only. It does not predict cycle life or provide a charging prescription.

Inputs and equation contract

• One consistent cell or battery-unit basis: reference Ah, supplied relaxed OCV versus SOC, effective DC resistance, cutoff, series/parallel counts and optional discharge current limit. Identical units are assumed. A 12 V monobloc can be one modeled unit if all its parameters are block-level.
• Sequential repeating segments of constant terminal current, output power or terminal resistance. Segment durations are seconds; optional battery temperature overrides the default for that segment.
• Pack OCV is unit OCV × series; pack R is unit R × series/parallel. For a current load, V = Voc − IR. For a resistor, I = Voc/(Rpack + Rload). For a constant-power output, solve Pout/η = I(Voc − IR) using the low-current quadratic root. Demand beyond the model’s maximum power or supplied current limit ends service.
• Converter efficiency applies only to output-power segments. Current and resistance segments are already battery-terminal loads.
• Supplied SOH scales capacity once. Supplied daily self-discharge is modeled as exponential charge loss; neither value is inferred from a chemistry label.

Temperature and retained charge

Reference charge Q remains a conserved state across temperature changes. With capacity factor f in (0,1], the lower inaccessible charge window is Qnom(1−f), and accessible SOC = [Q − Qnom(1−f)]/(Qnom f). Resistance is scaled by the entered temperature R factor. Temperature is prescribed, not calculated from heating. The reference capacity must bound the temperatures being modeled. No curve extrapolation occurs.

Cold-limited service can end with reference charge retained. That is reported as a temperature-capacity limit, not chemical exhaustion. The simulation stops at the first service failure; it does not pretend the load continued until a later warm interval. OCV-temperature shifts, recovery, nonlinear polarization and Peukert/rate-dependent capacity are not represented by this screening model.

Catalog and reference evidence

The manufacturer catalog contains 57 representative identities across 17 brand labels. It includes Eveready India and Energizer’s Eveready as distinct identities, Duracell, Nippo, Panasonic, Energizer, GP, VARTA, Maxell, Renata, Saft, Murata, Molicel, Lithium Werks/A123, GS Yuasa, CSB and Trojan. Ratings carry available test conditions; unknown capacity/voltage/test conditions remain null. It is not a complete global catalog or a purchasing recommendation.

See profile evidence for primary sources. Catalog entries do not automatically provide OCV, DC R or temperature-model coefficients. Selecting a manufacturer reference clears these fields so the previous illustrative model cannot silently become a product-specific prediction.

Two optional Energizer generic AA alkaline capacity-temperature references are coarse readings from a manufacturer handbook, with explicit test currents, cutoff, temperature span and reading uncertainty. They are not specific E91 calibration. Adoption replaces reference capacity/cutoff and temperature data; a constant R factor remains an explicit assumption unless independently supplied. Published pulse resistance and 1 kHz impedance are not silently copied as DC R.

One loaded-discharge reference provides 14 coarse manufacturer graph readings for generic AA/LR6 at 100 mA, 21 °C and 0.8 V cutoff. Explicit adoption uses the user’s R to estimate Voc ≈ Vloaded + Itest R. This reproduces the reference-current loaded curve within graph-reading/numerical precision; it is not measured relaxed OCV or validation at other currents or temperatures. The source’s approximately 2.4 Ah plotted endpoint and rounded 2.5 Ah prose example remain distinguished.

The initial Li-ion curve, cold-load example and ideal-voltage option are clearly labelled assumptions. None is advertised as a manufacturer fit. JSON exports retain selected source metadata, model provenance, config, limits and result. CSV exports contain the sampled numerical trace.

Numerical and UI checks

node --test tests/battery-analysis.test.cjs tests/battery-analysis-ui.test.cjs

Independent analytic checks cover constant current, resistance and power; pack scaling; voltage cutoff; cyclic loads; current/power infeasibility; temperature/charge conservation; self-discharge; integration refinement; and time/step bounds. In particular, a 1 Ah OCV=3+SOC V model stores 3.5 Wh; a 2 W output at 80% efficiency delivers 2.8 Wh in 5040 s with no internal R.

The UI fixture executes the actual scripts and fragment. It checks model initialization, source filters, clearing unknown product data, source-curve adoption, temperature examples, plot switching and stale-result clearing. Browser checks separately verify actual loading, interactions, mobile layout and the gold-dark/blue-light theme. Passing software checks is not cell validation.

Integration is bounded to 100,000 steps and 1600 plotted/exported points in the UI. Fractional cutoffs are located within the final step; horizon/step-limited runs report lower bounds. Sampled traces may omit some short-pulse detail even though integration extrema/energy are accumulated at integration steps.