battery-design · battery sizing studio

Cell

Configuration

Series (S)

Parallel (P)

Arrangement

Orientation

Cells per row (0 = auto)

Layers

Spacing & enclosure

Cell spacing / spacer Wall thickness Busbar / BMS headroom

Why these spaces exist — the standards behind them

Cell spacing is not padding: prismatic and pouch cells SWELL 10–20% in thickness over life (the design allowance under constraint — unconstrained cells measure far worse), and the gap doubles as the thermal break that propagation tests (IEC 62619 §8, GB 38031) probe. A spacer from the Components tab sets this number from a real part.

Wall thickness is what the crash and crush tests demand: no standard prescribes a millimetre value — UN ECE R100 Annex 4, GB 38031 and UL 2580 prescribe TESTS (mechanical shock, crush, integrity) the wall must pass for your market. The slider reserves the space; the proof is the test.

Busbar/BMS headroom keeps room above the cells for interconnects, sense wiring and the gas path to the vent — the sources give no vent-area sizing method (marked honestly, not guessed), but a blocked vent path fails every abuse test.

Application preset

Environment — climate & season

Class-typical ambient bands — the design should survive ALL seasons (All year), a season button just shows what that season does to the system. The system temperature is the ambient plus the pack's own heat rise at continuous load.

Requirements

Voltage min (V)

Voltage max (V)

Energy (Wh)

Continuous power (W)

Peak power (W)

Advanced settings — duty cycle, DoD, limits, stacks

Cycles / year

Target years

Usable DoD (%)

Charge rate (C)

Duty cycle and depth of discharge drive the cell-cost comparison and the CO₂ payback — the defaults are sensible, tune them when you know your duty.

Max mass (kg) Max dimensions L×W×H (mm, blank = open)

Temp min (°C)

Temp max (°C)

The climate & season picker above fills these — type over them any time.

Stacks / racks in parallel (blank = auto from the energy target)

For BESS containers, ships, buses and other big systems: the tool designs ONE stack and multiplies — set the count here if your plant layout already fixes it.

Suggestions

Pick a preset or fill the form, then hit “Suggest designs”.

What should the battery do?

Pick an application on the Usage tab first. Sizing will show only the relevant choices.

🧮 Size from the duty Pro

The route, the schedule, the charge windows and the life go in; the nameplate comes out with THE CONSTRAINT THAT DECIDED IT named — every constraint at end of life, winter as the design day. Then one click puts the answer on the bench: the sized pack runs through the simulated service day, and the two engines agree or say exactly why not.

Sizing basis

Choose an application and operating goal.

Your recommendation

Complete the sizing choices to see one clear recommendation.

Engineering details

The selected goal is converted internally into battery power over time. Engineers and software integrations can inspect or replace that trace here.

Measured data or another profile

Use this only when you have measured data or need to compare a non-default duty.

Profile library

Two columns: seconds, watts — positive = battery discharge, negative = battery charging. Any logger or Excel export works. Download a sample CSV.

Available space

L (x)

W (y)

H (z)

A Usage-tab preset pre-fills the typical bay for that application. Walls, spacer gap, busbar headroom and the selected cooling system's space are subtracted before packing.

Integration allowance 35%

Plan-area reserve for module walls, crash structure, manifolds and wiring that the packer doesn't model. Calibrated against the Tesla Model 3 LR pack, where the OEM realizes ~64% of the geometric ideal. Set 0 for bare-maximum geometry.

Cell cost basis

Estimates cell purchases over the target years and cost per nominal cycle-throughput kWh, using your Usage settings (80% DoD by default). Excludes BMS, enclosure, thermal hardware, labor, electricity, losses and calendar aging. Upfront is cells only.

Objective weights

Energy / fill 5

Cost 3

Weight 2

Relative priorities for the multi-objective score. Pareto-optimal designs (no other candidate is better on every objective) are marked regardless of weights.

Results

No search yet.

Market-representative component classes with example suppliers — properties are class-typical, not one SKU. Pickers filter to the current cell shape.

Mechanical

—

Thermal

—

Electrical

—

Architecture — modules, BMS & HV chain

How this pack is built up and switched: cells → modules (or cell-to-pack) → pack → parallel racks for big systems, plus the BMS topology, precharge, contactors, fuse, DC-DC and isolation the customer needs around the cells.

System — pack, HV chain & supervisory layer (tap to enlarge)

Inside the BMS — layer 2 (tap to enlarge)

BMS topology

Isolation bus context (UN R100 vehicle topology — no averaging)

UN R100 Rev. 3 §§5.1.3.1–5.1.3.2. The connected-bus 100 Ω/V exception requires documented protection on every AC bus and is not silently selected. Marine IT systems need their declared earthing topology, class rules and flag-state basis instead.

Assumptions (AFE channels, DC link, precharge duty, sensors)

AFE channels/IC

DC link (µF)

Precharge (s)

Precharges/hour

Cells/temp sensor

Module series (S_mod)

Set S_mod when the mechanics fix the module first — a 30S module at 16-channel AFEs correctly gets 2 slave ICs.

AFE ICs run 14–25 channels; the DC-link capacitance belongs to the load's inverter — replace the placeholder with its datasheet value. The precharge resistor is sized for the stated repetition rate.

—

HV startup, current shunt & fast protection

Sensata-based calculators for the parts around the DC link. The curves show what the circuit does; the plain-language verdict says what is qualified, missing or unsafe. Archived parts are reference models only.

Precharge resistor and contactor

—

Calculator inputs and supplier evidence

Selected resistor Ω

Tolerance ±%

Downstream load A

Voltage rating V

Pulse energy J

Pulse power W

Continuous power W

Design margin %

Precharge contactor

Make-current rating A

Switching life cycles

Evidence date

Supplier part(s)

Document revision(s)

Current-shunt selection

—

Duty, measurement and thermal model

Reference / supplier model

Continuous duty A

Peak duty A

Peak duration s

Resistance µΩ

Continuous rating A

Peak rating A

Rated peak duration s

Conductor area mm²

Max operating °C

Gain error ±%

Offset error mA

Noise error mA RMS

Thermal R K/W

Thermal τ s

Required accuracy %

Supplier part

Evidence revision

Evidence date

Fast-fault coordination

Run the short-circuit study first.

Detection and interrupter evidence

OCD threshold A

Total delay ms

Sensor peak range A

Interrupter voltage V

Interrupt current A

Evidence date

Interrupter part

Document revision

If it shorts

Expert detail

Busbar cross-section (mm²)

Fuse rating (A)

Per-cell fusible link (A) — blank = none fitted

Charging

Expert detail — charger classes, connectors, strategy

🔌 Charge session, simulated Pro

The charger connected to YOUR pack: every second is a negotiation between charger power, connector current, the pack's own acceptance and the voltage on the cell's actual curve. CC and CV fall out of the physics, the winter floor behaves exactly as the mission does, and the session names WHAT DECIDED THE CHARGE TIME.

Charger (kW AC) From (%) To (%) Ambient (°C)

⚡ Power electronics Pro

The converters around the pack — inverter, on-board charger, the 12 V DC-DC — as REAL efficiency curves: poor at light load, peaking mid-load, not one flat number. Draw the curves, then run your day through them: what each converter cost, the heat it hands the cooling system, and the ripple the pack absorbs. Switch-level PWM stays refused by name — export the FMU for device studies.

Inverter (kW) Aux DC-DC (W) Aux draw (W) Bus (V)

🧭 BMS mode charts Pro

The battery fraction of Stateflow, pre-assembled: modes, guarded transitions, latching faults. The chart is data — the diagram below is generated from it — and every transition explains itself with the signal values that fired it. Run it against the mission above, then take it to your controller as loop-free C, held to this interpreter by an open conformance test.

Supervisor

🔥 Runaway investigation lab Pro

One cell goes — watch the event the way an investigator watches a test: the cell map through time, the propagation order, which heat path carried the energy. Battery questions only, no fire-CFD authoring; and it can never say "safe" — only what the model observed.

Safety & standards

Standards referenced

The SYSTEM behind the cold plate: the loop (pump, radiator, refrigerant chiller / heat exchanger, valves, heater), the BTMS control unit that runs it, and what it costs the higher system. Derived from your selected cooling component, the pack's heat at continuous load and the climate window.

Thermal management system

Loop & control (tap to enlarge)

—

🌡️ The machines behind the loop Pro

The loop above budgets with flat figures; these are the curves those figures hide. Chiller COP against the ACTUAL lift — hot days genuinely cost more. Heat pump against PTC, with the advantage and its cold-weather cutoff stated. Pump and fan on the affinity cube law. Then the day ledger: the mission's own heat, priced through those curves — pump, fan, compressor, mode by mode — and winter preconditioning priced through either heater. No CFD, no refrigerant-cycle simulation — export the FMU for those.

Coolant target (°C) Radiator UA (W/K)

The design, run through TIME: the application's load profile drives an equivalent-circuit model of YOUR pack (OCV−I·R, coulomb counting) coupled to a lumped thermal model (I²R in, cooling out). Watch SoC, voltage sag and temperature over the mission — and see exactly when a pack runs empty, hits the voltage cutoff, overheats or refuses winter charge.

Test protocol bench

A cycler for the simulated pack: charge, discharge, rest, pulse trains, drive-cycle replay and repeats, written in a small protocol language or clicked together — both edit the same protocol, and the run names why every block ended.

Mission simulation

Profile passes Start SoC %

Charging in the mission

Minutes

Top-ups: short charges after every pass (opportunity charging). At base: one charge at the end (depot / overnight). Power comes from the cell's own charge ratings.

Mission traces (tap to enlarge)

—

What this model is (and is not)

⚙️ C code for your BMS Pro

This exact model — OCV, sag, cutoff hold, CV taper, winter charge inhibit, coulomb counting, thermal — generated as portable C99 with a fixed-step, reentrant ABI (init/step/reset/terminate), static memory, every loop bounded. The generated C is held to this simulation step by step by an open conformance test — the proof ships with the tool, not on a vendor's name. The header carries the same honesty as this tab: class-estimate OCV and published DCIR, to be replaced with your measured data.

📈 Fit the model from your test data Pro

Upload a cycler trace and the class estimates above become YOUR cell's measured curve: OCV from the rest points, resistance by least squares under load. The fit is deterministic and judges itself in millivolts; where the data never went, it says "interpolation", not numbers. Adopted, the measurement rides into the simulation and the generated C with the dataset's digest attached. Nothing leaves your browser.

🔗 Co-design: the system loop Pro

The garage's machine, YOUR pack, the charger and the BMS supervisor as one system: the pack's own mass rides in the road load, the machine's day drives the pack, the end-of-day drives the charger's recovery, and the supervisor watches it all. Change one part and see what it did to the others — and what it did NOT touch.

Machine Passes/day Charger (kW)

Every sensor THIS design needs, at its level — cell, module, system, cooling loop. What the system does not need is not shown at all: a wearable lists no coolant sensors, a drone no runaway detector. Counts feed the connector pin-out and harness budget.

—

Release checklist — standards per application & market

—

EU 2023/1542 — what applies to THIS design

Battery category

Assessment date

EV and LMT can be derived from a governed application. Other applications need an explicit Regulation category; energy alone never invents “industrial.”

—

Regulation (EU) 2023/1542 — timeline

Your recommendation

Complete the sizing choices to see one clear recommendation.

🔍 Where every number comes from

The effective-value ledger: every load-bearing value in this design, with its source named — measured, document-backed, class estimate, configuration, or derived with the formula shown. A value's origin decides whether a review can believe it.

Cell comparison

Tick two or more cells under Design → Browse and compare cells to see how their compromises differ.

CO₂ — what the delivered energy displaces

Manufacturing footprints are chemistry-class estimates; displacement assumes delivered energy replaces this source one-for-one. Duty (cycles/year, target years) comes from the Usage tab.

The visual report is a self-contained, sub-minute animated HTML file using this design's exact sizing, simulation, selected thermal hardware, semantic trace and versioned 3D asset. The architecture report is a layered HTML page — select and open exactly the level you care about (system → pack → module → cell, control, thermal, sensors). The workbook (Excel) computes with live formulas over named cells, so you can check every number against your own spreadsheet — edit it and send it back to teach the tool.

Copies the whole design as JSON plus a short written brief, so you can paste it into any chatbot and ask about it. To let an AI drive the designer directly — size a pack, run a mission, compare cells — run the included MCP server on your machine (see desktop/README.md).

View

Color

Show enclosure series path

Lemonergy account

Add your cell — privately