MY SIMULATOR

Grid Battery Yard

Liquid cooling, thermal derating and solar-wind dispatch.

INTERACTIVE 3D

Time 0:00Day revenue 0 EURFPS 0

renewables / charging discharge / grid hot coolant

STATE OF CHARGE30 %

BATTERY POWER0.0 MW

HOTTEST RACK25.0°C

GRID EXCHANGE0.0 MW

TIME☀☾

Grid Battery Yard: Liquid Cooling, Thermal Derating & Solar-Wind Dispatch

A containerised lithium-iron-phosphate battery soaks up midday solar and wind, then feeds the evening peak. Every cell makes heat, a liquid loop carries it to a chiller, and the battery management system backs the power off if it gets too hot.

Time speed

Scenario

Dispatch strategy

Yard and weather

Live numbers

Power flows, MW (+ to bus: PV, wind, battery discharge)

State of charge, %

Temperatures, °C

Price, EUR/MWh

Missions

Independent numerical checks

Also run outside the browser: node src/check.js (Thevenin step vs analytic and RK4, Coulomb counting, terminal-power solve, lumped thermal vs RK4, steady-state coolant balance, Arrhenius ratio, 24 h energy ledger).

Educational model, not an operating guide and not affiliated with any manufacturer. One aggregated container (4160 cells) stands for the whole yard; cell data, the OCV table, thermal constants, prices and load are illustrative [assumption] and must be verified against a source before any real use. No cell failure or thermal runaway is modelled.

Controls

Drag / one finger: orbit; wheel / pinch: zoom; Hero, Racks and Map buttons (keys H, R, M) jump the camera.
The TIME slider scrubs the whole day exactly (the day is precomputed); Start day plays it forward. Space: play/pause. Home: reset.
Sliders and the strategy re-run the day at once. Green streams charge the battery or carry renewables, blue streams are discharge or grid flow, orange pipes carry hot coolant.
Rack lamps: filled lamps show state of charge (blue low, green high); a rack turns orange as it heats.

The science

Energy vs power. A container of N = N_sN_p = 4160 cells of 280 Ah at 3.2 V stores about E = N·Ah·V = 3.7 MWh. The C-rate C = P/E says how fast it can be emptied; here the rating is 0.5 C, about 1.9 MW per container.

Cell model (Thevenin, one RC pair). V = OCV(SoC) − i R₀ − V₁ and dV₁/dt = −V₁/(R₁C₁) + i/C₁, with the exact exponential update over each 10 s step. The power command is turned into a current by solving R₀i² − (OCV − V₁) i + p = 0 per cell. State of charge follows Coulomb counting dSoC/dt = −i/(3600 Ah). The OCV table is a flat-plateau LFP curve [verify against source].

Heat. Q = i²R₀ + V₁²/R₁. R₀ rises when cold: R₀(T) = R_0,25 exp[E_a/R (1/T − 1/298 K)] with E_a about 20 kJ/mol [assumption]. The reversible (entropic) heat is ignored.

Seven racks and the coolant manifold. Each rack is a lumped node C_th dT_k/dt = Q_k − UA (T_k − T_c,k) − UA_amb(T_k − T_amb). Coolant flows through the racks in series, warming by Q_x,k/(ᴁc_p) in each; racks further downstream see warmer coolant, and older racks (higher resistance) make more heat. The chiller returns coolant at the set temperature until its capacity is reached; it draws Q_removed/COP from the bus, with COP falling as the day warms [assumption].

BMS derating. Allowed power falls linearly from 100 % at 35 °C to 0 at 50 °C, charging is limited below 10 °C, and the state-of-charge window is 5 to 95 %.

PCS and efficiency. The inverter efficiency is η = 0.985 − 0.015 (1 − p)² [assumption]. The AC-side ledger is E_in − E_out = ΔE_chem + E_{cell heat} + E_PCS; the cooling and control draw is a separate load on the bus.

Dispatch. Peak shaving discharges only when grid import exceeds the cap (with a morning reserve) and absorbs all surplus; arbitrage charges below the lower third of the price curve and discharges above the 70th percentile; frequency response follows a filtered regulation signal around 50 % SoC; manual is a fixed share of the rated power. Generation, load and price curves are made up for teaching.

Not modelled: electrochemistry, ageing, cell-to-cell imbalance, DC-link dynamics, fire safety systems, real grid codes.