Thermally Managed Battery Enclosures: A Design Guide

What Makes an Enclosure Thermally Managed

Start with the heat source. Each cell dissipates roughly I²R watts, where I is the per-cell current (pack current divided by the parallel count P) and R is the cell's DC internal resistance at the operating temperature and state of charge. A 13S4P pack of 18650 cells at 30 mΩ each, pulling 40 A, puts 10 A through every cell: 3 W per cell, about 156 W for the whole pack. That heat then crosses a chain of thermal resistances on its way out — cell can to air gap or pad, pad to enclosure wall, wall to outside air — and each leg is measured in °C per watt. Multiply the total °C/W by the watts and you have the steady-state temperature rise above ambient. If 156 W meets a 0.25 °C/W path, the pack settles 39 °C above ambient; on a 30 °C day that is 69 °C, far past the 45 °C longevity ceiling. Designing a thermally managed enclosure means shrinking that °C/W number until the arithmetic lands somewhere safe, rather than hoping the case is big enough.

The Danger of Heat Soak

Lithium cells generate heat as they discharge. In a tightly sealed plastic case with no ventilation, this heat gets trapped, causing the internal temperature to rise. As temperature increases, the internal resistance of the cells also increases, creating even MORE heat. For high-power packs, you must include thermal pads or leave air gaps between cells to prevent this cycle.

Sizing Passive Cooling: Surface Area, Air Gaps, and Pads

Passive cooling relies on two steps: conduction from the cells into the enclosure wall, then natural convection and radiation off the outside of that wall. Free convection to still air moves roughly 5–10 W per m² per °C; add radiation from a matte, non-reflective surface and 10–15 W/m²·°C is a reasonable planning figure. A 300 × 150 × 100 mm case has about 0.18 m² of outer surface, so at 12 W/m²·°C it sheds around 2.2 W for every °C of rise — about 0.46 °C/W to ambient, before you even account for getting the heat out of the cells. That budget is exactly why the 156 W pack above cannot be cooled passively in that box, while the same box carrying a 20 W continuous load sits about 9 °C above ambient and needs nothing further. Inside the case, the enemy is trapped air: a 2–3 mm gap between cells does very little on its own because that air is stagnant and stagnant air is an insulator at about 0.026 W/m·K. Bridge the gap instead. Thermal pads or potting compounds in the 1–3 W/m·K range tie the cells to the wall and make the enclosure part of the heat path rather than a blanket around it.

Enclosure Materials: Plastic vs. Aluminum

Plastic (ABS/Polycarbonate) is cheap, lightweight, and non-conductive, which is great for safety. However, it's an insulator that traps heat. Aluminum cases are much stronger and act as a giant heatsink, pulling heat away from the cells. If you use aluminum, you must be extremely careful with insulation to prevent the battery from shorting out against the metal casing.

When Passive Cooling Is Not Enough: Forced Air and Cold Plates

Two signals say you have outgrown passive cooling and need active cooling: continuous dissipation above roughly 10 W per litre of enclosure volume, or a passive estimate that lands within about 15 °C of the cells' 45 °C ceiling on a hot day. Forced air is the cheapest form of active cooling, and the airflow you need follows from a simple balance — CFM ≈ 1.8 × watts ÷ the air temperature rise you will accept. Shedding 156 W with a 10 °C rise across the pack therefore needs roughly 28 CFM, which is a 60–80 mm fan rather than the 40 mm fan most builders reach for. The catch is that ducting outside air through the pack destroys your ingress rating and drags in dust and moisture. The cleaner compromise for an ebike or scooter is to keep the enclosure sealed and blow air over the outside of a finned aluminium wall, so the fan never touches cell space. Above roughly 500 W of continuous dissipation, or wherever sealing is non-negotiable, a liquid cold plate bonded to one face of the pack is the standard answer: a plate with a water-glycol loop reaches 0.02–0.05 °C/W, an order of magnitude better than any passive wall of the same size.

Cell Spacing and Thermal Runaway Propagation

Steady-state cooling and propagation resistance pull in opposite directions, and the tension is worth being explicit about. A cell entering thermal runaway vents gas and particulate at several hundred °C; whether the pack loses one cell or all of them depends on how much of that energy reaches the neighbours in the first seconds. Cans in direct metal-to-metal contact — the shrink-wrapped brick — are the worst case, because conduction between touching cans is the fastest propagation path there is. Holding cylindrical cells about 2 mm apart in plastic holders and adding a 0.3–0.5 mm mica or aerogel barrier between rows raises the resistance sharply; aerogel sits near 0.02 W/m·K, so it blocks the runaway path while barely affecting a design that already dumps its steady-state heat sideways into the wall. Note the trap in potting: a compound conductive enough to help normal cooling also carries runaway heat straight into adjacent cells, so prefer holders plus pads on the wall side over burying the whole pack in one conductive block. The interconnect matters too — a thin fusible nickel tab per cell both limits fault current and removes a metal bridge that would otherwise conduct heat between cells.

Pressure Relief and Vent Gas Routing

A case sealed well enough to keep water out is also sealed well enough to keep vent gas in, and a single 18650 in runaway releases several litres of hot gas. In a half-litre sealed enclosure that pressure has to go somewhere, and an enclosure that ruptures at a random seam sprays hot electrolyte in an unpredictable direction. Give it a deliberate weak point instead: a burst disc, a lightly bonded panel, or a labyrinth vent placed at the end of the case furthest from the BMS and pointed away from the rider's legs and any fuel or wiring. This is a different mechanism from the GORE-tex breather described below — a breather equalises the slow pressure changes of daily temperature swings and cannot pass runaway volumes. A well-built pack has both.

Waterproofing and Condensation

An ebike battery will eventually be caught in the rain. While you want to seal the case against splashes, a 100% airtight seal can actually be bad. Changes in temperature can cause moisture to condense inside the case, leading to corrosion on the BMS and cell terminals. Use GORE-tex vents or breathable membranes that allow air to pass but block liquid water.

Vibration and Impact Protection

Batteries in vehicles are subjected to constant vibration and occasional impacts. Use high-density foam (like neoprene) to 'bed' the battery inside its case. This prevents the heavy pack from bouncing around and snapping the delicate BMS wires or spot welds. Never use standard Styrofoam, as it can generate static electricity and crumbles easily over time.

Validating the Design: Sensors, Test Runs, and Cutoffs

Model first, then measure, because the assumption that usually breaks is an air gap you believed was conducting. Place at least two NTC thermistors against the cells the model says run hottest — normally the geometric centre of the pack rather than the row touching the wall — and one on the ambient side for reference. The BMS temperature channel alone is not enough: it usually sits on the PCB and reads several degrees below true cell temperature. Then run the worst realistic case (a sustained full-power climb, or a complete CC charge at maximum current) and log until the reading plateaus, which for a pack of this thermal mass typically takes 20–40 minutes. Compare that plateau against your °C/W estimate; a measured rise more than about 30% above the calculation means the heat path is not what you drew. Finally, set the BMS over-temperature cutoffs with margin — commonly around 55–60 °C on discharge and 45 °C on charge — and never charge a lithium pack below 0 °C, where lithium plating causes permanent capacity loss and internal shorts regardless of how good the enclosure is.

FAQ