Battery cell spacing in sealed enclosures: thermal sizing

You cannot determine battery cell spacing from discharge rate and an air gap alone. First calculate cell losses, the thermal resistance to the enclosure wall, and the wall’s temperature rise above ambient. Then compare candidate layouts and verify the hottest cell under sustained load.

Spacing changes those heat paths, but it is not itself a cooling capacity. In a sealed enclosure, heat must eventually cross the enclosure boundary. A wider gap cannot compensate for an enclosure wall that already runs too hot.

What you need

Gather these inputs before choosing the cell pitch or ordering the enclosure:

• Cell-manufacturer data for capacity, permitted continuous current, temperature limits, mechanical support and vent clearance.
• Resistance or heat-generation data covering temperature, state of charge, aging and the intended operating current.
• Pack arrangement, current-sharing assumptions, carrier geometry and interconnect losses.
• Actual cell dimensions, dimensional tolerances and required electrical insulation.
• Thermal properties and thicknesses of carriers, interface materials and mounting structures.
• Maximum local ambient temperature, mounting orientation and available enclosure surface area.
• A defined temperature margin covering measurement error, modeling uncertainty and operating variation.

Confirm that the battery chemistry and assembly are suitable for the proposed enclosure. Environmental sealing does not establish suitability for battery gas release or fault pressure. The article on gas accumulation in sealed enclosures explains why that needs a separate assessment.

Develop the mechanical arrangement alongside the thermal model. Use circuit board and battery enclosure sizing for component envelopes, access and separation rather than treating the pack as a solid rectangular block.

Build the thermal model

1. Convert the duty into current per cell.

   A 1C discharge means a current numerically equal to the cell capacity expressed in ampere-hours. For an illustrative 2.5 Ah cell, 1C is 2.5 A, not 5 A.

   With four equal parallel branches, a 10 A pack current gives approximately 2.5 A per cell. Six series groups containing four parallel cells make 24 cells. Series count changes pack voltage and total heat, not that individual cell current.

   Equal sharing is an assumption to check. Contact resistance, temperature differences and cell variation can create unequal branch currents.

2. Estimate losses at the relevant operating conditions.

   Start with q_cell = I_cell² × R_cell as a screening estimate. At 2.5 A and an illustrative resistance of 0.040 Ω, this gives 0.25 W per cell.

   Across 24 cells, that is 6 W. Add interconnects, protection electronics, balancing circuits and other equipment operating during the same duty. For example, another 1 W makes the enclosure heat load 7 W.

   Do not treat a single resistance value as constant across the discharge. Battery heating also includes reversible electrochemical effects. Use manufacturer heat data or validated measurements when those effects materially change the result. A short-pulse resistance measurement may not represent sustained heating.

3. Calculate the enclosure wall temperature first.

   For a preliminary steady-state model, use T_wall = T_ambient + Q_total × R_wall-ambient.

   The effective wall-to-ambient resistance includes external natural convection and radiation under the specified mounting conditions. Establish it through a suitable thermal calculation, validated simulation or representative testing. It is not determined by stainless grade alone.

   At 25 °C ambient, a 7 W load and an assumed 2 K/W enclosure resistance give a 39 °C representative wall temperature. That assumption must match the proposed enclosure and installation.

   Outdoor installations need a separate solar-load allowance. Use the method in solar heat load on an outdoor enclosure before calculating the remaining cell-temperature margin.

4. Connect each cell to its actual heat paths.

   A simple screening expression is T_cell,i = T_wall + q_i × R_local,i. Here, local resistance includes the complete modeled path from the chosen cell temperature point to the wall.

   Define that temperature point explicitly: cell case, terminal or estimated internal hotspot. Match it to the manufacturer’s stated limit. A case measurement does not directly establish internal temperature.

   Include carrier conduction, permitted thermal interfaces, contacts and wall spreading. Internal air convection and radiation provide additional paths, but they need suitable geometry-dependent models.

   The article on using a sealed enclosure as a heat sink covers the wall path. The spacing calculation must additionally resolve how each cell reaches that wall.

5. Include thermal coupling between cells.

   Represent adjacent cells with thermal conductances rather than assigning every cell an independent ambient boundary. A steady-state cell-node balance can be written as:

   q_i = G_iw × (T_i − T_w) + Σ[G_ij × (T_i − T_j)]

   Conductance G has units of W/K. The first term represents the cell-to-wall path; the summation represents exchange with neighboring cells. Add separate air, carrier and wall nodes where one equivalent path is inadequate.

   Two equally hot cells exchange no net heat with each other. An interior gap therefore does not automatically become a useful heat outlet. Edge cells can run cooler because they have shorter or stronger paths to the enclosure.

6. Compare layouts against the available temperature margin.

   For the simple common-wall model, the maximum local resistance is R_local,max = (T_target − T_wall) / q_cell.

   If the wall temperature reaches the target, no positive local thermal resistance can satisfy that target. Change the total losses, enclosure dissipation or installation conditions before refining the cell pitch.

   Repeat the calculation for the hottest credible cell, not just the average cell. Include tolerances and uncertainty before accepting a layout.

Compare heat paths before choosing a gap

Continue the illustrative calculation with a 45 °C preliminary cell-temperature ceiling. This is a project assumption, not a universal battery limit.

The available local resistance is (45 − 39) / 0.25 = 24 K/W. Candidate A fails the preliminary ceiling. Candidate B passes the nominal calculation, but its 1 K margin does not establish an acceptable design.

These resistance values describe hypothetical heat paths, not particular cell gaps. A documented interface to a heat spreader might reduce resistance. Simply increasing spacing might not.

For a deliberately simplified, stagnant parallel air layer, use R_air = g / (k_air × A). With a 0.079 in [2 mm] gap, 1.55 in² [0.001 m²] area and air conductivity of approximately 0.026 W/(m·K) at 25 °C, the conductive resistance is about 77 K/W.

That calculation excludes radiation and convection. It also assumes parallel surfaces, unlike the curved spaces between cylindrical cells. It illustrates why a small air gap can remain a weak conduction path; it does not establish a recommended spacing.

Increasing gap width increases resistance in that conduction-only model. Natural convection may improve if the geometry supports circulation, but its behavior depends on orientation and channel geometry. Air circulating inside a sealed enclosure still has to transfer its heat to the wall.

Convert the selected geometry to pitch using center-to-center pitch = actual cell diameter + clear gap. Include cell tolerances, insulating sleeves, carrier tolerances and any manufacturer-required movement allowance.

Know when the simplified model is insufficient

Use a multi-node network or three-dimensional thermal model when wall temperatures vary significantly or the pack has several rows. Explicitly model central cells, end cells and nearby electronics.

Cylindrical cells can have different effective thermal conductivities along and across their axes. End cooling and side cooling therefore cannot automatically use the same internal resistance. Obtain appropriate properties rather than assigning the entire cell the conductivity of its metal case.

Symmetry boundaries are valid only where geometry, heat generation and boundary conditions genuinely mirror each other. A symmetry plane has zero normal heat flow; it is not a cooling surface. A rectangular cell array generally cannot be reduced to an axisymmetric cell surrounded by an arbitrary cold boundary.

For interfaces, estimate material resistance with R_interface = thickness / (conductivity × contact area). Then add contact resistance and spreading effects. Use properties at the intended temperature and compression.

Thermal interfaces must preserve electrical insulation, cell support and manufacturer-required vent clearance. Do not obtain better contact by applying unapproved loads to cell cans or terminals.

Keep nearby power electronics in the model. The guidance on component layout inside a sealed enclosure helps identify heat sources that can warm the pack independently of its own losses.

Verify the hottest cell under sustained load

Validate the proposed geometry in the intended enclosure, orientation and mounting arrangement. Use a controlled, protected test setup with suitable battery protection and independent temperature monitoring. Establish stop limits before testing.

Measure representative central cells, edge cells, the predicted hotspot, internal air, enclosure wall and ambient air. Use electrically insulated sensors with attachment methods suitable for the cells. Confirm that sensor placement does not change insulation, support or vent operation.

Log current, voltage and temperature together. Assess the complete intended duty, including relevant charging, balancing and repeated-cycle conditions. Start from the worst permitted initial temperature rather than assuming every discharge begins cool.

A steady-state calculation is an asymptotic check. A finite discharge may end before equilibrium, while repeated operation may begin with stored heat. Use thermal capacitance in a transient model when those differences matter. Do not extend discharge beyond manufacturer limits merely to seek equilibrium.

Compare measured temperature differences with the model. A larger-than-predicted cell-to-wall difference points toward missing local resistance or underestimated heat. A hotter-than-predicted wall points toward total-load or external-dissipation errors.

Set an acceptance criterion for temperature stabilization and model agreement before testing. A brief, nearly flat temperature trace is not sufficient evidence by itself.

Checks before accepting the layout

Accept the spacing only when the calculated and measured hotspot remains below the applicable manufacturer limit with your documented margin. Confirm the same result across the specified ambient range and operating duty.

Check the following before releasing the enclosure drawing:

• Cell pitch, orientation and tolerances match the validated arrangement.
• Carrier and interface materials match the thermal model and electrical requirements.
• The hottest cell location remains accessible for representative verification.
• Wall attachments provide the modeled contact area without distorting the enclosure.
• Cable bends, protection devices and service access fit without blocking heat paths.
• Gas-release and fault-pressure assessments remain separate from normal-operation thermal acceptance.

Use the electrical enclosure sizing guide for mechanical clearances and wiring space. Thermal clearance alone does not establish a complete panel layout.

For custom enclosures, specify material, width, height, depth, mounting details and every penetration on the drawing. Include the thermal contact locations and any surfaces that must remain unobstructed. The stainless enclosure size finder helps compare available envelopes after those requirements are established.

Mistakes that cost thermal margin or sealing

Do not specify a universal gap based only on C-rate. Capacity determines current, while chemistry, resistance, aging and geometry determine the resulting temperature.

Do not validate a densely packed center cell using only an edge-cell measurement. Likewise, do not treat surrounding cells as room-temperature heat sinks.

Do not add vents, drain holes or through-wall fasteners after thermal validation without reviewing the enclosure design. An environmental rating does not automatically survive an added opening. For our custom NEMA 4X designs, ratings are manufacturer ratings for the specified design; the enclosures are not UL listed.

Finally, do not treat normal-operation temperature acceptance as evidence against thermal-runaway propagation. Cell spacing, barriers and emergency venting require a separate battery safety assessment. If the enclosure cannot reject the expected heat within the temperature budget, re-specify it rather than assigning the air gap a cooling capability it cannot provide.

Enclosures for this job

Catalog sizes that suit the job this article answers, picked from its subject; open one for its full specifications.

Questions