How to Size a Pressure Relief Valve for a Battery Pack

“Sizing” is a slightly misleading word for what this tool does. You are not solving for a dimension. You are selecting among real parts, each with a setpoint and a flow curve fixed by its manufacturer, and checking whether the one you picked keeps the battery pack below its limit.

That distinction shapes the whole procedure.

What Is Fixed, and What You Choose

Everything in the left column is an input to the question, not a lever. Void volume follows from the pack layout. The venting curve is a property of the cell. The number of venting cells is the scenario you are required to survive. Adjusting any of them until the case passes changes the question rather than answering it.

You have exactly two degrees of freedom, and they are both in the right column.

Step 1: Fix the Scenario

Before opening the tool, write down the case you are required to survive.

• Void volume, from the pack layout, with cells and internals subtracted.
• Ambient conditions, including altitude if the pack will operate there.
• The venting curve, and where it came from.
• Number of venting cells, from your propagation testing or from the safety requirement you are demonstrating against.
• Gas temperature, or the decision to treat it as constant and bracket it later.

This set does not change while you screen valves. If it changes, every result you have collected is invalid and the screen starts again.

Step 2: Set the Pass Criterion

Enter your enclosure limit into Max Pressure Limit, relative to ambient.

It comes from structural analysis or a burst test, and it should be the pressure at which the enclosure stops doing its job, not the pressure at which it fails catastrophically. Deformation that breaks a seal or a busbar clearance is already a failure.

This field does not affect the simulation. It only draws the line your results are judged against.

Step 3: Screen Candidates

Now run the loop. For each valve you are considering, run the case at one valve, then at two, and so on, until the peak drops below the limit or you run out of mounting positions.

Here is what a screen looks like. The case is 30 L of void volume, five cells venting, and a 50 kPa enclosure limit, run against the two sample valves.

These figures come from the fabricated sample entries. They illustrate the method, not the performance of any real product.

Step 4: Read the Table, Not Just the Verdict

Three things in that table matter more than the pass column.

Extra valves show sharply diminishing returns. The second valve nearly halves the peak. The third takes off another 2.5 kPa. The sixth is worth less than half a kPa. If your design is relying on the fifth and sixth valve to scrape past the limit, it has no margin, because the next thing you learn about your venting curve will move the answer further than those valves did.

A membrane valve pins the peak at its setpoint. Every M001 row reads exactly 15.000 kPa, regardless of how many are fitted. Once that valve opens, its flow capacity so far exceeds the gas generation rate that pressure cannot climb past the point at which it opened. Adding a second one changes nothing, because the first one was never the constraint. For a valve in this regime, the setpoint is your peak pressure, and the only way to lower it is a valve with a lower setpoint.

The two valves fail in different ways. S001 opens early, at 5 kPa, but its relief flow builds gradually, so the peak depends on how many you fit. M001 holds nothing back once it bursts, but it does not burst until 15 kPa. Which is better depends entirely on where your enclosure limit sits. Below 15 kPa, no number of M001 valves will ever pass.

Step 5: Check the Margin, Then the Sensitivity

A pass is not a result until you know how fragile it is. Three checks, each one run in a minute:

Temperature. Re-run at a higher assumed gas temperature. If the peak moves a lot, your result is resting on an assumption rather than on data. See Temperature Modes.

Venting rate. Re-run with one more venting cell than your requirement states. This is the cheapest proxy for “the gas arrived faster than we thought” and it is usually the most revealing.

Void volume. Re-run at a smaller void volume, since layout changes rarely free up space. On the case above, this one turns out to matter very little: between 20 L and 100 L of void volume, the peak moves from 12.057 kPa to 11.950 kPa with two S001 valves fitted.

That result is worth understanding, because it generalises. Once the valve is relieving, the peak is set by the balance between the gas generation rate and the valve’s flow at that pressure. The void volume changes how fast pressure gets there, not where it settles. Void volume matters most in cases where the valve barely opens at all.

Step 6: When Nothing Passes

If every candidate valve at every plausible count exceeds the limit, the answer is not a different valve. The problem has moved outside the valve, and there are only two places it can go:

Raise the enclosure rating. A stronger lid, a better seam, or more fasteners buys pressure limit directly. This is a structures conversation, and the number you bring to it is the peak your best candidate valve produced.

Reduce simultaneous venting. Propagation barriers, cell spacing, and thermal insulation reduce how many cells vent together. This changes the scenario legitimately, because it changes the pack, and it is the intervention with the largest effect on peak pressure. It is also the longest lead time.

Both are decisions above the level of valve selection, which is exactly why it is useful to reach them early, with a specific pressure number attached.

Before You Commit

Two things this tool does not check, and which have sunk otherwise sound selections:

• The breathing duty cycle. A valve that passes the venting check may still be wrong for the pack if its everyday permeability does not match your thermal and altitude cycling. That is a separate calculation.
• Real ejecta. Vented gas carries particulates and condensable vapour, and a real relief path can underperform its clean flow curve. See Assumptions and Limits.

Use a comfortable margin as confidence to proceed. Use a marginal one as the justification for a test.