What Is Temperature Monitoring?

I spent three years doing pack validation at a Tier 1 supplier in metro Detroit. Temperature monitoring was maybe 40% of what I dealt with day to day. Not because it's complicated. Because people keep screwing it up in the same ways.

Validation testing in the lab environment.

The basic idea is you stick thermistors on cells and watch the numbers. That's it. NTCs mostly. Cheap, accurate enough, everyone knows how to use them. The problems start when you ask where exactly to put them and how many you actually need.

Cost pressure is brutal. Program managers want fewer sensors. Thermal engineers want more. I sat through dozens of those arguments. The compromise usually lands somewhere unsatisfying for everyone. Eight sensors for a module with 12 cells. Put them on the cells you think will run hottest and hope you guessed right.

Thermal simulation meets real-world constraints.

Simulation vs. Reality

You can model this stuff in ANSYS or STAR-CCM+ until your eyes bleed. Simulation tells you the center cells in a module run hotter than the edges. Tabs run hotter than can bodies. None of that is surprising. What simulation doesn't tell you is what happens when your production line puts thermal paste on inconsistently. Or when a cell from a bad batch has slightly higher internal resistance than the others. Or when the customer installs the pack in a vehicle with an air intake right next to the exhaust manifold. Seen all three.

Here's the thing though. Runaway is rare. What actually kills packs in the field is years of running 5 or 10 degrees warmer than the cells want to be. Capacity fade. Power fade. You won't see a fire. You'll see warranty claims at year 4 when the range drops below spec. Temperature monitoring matters for that too but it's less dramatic so people don't think about it as much.

 

The BMS samples temperature at something like 1 Hz. Sometimes slower. Thermal events in cells happen over seconds to minutes so that's plenty fast. What matters more is where the threshold limits get set.

 

Too conservative and you're cutting power when the pack still has headroom. Too aggressive and you're cooking the cells. Every OEM has different philosophies on this. Some aim for maximum performance, others prioritize longevity. You can usually tell which by how the vehicle behaves on a hot day at a fast charger.

The Cold Reality

Cold weather is actually harder to deal with than hot. A pack at minus 20 C has almost no available power. Internal resistance goes way up. The cells can't accept much charge current without risking lithium plating. So you need heating, which means burning energy to warm up the pack before you can use it. Some vehicles precondition automatically when you set a departure time. Others make you sit there waiting.

Temperature monitoring in that context means knowing when the pack is warm enough to actually use. And making sure the heating system doesn't create hot spots. Running a PTC heater right next to one cell while others stay cold is not great. Seen that too.

The evolution of sensing technology.

The industry is moving toward more sensors, not fewer. Partly because NTCs keep getting cheaper. Partly because the analytics people want more data to feed their degradation models. There's talk about fiber optics inside cells but I'll believe that's production ready when I see it on a line. For now it's thermistors on cell cans and busbars, same as it's been for 15 years.

That's temperature monitoring. Not the sensors themselves. The decision about what to do with the data they give you.