Now, the question is: How much of it do the powerful chips that generate a lot of heat ("heat" is thermal energy) transfer into the battery; that's a question of considering where the heat goes.
First thing to realize is that if you have a perfectly sealed, perfectly thermally isolated box, it's going to heat up forever, and at one point break down/melt/outshine the sun in its temperature.
Obviously, that's not happening, because the enclosure of your device isn't a perfect thermal isolator.
Conversely, your board itself isn't a perfect thermal conductor, either: if that were the case, all spots on the board would instantly have the same temperature!
So, in a first step, it'd be important to model how warm your overall device gets. Lets consider it as a black box: Inside, someone converts \$P\$ watts of electrical power to heat, and these will need to be dissipated to the environment in order to stop the infinite heating up.
Now, the way we model that actually uses similar terminology as we're used from Ohm's law: There's thermal resistance, that tells us how much something is in the way of heat flow. Its unit is typically "K/W", or "°C/W" and tells us how much hotter something gets if a specific power is converted to heat inside.
You'll often find IC datasheet specifying something like a "junction to environment thermal resistance 45 °C/W", and together with an estimate of how much power the IC uses (for example, voltage drop times current in a linear voltage regulator), you can tell how much hotter than ambient things get.
So, our process goes like the following:
- Estimate how much power is converted to heat in your system.
- Estimate the thermal resistance of your enclosure; that times the power from 1. gives you how much hotter the inside of the enclosure is than the outside
From here, I'd guess that in any typical device, improvements are minor by being more detailed: If you're already above 45 °C, then you're not colder anywhere inside the box (after a while, at least), and your device needs better cooling.
If you're sufficiently below, and there are enough places heat can go without going through the battery, you honestly don't need to worry too much.
Problematic would be if you're close below 45 °C inner-enclosure temperature; then you'd need to calculate further:
- Estimate how much warmer the components in close proximity to the battery are than the in-box environment: same procedure as above, but ambient temperature is the already elevated one of the inside of the box.
- calculate the heat transport that reaches the battery by putting all thermal resistances in parallel between the heat source and the battery and calculate how much heat will flow into the battery.
Step 3. and 4. are pretty often done in simulation, because estimating how much heat a complex PCB and a battery fixation will transport is hard.
Step 1. and 2. can be done pretty well by hand: For the outside of the box, you can often assume something like "well enough ventilated place" and hence assume cooling by convection and maybe radiation. There's ready-to-use formulas that relate horizontal and vertical surface area to the resulting thermal conductivity and resistance.