In a crowded art, the application has to say why the change works.

A fuel cell stack is a thoroughly documented machine. Separators, flow fields, membrane electrode assemblies, reactant gas passages and end plates have each been described many times over, and an examiner building an obviousness rejection can usually find every structural element of a new stack in the references. After KSR, a combination of familiar elements according to known methods is likely to be obvious when it does no more than yield predictable results, so the question in examination is rarely whether the parts are new. It is whether there was a reason to arrange them this way, and whether the result is one the references would have led anyone to expect.

That answer is strongest when it is already in the specification. An application that shows a separator geometry and stops gives the examiner nothing to weigh against the combination; one that explains the mechanism — the reactant it distributes more evenly across the active area, the water it keeps out of the flow field — and ties that effect to the claimed structure has made the argument before the rejection is written, and gives any later declaration evidence something to stand on. Operating methods raise the same problem in a different form. A start-up or activation sequence is a series of known steps, and its case for patentability rests on the condition it prevents or the result it achieves, which is what the claim and the specification both have to state.

Much of battery management is estimation: the temperature inside a cell that no sensor reaches, the state of charge, the capacity, the internal resistance. A claim that recites acquiring voltage and current and calculating one of those values reads, to an examiner applying Alice and the Office's eligibility guidance, as a mathematical concept performed on a generic controller, and it is likely to draw an eligibility rejection however careful the model behind it.

What holds is the control the estimate drives. A claim that uses the estimated temperature to decide when and how far to raise the battery's temperature, to control charging, or to balance cells integrates the calculation into the operation of a physical system, which is the kind of practical application the eligibility analysis looks for, provided the control step is recited with enough particularity that it is more than an instruction to apply the result. The specification should say why the estimate is needed — that the quantity cannot be measured directly in a sealed pack, and what goes wrong when the control acts on a poor one — so that the technical improvement is on the record before anyone questions it.