md2perpe said:
Does Noether's theorem really specify the value of the energy? Doesn't it just give an expression ##h(t,x,\dot{x})## that is conserved? I mean, if ##h## is conserved, then so are ##h+C## and ##Ch## where ##C## is a constant. Can you take ##E## to be any of these?
As in non-relativistic mechanics also in special-relativistic mechanics the choice of the "energy-zero point" is arbitrary. You can choose it as you like without any change in the physical meaning of energy as a conserved quantity following from temporal translation invariance a la Noether. However, it's convenient to group your quantities in terms of Poincare-covariant quantities, and in the case of energy it's clear that it should be grouped together with momentum, which are the conserved quantities due to spatial translation invariance. If you choose momentum in the usual way, as
$$\vec{p}=m \mathrm{d}_{\tau} \vec{x},$$
where ##\tau## is the proper time and ##m## the invariant mass of the particle, then it's convenient to choice your "energy-zero point" such that
$$E/c=m^2 c^2+\vec{p}^2.$$
Then
$$(p^{\mu})=(E/c,\vec{p})$$
are the contravariant components of a Minkowski four-vector. The invariant mass is then given by
$$p_{\mu} p^{\mu}=m^2 c^2,$$
and thus is a scalar, as it should be.
In the case of a battery of course also the chemical energy, mechanical stresses, the electromagnetic field etc. contributes to the total energy in the rest frame of the center of energy (not center of mass as in non-relativistic physics!) and thus to the invariant mass, ##m=E_{\text{COE}}/c^2##.
Concerning the rescaling of coordinates and/or momenta, it's not related to Noether's theorem since scaling invariance is not a symmetry of Nature. Even in models without any dimensional fundamental quantity (like massless electrodynamics or pure Yang-Mills theory) the quantization breaks the scaling and/or conformal symmetry ("trace anomaly").
Also I wouldn't count the redefinition of units as a symmetry, because all properly formulated theories of physics are independent of the choice of units to begin with (though in the case of electromagnetism it's more subtle since you've in principle the choice between 3, 4, or even 5 base units, leading to more complicated conversions between the electroamagnetic quantities).