Seems that it can be done, and here are instructions
Copying from the link
Strike an iron nail squarely and sharply several times with a hammer while keeping the nail positioned in a north-south orientation. The impact of the hammer with the iron nail causes the magnetic domains within the nail to break loose from their current orientation. The Earth's magnetic field will then reposition the domains into a new orientation parallel with the Earth's magnetic field.
It is evident that this can be done only with materials that have small domains with magnetization, which are randomly oriented, so the material has to be ferromagnetic.
You don't give any sources, and some of your statements are not universal in the best case. Iron's domains are not necessarily aligned along the external magnetic field (e.g., they can be in a metastable state with low average magnetization). Furthermore, striking iron with hammer can actually be used to magnetize iron in external magnetic field (see, e.g., http://www.youtube.com/watch?v=GaD9vAuj20s , where an iron pipe is magnetized in the weak magnetic field of the Earth using gentle persuasion with a hammer:-) ) - as a result of hammering, iron domains transition to a more stable state with larger average magnetization.
EDIT (10/15/2013): I am afraid I have similar problems with the edited question. Your statement "If an iron is exposed to an external magnetic field, the iron's domains are all aligned with that ext-B" is not universally correct. For example, a permanent magnet's magnetization typically does not follow the direction of the Earth's magnetic field.
In general, there is an equilibrium (stable) state of a ferromagnetic for fixed values of temperature, pressure, external magnetic field, etc. However, the ferromagnetic is not necessarily in this stable state, it can be in a metastable state, e.g., because of domain friction. As far as I understand, what hammering does, it helps overcome domain friction. As a result, the ferromagnetic tends to approach the stable state as a result of hammering. However, this stable state can have larger or smaller magnetization, or a different direction of magnetization than that of the metastable state, depending on the external parameters (temperature, pressure, external magnetic field). Therefore, hammering can either magnetize or demagnetize the ferromagnetic, depending on the relationship between the stable and the metastable state. The following link may be useful: http://capone.mtsu.edu/phys2020/Lectures/L12-L18/L17/Ferromagnetism/ferromagnetism.html
Best Answer
I believe the term you're looking for is "Magnetostatic Energy". Magnetostatics is the field that studies static (constant) magnetic fields, much like electrostatics.
For a uniform material the magnetostatic stored potential energy is:
$$E_{\mathrm{ms}} = \frac{1}{2}\mu_0 \int_V \mathbf{M} \cdot \mathbf{H}_{\mathrm{ms}} d^3 r$$
You can find a full derivation here. The Wikipedia article on magnetic domains also covers a lot of details about the field energy.