There are many other factors to consider when looking at what case style/power rating is going to work best.  But ultimately, increasing the rotor potential in this manner is the more efficient approach.  Sure, he did find a piece that should accomplish what he wants to do with the alternator within his spec, what happens if that burns out?  Fatigue fails, or some other malady that often besets old equipment?  

My suggestion, which I probably should have explained a little further initially is his next best option for time efficient and materials efficient approach for what he want to do.  That circuit I suggest replaces the integral regulator of the alternator in order to get the alternator output up to a level to charge the 24 volt battery as a single 24 volt battery, instead of as a pair of 12 volt batteries joined at the center post.  He can also save the money by NOT buying the external regulator which is redundant for that alternator series.(or if he has that external regulator already- save it for another project.)

Additionally, if he is requiring an exact midpoint of the ouput of the alternator, if he has his current requirement met, 2 precision resistors in series tied from ground to the main power rail give him his reference voltage at the point located between the two resistors.

If the battery is being charged at the full 24 volts, there is no need for a relay, nor is there a need for finding a way to disipate the theoretical excess current. Given his original parameters in his first post, that is what he was sounding like he was wanting to do beside the switching aspect.  Besides, 2 TO-3 type 2N3055's in parallel have a 200 watt rating, and would be able to handle 96 Watts with plenty of dissipation to spare.  and I did also indicate in that post if he did not feel it was adequate, I did indicate just a few sentences below the mention of 2N3055's "...he could add more in parallel.  

Ultimately there is little problem running a transistor rated for 100 Watts, at 50 Watts power level, or a pair rated 100 Watts each at 100 Watts. IF you provide adequate heat dissipation, and are also within the voltage ratings of the transistor.  

The reason my suggestion works is if the alternator rotor is being run at a potential to generate the 29 volts nominal for charging a 24 volt battery, the armature is not likely drawing much more than about 4 to 5 amps at max charging rate of 165 Amps at 14 volts or 82.5 Amps at 28 volts, in both cases at approximately 4500 RPM.  The extra energy of the motion of the alternator is where that extra energy is coming from off the 3 phase stator.  The rotor also is normally at a lower potential than the output voltage of the alternator.  I have a pending project with a similar alternator that once it is done I can take the exact measurements and make adjustments for my application.  Most of the time the current drawn by the rotor/armature of a typical automotive/truck alternator should be in the area of 2 to 3 Amps.  

The beauty of the LM117/317 circuit is the adjustability of it.  For extra current drive, you add as many power transistors as you would need for your expected power handling requirement in parallel as described loosely on page 14 of the Application notes I cited.  If you needed to disipate 1000Watts, you either add 2n3055's in parallel until you are comfortable with the performance envelope, or you switch to a different type with higher ratings.  

2N3055 is a general type,  and is available as either a plastic version (with lower power ratings) that can fit the space requirements for a metal can TO-3, or is available as a TO-3, which is manufactured by a number of different manufacturers.