As far as I can see you simply want to drive your glowplug.
By the circuit you are showing you would need about 12A through the 1 Ohm resistor for a fully-on driver.
If you assume a transistor gain of 250 the collector current vs. the base current dictates you need 12A/250=48mA flowing into the base of the transistor.
At 4V driving signal and assuming 0.85V Vbe drop, you need a base resistor of (4V-0.85V)/0.048mA ~ 65 Ohm
You stated the transistor is actually a darlington type. This means that when fully ON, the collector~emiter voltage drop is about 1V when fully ON. This means that out of 12V, the glowplug will only see 11V, the remaining voltage will be dropped on the transistor.
Considering all this, the powers will be:
P(battery)=12V*11A=132W
P(R1)=151mW
P(glow)=11V*11A=121W
P(transistor)=1V*11A=11W (HOT!)
You have to cool the transistor. the TO-220 package for TIP120 is NOT able to handle 11W of power with no heatsink. You will need a big one.
If you need to adjust the power, you should do it with PWM. This means quickly turning ON and OFF the transistor by cycling the input signal 0->4->0. But you have to do this quickly, probably in the kHz region.
If you plan to play with the glow current by varying the current through the glowplug in a linear fashion, the maximum power dissipation can be calculated by the max. power theorem, which states the maximum power dissipated in the transistor will occur when the impedance of the transistor matches the impedance of the glowplug.
In this case the max. power heating the transistor equals 36W (same power is also dissipated on the glowplug). This would require a massive heatsink. It's a waste, so don't do it.
Use a logic level mosfet and PWM it. this will do the trick, but you will have to include a diode (anode to collector, kathode to 12V) as a glowplug kicks back some voltage when you turn it off quickly.
All this is valid assuming you want to turn the glowplug fully ON. If you don't, the same principle applies, but be warned, that by setting the transistor base current to a fixed value and than expecting the collector current to stay stable is considered a poor design. This is probably why you found the transistor is dissipating more power as you expected. The gain was higher than you predicted, so more power was drawn from the battery.
Altough the collector current and base current are tied togather by the gain factor (called Beta for bipolar transistors), the transistor gain is by no means stable and should not be relied on. You can expect about 70% variation, affected by the collector current, temperature and frequency of operation. You must also take into account that the gain differs from batch to batch of manufactured transistors, even from the same factory.