Hi there! I'm a scientist working on fusion technology and a KSP player since 2013. About a month ago I saw the Twitter post with the LANTR engine that mentioned that y'all were still adding engines to the roster. Myself and two other PhD fusion scientists / KSP fans came up with a proposal, which is finally ready to share. I hope it's not too late! So here I present:
A Direct Fusion Drive for KSP2
Introduction and summary
Nuclear-fusion-powered rocket technology is at an early stage of development. As a potential near-future technology, we propose adding fusion rockets to KSP2. In particular, the ‘Direct Fusion Drive (DFD)’, which could have an ISP more than double that of the Dawn ion engine, is a good candidate. We select characteristics for the DFD to balance realism and playability, as well as game balance (with the KSP1 engines). In the implementation described here, it most directly competes with the NERV and Dawn, and also, due to its higher ISP, extends the space of possible missions. Table 1 summarizes the proposed implementation.
Table 1: Proposed parameters for the Direct Fusion Drive engine.
Engine
Thrust / kN
Vac ISP / s
Mass / t
TWR
Cost / spesos
Fuel Type
Size/m
DFD "PFRC"
40
9000
3.5
11.42
100k
Deuterium, Helium-3
2.5
KSP1 "Nerv"
60
800
3
20
10k
LiquidFuel
1.25
KSP1 "Dawn"
2
4200
0.25
8
8k
Xenon
0.625
The section ‘The Direct Fusion Drive for KSP2’ describes a few further proposed details as well.
In real life, fusion rockets are a ‘near-future’ technology that could enable transport of larger payloads to Mars, missions to the outer solar system, and the interception of long-period comets. One design in particular is the Direct Fusion Drive. It’s currently being developed by researchers at the Princeton Plasma Physics Laboratory and Princeton Satellite Systems. Figure 1 (in the above spoiler) shows an artist’s rendering of the engine.
The engine concept is as follows: a hot Deuterium-3He ‘Field-reversed configuration’ (FRC) plasma, confined using a set of electromagnets and stabilized by using radiofrequency waves which quickly rotate it, generates heat and fast particles. The hot plasma is about the size of a bathtub. The outer diameter of the magnets is about 2m, and the device is about 10m long. In order to generate more significant amounts of thrust, a cold deuterium gas is passed around the outside of the plasma where it is heated and directed out the back using a magnetic nozzle. In order to run the engine, about 10% of the power output is converted to electricity, which is re-injected into the plasma as RF waves. Another fraction is lost to heat. In principle the engine can be used as a standalone generator upon reaching the destination.
Figure 2 shows the exhaust velocity (proportional to ISP) and thrust ranges of a number of rocket technologies. The region between the two blue dotted lines indicates the ranges of exhaust velocity and thrust that DFD technology in general can produce: any specific engine might be able to trade off thrust and ISP along a diagonal line. This is an example of a ‘power-limited’ rocket.
As seen in Figure 2, when comparable exhaust velocities of 10 km/s to 100 km/s are chosen, the thrust is about 103 times higher than that of ion engines and Hall Effect thrusters, and also about 102 to 103 times lower than the NERVA. Figure 2 also shows the exhaust velocity and thrust of the KSP “Dawn” ion engine and “NERV”. For gameplay reasons, the Dawn’s thrust has been greatly increased relative to real life – coincidentally it is similar in magnitude to what a real-life DFD might provide.
The Direct Fusion Drive for KSP2
In KSP2, as in real life, a DFD engine would have thrust between the ion engine and nuclear thermal propulsion, but with higher exhaust velocities. It would also be further in the tech tree than the Dawn. It would be used for propelling larger ships on missions with larger dV requirements than the Dawn can provide, and gives a slightly shorter burn time. Compared to Dawn (mass 0.25t, ISP 4200s, thrust of 2 kN) it would have higher mass (3.5t, ISP of 9000s, and thrusts of say, 40kN). It would be a 2.5m diameter part rather than a 0.625 part.It would generate electric power rather than consume it, so no large solar panels are needed. But, it would also generate significant quantities of heat, requiring large radiators. Similarly to the Dawn, its thrust goes to zero with any trace of atmosphere.
It requires two new resources, deuterium and 3He, though not in equal quantities. The fusion reaction itself burns the two in a 1:1 ratio, but extra deuterium is flowed around the outside of the plasma to augment the thrust. The ratio of deuterium-as-propellant to deuterium-as-fusion-fuel is about 10000 to 1. The deuterium might be stored in standard-ratio tanks (8:1 fuel to tankage ratio) like LiquidFuel, but the 3He is stored in 1.3:1 gaseous tanks like Xenon.
We propose two further possible complications, to make things more interesting: one is that activating the engine requires a one-time burst of a large quantity of electric power, to represent the energy needed to start up the plasma. The engine could constantly burn Deuterium and 3He (in a 1:1 ratio) even when the thrust is set to zero until it is deactivated. It would also constantly generate a set amount of electric power and heat.
A further realism element would be the ability to inversely trade ISP and thrust by changing the amount of deuterium propellant once the engine is on. This opens the door to playing with power-limited rockets and brachistochrone trajectories. In this implementation, the 100% throttle setting might be a thrust of 80kN with an ISP of 4500s, 50% throttle would be the ‘basic implementation’ 40kN and 9000s, and 25% would be 20kN and 18000 s. One percent throttle would give 0.8kN at 450000s. At zero throttle, the engine would not generate thrust, but it could still be used (for a space station or base) as an electrical generator.
Game Balance
The DFD is a powerful new technology, but it should not be strictly better than its competitors. This can be achieved by thoughtfully setting its mass, thrust, and ISP, or in other ways, such as placing it further on the tech tree, give it high monetary cost, or by making its fuel difficult to obtain. We first discuss its physical performance at the proposed (basic) settings: mass = 3.5t, thrust = 40kN, and ISP=9000s.
Figure 3 shows an ‘Optimal Engine Chart’ for a 20 ton payload with the DFD added, and compared with Dawn, Nerv, and several of the highest-performing conventional engines. Each colored patch shows the (single) engine that will yield the stage with the lowest mass, given delta V and acceleration requirements. The blank region is infeasible.
The DFD allows missions with reasonable TWR with delta V of more than 20000 m/s. It covers most, but not all, of Dawn and Nerv’s parameter space. This is expected as it is an advanced technology! For this 20-ton payload, the Dawn is optimal only at very low Kerbin TWR and delta V up to 6000 m/s, and Nerv allows a higher TWR for the 20-ton payload mission when dV requirements are up to 3000m/s.
The DFD beats the Rhino and extends the overall feasible parameter space at delta Vs of 7000 m/s and above. Its TWR is slightly higher than that of Dawn (11.4 vs 8), but lower than that of the Nerv (20).
The DFD does not totally obsolete the Dawn for all missions. Figure 4 shows that for a 3 ton payload (and with up to 3 engines of a type allowed), Dawn is optimal up to 15000 m/s. The DFD does cover most of the Nerv’s parameter space, so it should be balanced in non-physical ways:
The DFD could be significantly more expensive than the Nerv (100k vs 10k); this would promote its reuse. For comparison, 20 Dawn engines, which would yield 40kN of all-electric thrust, cost 160k spesos.
The DFD is further along in the tech tree than the Nerv, so perhaps it’s relatively okay to obsolete the Nerv once it is developed. This make sense since it’s a near-future technology, while nuclear rockets are something that we probably could have flying today if there were no safety concerns.
If new resource requirements are on the table: it could require 3He, which is either ferociously expensive or totally unavailable on Kerbin, and which needs to be mined from certain locations in the solar system. It can be mined in very low concentrations from Mun or Minmus regolith, or in moderate concentrations from the atmospheres of ice-giant planets. (Deuterium is also more expensive than LiquidFuel but relatively easy to find.) 3He could also be generated from standard D-T tokamaks (or Z-pinch reactors!) if extra Tritium is allowed to decay over time. That's slow, though.
It’s also larger in diameter and length, which makes it awkward as a payload during an initial launch from Kerbin.
Conclusions
The Direct Fusion Drive is a plausible near-future propulsion technology that could be a great fit for KSP2. It has unique abilities to enable high delta-V missions, but also could have unique requirements.
...
I'd be happy to answer any questions that I can about this!
