Turboprops

[blowfishpro]

Most of the pieces required for turboprops are in place already - we would need a Turboshaft class implementing ITorqueProducer and a module supplying parameters to it. The core of the Turboshaft could reuse the compressor and turbine code from jets, and of course the actual propeller code could be reused with no modifications.

The remaining piece of the puzzle is modeling the power turbine, and specifically how power output changes with shaft RPM, since NASA's turbine model doesn't account for RPM at all.

[NathanKell]

Good. That was certainly my hope from writing it like that. :)

[blowfishpro]

As a first approximation we could say that shaft power doesn't depend on RPM, though I'm not sure how close to/far from reality that is.

[blowfishpro]

I guess the other unknown is that the compressor code requires a maximum turbine inlet temperature which doesn't seem to be available for turboshafts.

[gertsonderby]

It seems that a common concept in turboshaft engines is a cap at a thermodynamic limit - temperature and/or RPM-based - for the engine. This cap, meanwhile, is static as the plane climbs, and as more turbine speed is applied, leading to a steady power output over a wide span of altitude and velocity.

An engine may not be able to hit its full rated power in certain cases, however - early turboshafts seem to have had trouble with it on the runway, where intake temperature was relatively high.

I'm curious, now, so I'll dig a bit further.

[gertsonderby]

Initial results indicate that turboshaft output has ridiculously high RPM as a basis - on the order of 10K and up - but are usually geared down, and attached to variable-pitch constant-RPM propellers as a rule. So I think a reasonable approach might be to abstract away the internal RPM of the engine, instead resting mainly on the power output and prop RPM for our resulting thrust output. Now, I don't know how possible that is, code-wise, I haven't dug deep there yet. But if we can abstract this, I think it'd be worth it.

[blowfishpro]

The current jet solver code doesn't deal with RPM at all, so I think it's fine to abstract it.