Engine Design and/or customization

[Balto-the-Wolf-Dog]

Going to post as discussion because it's hardly developed enough to speak on meaningfully otherwise, and is a bit of a longshot at that.

What if we designed our own engines?

A healthy portion of rocketry, if not indeed the majority, is wrapped up in nozzle and engine design; that's not something we generally get access to outside of powder toy/trying it in our backyards. KSP is a rocketry game, and it strikes me the proper environment. I propose an engine designer separate from the vehicle designer, quite likely a 2d cross section designer, to put players in charge of engineering their own thrusters and nozzles. Presumably this would be done in addition to the stock offerings, or perhaps absorb the stock offerings as stock thruster blueprints for those not interested in that degree of micromanagement in their engineering.

Reasoning

As it stands now most of our engineering influence is limited to air frame. When it comes to propulsion we then pick from an array of engines of a few basic categories and configurations, with our largest influence being how many we use, which, and where. It's something we're at the mercy of, and thus we see a lot of begging and balancing to get the quirks and variety we want. That requires an ever-growing number of thrusters with very specific stats, and when it comes time to reap the benefits of them, we still aren't really the ones responsible for the success. Mods like Interstellar are effective in providing novel thrusters, but the benefits fall flat, because all we did was tack them on. Career mode is a step in the right direction, but ultimately science is nothing more than currency and not nearly as satisfying as reaping the benefits of things we've actually experimented with and learned. The Kraken drive and infiniglide phenomenon were good examples of this. While neither was realistic in any way, both devices took off with the community not so much because they worked well (they usually didn't) or offered easy delta V (so does hyperedit, etc). The reason people built kraken drives instead of editing config files for infinite delta V was that it was fun to try to make a kraken drive that worked, and, once you did, it was fun to try to make one that worked better. The physics were typical blow-your-own-sail broken, but the Kraken drive was effectively a player-designed engine that required its designers to run actual experiments and do what could be argued to be literal science to make it work. Personally I spent many hours trying to make a stable one and had a great time doing it. This was all based around an exploit, but engineering itself is effectively the practice of finding "exploits" in the real world. Giving players a tool to develop actual thrusters from scratch or a simple outline would offer the same kind of experimental gameplay in what could actually be a fairly realistic environment.

For the player this wouldn't necessarily be the easiest thing to approach (hence stock blueprints), but it wouldn't be terribly overwhelming. Rocket engines are generally very simple in principle and extremely difficult to prefect, something a player could learn and experiment with as he or she went along without really distracting much from traditional building and flying. Eventually you pull off a thruster that works a little better than the stock ones, or is maybe a little more suited to your needs. Maybe you strike gold with some eccentric design and pull off seriously meaningful thrust and DV. A fairly small number of offered components would allow experimentation with a virtually infinite variety of thruster types and configurations. The results aren't really a balance concern so long as the results of the designer are fairly realistic. Overpowered thrusters aren't fun because they remove challenge, and mean you aren't responsible for your own success. In this case you get out the effort you put in, and if you manage something that would be traditionally overpowered then you're responsible for it's success, and as such it wasn't any less fun and certainly wasn't any easier.

execution

Quantum Vacuum thrusters aside, rockets generally work through the reaction force resulting from throwing things out the back. There are a lot of ways to do this, namely heating for expansion, combustion, and magnetic acceleration, and there are, in turn, a lot of ways to accomplish each of those things. The diagrams above are fairly complicated, probably too complicated for integration with a general audience, but a designer really doesn't need to be that complicated to offer the same degree of player involvement. For instance, the player could be provided with the outline of a rocket nozzle or perhaps a way to create their own under reasonably simple terms. To accomplish this task of throwing things out the back they via magnetic acceleration/ chemical combustion/etc; they would be provided with an array of parts set up not unlike the way rocket parts are now, with offerings for fuel injectors, ignition systems, superconductors and magnets, cyclotrons, etc; then choose and arrange components to make an engine as they see fit. This could be done in 2d and then tested with a simple powder-toy-esque particle physics engine or an akin system that then generates the stats for the engine in simple terms and writes them to a config file for use as an engine is typically used now. A player could arrange a fuel injector to feed into a superconductor bank heated by a cyclotron and channeled into a magnetic nozzle for a simple VASIMR, or, as simply, just add fuel and oxidizer and ignite, designing the fuel flow and nozzle for the efficiency and thrust they're looking for. The beauty of this is something like, say, a hybrid jet isn't something you choose to use but rather something you come up with a way to do. Perhaps you design a jet engine with a secondary oxidizer injector that you can toggle on in flight. A solution that simple isn't going to work, but that just means going back and finding a way that does. All the designer does is take the fuel and watch what happens to it. If it oxidizes and combusts then it oxidizes and combusts, regardless of whether the player used intake air or oxidizer or any mixture of them to accomplish oxidation. A VASIMR design could be made more powerful by increasing the number our output of the heating elements, but this would of course draw more power. Etc. The simplest way I could think to describe this system is as a somewhat simplified powder toy that measures results and writes them to a part that can actually be used in vehicle assembly. Generating a visual model poses a problem, there wouldn't really be much wrong with just substituting a generic model or picking one in the editor. (Perhaps picking the model gives you the cross section you're going to work with?) Nozzle design also poses some difficulty, as nozzles have a lot to do with efficiency and coming up with an accessible way to design them would be something of a challenge. In theory one could simply offer a generic mechanical nozzle and a generic magnetic containment nozzle, but that would remove most reasons to experiment with straight LFO thrusters, whose performance deals mostly with the nozzle design. If a reasonable way for players to design nozzles could be implemented, that would allow generic parts to preform multiple duties, for instance generic magnets could preform whatever magnetic functions are needed, such that the part you're using to contain plasma for heating is the same one being used to channel the plasma for exhaust.

A brief list of parts that make sense to me in this context.

Liquidfuel injector

oxidizer injector

intakeair channel

ignition system

cyclotron

Reactor core

solid fuel

Superconducting magnets

Even with a fairly minimalist parts compliment, anything from an air augmented rocket to a hybrid rocket to a NERVA or any number of electric thruster configurations could be produced.

Something like this would be fairly severe, but I've been thinking about it a while and figured it bore mentioning seeing as I haven't seen any other mention of such a system. Thoughts? Suggestions?

[Rath]

It sounds cool for a official mod, or maybe a simple version in stock with adjustment of nozzle ratios.

[Frozen_Heart]

I agree with this. No matter how many engines are added there will always be gaps in capability. Allowing players to make their own would be fun and encourage experimentation.

[Azimech]

Yes, totally agree. But we need to expand the physics simulation for this to work. Not exactly fluid dynamics, but more complex than it is now.

[Balto-the-Wolf-Dog]

Yes, but not much. That kind of thing doesn't have to be calculated actively, just figured in the editor and then written to a static file. Compiling, basically. Tamer math than say, FAR, at least by my estimation, and performance doesn't matter if its just the process of saving the engine