In-game Simulator/Design Evaluation feature thoughts

[pandaman]

An in-game simulator feature has been requested many times already (including by me) to help with ship design and mission planning.  Whilst I think such a feature could be helpful if done sensibly, I had a re-think on what approach would be best for the game and how it could best be implemented, so...

As I see it there are two main areas that need to be covered, Design Evaluation and Test Flights.

Test Flights are exactly that, a real launch designed to check out and test some aspects of the planned mission or design.  As such they will very often only carry the bits needed to complete the test.  A good example is the early Apollo missions, they were very much not 'Moon landing attempts' but they were vital to test and evaluate different aspects that were needed for the eventual success of Apollo 11.  They were all 'Test Flights' but they were still very real flights with very real financial cost and very real risk.  And I think it's this aspect that the 'Trial and Error' approach replicates quite well,  no matter how good a simulator is there is still no substitute for trying it out for real.

Design Evaluation is where I think a 'simulator' feature should fit, and it's probably a better name too.  IRL we design probes for Mars, we calculate what we think they will behave like based on what knowledge and experience we have, but we don't 'know' if it will actually work until it gets there and either succeeds or fails.  This is the role I see for a Simulator/Design Evaluation feature, it should not replace Test Flights or the need for them, or guarantee success, but should help players to ensure that the basic design has a fair chance of doing the job it was designed for before committing to an expensive and risky test flight.  Effectively an in game equivalent to getting the boffins to re-check all the maths, running computer simulations and ensuring that the ship is balanced and the staging plan works as intended etc.  Yes it would need sensible design aids like Dv, TWR, CoM and CoL readouts by stage during initial construction, but also just a simple facility to run a dummy launch sequence, or selected portion of the flight 'in the lab', to check that the ship itself 'works', so that when it's launched either as a test flight or full mission the risk of failure due to design oversight is reduced.

Of course there are always the 'Revert' and 'Sandbox/Hyperedit' options for those that wish to use them, but for a 'No Revert Career' game this type of feature could fit in nicely and be less 'overpowered' and a bit more realistic than a 'fully accurate practice run' simulator.

 

Thoughts, comments and alternative suggestions welcome.

Edited February 22, 2016 by pandaman

[LorenLuke]

When I think 'Design Evaluator' I think of the wind tunnel test that they occasionally show on those luxury car commercials:

I think something like this with statistics and sliders and overlays would be neat tohave, especially by stage, such as what would be most prone to overheating (similar thermal graph like what F12 does), all the lift surfaces (F11), and some sort of statistics/feedback for how they'd perform at an adjustable speed and altitude/atmospheric density.

[pandaman]

Yeah, that's the sort of tools I have in mind.  When they designed the sky crane lander for curiosity they had to do all the calculations, check 'em, check the check just in case, then just hope it worked on Mars like it was supposed to.  That's the kind of design evaluation process I was thinking of.

I don't think an 'accurate flight simulator' would be appropriate, but you need to be able to check concepts and run through stuff like launch and staging sequences and get basic aerodynamic evaluation etc before committing to a live launch.  Just as IRL they design it on paper and work all the stuff out as best they can before even starting to build anything.   I see the VAB/SPH as a sort of 'rocket design CAD package' once it's designed you check it with the evaluation tools and tweak it as needed and when happy, or just impatient,  go for launch.

[Speadge]

just ONE other approach:

while i think it would be nice, i dont believe its something that will happen in stock since the DEVs have many other things more important to fix first.

Best chances are with a mod here

[pandaman]

A long time ago, once I'd got past the basics of KSP and started planning interplanetary stuff was when I (like many others) thought a simulator to help me test stuff before committing a lot of time to the mission would be a good idea.  But the more I though about it recently the more I thought that a proper 'simulator' would not be the right approach for the game.

What I think is needed really are a better, more comprehensive, set of design and evaluation aids that can give us the chance to design a bit more effectively.  The usual suspects being Dv and TWR info etc of course, but also simple ways of checking that stuff works as we want, like action groups working in the VAB/SPH, a 'wind Tunnel' setting that shows lift/drag arrows that can be tweaked to show the effect of different atmospheres and altitudes etc. and perhaps a 'zero G simulator' setting that allows us to check thrust balance etc and see directly how stuff is affected, without needing to do an 'actual' launch and hack gravity.  The kinds of tools that would be available to help RL designers either as computer simulations or physical tests, but represented in game in an easy to use way and ideally available in the VAB/SPH environments. 

Some of this would, I guess, not be too difficult to implement and other bits would be much harder, but this kind of approach would be better for 'in game' IMO than a 'Hyperedit to x' type simulator.

There are of course a lot higher priorities dev wise, especially at the moment, but maybe we could see things along these lines start to appear at some point.

[Buster Charlie]

I posted on this, if there is a simulator I think it should look very basic and uninteresting as not to take away they enjoyment of visiting a planet  for the first time.

 

I also think, as you mentioned simulators are limited, that any simulator should be dependant on researching a planet just like IRL.

 

So I've proposed the idea of ground based, orbital, and probe based sensors and telescopes that you could target a body with, and over time (based on duration of study,  number of sensors,and technology of sensors, and proximity of sensors) you developed more info and better simulators.

 

For example ground based telescopes  could let you guess the orbit size, but you need more advanced methods of discovering  mass, terrain, atmosphere, temperature,  etc.

 

So while early game you just know Duna is there, you have no way to simulate it to any degree.

 

So say you build a satellite in kerb in orbit, now you see its rocky,  not a gas giant, it's a certain size,  maybe guess it's density, but that could be wildly off, you don't see a thick atmosphere so you can make some aerodynamic guess.

 

So really the next step is a probe flyby,  you get a real good look at the surface,  some good spectral analysis. 

 

So you now have a good enough data set for a fair simulaton, and rough terrain mapping.

 

But if you really want to make sure jeb comes back alive, you drop an unmanned probe onto Duna. You don't know enough to even land it right, but even if it crashes you know almost everything you need to simulate a landing.

 

You know have an atmospheric density map as you transmit while falling, and you know the gravity acceleration, local temperature, whatever.

 

Basically, I'd make simulations and testings  as effective as your dataset, and as generic looking as not to replace real flights. 

 

I'd also use this to emphasise unmanned probe research on new planetary conquests. 

[LorenLuke]

10 hours ago, pandaman said:

A long time ago, once I'd got past the basics of KSP and started planning interplanetary stuff was when I (like many others) thought a simulator to help me test stuff before committing a lot of time to the mission would be a good idea.  But the more I though about it recently the more I thought that a proper 'simulator' would not be the right approach for the game.

What I think is needed really are a better, more comprehensive, set of design and evaluation aids that can give us the chance to design a bit more effectively.  The usual suspects being Dv and TWR info etc of course, but also simple ways of checking that stuff works as we want, like action groups working in the VAB/SPH, a 'wind Tunnel' setting that shows lift/drag arrows that can be tweaked to show the effect of different atmospheres and altitudes etc. and perhaps a 'zero G simulator' setting that allows us to check thrust balance etc and see directly how stuff is affected, without needing to do an 'actual' launch and hack gravity.  The kinds of tools that would be available to help RL designers either as computer simulations or physical tests, but represented in game in an easy to use way and ideally available in the VAB/SPH environments. 

Some of this would, I guess, not be too difficult to implement and other bits would be much harder, but this kind of approach would be better for 'in game' IMO than a 'Hyperedit to x' type simulator.

There are of course a lot higher priorities dev wise, especially at the moment, but maybe we could see things along these lines start to appear at some point.

I'd say being able to see things like force of drag, terminal velocity, etc. at (input speed, pressure, angle, temperature, and gravity) in the form of enterable fields would be more like modeled simulations, with some tools to interpret the data, and sliders to view charts of max thrust, acceleration, drag, ISP, etc. for your craft.

 

10 hours ago, Buster Charlie said:

[snip]

I'd say that's a bit out of scope, and while you could gain stuff like atmospheric density from exploring, I'd contend that you'd not be able to simulate a landing. That said being able to enter values and basically see 'how would my ship perform at this gravity/pressure/speed/angle/etc. would be useful to tell what the potential pitfalls of landing on the something like Duna would be (as you'd have some of that information from exploration).