Close

EMDrive realtime simulation

A project log for EMDrive/satellite

Developing a small fuelless microwave thruster

paul-kocylaPaul Kocyla 08/17/2016 at 05:5518 Comments

Finally finished the EMDrive photon based simulator.
It´s running realtime for millions of photons, so it´s also neat to look at. I´ll prepare a software package for download in the next time.
The cool thing is: The model is quite simple. Each photon has a momentum vector and numerically propagates through space in light speed. If it hits a wall, its momentum is changed according to reflection on the wall´s normal vector at this place.
To conserve momentum, the same (but negative) momentum change is applied to the EMDrive body (according to the radiation pressure rule). The corresponding code for the reflecting function is in the picture.
The yellow line shows the body´s momentum after some time.
And it shows that the small end is indeed leading. The more reflections there are the larger the EMDrive´s momentum becomes. This explains why a higher Q factor would produce a higher thrust.
So do I miss something or can it be that simple? I mean I just implemented the rules for reflection and momentum transfer like shown in a physics book.

Discussions

IslandPlaya wrote 10/28/2016 at 10:37 point

I wrote a molecular dynamics simulation as part of my PhD. There are at least a dozen ways to get the reflection calculation wrong. Your simple code snippet tells us nothing.
Your simulation is incorrect. The result should be no net force.

If you post the source I will point out where your errors lie.

  Are you sure? yes | no

Paul Kocyla wrote 10/28/2016 at 16:02 point

Show me this simulation, then I show you the code LOL, come on, you trolled better before :-D

  Are you sure? yes | no

IslandPlaya wrote 10/29/2016 at 08:49 point

Fine. I am 100% certain your code is flawed. Why is it not open-source? Is it that advanced?
Yet another claim you make with no justification or evidence whatsoever.

  Are you sure? yes | no

willemstaal wrote 09/02/2016 at 10:14 point

I don't know if im up to something, but i stumbled while reading a book about Fractals from  James Gleick where he quoted a dicovery from a Dutch scientist called Christiaan Huygens. He observed  synchronization of pendulum clocks, and he discovered that at some times they run in phase or in anti-phase  due to vibations trought walls or on a table.

I think that a similar behavour occurs also in a EM drive frustrum where the amplitude of the waves are truncated by the shape of the frustrum while the waves forced into sychronization phase. So the energy of the amplitude needs to escape somewhere when a wave goes into phase.. 

The power of the rf transmitter needs to go somewhere! So why not outside the frustrum? but not as wave but as a phase or anti phase synchronization event.

Im not a scientist, who juggles with Unruh theories, but this is the idea what im coming up with.

  Are you sure? yes | no

Paul Kocyla wrote 09/02/2016 at 12:37 point

I don´t know - I´m still trying to understand the new papers :)


  Are you sure? yes | no

TheTravellerEMD wrote 10/28/2016 at 13:43 point

Paul. 

Talk to me. I understand Roger's new papers and the independent review.

Phil

  Are you sure? yes | no

TheTravellerEMD wrote 10/28/2016 at 13:43 point

system error

  Are you sure? yes | no

TheTravellerEMD wrote 10/28/2016 at 13:43 point

system error

  Are you sure? yes | no

TheTravellerEMD wrote 10/28/2016 at 13:43 point

system error

  Are you sure? yes | no

willemstaal wrote 08/29/2016 at 14:22 point

A new paper was published! I think this look very promising!  especially the dev 71 settings! 

http://www.emdrive.com/

  Are you sure? yes | no

Krzysztof wrote 08/24/2016 at 13:40 point

Just a thought - depending on your values, this momentum can be a numerical error, millions of photons seems about right. If your simulator could support any geometry, you could rotate your simulated assembly by e.g. 20 degrees in two axes and test if momentum vector is still pointing at front. If it is, congratulations, if it is axis aligned, then it's just (rounding error x 1 million).

I can't wait to test your code myself... If this works it means just a strong light source in tapered cone with mirror finish would show some effects. It's like solar sail, but multiple reflections are equivalent to multiplicated surface area of single reflecting sail.

  Are you sure? yes | no

Paul Kocyla wrote 08/24/2016 at 18:36 point

Yes, the optical version would be also a great thing to try - maybe even more efficient than the radio based approach.
I am working out a theoretical paper based on the simulated results.
According the error: The photon path stays completely stable during reflections - but I assume a stationary cavity so the whole impulse is reflected. If I´d assume the loss of energa during reflection (which is something like 10^-32 the factor of the initial momentum) then things would get more complicated.
The simulation shows already positive results. The interesting thing would be a mathematical proof.

  Are you sure? yes | no

TheTravellerEMD wrote 10/28/2016 at 13:48 point

Needs a high Q cavity to build internal energy. IE internal energy = Q * Power. Momentum = Energy / c so need high internal energy to show a good effect.

Can high Q optical cavities be built? IE Q in the millions?

  Are you sure? yes | no

max.kargl wrote 08/22/2016 at 18:53 point

Congratulations for your success so far and many thanks for your ongoing efforts. It is so exciting! I wish you the best of luck for your next steps. Will your sim package contain sourcecode, or can it read geometry definition files ? I would like to try "genetic" optimization of geometry and other parameters (if I can ever find time to do so). Of course that will help nothing, if the underlying theory (which I'm afraid of)  is too simple. But would be a nice programming exercise ;-)

  Are you sure? yes | no

Paul Kocyla wrote 08/23/2016 at 07:07 point

Thanx a lot. The source for reflection is only a few lines of code - but the whole package is built around another project, so the code is quite messy.
When I find some time, I´ll post the important functions. Note that the code is just now only handling the cases of two flat walls and a static outer shell - there is no collision detection for free geometry yet.

  Are you sure? yes | no

Krzysztof wrote 08/17/2016 at 07:44 point

Now you can try out different geometries (will cone work?) and injector placements. Also you could try to estimate a force on your drive. Do photons only bounce in your simulation or are they also absorbed?

One nitpick: everyone builds emdrives to use resonance, but it is a wave behaviour. If your photons are of comparable size to your cavity, will your simulation still give the same results?

  Are you sure? yes | no

Paul Kocyla wrote 08/17/2016 at 15:54 point

1) I can define a value after how many reflections a photon will be absorbed. This value can be also randomized at creation of a new photon.2) I have no wave model implemented yet, so I can´t predict the behaviour

  Are you sure? yes | no

attiladtoth wrote 08/19/2016 at 02:34 point

Keep up the good work. You should also try some day (simply a test) of building four then making a drop with them. Like in all of the sci fi.

  Are you sure? yes | no