Posts Tagged ‘simulation’

Not So Fast, Model Might Work After All

July 19, 2013

I had decided (in the last post) that the model I was using couldn’t be right for several reasons in spite of some promising sim results. But upon thinking about it, I realized I was a little too hasty–I discovered a way that a potential (scalar) function could work in accelerating a twist ring in spite of the orientation problem and the curvature problem.    This is an important question because it gets at the heart of why a particle would move due to EM fields.  Conventional theory just asserts the Lorentz force laws, and this works under all relativistic situations.  Conventional theory also says that the electron is an immeasurably small particle.  I have worked out that the twist field particle, which would not be immeasurably small, shows that as it is accelerated relativistically, it stretches to approach the behavior of a linear twist–asymptotically approaching radius size zero.  My hypothesis is that scattering experiments make the electron appear to be infinitely small because of this stretching.  I do have to admit that this experimental result is the chief reason why other physicists discount any electron theories that require non-point like models.

Anyway, back to the sim conclusions–I’ve been trying to create a hypothesis as to why it moves as it does with the twist field theory, and created my simulation environment to test the hypothesis.  I needed to know how a particle knows which way to move when there is one or more nearby sources, and I need it to work right regardless of relativistic behavior.

A big question is whether the particle as a twist ring would sense a variation of field magnitude, or whether the field has to be a vector and the particle senses which way to move based on this vector (which would be a vector sum if there was more than one source).  The scalar field is preferable because then motion can result from the potential function, but I had thought that the orientation problem as well as the curvature problem of negative fields (see previous post) meant that the twist field ring would have to respond to a vector field, that the particle would have to accelerate independent of the orientation.  I also think the stretching of the ring in a relativistic situation might not hold up to correct behavior, but the scalar field is more likely to work than the vector field (simpler–fewer complications in different scenarios).

However, I realized that all of these reasons for thinking the twist sim model are wrong are not all-encompassing. There’s a way around them, which means I have to check those out.  First, the negative field situation, which uses curvature analysis to show a paradox (stronger curvature for a field component that is on the far side of the ring).  I had done the math and things worked correctly, but had reasoned that the math couldn’t be right because it implied a force that didn’t decay with distance.  Now I realize the math is right, because there are three components that add to create the normal acceleration that determines the local curvature of the ring.  The end result of this sum is that while a weaker far-side field cannot induce more curvature, a cancelling out of part of the sum of the near-side acceleration caused by the negated field would result in *less* acceleration there and would achieve the same acceleration (as the far-field stronger field)  for source particles that attract.  The sim was correct, I just wasn’t drawing the right conclusion.

Secondly, I realized that the orientation problem may cancel itself out.  I’ve reasoned that since some orientations cause every point on the twist ring to see exactly the same field, so a solution that depends on the particle sensing the delta field cannot work in that case, and thus invalidates that solution as a general one.  But it is possible that the potential is sensed whether or not the delta field is sensed.  There has to be different behavior between a constant potential field and a sloping potential.  If the orientation problem is real, then there would be no difference in what the particle sensed from source particles and what it would sense if there were no source particles.  The field component would be the same in the local neighborhood  in either case and there would be no information available that would indicate where the particle would move to.   But a solution that has acceleration also due to potential alone, regardless of a change in potential, would work–kind of a switching between normal and tangental effects.  I will pursue this more–this idea isn’t flushed out yet.

If the delta potential *is* sensed, then this means that particles like the electron must have non-zero size, otherwise the delta field that the particle sees would still be flat.  Then the only information where the sources were would come from a directional component resulting from the vector sum of source fields at that point (where the ring is located).  Current experiment appears to show that the base electron has no size, which means it cannot sense potential across the twist ring.  In addition, the notion that an electron is imeasurably small has a real problem with Heisenberg’s uncertainty relation.  It is fundamental to the twist ring theory that the electron does have physical size, the Lorentz transforms arise from that, and the E=hv quantization of the ring depends on that.  One way or the other, a determination has to be made whether we have a scalar or vector field inducing motion.  The sim model I have now depends on a scalar field (potential) and is qualitatively correct.

So, in summary–the question of whether motion results from travel through a potential is still possible–and unresolved.   More work ahead.

Agemoz

Sim Results Show Wrong Acceleration Factor

July 18, 2013

Well, it looked promising–qualitatively, it all added up, and everything behaved as expected.  But it’s a “close, but no cigar”.   The acceleration at each point should be proportional to 1/r^2, but after a large number of runs, it’s pretty clearly some other proportionality factor.  I’ve got some more checking to do, but looks like I don’t have the right animal here.  One thing is clear though–this model, which attracts and repels, is the first one that shows qualitatively correct behavior.  If twist rings have mass due to the twist distortion, this is the first model that shows it, even if the mass can’t be right.

So, I stepped back and ran through the list of assumptions, and see some flaws that might guide me to a better solution.  Many theories die in the real world because of the glossy effect, as in, I glossed over that and will deal with it later, it’s not a major problem.  I unintentially glossed over some problems with the model, and in retrospect I should have addressed them from the get-go.

First, twist rings (as modelled in my simulation) have a real planar component, but twist through an imaginary axis.  The twist acts as an E field in the real space and as a magnetic field in the imaginary space.  The current hypothesis is that the loop experiences different field magnitudes from the source particle, and this causes a curvature change that varies around the loop.  The part of the loop that is further away will experience less curvature, the closer part more curvature (curvature is a function of the strength of the magnetic field from the source particle).  This simulation shows that if that is the model, you do indeed get an acceleration of the ring proportionate to the distance from the source particle–and the acceleration is toward the source particle–attraction!  If you switch the field to the negative, you get the same acceleration away–repulsion.  So far, so good, and the sim results made me think–I’m on the right track!  I still think I might be on the right track, but the destination is further away than I thought.

First, as I mentioned, the sim results seem to show pretty clearly that the acceleration is not the right proportionality (1/r^2).  That might just be a computational problem or just indicate the model needs some adjustments.  But there are some things being glossed over here.  First, while the model works regardless of how many particles act as a source, there is always one orientation where every point on the ring is equadistant from the source particle–in this case, there is no variation in curvature.  The particle would have to act differently depending on orientation.  It could be argued that the particle ring will always have its moment line up with the source field, and so this orientation will never happen–fine, but what happens when you have two source particles at different locations?  The line-up becomes impossible.  OK, let’s suppose some sort of quantum dual-state for the ring–and I say, I suppose that is possible, some kind of sum of all twist rings, or maybe a coherence emerges depending on where the source particles are, but then we no longer have a twist ring.  In addition, the theory fixes and patches are building up on patches, and I’d rather try some simpler solutions before coming back to this one.  The orientation problem is a familiar one–it shoots down a lot of geometrical solutions, including the old charge-loop idea.

Here’s another issue:  I make the assumption that there is a “near side” and a “far side”, which has the orientation problem I just mentioned–a corollary to that is that it also could get us in trouble as soon as relativity comes in play since near and far are not absolute properties in a relativistic situation).  I then get an attraction by assuming the field is weaker on the far side and thus there is less curvature.  The sim shows clearly the repulsion acceleration away from the source when this is done.  Then I cavalierly negated the field and Lo! I got attraction, just like I expected.  But I thought about this, and realized this doesn’t make physical sense–a case of applying a mathematical variation without thinking.  This would mean that the field caused *greater* curvature when the twist point is further away (the far side).  Uhh, that does not compute…

While not completely conclusive, this analysis points out first, that a solution cannot depend on source field magnitude variation alone within the path of the ring.  The equidistant ring orientation requires (more correctly, “just about” requires, notwithstanding some of the alternatives I just mentioned) that the solution work even if all neighborhood points on the ring have exactly the same source field magnitude.  In addition, there’s another more subtle implication.  The direction a particle is going to move has to come from a field vector–this motion cannot result from a potential function (a scalar)  because within the neighborhood of the ring, the correct acceleration must occur even if the potential function appears constant over the range of the twist ring.

This is actually a pretty severe constraint.  In order for a twist ring to move according to multiple source particles, a vector sum has to be available in the neighborhood of the twist ring and has to be constant in that neighborhood.  The twist ring must move either toward or away from this vector sum direction, and the acceleration must be proportionate to the magnitude of the vector sum.  Our only saving grace is the fact that this vector sum is not necessarily required to lie in R3, possible I3–but a common scalar imaginary field of the current version of the twist theory is unlikely to hold up.

Is the twist field theory in danger of going extinct even in my mind?  Well, yes, there’s always that possibility.  For one thing, I am assuming there will be a geometrical solution, and ignoring some evidence that the twist ring and other particles have to have a more ghostly (coherent linear sum of probabilities type of solution we see in quantum mechanics).  For another, my old arguments about field discontinuities pop up whenever you have a twist field, there’s still an unresolved issue there.

But, the driving force behind the twist field theory is E=hv.  A full twist in a background state is the only geometrical way to get this quantization in R3 without adding more dimensions–dimensions that we have zero evidence for.   Partial twists, reverting back to the background state, are a nice mechanism for virtual particle summations.  We do get the Lorentz transform equations for any closed loop solution such as the twist theory  if the time to traverse the loop is a clock for the particle.  And–the sim did show qualitative behavior.  Fine tuning may still get me where I want to go.

Agemoz