Discussion Forum

Due to the nonlinear fluid nature of the navier stokes equations, and the finite representation of discrete approximations, ie numerical simulation, it is impossible to predict plasma flow beyond some limit into the future, and it becomes more uncertain as time progresses. Chaos theory (for an introduction see James Gleick, or Mandelbrot) means that the lyapunov exponents, the rate at which neighbouring infinitessimal cells diverge on a chaotic or smooth attractor, show that we might be able if we knew the attractor shape in phase space to compute bounds to our errors, but without a kalman filter or real time feedback like camera shots or measurements of the real system, we will not be able to keep it on track. This is the fundamental problem with predicting campfires and fusion.
If we have laminar flow, we can predict to a certain extent and quite acceptably, until we take into account material imperfections in wood, sap, structural imperfections and rate of change of structure during burn. a natural gas nozzle flame with laminar flow can be well modelled in the ideal, however with shifting wind patterns, unless we are measuring these and feeding them back in with negative feedback to our kalman predictor, we will not be able to predict such a flame either. There are also quantum limitations, that can't presently be computed for fusion or campfires, there is quantum noise at some level, or is it noise? There is also measurement noise, quantization noise and sensors not being quite as accurate over the whole range of measurment as they might be .... these are fundamental problems to prediction and control of fusion or campfires.
just food for thought...... any comments...... how has comsol tackled this problem?