4.10 poission_correction

• description: This command requests correction of the electric fields so that they satisfy Poisson’s equation. It is recommended in order to increase the accuracy of the field solutions. Without it, numerical errors tend to build up that correspond to fictitious concentrations of static charge on the grid.
• example:
  &poisson_correction  
      start_time = 1e-9,  
      step_interval = 32,  
      accuracy = 1e-4,  
      error_charge_threshold = 1e-15  
  &end

  This requests that Poisson correction be performed every 32 simulation steps starting 1ns after the start of the simulation. The fractional accuracy of the Poisson solver is 10^-4. It is invoked only when the amount of fictitious charge exceeds 0.001 pC.

• synopsis and defaults:
  &poisson_correction  
      double start_time = 0;  
      long step_interval = 0;  
      double accuracy = 1e-6;  
      double error_charge_threshold = 0;  
      long maximum_iterations = 1000;  
      long verbosity = 0;  
      double test_charge = 0;  
      double z_test_charge = 0;  
      double r_test_charge = 0;  
      STRING guess_type = "none";  
  &end

• details:
  • start_time: Time in seconds at which to begin Poisson correction.
  • step_interval: Interval between corrections in units of the simulation step.
  • accuracy: Fractional accuracy of the Poisson solutions.
  • error_charge_threshold: Amount of net error charge that must be present for Poisson correction to actually take place.
  • maximum_iterations: Maximum number of iterations of the relaxation loop for the Poisson solver.
  • verbosity: Flag requesting informational output about Poisson correction.
  • test_charge, z_test_charge, r_test_charge: For developmental use only.
  • guess_type: The model to use for the initial guess to start the Poisson iteration. Possibilities are, “none”, “line-charge”, “point-charge”, and “zero.” It is recommended that “none” be used.