General Advice

It is highly recommended that you monitor the progress of the simulation closely. Check your results during the simulation. In cases where the simulation takes multiple days, check the results at least once a day using ParaView and make use of the available info files.

Be aware that cases where high pressurization can occur (pumps, compressors, hydraulic simulations, and so on) are not ideally suited for the weakly compressible SPH method. While theoretically possible to do, build up of high pressure areas can lead to extreme forcing (via pressure gradient) that time step by default does not account for. This commonly leads to CFL violations and causes simulation instabilities. The only way around this is to set the time step through trial and error, often resulting in computationally expensive simulation.

Particle Resolution - how to decide dx

Resolution choice is typically established by answering the following questions:
  • What is the smallest fluid structure that I would like to resolve?
  • What is the smallest solid geometry feature that I would like to resolve?

By resolving, it is implied to use a minimum of three to five particles across that fluid/solid structure. This by no means implies high accuracy, but a basic minimum of information that the result will provide. For analogy sake, ask yourself if you would consider the structure resolved if you used a finite-volume CFD code and had five cells across a structure. That is the approximate level of accuracy you can expect from having equal number of particles.

As in all CFD, the more resolution the better.

Physics

Interaction Scheme
Choose whether to use Riemann or Weighted interaction schemes (interactionscheme). Weighted is the default choice and is capable of handling violent and chaotic multiphase flows. Riemann results in smoother pressure gradients and less breakup of sprays and jets but will experience local instabilities with more violent flows.
When using Weighted interaction, Transport Velocity and Free Surface should be set in simulation parameters. Riemann interaction offers some advanced controls, but these should not require adjustment.
Surface Tension
Choose if surface tension is active (surften_model) and which model to use. By default, surface tension is disabled, since it adds computational cost.
When creating a multiphase simulation, ADAMI should be chosen. This uses a physical value of surface tension coefficient (surf_ten) which can be calculated.
For single-phase simulations, SINGLE_PHASE or TARTAKOVSKY should be chosen. The Tartakovsky model has the advantage that a physical value of surface tension coefficient (surf_ten) can be calculated and used.
Restriction: Surface Tension requires reference curvature (ref_curv) to be set in Domain parameters, along with surface tension coefficient (surf_ten) in Material (phase) parameters.
Adhesion
Specifies whether the Akinci wall adhesion model (adhesion_model) is active. This is used in conjunction with surface tension to define how the fluid interacts with the wall and the contact angle of a fluid droplet on the wall.
Restriction: Adhesion Coefficient (adhesionCoeff) must also be set in Material (phase) parameters when using the adhesion model.
Temperature
Activates energy_transport so that the energy equation is solved by nanoFluidX for conduction and convection heat transfer.
Restriction: When Temperature is used, initial temperature (temp_0), evolve temperature flag (evolve_temp), specific heat capacity of the material (heat_cap), and heat conduction coefficient of the material (heat_cond) should be set in Material (phase) parameters. Temperature and Flux output should also be requested in Results Request.
Large negative pressures
Large negative pressures cannot be accurately handled by the current weakly compressible formulation without total particle volume preservation.

Simulation Parameters

Solution Parameters
All parameters have descriptive name and solver keywords. Most default values will give good results, although they are typically conservative. Solution Parameters: Basic Settings are most critical to set.
Result Requests
Typically, the main options to be changed are the Output Time Interval and Phase Information Output Frequency. Output Time Interval is set to give 20 outputs/second by default, which gives a smooth animation. It is advised to aim for 100-300 outputs for a simulation as a good compromise between animation quality and file size. For longer runs and larger models Particle Output Frequency can have a big impact on the size of the results file generated.