Simulation Setup

This guide walks you through configuring a CFD simulation in Gradient Dynamics Studio. Simulation setup is done in the Simulation tab of a CFD project.

Prerequisites

Before setting up a simulation, you need a mesh. You can:

1. Generate a mesh in a Meshing project and import it

2. Use the mesh generated within the same project

3. Use an existing mesh from a previous project

Setup Overview

A simulation configuration consists of five parts:

1. Solver type — Density-based (recommended) or pressure-based

2. Turbulence model — The physics model for turbulent flow

3. Boundary conditions — What happens at each surface (inlets, outlets, walls)

4. Solver settings — Numerical algorithm and convergence parameters

5. Run parameters — Maximum iterations, convergence criteria

Step-by-Step Configuration

1. Select Solver Type

Choose the solver formulation from the dropdown at the top of the Simulation tab. The available solver types are:

Solver	Description
Density-Based (recommended)	Compressible, explicit — best GPU performance for all standard CFD
Pressure-Based	Incompressible, implicit, segregated — SIMPLE, SIMPLEC, PISO, or PIMPLE algorithms
Coupled	Incompressible, implicit, coupled — solves momentum and pressure simultaneously
FSAC	Incompressible, hybrid explicit/implicit — fractional step with artificial compressibility
ACM	Incompressible, explicit — artificial compressibility method for GPU-efficient incompressible flows

Density-based is the default and recommended choice for the vast majority of applications. It is optimized for the GPU-native Cartesian AMR mesh and delivers the best performance across both compressible and incompressible flow regimes (via low-Mach preconditioning). See Solver Settings for a detailed comparison of all solver types.

Use the other solvers only if your application specifically requires an incompressible formulation or a particular coupling strategy.

2. Select Turbulence Model

Choose the turbulence model from the dropdown. The default is k-ω SST, which works well for most external and internal flow applications.

Available models depend on your subscription tier — see Turbulence Models for details on each model.

3. Set Boundary Conditions

Studio auto-detects boundary types from your surface names:

• Surfaces named inlet → Velocity inlet

• Surfaces named outlet → Pressure outlet

• All other surfaces → Wall (no-slip)

You can override any auto-detected condition. See Boundary Conditions for all available types and parameters.

4. Configure Solver Settings

For most cases, the defaults work well:

• Density-based: SSP-RK3 explicit time marching, CFL 1.5 with auto-ramping, AUSM+ flux scheme, MUSCL reconstruction

• Pressure-based: SIMPLE algorithm, model-appropriate relaxation factors, AMG preconditioner

• Coupled: Block-coupled momentum-pressure system, AMG preconditioner

• FSAC / ACM: Explicit time marching with artificial compressibility, CFL-controlled

See Solver Settings for advanced tuning options.

5. Set Run Parameters

Parameter	Description	Typical Value
Max iterations	Maximum solver iterations	500 – 2000
Convergence criterion	Residual threshold for completion	1e-5 (density) / 1e-4 (pressure)

Tip

Start with 500 iterations. If residuals haven’t converged, you can extend the run. Most RANS simulations converge within 300–1000 iterations.

Credit Cost Estimation

Before running, Studio shows the estimated credit cost based on:

• Mesh size (number of cells)

• Expected runtime

• GPU resources required

The estimate is shown when you click Run Simulation, before confirming.

Quick Setup with the AI Assistant

For the fastest setup, use the AI Assistant:

“Set up a density-based simulation at 30 m/s with k-omega SST”

The assistant will:

1. Select the density-based solver

2. Select k-ω SST turbulence model

3. Set inlet velocity to 30 m/s

4. Configure outlet as pressure outlet (0 Pa)

5. Apply default solver settings

6. Ask for your confirmation before running