Content

What are the required computer specifications?

The primary requirement is a stable internet connection, as the platform is entirely browser-based. All optimizations and simulations run on our secure, cloud-based High-Performance Computing (HPC) infrastructure, eliminating the need for expensive in-house hardware. However, for smooth post-processing and 3D visualization of results on your local machine, a dedicated (discrete) GPU is recommended. Results can be securely stored on the platform or downloaded locally, provided you have sufficient storage space.

Are the designs manufacturable?

Yes. You can incorporate specific manufacturing constraints directly into your case setup. ColdStream supports a wide variety of manufacturing techniques—including CNC milling, sheet metal forming, 3D printing, die casting, stamping, and injection molding—each with its own set of governing parameters. Because these constraints are applied mathematically during the generative process, we recommend that engineers critically review the final output to ensure all features meet real-world production standards.

Can the software only be used for electronic cooling systems?

No, ColdStream is a versatile platform capable of addressing a wide range of thermal and fluid dynamics challenges. It is highly effective for any application requiring thermal management or flow distribution, regardless of the industry. Current applications span across automotive, laser technology, aerospace, home appliances, medical imaging, and beyond.

How are the designs created on Diabatix Coldstream?

ColdStream utilizes a fully generative topology optimization approach. Unlike traditional shape optimization, topology optimization does not require an initial, pre-existing design to modify. Instead, it organically distributes material within a defined design region based on your specified objectives and constraints. This method allows for a much broader exploration of the design space, frequently yielding highly innovative and superior-performing geometries.

How can the resolution of the mesh affect the quality of the simulation results?

In Computational Fluid Dynamics (CFD), simulation accuracy is directly tied to mesh resolution. Finer mesh elements (higher resolution) can capture smaller, more complex physical flow phenomena, resulting in greater accuracy. However, this comes at the cost of increased computational requirements (processing cores, memory, and time). ColdStream streamlines this by automatically generating a mesh based on your chosen base resolution, dynamically adapting to both the geometry and the expected physical phenomena to deliver an optimal balance between accuracy and computational cost.

How can the resolution of the mesh affect the design?

Similar to simulation accuracy, mesh resolution dictates the quality of the generated design. A finer mesh produces higher-fidelity data, which the solver uses to evolve a more precise design. Additionally, the mesh resolution defines the minimum allowable feature size—the smallest physical structure (or gap) the solver can generate depends directly on the size of the underlying mesh cells. ColdStream automatically configures the design space resolution to strictly respect the minimum feature size criteria you provide.

Why does my design case take so long to complete?

Calculation times depend heavily on the case inputs, particularly the ratio between the overall design region volume and the specified minimum feature size. If a very small feature size is requested within a massive design region, ColdStream must generate a highly dense mesh to comply with those strict geometric constraints. This significantly increases the computational load. Allocating more credits to the case provides the solver with more computational resources, which decreases the total run time. For complex cases, our support engineers can help you determine the optimal feature size ratio.

Can I run supersonic cases on Diabatix Coldstream?

Currently, ColdStream operates on an incompressible flow solver, meaning supersonic flows are not supported. While liquids are virtually incompressible, gases can compress at high velocities. As a standard rule of thumb for forced or mixed convection cases, the Mach number should remain below 0.3 (approximately 110 m/s for air at sea level). Exceeding this threshold introduces compressibility effects that the current solver will not accurately capture.

Can Diabatix Coldstream handle external aerodynamic optimization problems?

While internal flow and thermal management are ColdStream's primary strengths, it is fully capable of external aerodynamic optimization. For instance, to minimize drag, you can simulate a virtual wind tunnel around your object and apply a powerLossMinimization objective between the inlet and outlet. Because the platform uses topology optimization, the design region must be strategically defined around the object to allow the solver space to add or remove material effectively. Our support team can assist you in configuring these specific setups.

How is turbulence taken into account on Diabatix Coldstream?

Turbulence is characterized by chaotic, swirling fluid motions known as "eddies," which occur across varying length scales. Because resolving every microscopic eddy (Direct Numerical Simulation, or DNS) requires immense computational power and is impractical for industrial use, ColdStream employs turbulence modeling.

Instead of filtering out only the smallest eddies like Large Eddy Simulation (LES) models, ColdStream utilizes Reynolds-Averaged Navier-Stokes (RANS) models. RANS models are the industry standard for complex engineering problems. They statistically average the turbulent fluctuations, allowing for coarser meshes and focusing computational resources on large-scale physics, multiphysics interactions, and complex industrial geometries. Coldstream has several different RANS models available:

KOmegaSST
KOmega
kEpsilon
RNGKepsilon
Laminar

What turbulence model should I choose?

Model selection depends on your specific flow regime:

kEpsilon: A robust and widely used model, excellent for fully turbulent flows and external aerodynamics. It performs best in free-stream areas away from walls.
kOmega: Superior for analyzing flow behavior close to walls, handling adverse pressure gradients, and predicting flow separation.
RNGkEpsilon: A refined variation of the standard kEpsilon model that accounts for multiple turbulence length scales, improving accuracy for swirling flows.
kOmegaSST (Default): A highly effective hybrid model. It utilizes the kOmega formulation near solid walls for accurate boundary layer resolution and seamlessly transitions to the kEpsilon formulation in the free stream.
Laminar: Strictly for flows with very low Reynolds numbers, where viscous forces heavily dampen out any chaotic movement, resulting in smooth, non-turbulent flow.

Is it possible to start from an existing design and modify it?

Yes. Because ColdStream uses topology optimization, it can actively transform a baseline geometry. Please refer to the Initialization section in the Case Type chapter.

How many objectives can be used in a single design optimization?

You are not limited to a single objective. ColdStream fully supports multi-objective optimization. To prioritize competing goals, you assign a weighting factor between 0 and 1 to each objective. These values do not need to add up to 1; the platform automatically normalizes the relative weighting during the calculation.

Can I restart an optimization from the same point after stopping it?

No. If an active optimization or simulation is manually aborted by clicking the 'Stop calculating' button, the intermediate data is permanently deleted. You will need to configure and submit a new case from the beginning.

How does the fanInlet/fanOutlet boundary patch work? Can I set my fan?

The fanInlet and fanOutlet boundary conditions simulate the physical behavior of a fan by dynamically altering the fluid momentum based on a specific performance curve. While ColdStream features a comprehensive library of commercial fan curves, you can easily define a custom fan. You must provide a tabulated curve with at least two data points mapping volumetric flow rate (m³/s) against total pressure loss (Pa). Ensure all inputs use strictly S.I. units and represent total pressure, not static pressure.

The mean outlet temperature is too low, can I trust the simulation results?

For every finished case, ColdStream outputs all key performance values to the tables tab. One of these automatically generated tables displays the maximum, mean, and minimum temperatures of each entity. These values are generated as follows:

maximum temperature ( $T_{max}$ ): mesh element with the highest temperature

minimum temperature ( $T_{min}$ ): mesh element with the lowest temperature

mean temperature ( $T_{mean}$ ): surface/volume average over the entire entity

$T_{mean} = \frac{1}{S}\int_S{T\space dS}$

$T_{mean} = \frac{1}{V}\int_S{T\space dV}$

Where:

S is the surface of an entity if that entity is a boundary

V is the volume of an entity if that entity is a region

T is the temperature distribution of an entity

From thermodynamics principles, we know that the theoretical mean outlet temperature of a coolant can be estimated using the following energy balance equation:

$T_{mean} = \frac{\dot{Q}}{\dot{m}\space c_p}+T_{inlet}$

Where:

$\dot{Q}$ is the amount of heat transferred to the fluid

$\dot{m}$ is the mass flow rate of the coolant

$c_p$ is the specific heat capacity of the coolant

$T_{inlet}$ is the inlet temperature of the coolant

In some cases, there is a small discrepancy between the mean temperature reported in the interface tables and the theoretical mean temperature calculated using the equation above. This difference primarily occurs because of poor fluid mixing within the bulk region of the heatsink. If the fluid does not mix well, the strict average may be skewed in the table view. A more precise representation of this mixed energy transfer is given by:

$T_{mean} = \frac{\int_S{\rho\space c_p\space T\space\vec{U}\cdot\vec{n}\space dS}}{\int_S{\rho\space c_p\space\vec{U}\cdot\vec{n}\space dS}}$

Where:

$\rho$ represents the density of the coolant

$c_p$ represents the specific heat capacity of the coolant

$\vec{U}$ represents the local velocity of the coolant on the outlet

$\vec{n}$ represents the normal on the outlet

$T$ is the temperature distribution on the outlet

This final equation evaluates the true amount of energy carried over to the coolant, independent of imperfect mixing. This accurate thermodynamic measure is available in ColdStream under the 'imbalance' table. The h_relImbalance value tracks all energy entering and exiting the coolant.

👍
In short
Yes, you can trust these simulation results as long as the h_relImbalance value for the fluid region is close to 0 and within acceptable margins, indicating a well-converged energy balance.

How can I interpret the intermediate custom design results?

Please refer to the Custom Design section in Design evolution chapter.