Turbulence & Flow Optimisation

Highly specialised requests are first checked for feasibility from this information. Please do not send confidential data through the public form; describe it first and agree secure transfer afterwards.

1. 01Enquiry and requirement assessment
2. 02Technical scoping and proposal
3. 03Execution
4. 04Quality review and delivery, with an agreed round of revisions

Read the full project workflow.

In depth: how the work is done

1. Comparing turbulence models for the problem at hand

   Different models behave differently for separation, swirl, jets and strong pressure gradients. A comparison study runs the baseline case with several models, compares predictions of key quantities with data where available and identifies which features are robust across models and which are model-dependent. Differences between models give a practical estimate of turbulence-model uncertainty and prevent overconfidence in any one result.

2. Defining the objective and design variables

   Optimisation requires a clear objective, such as minimum pressure loss at a given flow rate, and constraints, such as maximum size, minimum outlet uniformity or manufacturing limits. Design variables, for example parameterised wall shape, vane angles or inlet positions, must allow a meaningful range of designs without producing impossible geometry. A poorly formulated problem produces unusable optima, so we invest effort in the formulation before computing.

3. Surrogate-assisted and gradient-based search

   CFD runs are expensive, so direct search over many designs may be impractical. Surrogate models fitted to a designed set of simulations allow broad exploration at low cost, with new simulations added where the surrogate is uncertain or promising. Adjoint-based gradient methods can compute sensitivities for many variables at the cost of a few solves and are powerful for shape optimisation, though they require suitable tools. We choose the approach according to the number of variables and available software.

4. Flow-control and mixing studies

   Beyond shape, flow-control devices such as vortex generators, ribs or injection can reduce separation or enhance mixing. Parametric studies of such devices examine their placement and size, and measure benefits against added losses. For mixing, metrics such as the coefficient of variation of concentration downstream quantify performance. Unsteady simulation may be required if the control mechanism is inherently unsteady.

5. Verifying the optimised design

   The final candidate is re-evaluated with a refined mesh, an alternative turbulence model and, where possible, an experiment. Robustness to off-design conditions and manufacturing tolerances is tested, because a design that performs well at a single point but degrades rapidly nearby may not be valuable. The report states the gain achieved with its uncertainty and notes which conclusions depend on the model used.

Common pitfalls this service helps avoid

• Optimising against model error. Optimisers can exploit weaknesses of a turbulence model, producing gains that do not appear in reality.
• Poorly parameterised geometry. Too few parameters limit the improvement; too many make the search impractical or produce unmanufacturable shapes.
• Ignoring off-design performance. Designs tuned to one operating point may perform badly at neighbouring conditions.
• Treating small gains as significant. Improvements smaller than numerical and model uncertainty should not be claimed.

What good reporting looks like

The report states the objective, constraints and design variables, the baseline performance, the turbulence-model comparison, the optimisation method and the number of simulations, the results with the improvement and its uncertainty, the verification of the final design with refined settings and robustness checks, and the limitations. It documents parameterisation and settings to allow reproduction, and it makes no claim about performance in real hardware that has not been tested.

An illustrative scenario

Hypothetical example written for explanation. It is not a client case, a result or a testimonial.

Imagine a researcher aiming to reduce pressure loss in a diffuser. Three turbulence models are compared on the baseline and agree on the main flow structure but differ on the extent of separation. A surrogate built on a designed set of wall shapes suggests a shape with reduced separation. Re-evaluation of the candidate with a refined mesh and a second turbulence model confirms a smaller improvement than first predicted. The report presents the smaller figure as the conservative estimate and recommends testing a prototype.

Questions to settle before you begin

• What is the objective, and what constraints apply?
• Which geometric features can I vary and manufacture?
• How will I check that gains are not an artefact of the turbulence model?
• What off-design conditions must the design handle?
• Can I test a prototype to verify the result?

Key terms

Turbulence model

A set of equations that represent the effect of turbulent fluctuations on the mean flow.

Adjoint method

A technique for computing sensitivities of an output to many design variables efficiently.

Surrogate model

A fast approximation fitted to simulation results.

Parameterisation

Describing a geometry with a limited set of variables.

Coefficient of variation

Standard deviation divided by mean, used as a measure of mixing uniformity.

Working together on Computational Fluid Dynamics (CFD) projects

CFD engagements start with a flow-physics review: expected regime, quantities of interest, available geometry and the measurements that could validate results. A coarse pilot simulation can check set-up and cost before the full study. We ask for geometry in a standard exchange format or dimensioned drawings, fluid properties, boundary conditions and any test data. Reports include mesh and model sensitivity so that you can see how much the answer depends on numerical choices. Results are presented as engineering quantities with units, and limits of validity are stated explicitly.

Few projects need a single service in isolation. These are the services most often combined with this one.

• CFD Simulation Services

  Computational fluid dynamics studies of internal and external flows, thermal effects and mixing, built on an explicit statement of physics, boundary conditions, turbulence treatment and validation basis.

• Engineering Optimisation

  Formulation and solution of optimisation problems in engineering design and operation: defining objectives, variables and constraints carefully, then choosing an algorithm suited to the problem structure.

• Multi-Objective Optimisation

  Optimisation with conflicting objectives, such as cost, weight and performance, producing Pareto-optimal trade-off sets rather than a single answer.

• Bayesian Optimisation

  Sample-efficient optimisation of expensive or noisy black-box functions using probabilistic surrogate models and acquisition functions.

• Topology & Design Optimisation

  Computational determination of material layout and form within a design space to meet performance targets under constraints, with attention to manufacturability.