About

Why we exist

Industry tends toward one of two extremes: either no analysis is commissioned at all, or one is commissioned and nobody knows how far it can be trusted. Both end in the same place — the decision gets made on intuition even though a number is sitting on the table.

DYNABLA was founded to close that gap. Our job is not to produce numbers, but to produce numbers whose reliability is known.

What we do

We work wherever there is flow and heat: internal and external flow, multiphase, combustion, turbomachinery, conjugate heat transfer, aerothermodynamics, heat exchangers and electronics cooling.

The equations do not change from one industry to the next — mass, momentum and energy are conserved in a battery pack exactly as they are in a turbine. What changes is everything chosen around them: which physics enters the model and which stays out, which turbulence, combustion or multiphase model suits the regime, which solver and which order of discretisation the problem tolerates, and whether the boundary conditions really represent the installation.

The engineering is in those choices, and that is where a result is won or lost: the same geometry, solved with the wrong model or fed the wrong inlet condition, returns a number that looks finished and is wrong.

How we work

Every job goes through five steps — define, model, solve, validate, report. We do not change the method to suit the project; what changes is the physics model and the acceptance criteria.

Assumptions are written down. What was not modelled appears in the report, together with where it affects the result. The reader also learns what we do not know.

Grid independence is demonstrated. At least three meshes are run and the sensitivity of the result to the mesh is quantified. "It converged" is not on its own a measure of quality.

Deviations are not hidden. Where simulation and measurement differ, the size, the likely cause and the effect on the decision are stated plainly.

Uncertainty is given. The result is delivered as a band, not a single number, so you know which side of it is the safe side.

The report is the deliverable. On analysis work, model files and solver setups stay with us; what goes to you is the engineering content in full — the numbers behind every figure, where the bottleneck is and why it is there, and which change buys what. The report is written so the decision can be taken from it, and defended later, without opening a solver. Optimisation and automation work is the exception: there the pipeline we build is handed over, with run notes.

What we decline

We do not present an unverifiable result as if it were verified. Where no measurement exists we say so, and set the uncertainty accordingly.

We do not leave a vague scope vague. Before anything is modelled we sit down with you — often several times — and work out what the question really is, which geometry has to be in the model and which can come out of it, and what the boundary conditions should be. That is where our engineering experience earns the most, and it is the cheapest hour in the project.

Confidentiality

No client data — CAD, test results, operating conditions — is shared with third parties, used as a reference, or published for promotional purposes without written consent. On request we sign an NDA before the first call.

Every image on this site is a computation of our own; no client or third-party data has been used.