How Much Uncertainty in CFD Results Is Acceptable in an Investment Decision Concerning a Propeller Nozzle?
Author: Jeroen Berger • Publication date:
Uncertainty in Computational Fluid Dynamics (CFD) results rarely presents a problem as long as the analysis is used to explore design directions. At that stage, the calculation helps compare variants, reveal system sensitivities and better understand why a propeller nozzle profile behaves in a particular way under certain conditions.
The meaning of that uncertainty changes, however, once the same analysis becomes the basis for an investment decision. At that point, the role of the calculation shifts from a tool for insight to the substantiation of a design choice, an order or an implementation schedule.
The vulnerability rarely lies in the bandwidth itself. It only arises when that bandwidth becomes large enough to overturn the final decision. In every simulation, modelling assumptions, mesh generation and the selected operating points leave their mark on the outcome; that is unavoidable. The relevant question is therefore not whether these factors influence the result, but whether that influence can reverse the decision.
When a limited change in assumptions or numerical setup can already reverse the conclusion, what is missing is not another decimal place in the calculation but decision margin in the comparison. This applies both to the ranking between nozzle profiles, such as 19A and 37, and to the preceding decision between an open propeller configuration and a nozzle installation.
Acceptable Uncertainty Depends on the Type of Decision
The acceptable degree of uncertainty is determined not only by the analysis itself, but also by the type of decision involved. The same uncertainty margin may be workable in one project and too wide in another.
A comparison between standard reference profiles within a known installation situation usually requires less detailed certainty than a project in which a project-specific optimised nozzle is being developed. In the latter case, design and manufacturing effort are greater and later corrections are often more difficult.
The project phase also matters. In an early design exploration, a broader bandwidth may be acceptable as long as additional variants can still be investigated. As a project moves towards ordering, production and docking schedules, it becomes more important that the difference between alternatives remains sufficiently robust to support capital exposure and downtime risk.
Ranking Carries More Weight Than Absolute Values
In practice, investment decisions often show that the relative ranking between variants carries more weight than the absolute values of the calculated effect.
When two configurations are close to each other at one operating point and the order reverses under limited variation, the uncertainty margin may dominate the outcome. In such a situation, a single optimum says little about the behaviour of the system.
Decision certainty arises when the ranking between variants remains intact across several representative operating points and does not shift materially under realistic variation of model settings or boundary conditions. Absolute values may vary; as long as the ranking remains stable, the decision rests on hydrodynamic behaviour rather than on the chosen numerical setup.
Keep the Comparison Basis Constant
To assess that robustness, the comparison basis itself must be clearly defined. Unnecessary uncertainty arises when the comparison question shifts unnoticed.
An assessment based on required shaft power at equal vessel speed, for example, answers a different technical question than an assessment based on thrust at equal rotational speed. When such definitions are mixed, dispersion arises that does not result from flow behaviour, but from the chosen interpretation of the comparison.
For that reason, it must be clear in advance which quantity is kept constant, and that comparison basis must be applied identically to all variants. Only then does the remaining bandwidth say something about predictive uncertainty rather than about a shifting question.
Identify the Dominant Source of Uncertainty
Even with a fixed comparison basis, the question remains which factor in the calculation has the greatest influence. CFD results may be sensitive to mesh generation around the nozzle and aft ship, the selected turbulence approach, propeller modelling or the choice of operating points.
Which factor is most decisive depends on geometry and loading regime. An investment decision therefore becomes stronger when the dominant source of uncertainty in the configuration is explicitly identified in advance and it is then examined whether the ranking between variants is not determined solely by that factor.
The objective is not to eliminate all uncertainty, but to understand where the comparison reaches its tipping point.
Let the Operating Points Reflect the Operating Profile
The interpretation of uncertainty is directly linked to the selected operating points. An investment in a nozzle is rarely justified by one ideal operating condition.
When a vessel has a broad operating profile, a small modelling uncertainty at one point may translate into a larger spread in realised effect over the year. Uncertainty becomes manageable when the calculated operating points match the vessel’s actual operating profile.
An otherwise numerically consistent simulation may still provide a weak basis for decision-making if it analyses a boundary condition rather than the dominant operating profile.
Model Uncertainty and Installation Tolerances Belong Together
Uncertainty does not end with the numerical model itself. It carries through into the physical realisation of the installation. Tolerances in centring, tip clearance and alignment influence the resulting flow pattern, particularly in retrofit projects within existing installation space.
An advantage that only remains intact under very specific geometric boundary conditions may therefore be vulnerable from an investment perspective. A more moderate advantage that is less sensitive to practical tolerances often aligns better with limiting downtime and rectification risk.
When Classification Is Involved
When a nozzle modification becomes part of an assessment by a classification society, the emphasis shifts. Transparency regarding assumptions, boundary conditions and sensitivities then becomes essential, always subject to acceptance by the relevant society and, where applicable, the flag state.
In that context, traceability often carries more weight than one absolute advantage.
Conclusion
Acceptable uncertainty in an investment decision concerning a nozzle exists when the ranking between variants remains intact across representative operating points, when the dominant sensitivities have been identified in advance and when the expected effect proves robust against reasonable variation in model setup and installation tolerances within the same vessel configuration.
In that case, the decision rests on robust system behaviour rather than on a fragile numerical outcome.
This Article Within the Series
Within Propeller Nozzle: Design and Performance Validation, this article shifts the focus from methodological consistency to the question of how much uncertainty in CFD results remains compatible with an investment decision.
Where the preceding article How Can You Identify That a CFD Comparison of Propeller Nozzles Is Methodologically Inconsistent shows when a difference may originate from the calculation framework rather than from profile behaviour, the focus here is on when the remaining bandwidth still leaves sufficient decision margin to support a profile choice, an order or an implementation step responsibly.
The next step in the series shifts from uncertainty assessment to the substantiation of performance claims. In When Is a Performance Statement for a Propeller Nozzle Demonstrably Supported it is elaborated under which conditions an observed advantage remains sufficiently stable across representative operating points to qualify as a traceable performance statement.
Those who want to translate this methodological analysis into a concrete vessel configuration will find the practical application in Propeller Nozzle for Ships. There, geometry, operating profile, reference profiles such as 19A and 37 and project-specific design alignment come together in a traceable nozzle configuration for newbuild and retrofit.