At One Simulations we have a lot of experience with researching and assessing the local wind climate using CFD simulations. We have developed and checked an efficient and accurate method through a sensitivity analysis. In a sensitivity analysis, the same simulation is run several times and only one basic parameter is changed to see what effect this has on the result.
Wind comfort is assessed using CFD simulations to quickly and accurately provide insight into the wind climate before a new building is built. One Simulations has developed an efficient method which can be applied for assessing the wind climate according to various standards such as the NEN8100 or the Lawson Criteria. Extra knowledge has been gained for wind studies and the applied simulation method by means of a sensitivity analysis. The results of the sensitivity analysis are provided below.
The sensitivity analysis has been performed for the following five parameters:
- Size of the modeled area
- Refinement of the number of wind sectors in the wind rose
- Calculation grid refinement
- Time dependence (RANS / URANS)
- Reference speed
The results of the sensitivity analysis per parameter are shown below. A brief analysis of the results is provided below the figures. At the tab “Assessment of wind climate” a description of what the results show and the legend are provided.
Results sensitivity analysis
- According to the NEN8100 for testing the local wind climate, the buildings within a radius of 300 meters around the building of interest must be included in the CFD simulation. The expectation was that the radius and therefore the amount of buildings will influence the result. That is why the effect of a radius of 300, 400 and 500 meters has been investigated.
- The CFD results for the different radii show a clear effect on the wind climate. In particular, the difference in the wind climate between the radius of 300 m and 400 m shows that the extra buildings (on the south-west side) cause the air to accelerate less.
- It is concluded that the size of the modeled area can have a lot of influence on the local wind climate. It is advisable to check for each wind direction whether there are buildings upstream that can influence the air flow. If so, then the chosen radius of the modeled area should be adjusted accordingly.
- In NEN8100, the wind rose is divided into twelve wind sectors, each consisting of 30 degrees.
- The effect of dividing the wind rose into 36 wind sectors, each consisting of 10 degrees, is investigated.
- The wind roses above represent the percentage of wind from the relevant wind direction in an average year. The refinement of the number of wind sectors already shows a large difference between the wind roses.
- When comparing the two results, it can be found that adding wind sectors affects the wind climate.
- Refining the number of wind sectors ensures that the effect of streets that are in line with a common wind direction is less significant.
- The addition not only yields a more realistic result, but can also ensure a more positive wind climate.
- The calculation grid has been refined several times in the area of interest step by always halving the element size.
- The figures above show the wind factor (local speed relative to reference speed) for wind direction 270 with the corresponding scale bar in the last figure.
- As can be seen from the figures, the effect of the applied refinement is minimal on the wind factor and, therefore on the overall wind climate. The extra computational costs do not compare to the marginal difference of the grid refinement.
- If a very accurate wind climate is required in narrow areas such as balconies or a small courtyard, it is recommended to refine the calculation grid at that location.
- The air flow in a built environment fluctuates over time. Mainly behind and on a corner of a building, vortices are formed. Nevertheless, it is common to run simulations with the Reynolds-Averaged Navier Stokes (RANS) equations, where the result is a snapshot of the flow field.
- The influence of the application of Unsteady RANS (URANS) equations, in which a time component is included, has been investigated and, therefore, large flow fluctuations can be averaged over time.
- The result of a URANS simulation is therefore a time average of the flow field, while the result of a RANS simulation is a snapshot.
- The results between these two methods are comparable and no significant difference was found. A RANS simulation can be used for wind climate research.
- In order to assess the wind climate, a wind factor is calculated with CFD simulations. The wind factor indicates how much the air speed accelerates or slows down compared to a reference speed measured in the free air flow at a height of 10 meters.
- In this subject the effect of varying the reference or base speed is investigated. Despite the fact that the Reynolds number is already relatively high, the effect of changing the base speed between 2,5 and 10 m / s was examined.
- The CFD results show that varying the base speed produces small differences, but the conclusions of the assessment of the wind climate do not change.
In this study, the simulation results were analyzed on the basis of the wind factor and wind nuisance according to the Dutch standard NEN8100.
- The wind factor indicates how much the air speed accelerates or slows down compared to the reference speed.
- Wind nuisance is defined according to NEN8100 and is shown on the basis of the local amount of exceeding the threshold value for wind nuisance (5 m/s) over an average year. The results in the analysis were checked in a radius of 150 m around the new-built development.
All results, except for the tab “Calculation grid refinement”, show the wind climate based on wind nuisance. The above scalebar, defined according to NEN 8100, has been applied.
A separate scalebar is displayed under the heading “Calculation grid refinement”.