Weird noise propagation close to the ground

[sachabaclet]

Hi!

I am using NoiseModelling 3.4.1, and I have noticed a weird behavior of the noise propagation in a specific situation.

I am trying to evaluate noise levels on a façade at different levels. To test this possibility, I have created a very simple example with 2 parallel buildings, a source point (vehicle) between them, and a grid of receivers (1x1 meter grid) on the façade of one of the buildings.

Here is the configuration of the buildings (orange), source (pink) and receivers (green) as seen from above:

The source is 0.5 m above the ground, the receivers are 0.5 m from the façade. The receivers start at 1 m above the ground, up to 9 m.

The parameters of the NoiseModelling simulation are as follows:

    {"reflections": "3",
    "skip_lden": "true",  
    "skip_lnight": "true",  
    "skip_levening": "true",  
    "skip_lday": "false",
    "table_sources": "LW_ROADS",
    "threads": "16", 
    "max_src_dist": "200", 
    "vertical_diff": "true", 
    "table_receivers": "RECEIVERS", 
    "table_buildings": "BUILDINGS", 
    "export_source_id": "true",
    "horizontal_diff": "true", 
    "max_reflection_dist": "200"}

The distance between the buildings is 20 meters; I've made sure that the maximum propagation distance is much larger than the dimensions of this area.

Here are the resulting sounds levels on the grid of the façade:

As you can see, there is a "discontinuity" of sound levels on the row of receivers that is 1 meter above the ground. They have a lower sound level than the row that is right above them. It looks almost as if those receivers received fewer reflections than the other receivers.

With other source positions, I usually don't encounter this issue. For instance:

Do you have any idea what could be causing this? Is there a height "threshold" for reflections?

Thanks for your help!

PS: I could post the shapefiles if you'd like to reproduce the issue.

[pierromond]

In the CNOSSOS-EU model, you sometimes have some discontinuity in the equation as the next one for example. It depends from a ratio between zs, zs and the distance between sources and receivers. It is likely that your problem is related to this point of the method. This would have to be checked in more detail.

You can try to change the G factor of your floor, or the Gs factor above our sound source to check if it is related.

Keep in touch !

"When the source and receiver are close dp≤ 30(zs + zr), the distinction between the type of ground located near the source and the ground located near the receiver is negligible. If the receiver is very close to the edge of the platform, an absorbent ground receiver side should not be considered." p 86/180

or

[sachabaclet]

Hi Pierre, Thanks for your reply! This explanation sounds plausible.

I have not had time to check whether changing the ground has an effect, but I have also read that CNOSSOS-EU should not be used with receivers that are placed less than 2m above the ground, so I guess I should change my grid in the first place.