Behind Norstøy – Technology of a national tool for road traffic noise mapping

 

Rolf Tore Randeberg, Research scientist, SINTEF. Tel. 735 92668 / 416 70 206

 

 

The Norwegian Public Roads Administration (NPRA) has since 2004 commissioned the development of a new tool for calculation of (road) noise. Version 3.0 was released at the end of 2009, and supports the requirements of the new ”Støybygg 3” software for calculation of indoor noise.

 

Norstøy’s central calculation kernel, which calculates the point-to-point ground attenuation, is the Nord 2000 method. Compared to earlier methods, Nord 2000 has a very high computer power consumption. Hence, new procedures had to be developed to achieve a reasonable balance between calculation time and precision, when the method is used for large scale mapping.

 

The Norstøy software consists of a user interface and a calculation module, the latter developed by SINTEF. The calculation module consists of a central core, a number of calculation cells distributed on computers in the network, and a service which administers the tasks and the calculation cells.

 

The distribution of noise calculation on networked computers reduces the total computation time. To additionally reduce the total (CPU) computation time, further steps were required, as listed below.

 

             Advanced interpolation of receiver points. Noise contours are generated from an irregular grid of receiver points. This grid consists of a regular grid interspersed with points along constricting lines such as buildings and screens. The latter, non-regular points are calculated using the fast, but lesser accurate, Nordic method for industrial noise. The regular points are calculated using both Nord 2000 and the industrial noise method. The noise level difference between the two methods is then added to the non-regular points in the vicinity of the regular points.

 

             Significance testing of sources. The road sources are split into a large number of straight segments. These source segments are then tested for proximity and contribution to the total noise level at each receiver point, using the fast, but lesser accurate, industrial noise method. Only significant source segments are subsequently calculated using the Nord 2000 method.

 

             Reuse of reflected sound. For each source point along the road, subsources are created for each of the lanes of the road and for three source heights. For direct sound, the contribution from each of these subsources are calculated separately. However, for groups of subsources where the direct sound contribution is approximately equal, the reflection contribution is calculated only once, and the resulting reflection factor is applied to the reflection contribution from all subsources in the group.

 

The presentation will explain the technology and its influence on the resulting quality and calculation time of Norstøy.