We proposed here some validation of the DEUFRABASE by comparison with measurements:
- Measurements are the same ones than those detailed at this page.
- These measurements are compared to the DEUFRABASE calculation:
- using built-in road pavement (i.e. the mean road pavement spectrum for PC and HT that are included in the DEUFRABASE),
- using user spectrum (i.e. with the corresponding spectrum that have been measured in situ for PC and HT).
The table below presents the comparison in terms of LA,eq,1h. Only the 12 road configurations (over the 47 configurations) with impedance values equivalent to those that are introduced in the DEUFRABASE have been considered). DEUFRABASE deviations from measurements are given between parenthesis.
(number of vehicles and % of HT)
|BBDr 0/10||459 vehicles, 14% HT||65.1||66.0 (-0.9)||64.6 (+0.5)|
|BBDr 0/12||503 vehicles, 38% HT||70.6||69.7 (+0.9)||70.3 (+0.3)|
|BBDr 0/14||1403 vehicles, 17% HT||74.5||73.1 (+1.4)||73.7 (+0.8)|
|BBSG 0/10||244 vehicles, 32% HT||67.5||69.4 (-1.9)||67.2 (+0.3)|
|BBSG 0/14||943 vehicles, 14% HT||75.0||75.1 (-0.1)||74.7 (+0.3)|
|BBTM 0/6 type 1||680 vehicles, 15% HT||69.5||69.3 (+0.2)||68.3 (+1.2)|
|BBTM 0/10 type 1||381 vehicles, 14% HT||69.6||70.1 (-0.5)||69.4 (+0.2)|
|BBTM 0/14||1610 vehicles, 16% HT||76.8||77.8 (-1.0)||79.0 (-2.2)|
|BBUM 0/10||549 vehicles, 37% HT||73.3||72.6 (+0.7)||75.8 (-2.5)|
|BC||411 vehicles, 50% HT||73.6||75.1 (-1.5)||74.1 (-0.5)|
|ECF||1767 vehicles, 9% HT||75.9||76.0 (-0.1)||77.3 (-1.4)|
|ES 6/10||666 vehicles, 19% HT||74.9||73.8 (+1.1)||74.2 (+0.7)|
The two graphics below shows the distribution of the deviations between the DEUFRABASE and the measurements.
As expected, when considering the 12 road configurations that have similar impedance discontinuity values between in situ measurements and DEUFRABASE configurations, the deviation is reduced. In addition, the deviation is also reduced when considering the measured road spectrum.
More generally, in view of the results for all the 47 road configurations (i.e. which include configurations for which the impedance discontinuity can be different from the DEUFRABASE configurations), it seems not necessary to consider measured spectrum to obtain a significant enhancement of the results.
In conclusion, even with 'mean values' (both for the impedance and the road spectrum), the DEUFRABASE can predict LA,eq,1h with a good accuracy (in the present case, 87% of comparisons are less than 2 dBA). One can also remark that the DEUFRABASE tends to overestimate the real values, which goes well in the direction of an usual precautionary margin.