Comparison of Kilde and NORD2000 rail noise prediction methodologies

摘要

In Australia, predictions of rail noise are most commonly conducted using the Kilde Rep 130 methodology, which has been twice superseded. The most recent version of the Nordic methodology is NORD2000. This paper compares the practical results of the Kilde and NORD2000 methodologies in a range of simple scenarios which have been selected to isolate specific effects, such as source directivity, meteorological effects, shielding losses and reflections. A test case representing complex real-world conditions has also been studied. Significant differences have been observed in the results of the two methodologies, with NORD2000 generally predicting larger propagation losses. Determining which methodology provides more accurate predictions would require comprehensive measurements in tightly controlled conditions. However, it appears that Kilde uses a more conservative approach, to provide 'margin for error' due to its simplistic calculations.