Low Frequency Noise Prediction and Validation Study

Low frequency noise from mining activities has long been an issue for both regulatory bodies and industry. The impacts of low frequency noise (LFN) are subjective and difficult to both predict and regulate with current noise modelling methodology. This presents issues for legislation, as well as limitations in predictive modelling. This study has focused on the sub 200Hz component of noise or sound, which has been referred to from here on as LFN.

In NSW the Environmental Protection Authority Noise Policy for Industry (NSW EPA, 2017) has developed noise criteria to assess LFN impacts between the 10-160Hz frequency bands which when exceeded trigger a penalty for noise emitters which have licensed noise limits.  Measurement of noise impacts at these lower frequencies can be undertaken with modern sound level meters. However, conventional noise model predictions of mining noise impacts at sensitive receptors are typically limited to frequencies at and above the 31.5 Hz or 63 Hz octave bands. This means that lower frequencies cannot be modelled with certainty and subsequently assessed against LFN criteria.

This results in uncertainty when undertaking noise modelling for mine planning, particularly as to whether an LFN penalty may be triggered. Where a low frequency impact is identified in a mining assessment, areas of land acquisition can potentially double in area as low frequency penalties of 5dBA are applied to predicted noise impacts. The implications of this uncertainty on the requirement for land acquisition have a material impact on the mining approval process, community perception of impacts and project costs.

The objective of this study is to improve confidence in the modelling and prediction of low frequency noise impact from mining activities by:

• Better quantification of LFN emissions and prediction of LFN noise propagation.
• Improvements to modelling methodologies in the lower frequency (sub 63Hz) spectrum, using
• validation against measured LFN levels.

The CONCAWE calculation method as implemented in CadnaA (version 2019) allows prediction down to the 25Hz 1/3 octave band. Comparison of the total attenuation applied by the model from the source sound power level (LW) to the noise measurement location indicated that below the 80Hz 1/3 octave a consistent attenuation is applied at each 1/3 octave band.  To assess lower frequencies, an attenuation K value consistent with the 25 - 80 Hz model outputs was applied to 1/3 octaves from 6.3 - 20Hz.