Abstract
The spectral characteristics of turbulent combustion noise is dictated by the temporal correlation of heat release rate fluctuations, which has not received sufficient attention in prior studies. In this work two temporal correlations, the two-point space-time correlation of locally fluctuating heat release rate and the two-time correlation of the overall change of heat release rate fluctuations within the flame brush, and their roles in modelling combustion noise spectrum are investigated by analysing direct numerical simulation (DNS) data of turbulent premixed V-flames. Both temporal correlations can be well represented by Gaussian-type functions, whereas only the two-time correlation accurately captures the slow global variation of the fluctuating heat release rate and hence the low-frequency noise sources. The resulting correlation model is applied to predict the far-field noise spectrum from open flames, and different reference time scales are used to scale this correlation from the DNS data to the test flames. The comparison between predictions and measurements indicates that all reference time scales are capable of reproducing the essential spectral shape including the low- and high-frequency dependencies. Reasonable agreement in the peak frequency, peak sound pressure level, and the Strouhal number scaling of peak frequency is also achieved for two turbulent time scales. A promising convective time scale shows great potential for characterising the spectral content, yet requires further verification through the DNS of a bounded flame configuration.