Machine-Learning Clustering Methods Applied to Detection of Noise Sources in Low-Speed Axial Fan

Abstract

The integration of rotating machineries in human-populated environments requires to limit noise emissions, with multiple aspects impacting on control of amplitude and frequency of the acoustic signature. This is a key issue to address and when combined with compliance of minimum efficiency grades, further complicates the design of axial fans. The aim of this research is to assess the capability of unsupervised learning techniques in unveiling the mechanisms that concur to the sound generation process in axial fans starting from high-fidelity simulations. To this aim, a numerical dataset was generated by means of LES simulation of a low-speed axial fan. The data set is enriched with sound source computed solving a-posteriori the perturbed convective wave equation (PCWE). First, the instantaneous flow features are associated to the sound sources through correlation matrices and then projected on latent basis to highlight the features with the highest importance. This analysis in also carried out on a reduced dataset, derived by considering two surfaces at 50% and 95% of the blade span. The sampled features on the surfaces are then exploited to train three cluster algorithms based on partitional, density and Gaussian criteria. The cluster algorithms are optimized and their results are compared, with the Gaussian Mixture one demonstrating the highest similarity (>80%). The derived clusters are analyzed, and the role of statistical distribution of velocity and pressure gradients is underlined. This suggests that design choices that affect these aspects may be beneficial to control the generation of noise sources.

An Affordable Acoustic Measurement Campaign for Early Prototyping Applied to Electric Ducted Fan Units

“…Having discussed what limitations apply to the affordable measurement campaign, gaining first insight in early prototyping is valuable and can be combined with more detailed computational [24,25,35,66] or experimental studies [30,46,67,68]. Simulations in combination with optimization can especially serve as a complementary noise reduction potential [69,70]. The value of the method was proven by identifying the EDF's noise characteristics of an improved design version and lowering the overall noise signature by several dB across a wide operating range.…”

Implementation of Ensemble Machine Learning Techniques on Hotel Reservation System

“…ML approaches are commonly used for different purposes such as grouping objects using common behavior [12], [13], extraction of interesting facts [14]- [16], classifying data using multiple factors [17]- [19], and detecting outliers [20]- [22]. Similarly, it has been used in the hotel industry to predict various aspects and determine the reason for room cancellation.…”

Surrogate Modeling of the Aeroacoustics of an NM80 Wind Turbine

“…Other major limitations to the application of CAA methods come from the limits of URANS in reproducing the turbulence spectrum and the typical necessity to move at least to hybrid LES-RANS, with the inevitable increase in computational costs [8]. On the other hand, the current involvement of unsupervised machine learning algorithms for predicting aerodynamic noise in rotors is still far from being a competitive alternative to CAA [9]. The common strategy for engineering purposes is to trade between accuracy and computational costs.…”