Aeroacoustics -Noise Prediction by Using "LES" for Signal Processing. (March 2017)

• Record Type: Journal Article
• Authors: Hodor, Victor; Birle, Dan; Nascutiu, Lucian; Deac, Ioan
• Is Part Of: Energy procedia. Volume 112(2017)
• Page Start: 322
• Page End: 329
• Publication Date: 2017-03
• Subjects: Aeroacoustics -- CFD prediction -- LES -- FW-H -- CAA -- Noise -- Octave; Power resources -- Congresses; Power resources -- Periodicals; Power resources; Conference proceedings; Periodicals; 333.7905
• DOI: 10.1016/j.egypro.2017.03.1104
• Languages: English
• ISSNs: 1876-6102

Abstract

Abstract: "Less noise, More sound": Starting tests were focus on boundary conditions, in order to reduce running time without losing convergence and stability. To be noticed that in the present particular application, the most significant noise generation comes from the imposed "primitive geometry" of the fan. Its blades should be straight (i.e. not aerodynamic profile) –for this specific electric motor –as it should have reverse spinning directions. The predictions refers only to the difference coming from a regular versus irregular blades disposal. The overall spatial discrimination was done using 12 million tetrahedral elements and with 1e-6 seconds time-step For these, we have done some pressure field vs. time scaling (frequency) prediction –with Ansys FLUENT and the appropriate FFT transform in certain monitoring points. LES with FW-H model was used in order to achieve higher accuracy in terms of a desired psychoacoustic discrimination between noises vs. sound attributes. First engaged steps, regards some classical DSP (tone, octave, etc.) discrimination. MatLab was used for the database manipulation in order to extract and export the relevant groups of peaches -from each (CFD with CAA) predicted octaves. WolframAlpha –is suited for an online proper pre-evaluation, between consonant and dissonant audio relevance's.