Analysis Of Characteristic Parameters In An Axial Flow Fan With Abnormal Blade Angle

Because of the presence of abnormal installation,stall and surge phenomena are generated during the running of variable pitch axial flow fans, and these phenomena are generally accompanied with the significant increase of fan noise which seriously affects the physical and mental health of operators. Further research on aerodynamic noise characteristics of an axial flow fan with abnormal blade angle is crucial for high efficiency, low noise and safe operation and for developing high performance of an axial flow fan.Taking a variable pitch axial fan of OB-84 type as an example, the distribution of sound source intensity and characteristics of monitoring regions in time and frequency domains are simulated using the large eddy simulation(LES) and FW-H noise model, wavelet decomposition and reconstruction is applied on the original sound pressure signal with power spectral density analysis results. The abstracted characteristic entropy with reconstruction signal, three kinds of complexity algorithms and characterization capabilities are examined, in addition, sound pressure signal energy features are extracted using wavelet packet and EMD methods.Results show that, main noise sources of the fan are located at the blade leading edge and the tip of the suction surface; the sound pressure in different regions of the axial fan are obviously different in time domain, and the morphology, interval and amplitude of pulsation varies with abnormal installation angle; the features of discrete peaks in low-frequency region is exhibited, the broadband noise of blade region is enhanced under abnormal installation angle, and the discrete peak of aerodynamic noise transits from low-frequency region to high-frequency region, while the increase of broadband noise in middle-high frequency region is visible distinctly. The characteristic entropy in blade region is the least due to uniform regularity, and the aerodynamic characteristics and noise signal remain relatively stable in afflux section and diffusion section, thus the abnormal degree of blade installation angle could be evaluated by the approximate entropy, sample entropy and Lempel-Ziv complexity. With the exception of rotating blade, the wavelet packet energy in frequency band one contained fundamental frequency under abnormal condition is much lower than that under normal condition, and, with the increase of ??, the difference between them is greatly significant; the EMD energy features in different regions recognizing abnormal installation angle exhibit as prominently obvious change of relative IMF energy; energy entropy of wavelet packet and EMD energy entropy raise with increasing ??, and the energy entropy is closely related with the deviation of installation angle.