Forced and unforced unsteadiness in an axial turbomachine

Abstract

Different sources of unsteadiness in low-speed axial turbomachinery are identified and classified in the present paper. From the classical picture segregating non-periodic mechanisms (turbulence) from periodic phenomena (unsteadiness), a further decomposition is outlined to distinguish between forced (deterministic periodicities) and unforced (non-deterministic) unsteadiness. Raw velocity traces, measured for several test conditions in a typical industrial fan with hot-wire anemometry, are ensemble-averaged to obtain statistically-stationary fluctuations. Then, a frequency-based filtering procedure is employed to isolate non-deterministic, but also non-chaotic, disturbances from such fluctuations, resulting in the so-called unforced unsteadiness. This term reveals coherent flow structures that involve "large-scale" unsteadiness with other periodic features different to the blade rotation scales (BPF). As a starting point, the kinetic energy associated to the total unsteadiness (in terms of a percentage of the kinetic energy of the time-averaged flow) is analyzed as a function of the operating conditions. Next, the different components contributing to the total unsteadiness of the flow are also observed, in order to determine their particular significance on the global unsteady scenario. It is shown that the turbulent kinetic energy reaches up to approximately a 50-60% of the total unsteady energy, while both forced and unforced components contribute equally to the rest of the energy. In addition, it is observed that higher levels of unforced unsteadiness are concentrated towards the endwall boundary layers where forced unsteadiness is notably reduced due to the loss of the jet-wake structure. Conversely, forced unsteadiness is more evident at inner regions of the rotor passage. Furthermore, unforced unsteadiness is especially intense in the tip regions where large scales associated to the tip leakage vortex are established. It is demonstrated that the estimation of the unforced unsteadiness constitutes an accurate indicator of the presence of tip leakage flows for axial turbomachinery. Moreover, this is confirmed through the representation of the degree of anisotropy, where typical anisotropic structures are revealed. Finally, with the introduction of power spectrum densities for the unforced components, it is possible to identify typical eddy sizes of these large fluctuations.

Non-deterministic kinetic energy within the rotor wakes and boundary layers of low-speed axial fans: frequency-based decomposition of unforced unsteadiness and turbulence

“…On the contrary, the unforced component remains practically constant, in the order of a 17-19% whatever flow rate considered. This finding is in consonance with previous observations of the authors for this type of industrial axial fans: while both turbulent and forced mechanisms show important quantitative variations with the operating conditions, the unforced component is practically independent to this parameter [40]. Along the radial coordinate, as we move away from central positions, the forced contributions are also continuously reduced and much of the unsteady energy is again transferred to the turbulent scales.…”

“…This means that the timeunresolved fluctuation is mainly composed by turbulent structures, but with additional periodic contributors superimposed to the baseline spectrum of turbulence. This largescale mechanisms, redefined by the authors as "unforced unsteadiness" [40], include vortex shedding, tip vortex instabilities (like an unsteadiness on the unsteadiness), unsteadiness of separated points, rotating stall and rotating instabilities [55], flutter and blade misalignments. Previous investigations have analyzed in detail the flow structures of the deterministic fluctuations [36][37], especially concerning rotor-stator interaction in the axial gap [38].…”

“…In these studies, complemented also with numerical simulations [33][34][35], a multi-point traversing system was employed to obtain detailed distributions of the velocity field in several windows covering the passage span for a complete stator pitch. The data set obtained from this facility has been used to investigate many different flow phenomena including wake dispersion and wake transport [35][36], mean flow, turbulent structures and integral length scales [36][37], deterministic stresses [38] and secondary flows and tip vortex in similar fans [39][40]. At present day, as an alternative to the costly experimental campaigns, computational modelling has become an extended, well-mature discipline, especially useful in the design process.…”

Experiments on turbocharger compressor acoustics

“…Other tests have been performed to directly measure the flow velocity fields through different velocimetry techniques such as Laser Doppler Anemometry (LDA) [53], hot-wire anemometry (HWA) [54,55] and Particle Imaging Velocimetry (PIV) [53,56,57].…”

“…The latter technique is desirable for CFD validation as it allows the measurement of whole sections of the field instead of a single point such as in the HWA and LDA techniques. Regarding the use of PIV in turbomachinery, many of the studies available in the literature have been performed on large centrifugal pumps [53,58,59], axial fans [54,60] and axial compressors [56,57,61,62], usually with the aim to study the evolution of the flow through the passages and the interaction with the diffuser vanes.…”