In this work, a numerical 3D simulation of a longitudinal ventilation system (LVS) is developed to analyze the fire behaviour inside a road tunnel. The numerical modelling reproduces the Memorial Tunnel, a two-lane, 853 m long road tunnel, used for experimental purposes. On this tunnel, 98 full-scale fire ventilation tests with different ventilation systems were conducted, constituting the first significant experimental approach to analyze fire incidents inside road tunnels. A total number of 24 reversible jets fans were installed in groups of three, nearly equally spaced over the length of the tunnel, and cantilevered from the ceiling of the tunnel. The validation of a numerical model is developed in the present paper. For that purpose, the behaviour of the smoke generated during a fire incident inside a road tunnel is predicted and compared with previous experimental data collected in the Memorial Tunnel Project. The smoke evolution and the performance of the LVS is simulated with a commercial code, FLUENT, which allows 3D unsteady simulations of the Navier-Stokes equations for multispecies mixtures of gases. A sufficient mesh density was introduced for the spatial discretization in order to obtain accurate results in a reasonable CPU time. Hence, typical ratios between total number of cells and the overall tunnel length were employed in the modelling. As a result, good agreement was achieved in all the tested cases, defining an accurate methodology to predict the performance of a LVS in case of fire inside a tunnel.

This paper analyzes the structure of turbulence in a single stage, low-speed axial fan with inlet guide vanes. Turbulence intensity values and integral length scales have been obtained using hot-wire anemometry for three different operating points and two different axial gaps between the stator and the rotor. These measurements were carried out in two transversal sectors, one between the rows and the other rotor downstream, covering the whole span of the stage for a complete stator pitch. Since total unsteadiness is composed of the contribution of both periodic and random unsteadiness, a processing data method was developed to filter deterministic unsteadiness in the raw velocity traces. Velocity signals were transformed into the frequency domain by removing all the contributions coming from the rotational frequency, the blade passing frequency and its harmonics. Consequently, coherent flow structures were decoupled and thus background levels of turbulence-RMS values of random fluctuations-were determined across the stage. Additionally, this unsteady segregation revealed further information about the transport of the turbulent structures in the unsteady, deterministic flow patterns. Therefore, anisotropic turbulence, generated at the shear layers of the wakes, could be identified as the major mechanism of turbulence generation, rather than free-stream, nearly isotropic turbulence of wake-unaffected regions. Finally, spectra and autocorrelation analysis of random fluctuations were also used to estimate integral length scales-larger eddy sizes-of turbulence, providing insight on the complete picture of the turbulent flow.

The present study is focused on the analysis of the dynamic and periodic interaction between both fixed and rotating blade rows in a single stage, low-speed axial fan with inlet guide vanes. The main goal is placed on the characterization of the unsteady flow structures involved in an axial flow fan of high reaction degree, relating them to working point variations and axial gap modifications. For that purpose, an experimental open-loop facility has been developed to obtain a physical description of the flow across the turbomachine. Using hot-wire anemometry, measurements of axial and tangential velocities were carried out in two transversal sectors: one between the rows and the other downstream of the rotor, covering the whole span of the stage for a complete stator pitch. Ensemble- and time-averaging techniques were introduced to extract deterministic fluctuations from raw data, both of which are essential to understand flow mechanisms related to the blade passing frequency. An exhaustive analysis of the measured wakes has provided a comprehensive description of the underlying mechanisms in both wake-transport phenomena and stator-rotor interaction. In addition, unmixed stator wakes, observed at the rotor exit, have been treated in terms of dispersion and angular displacement to indicate the influence of the blades loading on the transport of the stator wake fluid. The final aim of the paper is to highlight a complete picture of the unsteady flow patterns inside industrial axial fans.

There is still discrepancy regarding the verification of CFD U-RANS simulations of vertical-axis wind turbines (VAWTs). In this work, the applicability of the Richardson extrapolation method to assess mesh convergence is studied for several points in the power curve of a VAWT. A 2D domain of the rotor is simulated with three different meshes, monitoring the turbine power coefficient as the convergence parameter. This method proves to be a straightforward procedure to assess convergence of VAWT simulations. Guidelines regarding the required mesh and temporal discretization levels are provided. Once the simulations are validated, the flow field at three characteristic tip-speed ratio values (2.5-low, 4-nominal and 5-high) is analyzed, studying pressure, velocity, turbulent kinetic energy and vorticity fields. The results have revealed two main vortex shedding mechanisms, blade-and rotor-related. Vortex convection develops differently depending on the rotor zone (upwind, downwind, windward or leeward). Finally, insight into the loss of performance at off-design conditions is provided. Vortex shedding phenomena at the low tip-speed ratio explains the loss of performance of the turbine, whereas at the high tip-speed ratio, this performance loss may be ascribed to viscous effects and the rapid interaction between successive blade passings.

In this paper, the aerodynamic field around a FX 63-137 airfoil for four angles of attack and low Reynolds numbers was simulated with a Large Eddy Simulation (LES). Following, an acoustic analogy method was employed to calculate the airfoil trailing edge (TE) noise. In this second scheme step, the far-field acoustic pressure was predicted from the LES source terms using two different methods based on Lighthill's analogy: Curle's surface approach and Ffowcs-Williams and Hall's volumetric analogy (FW-Hall). Numerical results have been validated with hot-wire anemometry for the aerodynamic fields, thus verifying the accuracy of the CFD simulation for the prediction of noise propagation to the far field. Additionally, aeroacoustic results were validated with experimental measurements carried out in an anechoic wind tunnel using a frequency analyzer. The FW-Hall formulation shows a better agreement with the experiments, especially in the range of frequencies corresponding to the trailing edge, whereas Curle's analogy overpredicts airfoil sound. An exhaustive analysis of the aerodynamic flow field has been performed in order to better understand the generation mechanisms of the TE noise. The aeroacoustic calculations presented in this work contribute to develop a more reliable and efficient prediction methodology based on the Computational Aeroacoustics Approach (CAA).