Search Results (9)

In this study, an optimization framework for turbomachinery blades using a hybrid surrogate model assisted by proper orthogonal decomposition (POD) is introduced and then applied to the aero-structural multidisciplinary design optimization of a transonic fan rotor, NASA Rotor 67. The rotor blade is optimized through blade sweeping controlled by Gaussian radial basis functions. Calculations of aerodynamic and structural performance are achieved through computational fluid dynamics and computational structural mechanics. With a number of performance snapshots, singular value decomposition is employed to extract the basis modes, which are then used as the kernel functions in training the POD-based hybrid model. The inverse multi-quadratic radial basis function is adopted to construct the response surfaces for the coefficients of kernel functions. Aerodynamic design optimization is first investigated to preliminarily explore the impact of blade sweeping. In the aero-structural optimization, the aerodynamic performance, and von Mises stress are considered equally important and incorporated into one single objective function with different weight coefficients. The results are given and compared in detail, demonstrating that the average stress is dependent on the aerodynamic loading, and the configuration with forward sweeping on inner spans and backward sweeping on outer spans is the most effective for increasing the adiabatic efficiency while decreasing the average stress when the total pressure ratio is constrained. Through this study, the optimization framework is validated and a practical configuration for reducing the stress in a transonic fan rotor is provided. Full article

A large-sized backward-curved fan with high shape difficulty was designed, and fan performance was roughly predicted from computational fluid dynamics. Three gating systems of aluminum sand casting were designed to fabricate the fan. The flow pattern and solidification process of molten metal were analyzed by casting simulation. Three types were applied: bottom-up with four gates, bottom-up with ten gates, and top-down with a feeder. The simulation results of the bottom-up with four gates show that a large temperature loss occurs while molten metal flows into thin blades, and there is a temperature range below the liquidus temperature. Due to nonuniform temperature distribution, the solidification pattern is also not uniform. The bottom-up with ten gates shows almost similar flow and solidification patterns but has the effect of slightly reducing the temperature loss of molten metal. The top-down type has a much smaller temperature loss, while molten metal flows into the mold cavity compared to the bottom-up type and has a directional solidification pattern. As the feeder also acts as a riser to compensate for the shrinkage of the thick part, the simulation results regarding porosities are also significantly reduced. The fan cast as a top-down type has soundness without any unfilled parts. Full article

The existing inherent strain method is improved in this paper to address the shortcomings of the existing inherent strain method in the process of loading inherent strain. Unlike the traditional inherent strain method, which uses one-step loading inherent strain for each weld seam for one-time elastic calculation, the improved inherent strain method uses step-by-step loading inherent strain for each weld seam for multiple elastic calculations to predict welding deformation. The step-by-step loading inherent strain method (SBS-ISM) is more in line with the actual welding deformation generation process. Firstly, the local finite element model of the T-joint was used to analyze the welding deformation and extract the inherent strain by using the thermal elastic–plastic finite element method (TEP-FEM). Subsequently, the one-step loading inherent strain method (OS-ISM) and the step-by-step loading inherent strain method (SBS-ISM) were used to predict the welding deformation for the same local finite element model, respectively. The comparative results showed that the trend and magnitude of welding deformation calculated using SBS-ISM was much closer to those calculated using TEP-FEM. The OS-ISM and SBS-ISM were used to predict the welding deformation of the backward centrifugal fan impeller under different welding sequences, respectively. By comparing the welding deformation results calculated using the two inherent strain methods with the experimental results, it was demonstrated that the step-by-step loading inherent strain method (SBS-ISM) provides more accurate and reliable predictions of welding deformation for large and complex thin-walled T-shaped structural components compared to the one-step loading inherent strain method (OS-ISM). Full article

For a backward curved centrifugal fan, reducing volume flow rate from design operating point towards part load yields an increase in noise emission together with a reduction of efficiency. The spectral content of the emerging noise emission can be characterized by a subharmonic hump with several harmonics. Based on narrow-band acoustic measurements and numerical Lattice-Boltzmann simulations, a deeper insight is sought after. Downstream unsteady flow patterns are identified to play a central role in this performance deterioration; and geometrical flow stabilization means are investigated. Full article

The effect of blade sweep has been studied numerically with the Lattice Boltzmann Method on a family of low-speed free-vortex axial fans with sweeps of ±45°. Good overall aerodynamic agreement is first demonstrated on all fans at the design condition, particularly in the tip gap. The local larger wall-pressure fluctuations seen in the unswept and backward swept fans compared to the forward case are traced to the stronger tip vortices that remain in the rotational plane or even move upstream. These stronger and faster vortices interacting with the fan blades are then responsible for the larger noise levels observed in the acoustic spectra of these fans, and particularly for large subharmonic humps. Excellent agreement between experimental and numerical noise predictions is finally reported stressing the dominant tip noise. Full article

We apply the Support Vector Regression (SVR) machine learning model to estimate surface roughness on a large alluvial fan of the Kosi River in the Himalayan Foreland from satellite images. To train the model, we used input features such as radar backscatter values in Vertical–Vertical (VV) and Vertical–Horizontal (VH) polarisation, incidence angle from Sentinel-1, Normalised Difference Vegetation Index (NDVI) from Sentinel-2, and surface elevation from Shuttle Radar Topographic Mission (SRTM). We generated additional features (VH/VV and VH–VV) through a linear data fusion of the existing features. For the training and validation of our model, we conducted a field campaign during 11–20 December 2019. We measured surface roughness at 78 different locations over the entire fan surface using an in-house-developed mechanical pin-profiler. We used the regression tree ensemble approach to assess the relative importance of individual input feature to predict the surface soil roughness from SVR model. We eliminated the irrelevant input features using an iterative backward elimination approach. We then performed feature sensitivity to evaluate the riskiness of the selected features. Finally, we applied the dimension reduction and scaling to minimise the data redundancy and bring them to a similar level. Based on these, we proposed five SVR methods (PCA-NS-SVR, PCA-CM-SVR, PCA-ZM-SVR, PCA-MM-SVR, and PCA-S-SVR). We trained and evaluated the performance of all variants of SVR with a 60:40 ratio using the input features and the in-situ surface roughness. We compared the performance of SVR models with six different benchmark machine learning models (i.e., Gaussian Process Regression (GPR), Generalised Regression Neural Network (GRNN), Binary Decision Tree (BDT), Bragging Ensemble Learning, Boosting Ensemble Learning, and Automated Machine Learning (AutoML)). We observed that the PCA-MM-SVR perform better with a coefficient of correlation (R = 0.74), Root Mean Square Error (RMSE = 0.16 cm), and Mean Square Error (MSE = 0.025 ${\mathrm{cm}}^2$). To ensure a fair selection of the machine learning model, we evaluated the Akaike’s Information Criterion (AIC), corrected AIC (AICc), and Bayesian Information Criterion (BIC). We observed that SVR exhibits the lowest values of AIC, corrected AIC, and BIC of all the other methods; this indicates the best goodness-of-fit. Eventually, we also compared the result of PCA-MM-SVR with the surface roughness estimated from different empirical and semi-empirical radar backscatter models. The accuracy of the PCA-MM-SVR model is better than the backscatter models. This study provides a robust approach to measure surface roughness at high spatial and temporal resolutions solely from the satellite data. Full article

Two blade curvatures representative of those found in automotive fans are compared. Measured performances are analyzed for forward and backward curved blades, either with or without heat-exchangers placed in front of them. The backward fan demonstrated good efficiency but poor acoustics, whereas it is the contrary for the forward fan. Investigations are completed by a numerical analysis of the flow in the cooling module. Different integration effects are highlighted depending on the blade curvature, showing variation in pressure, torque and efficiency. Analyses of blade loadings show that the flow is more homogeneous with a forward curved fan and it produces less unsteadiness at the blade tip. Post-processing of detached eddy simulations (DES) shows density fluctuations on the blade wall and confirms the correlation between the large vortical structures and the acoustic sources for both fans. In addition, with the forward fan, the sound propagation is less directed towards the axis of rotation and it yields up to −3.6 dB of sound pressure level (SPL) measured in front of the cooling module. As a conclusion, any choice for a fan must result from a compromise between aerodynamics and aeroacoustics, and the final performances must be carefully checked on the module. Full article

The steady and unsteady flow characteristics of internal flow in a backward centrifugal fan of double inlet at low flow-rate condition are investigated by computational fluid dynamics in this paper. The investigation aims to reveal insights into generation mechanisms and our physical understanding of the rotating stall and surge. The numerical results mainly demonstrate that, with decreasing flow rate, a large number of vortex flows almost increasingly occupy the internal flow of the impeller. The reverse flow and separation vortices increasingly appear near the outlet of volute, and the internal flow of the impeller is completely blocked by the separated vortex flow at low flow-rate conditions. Results indicate that, due to a synchronization of the impeller rotation and separation vortex, these separated vortices act intensely on the pressure surface of the blade with time evolution, and the interaction between the separated vortices and surface of blade increasingly yields small-scale eddies. It is further found that the amplitude of pressure and velocity fluctuations gradually increase with the decrease of flow rate in a certain range. The unsteady characteristics acting on the volute tongue gradually increase in a range of Q_d to 0.3 Q_d (Q_d is the design volume flow rate) with the decrease of flow rate, and the unsteady characteristics acting on the volute tongue are weakened at the working condition of 0.15 Q_d. These insights clearly explain the unsteady nature of the rotating stall and surge phenomenon in the double inlet backward centrifugal fan. Full article

In this paper, nonlinearity associated with intense ultrasound is studied by using the one-dimensional motion of nonlinear shock wave in an ideal fluid. In nonlinear acoustics, the wave speed of different segments of a waveform is different, which causes distortion in the waveform and can result in the formation of a shock (discontinuity). Acoustic pressure of high-intensity waves causes particles in the ideal fluid to vibrate forward and backward, and this disturbance is of relatively large magnitude due to high-intensities, which leads to nonlinearity in the waveform. In this research, this vibration of fluid due to the intense ultrasonic wave is modeled as a fluid pushed by one complete cycle of piston. In a piston cycle, as it moves forward, it causes fluid particles to compress, which may lead to the formation of a shock (discontinuity). Then as the piston retracts, a forward-moving rarefaction, a smooth fan zone of continuously changing pressure, density, and velocity is generated. When the piston stops at the end of the cycle, another shock is sent forward into the medium. The variation in wave speed over the entire waveform is calculated by solving a Riemann problem. This study examined the interaction of shocks with a rarefaction. The flow field resulting from these interactions shows that the shock waves are attenuated to a Mach wave, and the pressure distribution within the flow field shows the initial wave is dissipated. The developed theory is applied to waves generated by 20 KHz, 500 KHz, and 2 MHz transducers with 50, 150, 500, and 1500 W power levels to explore the effect of frequency and power on the generation and decay of shock waves. This work enhances the understanding of the interactions of high-intensity ultrasonic waves with fluids. Full article