Hossein Ali Kamali

While artificial supercavitation serves as a viable method for minimizing skin friction in underwater vehicles, there is a notable lack of comprehensive research into ventilated cavitation with constant temperature and its repercussions on such vessels. Additionally, Machine Learning models stand out as a contemporary approach for addressing complex issues in today's landscape. In the current study, the effect of increasing the aft body diameter on ventilated supercavitation flow with constant temperature using experimental, numerical, and optimized machine learning methods has been studied. In this study, incompressible air with a constant temperature was used for injection into the cavity. The data was used to validate the numerical simulation and machine learning model. The machine model used is optimized random forest model, with acceptable accuracy (About 10%) predictions and validated by numerical and experimental results. Also, using machine learning interpretability techniques, the importance and impact of parameters on cavity length have been investigated. The results show that in an aerated cavity with a constant temperature, the cavitator aft body diameter ratio affects the length of the cavity in a nonlinear way. First, it increases the length due to body-cavity interactions, and then decreases the length. Also, by increasing the ventilation coefficient, the maximum cavity length is achieved at a smaller aft body diameter ratio. Specifically, in low ventilation coefficients, doubling the rear body diameter results in an increase in cavity length by more than 70%. In high ventilation coefficients, the length of the cavity increases by more than 20%. After reaching the maximum cavity length, further increasing the rear body diameter by 50% can decrease the cavity length by up to 30%, depending on the Froude number and the ventilation coefficient.

Icing on airfoils is one of the major challenges affecting the aerodynamic performance of aerial vehicles, especially under cold weather conditions. Most previous research has focused on simulating icing on conventional geometries such as aircraft wings, wind turbine blades and thick airfoils, primarily under high-speed conditions. However, the impact of icing on thin airfoils at low speeds and Mach numbers between 0.1 and 0.3 has not been widely studied, even though it plays an important role for low-altitude and light aircraft. In this work, the NACA 0009 thin airfoil is chosen, and its aerodynamic performance under icing conditions is examined through numerical simulation. After validating the numerical model against experimental data, the effects of increasing angle of attack ([ ] to [ ]) and Mach number (0.1 to 0.3) on ice shape and thickness, flow field characteristics and aerodynamic coefficients (lift and drag) are evaluated. The results show that at a fixed angle of attack, increasing the Mach number under low-speed conditions significantly expands the iced region. Also, increasing the angle of attack changes the icing pattern and makes it asymmetrical. Additionally, in Mach numbers 0.3 and angles of attack [ ]–[ ], flow separation and a sharp drop in aerodynamic performance occur due to ice growth on the leading edge, effectively reducing the stall angle.

Improving the aerodynamic coefficients of airfoils is an important and practical issue that has been the subject of research and development over the years. The use of porous media is one of the passive methods in fluid mechanics aimed at enhancing the aerodynamic coefficients of wings. This study investigates the effect of porous media at the leading edge of the NACA 0012 airfoil on its performance in the stall region, utilizing numerical simulation of the porous medium with the assumption of a medium containing circular micro-grains at low Reynolds numbers. In this study, various parameters of the porous medium are examined individually for their effects on the aerodynamic performance of the airfoil. The results indicate that the use of porous media improves the aerodynamic performance of the airfoil in the stall region and delays the onset of the stall. Additionally, parametric studies on the variables of the porous medium show that reducing the length of the porous medium from 0.3c to 0.1c results in a 25% increase in the airfoil’s performance in the stall region. Furthermore, decreasing the depth of the porous medium from 0.05c to 0.01c leads to more than a 50% improvement in airfoil performance, while doubling the porosity leads to an increase of more than 80% in airfoil performance.

The interaction of high-speed flows and solid particles is one of the complex studies among multiphase flows that has been less investigated due to the complexity of the simulation. In this study, the effects of injecting solid particles with different temperatures on the characteristics of high-speed gas flow inside a three-dimensional cylinder are investigated. For this purpose, the Eulerian–Lagrangian method is used to study the interaction of solid particles and gas flow with each other. Also, the accuracy of the machine learning model in predicting the properties of gas flow in interaction with the solid particles with different temperatures is investigated. For this purpose, the optimized random forest model is used to predict the changes in the total flow temperature along the cylinder. The results show significant effects of the temperature of solid particles on the characteristics of high-speed gas flow inside the cylinder. Injecting solid particles with a temperature equivalent to the temperature of the total gas flow increases the temperature of the total gas flow along the cylinder and reduces the temperature of the solid particles. Also, injecting solid particles with a lower temperature reduces the temperature of the total flow and increases the temperature of the solid particles. The temperature of solid particles with [ ][ ]K increases from 300[ ]K to 550[ ]K in a very short time of about 0.00074[ ]s. At the same time, the temperature of particles in another case with [ ][ ]K decreases from 800[ ]K to about 700[ ]K. After the injection of solid particles, their velocity increases. This variation occurs very quickly, so the particle speed increases from 100[ ]m/s to 550[ ]m/s in just 0.0002[ ]s for both cases. Also, the results of the random forest model indicate the appropriate accuracy of this model in predicting changes in the total temperature of the flow along the cylinder. The results show that the maximum error between CFD results and random forest prediction is about 2%.

Ventilated supercavitation, a complex two-phase flow, has mostly been explored through experiments and simulations, with machine learning yet to emerge as a complementary research approach. This study combines experiments, numerical simulations, and machine learning models to explore the characteristics of ventilated supercavitation and the different cavitating regimes behind a disk-shaped cavitator. The experimental data were used to validate the simulation models. Subsequently, an optimized Random Forest model, enhanced using the Bayesian Optimization Algorithm (BOA-RF), was trained on the simulation results to predict cavitation length, cavitation number, and classify the cavitation regime type. The BOA-RF model exhibited high accuracy when compared with both experimental and numerical results. The results indicate that the supercavity length increases with the ventilation coefficient (CQ). However, at the formation ventilation coefficient (CQf), the cavity length undergoes a significant and rapid increase. Moreover, as the ventilation coefficient (CQ) increases, the cavitation regime evolves from foamy cavitation to a continuous transparent and asymmetric cavity, and when CQ ≥ CQf, it transitions to clean supercavity. Additionally, the results indicate that as the Froude number increases, the CQf initially increases and then decreases.

The phenomenon of artificial cavitation is a significant and practical aspect of reducing drag forces on devices interacting with water. Among various regimes, the supercavity regime, where the entire body or a substantial portion of it is enveloped by a cavity, plays a crucial role. This study investigates the cavitation phenomenon using a combination of numerical experimental methods and machine learning techniques. Specifically, the Support Vector Machine (SVM) classification model is employed to identify the type of artificial cavitation regime and to determine the occurrence of the supercavity regime based on input variables. The findings indicate that the Radial Basis Function (RBF) model outperforms other machine learning models in accurately detecting the cavity regime type, achieving an accuracy of over 95% in identifying the supercavity regime. Additionally, results from the RBF model demonstrate that the supercavity regime is observed at low cavitation numbers, exhibiting the longest cavity length.

Minimizing drag on the cavitator is essential in hydrodynamic research and for improving the performance of objects used in marine environments. This study focuses on optimizing a parabolic cavitator and analyzing the cavities it generates in detail. First, the key factors for optimizing the cavitator were identified using the Taguchi method. Based on these factors, the three-dimensional shape of the cavitator was numerically simulated, and the hydrodynamic forces acting on it were calculated with consideration of cavitation. The optimized cavitator shape was then identified through further analysis using the Taguchi method and was experimentally tested to confirm its real-world performance. Subsequently, the characteristics of artificial cavitation behind the improved cavitator were examined both experimentally and numerically across various ventilation coefficients. The experiments included high-speed imaging and pressure measurements to capture the dynamics of cavity formation and collapse, while numerical simulations were performed using a k-omega shear stress transport turbulence model and a volume of fluid approach to accurately predict the phase interface. The results highlight the importance of the cavitator's incidence angle and the distance from its nose to its base in the optimization process. Moreover, the analysis shows that pressure fluctuations are significantly more intense at the point where the cavity closes than within the cavity itself. Additionally, the findings indicate that the supercavity characteristics generated by this optimized cavitator are 10% better than those produced by other cavitators, contributing to reduced drag and improved hydrodynamic efficiency.

One of the complex challenges in ventilated cavitating flow studies is analyzing hysteresis behavior and the formation air entrainment coefficient (Cqf) under different conditions. This study explores the formation and collapse processes of the supercavity using experimental observations, numerical simulations, and machine learning (ML) models to reveal the hysteresis behavior of air entrainment at different flow conditions. Initially, the research focused on studying air entrainment hysteresis for a disk-shaped cavitator under different Froude numbers (Fr) through experimental and numerical methods. The study identified two key air entrainment coefficients in the hysteresis curves, which are important for gas generator design. In the ML section, hyperparameter optimization for the random forest (RF) model is performed using genetic algorithm (GA) and particle swarm optimization (PSO). The results demonstrate that the GA-RF model is more accurate than the PSO-RF model in predicting experimental data. The GA-RF findings show that, for a fixed cavitator diameter, Cqf increases with Fr, reaching a maximum value (MaxCqf) before decreasing as Fr continues to rise. Additionally, with an increase in the cavitator diameter, MaxCqf increases and Fr belonging to this ventilation coefficient (FrMaxCqf) decreases.

The aim of this paper is to investigate the ventilated cavitating flow behavior, through experimental and numerical methods. The experiments are performed for a test model placed in a water tunnel and a high-speed camera is used to image the formation and collapse of the supercavity. For the numerical simulations, Ansys CFX commercial software is employed to solve the Reynolds-averaged Navier–Stokes (RANS) equations and additional transport equation for the liquid volume fraction in the unsteady condition. Both the experimental observation and numerical prediction show a hysteresis behavior for ventilated supercavity in which the value of the air entrainment coefficient required to maintain the supercavity in the formation process is less than the amount required to form it. The good agreement observed between the numerical prediction and experimental data, revealed the accuracy and capability of the numerical scheme. Also, the numerical simulation shows that the variations in the supercavity length, in addition to the air entrainment, are a function of the air leakage regime from the closure region of supercavity.