چكيده به لاتين
Cavitation phenomenon is one of the most important phenomena in fluid flow and occurs when the local pressure in the fluid drops near the vapor pressure of the fluid at that temperature. Cavitation is divided into different categories according to the type and location of formation, of which partial cavitation has a special place of study because of its high subject matter and higher occurrence in real industries. In the present study, a hydrofoil in the cavitation tunnel is experimentally tested and the cavitation event is visualized by photodetection. Using image processing techniques, the length of the cavity created around the hydrofoil was obtained, and for a more detailed study of cavitation behavior and due to the dynamic behavior of partial cavitation, twelve images were taken at each angle and by averaging the length of the cavity created was determined in each specific flow conditions. The main innovation in this study is the study of the cavity behavior when increasing and decreasing the hydrofoil's angle of attack which led to the definition of hysteresis with a new approach in cavitation literature, which is the difference between the cavity to chord length ratio in increase and decrease cycle for a particular angle. It should be noted that the hysteresis phenomenon was experimentally observed and reported with the new approach defined in this study. To show the reproducibility of the results, experiments were performed for three different mass flow rates and three cycles were run in each mass flow rate. Due to temperature changes during the experiment, the cavitation number was also changed per cycle, so the results were tested and reported for nine different cavitation numbers, and it was shown that the hysteresis phenomenon occurred in all cavitation numbers. Another innovation in the present work is the statistical study of the partial cavitation formed around the hydrofoil. It is shown that the total number of 24 images taken for each specific angle of attack and flow condition (12 images in increasing and 12 images in decreasing), respectively, follows the Normal or Gaussian distribution. 1296 data obtained from a series of experiments were used to develop the artificial neural network and to use in response surface methodology, which were able to estimate the cavity length with regression coefficients of 0.93 and 0.82, respectively. These amounts were highly accurate with respect to the dynamic behavior of the cavity.
