نوع مقاله : مقاله پژوهشی
گروه مهندسی مکانیک، دانشکده مهندسی، دانشگاه شهید چمران اهواز، ایران
عنوان مقاله [English]
In this paper, a new correlation is developed to predict sand particle erosion in standard elbow for churn flow pattern. Using the Buckingham Pi theorem, dimensional analysis is performed, and the dimensionless groups governing the erosion phenomenon in churn flow are derived. The dimensionless groups include dimensionless erosion, mixture Reynolds number, ratio of the superficial velocities, and the ratio of particle diameter to elbow diameter. Using the Gaussian process regression (GPR), a parametric erosion correlation for churn flow is then constructed. In the proposed model, the ratio of the superficial velocities is considered as the deviation factor of actual properties in churn from the mixture properties and its effect on the predicted erosion rate. For different operating conditions, the model results are validated with experimental data. The results show that more than 83% of the predictions have an error of less than 30%. Also, the model performance is compared with previous existing models. It is observed that the proposed model error is much less than other existing models. These results indicate the accuracy and efficiency of the presented model in solid particle erosion for churn flow. Finally, the proposed model is utilized to obtain the threshold erosional velocity curves in the churn flow pattern. The effect of various parameters such as elbow and particle size, liquid viscosity as well as gas flow pressure on the erosion values and threshold erosional velocity in churn flow are investigated.