Abstract
Ways to improve and optimize the Tesla valve design, the performance of which is extensively studied by combining artificial neural network (ANN) and response surface methodology (RSM) for a deeper analysis of valve performance are considered. Some key valve design parameters are selected for optimization using dipolarity (Di) and absolute pressure drop ratio (APDR) as response factors. Experimental data are obtained by numerical simulation. Forecasting models using ANN and RSM are built and their accuracy is compared. The developed two-stage Tesla valve demonstrated high efficiency in real-world conditions and can be effectively used as a check valve. The prediction results using the ANN model were more accurate than with the RSM model, and no retraining is required. Keywords machine learning, Tesla valve, optimization of structural parameters, optimization of numerical simulation, one-way valve design
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call