Application of artificial neural networks modeling to sprays and spray impingement heat transfer

摘要

Artificial neural networks (ANN) mdels have been developed and applied to free propane sprays and to water spray cooling beat flux predictions. For the propane spray conditions the ANN model is trained against the computational fluid dynamics (CFD) results and verified against experimental data for drop diameter at the centreline 95 mm from the nozzle. It is shown that an ANN model trained on CFD gives results comparable to the CFD predictions and that it can therefore be dmployed online in industry to investigate and limit the consequences of a depressurization accident. When trained and proves to be an alternative numerical modeling technique to CFD,with the numerical predictions comparable to the CFD predictions,but in real-time mode.