Artificial neural networks analysis of thermophoresis in surface tension gradient tetra hybrid nanofluid for semiconductor processing and crystal growth applications

Abstract

The effects of activation energy on thermophoretic particle deposition in axisymmetric Marangoni convective flow of a tetra-hybrid nanofluid across a disc are studied in this work using an integrated numerical approach that makes use of artificial neural networks backpropagated with the Levenberg-Marquardt algorithm (ANN-BLMA). The importance of thermal radiation and Joule heating are discussed. The established model of thermophoretic particle deposition in a tetra hybrid nanofluid, which incorporates surface tension gradients (Marangoni effect) and activation energy, has numerous practical applications. It can be utilized to improve microscale coating and deposition processes in semiconductor manufacturing and photovoltaic cell manufacture, where accurate particle placement is required. In thermal management systems, including heat exchangers and microfluidic cooling devices, that enhances heat transmission and reduces particle clogging. The system of partial differential equations is transformed into nonlinear ordinary differential equations by using the appropriate transformations. This problem is theoretically solved using the Bvp4c algorithm. The proposed method's accuracy is determined using numerical tools such as regression-based statistical graphs and error histograms. The results, obtained using the Levenberg-Marquardt technique, demonstrate that artificial neural networks have consistent, trustworthy derivation, convergence, and validation. As the thermophoresis parameter values rise, the concentration profile decreases.