Artificial Neural Networks for the Prediction of Electrochemical Etched Micro channel Dimensions

Abstract

In this study, artificial neural network was used to model the micro channel size created with electrochemical etching method in a specific pattern. Special series 5754 aluminum surfaces were coated with employing mask. The pre-designed pattern was then marked to the masked surface with laser, then it was subjected to electrochemical etching process. In this way, micro-patterned channels are formed on the aluminum surface. Various experiments were carried out based on the electrochemical etching parameters, such as concentration (0.1-2.5 M), distance between the electrodes (5-15 cm), operating voltage (15-48 V) and time (6-30 min). And the depth and width of the channels were investigated. Studies conducted under various conditions were modeled with ANN and the synergistic effects of the input and output parameters were explored by the surface graphics obtained as a result of the modeling. This modeling study is a powerful tool in terms of providing a prediction of the channel dimensions of the micro channel fabricated by electrochemical etching for the future related studies. In addition to the modeling, some impressions and inferences obtained from the experiments were also yielded in the conclusion part.