Optimization of Rotary-Magnet assisted ECSM on borosilicate-glass using machine learning

Abstract

ABSTRACT Owing to the incapability of the existing techniques, Electrochemical Spark Machining (ECSM) is transpiring to micro machine the glass for potential applications in micro domain. However, in ECSM, the bubble accumulation and poor flushing lead to the generation of a non-uniform gas film, resulting in the deterioration of Hole Circularity (HC) and Surface Roughness (SR). Therefore, present communication is an attempt toward the concept of Rotary-Magnet assisted ECSM (RM-ECSM) to produce the micro holes in the borosilicate glass with better quality characteristics (higher HC and lower SR), which are required for better performance of glass vias in RF-MEMS. The evolutionary algorithm and Machine Learning (ML) have been adopted for optimization. From optimization, the values of optimal input parameters were obtained as follows: Applied voltage = 43.46 V; Electrolyte-temperature = 45.93 ֯ C; Pulse on time = 3.28 ms; Magnetic rotation = 551 rpm with corresponding values of HC and SR as 79.28% and 3 um respectively. Finally, these outcomes were validated by Levenberg-Marquardt algorithm (ML).