Prediction and Optimization of Ultrasonic Vibration Assisted Wire EDM Process for AISI P20 + Ni Using COOT Optimization Algorithm Based Deep Neural Network

Wire Electrical Discharge Machining (WEDM) is the most commonly used machining method due to its versatility towards complex machining projects, especially in mould-making industries. But the efficiency and surface finish of this WEDM process is questionable for mass production. Therefore, it is necessary to optimize WEDM process parameters for mouldmaking industries.

The experiments are conducted on AISI P20+Ni material which is prevalent in plastic mould-making industries. By introducing Ultrasonic Vibration (UV) into cutting wire, this study attempts to improve the efficiency and surface finishing of WEDM process. Further, the dynamics of key WEDM parameters such as peak current, servo voltage, pulse on time, and pulse off time on output responses are analyzed through Response Surface Methodology (RSM). Furthermore, the output parameters such as, Material Removal Rate (MRR), Micro Hardness (MH), Surface Roughness (SR), and Recast Layer Thickness (RLT) during UV-assisted WEDM are predicted using a novel hybrid deep learning technique called Deep Neural Network based COOT optimization algorithm (DNN+COOT).

The results shown that the proposed DNN+COOT algorithm learned better to predict output responses with an accuracy of 98.77% compared to the usual DNN. Finally, the UV-assisted WEDM process parameters are optimized to obtain maximum MRR, MH, and minimum SR, RLT at an average desirability of 0.65.

This study concludes, the UV-WEDM process as the most effect machining method for AISI P20+Ni with an optimal input settings of 12A peak current, 5.999V of servo voltage, 106.21μs pulse on time, and 50.2683μs pulse off time.