Cost Reduction for Electrochemical Machining Through Hybrid Machine Learning Simulation

Electrochemical machining has multiple benefits for manufacturing of turbomachinery including high quality surface finish, an ability to process advanced metal alloys, no imparted stress, and scalability to deliver production affordability. Unfortunately, ECM’s adoption is limited by the up-front time, cost, and risk to demonstrate an ECM solution. This problem is particularly difficult during new product development when customers require frequent design changes and rapid iteration. Even when ECM can produce the best quality part at the lowest price, it can lose out to conventional machining methods during critical product development milestones, especially with small production volumes on which to amortize the development expenses. The upfront investment in ECM is driven by the process complexity. It is a multi-physics environment involving fluid flow, current density distribution, electrochemical reactions, heat transfer, and gas evolution – all of which change dynamically over time. Most ECM practitioners rely on experience and repeated iterations of electrode design and process parameter changes to achieve the desired outcomes. More advanced methods such as pulsed or precision ECM (PECM) can reduce that design and iteration burden but come at the expense of significant additional cycle time. Others have proposed and developed multi-physics simulation methods however these approaches suffer from long solution times and inaccuracy. Therefore, to address this challenge, Voxel, in collaboration with Notre Dame University, proposes a hybrid simulation software tool combining multi-physics simulation and machine learning which can be used to predict and optimize the ECM process parameters and electrode design. The resultant solution has the potential to reduce process development times from months to days, enabling the use of ECM for a variety of Navy, DoD, and commercial applications.