Machine Learning-Based Predictive Modeling of the EDM Process
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Conventional machining is often not ideal for certain machining tasks. Therefore, electro-discharge machining (EDM) is one of the best choices due to its extensive capability to cut electrically conductive hard materials. The interactions among the machining parameters (peak current ( I p ), spark voltage (Sv), pulse on time ( T on ), and powder concentration ( C p ) in the dielectric fluid) improve the sensitivity of these parameters. Additionally, the cryogenic treatment (CT) of electrodes adds complexity to the process, restricting the effectiveness of traditional simulation approaches for numerical modeling interrelationships. This complexity necessitates experimental exploration and systematic investigation to model issues in EDM (such as overcut (OC), Kerf width (KW), tool wear rate (TWR), and electrode wear rate (EWR)). To achieve this, the development and evaluation of accurate machine learning models are required. This chapter delves into the theoretical foundations, methodologies, and applications of various models, offering a comprehensive overview of their capabilities and limitations. It starts by exploring the fundamental principles of machine learning, differentiating between supervised, unsupervised, and reinforcement learning paradigms. Thus, supervised learning algorithms (artificial neural network (ANN), neural network (NN), Multi-Objective Genetic Algorithm (MOGA), Particle Swarm Optimization (PSO), convolutional neural network (CNN), etc.) are used to predict the correlation between the dependent (performance parameters) and independent (process parameters). This chapter illustrates the versatility of predictive models with case studies from the realm of EDM. Previous studies have demonstrated that predictive models can achieve an accuracy of up to 99%.
Fields of Study & Subjects
Advanced Machining and Optimization Techniques
Advanced machining processes and optimization
Machine Learning in Materials Science
Publication details
- Journal / source
- CRC Press eBooks
- Pages
- 74–87
- Publication date
- 30 March 2026
- Language
- EN
- Document type
- Book chapter