Open Access Open Access  Restricted Access Subscription Access

Artificial Intelligence and Reinforcement Learning for Real-Time Process Control in Non-Traditional Machining

Riyon Ignatius, Sampreeth N Ganganagoudar, Saarthak Dubey, Rishi Chaudhari

Abstract


Manufacturing industries face mounting pressure to produce precision components from exotic materials while maintaining cost-effectiveness. Non-traditional machining (NTM) techniques—spanning electrical discharge, laser-based, and ultrasonic approaches—provide pathways to handle materials conventional methods cannot address. The fundamental limitation plaguing current NTM adoption lies in maintaining process stability when parameters shift during production runs. This research presents an adaptive control strategy leveraging reinforcement learning algorithms to manage machining operations in real-time by continuously monitoring sensor inputs and modifying operational settings automatically. Using deep Q-learning networks trained on industrial sensor data, the system learns to balance competing manufacturing objectives: achieving target surface finishes while limiting tool degradation and preventing thermal distortion. On EDM and LBM operations, the developed controller demonstrated a 23% consistency improvement in surface characteristics and reduced overall variability by 31% versus conventional parameter-fixed approaches. Processing decisions complete within 50-millisecond windows, enabling the system to respond to sudden process changes. This investigation validates the utility of adaptive RL-based controllers for advancing manufacturing automation capabilities.


Full Text:

PDF

References


Mnih, V., Kavukcuoglu K., et al. (2013). Playing Atari with deep reinforcement learning. arXiv preprint arXiv:1312.5602.

Deghatipour, S., & Nair, R. (2020). Machine learning approaches for predicting machining performance. Advances in Manufacturing Technology Journal, 28(3), 156-172.

Kuppuswamy, R., & Srinivasan, M. (2019). Challenges in electrical discharge machining process control. Precision Engineering International, 42(1), 78-94.

Watkins, C. J., & Dayan, P. (1992). Q-learning. Machine Learning, 8(3-4), 279-292.

Zhao, J., & Thompson, M. (2021). Sensor fusion for adaptive machining control. IEEE Transactions on Industrial Systems, 15(4), 512-528.

Patel, S., Kumar, V., & Singh, A. (2018). Metaheuristic parameter optimization in precision machining. Journal of Manufacturing Excellence, 31(2), 45-61.

Chen, L., & Wang, H. (2020). Adaptive feedback control mechanisms in non-traditional machining. Precision Manufacturing Review, 38(7), 234-251.

Rodriguez, M., & Garcia, P. (2019). Surface integrity in electrochemical machining operations. International Journal of Advanced Manufacturing, 51(9), 891-908.

Kim, T., Lee, S., & Park, J. (2021). Neural network applications for manufacturing process optimization. Smart Manufacturing Today, 44(1), 112-129.

Hassan, A., & Mahmoud, S. (2020). Real-time control strategies for complex manufacturing systems. Advanced Automation Quarterly, 25(3), 178-195.


Refbacks

  • There are currently no refbacks.