Manufacturing industries are experiencing significant transformations in recent times due to the onset of Industry 4.0 concepts. The newer set of technological solutions necessitate real-time monitoring of manufacturing processes using sensors followed by data analytics to evaluate the status and adjustment of parameters. It will be necessary to have appropriate process knowledge embedded into the decision-making system for the adjustment of settings. The primary objective of Smart Manufacturing activities at IIT Jodhpur is to conduct interdisciplinary and translational research relevant to the next generation shop floors. The overall focus is on improving the practicing engineer’s ability to produce highly accurate and precise components on time. The focus is on strong analytical/numerical modelling capabilities coupled with fundamentals accompanied by experimental techniques and data analysis skills, including physics-based machine learning. Some of the ongoing research efforts include development of physics-guided data-driven models to predict the performance of metal cutting operations, vision-based smart machining platform for in-situ inspection and monitoring and exploring solutions for leapfrogging from Industry 2.0 to Industry 4.0.
Cutting force is the primary source of multiple disturbances contributing to the deterioration of component accuracy significantly during metal removal operations. The prediction, monitoring, and control of cutting force are imperative to avoid or minimize faults such as tool breakage, tool wear, selection of cutting parameters, fixture errors, etc. As cutting force is linked with multiple process faults during metal removal operations, it is essential to have a reliable predictive model to assist in the decision making related to process faults. The group is attempting to develop reliable process models for end milling operation, which is commonly employed in most of the manufacturing industries to fabricate complex shapes in a variety of materials at higher accuracy and productivity. One of the recent works attempted to combine physics-based approaches with machine learning to enhance the prediction accuracy of computational models. It has been observed that the developed hybrid model can predict cutting forces accurately over a wide range of cutting conditions. It is planned to strengthen the present models subsequently by using new generation networks for better realization of relationships.
Department of Mechanical Engineering