Title: Data-Driven Modelling: Revolutionizing Manufacturing Excellence - A Sand Casting Case Study
Speaker: Dr. Janak Suthar, Data Analyst at Zensung Pvt Ltd, Mumbai.
Day, Date, and Time: Thursday, August 24th, 2023, 9:30 AM to 10:30 AM.
Venue: Seminar room IE 211, Second Floor, IEOR Building.
Abstract: Data-driven modelling in manufacturing is a transformative approach that utilizes data and advanced analytics to revolutionize the optimization of processes and the improvement of product quality. In our study on marine propeller manufacturing, we harnessed vast and diverse data collected from various process variables and quality parameters to develop data-driven models, providing valuable insights and driving informed decision-making for proactive management of manufacturing operations.
In the realm of foundry environments, where casting processes are inherently complex and variable, data-driven modelling becomes even more critical. The interactions among multiple process variables present challenges in identifying the most influential factors impacting product quality. Conventional approaches, often limited in scope, may struggle to effectively control process variability, leading to diminished product quality and escalated production costs.
In our case study, meticulously explored machine learning to build models linking quality characteristics, such as surface roughness and mechanical properties, to their corresponding casting process variables. By leveraging an ensemble learning approach with the robust random forest algorithm, we achieved remarkable precision in predicting surface roughness and mechanical properties. Additionally, our expertly calibrated classifier delivered superior outcomes in identifying defects.
The data-driven modelling outcomes empowered our foundry with several pivotal advantages:
Predict Quality Parameters: By identified robust relationships between process variables and quality characteristics, foundry could accurately predict product quality. This proactive approach enabled early issue detection, reducing defects, rejections, and generating cost savings while enhancing customer satisfaction.
Monitor and Control Processes: Real-time process variable monitoring allowed us to exercise precise control over the casting process. Continually analysing data and capitalizing on data-driven insights, we adeptly adjusted operating conditions, ensuring consistent desired quality levels while complying with stringent standards.
Informed Decision-Making: Our data-driven approach equipped us to make informed decisions regarding process optimization and product improvements. Armed with a comprehensive understanding of critical process variables and their implications on product quality, we optimized processes and boosted product performance.
Informed Decision-Making: Our data-driven approach equipped us to make informed decisions regarding process optimization and product improvements. Armed with a comprehensive understanding of critical process variables and their implications on product quality, we optimized processes and boosted product performance.
Our case study exemplifies the immense potential of data-driven modelling in empowering industries to achieve superior product quality, optimize operations, and bolster competitiveness. Embracing data- driven approaches is key to manufacturing excellence in the digital age.
Speaker Bio: Janak Suthar is a skilled Data Analyst at Zensung Pvt Ltd, Mumbai, specializing in operations management. His expertise lies in developing ML/DL models for AI-based claim settlement, vehicle damage detection, and driving score generation. As a Visiting Professor at DSIMS Mumbai, he adeptly taught operation analytics with Python and data visualization. He also served as an Assistant Professor at SJCEM, Palghar, where he taught machine learning, basic statistics, Production planning and control, and project management.
Janak holds a PhD in Operations Management from NITIE, Mumbai, and a Master's in Engineering from Mumbai University. His research focuses on quality management, manufacturing process optimization, and AI/ML applications in manufacturing. His contributions have been published in reputable ABDC, SCI, and ABS3 indexed journals. With diverse teaching experience in operation analytics, basic statistics, Machine Learning using Python, and operation management.