Predictive Maintenance: Zero to Deployment in Manufacturing
Predictive Maintenance: Zero to Deployment in Manufacturing

Abstract: 

The manufacturing industry is going through it’s IV industrial revolution where machines are connected, and data is harvested to discover deeper insights and solved problems to achieve a competitive edge in the market. There are various applications in industry 4.0, such as machine connectivity, machine productivity monitoring, predictive maintenance, predictive quality, inventory optimization, supply chain optimization, data discovery, and the possibilities are unlimited. In one estimate, it was found that by 2020, over 50 billion machines would be connected to the network and could generate as much as $3.7 Trillion in value. With all these estimated benefits one of the common problems manufacturers still face today with Industry 4.0 is connecting its legacy and process critical machines where it’s too expensive to integrate machine connectivity, and in some cases it’s impossible to achieve that machine connectivity. This lack of connectivity in some cases leads lack of ROI and in other cases to a failed Industry 4.0 activity. Hence, there is a significant need for more reliable and robust approaches to achieve the benefits of various applications. Among these applications in manufacturing, predictive maintenance is seen as the magic unicorn for achieving a competitive edge. In a study by PWC, it is estimated that by implementing predictive maintenance, manufacturing industries could reduce cost by 12 percent, improve uptime by 9 percent, reduce various risks by 14 percent and extend the life of the machine by 20 percent. By integrating the philosophy of lean manufacturing and “A small success eventually leads to bigger success,” one case of predictive maintenance (PdM) led to a global scale implementation by developing new technologies using ML & DL models and deploying them on an industrial scale. This case study discusses different levels of analytics, basics of predictive maintenance, anomaly detection, remaining useful life prediction techniques, combining data and SME knowledge, ML models, and how they can be deployed in manufacturing for different use cases. This case study also discusses various modeling and deployment challenges of implementing predictive maintenance in manufacturing.

Bio: 

Dr. Nagdev Amruthnath currently works as Data Scientist III at DENSO North America. He has earned his Ph.D. and Master’s in Industrial Engineering Department, Western Michigan University (WMU), and a Bachelor's degree in Information Science and Engineering from Visvesvaraya Institute of Technology, Karnataka, India. He has four years of experience working in the manufacturing industry, specializing in the implementation of lean manufacturing, JIT technologies, and production system, three years of full-stack data science experience in manufacturing and undergraduate and graduate teaching experience. He has also authored in journal and conference proceeding publications in production flow analysis, ergonomics, machine learning, and wireless sensor networks. Currently, his research focus is on developing new machine learning models and AI technologies for manufacturing applications. Nagdev continues to serve as a reviewer of scholarly journals, which includes the Journal of Electrical Engineering, IEEE Transactions on Reliability, and Machine Learning and Applications: An International Journal. He has also open-sourced all his projects, including various R-packages on GitHub.