Predictive Maintenance For Industry 4.0 - Key Challenges And How Machine Learning Can Solve Them
Santa Clara Convention Center, CA, USA
Grand Ballroom A
17:15 - 17:35
The presentation will focus on real world scenarios where anomaly detection and machine learning has been used to predict failures in industrial assets. Traditional approaches to analytics cannot be applied to IIOT scenarios as the scale and the nature of data is very different. Further, failures are very few. Hence, there is not enough training data to create accurate predictive models for failures. Therefore, to ensure that accurate predictive models are created, a new approach is needed which operates at scale and models for asset behaviour. In such a scenario, failure becomes a context in the behaviour and has a better accuracy since more information is available to train. Lastly, the way IOT projects are being designed are setting them up for failure right at the start. The last part of the presentation focuses on what is needed to build a successful IOT project structure and roadmap.
Abhishek Tandon brings 10 years of experience in applying big data analytics and Machine Learning solutions to a variety of Industrial IOT use cases, such as predicting - machine failure, inventory stock-outs, quality control issues etc. Through early detection using predictive analytics he has helped Fortune 100 global industrial companies save millions of dollars. He also brings significant experience in understanding business impact of these decisions.
DataRPM is an independent subsidiary of multi-million dollar software company - Progress Software. It provides a Cognitive Predictive Maintenance Solution that helps reduce asset failures, optimize yield and improve product quality. Through its patent pending meta learning on machine learning approach DataRPM automates the process of finding anomalies and predicting failures. It reduces the amount of time to deploy a predictive maintenance solution in production by months. DataRPM has several fortune 100 clients who are saving millions of dollars by introducing its Cognitive approach to Predictive Maintenance