TOP THREE TAKEAWAYS FROM THE NIST INDUSTRY FORUM
MAY 17, 2018 - DAVID SIEGEL, CTO
Predictronics Chief Technology Officer David Siegel was a speaker and panelist at The National Institute of Standards and Technology (NIST) Industry Forum May 8-11 in Gaithersburg, Maryland.
During his panel, Dr. Siegel discussed monitoring and analysis technologies for prognostics and health management (PHM). He also presented several manufacturing case studies covering robot predictive monitoring, steel manufacturing, semiconductor manufacturing and stamping machines.
Below are Dr. Siegel’s top three industry forum takeaways:
1. The Need for Standards
Standardization of formats, connectivity and the integration of data sources for manufacturing PHM were heavily discussed topics throughout the event. Another major point of discussion was whether or not the presentation of health monitoring results could be standardized to facilitate end-customer interpretation. Many of the valid points made on this issue of standardization were centered around the common objective of facilitating the adoption of this technology.
2. Common Problems and Business Needs
Although the conference featured various talks from different industry and academic leaders, the use cases and problems presented were strikingly similar. In addition to monitoring applications, case studies on industrial robotics, stamping machines, product quality prediction and semiconductor manufacturing were presented. This makes it clear that different industries are facing similar problems and have several business needs in common.
3. Sustainability of PHM Manufacturing Solutions
Maintaining software and analytic solutions over time can be a challenge, mainly because solutions have a hard time providing actionable insights over the course of several years. To address this issue, machine learning models need to continue to adapt and learn over time, which is something to be considered when designing and deploying these types of technologies.
For more information and weekly updates, follow Predictronics: