Continuing from Part 1
where we introduced the fit-for-purpose layered approach to analytics, in this session we will review the built-in data engineering capabilities in the PI System.
Raw time-series sensor data is often transformed to another metric – say, convert flow rates to standard volumes and tank levels, derive statistical aggregates such as hourly/daily averages, hourly/daily totals, convert using first principles physics/math equations, and others. Also, datasets can be filtered for process/equipment events - turbine trip, compressor failure, pharmaceutical batch, catalyst poisoning, heat-exchanger fouling etc. And, you can incorporate material flow related latencies in the dataset, for example, the 14-hour lag as wood chips are cooked and flows through various process units to be converted to paper or in the metals industry where iron ore is mined and goes through a number of steps to be converted to ingots, thus allowing you to capture SME (subject-matter-expertise) knowledge.
With illustrative use cases, we will review how such preprocessing for industrial time-series data is critical for data visualization and for advanced analytics, including machine learning with R/Python libraries or cloud platforms such as Microsoft Azure, Amazon AWS and Google GCP.
About the Presenters
Gopal has been involved in several roles at OSIsoft for over 20 years and has been involved with software development, technical and sales support, and field services. Prior to being a Solutions Architect, he was a Product Manager with a focus on Enterprise and Asset Integration. Gopalkrishnan has a master's degree in chemical engineering, continuing education in business administration, and is a registered Professional Engineer in Pennsylvania. He is also active in the MESA Technical Committee and the MESA Continuous Process Industry Special Interest Group, and active in topics such as data mining, energy efficiency, manufacturing intelligence, sustainability, including green initiatives in facilities and data centers.
Curt Hertler is a Global Solutions Architect based in Cleveland. He has had several roles at OSIsoft, including the establishment of the OSIsoft’s Developers Network in 1989, which might have been the world’s first app store. Before joining OSIsoft, Curt had 19 years of petroleum refining experience working for Marathon Oil and British Petroleum. In addition to having advanced control and process engineering responsibilities, he has taken on projects to develop and operationalize analytical models for online process optimization. Curt represented BP for five years on the National Petroleum Refiners Association Computer Applications Committee. He graduated with a BS and MS in Chemical Engineering from the University of Michigan.