Is the PI System an application development platform?
The PI System is a data infrastructure for industrial operations. It's been enhanced over the years to include contextualization and digital modeling (Asset Framework) and data visualization (PI Vision) but the mission remains: customers use it to tap into the vital signs of their organization and get a single, comprehensive version of the truth. Think of us as a nervous system for the working world.
But many customers additionally use the PI System to conduct equipment diagnostics and predictive analytics (Syncrude, Sempra Renewables), energy efficiency (Covestro), regulatory compliance (Barrick Gold) or facilities management (NASA, Harvard). Thus, it has application-like qualities that go beyond what a historian or data management platform would ordinarily provide.
And in the past few years we've seen the growth of another use case: the PI System as a foundation for in-house applications. The PI System essentially serves as a data engine to accelerate their own software development. In the past few PI Worlds we've seen:
Air Liquide's SIO: SIO lets Air Liquide tune plants to an economically optimal production point by analyzing factors such as current performance, inventory and past performance. The PI System, PI Vision and other PI System technology feed into SIO. Besides using SIO internally, Air Liquide and its subsidiary Alizent have developed digital services for managing the chemical plants of its customers.
Uniper's Tiresias: advanced analytics for turbines and other complex equipment. Uniper has also received a 'green light' to explore commercializing it, i.e. selling or licensing the technology to others. If Uniper goes through with those plans, it could serve as a blueprint for utilities looking to expand beyond therms and electrons and soar out of the so-called utility death spiral.
Petronas' PROTEAN: an in-house app for predictive maintenance and real time monitoring for IIoT in oil and gas. Petronas engineers have proposed charging OEMs for data and analyses developed by PROTEAN (which stands for Petronas Rotating Equipment Analysis) that they use to improve their own products.
"Why pay the OEM for remote monitoring and diagnostic services when the OEM is receiving the data for his R&D," said Gavin Halls of Petronas at PI World London, 2017, to applause. "The OEM should be paying the operator, us, for the data."
PJM's DIMA: DIMA gives remote maintenance technicians a control room view on their cell phones and laptops. PJM also has mulled commercialization.
Evides GAMEs: an application sandbox that combines operational and equipment data from the PI System with geographical and other information to give employees a richer sense of Evides' operations. In the first year, Evides created 50 GAMEs functionalities and has been working to graduate around 20 to Serious GAMEs, where they could be deployed across operations. (the picture shows Evides tracking the progress of a burst water main.)
“There is useful information hidden in the data,” said Jan Urbanus of Evides. “A picture is worth a 1,000 words, or in this case 1,000 figures.”
Rezonance: Operations management and prescriptive decision making from Dalkia, the district heating subsidiary of EDF. With Rezonance, Dalkia believes it can improve the efficiency of district heating systems by 3 percent and increase renewable biomass sources by 5 percent by exploiting data feeds to more efficiently manage peak heat demand. Dalkia has deployed Rezonance at 55 networks and will expand it to 100 over the next two years.
Stream. A application initiative for predicting sewer blockages, energy consumption and other tasks from United Utilities, several UK universities and OSIsoft. A neural network-based application created by United and the University of Exeter, for instance, can predict water demand 24 hours ahead with 98% accuracy: the work led to United winning the Data Demon award at Utility Week.
Aorta. Thames Water. Real time diagnostics and analytics. In one instance, Thames--which serves 15 million customers and processes enough water and wastewater a year to fill Sydney Harbor six times--thought it saw a maintenance problem. A pump was consuming power, but not performing much work. It turned out to be caused by a localized gaffe in operating procedure. The correction saved Thames 10,000 GBP in energy costs a year.
Tomoni. Mitsubishi Hitachi Power Systems has created an add-on service that monitors the health of power generating assets.
And these are just the examples with catchy product names. Let's not forget the City of Riverside, TEPCO, KEPCO and others. (I've also got a sequel coming of applications that partners have developed that leverage the PI System as a data engine. If you have suggestions, send them along.)
What's the upshot? Internal software development is alive and well and the PI System is keeping it thriving by reducing the time, cost and busywork to develop tailored apps. Industrial customers rely on third party software, but there's a persistent tension. Often, the third party applications don't meet their specific needs and have to be tailored. Some industries--like oil and gas-are also worried that handing off their analytics and software requirements to giants like Amazon could become an uncomfortable trade off, as the Wall Street Journal reported.
With the PI System, industry can have it both ways.