Dave DollDavid Doll will be speaking at CaFFEET’2014 about the impact of Smart Factory on the supply chain. He is the Industry Principal, Facility Analytics for OSIsoft where he is accountable for setting the strategic direction for the use of advanced, real-time analytics for smarter facilities.

Could you tell us a bit more about OSIsoft and especially about the PI System?

OSIsoft is a software company, based in California, and makes the PI System (or “Pi” as it’s commonly called), an open data infrastructure for real-time operational intelligence. While the world is buzzing about the “Internet of Things,” OSIsoft has been connecting sensor data, control systems, and people for over 30 years. The PI System is the standard real-time data technology in many industries including Power and Utilities, Critical Facilities, Oil and Gas, Process Chemicals, Pharmaceutical manufacturing, Mining, and other process industries.  As new businesses expand into smart equipment, more and more assets are producing streaming data, and the PI System is being used in new ways all around the world.

How do you collect the data that are then analyzed?

OSIsoft has hundreds of supported interfaces to make sure we can collect data from any source.  There is a saying, “The good thing about standards is that there are so many to choose from.” We have a development team dedicated to providing interfaces needed to connect and collect the data – whether it’s a standard convention or a propriety protocol.

These interface “nodes” are critical not only because they connect to the data source wherever it may be, they are able to buffer the data in case of network connectivity issues.  They also compress data to reduce network traffic and enable the collection of high fidelity data.  In Smart Factories, you need to be able to reply on the data you’re using to make decisions.  Garbage in, garbage out is completely true.

And speaking of decision making, we believe it’s critical to collect and use the data is its raw fidelity.  A single PI Server can handle tens of millions of data streams at rates of several hundred thousand values per second and maintain the data in its original fidelity. No other system can handle as much data. This is vital to accurate decision making. Many other solutions rely on averaging or sampling data – say once every 5 or maybe 15 minutes – or they only store data for a few months before it has to be purged. In a manufacturing environment, these approaches actually hide many problems, which means the system you are relying on is inaccurate. We believe that inaccurate data is worse than no data at all.

OSIsoft PI system

How does the PI System contribute to smarter factories?

The key to a smarter factory is continuous improvement. There is no end-state where an operation can look around and say, “Now I’m a smart factory. I’m done.” The culture and the systems need to enable people to find and execute on the next idea and the next improvement goal.

Typical data systems have limitations that impact effectiveness. Solutions provided by equipment vendors are often limited to that vendor’s equipment – yet few enterprises are really homogenous.  Vertical solutions are also usually limited to a specific business area – a unit or process or perhaps a fixed set of reports. By solving one problem at a time, companies create an unwieldy set of software solutions with islands of data.  Gaining insight across different systems, different vendors, and different processes becomes difficult or maybe impossible. Here’s a short video of a customer explaining how the PI System helped their water utility unlock its data silos to achieve operational transparency.

By taking an “infrastructure” approach, companies can collect all operational data and enable not only “day-one” requirements, but the ideas and projects yet to come. A common question is, “What data should I collect?”  And the answer is, “All of it.”

By providing customers with easy access to all of their data, combined with analytics, notifications, and flexible visualization tools, the Smart Factory is able to meet regulatory compliance issues, minimize energy and resource consumption, optimize production and uptime, and reduce maintenance costs.

The next great idea doesn’t have to be a six month IT project; subject matter experts are empowered to create their own solutions. The enablement fosters a continuous improvement culture and that’s the real meaning of “smart factory.”

Does it impact energy management in the factory? What percentage of energy saving have you been able to reach?

A continuous improvement culture impacts everything.  By empowering people with the data and tools, systems can provide insight and drive better data-driven decision making.

Many customers are seeing 10% in reduced energy just by creating visibility and addressing low hanging fruit and smarter maintenance.  With customers that take it further into predictive analytics, they are reporting over 20-30% improvement.  And this applies to energy, water, batch optimization, asset uptime, and virtually every other data-driven aspect of manufacturing.

Could you give us some of your last success stories?

A microchip plant in Vermont is seeing a savings of $10 million per year just in water-related costs against their baseline ten years ago. Their production has increased 27% while water consumption has gone down 30%.  They used to act like water was free.  Now they are smarter.

Some customers are using the PI System to drive Demand Response solutions with their local utilities. Whereas they used to make production decisions based strictly on output optimization, now they optimize on monetary return.  One Alcoa aluminum plant reported their ROI was only 4 months. Now they collect a check each month from their local utility and go, “Cha-ching!”

To register for CaFFEET’14, go to EventBrite