A Primer on the Industrial Internet for Downstream
In recent years, the term "Industrial Internet" has entered the lexicons of myriad businesses. The downstream oil and gas industry is no exception to this trend. The development of the Industrial Internet, which relies on the collection of larger and more robust sets of operational data also known as "Big Data," could help refineries and petrochemical plants run more efficiently, safely and economically.
Alan Hinchman, infrastructure global market director with General Electric's Intelligent Platforms business, recently gave DownstreamToday an overview of the Industrial Internet and its implications for refiners, petrochemical manufacturers and others in the downstream. A transcript of the conversation follows.
DownstreamToday: What is the Industrial Internet, and how does it differ from the "regular" Internet?
Alan Hinchman: The Industrial Internet combines and maximizes the capabilities of intelligent machines, advanced analytics and people at work. It refers to complex machinery communicating automatically with networked sensors and relaying the information to plant operators to enhance operational efficiency, maximize assets and reduce downtime. Using big data and real-time predictive analytics allows infrastructure-heavy industries to become much more efficient and allows operators to make more informed decisions regarding their assets.
DownstreamToday: How does the Industrial Internet function?
Hinchman: The Industrial Internet functions via a series of networked sensors, controls and software that connect all the equipment within an operation. The sensors monitor all aspects of a machine's functionality and "learn" what normal activity is, alerting operators when the data moves out of a specified comfort zone. All data is aggregated and viewed from a monitoring interface, so operators or data scientists can analyze a complete snapshot of asset activity or inactivity. Problems are detected much faster and can save substantial amounts of money for an operation by avoiding downtime and scheduling maintenance at a convenient time avoiding an emergency shutdown.
Big Data: A term that refers to very large, robust datasets that are acquired and used very quickly – often in real time. Such datasets, which are in essence the "feedstock" of the Industrial Internet, are too complex to be processed with conventional data processing software.
Cloud-based application: An Internet-based application or service that is hosted remotely from the user's physical location.
Hadoop: Administered by the Apache Software Foundation, Hadoop is an open-source system for processing "Big Data." The foundation defines the software library project as "a framework that allows for the distributed processing of large data sets across clusters of computers using a simple programming model."
Historian: A software application that collects and stores all of a facility's data – obtained via sensors deployed at different onsite locations – and makes it available for analysis by data scientists.
Industrial Internet: The integration of complex physical machinery with networked sensors and software. The Industrial Internet applies machine learning, Big Data and machine-to-machine communication to retrieve data from machines, analyze it (usually in real-time) and tailor operations accordingly.
Machine learning: Using algorithms so that a computer can analyze data and "learn" what to do when it detects a given pattern or event.
DownstreamToday: What are some applications of the Industrial Internet in the downstream oil and gas industry?
- Improved asset uptime. Predictive analytics provide superior monitoring over traditional alarm set points. Software takes multiple inputs and predicts what the outcome should be, then compares that to the actual. If a difference between the real and the calculated are detected, the operator is alerted, resulting in smaller problems being detected before they become major problems.
- Better asset utilization. When machines are connected it results in improved operations because the support systems can be maximized. These assets can also be compared across the entire enterprise so operators can understand if they are operating sub par to the fleet.
- Simplified environmental reporting. With smarter discharge and air monitoring, improved steam control can be monitored and discharges can be tracked closer and less expensively.
DownstreamToday: How can the Industrial Internet influence cost control and efficiency at refineries, petrochemical plants and in other downstream oil and gas settings?
Hinchman: When oil and gas machinery goes down, it can be overwhelmingly expensive to fix. The cost of downtime can have a huge impact on a bottom line. Since the technologies of the Industrial Internet give advanced warning when a piece of equipment needs maintenance or will fail, operators can plan on when to shut down operations, fix the equipment, or replace parts.
DownstreamToday: What are some of the key trends in the deployment of the Industrial Internet downstream?
- Mobility. Operators want a "single view of data." This means that everyone needs to be looking at the same data no matter where they are in the operation. This can be accomplished with mobile applications that allow the crew in the control room to see the same information as the crew on the rig.
- Centralized data historization. Lower cost of automation and monitoring drive higher data collection and analysis. Having one historian application that collects and stores all of your data and makes it available for analysis by data scientists can have a huge impact on bottom line.
- Increased cloud-based applications. Cloud applications are the wave of the future. Access to information no matter where you are is made possible by the Industrial Internet and the advent of cloud.
DownstreamToday: On a human level, what impact is the Industrial Internet having on how downstream oil and gas personnel do their jobs?
Hinchman: The emergence of the Industrial Internet has given downstream personnel much more detailed information to work with when managing their infrastructure. Personnel can spend much more time analyzing data and responding to problems ahead of time and much less time responding to emergencies and unplanned downtime.
DownstreamToday: Looking ahead 5 to 10 years, what might be different about the downstream oil and gas industry as a result of the Industrial Internet?
Hinchman: Unprecedented operational information from the rise of Industrial Big Data powered by the Industrial Internet is about gathering much more data than we have ever been able to accumulate— from multiple sources, over longer periods of time—and doing it much more quickly. Comparing and correlating years of diverse historical data to real-time data allows for a myriad of new analysis possibilities, allowing operators to rapidly detect trends and patterns never before possible to better understand how equipment and processes are running versus how they should be running and to help businesses make even better and quicker decisions to improve operational performance. To accomplish this, GE built an Industrial Big Data Historian built on Apache™ Hadoop. Hadoop is a technology specifically designed to handle large data sets by clustering large numbers of low-cost commodity computers together to act as one, in the "cloud." This makes it possible to analyze much more data, to store it economically and to get answers in a much more expedited fashion. A historian built on Hadoop can compare data across time and across the enterprise, and it can scale both horizontally for any data volume and variety, and vertically for any velocity.
Matthew V. Veazey has written about the upstream and downstream O&G sectors for more than a decade. Email Matthew at email@example.com. Twitter: @The_Mattalyst