Use IIoT to improve operations

Data analytics for IIoT: More data is just more data. Data analysis software is the key to extracting insights and creating value from the Industrial Internet of Things (IIoT) opportunities in production facilities. See two implementation examples and the four needs of data analytics.

By Michael Risse March 9, 2016

Thanks to a new generation of wired and wireless sensors, data can now be economically generated and gathered in quantities never previously available and then sent to process control and monitoring systems via plant networks or through the Internet. Data can then be used to improve automated real-time control and to help plant engineers and operators make better decisions regarding operation and maintenance. It is also available to data analysis software, which can be used by plant personnel to increase efficiency, diagnose equipment problems, and improve safety.

As a result of this opportunity for new insights, terms such as the Internet of Things (IoT), the Industrial Internet of Things (IIoT), big data, and Industrie 4.0 are now common. Recent advances in sensors, connectivity, and data analysis software combine to make it easier and less expensive to acquire, send, store, and analyze information. But no matter the terminology, the objective is the same: better insights faster. 

Green and brownfield applications

For the sake of simplicity, there are three common scenarios in which the IIoT, to choose one of the terms, can improve plant operations. They may be summarized as brownfield, greenfield, and servicization. All three scenarios can be described independently, but will frequently co-exist within the same plant.

Brownfield refers to existing plants and operations where new sensors are added to existing control or plant networks. Common brownfield scenarios include adding a wireless system and sensors to expand operator visibility and asset monitoring capabilities or adding sensors to replace the eyes and ears of engineers being transitioned to centralized remote monitoring centers or integrated operations facilities.

Greenfield scenarios are deployments in plants or facilities just coming online with IIoT projects. This is the most common scenario for smart city or public sector projects, and it’s where the association of IIoT with cloud-based monitoring systems originates because the project isn’t designed around an on-premise control and monitoring system. The key difference with most greenfield deployments is this lack of a control system infrastructure, but there are many examples of these systems monitoring remote tank farms, pump stations, and vehicles as a complementary system for an existing facility.

Cash in on services

Finally, "servicization" is one of several monikers describing the inclusion of a remote monitoring capability for an asset. This is the future of your personal vehicle-unless you already have a Tesla and your vehicle information is piped continuously back to the dealership-and of many assets in process plants. Vendors of pumps, valves, and many other asset types are introducing subscription services for monitoring equipment installed on customer promises. The business benefit is asset reliability and uptime, but the real driver is the opportunity for the vendor to provide expertise in asset performance and management.

All of these models follow a common architecture of sensor, communications network, and analysis that is very familiar to the process industry and noted by many prominent industry speakers. Even the difference between the brownfield and servicization models may be summarized as a question of where the data lands for analysis-at the plant or with the asset vendor.

But familiarity with the IIoT architecture and vendor advocacy over what is or isn’t a part of the IIoT architecture—cloud systems, fog computing, big data, and Internet standards being the most common examples—should not obscure the fact that there are new opportunities for improved plant performance enabled by these new technologies, and available at drastically lower price points.

The key question for manufacturers with existing plants then becomes: "How do we bring our facilities forward into a smarter state?" The answer may require any or all of the IIoT deployment models mentioned, but should always be framed in the context of the end benefit: better insights faster. 

Start with sensors

Sensors are the starting point in the data collection process. They monitor operation of the "things" in the IIoT: pumps, valves, and other assets. Their cost of implementation and use is dropping rapidly, making it cheaper to acquire more data. Plant personnel were once limited to 4 to 20 mA, HART, or various fieldbus protocols to connect these sensors to control and monitoring systems and software. But today, they can use many types of wired and wireless data connection methods, often employing multiple networks simultaneously in one plant (see Figure 1).

These sensors and connections are what enable new data from new sources to be accumulated quickly and inexpensively, and, as shown, there’s a wide range of modern networking options for deployment. Battery-powered transmitters require no signal or power wiring infrastructure, so they can be installed in locations far away from a process unit’s wired signal termination points. They can also operate safely for years in hazardous and other areas.

Wireless instrumentation also makes it possible to monitor a wide variety of equipment and systems previously too difficult or expensive to reach with wired solutions, such as 4 to 20 mA or fieldbus. Pump monitoring is an excellent example because these assets are found in most every process plant in large numbers and because their reliable operation is critical. 

Learn more about brownfield IIoT monitoring and the four requirements for data analytics.

Brownfield IIoT: using pump data

Retrofitting a pump with sensors, a network, and pump analysis software makes it possible for process plants to monitor pumps and detect problems long before a pump fails and shuts down, and it is an excellent example of a brownfield IIoT solution.

Only a few years ago, the cost of installing a dedicated online monitoring system kept it from being used on anything except the most critical pumps. But with the relative ease of adding pump condition monitoring using wireless sensor technology, online monitoring can be done on all of a plant’s important pumps (see Figure 2).

A pump monitoring system gathers data on temperature, pressure, level, and other variables in real time and transmits via a wireless mesh network to a gateway, which sends it to the control room via a hardwired link, usually Ethernet-based. There, pump monitoring software analyzes data from dozens or hundreds of pumps and alerts operators when it finds potential problems.

As to the business benefit, a pump health monitoring system can pay for itself in a matter of months. At one 250,000 bpd refinery, for example, pump monitoring systems were installed on 80 pumps throughout the complex. The annual savings was over $1.2 million after implementation, resulting in a payback period of less than six months.

Similar results have been accomplished across a range of asset types: valves, steam traps, and others. Cost/benefit ratios continue to improve because of ongoing downward cost pressure on the components used.  

Brownfield IIoT: more monitoring

In another example, a food company added temperature monitoring in ingredient/supply train railcars lined up prior to use in a production environment. Ingredient temperature variance was found to create deviations in product quality. The facility therefore needed to track this variance and change the production process to maintain desired product quality.

To address this issue, the company put wireless sensors in the railcars, which were then connected into the plant data infrastructure and analytics software for improved process outcomes. This scenario—new sensors bringing together what is now needed versus what was in the original plant design—is a very common approach because it often yields great value and quick payback. Risk is also low because the benefit can usually be proven before the IIoT deployment is implemented.

Both examples improve operational equipment effectiveness within an existing plant. Users of such technologies suggest that IIoT benefits can easily extend beyond one facility to include comparing effectiveness across plants, understanding the impact of suppliers and ingredient quality across geographies, and implementing best practices for an entire organization.

IIoT last mile: data analysis

The scenarios and components summarized share a common goal of delivering better insight faster, but it’s the last mile of IIoT that unlocks value. To deal with the challenges of data from IIoT deployments, specialized data analysis software is required to handle the high data volumes and provide the integration platform for data from disparate sources.

Data analysis software should enable rapid insights for employees who know the process, assets, and operations (see Figure 3). The software must also be easy to use so as not to require the intermediate and time consuming step of programming by developers or translation by data scientists.

The value of insight degrades quickly if developers and data scientists need to become involved as they introduce time lags to the data analysis process. If changes can’t be found and implemented in time to impact production outcomes, they have little value.

For example, the pump monitoring system described above sends data to software designed to analyze problems inherent to pumps. The results are of immediate use by maintenance engineers who understand exactly what the software is telling them.

Data analytics: four needs

Four requirements for data analytics to deliver value to IIoT implementation are: 

Productivity: This requires an application approach such that engineers and analysts can easily use the software in their investigation and discovery efforts to quickly improve outcomes. The analysis tools must fit the person who has the expertise and perspective on the plant, not the programmer or data scientist who is a software domain expert but doesn’t understand the process.

Time series expertise: At the core of any IIoT opportunity is analog data from sensors. This data is often hard to manipulate, cleanse, and contextualize. Typical manual solutions require hours of custom work in Excel or programming scripts. Data analytics tools for IIoT should facilitate and accelerate the investigation with time series data so that engineers can focus on discovery and insights instead of data wrangling or manipulation.

Data variety and contextualization: This allows sensor data to be organized by the batch or asset it is associated with, or can be easily broken into machine states and conditions within a continuous process. This is important because the most typical scenario is a blend of new data and old data, or a mix of control data from various sources. 

Collaboration support: This enables shared discovery and discussion among team members, both within a group and across one or more facilities. This requires access to the data analysis software from any browser running on any platform including PCs, tablets, and smartphones. Therefore, to get the most from acquired data, process companies should look for systems and suppliers that understand the data context as well as the business and process challenges the company faces. These suppliers can do the heavy lifting of integrating protocols, deployment, and architecture—enabling process plants to focus on getting the most value from opportunities represented by various IIoT scenarios.

Realizing the benefits of the IIoT requires a fresh look at software analytics offerings. The goal is to find a product providing a complete and agile approach to extracting insights from production data. We believe that Seeq’s data analysis software, and possibly competing products down the road, will give process experts first hand insights to their data, enabling them to customize analysis and improve production outcomes. 

Michael Risse is a vice president at Seeq Corp., a company building productivity applications for engineers and analysts that accelerate insights into industrial process data. He was formerly a consultant with big data platform and application companies and prior to that worked with Microsoft for 20 years. Risse is a graduate of the University of Wisconsin at Madison and lives in Seattle. Edited by Mark T. Hoske, content manager, Control Engineering, CFE Media,


Key concepts

  • IIoT in action: Data acquisition is easier than ever, with many wired and wireless networks available.
  • Important assets, such as pumps in a refinery as well as valves, steam traps can be monitored for maintenance prior to failure.
  • Ingredient tracking improves product quality.

Consider this

Information integration works best when information is put to use within an appropriate period of time. 

ONLINE extra

This online version is longer than what appears in the March Control Engineering print and digital edition.

See other IIoT coverage below.

Author Bio: Michael Risse is the CMO and vice president at Seeq Corporation, a company building advanced analytics applications for engineers and analysts that accelerate insights into industrial process data. He was formerly a consultant with big data platform and application companies, and prior to that worked with Microsoft for 20 years. Michael is a graduate of the University of Wisconsin at Madison, and he lives in Seattle.