Mobile Navigation

Chemical Engineering

View Comments PDF

Process Control: Optimization at the Edge

| By Mary Page Bailey

The key technologies shaping industrial process control all work in tandem to support the ultimate goal of plant optimization

Process control technologies are at the heart of all chemical process facilities, providing the foundation for safe and efficient plants. As traditional process control systems are evolving to integrate more automated elements, including artificial intelligence (AI), edge connectivity, faster communications and “smarter” field devices and sensors, plants can reach new levels of optimization on the path toward more autonomous operations (Figure 1).

FIGURE 1. The advancement of process control involves integrating the physical control devices with digital innovations via cloud and edge connectivity, sophisticated data analytics and artificial intelligence

Defining autonomy

“Autonomy is a part of a larger topic, which is the optimization of assets. When we talked about optimization in the past, it was about getting better sensors with higher accuracy and repeatability, but when we talk about optimization today, including the journey to autonomy, it’s more about generating the right data and how to make that data available in a way that can actually be used to further optimize operations,” says Claudio Fayad, vice president of technology, process systems and solutions at Emerson Co. (St. Louis, Mo.; www.emerson.com).

Kevin Finnan, advisor of market intelligence and strategy at Yokogawa Corp. of America (Sugar Land, Tex.; www.yokogawa.com/us) defines autonomous operations as “the state in which assets and operations throughout possess learning and adaptive capabilities that enable responses with minimal human interaction, thus empowering operators to perform higher-level optimization tasks.” In theory, with a well-designed autonomous control platform, operators and subject-matter experts would not even need to be present at a production facility and could instead reside at an integrated operations center serving multiple facilities globally.

Yokogawa has executed several proof-of-concept projects for fully autonomous operations. In 2022, the company collaborated with Eneos Materials to launch a field test using a reinforcement-learning AI platform to autonomously operate a chemical manufacturing plant for 35 days. “This was a world-first project. The test confirmed that reinforcement-learning AI could be safely applied in an actual plant and demonstrated that the technology can control operations that had been beyond the capabilities of other control technologies. Until now, such tasks have necessitated the manual operation of control valves based on the judgment of plant operators,” says Finnan. Since the initial test in 2022, the application has now run successfully for over a year and has been officially adopted by Eneos at the facility. For more information on this project, see Autonomous Control of a Distillation Column, Chem. Eng., July 2023, p. 10.

Finnan also cites examples where AI-based asset-management can be used to replace traditional process control technologies, such as proportional–integral–derivative (PID) loops, to improve energy efficiency. For instance, Yokogawa offers a holistic control system to manage fired assets, such as boilers and furnaces, to increase fuel efficiency, extend asset life and minimize downtime. “The PID control in this platform could be replaced by reinforcement-learning AI technology to add additional stability that conserves more energy,” he notes.

 

Next-generation networks

In another proof-of-concept, Yokogawa collaborated with NTT Domoco to test an autonomous AI and cloud-based control solution, along with a 5G mobile communications network. The test successfully controlled a simulated process plant operation and demonstrated that 5G is suitable for remote control of actual plant processes. “In conjunction with AI and cloud, 5G communication, which offers low latency and the capability to connect numerous devices, promises to be a core technology for autonomous operations. Compared to 4G communication, the test demonstrated that, especially with high-speed control, 5G provides lower latency and less overshoot relative to the target,” says Finnan. He also mentions that 5G can handle control cycles as short as 0.2 s, providing faster process control that can result in higher product quality and energy efficiency.

A partnership between Nokia Corp. (Espoo, Finland; www.nokia.com) and i.safe Mobile GmbH (Lauda-Koenigshofen, Germany; www.isafe-mobile.com) will see i.safe Mobile’s IS540.x intrinsically safe smartphone — said to be the world’s first 5G-enabled smartphone for hazardous-area use — integrated with Nokia’s end-to-end Nokia Digital Automation Cloud (DAC) platform, as well as the company’s private wireless infrastructure and edge applications. “In hazardous areas, it is vital that workers are connected using devices that are Ex certified. As a result of this international partnership, Nokia will extend its industrial device portfolio with our products to support the connectivity needs of organizations in the chemical, pharmaceutical and mining sectors,” according to i.safe Mobile. The company emphasizes the benefits of industrial 5G communication, including the secure integration and consolidation of complex workflows. The company is also developing the industry’s first 5G-enabled Android and Windows tablets (Figure 2) for Zone 1/21 with charging options for hazardous areas, which are expected to be available later in 2024.

 

FIGURE 2. For plants to fully take advantage of the faster response enabled by 5G communication, it is crucial that mobile communication devices be designed for hazardous-area use

Data in context

It is well-known across industry that the usefulness of process data is limited if there is little or no context given. Because of the way that legacy distributed control systems (DCS) have been created, there are often many kinds of independent systems where the data and its associated models may only exist within that system, with no context linking it to other parts of the plant. “These systems are very good at transferring data like temperature or flow, but they are not very good at transferring the meaning of the data. So even when you build a system of systems, to try to bring the data together, it doesn’t include the context that allows an autonomous system to make sense of the data,” says Fayad. One way that plants can move away from this “system of systems” approach is with an integrated control platform that includes edge connectivity. The main principle of edge computing is to bring data from disparate sources closer to the actual points where it will be processed.

Emerson recently launched its DeltaV Edge Environment (Figure 3), which replicates the data gathered from the DCS and makes it available for other applications, such as data analysis or optimization platforms. “The Edge Environment is connected to the rest of the DeltaV control system, so there is only a unidirectional path for the data from the controllers to the Edge Environment. People can actually view and manipulate the data with context, making queries about the status of the plant or feeding a digital twin, without actually needing any physical access to the controllers and the critical parts of the DeltaV system,” says Fayad. While obviously promoting improved data utilization, this setup is also a boon for plant security, since it provides broader access to DCS data without expanding access to the control system itself.

FIGURE 3. The DeltaV Edge Environment consolidates DSC data into a single outbound, unidirectional connection to maximize accessibility to secure, real-time contextualized data

Another facet of next-generation control systems is the integration of AI and machine learning. Through Emerson’s partnership with Aspen Technology, Inc. (Bedford, Mass. www.aspentech.com), the MTell platform helps to support advanced automation through AI-supported anomaly detection. “Models are created to understand how a certain process needs to operate, and if something goes wrong, they can identify when the process is out of its operating envelope. If we can get enough data in a secure way to have a digital twin and they can check if the plant is operating with the right parameters, and adjusting setpoints as needed, that is how we close the loop to the autonomous operation,” explains Fayad.

Furthermore, says Chris Jones, LNG and midstream vertical leader at Honeywell Inernational, Inc. (Charlotte, N.C.; www.honeywell.com), smart edge devices are driving more insight and intelligence at the sensor level, which will, in turn, drive more decision-making away from human interaction, placing more reliance on smart sensors. “Improved integration of disparate systems seems next on the horizon. If we compare the process control industry to the auto industry, we can see how previously separate technologies like adaptive cruise control, lane keep assist and blindspot monitoring were integrated to create self-driving cars. In process control, we should see integration of process/operations parameters with emissions data with asset data. The processing and control of these parameters as a single enterprise, versus separate maintenance and emissions programs, will drive the next generation of business performance,” adds Jones.

As part of their process-control objectives, many manufacturers are implementing overall equipment effectiveness (OEE) analytics (Figure 4), which can quantify production losses due to availability, performance and quality. Such process-control improvement can be the key to enhanced process performance that can maintain operations at target rates, even in the face of process upsets and quality-related OEE losses, says John Cox, principal analytics engineer at Seeq Corp. (Seattle, Wash.; www.seeq.com).

FIGURE 4. Overall equipment effectiveness is a useful metric to quantify production losses, which can be easily visualized using smart data analytics

Being able to intelligently analyze, identify and respond to upset conditions is a major step toward the data contextualization required for more autonomous operations. “Analytics insights drive autonomous operation through actionable monitoring and diagnostics focused on process-control performance and service factors, and crew-by-crew interactions with the regulatory control layer,” explains Cox. Once implemented, these monitoring and diagnostics capabilities can be easily scaled across controllers, control valves and similar processing equipment to help rapidly identify and prioritize optimization opportunities. “Analytics can be used to contextualize everything from leaky control valves to site-wide carbon intensity, aggregated by product and energy or utility streams. Process-control expertise can transform these insights into process-control improvements,” says Cox.

 

Cloud-based operations

Cloud computing technologies are making vast amounts of process data more accessible than ever before, helping engineers to make better-informed choices. “Many products are contributing toward fully automated industrial operations. Cloud-based platforms in particular enhance centralized control and data analysis, thus improving decision-making in autonomous operations. For instance, cloud solutions streamline engineering and simulation in automation, while intelligent analysis of field-device data aids in optimizing plants,” says Rebecca Vangenechten, head of automation and engineering, process industries, at Siemens AG (Munich, Germany; www.siemens.com). This unprecedented access to data means that engineers from any location can connect seamlessly and work on several projects at the same time. “This is where the advantages of a web-based control system like Simatic PCS neo come into play. This system facilitates cross-regional and cross-functional collaboration, allowing rapid development of new systems,” says Vangenechten.

Siemens recently demonstrated this global reach in a partnership with Dow (Midland, Mich.; www.dow.com) at MxD, an advanced manufacturing institute and innovation center in Chicago, Ill., with a new test bed for the process industries equipped with Simatic PCS neo to exhibit the efficiencies of using a uniform source for engineering data across different disciplines. “When such a setup becomes part of the digital twin technology for production, plant engineers get instant access to live process data, up-to-date plant drawings and documentation. The maintenance aspect of engineering has also benefited from these advancements, as it has led to the creation of digital work orders, significantly reducing paperwork and minimizing confusion. This shift can potentially reduce engineering time by approximately 30%,” adds Vangenechten.

 

Interoperability

Also at the forefront of next-generation process-control technologies is a greater focus on interoperability and open process automation (OPA). “Many chemical companies have found that closed systems are expensive to upgrade and maintain. With proprietary systems, it also becomes more challenging to integrate new technologies. Digital transformations demand enterprise-wide operability, along with cybersecurity, agility and sustainability,” notes Yokogawa’s Finnan. The OPA Forum (www.opengroup.org) began work in 2016 to develop process automation standards that aim to ensure the adoption of truly interoperable automation systems while providing built-in security, future-proof updates and simplified migration. “These standards will allow users to select ‘best-of-breed’ software, hardware and other technologies from multiple suppliers and realize much more value from their operations. There are currently several OPA proof-of-concept tests in process,” says Finnan.

For more on interoperability and OPA standards from Chem. Eng., read:

Furthermore, adds Siemens’ Vangenechten: “Increased interoperability and standardization allow for enhanced compatibility between different systems and brands, facilitating more seamless operations across various platforms.” Interconnection between digital systems is just the beginning, though. “The seamless integration of the physical and virtual worlds will continue to be a key factor in designing distributed control systems more efficiently and operating them even more sustainably in the future,” she notes. And the benefits of next-generation control systems are not limited to brand-new facilities. “Collaborative development of AI applications facilitates proactive maintenance, and also supports digitizing of documentation for older facilities, laying the groundwork for more efficient, targeted optimization and modernization planning.”

Modularization and standardization are additional areas of growth driving progress in industrial control and automation technologies. “As automation systems and processes grow larger and more complex, there is a natural drive to divide those processes into manageable components with well-defined connection points. Modularization helps achieve that through what are essentially standalone “mini-systems” that can operate independently during the initial stages of a project, and easily integrate into a larger system later in the project lifecycle. Standardization has a focus on simplifying the engineering and procurement process by abstracting sets of control gear into functional pieces,” explains Joe Bastone, director of offering management at Honeywell. Honeywell is moving forward the modularization trend with its Experion PKS Hive product line, which gives the project the necessary flexibility to deploy the solution as it is ready, instead of as a monolithic system. Different Hive products can represent subsystems that are engineered independently and incorporated holistically into a single view of the entire process. “The focus on standardization will continue to drive innovation for the next several years in process control. This will obviously benefit new installations, but it will also have an impact on the modernization of aging assets in the field,” Bastone continues.

Field devices support insight

FIGURE 4. Communication must be fast and reliable for plants to fully take advantage of the data collected by field devices. Gateway devices can convert and store data from wireless field devices, providing capabilities for remote monitoring and diagnostics

Beyond the software and fully digital elements of control systems, the physical devices that provide process data to control systems are also undergoing a rapid evolution to meet new demands for process automation and optimization. “Field sensor technology is quickly progressing to more sophisticated, wireless and faster communication protocols,” says Kris Worfe, industry marketing manager — chemical, Endress+Hauser USA (Greenwood, Ind.; www.us.endress.com). Wireless gateway technologies, such as the FieldGate SWG50 (Figure 5), are designed to provide secure communication from field devices and help users to monitor measurement and device health status.

“On the horizon are internal improvements to physical devices to include advanced diagnostic and health functions, more intuitive information and data sharing. Field devices are continuously trending in the ‘smarter’ direction, becoming more capable of sharing their health status and providing process insights,” explains Worfe.

But as a foundational step for remote access to the data from these advanced field devices, cloud and edge technologies remain key. “Cloud-based infrastructure supports storing, analyzing and sharing the vast amounts of data generated across the process industries. Interconnected devices and edge computing capabilities are what allow for real-time data processing, enabling faster decision-making and reducing latency in autonomous operations,” says Worfe. However, he points out that the adoption of interconnected sensor networks can be a slower process for some organizations, as it may require a change in company culture. “A challenge for end users in the industry could be aligning with internal IT personnel and infrastructure with new features and capabilities.”

 

Sustainable practices

One area where many company cultures are indeed aligning with advances in process control is in the capabilities to improve upon environmental sustainability. “Advanced control systems are optimizing energy consumption by regulating machinery and processes according to real-time demand, reducing unnecessary energy usage,” explains Worfe. This optimization of resource consumption stretches beyond energy to other essential parts of a process, including raw materials and water. “Precise control over production processes minimizes material wastage, ensuring materials are used efficiently, consequently reducing the overall environmental impact associated with waste disposal,” adds Worfe. He highlights water management as a particular application that can benefit from the remote-monitoring capabilities and waste-reduction realized by modern control systems. “Integrating eco-friendly practices into process control to reduce environmental impact is mainly driven at the corporate level with ‘green’ or ‘drive-to-zero’ initiatives,” he notes.