Q&A: Digital Ecosystems Optimize Plant Efficiency & Safety

Jim Nyenhuis, the Performance Consultant at Emerson’s Power and Water Solutions Business, spoke with Turbomachinery International about the various benefits and uses of software, including artificial intelligence (AI) and machine learning (ML) in a power plant.

Q: How does a digital ecosystem improve operational efficiency, plant safety, etc.?

Nyenhuis: Digital technologies allow us to address all those issues. We work with customers to understand what operational efficiency means to them based on their objectives. Plant safety is without question built into everything we do. We want to understand and leverage what the underlying technology can tell us and let that information feed up through the digital ecosystem—how do we understand what’s working well at the plant level, how do we identify if there are control problems or if individual pieces of equipment are having issues? A lot of that information is lying within the automation infrastructure, and I don’t think we’ve leveraged that to the degree we want to. We find many customers do not understand the performance characteristics of their closed-loop controls or areas to drive additional improvement, efficiencies, etc.

Performance and control management is another big area that we look at. Our information, such as the automation system, will oftentimes track what operators are doing within the system. In the past, that information was stored in text-based logs, but there’s a lot of valuable information stored in those log files that if we expose it, we can understand how operators are interfacing with the system during different periods of operation, be it startup, shut down, various operating modes, etc., and based on that we can determine where operators are spending a lot of their time.

Q: How do intelligent detection and monitoring technologies increase a turbomachine's life?

Nyenhuis: We want to ensure that we’re monitoring the right things and keeping the operations within the OEM limits of that asset—gas turbine, steam turbine critical pumps, etc. We want to coordinate with our customers on how to provide the right guidelines and additional levels of automation to ensure that those assets are not operated outside their design specifications and parameters. In many cases, this means determining what additional levels of monitoring we can deploy to track things that may not be necessary for control but could tie into the larger picture of the evolving life expenditure and the amount of critical stress being put on that asset. These monitoring levels can be deployed and then folded back into technologies to understand the operational impact of the asset’s long-term life.

Q: What about cybersecurity—what kinds of risks are power plants exposed to?

Nyenhuis: That’s a critical issue, and it comes up in everything we’re doing—and certainly within this digital ecosystem. There is an expectation from our customers that cybersecurity is thought of at all levels of deployment, even individual things driving valve positioning. Those risks are real in the plants and critical infrastructure assets that we deal with. As solution providers, we’re looking to build cybersecurity capabilities all along that chain—from the lowest-level field device implementation to software that manages those devices with real-time control systems and the cloud. This capability becomes a bit more efficient and streamlined when we have a seamlessly integrated digital ecosystem where we can share some of the expectations of what is running and how we manage security throughout.

Q: Cost and integration are two barriers that can keep an organization from upgrading or implementing new digital tools. How does Emerson address these and other concerns?

Nyenhuis: We primarily want to ensure that the current solutions are functioning as they should. As we engage with our customers, it’s very common that we are not fully deploying and realizing the benefits of the solutions already there.

As new solutions in monitoring, AI, and ML come out, there are two components we are seeing: On the one hand, we need to have templated solutions that we can drop in—a standard and automated approach to develop models and deploy solutions against critical templated assets. This gives us an efficiency of deployment from a cost and value perspective to the end users. On the other hand, customers need a level of customization. Whatever solution platforms we deploy, there needs to be a component that allows a customer to, when they find situations that are specific to them, have the means to step outside those templated deployed solutions that cover 80 - 90% of the situations and be able to customize things to meet their needs.

Q: From your point of view, what role will software play—whether it’s increased safety, automation, optimized operations, etc.—in a sustainable energy market in the coming years?

Nyenhuis: Software is going to touch everything we do, and we want to better manage those safety-related situations to the degree that we can. We want to detect if critical assets are having an issue that might put plant personnel in a dangerous situation. The earlier we can detect those things, the better we can manage those types of situations. To help accomplish this, the applications that are driven by that software are going to become more sophisticated. As each piece of the ecosystem becomes smarter, we filter out more noise and deliver more meaning. Issues identified by software at one level will help teams better interpret what is going on at the next level.

For example, our colleagues at Fisher Valve said sophisticated software automatically detects when things start to evolve at the valve positioning level, and that information needs to be filtered up so it can be accessed across that entire digital ecosystem. The value comes when the valve supplier tells me, “Hey, there's something critical going on with this valve”, and teams interpret that at the real-time control level so we can make decisions and alert the right people. We can then correlate and say we, too, are seeing degradation in control performance in that area. Leveraging software capabilities in the digital ecosystem helps us understand what is evolving that is off-design and that I need to be aware of.

The other side of the coin is how to optimize and drive additional performance and make better decisions. For example, optimizing output from wind turbine controls through automated yaw calibrations and ensuring maximum output from a fleet of wind turbines at a farm. Another example is using software to ensure that we're automatically able to start up a complicated power plant with minimum time and minimum fuel usage. These things will eventually extend out to higher-level cloud management power systems. It becomes a question of how we blend and optimize what's available from an energy storage capacity in real-time—from batteries and other technologies—to better forecasting renewable generation and the various assets that generate electricity and blend those most efficiently with what the overall system and end users of electricity need.

Q: How does Emerson see the future of AI-based software and envision the role of cloud-based analytics in shaping the future?

Nyenhuis: As we continue to evolve toward the future, AI and ML will become an inherent part of what we are already deploying. They are the enabling technologies that allow us to take the vast amount of information available and encapsulate and interpret the relationships that those things tell us. As we actively work in this area, we see the need for overarching contextualization of the various data. In a traditional plant scenario, we might organize things at an asset level in our ERP and computer maintenance management systems or folks only maintain individual parts and assets. But we want to work toward a consistent frame of reference that allows the data from these disparate pieces, pieces of that digital ecosystem, to efficiently flow together so these enabling technologies allow us to pull the right information from different systems and aggregate them together and put them in the right context to make decisions.

To find out more, listen to the full podcast with Jim Nyenhuis of Emerson.