The industrial sector is both enamored and perplexed by how to best use artificial intelligence (AI) in process control. To get answers to the most common questions for industrial operators, Control talked to Curtis Thompson, DeltaV product marketing manager for AI at Emerson, who works on AI product strategy and research for applications in process control.
Q: AI means different things to different people. How do we define AI, particularly in the industrial sector?
A: There are two parts to this answer. The first is that AI is simply a tool that takes the form of intelligent machines using fancy math, complex algorithms and large amounts of data to simulate human intelligence. What that means for us is we can leverage AI to perform complex human-like tasks. That's big picture.
The second part is in an industrial context.Where general AI may deal with vast and diverse data sets and broad applicability. Industrial AI specifically deals with existing operating technology systems, where real-time or near real-time decisions need to be made to enhance efficiency, reliability and safety concerns. It's not only the operational aspects, but also engineering challenges, where we can use generative tools to accelerate engineering productivity and timelines.
Q: What are the primary drivers accelerating AI adoption in the process control sector?
A: There are several challenges where AI is perfectly suited to help with brain drain from the retiring workforce. Knowledge is leaving with those workers. There are also competitive pressures, such as the need to squeeze out more with less. I think modernization is a big driver, too. For each of these challenges, there are specific AI solutions that can help address them.
On the flip side, there are key tailwinds in the technological advancements that have been made. On the machine learning side, we've seen considerable improvements to the adoption of these algorithms, especially for deep learning and reinforcement learning. We also have the Industrial Internet of Things (IIoT) where enhancements to sensor intelligence, connectivity and affordability have driven easier access to data. Then, access to computing certainly, especially at the edge, has matured, so real-time AI processing can be available wherever needed.
We have smaller, large-language models (LLM) that that can be hosted locally, and they’re advancing at light speed in terms of intelligence and capability. So, from a technological perspective, AI has moved from conceptual dream to a more practical, achievable tool with tangible benefits.
Q: Can you highlight some AI solutions Emerson is delivering in the industrial setting?
A: Where to begin? First, we should talk about enabling technology as a part of our larger boundless automation vision. We've laid out what the future of automation looks like, and the importance of AI and data as cornerstones. Starting at the top, we have our cloud-enabled, enterprise-operations platform, where we not only have access to unparalleled AI computing, but also the leverage of our unifying data fabric with AspenTech Inmation and its enterprise-wide insight, where we can make sure to drive informed decision making.
We have the DeltaV Edge environment coupled with our software-defined control solution, which is the IQ controller. It provides a modern, containerized computing environment and a seamless flow of data. The final piece of this balanced automation vision is the intelligent field. We've enhanced onboard analytics and streamlined data connectivity from anywhere.
This industrial data fabric for DeltaV users, specifically, can progress that data through DeltaV Edge, and aggregate other data sources for a higher-order data context, which is crucial on the AI journey. Also, internally we have cloud-based tools like DeltaV Revamp where we're using AI to drive.
Get your subscription to Control's tri-weekly newsletter.
Q: What challenges may hinder adoption and how can DeltaV help solve those challenges?
A: I often say AI products are data products. In our served industries, the biggest challenge is data. We're often dealing with proprietary data, whether it's sensor data or other control strategies. Accessing that data can be a huge challenge.
We know the term IT/OT convergence, which kind of morphed into digital transformation. Much of the industry is still going through it, so there's a spectrum where existing facilities that have been running for decades face a modernization challenge. On the other side is the born-digital concept, where greenfield projects make it easier to get that data infrastructure in place from the ground up, but then you don't have decades of historical data. That’s where we're leveraging our simulation expertise to help bridge that gap. There's a vast amount of critical operational knowledge locked within DeltaV as well as in the minds of experienced engineers and operators.
There are also larger cybersecurity challenges. I mentioned the brain drain and retiring workforce. So, whether you can access the data, even if you can, it's just not always easy for humans to understand or quickly reference. So, you have huge complexity issues.
Q: How is AI changing this industrial automation industry?
A: I think there are three key promises of industrial AI. The first is increasing agility by embedding deep learning and hybrid models that can learn to optimize and boost production.
The second is eliminating complexity with intelligent use of data where we can use AI to capture and embed knowledge, making it accessible to all workers. I think about this in terms of humanizing the data to make us smarter operators of our facilities and to be more predictable and efficient.
The final promise is automating workflows—generative tools for engineering activities and automating other key workflows. It’s a tool that can do actual work for those common tasks, where we have a multi-AI agent system and agents are highly connected to work collaboratively.
All the AI tools and capabilities map to these three pillars. They are the building blocks on the journey to autonomous operations.
We talked about how we use artificial intelligence to model behavior and understand relationships between key process parameters, even including external forces spanning. This is how we see the DeltaV automation platform specifically driving this evolutionary change in automation.
Q: Where can people more about DeltaV and how it works?
A: We've barely scratched the surface here, but I encourage people to visit here.
