The Future of Manufacturing Plants: Self-learning, Self-adapting, and Self-sustaining

Summary

Though AspenTech has traditionally operated as a conservative company in the automation space for the past 39 years, the rapid changes that have swept the globe this year have inspired the company to establish a different plan moving forward. “We are moving toward self-optimization of manufacturing facilities rather than just automation,” said Ron Beck, energy industry director with AspenTech. “Automation typically involves working toward a fixed objective, while self-optimization involves looking at how the plant is performing and adjusting to do better in a changing environment.”

Ron Beck has more than 30 years’ experience in providing software solutions to the process industries and has worked with AspenTech for 13 years. He has acted in a strategic consulting role for several of the world’s largest process manufacturers, including Dow Chemical, Duke Energy, Israel Electric, Shell and Celanese.

The self-optimizing plant (SOP)

AspenTech is working toward what is called an SOP, or Self-Optimizing Plant. “People used to think of this as science fiction, the idea that a plant can run itself,” Beck said.

However, some of us use forms of self-optimizing technology every day. For example, Google Maps, an app many people have on their cell phones, provides directions to various geographical locations. “Every once in a while, the app will give you a message saying it has generated a faster route than the one you are currently following,” Beck said. “The app has been learning about traffic patterns based on data it gathers from other people who are currently using Google Maps.”

Obviously, a manufacturing plant involves more complicated data streams. That’s why sophisticated technology is required to self-optimize a plant. Now matter how sophisticated the algorithms, however, they won’t replace engineers. “People will always be involved,” Beck said. “If we combine data science with engineers’ knowhow, we get a hybrid model.”
 

Three concepts driving the self-optimizing plant

To create the ideal SOP, AspenTech says there are three main concepts or characteristics that it should embody: self-learning, self-adapting, and self-sustaining.

• Self-learning: Utilize data from across the environments to get smarter and increase accuracy and scope of predictions.

• Self-adapting: React in real-time to changing conditions by making adjustments to meet targets.

• Self-sustaining: Detect anomalies and trigger actions to improve longevity and prevent performance degradation.

“Self-learning involves taking into account a large amount of data from a long period of time and evaluating it in an advanced manner to gain insights regarding how production is going,” Beck explained.

Ideally, engineers would set up the plant, and it would continue to run smoothly indefinitely. However, the plant gets dirty, machinery breaks down, and other things about the plant change. “When things start going wrong, we need more sophistication in the automated systems to alert the operators,” Beck said. Ultimately, such systems can and will predict when parts of the plant would break, or point out when certain materials would run out or fail to be available. This sort of data would help engineers and operators minimize interruptions in operations, increase output, and save money.



“Most companies are on stage two of the plant digitalization journey,” Beck said. “They have optimized individual disciplines to bring about isolated improvement. While improving operations in a particular department or for a particular function is desirable, the idea outcome of self-optimization is to integrate self-learning, self-adapting, and self-sustaining across disciplines at the plant.”
 

Aspen Unified: The first deliverable on the journey to SOP

AspenTech has already begun the process of helping plants integrating self-learning, self-adapting, and self-sustaining attributes across the plant. “The Aspen Unified solution brings planning and scheduling together in one environment to connect and automate isolated processes, maximize the yield of high value products, push closer to limits and drive improved margins,” according to an AspenTech release.

The planning and scheduling software “connects silos and increases visibility to improve decision making,” David Arbeitel, senior vice president, product management for Aspen Technology said. “Bringing planning, scheduling and operations together allows accelerated and more consistent model building.”

A video from AspenTech provides more information.
 

Adapting to the new normal: Remote workers

Beck also discussed the future of self-optimization in light of COVID-19, increased environmental awareness, and the up-and-coming workforce.

Part of what spurred AspenTech’s journey toward SOP involved the rise of remote work brought on by the COVID-19 pandemic. “People are redoubling their efforts to digitalize because they know all of this isn’t going away,” Beck said. “It’s been a wakeup call.”

In addition, companies have increasingly prioritized environmental awareness. “Self-optimizing technologies are exciting in the field of manufacturing not only to save money and create a higher quality product but also to reduce carbon loads,” Beck said.

Even though plants will ideally be moving toward increased self-optimization, there will still be jobs; the jobs will just change. “Those in the manufacturing industry will need to be on a career-long journey to re-educate themselves,” Beck said.