Honeywell UOP pioneering 'optimization-as-a-service'
By Tom Wilk, Plant Services editor in chief
Refineries and petrochemical facilities are among the most complex process optimization challenges in the industrial world. Sure, the building of models and tuning of interrelated control loops is an ambitious
undertaking, but in the end all works great. For a while. Exchangers foul, feed compositions change, catalysts lose potency and the online optimizers developed with so much care and investment often are switched to manual mode, awaiting some future re-optimization project.
But it doesn’t have to be this way.
In a presentation at the 2017 Smart Industry event, Zak Alzein, VP & GM of Honeywell UOP Connected Performance Services outlined a vision of how the Industrial Internet of Things (IIoT) is helping process facility operators and reliability team members evolve their process optimization efforts from an irregular series of fix-it opportunities to a position of proactive continuous improvement in real-time.
The solution? A secure cloud-based software platform that can collect data continuously and in real-time from internet-connected assets, underpinned by analytics engines that leverage deep domain expertise. Ultimately, argued Alzein, this type of platform can help organizations overcome not just technological hurdles, but human capital issues as well.
"If you think about petrochemical plants or refineries, they come to UOP with their market need, their demand," said Alzein. "We first work with them on a feasibility design, a configuration, then they select a technology. They go and start working on the permitting and the funding and so on while we get busy doing the basic design, and start to define the configuration in a lot more detail. Then we help them commission the plant, and expand it over time, apply the next generation of catalysts, help them tweak the configuration on an ongoing basis to continue to adapt to market changes."
He added: "We also do the troubleshooting. These assets are very complicated, as you can imagine, and they run through a lot of challenges, whether equipment failures or process conditions that lead to degradation of yields or a shutdown. We're deploying digital transformation throughout this, our entire value chain."
Based on the company’s deep domain expertise, Alzein outlined four operational challenge areas that Honeywell observed their customers as moving through as they shifted from reactive to proactive operational positions:
- Unplanned downtime. "We do a great job solving problems after the fact," said Alzein. "But the customer's desire is, 'How can I avoid this to begin with? Given all the complexity of operations, how do I mitigate problems very quickly before they become a decline in yield rates or a unit going down?' "
- Underperforming assets. When asked by a customer to help optimize their facilities, Honeywell UOP often is able to identity millions of dollars in operational savings. However, optimization efforts need to evolve from occasional or annual efforts to more sustained, continual improvement processes: "Instead of doing this once a year, or every couple of years, ask, How do I do this every day in a more sustainable way?"
- Energy and emission. In some cases, producers that want to increase capacity cannot obtain the necessary permits, so they have to reduce emissions in order to increase production, a bottleneck situation that Alzein argued could be relieved if an organization were able to leverage machine learning platforms that would learn over time to adjust and optimize processes in real-time.
- Human capital challenges. "We've all been seeing a lot of retirements of expertise in the developed world," said Alzein. "In the developing world, that challenge is even more profound. Today's engineers even halfway around the world don't want a job for a specific area for 15 years so they become an expert. They want to keep moving around, they want a different experience, than the last generation."
Honeywell's response to these challenges is its Connected Plant concept, which combines cloud-based analytics with a unified open architecture platform (termed "Sentience") to create an industrial application implementation that leverages Honeywell's process domain expertise and know-how. And, current human capital dynamics are specifically influencing the design and implementation of these platforms, to accommodate worker flexibility demands while optimizing facility design and operation.
"We’re taking our knowledge in connectivity and cybersecurity, to build a secure digital twin in the cloud," said Alzein. "It’s a digital replica of what that refinery should be operating like," which enables customers to compare current operational performance against theoretical capabilities. Customers also can start to predict potential issues via embedded predictive analytics suites.
The cloud's primary benefit, argued Alzein, is scalability. "The cost of sensing capability is dramatically improving. Meanwhile, cloud capabilities are advancing even as the costs are coming down. The cloud really enables us to solve a lot of problems in a new, innovative way."
