Another Tool in the Toolbox

DuPont’s XCELIS® AI enhances ethanol plant operational assessment with data-driven analyses and predictive modeling driven by science and engineering fundamentals. The new tools will help de-risk decision making and enhance plant optimization.
By Matt Thompson | October 28, 2020

Dr. Tony Schindler, regional application development leader at DuPont, strongly believes that enabling ethanol plants to make more real-time decisions based on data and quantitative predictions is valuable. DuPont recently announced a new set of services from its XCELIS® Ethanol Solutions platform that will help producers do just that.

The new offering, XCELIS® AI as Schindler describes, has two goals. “One of them is to help plants make the most sense of their data so they can see what will improve efficiencies and boost yield and help them achieve their goals right now. The second one is to use XCELIS® AI Virtual Plant Technology to model how future changes to the plant will impact operations.”

Making the right decision is increasingly important, Schindler says, as benefits seen from plant upgrades and process improvements are incremental. “Over the years, I’ve seen yields go up, and as you get closer and closer to theoretical, I think it’s harder and harder to get that next bump in performance. Having access to tools like XCELIS® AI lets you really fine tune the plant.”

Schindler says XCELIS® AI Virtual Plant Technology uses predictive models based on science and engineering fundamentals to help preview potential changes. “We’re helping them de-risk decision making in how they’re running their process right now and, in the future, when they look to optimize and maybe make some changes.”

Schindler gives an example of a plant that’s considering an upgrade to increase grind capacity. “If they’re considering that upgrade, what we’re able to do with the model is say, ‘If you grind a certain percent more corn, here are different ways you could handle that corn in your process. Here are all the energy efficiencies, here’s how it’s going to impact your economics, your Carbon Intensity (CI) score, etc.’ So now the plant has a better understanding of what the process will look like and they have actual numbers to make that decision,” he highlights.

The models can also be used to create plant simulators. “We can make a DCS style interface, but what’s behind it isn’t a plant, it’s a model,” Schindler shares. “Operators can sit in front of the simulator and run the process but it’s in a nice, safe virtual environment. So, if they overfill a tank or forget to feed enzyme into a fermenter, they just hit reset on the simulator and they’re not causing any damage or yield loss.”

Both the data processing and Virtual Plant Technology sides of XCELIS® AI are customizable to individual plants, Schindler says. “On the data side, we’re working with some plants where they’re sending us the typical smaller data sets that they’re used to sending vendors and we can apply our analytics tools to those,” he explains. “Then we have other producers who are more interested in having us take a more detailed look at their process, so we work with their data infrastructure provider to send us data and we implement a tailored solution, pulling process and lab data from all across the plant, doing complex analyses and then sending back custom reports geared at showing insights.”

Schindler says the response from customers has been very positive. “Our customers see value in having both data-driven analytics and predictive models driven by science and engineering fundamentals.  Using both, as XCELIS® AI does, is a way to help our customers get the most out of their plant now and in the future.” concludes Schindler.

For more information about XCELIS® AI, please visit  www.xcelis.com/xcelis-ai.