Partners In Prospecting

FROM THE JULY ISSUE: A predictive technology from the paper mill industry is making its way into ethanol, tying into main databases to plan and troubleshoot.
By Matt Thompson | June 17, 2019

Patti Sprouse, co-founder and president of Bonanza Associates, describes her company’s predictive modeling software, Prospector, as both a flight simulator and GPS for ethanol plants. It’s similar to a flight simulator, she says, because “it allows the clients to really practice running the plant in different ways, without anybody getting hurt or bad decisions [being] made that are going to have costly implications.”

And Prospector is comparable to a GPS unit because it allows plants to see how changes within the plant will affect future operations, she says. “You get off track and you want to reset. So maybe you had to take a little detour and you still want to go to this destination. You just reset your GPS and say, ‘Take me from here,’ and the same thing with the model.”

The software has been used in the paper mill industry, but was recently adopted by Show Me Ethanol in Carrollton, Missouri. Sprouse says the decision to explore using Prospector in ethanol plants came from Ronnie Hise, Bonanza Associates vice president, whose son has experience with biofuels.

“He was working on his master’s degree and working with a group of people looking at algae as a route to biodiesel,” Hise says. “He and I got to talking about what the plants looked like that did that sort of thing, and then I paid him to do a little research for us, in fact, to look at ethanol plants while he was a grad student.”

Sprouse says the similarities between the paper and ethanol industries made it easy to adapt Prospector to ethanol plants. Producers in both the paper and ethanol industries have inventories of partially finished products, and Prospector is able to accurately predict progress. One difference, she notes, is the fermentation process. “We had to do quite a bit of changing, especially around the fermentation, because we had never done something like fermentation processes,” she says. “We have modeled batch digesters in the paper industry, but the batch time is two to three hours, not two to three days. The scheduling of the filling and emptying of each fermenter, along with modeling multiple chemical additions were a challenge.”

“Another characteristic that we thought fit very well with the fuel ethanol manufacturing plants and the paper industry is, there didn’t seem to be any one tool that would allow people to look at both the operations and the financial aspect,” Hise says.

The software is customized for each mill or plant that uses it, and Sprouse says the first step involves setting up a detailed mass and energy balance of the plant for the software to work with. “If you have a historian or if you have an external schedule, we’ll bring that into the model to start with,” she says. “The first hour of the model shows where everything is running right now. And then as time goes on, the next hours are predicted based on the past using mathematical and statistical formulas.” Sprouse adds that it takes six to eight weeks to create an individualized model for a plant. “It sounds fairly complicated, but our model is robust and quick, and so it can do a lot of these analyses for us,” she says.

And once that model is built, the software allows the plant’s entire staff to be on the same page. “You have a plan and anybody in the whole plant can open their computer and see this plan that’s predicted, and you can see the past and the current time and the future all in one shot,” Sprouse says. “It really promotes a lot of collaboration and coordination across the workforce in the facility.”

In Action
Brian Pasbrig, plant manager at SME, says the plant has been using Prospector since late last year. “We’re very focused on using data that’s available to us and using technology to make our operation as efficient and successful as it can be,” he says. “When people approach us with technology-driven ideas and proposals, we generally will take those meetings. So that’s kind of how it started. They came to us, and technology’s exciting for us, so we decided to go ahead and collaborate with them.”

And so far, Pasbrig says, the experience has been positive. “We’ve used the tool very successfully with our spring outage,” he says. “We were able to better plan the entire coming in and coming out of the shutdown and managing the process, the tanks and just managing the entire operations side of the shutdown.”
It was particularly handy in emptying tanks to ready them for inspection and cleaning, he says. “Using a common tool for everybody to refer to made it a lot smoother; made the production planning going in and coming out of shutdown a lot more transparent. It really helped a lot.” He adds the plant has also used the tool for production planning, as well as troubleshooting.

“Since it’s a mass balance, if some part of the process—say the centrifuge aspect of an ethanol plant—if that’s not running quite right and it’s slowly degrading in performance, you can pick that up much more quickly because Prospector has charts that are monitoring things on an hourly basis,” Hise says. The troubleshooting part of the software, he adds, is one of the more advanced functions. “Our experience with the paper industry is that over time, people become more comfortable using it and start to use some of the more advanced features, beyond planning for outages.”

And Prospector is currently being updated to help plants evaluate chemical trials, Pasbrig says. “When a supplier comes to us and has a trial proposal, we can use the predictive tool to simulate the trial before we even think about implementing it.” It’s a suggestion from SME that wasn’t considered by the paper industry, Sprouse says. “With ethanol plants, it seems that trials are a pretty big part of their operations, where you’ve got very expensive chemicals—the alpha-amylase, glucoamylase, yeast and other important enzymes” she says. “So, we’re really focusing right now on building a great add-in for him [Pasbrig] where he can try adding a trial chemical at different points in the system, allowing him to predict the operational and financial outcomes.”

Hise says the trial evaluation aspect of the software is still in the prototype stage. “But we’ll be testing it very soon with SME and we feel very comfortable it’s going to work.”

In addition to forecasting plant operations, the software can also forecast financial outcomes of operational changes, Sprouse says. “We also have financial reports, so users can quantitatively compare between different plans, and that’s really valuable if you’re coming up on an outage and you want to try different scenarios,” she says.

A New Approach
Pasbrig says Prospector differs from other systems ethanol plants are currently using in that it can predict plant outcomes from small operational changes. “I don’t know that there’s a lot of facilities that have an active mass balance predictive model that says, ‘If I do this very small process change, here’s how it affects the entire plant,’” he says. “Usually, the operator in the control room can take a stab at that, but this tool actually tells you if I change just one small thing, the—centrifuge feed—it tells me how it’s going to change the water balance on the front end in four days.”

Sprouse agrees. She says in both the ethanol and paper industries, prospective customers tell her they already have a spreadsheet that can do what Prospector promises. But Prospector is different, she says. “Prospector models the entire plant, so it’s not just optimizing one department or one piece of equipment. It’s optimizing the entire facility so everybody’s on the same page and there’s one vision for how the plant plans to run.”

Pasbrig says, “We can go out and look three or four or five days ahead, or even two weeks ahead, and predict what the plant’s going to be doing a lot quicker, easier, and again, we have this common tool that we can use instead of everybody having their own way of looking at it and their own spreadsheets and their own outlooks.”

Although Prospector uses historical plant data to predict future operations, it isn’t tied directly into a plant’s distributed control system (DCS), Sprouse says. “We don’t control anything. We’re more of an advisory product for the people in the plant to use,” she says.

“Specifically, it’s not tied in to the DCS, but it’s pulling historian data,” Pasbrig says. “All of our operations are combined into one central database, and Prospector’s tying into that database.

“I think a tool like this would be beneficial in most plants,” he says.

Hise says that while Prospector has been used primarily in the pulp and paper industry, it was designed to be a tool used by any manufacturing industry. “It just seems like ethanol plants are a good fit for Bonanza Associates,” he says.

Sprouse notes that Prospector allows plant staff to use their time more efficiently. “A lot of plants tend to look back, and they really don’t have a lot of time for looking back. Prospector is really a forward-looking product, so you’re constantly saying, ‘OK, this is what’s coming up next,’ and make decisions. So we hope that using our product allows them to focus on other important decisions.”

Author: Matt Thompson
Associate Editor, Ethanol Producer Magazine