The Quest for Maximum Yield: What’s Math Got To Do With It?

By Holly Jessen | August 15, 2011

The Novozymes North America Inc. presentation at FEW was about a new data analysis system the company is now using to help an ethanol plant find its optimal enzyme dosing level. The presenter, Nicholas Giffen, a scientist and data analyst, developed what Novozymes is calling its Neural Network Model, which uses complex modeling to identify linear colorations between several variables at once. Several years ago, Novozymes analyzed mashes from more than 40 ethanol plants, in an effort to discover how to increase yield, Jack Rogers, regional marking manager, tells EPM. The results showed that, on average, about 7 percent of available starch was not being converted to ethanol. “That told us that there was a lot of room to go for improving yields for our customers,” he says.

Novozymes has taken two paths in an effort to help improve yields for ethanol plants. First, the company has been developing and bringing new enzymes to market that break down starch more effectively. The company has a product line of enzymes with three levels, the last of which is its premium product. It costs more but also has the ability to access more starch that other enzymes may not be able to convert. “Not all mashes respond the same way to all products, so you have to get a good match for the plant and the type of conditions,” Rogers says.

Novozyme’s second strategy is about finding the right dosing level for each individual plant—not too much and not too little. “It’s a matter of optimizing both types of glucoamylase or saccharification enzyme and the dose,” he says. That’s where Giffen’s mathematical model comes in. Instead of wasting time and money finding the right enzyme dose by trial and error or educated guesses, Novozymes can help customers come up with solid numbers based on individual plant data. And, if a plant’s process parameters change, the model can be used to quickly recalculate. In the past, Novozymes has used simpler models to find optimal dosing but “they really didn’t capture the complexities of the ethanol process, like this one,” Rogers says. Giffen spent about six months developing a model that would help capture what’s happening in a dynamic environment and, as of FEW, the company was starting to offer it as a service to customers. “It gives you a lot better insight into what’s actually happening in the process,” he says.

Increasing ethanol yield at an ethanol plant can have a big impact on the bottom line. That’s why it’s so important for plants to find their optimal dosing levels as enzymes are a great lever for increasing ethanol yield, Rogers says. “In the end, when you run the numbers, just about any slight gain in ethanol yield is going to pay off for a plant,” he says. “The value of the ethanol produced far outweighs the cost of the enzyme.”

—Holly Jessen