Informed Decisions

FROM THE NOVEMBER ISSUE: A practical guide for plant managers on recognizing strong lab data, and what to do when the results are unexpected.
By Matt Thompson | October 17, 2019

TaNeal Boer, technical science director at Bion Cos., describes an ethanol plant’s lab as “the nuts and bolts of the plant.” And because of the critical role the lab plays, Boer says it’s important for plant managers to understand and recognize the strong data it produces. “You can generate tons and tons of data, but if it doesn’t fit some of the requirements to make sure that the data’s strong and accurate, you could be making decisions on bad data, even though it looks good.”

The key to strong lab data is consistency, she says. Testing samples as quickly after collection as possible, having procedures in place to make sure samples are collected consistently, and cleaning sample bottles and equipment can all help contribute to strong data, she says. Those types of procedures should be clearly documented.

“Facilities will have to have procedures in place so that they make sure their teams are sampling consistently at all times, which will result in a good quality sample,” Boer says. “Purging the sample port, making sure that sample bottles are clean, that there’s no water in them, those kinds of things. Things that are small, but they can possibly contribute to error. And if you are making a decision on a bad sample, you could be costing that production facility hundreds, if not thousands of dollars every day.”

Andrew Hawkins, director of laboratory services at Phibro Animal Health Corp., agrees. “We have SOPs [standard operating procedures] here for the lab, and the best ones are those that have been through three or four revisions where somebody actually prints it out, takes it with them, and jots down some handwritten notes while they’re doing it, saying, ‘Hey, I don’t understand this stuff,’ or, ‘This step is confusing,’ or, ‘This actually is like four steps in one. Why is it all in one?’ Those types of feedback are the only way we can make our protocols better,” Hawkins says.

Communication, Questions, Quality
If data is inconsistent, or includes an error, the next step is to determine why. This is where communication between the lab and the operations side of the plant is important. “It’s just keeping that communication open and asking a lot of questions,” Boer says.

Those questions, she says, should focus on determining what factors might have led to the unexpected results. Asking the lab about instrument calibration, if the sample collected was a good one, and whether SOPs were followed are good starting points, she says.

Yan Zhang, director of research at the National Corn-to-Ethanol Research Center  (NCERC) agrees. She says the first step in analyzing unexpected data is verifying its quality. “Check the data quality first before jumping to making any conclusions,” she says. “So that’s the first step. You need to make sure there’s no foul play in the lab side.”

It’s also important for a plant manager to know and understand the plant’s idiosyncrasies. Plants that slow production will see changes in their lab data, Boer says. “Understanding how the plant data changes at the reduced rate or the slower rate becomes more critical so that you have trust in that data, because some things might look very strange since you’re operating differently.”

That can be accomplished by spending time with data, a strategy Hawkins recommends. “I think looking at data regularly, even if there’s not an emergency, is probably one of the best tips I could provide,” he says. “As a scientist, I’ve done that a lot and that really gives you a good foundation to know when things are off, and usually that’s critical.”

Zhang agrees. “For the plant manager, the combination of staying very connected with the lab data and also, if it’s possible, building certain scientific background to understand the process and the biochemical processing of corn to ethanol will definitely help you.”

Room for Improvement
Hawkins says monitoring data, even when there isn’t an emergency, can be helpful to spot opportunities for efficiency improvements. “I think a great place to look at to improve efficiencies is really to monitor the liquefaction system for fusels and ethanol,” Hawkins says. “Because these plants are so integrated—and there are quite a few plants that are moving to zero-discharge—a lot of what comes back around from the back end of your plant can actually inform you about what’s happening in general.”

Hawkins says regularly reviewing data gives plant managers an opportunity to recognize trends. “Really investigate those trends because those trends can be indicators of other things that are going on in your plant,” he says. “Maybe a piece of equipment is going to fail or maybe there’s a pump that’s not delivering what you think it’s delivering at the right time. Those trends, especially if they’re repetitive, can usually give you clues.”

Boer says another opportunity plants have when data indicates everything is operating normally is a lab proficiency program. She says plants that work closely together or are sister plants can send samples to each other, to compare calculations for acidity or moisture. “There are also proficiency programs out there where it’s a known lab sample and then a laboratory would then run a solids, an HPLC, maybe a moisture or acidity, depending on what that sample is, and it compares them to a known, and it tells their percent recovery,” she says. “It’s just a really nice check to see how a lab is doing independent of their normal day-to-day operations.”

Boer also cautions not to take lab equipment for granted. Lab equipment should undergo preventative maintenance programs (PMs), just like any other plant equipment, she says. “Whatever PMs go on in a facility, you want to do the same thing for your lab, because that lab equipment never shuts off; it’s running all the time,” Boer says. “If you have that PM program, which is through the vendor of that instrument, and you have equipment go down, you get moved up on the repair list.”

And Zhang says the importance of accurate, strong lab data is a key for running a successful plant. “There’s no other way for you to know whether you lost a lot of money or actually are in good shape,” she says.


------------------------------------ SIDE BAR-----------------------------------------------------------------

Tips for ensuring strong lab data
“Making sure that the data is completely trustworthy is always very important so you make the right decision and move in the right direction, as opposed to having some bad data that sends you off on the wrong path.” - Andrew Hawkins, Phibro

• Develop strong SOPs. Solicit feedback and ask how SOPs are working. “The best ones are the ones that have been through three or four revisions where somebody actually prints it out, takes it with them, and jots down some handwritten notes while they’re doing it.” - Andrew Hawkins, Phibro

• Ensure employees are trained in proper lab techniques. “Train everyone that does the work. That could be an operator on the evenings or weekends when the lab staff may not normally be there. Have everyone trained and validate the individual that’s doing it.” - Andrew Hawkins, Phibro

• Establish check standards. “Analytical method needs to be validated and, for daily operations generating the data, you need an independent check standard to verify the whole batch was good enough.” - Yan Zhang, NCERC

• Ensure instruments are cleaned, calibrated and verified. “Make sure that all the equipment has been validated and calibrated so the data is accurate.” - Andrew Hawkins, Phibro

• Put a preventative maintenance program for lab equipment in place. “If you have that PM program, which is through the vendor of that instrument, you get moved up on the repair list.” - TaNeal Boer, Bion

• Take part in a lab proficiency program to double-check data. “It’s just a really nice check to see how a lab is doing independent of their normal day-to-day operations.” - TaNeal Boer, Bion

Tips for unexpected data
“Troubleshooting always takes a whole team.” - Yan Zhang, NCERC

• Verify SOPs were followed. “Check the data quality first before jumping to making any conclusions.” - Yan Zhang, NCERC

• Verify the check standard was acceptable. “For daily operations generating the data, you need an independent check standard to verify the whole batch was good enough.” - Yan Zhang, NCERC

• Determine if there were changes in the plant that could affect the data. “Understanding how the plant changes at the reduced rate or the lower rate becomes more critical so that you have trust in that data, because some things might look very strange since you’re operating differently.” - TaNeal Boer, Bion.

• Verify instruments are calibrated and validated. “When you’re validating, you want to make sure that that validation check is part of your standard operating procedure. You want something independent so that it’s just not reinjecting the same sample that may be contributing to the calibration issue.” - TaNeal Boer, Bion

• Engage the whole team to determine why a sample may be off. “A lot of times, when things are not going right, there is more than one line of evidence.” - Yan Zhang, NCERC

Tips for data within norms
“I think the best and most positive thing a plant manager can do to help encourage good data, and to feel confident in the data that the lab is generating and reporting out back to operations, is to occasionally ask to take a look at … the logbooks to show that that equipment is working
correctly.” - Andrew Hawkins, Phibro

• Don’t change. Keep doing what works. “From a practical point of view, if things work, don’t change it. Especially for experienced people.” - Yan Zhang, NCERC

• Look for opportunities to improve, particularly in data from the liquefaction system. “Because these plants are so integrated—and there are quite a few plants that are moving to zero-discharge—a lot of what comes back around from the back end of your plant can actually inform you about what’s happening in general.”-  Andrew Hawkins, Phibro

• Regularly review the data. “I think looking at data regularly, even if there’s not an emergency, is probably one of the best tips I could provide.” - Andrew Hawkins, Phibro

• Look for trends in the data. “Really investigate those trends because those trends can be indicators of other things that are going on in your plant. Maybe a piece of equipment is going to fail or maybe there’s a pump that’s not delivering what you think it’s delivering at the right time.”
- Andrew Hawkins, Phibro

• Understand the technical, scientific side of data and ethanol processing. “For the plant manager, the combination of staying very connected with the lab data and also, if it’s possible, building certain scientific background to understand the process and the biochemical processing of corn to ethanol will definitely help you.” - Yan Zhang, NCERC

------------------------------------ SIDE BAR-----------------------------------------------------------------

Author: Matt Thompson
Associate Editor, Ethanol Producer Magazine
701.738.4922
mthompson@bbiinternational.com