NIR takes on the particle size challenge

By Richard Mills,
It’s a great thought, but can it work? A new study from North Carolina State University shows that a near infrared test for particle-size in feed is within our grasp.

A particle size test that is about 30 times faster than current methods, yes please! Most feed laboratory managers do not need to think long and hard about whether they would like a more convenient test for particle size in animal feed. In pursuit of this goal, Dr. Adam Fahrenholz at North Carolina State University has teamed up with analytical solution providers FOSS to look at how near infrared (NIR) instruments could take on the job. 

A quick review of the current analysis method reveals why a faster analysis would be so welcome. Involving equipment such as sieves and sieve-shakers, the particle size test typically takes 15 minutes, plus five to ten minutes of preparation time and analysis time, not including the recording of time and date information. A single test can take a well-trained operator anywhere from 20 to 30 min depending on what they are testing for. At the same time, particle size is so important to the feed mill that multiple analysis may be required per day when setting up equipment, and often once per shift, or several times a week under normal operations.


Importance of particle size
One use is to gauge the performance of equipment. “If, for example, I have targeted a specific micron size, let’s say 700 microns, when I put new hammers and new screens on my hammer mill, I’m targeting these 700 microns based on screen size and hammer speed that I have chosen. So I know that my system is doing what it is supposed to do,” explains Fahrenholz. “By measuring the particle size continuously, I can evaluate when it’s time to change those wear components, or when to make a change because I start to receive a slightly different physical characteristic of grain that makes the particle size start to deviate from the target.” 

Particle size is also important for the physical flow of the feed, especially for meal feeds. For instance, if the particle size is overly fine it can lead to problems in flow through feeders and out of bins.  And perhaps most importantly, on the nutrition side, particle size affects feed conversion for both pigs and chickens either in terms of intake or digestion or both. For poultry, some coarse particles in the diet are necessary for gizzard health “At the end of the day it all comes down to what gives us the most efficient pounds of feed per pound of gain of that species,” concludes Fahrenholz. Yet another factor for feed millers to consider is the variability of raw material, for instance when a mill switches over from so-called old crop corn harvested the previous year to the new crop corn that will grind differently.


Tracking particle size 
The bigger the operation, the bigger the potential impact of variations in particle size becomes. Some manufacturing facilities are turning out huge amounts of feed, perhaps around 20 to 25,000 tons of feed per week. Running 24/7, the mill is getting more mileage than a yellow cab and wear parts such as screens and hammers need replacement every few weeks. 

With the current method, if the particle size starts to drift out of spec, a lot of feed can go through before a control test shows the trend. The feed is still usable and no animals are going to starve, but can it live up to the all-important feed conversion promised to customers? In this situation, the feed mill manager has to decide whether to live with the trend or change screens and hammers earlier than expected. “If you had the ability to generate ten particle size samples a day, anyone could easily run them and track it, and when the control parameter drops below a certain point, you would know it’s time to make a change,” says Fahrenholz. “This would bring value not only to the animal, but also to the feed mill bottom line.” 

Near infrared test for particle-size in feed

The promise of NIR 
Given that many feed mills today already have NIR analysis equipment in place it would make perfect sense to use it as a rapid method – no sieving, no manual recording, just put the sample in the sample cup and push the start button. 

To investigate the possibility, Fahrenholz embarked on a calibration project aimed at getting enough data on particle size variation to program an NIR analyser. Firstly, he teamed up with NIR product specialist Ole Rasmussen from FOSS and then Fahrenholz and his graduate student J.T. Pope began to do reference tests and collected data on particle size, including average particle size (DGW) and standard deviation by weight (SGW). The SGW represents the distribution, and the larger the SGW the more varied the sample.


“We did 100 samples, which was enough to show that we had a fairly decent ability to replicate the sieve analysis,” says Rasmussen. “It was not as tight as we would like it to be, but we have shown that it can be done.” Fahrenholz concurs. “The prediction is relatively tight,” he says. “Particle size analysis is somewhat variable so when you get very small, you can get very repetitive with the sieve analysis. A good operator can do one sample ten times and only get a 20-micron variation. As you get larger and there’s more variation, you can pick up some small differences. For me the NIR is consistently within 50 microns of what the sieve analysis is telling me. The value of this is that I get in the range or ballpark of where I am supposed to be. The data shows that it is feasible.”


Particle size in feed
Average particle size is the particle size at which 50g of a 100g sample are larger and 50g are smaller. Geometric standard deviation by weight (SGW) is the distribution of particles in the sample - the larger SGW the more varied the sample.


Next steps

In conclusion, it is clear that the NIR solution works, but more reference data is required to handle the variability involved in testing particle size. It will take time, but the good news is that this job can easily be worked into existing everyday procedures. “If a lab has a robust particle size analysis program, they are already producing a lot of samples. So now it’s just a question about putting these on the NIR while they are doing it and creating a catalogue. And the more samples you develop that can be sent back to FOSS, the tighter it will become,” explains Rasmussen. “It doesn’t really increase the workload to go from the sieve analysis to run the same sample through the NIR. And the reward would be worth it.”


Rewards in sight
Research into nutrition is one particular area that can benefit. “There is a big interest in particle size right now, says Fahrenholz. “We are constantly improving the genetic lines, and so some research done 20 years ago may need to be done again related to particle size.” In this light, particle size is becoming an increasingly important metric both for the grind characteristics of raw material and for the feed delivered by the mill. Fahrenholz concludes: “If I can do this easier with something I already have, and not tying up quality personnel, I can do it more often because it’s less time-consuming. That would be a real benefit for me because I can always say that we are within spec.”


Near infrared test for particle-size in feed


The standard method
Equipment: Stack of Sieves including pan and cover, Balance (with accuracy to 0.01 g), Mechanical sieve shaker.
Many steps involving sieving, mass sample calculation, cleaning of sieves, weighing of separate sieves, calculation of result.
Typical analysis time: 30 minutes.


The potential of NIR
Equipment: An NIR instrument already in use for other purposes.
Method: A sample is poured into a sample cup and the operator pushes the start button. The result is delivered simultaneously with other results for protein, fat, moisture and so on.
Analysis time: less than a minute.

See more industry-related articles and videos

Load more