How granulation curves may help choose optimal time to change rolls.
In this column, Nathan Watson, a senior majoring in milling science in the Department of Grain Science and Industry at Kansas State University (KSU), Manhattan, offers a review of his study of how granulation curves can be used to monitor break roll wear. Watson first presented his findings at the International Association of Operative Millers (IAOM) Wheat State, Texoma, and Central district meeting, March 7, Manhattan, KS. He also made a presentation at the 2013 IAOM Conference and Expo, May 2, in Niagara Falls, ON.
Watson’s study and presentation were supervised by Chris Miller, instructor, Department of Grain Science and Industry, and Mark Fowler, a regular contributor to Milling Journal and associate director of the International Grains Program (IGP), Department of Grain Science and Industry, KSU.
Roll wear is a condition that every mill has to manage. Roll wear occurs on rolls when the corrugations wear down as a result of wheat or stock running through the rolls over a period of time.
Maintaining the corrugations on a roll is important for a mill. It helps to keep the mill in balance, maximize farina production in the break system, improve roll performance, and has the potential for energy savings.
On the other hand, roll maintenance is costly for the mill, as well. Changing out a roll pair causes downtime for a mill. There is expense in shipping and re-corrugating the roll pair, labor to replace the roll, and inventory to keep spare rolls and parts on hand for quick changes.
Therefore, finding an optimal time between roll changes to minimize maintenance costs and minimize impact to milling performance could be helpful to mill managers.
The granulation curve is a useful tool that has the potential to examine changes over time in the mill. It illustrates the particle size distribution of the milled product for a specific ground stock.
The granulation curve is helpful in showing the mass flow of product throughout the mill.
There are many factors that can change the granulation curve of a milled product, including wheat type, temperature of the wheat and mill, tempering time and moisture, roll wear, mill load, and balance.
In this study, the theory was tested that roll wear should cause noticeable shifts in the granulation curve over time.
Specifically, as a roll wears down, the miller should have to decrease the roll gap, in order to meet the specified break release.
This wear should cause more compression on the stock through the roll. With an increase in compression, the granulation curve shift should show an increase in fine particles and a decrease in coarse farina over time.
Two local mills that run hard red winter wheat flour agreed to provide samples in this study.
The mills are labeled “Mill A” and “Mill B.” The samples are of ground stock from first break and second break roll stand in each mill on a specific day. A sample of wheat going into the first break roll was included as well.
Sifting and analysis were performed at KSU’s Department of Grain Science and Industry milling laboratories.
First, the wheat was analyzed with the Single Kernel Characterization System (SKCS) for hardness, moisture, weight, and diameter.
Then, each stock for individual passages was sifted. Sifting was performed on a Great Western tabletop sifter box.
The first sifting used sieves with 1041-, 355-, 240- and 132-micron screens. The sample was sifted for two minutes. The weight of each stock on each sieve then was recorded, and the material on top of a 355-micron sieve was saved for a second sifting.
This second sifting analyzed the sizing stock further using 900-, 750-, 630-, 500- and 425-micron sieves. This stock also was sifted for two minutes. The weights then were recorded for each sieve.
One data set that had to be generated was the age of a roll pair, when each stock was collected.
Each stock was collected on a specific day, and the previous roll change was recorded and associated with that sample.
The difference between the stock collection date and the date of the previous roll change gave the age of the roll pair in days after a roll change. This helped to assign consistent ages to compare each curve.
When reviewing the data, it was concluded that Mill A should be our primary focus in this study due to longer average roll life for first break and second break, up to 650 days. This longer time period would be better for our analysis on how changes that occur to the rolls over time.
Granulation Curve Analysis
Once all the data were collected, granulation curves were generated for each break system in each mill.
One observation made was that because of how staggered the roll changes are, and how short of a collection period was utilized, there were several gaps in the roll age that could not be observed. Therefore, selecting data for granulation curve generation was difficult.
However, points were selected and analyzed to space the days evenly over the roll life. Each day includes all data points within +/-5 days of that selected point. The cumulative percent overs were averaged and made into a granulation curve for that period.
In Fig. 1, the granulation curve is generated for Mill A first-break stock. The time periods selected for analysis were days 0, 210, 420, and 600.
One major observation for this stock is some variability in the break release. For example, break release has a major impact in the granulation curve, and this may have impacted the efforts to determine how roll wear could shift the granulation curve.
Fig. 2 shows the granulation curves for second-break stock on Mill A over the roll life.
The break release for this stock is more consistent than the first-break stock. This granulation curve shows much tighter fitting curves. It is difficult to see any changes that determine any effects roll wear has on the curves.
Quantitative Stock Analysis
We moved ahead hoping that the amount of stock produced on specific sieve sizes would indicate better how the individual rolls performed.
For this analysis, the focus was on the second-break stock of Mill A and reducing the variability introduced by the wheat.
The theory is that second-break roll stock may be more consistent by being ground previously, as compared to wheat fed to the first-break roll with variation in hardness and kernel size.
The chart in Fig. 3 shows two trend lines. The top or the decreasing trend line is the percentage of second-break stock that passed through the 1,041-micron sieves and remained over the 500-micron sieve in the same samples.
What is observed is a small, approximately 2%, downward shift in particle size over roll life supporting the theory that roll wear impacts the granulation of ground stock.
Limitations of Study
Wheat variation, sampling procedures, and the changes in mill environment over the period of a year adds variability into this study.
Mill operation factors, including required adjustments to balance the mill flow and product quality, are other elements that introduce some variability to this study.
In the final graph, Fig. 4, the data shows how a change in break release can correlate to a change in flour production.
Over time, the second-break release varies, using a 20-point moving average. This change results in corresponding increases and decreases in flour production for these roll stands.
This shows how important a factor break release can have on the granulation of the stock.
Granulation curves are a way that roll wear can be monitored by a mill.
However, the results from this experiment are not conclusive. There are too many uncontrolled variables in the wheat and milling conditions to see a shift in the granulation from roll wear.
However, a shift in coarse and fine farina production was observed that could indicate roll wear.
Even still, too many variables play a factor in the determination. A controlled laboratory study would be necessary to eliminate some variables and to obtain more conclusive results.