A control wheat sample often is selected for tempering and milling at the onset of milling wheat samples for crop quality or incoming wheat evaluation. The objective is to confirm laboratory mill settings are a fixed parameter during wheat testing and evaluation. Ideally, mill settings and conditions are well-established and constant, not an uncontrolled variable changing between test runs. Undoubtedly, the same would be expected in a commercial mill setting; however, variations in mill conditions and settings are very well-known and expected despite the operative miller’s attempts to control the variables and bring order to the mill settings and environment.
This year, control samples of wheat were milled every day after mill warm-up to monitor or check for potential changes that might impact wheat evaluation significantly. Milling data are provided in Table 1 on p. 44, including laboratory conditions, tempered wheat moisture, milling results, and flour analysis. The average, standard deviation, maximum, minimum, and coefficient of variation (percent CV) is provided for each measurement.
The laboratory conditions identify the temperature (degrees F) and relative humidity (percent) of the room during the test run. Tempered wheat moisture was measured on the available laboratory moisture tester. Additionally, tempered wheat moisture content was calculated using the dry wheat moisture content and the weight difference (water) between the 4,000g dry wheat sample and the tempered wheat sample. Water addition was calculated mathematically to achieve 16% moisture content tempered wheat. A small amount of additional water was added during tempering to compensate for moisture losses coating the drum and measurement errors.
Recovery of feed and flour was reported based on tempered wheat to the mill with the goal of achieving approximately 95% weight recovery. Flour product was reported based on both tempered wheat to the mill and the total products recovered. The total products yield fails to take into consideration milling and moisture loss, as well as measurement errors.
Straight grade flour was analyzed using an available NIR analysis machine and calibrations. Measurements taken included flour moisture, flour ash, and flour protein content. Ash and protein content were reported on a 14% moisture basis (mb) as is often common in the milling and baking industry.
A correlation matrix for the variable measured was conducted using Microsoft Excel add-in Analysis ToolPak. The correlation matrix is provided in Table 2 on p. 44, while Table 3 on p. 45 provides a brief description of variable strength of association, given the identified correlation coefficient. Most researchers in physical science often discount lower correlation coefficients below 0.8 or 0.9. However, even weak or moderate correlations can be statistically significant. More importantly, they can have a significant practical impact, as is the case with moisture or milling loss.
The light gray cells under “Laboratory Conditions” in Table 3 should be discounted, as there is no reasonably expected correlation between milling laboratory room conditions (i.e., temperature and relative humidity) and tempered wheat moisture measured or calculated by difference. The relative humidity of the milling room has a very strong positive correlation with flour moisture (r=0.85) as would be expected. Flour ash has a negative strong correlation to laboratory relative humidity (r=-0.64). As one might expect, lowered relative humidity allows product to dry out during milling and allows bran to shatter more easily, creating bran powder which has an ash content of approximately 7%.
Surprisingly, the correlation coefficient between the calculated moisture content and measured moisture content of tempered wheat to the mill is poorly associated (r=0.32). It seems reasonable to expect a strong correlation between measured and calculated moisture content for tempered wheat.
The correlation between measured tempered wheat moisture and flour moisture (r = -0.61) is especially interesting. The loss of moisture between tempered wheat and flour in milling is well known. It appears that relative humidity of the milling environment is a variable that impacts the measured tempered wheat moisture and flour moisture, somewhat confounding this relationship. What would happen if the relative humidity and moisture addition were better controlled and observed at different relative humidity and tempered wheat moisture levels?
The two measures of flour yield, wheat basis, and total product bases are moderately positively correlated (r=0.53). Flour yield based on total products has a strong negative correlation to product recovery (-0.67). The better job done recovering products from the laboratory mill, the lower the total product flour yield. Both flour yield measures based on tempered wheat and total products have a moderately positive correlation to flour protein content with correlation coefficients of r=0.57 and r=0.54, respectively.
The concept of protein recovery or protein loss between wheat and resulting flour can have a significant economic impact, especially when protein recovery is poor or protein spread between wheat and flour is too great. Making up low flour protein with high-protein wheat or gluten can be costly. As expected, there is a moderate relationship between flour ash content and flour yield.
It appears that relative humidity of the milling environment is a variable that impacts the measured tempered wheat moisture and flour moisture, somewhat confounding this relationship.
Flour ash and protein have a moderate-to-strong negative correlation related to flour moisture content with r=-0.53 and r=-0.68, respectively. As one might expect, flour ash and protein increased as flour moisture decreased. Ash and protein have a strong positive correlation of r=0.78.
Given a resulting flour with low protein, one might try to grind a bit more aggressively to increase flour protein, but if the tempered wheat moisture and/or relative humidity is low, it will be difficult to increase flour protein without increasing flour ash. More importantly, the increase in protein is not likely to contribute to improved performance, as the additional protein recovered is not likely to be the gluten protein desired in the flour.
Laboratory room temperature was tightly controlled with a coefficient of variation of less than 1.5%. This variation had little, if any, impact on milling results or flour quality. Laboratory relative humidity is impactful with respect to milling and flour quality and should be controlled more tightly, perhaps with at a 5% CV or less. The difference between measured and calculated moisture average (0.47%) and variation is too great for a controlled system used to evaluate crop performance.
The greatest variability was observed in the measured moisture rather than the calculated moisture content. Clearly, work is needed to improve moisture measurement and control. Milling results based on wheat to the mill appear to be more uniform than those based on total products yield. Yield based on wheat to the first break roll is a better indicator and less misleading than a total product yield, and it relates better to commercial practice.
Clearly, the yield variation is greater than would be tolerated in a commercial setting. The solution would be to run a larger sample to achieve a steady-state operation over a long period of time, which is not a practical solution due to sample size and time constraints. Flour quality attributes for the control sample appear to reflect variability observed in the commercial setting reasonably. Milling of a control wheat sample during crop evaluation provides an opportunity to assess laboratory operational performance and reveal improvement opportunities for future crop analysis.
So what is going on in your mill? How well is tempering monitored and measured? Is your tempering process under control? What about changes in temperature and relative humidity? Undoubtedly, temperature changes and perhaps relative humidity in your mill are far greater than observed in the laboratory. How do these changes impact mill performance, protein recovery, and flour quality?
Perhaps the Internet of Things, automation, and data analysis can provide an opportunity to identify significant variables and their impact for improved monitoring and control.
Dr. Jeff Gwirtz is CEO of JAG Services, Inc., an international consulting company in Lawrence, KS; 785-341-2371; firstname.lastname@example.org. He also is adjunct professor in the Department of Grain Science and Industry at Kansas State University, Manhattan.