Single kernel analysis for fractionating cereal batches

Grading is conventionally based on differences in weight, colour and/or size. A relatively novel method for grading batches relies on evaluation in the near infrared spectrum. This is used to calculate protein and water contents.

Using this technology, a Northern European manufacturer has developed a grading machine that can scan, analyse and subsequently fractionate each and every kernel of a cereal batch. This serves to separate inhomogeneous batches from homogeneous ones; this can potentially raise quality significantly and considerably facilitate processing. In 2011, this innovative technology has been tested at the Research Center Weihenstephan, Freising, and the grading result has been evaluated in terms of quality. A technological description is given below, and selected results from process and equipment validations are then presented, concluding with a final discussion.

Equipment setup and operating principle

The patented equipment (figs. 1, 2) comprises three main components. The core is a rotating cylinder into which the material to be graded is introduced continuously. In order to capture and analyse each and every kernel, the cylinder is composed of rows of recesses having an array of pockets, each receiving a single kernel.

The shape and thus the number of pockets per circular path is a function of the cereal to be tested. Kernels are kept in position and moved upwards by the centrifugal force resulting from the rotational movement. At the top, the individual kernels are detected by near infrared spectroscopy, to be precise by near infrared transmission spectroscopy (NIT). In contrast to reflection measurement, each and every kernel is completely screened at numerous measurement points distributed over the length of the kernel. Spectral information goes to a computer at high speed where it is evaluated. It is thus possible to measure protein content and water content of each individual kernel. However, prior to measurement, the unit should be calibrated. Depending on the limit values previously set, the kernels are subsequently blown off into the fractions selected. Each of these modules can grade up to three tons of cereal per hour, and several units can be combined into larger systems.

The unit is designed such that access to the most important components is easy and unproblematic for repair and maintenance.

Validation of the unit

The unit was validated in several tests with 2-rowed spring barley and 6-rowed winter barley. Homogeneous samples of the different fractions were taken on-site while analyses were carried out in the accredited laboratories of the Weihenstephan Research Center. The material presented here is limited to that relevant to the particular investigation. Thus, a number of selected results for spring barley only are described below. Figure 4 is an example of the protein content measured for ten fractions that had been separated out successively during calibration, as well as of the average measurement result for the ungraded batch and the grading result for the whole batch.

Figure 4 shows that protein content in the fractions increases almost linearly. It is also obvious that protein content of the first half of the fractions is below and of the second half of the fractions above the protein content of the ungraded unmalted barley.

The picture is also an obvious one for kernel size: Figure 5 shows that the proportion of larger kernels drops with increasing fraction. The proportion of kernels larger than 2.8 mm is 85.7 per cent in fraction 1-10 per cent while only 60.2 per cent contain such kernel size in fraction 90 – 100 per cent. Figure 5 also shows that fraction 2 of the overall batch grading exhibits the same grading pattern as the ungraded unmalted barley in all categories, except for a difference of 0.1 per cent. As expected, the result for the thousand kernel weight correlated with that of the hectolitre weight in terms of size and, thus, dropped with increasing fraction.

Manual grading showed that the number of half or damaged kernels goes up drastically in the higher fractions as of 80 – 90 per cent. It also became apparent in manual grading that these fractions contained increased amounts of foreign kernels. In addition, husk damage with loss of the embryo was also noted more frequently ).

No significant differences between fractions were found when evaluating germination ability and germination energy. However, mention should be made of the fact that the quality of the particular batch was very high.

Moreover, no clear-cut result was obtained in terms of red kernels, relevant red kernels, open and hidden sprouters as well as the mycological state for all cereals analysed here.

Conclusion

In summary, this highly complex and ultramodern technology functions surprisingly accurately and reproducibly (and that has been tested). It makes it possible to fractionate cereal batches according to protein content and, in this process, analyses every single kernel.

Using such processes, the quality of lower quality batches can also be improved, thus reducing losses and optimizing processability.

It may also be conceivable in future to vary selectivity of the unit by changing the settings. It may thus be possible to remove infected or gushing-relevant kernels from a batch. Additional detectors could also be integrated in the unit. Both the processing industry and the seeds industry could benefit from such processes.

The manufacturer has meantime developed a new system generation. All sorts and types of correlations are being tested in parallel on a laboratory scale.