Colour Grading of Flour at Southern African Mills – It all started with Pioneer Foods and Narich (Pty) Ltd
COLOUR GRADING CAKE AND BREAD FLOUR
Harvesting food grains for human consumptions goes back to the dawn of history and was already well established from before biblical times. Egypt was a major grain source for the Roman Empire, and Marco Polo learned about Pasta from the Chinese who are reported to have consumed noodles as early as 3000 BC.
Modern Food Processing has come a long way since early manual labour days and consistency has become a key requirement in all stages of production. Yield and efficiency are already a given, but as countries may import or export grains, methods have to be developed that can be universally applied.
An emerging parameter that has become increasingly important since the 1950’s is the measurement of the colour of milled grain products, such as cake and white bread flour.
The whiter the flour, the more valuable it became. Longer milling systems yielded whiter and finer flour, however at some point milling extraction and flour whiteness is optimized and over-milling reduces flour brightness and may even change the intrinsic value of the flour. It was time to manage the colour of milled products with reliable tools.
In the 1930’s, spectroscopy emerged as a reliable measurement of visible (VIS) light frequencies, and later infrared (IF) and ultraviolet (UV) became possible as well.
As early as the 1600’s, Sir Isaac Newton did a series of experiments with prisms and light. He showed that prisms split light into the colours of the rainbow, and that the wavelengths for each colour were reliably reproducible.
Whenever light strikes an object, the light changes, not the object. If the change is measurable, repeatable and diagnostic of any property of the object, then light can be used as a convenient way to measure and report colour grading and even today, “Doppio Zero” or double 0 is the Italian standard for ideal Pizza Pasta colour.
The drawback of spectral data however, was that while a certain reading gave a good idea of where the green – white relationship was, the technology did not give good reproducibility from instrument to instrument, mainly because colour measurement is completely reliant on the conditions applied at the time of measurement. Any drift or changes to conditions would result in variances.
Colour is a perceptual sense, and is not an inherent property of the object measured, but subject to the conditions under which the colour was sensed. It became clear that these conditions which may have an influence have to be well defined and managed. So, while a single spectroscope gave fairly good relative information, it was not providing meaningful and comparable information when conditions changed.
Around the early 1930’s, the C.I.E. (the International Commission on Illumination) started to add mathematical algorithms to the raw spectral data, to create “Colour Spaces”. These colour spaces better described the abstract concept of colour, and included the ability to calculate variable conditions within the algorithms. This lead to the most used colour space used in food processing today, namely the C.I.E. colour space L*a*b.
“Ring testing” also became a common and costly practice, where millers would submit their readings to a central approved authority, like in South Africa, the Southern African Grain Laboratory (SAGL).
The SAGL had their own samples as well as all the miller samples and would then “Adjust” the single value numbers to a “Standard” sample. This was to compensate for the lack of Instrument Standards.
A further complication in the process was humidity, so it was deemed that a “Dry” and “Wet” test was required.
Lastly, depending on where the grain was actually harvested, residual soil colour could influence the actual readings, and research looked at the possibility of measuring the flour colour as well is its inherent “Ash” colour.
With the development of the Konica Minolta Colorimeter, commercially called Chroma Meters, a number of companies and countries adopted the CR-200, CR-300 and to still use the CR-410. Again, these devices were much more reliable, but when comparing values from instrument to instrument, one had to allow for a 0.6 to 0.8 ∆ab inter instrument agreement discrepancy. These values are just on the edge of human perception, and together with the experienced miller’s eye, give reasonably reliable results in the single site scenario.
More than ten years ago (2008), Mr Arie Wessels, R&D Manager and now also Manager Grain: Technical at Pioneer Foods approached us to discuss a better way to grade cake and white bread flour.
As Konica Minolta specializes in precise colour measurement and due to our experience with the Japanese Grain company Satake, we agreed to help develop a user friendly solution for quality assurance for colour. Campden BRI were also using Konica Minolta instruments research on food colour (around the same time) and we were able to correlate the millers needs from scientific point of view. Narich (Pty) Ltd specializes only in light based colour applications, and our experience helped to shape a solution.
We decided upon a number of parameters:
- We would use our most accurate and factory friendly Instrument – the Konica Minolta CM-5 Spectrophotometer for R&D and the Konica Minolta CR-5 Colorimeter for in mill measurement
- We settled on the CIE Colour Space Standard L*a*b with colour difference ∆*ab
- We settled on DRY readings only as wet measurements could not be easily standardised. Some researchers incorporated an “Adjustment” factor for their wet measurements which just returned us to manual massage of values rather than actual read values.
- We discounted Ash as a component as a colour measurement device measures what it “Sees” and cannot discern which component returns which colour values, only the whole appearance.
Pioneer foods, in the person of Arie Wessels, wanted a comprehensive scale by which a single measurement would return a single grade, from the whitest to the darkest blends of flour sold by the company. One criterion was that values should correlate throughout the group.
One criterion that Narich (Pty) Ltd insisted upon was the use of a “Target” and “Sample” approach, and to abandon the concept of using absolute values in favour of “Samples” automatically searching by Colour Values for the nearest “Target” values in the Colour Grade range.
This procedure allowed for visibly accurate colour grading as well as normal errors, which are encountered in sensory science. These errors are numerically visible, but visually meaningless, allowing the QA team to deal with visible variances rather than guessing what a numerical difference would “Look like”.
Pioneer Foods agreed to purchase the newly released Konica Minolta CM-5 for their Research and Development Laboratory, together with specific accessories to optimize dry powder colour measurement, as well as SpectraMagic NX Pro, the professional software was used to define settings/conditions, interpret data and to develop the actual colour grades.
We had discussions with the SAGL who also agreed to purchase the same Konica Minolta CM-5 and setup, to be in a position to continue the “Ring testing” procedure, as an independent arbitrator.
Pioneer Foods, R&D staff then carried out what was about a yearlong development of practical colour grade system for their application. At the same time, Narich (Pty) Ltd provided Pioneer Foods with a Konica Minolta CR-5 (the non-spectral data version of the CM-5) to verify correlation in the mill of CM-5 and CR-5 colour grading accuracy. In theory both Instruments had the same “Engine” and only different reporting screens, but both parties wanted to be sure of this. On validation Pioneer Foods Kindly purchased this instrument as well.
There were many meetings with the R&D team, to explain the hardware, the software, the firmware and colour physics in general. At all times we adhered to a policy of actual data only, not “Adjustment” options.
The trickiest part from a sensing point of view is that the colours had to be separated by a tolerance level that would find a colour, but not multiple colours (due to the tolerance being too wide) or no colour (due to the tolerance being too narrow).
The principal we applied for the tolerances was that of a Venn diagram.
A Venn diagram is a diagram that shows all possible logical relations between finite collections of different sets. These diagrams depict elements as points in the plane, and sets as regions inside closed curves. A Venn diagram consists of multiple overlapping closed curves, usually circles, each representing a set. (Thanks to Wikipedia)
Translated into English, as set of L*a*b colour values should be valid for only one circle, but may be secondary in accuracy in surrounding Venn circles. In practice it WORKS!
After the year of R&D and numerous software trials, Pioneer Foods declared their Colour Grade Chart fit for application. It was also shared with SAGL for both their use, as well as that of other even competing mills. This generous stance allowed a “New Standard” to be shared with any group requiring Group Wide continuity, and has in practice been adopted throughout Southern Africa and is growing in popularity in Nigeria and East Africa as well.
The scale has proven to be reliable and accurate.
The Inter Instrument agreement (IIA) between CM-5 and CR-5 models both within models as well as inter model agreement is a key Konica Minolta advantage.
To date there has not been ONE incidence of miss colour grading save if an Instrument is out of certification standards. Fortunately Narich (Pty) Ltd offers comprehensive certification and repair services in the region, backed and certificated by Konica Minolta Sensing Europe. (Bremen Germany)
The software for R&D is robust and once purchased attracts no further charges.
The accessories required make little difference to the overall cost, and ensures the highest IIA possible.
The CM-5 units in the laboratory and the CR-5 units in the field have proven to be robust and reliable, with 99.9% correlation results. In addition, NO PC is required for the in factory CR-5 units, saving time, money and there’s no PC issues holding up production.
SAGL has adopted the Pioneer Foods Colour Grade System as a Standard SAGL procedure.