Achieving blend uniformity

COMMON to all processors is the concept of a recipe. Wherever two or more materials are combined, a recipe exists. Whether it's called a recipe, a formulation, blend, or compound, the notion is the same: to specify ingredient proportions to create a product possessing certain carefully defined properties, attributes or characteristics. Small-scale processing operations may proportion ingredients manually (either by volume or weight), but continuous or batch feeders usually perform the proportioning operation where throughputs are higher.

Whilst many factors contribute to end-product consistency, the ingredient proportioning system ranks especially high in importance. Proportioning is the point where ingredients merge, and no matter how thoroughly the blend stream is then mixed, melted, conveyed or otherwise processed, the proportioning point remains the best opportunity to achieve end-product uniformity.

However, a search of available literature shows remarkably little attention has been paid to this important aspect of the processing operation.

The challenges

Where continuous feeders proportion a recipe's components, characterising the uniformity of the resulting blend stream's constituent ingredients presents some challenges. Even though each feeder may be set to its targeted proportion, each feeder possesses its own innate variability as revealed by the measured repeatability of its discharge rate (+/-0.5% of set rate @ 2 sigma, for example). These ingredient variations occur independently from feeder to feeder. How can these variations be combined to produce a single composite measure of blend uniformity?

Adding to the uncertainty is the fact that, by the very nature of the proportioning operation, some ingredients will dominate the blend, whilst others will be present in only small proportions. How can account be taken  for the widely different ingredient percentages that exist in a recipe?

And lastly there is the issue of timescale. It is reasonable to expect that  samples of the blend stream taken over a one-second period would differ in their ingredient proportions more so than would those from oneminute samples where momentary variations have had an opportunity to settle to an average.  How can this timescale effect be handled?

Constructing a Blend Uniformity Index (BUI)

Any single index of multi-ingredient blend uniformity should meet the following requirements:
1. The index's range should be defined to span the full range of potential variation within the subject recipe and be standardised to permit comparison  between recipes and proportioning systems.
2. The measure should appropriately account for the different contributions to overall formulation variability from ingredients depending on their associated feeder accuracies.
3. The measure should appropriately account for the different contributions to overall formulation variability from ingredients depending on their respective recipe proportions.
4. As part of its expression the measure should be qualified as to timescale and statistical confidence level.

These four requirements suggest a clear path to the construction of a Blend  Uniformity Index (BUI), which establishes a practical measure of proportioning system performance.

Boosting uniformity through feed accuracy

Simply operating within established ingrwomens nike shoes