Preserving feed quality through post-pelleting application
July 01, 1996
by Teresa Acklin
By Ruud C.M. Barendse
Part two of two describes equipment and feed handling requirements to assure consistent results.
Post-pelleting application involves incorporating a liquid additive into feed after pelleting. In its most basic form, P.P.A. consists of spraying a modified enzyme formulation onto the cooled feed pellets or crumb, an approach that prevents the loss of the additive's effectiveness during feed pretreatment and pelleting. With due care of process characteristics, P.P.A. is a realistic and economical alternative to incorporating dry additives before pelleting.
In addition to sampling and evaluation procedures to assess homogeneity (see June 1996 World Grain, page 34), P.P.A. processing requires attention to distribution and droplet size of the liquid spray, maximization of the sprayed fraction of feed, thorough mixing of the sprayed feed, and presence or formation of fines during and after P.P.A. processing.
The spraying process in P.P.A. is essentially governed by four parameters: liquid dosage level, droplet size, spray distribution and nozzle throughput. Nozzle throughput usually is fixed, either by capacity in batch processes or residence time in continuous processes, but the other parameters are variable.
In Figure 1, the number of droplets per gram of feed is shown as a function of the dosage level and droplet size. Figure 2 indicates the range of attainable droplet sizes for different nozzle types. Large numbers of droplets per gram of feed can be generated by fairly simple equipment even at the low dosage levels used for undiluted enzyme products.
In Figure 3, a dispersion of additive over feed is depicted assuming Poisson distribution of only 57 droplets per gram of feed; this corresponds to a 100 parts per million dosage level and a nozzle generating 150-micron droplets. From this distribution pattern, a coefficient of variation of 2.7% can be calculated for 25-gram samples.
In practice, the coefficient of variation will be higher because Poisson distribution assumes the droplets are dispersed over the feed independently of each other. In reality, they are applied to the feed as a mist or curtain of droplets, causing them to behave less independently.
To approximate as much as possible the optimal situation shown in Figure 3, a nozzle type with as wide a spraying pattern as possible should be used. This assures a maximum spraying area and dispersion of liquid spray, as well as a higher number of droplets per gram.
The homogeneous distribution in Figure 3 may seem very promising, but it requires more than appropriate spraying equipment to approximate it.
In practice, it is hardly feasible to disperse the liquid additive over all pellets. So the feed bulk will consist of two fractions: the feed pellets that have been sprayed and the pellets that have not.
In Figure 4, the minimal attainable coefficient of variation is depicted as a function of the fraction of feed particles sprayed. Note the number of pellets in the sample also influences the results. The figure assumes that all sprayed pellets have the same amount of additive and that sprayed and non-sprayed pellets are mixed completely.
The figure indicates that for 500 pellets, the equivalent of a daily ration of chick feed, a minimum of 20% of pellets must be sprayed to reach a coefficient of variation of less than 10%. Larger rations set much lower demands; only 5% of 2,000 pellets must be sprayed to reach the same homogeneity.
Figure 4 makes clear that serious attention must be paid to the "hit level," or the fraction of feed pellets sprayed. One or more wide angle nozzles is the first step in obtaining a higher fraction of sprayed pellets.
A second requisite is exposing the maximum number of individual pellets to the spraying process. This can be accomplished by forming a curtain of feed as wide and turbulently moving as possible. In agitated processes, prolonging spray time in batch processes or prolonging feed residence time in continuous processes can provide the desired results.
The data depicted in Figure 4 are based on complete mixing of the feed bulk after spraying has taken place. The importance of this mixing step in the total P.P.A. process usually is seriously underestimated.
Figure 5 indicates the influence of the mixing process on feed homogeneity by comparing the variation coefficients of virtually unmixed and completely mixed samples. Because variation coefficients of more than 100% are impossible, this figure is not completely correct.
However, the message it conveys is clear; thorough mixing of the sprayed and unsprayed feed pellets is essential. Although mixing to some extent occurs spontaneously during feed handling and transport, a greater homogeneity occurs when additional measures are taken to ensure virtually complete mixing.
Pellet size also plays a role in the P.P.A. process. An increase in pellet size means that there are fewer pellets in the same sample weight. Additionally, when feed pellets are coarser, it is more difficult to reach the optimal "independent" distribution of liquid over the feed. These factors tend to reduce homogeneity as pellet size increases.
On the other hand, increased pellet size makes it easier to spray a larger fraction in a given volume of feed pellets, a quality that in most cases compensates for the factors that reduce homogeneity. Because the variation coefficient is more strongly dependent on the fraction of sprayed pellets than on sample size, larger feed pellets may be expected to result in greater homogeneity — assuming use of the same equipment, identical sample sizes and virtually complete mixing.
Because animals with lower daily rations are usually fed smaller pellets, P.P.A. processing of these feeds requires extra attention. The smaller daily rations and the possible reduction of the "hit fraction" make it harder to reach the desired levels of homogeneity.
Another important issue is segregation of feed particles. In crumb feed, the wide range of particle size and the presence of a substantial amount of "fines" make segregation practically unavoidable.
Care should be taken to minimize fines in pelleted feed, as well. Because the spraying process distributes the liquid components over the available surface, fine particles, which have comparatively large surface-to-weight ratios, will absorb a relatively high percentage of the additive.
For example, in a feed with 95% by weight of 3-millimeter particles and 5% by weight of 0.5-mm particles, the fine fraction can be expected to contain up to 24% of the additive. Obviously, sieving off the fines after the spraying process will cause substantial additive losses, resulting in the bulk feed's additive content falling 20% below the target level.
Because of their high additive content, segregation of the fine particles effectively reduces homogeneity. Segregation also increases the chance of contamination of other batches, as fine particles are prone to dusting and can remain present in storage and transport equipment.
If it is not feasible for the fines to be removed from the feed, substantial improvement might be obtained if the P.P.A. system is designed to spray only the coarser fractions, such as those on top of an already segregated mixture. Another solution could consist of assuring the fines adhere to the coarser particles by means of fat addition.
Attrition of feed pellets after P.P.A. processing also will generate fine particles, which have higher additive levels than bulk because these fines are generated by attrition of the sprayed surface. Although probably lower in additive content than fines already present during spraying, they still present the same risks, and care has to be taken when considerable amounts of these fines have formed.
Ruud C.M. Barendse is in the biospecialties division of Gist Brocades, Delft, the Netherlands. This article is based on his 1995 presentation at the Victam '95 international feed and food industries show and conference in Utrecht, the Netherlands.