|
Potential of Dimensional Measurements of Individual Pellets for Evaluating Feed Pellet Quality
Lester O. Pordesimo1,*, C. Igathinathane2, Basil D. Bevans3, David P. Holzgraefe3
Published in Applied Engineering in Agriculture 38(5): 777-785 (doi: 10.13031/aea.14845). 2022 American Society of Agricultural and Biological Engineers.
1 Stored Products Insect and Engineering Research Unit, USDA ARS Center for Grain and Animal Health Research, Manhattan, Kansas, USA.
2 Department of Agricultural and Biosystems Engineering, North Dakota State University, Fargo, North Dakota, USA.
3 ADM Animal Nutrition, Quincy, Illinois, USA.
* Correspondence: lester.pordesimo@usda.gov
The authors have paid for open access for this article. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License https://creative commons.org/licenses/by-nc-nd/4.0/
Submitted for review on 9 September 2021 as manuscript number PRS 14845; approved for publication as a Research Article by Associate Editor Dr. Ajay Kumar and Community Editor Dr. Sudhagar Mani of the Processing Systems Community of ASABE on 2 August 2022.
USDA, NDSU, and ADM are equal opportunity providers and employers. The mention of trade names of commercial products in this article is solely for the purpose of providing factual information and does not imply recommendation or endorsement by the aforementioned organizations.
Highlights
Abstract. Pellet durability index (PDI) detailed in ASAE Standard S269.4 (ASAE Standards, 2007) has been the widely accepted measure for pellet quality in the U.S. and, by extension, the measure for evaluating the effect of ingredients and pelleting process variables on the quality of pelleted feed products. The PDI is calculated as the mass percentage of intact pellets remaining after tumbling a 500 g sample in a tumbling can durability tester for 10 min. In the case where pellet quality is good, oftentimes the resulting PDI for different experimental treatments are very close in magnitude. In these situations, it is desirable to have another measurement that would allow for finer discrimination among treatments. It was hypothesized that the average linear dimensions of animal feed pellets in a unit mass sample would vary as a function of formulation and pelleting process variables for a consistent knife setting in the pellet mill. This hypothesis was tested in a study involving effectiveness testing of varying inclusion levels of a potential pellet binder in a typical corn-soy swine diet pelleted by both conventional and cold pelleting processes. Pellet lengths and diameters measured by a machine vision implemented in ImageJ matched to measurements taken manually and varied with treatments. Pellet length varied with treatments but could not be a good discriminator of pellet durability between pelleted products because of the narrow range in measurement numbers (8.53 to 11.15 mm by machine vision). With the wider range in numerical values of the PDIs obtained through aggressive durability testing (23.0% to 81.0% in a tumble can with steel hexagonal screw nuts versus 91.0% to 97.5% in a tumble can without hexagonal nuts), PDI from aggressive testing is the better discriminator of quality among pelleted products because of its greater resolution.
In the animal feed manufacturing industry, the pelleting process provides a means of molding a blend of ground nutritional ingredients into dense free-flowing agglomerates known as pellets (Behnke, 2009). Pelleting is accomplished through a mechanical process combining moisture, heat, and pressure. The benefits of pelleting include enhanced handling characteristics of feeds (increased bulk density and flowability, and reduced ingredient segregation), improved feed efficiencies, and destruction of some deleterious organisms (Behnke, 2001; Fairfield, 2005). Recognizing the former two benefits in particular, livestock and poultry producers and those others raising animals and poultry have come to expect better quality in the pelleted products they produce or purchase. Justifiably so because feed is the largest single cost item for livestock and poultry production, accounting for 60% to 70% of the total cost in most years (Lawrence et al., 2008). For producers, the quick indicator of pellet quality is the absence of an inordinate amount of fines in the product. In 2019, the quantity of feed fed to beef cattle, hogs, broilers, dairy cattle, egg layers, and turkeys in the U.S. was 241.9 MT (Decision Innovation Solutions, 2020). Based on research on the effect of feed form on the feeding efficiency of poultry and advice given by poultry nutritionists, most of the 82.2 MT fed to poultry is pelleted. With this amount of feed, any fractional improvement in pellet quality leads to better feed efficiency and a significant improvement in the cost position of the producer. For this assessment, a reliable and sensitive measurement of pellet quality is required.
To minimize fines production, pellet durability index (PDI), detailed in ASAE Standard S269.4 (ASAE Standards, 2007), has become the widely accepted measure for pellet quality and, by extension, the measure for evaluating the effect of ingredients and pelleting process variables on the quality of pelleted animal and poultry feed products (Tabil and Sokhansanj, 1996; Winowiski, 1998; Cavalcanti and Behnke, 2005; Kaliyan and Morey, 2009). The PDI is calculated as the mass percentage of intact pellets remaining after tumbling a 500 g sample in a tumble can pellet durability tester for 10 min. In the cases where pellet quality is good, oftentimes the resulting PDIs for different experimental treatments are numerically very close to each other. Thus, PDI becomes ineffective in selecting formulations, process variables, or combinations of these that produce the best quality pellets. In these situations where measurements are nuanced, it is desirable to have another measurement that would enable a finer discrimination among treatments in a pelleting study. It was hypothesized that linear dimensions of animal feed pellets would vary as a function of formulation and pelleting process variables for a consistent knife setting in the pellet mill. In this regard, the mean linear dimensions, namely length and width (diameter), of pellets in a unit mass sample could be used as a secondary and maybe even another outright measure for pellet quality. It was also hypothesized that the number of pellets and pellet bits in a unit mass sample would be sensitive to changes in ingredient and processing variables to make them effective in discriminating pellet quality. These hypotheses were tested in a study involving effectiveness testing of varying inclusion levels of a potential pellet binder in a typical corn-soy swine diet pelleted by both conventional and cold pelleting processes.
The difficulty with manual measurement of pellet dimensions is that it is tedious and can be difficult to accomplish especially in the case of what are called in the feed industry as mini-pellets, pellets that have diameters of about 2.39 mm (3/32 in.). This same difficulty applies to accomplishing any kind of counts or selections from a sample through visual examination. This difficulty can be overcome by employing machine vision for the linear measurements, which may be implemented in an instrument and even be automated. Application of machine vision in evaluating animal feed pellet quality has not been widely investigated so its use in this study provides an initial assessment of the value of this technology in pellet quality evaluation in the feed industry. The objective of this research was therefore to evaluate pellet length, pellet width, number of pellets, and number of pellet bits from a unit mass sample as indicators of pellet quality in addition to PDI. Measurements of pellet linear dimensions were taken manually and through the machine vision and then compared. The impact of this research relates to the efficient utilization of feed in livestock and poultry production agriculture.
Materials and Methods
Pellet Production and Quality Evaluation
Animal feed companies and poultry integrators producing large numbers and quantities of products in pellet form are proactive in testing pellet binders that are commercially available, in the process of being commercialized, or in development for efficacy and cost. These binders improve the durability of pelleted products (Acar et al., 1991; Boney, 2019; Abdollahi et al., 2012) and their quality is perceived through visual inspection. In this study, a spray-dried product from the wood processing industry with perceived pellet binding potential was tested in a typical corn-soy swine diet at mass fraction inclusion levels of 0.0%, 1.0%, 2.0%, and 4.0%. Pellets of each formulation were produced in a 22.4 kW (30 hp) CPM Master Mill (California Pellet Mill Co., Crawfordsville, Ind.) equipped with a 4.76 × 50.80 mm (3/16 × 2.00 in.) die by conventional pelleting with pellets exiting at increased temperature and through a cold pelleting process with pellets at relatively colder temperature (Gao et al., 1999). Cold pelleting is the pelleting of a feed mash without steam conditioning of the meal that was termed pelleting without steam conditioning by Skoch et al. (1983) and dry pelleting by Summers et al. (1968) much earlier. Testing of the binder through both conventional and cold pelleting processes was undertaken to provide a comprehensive evaluation of the binder capability. Temperature of pellets from conventional pelleting averaged 77°C (170°F) whereas pellets from the cold pelleting process had a lower average temperature of 49°C (120°F). Thus, products from each pelleting process are referred to henceforth in this paper as hot and cold pellets. For each formulation-pelleting process treatment combination, 90 kg (200 lb) of product was produced.
The study was a 2-factor experiment [pelleting process (2), binder inclusion (4)] with eight treatment factor combinations. It was replicated twice and conducted following a completely randomized design. Based on preliminary runs and experience in commercial pelleting and the consistency of the process, two replications were found reasonable given the large sample size (90 kg). Mean values are reported. All statistics were calculated with the MIXED procedure of the SAS/STAT® software, Version 9.4 for Windows [Copyright © 2014 SAS Institute Inc. SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc., Cary, N.C.]. Samples for quality analysis were collected from the middle of each production run. Measurements taken were PDI (standard pellet durability testing – PDI without nuts), PDI with steel hexagonal nuts [three 19 mm (¾ in.) and three 13 mm (½ in.)] added into the tumble can for a more aggressive pellet durability testing (aggressive pellet durability testing – PDI with nuts), and pellet linear dimensions (length and width) before PDI testing. This aggressive pellet durability testing is the same modification to the standard tumble can pellet durability test making it more aggressive employed at Kansas State University (Stevens, 1987; Moritz et al., 2002; Cavalcanti and Behnke, 2005) where the measurement is referred to as modified pellet durability index (MDPI). Such an aggressive pellet durability testing is also provided by the Holmen (Thomas and van der Poel, 1996; Winowiski, 1998), Ligno tester (Winowiski, 1998), and DURAL (Larsen et al., 1996; Sokhansanj and Crerar, 1999; Adapa et al., 2005) test devices with the test completion in about 60 s.
A comparison of pellet linear dimensions (length and width) measured before and after pellet durability testing could have provided another measure of pellet destruction during handling but the test material after durability testing was not retained. Counts of pellets and pellet bits in a 10 g sample were also taken as potential pellet quality discriminators. Pellet bits were determined as any particle that did not exhibit the cylindrical form associated with whole pellets – no defined diameter. The 10 g sample was obtained by scooping out pellets with a spoon from the top center of the mass that had been poured into a polystyrene antistatic weighing dish. By collecting the measurement material in this manner, fines in the product would not be included in the 10 g. Since the collection of a sample for the counts could never be exactly 10 g, the results were calculated to 10 g by ratio and proportion.
Linear dimensions of the whole pellets in the 10 g sample collected for pellet and bits counts were measured manually using a digital Vernier caliper. Pellet dimensions were also measured through machine vision from a totally separate sampling of the test material. Granted that pellet length could be used as an indicator of pellet quality, the speed of machine vision will be needed for application in commercial practice. In this study, a system developed in academia as an alternative to expensive commercial machine vision systems was employed. Wood (1987) and Colley et al. (2006) also manually measured pellet dimensions. Wood (1987) related pellet length to pellet durability, which is the objective of this study as well, while Colley et al. (2006) took pellet length as a response factor to pelleting process variables to evaluate pellet quality.
Pellet Dimensional Measurements by Machine Vision
The machine vision method for measuring pellet dimensions involved a plugin coded in Java for use within ImageJ for the determination of orthogonal length and width of singulated particles from digital (Igathinathane et al., 2008, 2009a,b). The process for particle size measurement and distribution analysis is outlined in figure 1. ImageJ is a Java-based public domain image processing program developed at the National Institutes of Health in the U.S. Department of Health and Human Services (Bailer, 2006; Burger and Burge, 2008; Rasband, 2008). The plugin had been previously applied to measure orthogonal dimensions of eight types of food grains (Igathinathane et al., 2009b). In that testing, the plugin had overall accuracy greater than 96.6%, computational speed of 254 ± 125 particles/s, capability to handle all tested shapes and particle orientations, and repeatability in measurements. The plugin has been recently applied to evaluating fish feed pellets (Igathinathane et al., 2011).
 |
| Figure 1. Measurement of individual pellet dimensions and using machine vision methodology through ImageJ plugin. |
A consumer-grade flatbed scanner (4800×9600 dpi; Cano-scan 4400F, Canon U.S.A. Inc., Lake Success, N.Y.) was employed to acquire color images of a sample of each the treatment factor combination pellets. The test material was spread out in a single layer across the scanner window and, as best possible, without particles touching or overlapping one another (singulated arrangement). Particles touching one another could have been addressed with advanced image processing techniques, for example, successive erosion-scrap-dilation sequence (Shahin and Symons, 2005). However, for simplicity, the particles were manually singulated. While this required additional care and effort in spreading the particles, it greatly simplified image preprocessing. It was this manner of material presentation that was considered in this study.
The “pixel-march” method, which compares pixel colors to determine object boundaries, was the procedure used in the plugin for determining dimensional measurements (Igathinathane et al., 2009b). The pixel-march routine in the plugin utilized the centroid coordinates of an ellipse fitted by ImageJ around the particle image plus the major axis inclination (fig. 2). The pixel-march starts from the centroid of each identified pellet and proceeds along the fitted-ellipses’ major and minor axes for boundary identification. A sample color image of scanned pellets and the preprocessed binary image from which the dimensional measurements were made are presented in figure 3. The actual length and width measurements of cylindrical pellets were performed using the pixel-march method and the measurement directions from the centroid were indicated by crosshairs (fig. 4).
Results and Discussion
Pellet Durability Index
Results of the pellet durability testing (PDIs) are presented in figure 5. Durabilities of the pellets were affected by both binder inclusion and pelleting process (P < 0.001). For reference, the moisture profile of these pellets produced and tested for durability is given in table 1. The moistures are in line with measurements of other feed products reported (Moritz et al., 2002; Hott et al., 2008; Fahrenholz, 2012). The pellets were tested for durability after a lapse of one day stored in paper feed bags at room temperature. For both pellets produced conventionally and by cold pelleting there was increase in PDI with increasing inclusion levels of the binder (fig. 5). After a fairly larger increase in durability from no binder to a 1% binder inclusion, the resultant PDIs were increasing in a linear fashion within the inclusion levels tested (P < 0.001). The slope of the trendline of PDI with binder inclusion was greater for the PDIs from the aggressive testing with the numerical range of the results being wider (23.0% to 81.0% vs. 91.0% to 97.5%). It is evident that a more definitive discrimination between treatment effects could be made from the results of the aggressive PDI testing because of its greater resolution (larger range in measurement values). This study demonstrates that aggressive durability testing with the inclusion of steel hexagonal nuts into the standard tumbling can tester provides pellet durability index results that are more reflective of adjustments to improve pellet quality.
 |
| Figure 2. Pixel-march method for length and width determination in machine vision methodology. |
 |
| Figure 3. (a) Original scanned color image at 1270 dpi [= 0.02 mm resolution]; (b) Binary image mask after thresholding in the range of 50 to 255 and eliminating particles smaller than 10,000 pixel area [= 2 mm × 2 mm = 4 mm2]. |
 |
| Figure 4. Pixel-march method-based measurements of length and width of pellets (Cross-hairs indicate the directions of measurement; Length = Width; only a small window of the image is shown). |
| Table 1. Moisture profile of mash and pelleted test products. | |||
| Binder Inclusion (%) | Product Form | Moisture Content (% w.b.)[a] | Water Activity (aw)[b] |
| 0 | mash | 14.72 | 0.72 |
| 1 | mash | 14.34 | 0.71 |
| 2 | mash | 14.24 | 0.70 |
| 4 | mash | 14.09 | 0.68 |
| 0 | cold pellet | 13.25 | 0.68 |
| 1 | cold pellet | 13.34 | 0.68 |
| 2 | cold pellet | 13.20 | 0.68 |
| 4 | cold pellet | 13.10 | 0.66 |
| 0 | hot pellet | 14.04 | 0.71 |
| 1 | hot pellet | 14.06 | 0.70 |
| 2 | hot pellet | 14.16 | 0.68 |
| 4 | hot pellet | 13.76 | 0.69 |
[a] Air-oven method - ASAE S269.4 [b] Pawkit water activity meter (Decagon Devices, Inc., Pullman, Wash.) | |||
It is interesting to note that the durability of pellets produced conventionally was consistently higher than for the corresponding pellets produced by cold pelleting. This further confirms that the transformation of the proteins and starch in the ingredients into a melt at high temperature [particularly the plasticization of the proteins (MacBain, 1966; Winowiski, 1985; Abdollahi and Ravindran, 2013)], and their setting upon cooling (particle bonding) is a significant factor in creating the physical integrity of pellets.
 |
| Figure 5. Effect of binder inclusion and pelleting temperature on pellet durability index. |
Pellet Dimensions
 |
| Figure 6. Mean pellet lengths obtained through manual and machine vision measurements. |
Means of pellet length measurements are presented in figure 6. The general trend shown is that pellet length increases linearly with binder inclusion (P < 0.001), and by extension, with durability. Wood (1987) likewise found the relationship between pellet length and durability approximately linear. Considering the machine vision mean length because of the greater number of measurements comprising that mean, the increase in pellet length was more pronounced for the cold pellet than for the hot (P < 0.01). On average, the cold pellets were 1.36 mm longer than the hot pellets [10.36 vs. 9.00 mm (0.41 vs. 0.35 in.)]. It could be that since these pellets are colder at production such pellets can better retain the moisture in the mash and any added moisture in the pellet mill conditioner and thus they are more flexible and do not break at shorter lengths upon exiting the die. However, the moisture data presented in table 1 does not support this supposition. Rather, the cold pellets had a moisture content 1% less than the mash. This fact thereby indicates that it is not flexibility in the cold pellets that is the explanation for the differences in length but rather the brittleness of the hot pellets. With starch gelatinization, even though studies indicate that it is mostly surface (Abdollahi and Ravindran, 2012), the hot pellets are more brittle and thus tend to fracture at a shorter length upon exiting the die. From the feel of the pellets by hand after production in the pellet mill, it is clear that brittleness of the hot pellets in contrast to the flexibility of the cold pellets resulted in breakage during the handling after production that then led to shorter pellet lengths. Elastic recovery of the hot pellets may have occurred but not to extent of simple fracture of the hot pellets during handling.
The length measurements were close between the manual and machine vision measurements with the machine vision grand mean for all treatments lower by 1.0 mm [10.66 vs. 9.68 mm (0.42 vs. 0.38 in.)]. The mean number of pellets manually measured for the two replicate samples for each treatment are listed in table 2. The numbers do not belie the detail and amount of time needed to complete the length measurements manually.
| Table 2. Mean number of pellets manually measured for length. | |||
| Binder Inclusion (% w/w) | Pelleting Process | Number of Pellets in Sample | Pellet Length (mm) |
| 0 | Cold | 91 | 11.16 |
| 1 | Cold | 84 | 12.12 |
| 2 | Cold | 99 | 11.08 |
| 4 | Cold | 85 | 12.14 |
| 0 | Conventional | 128 | 8.81 |
| 1 | Conventional | 126 | 9.27 |
| 2 | Conventional | 114 | 9.93 |
| 4 | Conventional | 108 | 10.84 |
Since the length measurements for pellets produced by conventional and cold pelleting process varied by only 2.03 mm and 1.06 mm, respectively, pellet length does not seem to be a good secondary discriminator of pellet durability among the test pelleted products without a large number of replications in combination with statistical analysis. However, Abdollahi and Ravindran (2013) found young broilers (7-14 d of age) showed a preference for shorter pellets (3 mm) compared with longer pellets (5 mm). This finding indicates that pellet length could by itself be a factor in pellet physical quality, albeit, only for poultry. Abdollahi and Ravindran (2013) even forwarded that it may be possible to improve broiler performance if the appropriate pellet size for different growth periods is identified. Manually measuring pellet length from even a small sample is tedious, thus this presents an opportunity for the use of machine vision method applied in this study for quality assurance and control.
Pellet widths (diameters) measured through machine vision because of the greater sensitivity in that method are plotted in figure 7. As evident in the figure, widths were not markedly different between formulations and had very narrow standard deviations (fig. 7). This was expected because pellet width was physically constrained by the die hole. The measurements 5.13 and 4.91 mm (0.20 and 0.19 in.) for hot and cold pellets, respectively, indeed closely matched the 4.76 mm die hole opening. This measurement further reinforces the fact that the moderate temperatures and pressures in the pelleting process are not high enough to cause significant expansion of whatever starch-protein melt is created in the process as is the case in extrusion (Patil et al., 2007; Kristiawan et al., 2020).
 |
| Figure 7. Mean pellet widths from machine vision measurements. |
Pellets and Pellet Bits Counts
By visual examination, the number of pellets per 10 g and number of pellet bits per 10 g seemingly exhibit an inverse relationship with binder inclusion (figs. 8 and 9), particularly for conventional pelleting. Since PDI has been shown to have a direct relationship with binder inclusion, the number of pellets per 10 g and the number of pellet bits per 10 g have an inverse relationship with PDI. That is, when PDI was increasing with binder inclusion, pellets and bits numbers were decreasing with binder inclusion. There were therefore a larger number of pellets of shorter length (table 1) when binder levels were less, and there was also a larger number of pellet bits. However, analysis of variance showed that it was only the results for the number of pellets per 10 g that tended to vary with binder inclusion (0.05 < P < 0.10) with the trend being a decrease in a linear fashion for pellets from conventional pelleting (P < 0.05). These results only mean that with a reduced level of binder included, the pellets produced tended to break into shorter pieces after production and, in the process, bits were also generated. This is consistent with the PDI results already presented. Thus, these measurements could be indicative of pellet quality in an inverse manner, but application in commercial practice would be difficult because selection and counting from a sample without the benefit of machine vision method is tedious.
 |
| Figure 8. Mean number of pellets in a 10 g product sample. |
 |
| Figure 9. Mean number of pellet bits in a 10 g product sample. |
Conclusions
In evaluating the effectivity of a potential pellet binding material from the wood processing industry in pelleting tests with a typical corn-soy swine diet, the linear dimensions of pellets were measured from products of the experimental treatments along with pellet durability index (PDI). Numbers of pellets and pellet bits in the sample were also counted from unit mass samples. Binder inclusion and pelleting process treatments in the experiment affected PDI as well as pellet length. The PDIs obtained indicated that the proposed pellet binding material was effective in increasing pellet durability and the improvement was increasing with inclusion level up to the maximum 4% inclusion tested. This trend was evident in pellets produced through both conventional pelleting and by pelleting using a patented cold pelleting process (Gao et al., 1999), but for the same formulation, PDIs from conventional pelleting were higher than corresponding pellets produced by cold pelleting. Variation in mean pellet length followed the same trend exhibited by the PDI data. Pellet lengths were more and pellet numbers per 10 g sample were less for cold pellets. The direct opposite trend was found for hot pellets. Pellet widths (diameters) were not significantly different between diet formulations likely because pellet width was physically constrained by the die hole. The 5.02 mm grand mean width measurement closely matched the 4.76 mm die hole size. Pellet length and width measurements through the machine vision method closely compared to measurements taken manually thereby confirming the accuracy, convenience, and value of machine vision should pellet length become a consideration in evaluating pellet quality as has been suggested for poultry diets. However, since length measurements for the cold and hot pellets varied only by 1.06 and 2.03 mm, respectively, pellet length cannot be a good discriminator of pellet durability among the test pelleted products. With the wider range in numerical values of the PDIs obtained aggressively, PDI from aggressive testing appears to be the better measurement for discriminating quality among pelleted products and assessing the effects of ingredients and process variables on pellet quality. Considering that aggressive pellet durability testing provides better discrimination between pellets produced with different treatment factor combinations and that durability testing with the tumble can takes 10 min. to complete, it may be worthwhile to consider using the commercially available aggressive pellet durability test devices in quality assurance because tests in those devices are completed in about 60 s.
Acknowledgments
The assistance of Frank Bockhold, Feed Technology Technician at ADM, in the production and testing of the pelleted feeds is gratefully appreciated.
References
Adapa, P. K., Schoenau, G. J., Tabil, L. G., & Sokhansanj, S. (2005). Cubing characteristics of fractionated sun-cured and dehydrated alfalfa chops. Appl. Eng. Agric., 21(4), 671-680.
Abdollahi, M. R., & Ravindran, V. (2012). The quest for the perfect pellet: Issues and challenges. Proc. New Zealand Poultry Industry Conference, 11.
Abdollahi, M. R., & Ravindran, V. (2013). Influence of pellet length on pellet quality and performance of broiler starters. J. Appl. Poult. Res., 22(3), 516-522. https://doi.org/10.3382/japr.2013-00736
Abdollahi, M. R., Ravindran, V., Wester, T. J., Ravindran, G., & Thomas, D. V. (2012). Effect of improved pellet quality from the addition of a pellet binder and/or moisture to a wheat-based diet conditioned at two different temperatures on performance, apparent metabolisable energy and ileal digestibility of starch and nitrogen in broilers. Anim. Feed Sci. Technol., 175(3). https://doi.org/10.1016/j.anifeedsci.2012.05.001
Acar, N., Moran Jr., E. T., Revington, W. H., & Bilgili, S. F. (1991). Effect of improved pellet quality from using a calcium lignosulfonate binder on performance and carcass yield of broilers reared under different marketing schemes. Poultr. Sci., 70(6), 1339-1344. https://doi.org/10.3382/ps.0701339
ASAE Standards. (2007). S269.4: Cubes, pellets, and crumbles-definitions and methods for determining density, durability, and moisture content. 624-626. St. Joseph, MI: ASABE.
Bailer, W. (2006). Writing ImageJ plugins - A tutorial, Version 1.71. Retrieved from https://imagej.nih.gov/ij/download/docs/tutorial171.pdf
Behnke, K. C. (2001). Processing factors influencing pelet quality. Proc. 2001 Anim. Feed Manuf. Assoc. Forum. Retrieved from https://www.feedmachinery.com/?????articles/feed_technology/the-art-science-of-pelleting/
Behnke, K. C. (2009). The art (science) of pelleting. Retrieved from https://www.feedmachinery.com/articles/feed_technology/the-art-science-of-pelleting/
Boney, J. (2019). Feeding pellets for broiler performance improvements. Extension Bulletin. State College, PA: Penn State Extension.
Burger, W., & Burge, M. J. (2008). Digital image processing: An algorithmic introduction using Java. Springer. https://doi.org/10.1007/978-1-84628-968-2
Cavalcanti, W. B., & Behnke, K. C. (2005). Effect of composition of feed model systems on pellet quality: A mixture experimental approach. I. Cereal Chem., 82(4), 455-461. https://doi.org/10.1094/CC-82-0455
Colley, Z., Fasina, O. O., Bransby, D., & Lee, Y. Y. (2006). Moisture effect on the physical characteristics of switchgrass pellets. Trans. ASABE, 49(6), 1845-1851. https://doi.org/10.13031/2013.22271
Decision Innovation Solutions. (2020). Animal feed/food consumption and COVID-19 impact analysis. Report prepared for the Institute of Feed Education and Research. Retrieved from https://ifeeder.org/wp-content/uploads/210301-FINAL-REPORT-IFEEDER-Animal-Feed-Food-Consumption-COVID-19.pdf
Fahrenholz, A. C. (2012). Evaluating factors affecting pellet durability and energy consumption in a pilot feed mill and comparing methods for evaluating pellet durability. Unpublished Ph.D. diss. Manhattan, KS: Kansas State University.
Fairfield, D. H. (2005). Pelleting system and operations. In E. K. Schofield (Ed.), Feed Manufacturing Technology V (pp. 142-156). Arlington, VA: American Feed Industry Association.
Gao, Q., Moechnig, B. W., & Crenshaw, J. D. (1999). Animal feed pelleting process and animal feed pellets produced therefrom: U.S. Patent No. 5,871,802. Washington, DC: U.S. Patent and Trademark Office.
Hott, J. M., Buchanan, N. P., Cutlip, S. E., & Moritz, J. S. (2008). The effect of moisture addition with a mold inhibitor on pellet quality, feed manufacture, and broiler performance. J. Appl. Poult. Res., 17(2), 262-271. https://doi.org/10.3382/japr.2007-00083
Igathinathane, C., Melin, S., Sokhansanj, S., Bi, X., Lim, C. J., Pordesimo, L. O., & Columbus, E. P. (2009a). Machine vision based particle size and size distribution determination of airborne dust particles of wood and bark pellets. Powder Technol., 196(2), 202-212. https://doi.org/10.1016/j.powtec.2009.07.024
Igathinathane, C., Pordesimo, L. O., & Batchelor, W. D. (2009b). Major orthogonal dimensions measurement of food grains by machine vision using ImageJ. Food Res. Int., 42(1), 76-84. https://doi.org/10.1016/j.foodres.2008.08.013
Igathinathane, C., Pordesimo, L. O., Columbus, E. P., Batchelor, W. D., & Methuku, S. R. (2008). Shape identification and particles size distribution from basic shape parameters using ImageJ. Comput. Electron. Agric., 63(2), 168-182. https://doi.org/10.1016/j.compag.2008.02.007
Igathinathane, C., Rugroden, P., & Pordesimo, L. O. (2011). Fish feed pellets orthogonal dimensions through computer profile search approach, Paper No. 1110624. Proc. 2011 ASABE Annual International Meeting.
Kaliyan, N., & Vance Morey, R. (2009). Factors affecting strength and durability of densified biomass products. Biomass Bioenergy, 33(3), 337-359. https://doi.org/10.1016/j.biombioe.2008.08.005
Kristiawan, M., Chaunier, L., Sandoval, A. J., & Della Valle, G. (2020). 7 - Extrusion: Cooking and expansion. In A. A. Perdon, S. L. Schonauer, & K. S. Poutanen (Eds.), Breakfast cereals and how they are made (3rd ed., pp. 141-167). Oxford: AACC International Press. https://doi.org/10.1016/B978-0-12-812043-9.00007-2
Larsen, T. B., Sokhansanj, S., Patil, R. T., & Crerar, W. J. (1996). Breakage susceptibility studies on alfalfa and animal feed pellets. Can. Agric. Eng., 38(1), 21-24.
Lawrence, J. D., Mintert, J. R., Anderson, J. D., & Anderson, D. P. (2008). Feed grains and livestock: Impacts on meat supplies and prices. Choices, 23(2), 11-15.
MacBain, R. (1966). Pelleting animal feed. Arlington, VA: American Feed Manufacturers Association.
Moritz, J. S., Wilson, K. J., Cramer, K. R., Beyer, R. S., McKinney, L. J., Cavalcanti, W. B., & Mo, X. (2002). Effect of formulation density, moisture, and surfactant on feed manufacturing, pellet quality, and broiler performance. J. Appl. Poult. Res., 11(2), 155-163. https://doi.org/10.1093/japr/11.2.155
Patil, R. T., Berrios, J. D., Tang, J., & Swanson, B. G. (2007). Evaluation of methods for expansion properties of legume extrudates. Appl. Eng. Agric., 23(6), 777-783. https://doi.org/10.13031/2013.24044
Rasband, W. S. (2008). ImageJ. Bethesda, MD: National Institutes of Health, U. S. Dept. of Health and Human Services. Retrieved from http://rsb.info.nih.gov/ij/index.html
Shahin, M. A., & Symons, S. J. (2005). Seed sizing from images of non-singulated grain samples. Can. Biosyst. Eng., 47(3), 49-55.
Skoch, E. R., Binder, S. F., Deyoe, C. W., Allee, G. L., & Behnke, K. C. (1983). Effects of pelleting conditions on performance of pigs fed a corn-soybean meal diet. J. Anim. Sci., 57(4), 922-928. https://doi.org/10.2527/jas1983.574922x
Sokhansanj, S., & Crerar, W. (1999). Development of a durability tester for pelleted and cubed animal feed, SAE 1999-01-2830, Agricultural Machinery, Tires, Tracks, and Traction-SP-1474.
Stevens, C. A. (1987). Starch gelatinisation and the influence of particle size, steam pressure and die speed on the pelleting process. Unpublished Ph.D. diss. Manhattan, KS: Kansas State University.
Summers, J. D., Bentley, H. U., & Slinger, S. J. (1968). Influence of method of pelleting on utilization of energy from corn, wheat shorts, and bran. Cereal Chem., 45, 612-615.
Tabil Jr, L., & Sokhansanj, S. (1996). Process conditions affecting the physical quality of alfalfa pellets. Appl. Eng. Agric., 12(3), 345-350. https://doi.org/10.13031/2013.25658
Thomas, M., & van der Poel, A. F. (1996). Physical quality of pelleted animal feed 1. Criteria for pellet quality. Anim. Feed Sci. Technol., 61(1-4), 89-112. https://doi.org/10.1016/0377-8401(96)00949-2
Winowiski, T. (1985). Optimizing pelleting temperature. Feed Manage., 36(7), 28-33.
Winowiski, T. (1998). Examining a new concept in measuring pellet quality: Which test is best? Feed Manage., 49(1), 23-26.
Wood, J. F. (1987). The functional properties of feed raw materials and their effect on the production and quality of feed pellets. Anim. Feed Sci. Technol., 18(1), 1-17. https://doi.org/10.1016/0377-8401(87)90025-3