International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 1117
ISSN 2229-5518
Artificial Neural Networking Model an Approach for the Coagulation Properties of Milk
Vesna K. Hristova1, M. Ayaz Ahmad2, Biljana Trajkovska3, Stefce Presilski4, Georgi Bonev5
Abstract — The analysis and sequence of some technological parameters and milk coagulation properties (MCP) of Holstein Friesian dairy cows have been studied in the present research article. The milk samples have been collected from a local farm’s at Pelagonia region, Republic of Macedonia and the experimental work was conducted in the laboratories of the Faculty of biotechnical sciences, R. Macedonia and Tabuk University, KSA. The study illustrates the MCP of cow’s milk as well as the effect of milk urea nitrogen level and pH on the coagulum development. The scanning electron microscopy (SEM) of raw milk samples and after rennet addition also has been studied. All the said results/parameters have been compared with the soft computing approach so called artificial neural networking model, and the predictions of soft computing (ANN) model’s outcomes were found in a good agreement with the experimental data.
Index Terms —milk coagulation properties (MCP), milk urea nitrogen (MUN), artificial neural networking (ANN) model.
—————————— ——————————
Milk is the normal mammary secretion of milking animals obtained from one or more milkings without either addition to it or extraction from it, intended for consumption as liquid milk or for further processing (Codex Alimentarius, 2011) [1]. The composition of milk varies by species, but it always con- tains significant amounts of proteins (approx. 3.3%), carbohy- drate (approx. 4.6%) and fat (approx. 4.3%), as well as a great source of calcium and other components, organic acids, pep- tides, and vitamins (Heck et al., 2009, Hettinga, 2009) [2]. The milk delivered to dairies is converted into a number of fresh products and manufactured dairy products. Some 68.2 million tonnes of raw milk were used to produce 9.3 million tonnes of cheese in the EU-28 in 2013, while 31.5 million tonnes of raw milk were turned into a similar amount of drinking milk. (EUROSTAT, 2015) [3].
The process of milk coagulation and development of rennet-induced gel (coagulum) is the most significant and the most sensitive process in the production of the rennet curd cheese varieties(Tofanin et al., 2012) [4]. In terms to rennet co- agulation the main active component is an enzyme (chymosin) that hydrolyzes the k-caseins which largely contribute to the colloidal stability of casein micelles [5].
According to Lucey J. A., 2002 in reference [5], the percent- age of k-caseins proteolysis is about 90% of the unheated milk. Briefly, the particular hydrolysis of κ-casein by chymosin (rennet) leads to a dynamic decrease in the extent of the repul- sive forces among the casein micelles and formation of strands composed of aggregated casein micelles that further associate to build up a three-dimensional continuous network (Tuinier
————————————————
1. Faculty of Biotechnical Sciences, University of St. Clement Ohridski, Partizanska, bb 7000 Bitola, Macedonia. E-mail: v.primavera79@gmail.com
2. Physics Department, Faculty of Science, P.O. Box 741,
University of Tabuk - 71491, Saudi Arabia. E-mail: mayaz.alig@gmail.com
3. Faculty of Biotechnical Sciences, University of St. Clement Ohridski,
Partizanska, bb 7000 Bitola, Macedonia. E-mail: trajkovskab55@gmail.com
4. Faculty of Biotechnical Sciences, University of St. Clement Ohridski,
Partizanska, bb 7000 Bitola, Macedonia. E-mail: presilskistevo@yahoo.com
5. Faculty of Agriculture, Student Campus, Trakia University, Stara Zagora,
Bulgaria. E-mail: gbonev@uni-sz.bg
and de Kruif 2002; Lucey et al. 2003). [6, 7]
Due to the advancement of technology, so called modern
age of technology, the theoretical facts and experiment results
were found completely closer to each other. Predictive soft
computing methodologies have been used for prediction and
controlling many dairy production processes “from farm to
dairy industrial scale” providing significant advantages in
increased manufacturing efficiency and quality control. There-
fore the new technique named artificial neural network (ANN)
model would be a very powerful tool to determine some ex-
perimental observables and/or facts perfectly in the field of
dairy science. Recently, the artificial neural network (ANN)
model / technique are frequently applied for identification and
in checking of the complex interactions of raw materials, man-
ufacturing settings and attributes of dairy products such as
chemical composition, sensory traits, functionality and shelf
life [8 – 10].
By using some predictions of such mathematical models,
i.e. ANN model, one can decrease the cost of dairy products as
well as can be increase the production rate for mankind [11,
12].
From a research point of view the utilization of ANN sys-
tems to relate target responses back to input variables settings
is providing exceptionally efficient approaches to study the
complexities of interactions in dairy products [11, 12].
The present research work has been carried out to obtain a
better insight into the problem of rennet coagulation of
Holstein-Friesian cow’s milk in a local farm’s in Pelagonia
region, Republic of Macedonia. The study illustrates the MCP
of cow’s milk as well as the effect of milk urea nitrogen level
and pH on the total coagulation time. The scanning electron
microscopy (SEM) of raw milk samples and after rennet
addition also has been studied. All the said results/ factors
have been compared with the soft computing approach so
called artificial neural networking model, and the predictions
of soft computing (ANN) model’s outcomes were found in a
good agreement with the experimental data [13, 14].
IJSER © 2015 http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 1118
ISSN 2229-5518
Holstein-Friesian breed milk samples were taken at one
dairy farm in the Bitola district in the Republic of Macedonia
which is situated at an altitude of 1,233 metres above the
Adriatic sea level. The dairy cows on the farm were fed ad
libitum throughout the year as a total mixed ration, supple-
mented with concentrate according to standard practice and
the cows were never turned out to graze. Rations before and
after calving were formulated to exceed National Research
Council recommendations (NRC, 2001) [15], and the residues
um sensitivity was evaluated by measuring the coagulation time after addition of an appropriate amount of calcium chlo- ride.
Coagulation time identified as time from rennet addition to
the formation of the first visible floccules was measured visu- ally. Actually the coagulation time was defined as the time required for the first appearance of graininess in the moving film of milk samples on the surface of the glass walls of the beaker [16 -18]. A quantity of 25 ml by volume of each sample was measured into a beaker of 125ml and placed it in a water
°
of the dietary feed were generally observed in the herd. All
bath at 35
C up to 45 minutes. After that we added an appro-
cows received the same lactation diet for ad libitum intake throughout the experimental period post calving. The milk samples were collected from the morning milking of the dairy cows (6.00 - 7.00 hours). In accordance with the rules for milk sampling, the milk samples were manually taken from the individual collector of the milking De Laval system in with a special sterile plastic cups (50ml) [16]. Samples were trans- ported to the laboratory by movable refrigerator and kept in at the same temperature < 10 °C during the determination of milk quality parameters. The analysis of MCP was carried out
priate volume of rennet solution to it and then we measured
the coagulation time with the help of visual method.
Scanning electron microscopy (SEM) was used successfully
by many investigators to reveal the microstructures of curd as
well as to observe the modification in microstructure of the coagulated milk.
For SEM analysis, freeze dried, reconstituted samples of cow milk have been used. After adjusting the pH values 6.6 -
°
following M. Mele, R. Dal Zotto et al. (2009); [17] briefly, milk
6.8 the samples were equilibrated at 35
C and coagulation
samples (10 mL) were heated to 35 °C and 200 μL of rennet was added.
Examination and coagulation of milk were performed with-
in 3 hours after milking.Milk from the selected experimental
monitored by visual observation. The samples were taken be-
fore and after adding the rennet at time intervals up to the
visually observed coagulation time. The freeze-fracturing
technique has been applied to the milk samples for SEM anal-
ysis without adding any cryoprotectives. The freeze-fracturing
was carried out in a modified Batzers BAF 300 unit, in an ob-
°
cows was transported in containers from inox-steel, pre-
ject’s temperature up to -150
C. One can much more details
chilled at a temperature up to 80 C at the “Laboratory of dairy chemistry and technology”, Faculty of biotechnical sciences, Bitola. In the double bottom stabilizer, 5 kg of milk from each selected cow was heated to the renneting temperature of 35-37
0C and was added rennet powder, according to the rules of the company-manufacturer. The mixture was manually mixed well and the initial clotting time T (min), the time period start- ing from the addition of rennet to the first appearance of clots of milk solution, was recorded. Total clotting time T (min) was also measured and for normal coagulation was reported for extended testing period of 45 -60 minutes [18].
The rennet was commercial powder (CHY–MAX POWDER
EXTRA NB, CHR. HANSEN, the strength of the enzyme being
2235 IMCU/g) Rennet solution of 0.4% was prepared and an
appropriate amount of this solution was taken to give a visual-
ly observed coagulation time of approximately 10 minutes in
milk. For measurement of the milk coagulation time the fol-
lowing two steps were used.
The milk samples were adjusted to pH 6.6 -6.8 by slow ad-
dition of 1 M HCI, placed in a water bath, and the coagulation
time was measured at temperatures over the range at 30 °C to
40 °C.
All samples were equilibrated at 35 °C and the coagulation
time determined at pH between 6.20 and 7.00. And the Calci-
related to such experimental in reference (Muller et. al., 1982)
[13 - 14]. The scanning electron microscope (SEM) images have
been taken out of Japan made, JSM 6390A (JEOL Japan) at a
dissimilar magnification of the above prepared samples at
Physics Department, Faculty of Science, University of Tabuk,
Saudi Arabia. Before SEM examine, the prepared samples
were layered with gold in a vacuum coating unit. The cross
section areas ranging of samples were approximately 1 cm to 5
microns in width and the magnification ranging of SEM was
the order of 20X to approximately 30,000X, with spatial resolu-
tion of 50 to 100 nm [14].
Amongst all 148 cows, only 25 selected cows were followed and re-sampled on several occasions to evaluate possible changes in coagulation properties according to the milk urea nitrogen level. On the basis of milk urea nitrogen (MUN) level, the cows/samples were divided in two groups, group 1st (milk urea nitrogen level < 6.5 mmol/L) (10 cows) i.e. normal milk urea nitrogen level (MUN) and group 2nd (milk urea nitrogen level > 6.5 mmol/L) i.e. high milk urea nitrogen level (15 cows). The observed MUN levels, pH values and total coagulation times of cow’s milk have been depicted in Table 1. In the same table the predicted values of the total coagulation times by soft computing / simulations named artificial neural networking model also have been shown.
The lowest MUN level was measured 3.60, the pH value was 6.58 ± 0.01 and the coagulation time was 45 minutes for
IJSER © 2015 http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 1119
ISSN 2229-5518
the (10) cows with normal MUN levels, whereas the highest values for MUN levels was 5.40, the pH was 6.84 ± 0.15 and the coagulation time was 55 minutes for the same group of cows (1st group).
TABLE 1
DIRECT MEASURED COALGULATION TIME ALONG WITH SIMULATED
RESULTS OF ANN MODEL
(normal and high). In all the samples the visually measured total coagulation time for both the milk samples were found approximate twice in the comparison of prediction of ANN model. In both the cases / groups of cow’s milk, the predicted total coagulation times, found arbitrary in order of time value with standard errors and also its values found to be approxi- mated within the order of 28.5 ± 6 minutes. These results sug- gested that the milk composition was influenced by the die- tary treatment and some other genetic and paragenetic factors.
To acquire a better understanding on the extent of the ag- gregation of the casein micelles during renneting, the structur- al state of the raw milk proteins and casein curds after adding rennet were investigated using SEM technique (Figure 1. a-d). Practically, the prime initiative was to find out the structural changes and protein degradation after coagulation process, at both milk samples with normal and with high milk urea ni- trogen level.
In the present SEM analyses, the main goal was to find a visual illustration of the coagulation processes and to demon- strate the structural changes in cow’s milk.
(a)
Fig. 1 (a). A Scanning Electron Microscopic picture of normal MUN levels milk sample before the addition of rennet.
(b)
Further, for the 2nd group of (15) cows with high MUN lev- els, the lowest MUN level was measured 7.15, the pH value was 6.56 ± 0.09 and the coagulation time was 48 minutes, whereas the highest values for MUN levels was 12.75, the pH value was 6.81 ± 0.13 and the coagulation time was 60 minutes for the same group of cows (2nd group).
Therefore, one can conclude from Table 1, that the studied parameters, i.e. MUN levels and the total coagulation time were found little bit in higher order for the 2nd group with (MUN) level > 6.50 m-mol/L rather than the 1st group with (MUN) level < 6.50 m-mol/L. Such obtained results were also found in better agreement with the other worker in the field of dairy science [8 – 17 and references therein].
The total coagulation time was also predicted by soft com- puting of ANN model and it is depicted in the same Table 1 for each individual cow’s in both the cases of MUN levels
IJSER © 2015 http://www.ijser.org
Fig. 1(b). A Scanning Electron Microscopic picture of normal MUN levels milk sample after the addition of rennet.
(c)
International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 1120
ISSN 2229-5518
The total coagulation time measured in present milk sam- ples were found twice to the simulated results, which suggest- ed that the milk composition was influenced by the dietary treatment and some other factors which would be our future field of interest.
In the present research work, we did not find a significant effect of temperature and the pH values over the total coagula- tion time and also coagulum properties, but it is exited by some other authors.
Based on the present experimental work, one can say that such types of study are worth mentioning for the advance- ment of recent dairy technology. The produced results show a lot of good agreement with the results obtained by various authors.
(d)
Fig. 1(d). A Scanning Electron Microscopic picture of high MUN level milk sample after the addition of rennet.
Scanning electron micrographs obtained have been depict- ed in Fig.1 (a) and Fig. 1(b), for the milks sample of group 1st cows (10 cows), before and after the addition of rennet respec- tively. And the same Fig.1 (c) and Fig. 1 (d) were scanned for the milk of high urea nitrogen level > 6.5 mmol/L) i.e. 2nd group of cows (15 cows).
From all the figures it is clear that the cow’s milk appeared as almost special shape particles, and composed of numbers of submicelles. The visually obtained coagulation times of cow’s milk were 5 minutes. In the samples the onsets of aggregation were also pragmatic in SEM approximate of 70% of the coagu- lation time.
Based on our present experimental work entitled “artificial neural networking model an approach for the coagulation properties of milk” of the Holstein-Friesian Cows’ Milk in Re- public of Macedonia”, we can draw the following valuable conclusions:
The influence of the technological parameters of the total coagulation time in milk coagulation process as milk urea ni- trogen (MUN) level and pH values has been concluded and found it in a good agreement with the other worker in the same research field.
The authors are very much gratefully acknowledge to farm workers and also farm authority to use the valuable facilities from the dairy farm, in Bitola region, R. Macedonia. And also the corresponding author Vesna Karapetkovska Hristova is highly thankful to the faculty members at the Physics Depart- ment, Faculty of Science, University of Tabuk, Saudi Arabia for providing her the most valuable opportunity to make part of the experimental work in the faculty labs.
[1] Joint FAO/WHO Codex Alimentarius Commission, “Codex alimen- tarius. Rome: Food and Agriculture Organization of the United Na- tions.Milk and milk products”. Second Edition, (2011).
[2] Kasper Arthur Hettinga, “Quality control of raw cows' milk by head- space analysis, a new approach to mastitis diagnosis PhD thesis, Wa- geningen University, (2009). The Netherlands, 128 pp. ISBN: 978-90-
9585-300-8
[3] Statistical Office of the European Communities. “EUROSTAT: Agri- culture, forestry and fishery statistics 2014 edition”, Luxembourg: Eurostat. ISSN 2363-2488, (2015).
[4] V. Toffanin, M. De Marchi., M. Penas., D. Pretto, M. Cassandro “Characterization of milk coagulation ability in bulk milk samples”, Acta agriculturae Slovenica, Vol. 100, pp. 93-98 (2012).
[5] J. A. Lucey, “Formation and physical properties of milk protein gels”, Journal of Dairy Science, 85, pp. 281–294 (2002).
[6] Lucey J. A., Johnson M. E., Horne D. S., “Invited review: perspectives on the basis of the rheology and texture properties of cheese”, Jour- nal of Dairy Sci 86, pp. 2725–2743 (2003).
[7] Tuinier R., de Kruif C. G., “Stability of casein micelles in milk”, Jour-
nal of Chem. Phys., 117, pp. 1290–1295 (2002).
[8] Vesna Karapetkovska Hristova, M. Ayaz Ahmad, Julijana Tomovska, Biljana Bogdanova Popov, “Study of coagulation properties of Holstein cow’s milk depending on the level of milk urea nitrogen in Macedonia dairy farms”, Inter. J. of Enhanced Res. Tech. and Engg. (IJERSTE), vol. 3 (3), pp. 522-529 (2014).
[9] Vesna Karapetkovska Hristova, Julijana Tomovska, Georgi Bonev, Stanimir Dimitrov, M. Ayaz Ahmad, “Study of Seasonal Dynamics of Blood Metabolic Profile and Milk Urea Nitrogen (MUN) Level of Cows with Reproductive Disorders” Asian Academic Research Journal of Multidisciplinary (AARJMD), vol 1 (22), pp. 246-260 (2014).
IJSER © 2015 http://www.ijser.org
International Journal of Scientific & Engineering Research, Volume 6, Issue 4, April-2015 1121
ISSN 2229-5518
[10] Vesna Karapetkovska Hristova, Ph.D. Submitted at Faculty of
Agriculture, Trakia Uniersity, Stara Zagora Bulgaria, 2014.
[11] M. Ayaz Ahmad, Mir Hashim Rasool, N. Ameer Ahmad, Shafiq Ah- mad, “Neural networking (N-N) model in relativistic heavy ion colli- sions” Proceedings International Nuclear Physics Symposium (2013), Vol. 58, pp. 762-763 (2013).
[12] Vesna K. Hristova, M. Ayaz Ahmad, Julijana Tomovska, Biljana
Trajkovska, Georgi Bonev, “Assessment of raw milk quality by neural networking (N-N) model in Macedonian dairy farms”, Int. J. of Ethics in Engn. & Mang. Educ. (IJEEE), Vol. 1 (10), pp.58-61, (2014).
[13] Egerton, R. F., “Physical principles of electron microscopy : an intro- duction to TEM, SEM, and AEM”, Springer Book, 202, 2005.
[14] Alveera Khan, M. Ayaz Ahmad, Shirish Joshi, ”A systematic study
for electrical properties of chemically treated coir fiber reinforced epoxy composites with ANN model”, Int. J. of Sci. and Research (IJSR), Vol. 4(1), pp. 410-414 (2014).
[15] National Research Council recommendations (NRC, 2001); ex- plore.noaa.gov/sites/OER/.../national-research-council-voyage.pdf
[16] J. M. L. Heck, H. J. F. Van Valenberg, J. Dijkstra and A. C. M. Van
Hooijdonk, “Seasonal variation in the Dutch bovine raw milk
composition”, Journal of Dairy Science, Vol. 92 (10), pp. 4745-4755, (2009).
[17] De Marchi M., C. C. Fagan, C.P. O’Donnell, A. Cecchinato, R. Dal Zotto, M. Cassandro, M. Penasa & G. Bittante, “Prediction of coagu- lation properties, titratable acidity, and pH of bovine milk using mid- infrared spectroscopy”, Journal of Dairy Sci., 92, pp. 423-432 (2009).
[18] Vesna K. Hristova, “Study of chemical composition and coagulation
properties of Holstein cow’s milk in R. Macedonia”, Asian Academic
Research Journal of Multidisciplinary (AARJMD), vol 1 (21), pp. 222-
238, (2014).
IJSER © 2015 http://www.ijser.org