Modeling of pH and acidity for industrial cheese production

A three-layer feedforward neural network was successfully used to model and predict the pH of cheese curd at various stages during the cheese-making process. An extended database, containing more than 1800 vats over 3 yr of production of Cheddar cheese with eight different starters, from a large cheese plant was used for model development and parameter estimation. Neural network models were developed with inputs selected among 33 quantitative and qualitative process variables for final pH of cheese, pH at cutting, and acidity at whey drawing-off and at pressing. In all cases, very high correlation coefficients, ranging from 0.853 to 0.926, were obtained with the validation data. A sensitivity analysis of neural network models allowed the relative importance of each input process variable to be identified. The sensitivity analysis in conjunction with a priori knowledge permitted a significant reduction in the size of the model input vector. A neural network model using only nine input process variables was able to predict the final pH of cheese with the same accuracy as for the complete model with 33 original input variables. This significant decrease in the size of neural networks is important for applications of process control in cheese manufacturing.     

Improving rennet coagulation and cheesemaking properties of reverse osmosis skim milk concentrates by pH adjustment

The use of reverse osmosis (RO) for cheese milk concentration has advantages including obtaining reusable low pollutant permeates and reducing milk transportation costs. However, high levels of lactose and salts in RO concentrates impair their cheesemaking abilities. The objective of this work was to optimise the use of RO concentrates (5–11% protein content) for cheesemaking by pH adjustment. Rennet coagulation kinetics, salt partitioning and cheesemaking properties were studied in comparison with ultrafiltration concentrates. Results showed that concentration by RO induced an increase regarding the coagulation time and the gel maximal firming rate that reached a plateau at 9% protein content. Increases in calcium mineralisation of casein micelles as well as in yield, moisture and lactose content in model cheese were observed. Lowering renneting pH was found to improve the cheesemaking properties of RO concentrates by promoting partial demineralisation of casein micelles, accelerating coagulation kinetics and increasing curd drainage.     

A method for the inline measurement of milk gel firmness using an optical sensor

At present, selection of cutting time during cheesemaking is made based on subjective methods, which has effects on product homogeneity and has prevented complete automation of cheesemaking. In this work, a new method for inline monitoring of curd firmness is presented. The method consisted of developing a model that correlates the backscatter ratio of near infrared light during milk coagulation with the rheological storage modulus. The model was developed through a factorial design with 2 factors: protein concentration (3.4 and 5.1%) and coagulation temperature (30 and 40°C). Each treatment was replicated 3 times; the model was calibrated with the first replicate and validated using the remaining 2 replicates. The coagulation process was simultaneously monitored using an optical sensor and small-amplitude oscillatory rheology. The model was calibrated and successfully validated at the different protein concentrations and coagulation temperatures studied, predicting the evolution of storage modulus during milk coagulation with coefficient of determination values >0.998 and standard error of prediction values <3.4 Pa. The results demonstrated that the proposed method allows inline monitoring of curd firming in cheesemaking and cutting the curd at a proper firmness to each type of cheese.     

Response surface methodology modeling of protein concentration,coagulum cut size,and set temperature on curd moisture loss kinetics during curd stirring

The effects of the independent variables protein concentration (4–6%), coagulum cut size (6–18 mm³), and coagulation temperature (28–36°C) on curd moisture loss during in-vat stirring were investigated using response surface methodology. Milk (14 kg) in a cheese vat was rennet coagulated, cut, and stirred as per semihard cheesemaking conditions. During stirring, the moisture content of curd samples was determined every 10 min between 5 and 115 min after cutting. The moisture loss kinetics of curds cut to 6 mm³ followed a logarithmic trend, but the moisture loss of curds from larger cut sizes, 12 or 18 mm³, showed a linear trend. Response surface modeling showed that curd moisture level was positively correlated with cut size and negatively correlated with milk protein level. However, coagulation temperature had a significant negative effect on curd moisture up to 45 min of stirring but not after 55 min (i.e., after cooking). It was shown that curds set at the lower temperature had a slower syneresis rate during the initial stirring compared with curds set at a higher temperature, which could be accelerated by reducing the cut size. This study shows that keeping a fixed cut size at increasing protein concentration decreased the level of curd moisture at a given time during stirring. Therefore, to obtain a uniform curd moisture content at a given stirring time at increasing protein levels, an increased coagulum cut size is required. It was also clear that breakage of the larger curd particles during initial stirring can also significantly influence the curd moisture loss kinetics. Both transmission and scanning electron micrographs of cooked curds (i.e., after 45 min of stirring) showed that the casein micelles were fused at a higher degree in curds coagulated at 36°C compared with 28°C, which confirmed that coagulation temperature causes a marked change in curd microstructure during the earlier stages of stirring. The present study showed the dynamics of curd moisture content during stirring when using protein-concentrated milk at various set temperatures and cut sizes. This provides the basis for achieving a desired curd moisture loss during cheese manufacture using protein-concentrated milk as a means of reducing the effect of seasonal variation in milk for cheesemaking.     

Effect of goat milk composition on cheesemaking traits and daily cheese production

Cheese yield is strongly influenced by the composition of milk, especially fat and protein contents, and by the efficiency of the recovery of each milk component in the curd. The real effect of milk composition on cheesemaking ability of goat milk is still unknown. The aims of this study were to quantify the effects of milk composition; namely, fat, protein, and casein contents, on milk nutrient recovery in the curd, cheese yield, and average daily yield. Individual milk samples were collected from 560 goats of 6 different breeds. Each sample was analyzed in duplicate using the 9-laboratory milk cheesemaking assessment, a laboratory method that mimicked cheesemaking procedures, with milk heating, rennet addition, coagulation, curd cutting, and draining. Data were submitted to statistical analysis; results showed that the increase of milk fat content was associated with a large improvement of cheese yield because of the higher recovery of all milk nutrients in the curd, and thus a higher individual daily cheese yield. The increase of milk protein content affected the recovery of fat, total solids, and energy in the curd. Casein number, calculated as casein-to-protein ratio, did not affect protein recovery but strongly influenced the recovery of fat, showing a curvilinear pattern and the most favorable data for the intermediate values of casein number. In conclusion, increased fat and protein contents in the milk had an effect on cheese yield not only for the greater quantity of nutrients available but also for the improved efficiency of the recovery in the curd of all nutrients. These results are useful to improve knowledge on cheesemaking processes in the caprine dairy industry.     

Validation of a curd-syneresis sensor over a range of milk composition and process parameters

An online visible-near-infrared sensor was used to monitor the course of syneresis during cheesemaking with the purpose of validating syneresis indices obtained using partial least squares, with cross-validation across a range of milk fat levels, gel firmness levels at cutting, curd cutting programs, stirring speeds, milk protein levels, and fat:protein ratio levels. Three series of trials were carried out in an 11-L cheese vat using recombined whole milk. Three factorial experimental designs were used, consisting of 1) 3 curd stirring speeds and 3 cutting programs; 2) 3 milk fat levels and 3 gel firmness levels at cutting; and 3) 2 milk protein levels and 3 fat:protein ratio levels, respectively. Milk was clotted under constant conditions in all experiments and the gel was cut according to the respective experimental design. Prediction models for production of whey and whey fat losses were developed in 2 of the experiments and validated in the other experiment. The best models gave standard error of prediction values of 6.6 g/100 g for yield of whey and 0.05 g/100 g for fat in whey, as compared with 4.4 and 0.013 g/100 g, respectively, for the calibration data sets. Robust models developed for predicting yield of whey and whey fat losses using a validation method have potential application in the cheese industry.     

Cheese yield,cheesemaking efficiency,and daily production of 6 breeds of goats

Little is known about the complex process of cheesemaking at the individual level of dairy goats because of the difficulties of producing a high number of model cheeses. The objectives of this work were (1) to study the cheesemaking ability of goat milk; (2) to investigate the variability of cheesemaking-related traits among different farms; (3) to assess the effects of stage of lactation and parity; and (4) to compare 6 breeds of goat (Saanen and Camosciata delle Alpi for the Alpine type; Murciano-Granadina, Maltese, Sarda and Sarda Primitiva for the Mediterranean type) for their cheesemaking ability. For each goat (n = 560) we studied (1) 8 milk quality traits (fat, protein, total solids, casein, lactose, pH, somatic cell score, and bacterial count); (2) 4 milk nutrient recovery traits (fat, protein, total solids, and energy) in curd; (3) 3 actual cheese yield traits (fresh cheese, cheese solids, and cheese water); (4) 2 theoretical cheese yield values (fresh cheese and cheese solids) and the related cheesemaking efficiencies; and (5) daily milk yield and 3 daily cheese yield traits (fresh cheese, cheese solids, and water retained in the curd). With respect to individual animal factors, farm was not particularly important for recovery traits or actual and theoretical cheese yield and estimates of efficiency, whereas it highly influenced daily productions. Parity of goats influenced daily cheese production, whereas DIM slightly affected recovery as well as percent and daily cheese yield traits. Breed was the most important source of variation for almost all cheesemaking traits. Compared with those of Alpine type, the 4 Mediterranean breeds had, on average, lower daily milk and cheese productions, greater actual and theoretical cheese yield, and higher recovery of nutrients in the curd. Among Alpine type, Camosciata delle Alpi was characterized by greater nutrients recovery than Saanen. Within the 4 Mediterranean types, the 3 Italians produced much less milk per day, with much more fat and protein and greater recovery traits than the Murciano-Granadina, resulting in greater actual cheese yield. Within the Italian breeds, milk from Sarda and Sarda Primitiva was characterized by lower daily yields, higher protein and fat content, and greater recoveries of nutrients than Maltese goats. These results confirmed the potential of goat milk for cheese production and could be useful to give new possibilities and direction in breeding programs.     

Transfer of protein from milk to cheese

This report concerns measurement of paracasein in milk and transfer of protein from milk to cheese. In the main experiment, two vats of Cheddar cheese were made from each of 11 lots of milk from one large herd over a period of 7 mo. Exclusion of solutes from moisture in paracasein micelles in milk and cheese was central to estimation of paracasein and to the transfer of protein from milk to cheese and whey. Solute-exclusion by paracasein and its changes during cheesemaking could be visualized by considering paracasein micelles to be a very fine sponge. The sponge excludes solutes, especially the large solutes like whey proteins. The sponge shrinks during cheesemaking and expels solute-free liquid, thereby slightly diluting the whey surrounding the micelles inside the curd. Paracasein N in milk was calculated as the difference between total milk N and rennet whey N, the latter adjusted to its level in milk. Adjustment used appropriate solute-exclusion factors (h) of the protein fractions of whey and 1.08 for paracasein and associated salts. They were combined to give a factor Fpc, which adjusted the level of rennet whey N to its level in milk: 1.001 x (1 - 1.01 x FM/100 - Fpc x pc/100), where FM = fat in milk, pc = estimated paracasein, and 1.001 = dilution of milk by chymosin and CaCl2. The mean Fpc was 3.03. Differences in values were small among different procedures for calculating paracasein, but they are considered to be important because they represent biases, which, in turn, are important in analyses commercially. We conclude that solute exclusion by moisture in paracasein must have decreased during cheesemaking because the ratio of moisture to paracasein in the final cheese was 1.5, much less than the h of 2.6 for serum proteins by paracasein. Release of solute-excluding moisture from paracasein during cooking was likely the reason for lower total N in cheese whey than in the rennet whey in the paracasein analysis. Paracasein, estimated to be in cheese, curd fines, salted whey, and whey during cheddaring, was 98.21, 0.20, 0.25 and 0.19%, respectively, of the paracasein in milk for a total of 98.85% (SD of 22 vats = 0.46); the location of the missing paracasein is not known. On the other hand, recovery of milk N in cheese and wheys was 99.92% (SD = 0.37%). Nitrogen in paracasein and its hydrolysis products in cheese was estimated to be 98.51% of total cheese N. Proteose-peptone from milk appeared not to be included with the paracasein in appreciable amounts. Some was apparently included with denatured serum proteins during Rowland fractionation of whey, perhaps as a coprecipitate. Measured paracasein would include fat globule membrane proteins in milk containing fat, and denatured whey proteins in heated milks. It was concluded that the method of measurement and the associated calculations are integral parts of the definition and quantification of paracasein in milk.     

Galactose concentration in pizza cheese prepared by three different culture techniques

A study was conducted to determine the most suitable lactic culture combinations and the techniques for the preparation of low moisture part skim (LMPS) mozzarella cheese (pizza cheese) with a low level of galactose. The cheesemaking time tested for all 12 culture combinations was less than 3 h and hence was found suitable for pizza cheesemaking. The initial accumulated galactose concentration was lower in unstretched curd followed by no-brine curd and stretched curd in ascending order. During storage of the cheese for 7 days, the galactose percentage in the cheeses prepared using cultures comprising galactose-fermenting strains of Lactobacillus helveticus and Streptococcus thermophilus was 0.30 in stretched curd, 0.04 in unstretched curd and 0.03 in no-brine curd. The degree of reduction in the level of galactose in pizza cheese during storage was maximum with the no-brine curd technique followed by the unstretched and stretched curd techniques in that order. This study provides information regarding selection of starter culture combinations and techniques for pizza cheesemaking based on consumer preference for low or moderate browning mozzarella cheese as a result of its galactose concentration.     

Effect of homogenization, microfiltration and pH on curd firmness and syneresis of curd grains

The impact of the independent variables, homogenization pressure (p₁), concentration factor of microfiltration (i) and pH on curd firmness (CF) and syneresis of curd grains was studied. Texture analysis was used to characterize CF of the rennet-induced gels. The analysis of a two-level factorial design revealed that i, p₁, pH and the interaction of i and pH had the most important influence on CF. Cutting time was therefore individually determined for each milk system using small amplitude oscillatory rheometry for generating comparable conditions for the syneresis experiments. Syneresis of curd grains with a diameter of 11 mm was followed at 35 °C close to semi-hard cheesemaking conditions. The permeate release during microfiltration was taken into consideration, allowing an evaluation of syneresis of grains made from concentrated and unconcentrated milk. It was shown that with increasing milk concentration less curd treatment time was needed to reach a certain syneresis value. Hence, total processing time in cheesemaking is decreased. Analysis of variance revealed that syneresis was affected by the individual variables. Kinetic parameters were satisfactorily estimated through regression (R²>0.98) and it was shown that milk composition and concentration due to microfiltration markedly influenced the endpoint of syneresis, RWR_max. The experiments demonstrate that microfiltration and homogenization can be combined to reach CF and syneresis comparable to untreated milk used in conventional cheesemaking. This meets one claim of the cheese industry when implementing both technologies in the manufacture process, since consistency and quality of the ripened cheese are expected to be unchanged.     

Effects of milk composition, stir-out time, and pressing duration on curd moisture and yield

A study was undertaken to investigate the effects of milk composition (i.e., protein level and protein:fat ratio), stir-out time, and pressing duration on curd moisture and yield. Milks of varying protein levels and protein:fat ratios were renneted under normal commercial conditions in a pilot-scale cheese vat. During the syneresis phase of cheese making, curd was removed at differing times, and curd moisture and yield were monitored over a 22-h pressing period. Curd moisture after pressing decreased with longer stir-out time and pressing duration, and an interactive effect was observed of stir-out time and pressing duration on curd moisture and yield. Milk total solids were shown to affect curd moisture after pressing, which has implications for milk standardization; that is, it indicates a need to standardize on a milk solids basis as well as on a protein:fat basis. In this study, a decreased protein:fat ratio was associated with increased total solids in milk and resulted in decreased curd moisture and increased curd yield after pressing. The variation in total solids of the milk explains the apparent contradiction between decreased curd moisture and increased curd yield. This study points to a role for process analytic technology in minimizing variation in cheese characteristics through better control of cheesemilk composition, in-vat process monitoring (coagulation and syneresis), and post-vat moisture reduction (curd pressing). Increased control of curd composition at draining would facilitate increased control of the final cheese grade and quality.     

Short communication: Fourier-transform mid-infrared spectroscopy to predict coagulation and acidity traits of sheep bulk milk

Sheep milk is mainly transformed into cheese; thus, the dairy industry seeks more rapid and cost-effective methods of analysis to determine milk coagulation and acidity traits. This study aimed to assess the feasibility of Fourier-transform mid-infrared spectroscopy to determine milk coagulation and acidity traits of sheep bulk milk and to classify milk samples according to their renneting capacity. A total of 465 bulk milk samples collected in 140 single-breed flocks of Comisana (84 samples, 24 flocks) and Sarda (381 samples, 116 flocks) breeds located in Central Italy were analyzed for coagulation properties (rennet coagulation time, curd firming time, and curd firmness) and acidity traits (pH and titratable acidity) using standard laboratory procedures. Fourier-transform mid-infrared spectroscopy prediction models for these traits were built using partial least squares regression analysis and were externally validated by randomly dividing the full data set into a calibration set (75%) and a validation set (25%). The discriminant capacity of the rennet coagulation time prediction model was determined using partial least squares discriminant analysis. Prediction models were more accurate for acidity traits than for milk coagulation properties, and the ratio of prediction to deviation ranged from 1.01 (curd firmness) to 2.14 (pH). Moreover, the discriminant analysis led to an overall accuracy of 74 and 66% for the calibration and validation sets, respectively, with greater sensitivity for samples that coagulated between 10 and 20 min and greater specificity to detect early-coagulating (<10 min) and late-coagulating (20–30 min) samples. Results suggest that Fourier-transform mid-infrared spectroscopy has the potential to help the dairy sheep industry identify milk with better coagulation ability for cheese production and thus improve milk transformation efficiency. However, further research is needed before this information can be exploited at the industry level.     

Short communication: Prediction of milk coagulation and acidity traits in Mediterranean buffalo milk using Fourier-transform mid-infrared spectroscopy

Milk coagulation and acidity traits are important factors to inform the cheesemaking process. Those traits have been deeply studied in bovine milk, whereas scarce information is available for buffalo milk. However, the dairy industry is interested in a method to determine milk coagulation and acidity features quickly and in a cost-effective manner, which could be provided by Fourier-transform mid-infrared (FT-MIR) spectroscopy. The aim of this study was to evaluate the potential of FT-MIR to predict coagulation and acidity traits of Mediterranean buffalo milk. A total of 654 records from 36 herds located in central Italy with information on milk yield, somatic cell score, milk chemical composition, milk acidity [pH, titratable acidity (TA)], and milk coagulation properties (rennet coagulation time, curd firming time, and curd firmness) were available for statistical analysis. Reference measures of milk acidity and coagulation properties were matched with milk spectral information, and FT-MIR prediction models were built using partial least squares regression. The data set was divided into a calibration set (75%) and a validation set (25%). The capacity of FT-MIR spectroscopy to correctly classify milk samples based on their renneting ability was evaluated by a canonical discriminant analysis. Average values for milk coagulation traits were 13.32 min, 3.24 min, and 39.27 mm for rennet coagulation time, curd firming time, and curd firmness, respectively. Milk acidity traits averaged 6.66 (pH) and 7.22 Soxhlet-Henkel degrees/100 mL (TA). All milk coagulation and acidity traits, except for pH, had high variability (17 to 46%). Prediction models of coagulation traits were moderately to scarcely accurate, whereas the coefficients of determination of external validation were 0.76 and 0.66 for pH and TA, respectively. Canonical discriminant analysis indicated that information on milk coagulating ability is present in the MIR spectra, and the model correctly classified as noncoagulating the 91.57 and 67.86% of milk samples in the calibration and validation sets, respectively. In conclusion, our results can be relevant to the dairy industry to classify buffalo milk samples before processing.     

A preliminary study of the use of a Raman laser sensor to monitor coagulation and syneresis for the online control of cheesemaking

The objective of this article was to apply a novel laser Raman sensor for the control of on line cheesemaking, mainly focusing on coagulation but with an initial test of syneresis. By means of a novel Raman laser sensor, the spectrum of cheese curd was seen to be slightly higher than that of milk in the 800–1200 cm^?1 Raman shift range, with a sigmoid increase in intensity until a constant value was reached. As regards syneresis, the response of the Raman laser sensor and the moisture content both followed first‐order kinetics. These initial results confirm that the laser sensor is able to determine the changes that take place in the cheesemaking process, during the coagulation and whey drainage steps, and point to the potential usefulness of this sensor for the online control of cheesemaking.     

Diversity of traditional Caciocavallo cheeses produced in Italy

This paper presents the results of a survey carried out in 68 dairies in southern Italy on the manufacturing processes of traditional Italian Caciocavallo cheese varieties. Following a study of the relevant literature, the various cheesemaking processes were analysed and the implications of different cheesemaking procedures were explored. The manufacturing variations able to influence the organoleptic characteristics of Caciocavallo cheese were milk and rennet types, procedures for curd acidification and stretching, salting and ripening conditions, and smoking treatment. This survey is designed to guide producers and consumers alike with respect to the perceivable effects of manufacturing variants on cheese quality.     

Characterization of curd grain size and shape by 2-dimensional image analysis during the cheesemaking process in artisanal sheep dairies

The size and shape of curd grains are the most important parameters used by cheesemakers to decide when to end the cutting or stirring processes during cheesemaking. Thus, 2-dimensional image analysis was used to measure the characteristics of curd grains in commercial cheese productions carried out by artisanal sheep dairies. Dairies used different technical settings for cutting and stirring steps, causing differences in the size and shape of curd grains. A linear relationship between total revolutions used for cutting and stirring and curd particle size was established. However, particle size distributions after curd cutting and stirring were highly heterogeneous. Actual cheese yield was correlated with particle size and cutting revolutions, whereas curd grain shape and fat loss were associated with stirring conditions by a multivariate approach. Image analysis of the size and shape of curd grains gives useful information for determining characteristics related to cheese yield and quality and may contribute to improving and controlling the cheesemaking process in small artisanal dairies.     

Standardisation of milk for cheesemaking at factory level

The percentage of milk fat recovered as cheese varies between 85 and 93 per cent, depending on the system used, and this must be taken into account when the casein to fat ratio of milk for cheesemaking is selected. Seasonal variation in the composition of milk protein can have a significant influence on the potential cheese yield. Prolonged storage of milk may cause casein losses while heat precipitation can facilitate the incorporation of whey proteins in cheese curd. The economic consequences of seasonal variations in Ireland on the price of milk for cheesemaking are discussed. The economics of standardisation may be marginal, but it is a useful aid in achieving uniform cheese quality.     

Coagulation process in Manchega sheep milk from Spain: A path analysis approach

Characteristics of sheep milk are of great interest for the dairy industry, as almost the totality of production is intended for cheesemaking. However, the existing relationships between these variables are complex. This study assessed composition, hygienic quality, coagulation properties, and curd yield of 1,200 individual Manchega sheep milk samples. The aim was to compare the effect of composition and hygienic quality on coagulation and curdling, and to evaluate the relationship between curd yields and the coagulation process and the effect of other features by using path analysis methodologies. Outcomes proved path analysis to be a useful and effective tool to assess these relationships through direct and indirect paths within the same model. Results showed that the factors that had a direct influence on milk coagulation were lactose concentration, casein content, and initial pH of milk. Contrastingly, somatic cells did not seem to have any effect (direct or indirect) on the coagulation process. Factors that directly affected curd yield were fat content, lactose concentration, casein content, and curd moisture. However, technological parameters showed little effect over curd yield.     

Evaluation of on-line optical sensing techniques for monitoring curd moisture content and solids in whey during syneresis

This study focuses on the prediction ability of several optical sensing techniques, namely single wavelength (980 nm), broad spectrum and colour coordinates, for monitoring key syneresis indices during cheese manufacture. Three series of trials were undertaken in which milk gel was cut and stirred in an 11 L cheese vat. Three full factorial designs were employed with experimental variables consisting of: (i) three curd stirring speeds and three cutting programmes; (ii) three milk fat levels and three gel firmness levels at cutting; and (iii) two milk protein levels and three fat:protein ratio levels in the respective experiments. Models developed using the range of techniques investigated demonstrated that an on-line visible–NIR sensor was able to predict curd moisture content. However, the broad spectrum technique was the only one capable of predicting whey solids. The findings show that on-line sensing techniques can significantly improve the control of curd moisture content in cheese factories, across the range of experimental variables used in this study.