Research Projects in 2001-2002

Evaluation of Melting, Oiling-off and Browning Properties of Cheddar and Mozzarella Cheeses with Computer Vision

Hai-Hong Wang and Da-Wen Sun

Functional properties such as melting, browning, oiling-off, stretchability and shreddability are of crucial importance to cheese, especially when cheese is used for manufacturing consumer foods such as pizza. Many attempts have been made to improve these properties. However, one of the major obstacles is the lack of objective methods to evaluate these functional properties, as traditional methods are qualitative, subjective, time-consuming, laborious, or subjected to artefacts.

Computer vision using image analysis techniques has been intensively used to assure quality of a wide-range of products, including classification and quality evaluation of fruits, vegetables, cereals, consumer foods, and even cheese. It has shown great advantages in objective, rapid and automatic quality evaluation.

The objective of this study was to develop computer vision methods and apply them to evaluate the most important functional properties, i.e., melting, browning and oiling-off, of cheddar and mozzarella cheeses, which are two main varieties of cheese consumed individually and used in consumer foods. The influences of cooking conditions and sample dimensions on these properties were investigated.

The meltability of cheese was significantly influenced by cooking conditions and sample dimensions. Cheese melted rapidly at the initial stage of cooking and then melting slowed down. Maximum melting degree and melting rate occurred at certain temperature. Extended cooking and cooking at high temperature may cause reduction in melting. Cheddar cheese melted more rapidly and considerably than mozzarella cheese at all cooking temperatures from 70 to 200^oC. The meltability of cheeses increased linearly when the side lengths of square cheese slices increased from 25 mm to 70 mm. The meltability of cheeses also increased proportionally as the thickness of cheese slices was increased from 2 mm to 4 mm. The influences of the area/thickness of cheese slices on meltability were similar for both cheddar and mozzarella cheeses.

The influence of baking temperature and time on the browning property of cheese was also significant. However, the time and temperature dependencies of browning factor (BF) were different for cheddar and mozzarella cheeses. Cooking at 70-200^oC for 2-4 min, the BF of mozzarella cheese increased almost linearly with baking temperature. As for the cheddar cheese, after heating for 8-12 min, a linear relationship between the BF and baking temperature was observed from 70^oC to 130^oC. Further browning hardly occurred when cooking temperature increased from 130^oC to 160^oC. When heating at temperatures above 160^oC, the colour of cheddar cheese darkened dramatically due to scorching. The browning property of cheese was less affected by sample dimensions as was cheese meltability.

As for the oiling-off property of cheese, more free oil was formed with increasing temperature and time. However, while baking at higher temperatures or for prolonged time, the free oil formed tended to dry off. Oiling-off was intensified as cheese area and thickness increased. The trend was similar for the two cheeses. However, oiling-off was more serious for cheddar cheese than for mozzarella cheese, as free oil formed at lower temperature for cheddar cheese and larger amount of oil was separated from the cheese body.

The results of this research show that the computer vision methods developed provide a quantitative, objective, consistent and efficient way for evaluating the melting, browning and oiling-off properties of cheddar and mozzarella cheeses. The data available are beneficial to the manufacturers of both cheeses and consumer foods.

Sponsors: EU Food Research Programme

---

Extension of Vase Life of Cut Daffodil and Lily Flowers by Vacuum Precooling and Evaluation of Flower Quality with Computer Vision

Tadhg Brosnan and Da-Wen Sun

Cut flowers deteriorate rapidly at field temperatures due to their high rates of respiration and low tolerance to heat, hence in order to maintain their quality and freshness cut flowers are pre-cooled after harvest prior to cold storage or transport. Therefore the fundamental objective of this project was to assess vacuum cooling as a practical and viable precooling method for different cut flowers, as compared with cooling by forced air and cold room cooling. Vacuum cooling is used commercially for a variety of vegetables. It is achieved by the evaporation of moisture off the product surface. A secondary aim was the evaluation of cut flower life by the objective assessment technique of computer vision.

Initial trials focused on the vacuum cooling of cut daffodil flowers. In this research, freshly cut daffodil flowers were vacuum cooled to 2^oC and their cooling process monitored. The average cooling time for the flowers was 65 s with the longest cooling time being 100 s. Further studies examined the vase life of the vacuum cooled cut flowers with and without cold storage. The vase life was determined subjectively by various criteria. Results indicated that vacuum cooling gave an increase in vase life of 33% (from 3 to 4 days) in the case of cooled flowers stood in water at room temperature.The vase life of the cut flowers vacuum cooled and stored at 2^oC for 7 days was doubled, from 2 days for controls to 4 days for cooled flowers.

In subsequent research different treatments were assessed to determine if the inherent weight loss associated with vacuum cooling could be reduced. The techniques examined were vacuum cooling unwetted flowers, water sprayed flowers, and flowers adsorbing water through the stems for cut lilies flowers. Cold room cooled and uncooled flowers were used as controls. The cold room technique gave the slowest cooling and the greatest weight loss of the methods tested with mean values of 18.9 min and 5.3% (s.e. 0.11) respectively. The vacuum cooled unwetted flowers gave the highest weight loss among the vacuum cooling techniques examined with a mean value of 5.1% (s.e. 0.33). Initial cooling rates for all the vacuum cooling techniques were extremely quick with a temperature of 10ºC reached in less than 60 s. However for the unwetted vacuum cooled and covered absorption methods the final stem temperature reached was high at 5-7ºC. Vacuum cooling water sprayed flowers gave the best results with a cooling time of 62-87 s to reach the storage temperature of 2-3ºC. This method gave no weight loss unless the entire surface of the plant was not wetted. The water sprayed technique also yielded the best results as regards to vase life extension with a vase life of 11 d (s.e. 0.37) for room held flowers, and the cold stored samples having a vase life of 13.3 d (s.e. 0.26). This was an increase in vase life of 16% and 18% respectively over the controls. The results establish that water spraying cut lilies prior to vacuum cooling gives the most effective and beneficial results.

The effects of controlling the evacuation rate during the vacuum cooling of cut flowers were studied. The influence of this technique on mass loss, temperature reduction per mass loss (TRPML), vase life and cooling rate were determined. Four different evacuation rates were analysed. Results showed that the greatest mass loss of 5.4% was found for the quickest evacuation rate of 374 mbar/min in comparison to a mass loss of 3.7% for an evacuation rate of 8.5 mbar/min. The evacuation was seen to have no notable effect on the vase life and the final temperature achieved by the cut flowers for the cooling period examined. However the cooling rate was influenced by the evacuation rate used.

Further research investigated the use of computer vision for the objective assessment of cut flower vase life in comparison to the subjective criteria used in the above trials. Initial tests focused on the vase life of cut flowers that were not subjected to any cooling treatment. Results show that the computer vision was well able to map the colour, size and shape changes in the flowers examined over their life. There is a gradual decrease in grey value for each flower which indicates that it is fading. The subsequent cooling experimental results show that the colour assessment by the computer vision system and consumer panel gave good agreement. Misclassification by computer vision for colour assessment was 10%.

The extension of vacuum cooling to the floricultural industry is favorable to the cut flowers examined. Its benefits are that the cut flowers will have a longer vase life and hence provide the consumer with a better product. The most significant influences on the vase life of the flowers is seen especially for the stored situation. This is important because the majority of cut flowers are transported long distances to their markets in cold storage containers, and vacuum cooling would maintain their quality during storage and thus extend their subsequent vase life. The computer vision method provides a potential tool for evaluating the vase life of cut flowers.

---

Numerical and Experimental Investigation of Heat and Mass Transfer during Conventional Cooling and Vacuum Cooling Processes of Cooked Meat Joints

Lijun Wang and Da-Wen Sun

Cooling in the meat industry is usually accomplished by using conventional cooling methods such as slow air (SAC), air blast (ABC) and water immersion (WIC). Large cooked meat joints are inherently slow to be cooled by these cooling techniques due to their low thermal conductivities. Therefore, vacuum cooling (VC), which is different from the conventional cooling in the internal generation of cooling source due to water evaporation is used to rapidly chill large cooked meat joints in order to improve the quality of the meats

The conventional cooling and vacuum cooling processes of large cooked meat joints were mathematically modelled. Heat and mass transfer through the meats, and the different cooling systems were firstly described mathematically. The effects of meat shapes and sizes, physical properties of the meats and operating conditions during cooling were included in the modelling. The finite element method was then used to numerically formulate the models. Finally, an object-oriented software package was developed using Visual C++ programming language for the models.

Experiments on cooling large cooked meat joints using conventional and vacuum cooling methods were carried out to validate the models and to provide valuable data on practical performance of the cooling systems.The numerical predictions were compared with the experimental results. For cooling cooked meat joints with weights of 5 - 8 kg from the maximum initial temperature of around 74^oC to the maximum final temperature of 10^oC, the maximum deviations between the predicted and experimental temperatures for the conventional cooling methods was within 2.8^oC and 2.5^oC for the vacuum cooling. The deviations between the predicted and experimental total weight losses were less than 2.5% during SAC and ABC, and 7.8% during VC. The experimental results show that the total cooling time for VC was less than 2 h, compared with over 5 h, 7 h and 9 h for WIC, ABC and SAC, respectively. Furthermore, meat weights had no significant effect on the cooling rate and the location of the maximum temperature was not always in the geometrical centre of the meat body during vacuum cooling, which is different from the conventional cooling processes. However, vacuum cooling resulted in big weight loss. For each 6-6.5^oC temperature reduction, an average weight loss of about 1% of the weight before vacuum cooling occurs. This weight loss can be reduced by manipulating the injection levels of brine solution.

The validated models were further used to evaluate the performance of conventional cooling and vacuum cooling systems. Simulations show that it is difficult for conventional cooling methods to cool large commercial cooked meat joints. It takes about 8.5, 6.5 and 3.5 h for ABC with an air velocity of 3 m/s and temperature of 0^oC to reduce the meats in a typical ellipsoid shape with weights of 8, 5 and 2 kg from the core temperature of 74^oC to 10^oC, respectively. For WIC with a water velocity of 0.15 m/s and temperature of 0^oC, the cooling times can be reduced to 7, 5 and 2.8 h, respectively.

Simulations show that it normally takes less than 2 h for vacuum cooling to reduce from the maximum initial temperature of 74^oC to the maximum final temperature of 10^oC, resulting in 8-9% weight loss. The effects of weights, sizes and shapes of the meats on cooling rate and percentage weight loss is negligible. The pore size of the meats has a significant effect on the pressure distribution within the meats. For commercial cooked meat joints with an average pore diameter over 1 mm, the effects of pore size on cooling rate and weight loss is also negligible. However, the cooling rate, temperature distribution and the percentage weight loss significantly depend on the porosity and pore distribution within the meats. Better performance of vacuum cooling for cooked meat joints, in terms of high cooling rate and low percentage weight loss, can be achieved if the porosity is big enough and the pores distribute homogeneously within the meats.

Sponsors: EU Food Research Programme

---

Immersion Freezing of Potato Assisted by Power Ultrasound to Improve Freezing Efficiency and Microstructure

Bing Li and Da-Wen Sun

Power ultrasound was applied to assist freezing process. The effects of power ultrasound on freezing efficiency and quality of frozen food during freezing process were highlighted.

The freezing rate of potato sample was improved with the application of ultrasound, compared to that without the application of ultrasound. The enhancement effect of ultrasound was related to ultrasonic power and exposure time. Higher ultrasonic power and longer exposure time to ultrasound would be desirable to produce strong sonication thus enhancing the freezing rate. However, the heat produced when ultrasound passed through the medium limited the power applied and the exposure time. In this work, it was indicated that the freezing rate was fastest with the ultrasonic power of Level 4 and the exposure time of 2 min. Power ultrasound was applied to the phase change period during freezing process, the freezing efficiency was increased significantly (P<0.05), while the effects were not significant when applied to the chilling phase and the tempering phase (P<0.05).

Investigation on the freezing rates on different surfaces of a bulky sample immersed in forced flowing coolant indicated that when power ultrasound was applied with low ultrasonic power level (Level 3), the agitation produced by power ultrasound to boundary layer interacted with that by forced flow in a complicated manner and the enhancement effect of power ultrasound was not significant (P<0.05). Under high ultrasonic power level (Level 4), the agitation produced by power ultrasound predominated, and the freezing rates near both the top and bottom surfaces were improved greatly (P<0.05). However, a cloud of cavitation bubbles produced under high power levels retarded slightly heat transfer near the bottom surface.

The analyses conducted on the microstructure of plant tissue using cryo-scanning electron microscopy technique showed that the plant tissue exhibited a better cellular structure under ultrasonic power Level 4. Much less intercellular void and cell disruption was observed. This was attributed to high freezing rate obtained under high ultrasonic level and thus the domination of intracellular small ice crystals.

Sponsors: EU Food Research Programme

---

Glass Transition Temperature for Chicken Meat

Adriana E Delgado and Da-Wen Sun

The poultry industry has experienced tremendous growth in the late 20th century. Chicken may well be the most universally accepted and consumed meat in the world. Consequently, breast chicken was chosen in the present work for determining the glass transition temperature, an important characteristic for understanding many aspects of food stability and processibility. The glass transition temperature of the maximum cryoconcentrated solution, T_g is often considered an indicator of the temperature below which food will be well protected from deteriorative reactions that are diffusion limited. The objective of the present study was to determine the T_gvalue for chicken muscle tissue using differential scanning calorimetry (DSC) as a means of assessing stability during frozen storage.

The T_g value, was found to be –16.8°C. The DSC annealing procedure used is suitable for detecting T_g values for chicken muscle tissue.

Sponsors: National Council of Technological and Scientific Research of Argentina, EU Food Research Programme

---

Modelling of Vacuum Cooling Process for Ready Meals Using Computational Fluid Dynamics (CFD)

Bin Xia and Da-Wen Sun

Vacuum cooling is an efficient cooling method for ready-meal components and finished products. The aim of the current research is to model the vacuum cooling process of ready meals by using the commercial CFD package CFX for a better understanding of the cooking mechanism and to facilitate optimisation of the cooling process. The modelling may be beneficial to the quality and safety of the product as well as to the reduction of the cooling time. A model has been developed and an experiment has been designed to test the model.

Comparisons will be made between the model and the experimental data and the model will be modified as required.

Sponsors: EU Food Institutional Research Measure, Department of Agriculture and Food

---

Weight Loss in Vacuum-cooled Mushrooms

Bryan Tobin and Da-Wen Sun

Vacuum cooling is a process that cools products through evaporation of a small amount of the contained free moisture. The shelf life of produce is extended but the disadvantage is a resulting high weight loss. The aim of this work was to identify the most suitable technique for vacuum cooling of mushrooms, while preventing weight loss. A vacuum cooler was used to cool mushrooms. Mass losses were recorded to evaluate the parameter that most affected the loss in weight. Factors studied include initial mushroom mass, initial moisture content, and initial temperature of mushrooms. The temperature profiles were recorded by a data acquisition system and the cooling cycle stopped at 1-2°C. The results showed that spraying with water or short-term immersion of mushrooms in water (from 5 to 30 s) prior to cooling reduces weight loss during vacuum cooling (to as low as 0.7 g) and increases their shelf life. Further studies will examine the effect of surface area on weight loss in vacuum cooling and the effect of packaging in preventing weight loss during vacuum cooling.

Sponsors: University College Dublin

---

Effect of Vacuum Cooling on Cooling Rate of Diced Beef in Tray

Zhihang Zhang and Da-Wen Sun

This research was designed to investigate the vacuum cooling process with of a tray of cooked small brisket samples so as to analyse the factors controlling the vacuum cooling rate. During vacuum cooling of a tray of small dice-formed cooked brisket beef, the temperature profile within the tray did not relate to position but to the physico-chemical properties of the individual samples. Based on cooling time, samples were divided into fast and slow cooling groups corresponding to their texture properties. The controlling factor on cooling time was deduced to be the porosity of the sample.

Sponsors: EU Food Institutional Research Measure, Department of Agriculture and Food

---