2014 Volume 20 Issue 1 Pages 43-49
Rapidity is the biggest advantage of vacuum cooling in the food industry, especially for cooked meat. However, the largest volumetric displacement does not guarantee the best temperature uniformity because the structure of the food can affect the mechanisms of heat and mass transfer in food during the vacuum cooling process. Because the temperature uniformity is one of the most important factors in the cooling effect, it is necessary to achieve a uniform temperature distribution across the portion of food by choosing the best evacuation scheme. In this paper, the volumetric displacement was varied with the pressure to conduct a vacuum cooling experiment for heated ham. The results were evaluated by measuring the real-time temperature distribution on the ham section using a sensitive infrared thermal imaging system. The results indicated that properly varying the volumetric displacement during the cooling process can not only uniformize the temperature distribution across the ham section but that it can also reduce both the weight loss and energy consumption. Consequently, a reasonable varying volumetric displacement can help improve the vacuum cooling effect.
Cooked meat is becoming increasingly popular all over the world. In particular, ready-to-eat meat has a very large market in developed cities. However, with the development of industrial-scale processing of ready-to-eat meat, the safety problems caused by bacteria become an increasing threat to the consumer. The shelf life of food stored at higher temperature will be shortened because most of bacteria are thermophilic (Juneja et al., 2011). To avoid this danger of high temperatures, standards have been issued to control the cooling rate of foods. For example, the Food Safety and Inspection Service of the Department of Agriculture of the US determined that 90 minutes was the upper limit of the time allowed to cool the center of uncured meat from 54.4°C to 26.6°C. In the UK, less than 1.5 hours was permitted to cool the center temperature from 74°C to 10°C for cooked meat (Anonymous, 1989). As a result, various methods such as air blast, cold room, water immersion cooling and vacuum cooling have been studied in an attempt to cool hot cooked meat (70 - 90°C) to a low temperature (4 - 10°C) in a required time (Zhang et al., 2013; Zheng et al., 2004; Sun and Zheng, 2006; Drummond et al., 2009; Schmidt et al., 2010; Feng et al., 2012b; Dong et al., 2012;). Of these methods, vacuum cooling was the most effective.
The basic principle of vacuum cooling depends on evaporating a portion of the free water inside food to cool the remaining portion under low pressure, which is lower than the saturated evaporative pressure corresponding to the temperature at the food surface. Vacuum cooling was first applied to leafy vegetables (Webb, 1954). Because water can evaporate from any point in a food that has enough useful pores or void space, the temperature distribution of foods cooled by the vacuum cooling is more uniform compared to other methods. However, this uniformity comes at the cost of weight loss, especially for minced meat. Because of the anisotropy elements of cooked meat, different parts of the same food need different amounts of time to reach the desired temperature during the cooling process (Wang and Sun, 2002; Wang and Sun, 2003; Jin et al., 2006; Jin, 2007). Consequently, there are often temperature differences among different parts of the same food at the end of the vacuum cooling process. A faster cooling rate at the outer surface of the food results in a greater temperature difference. A too low temperature can make the food harden, thus destroying its taste. A too high temperature can allow bacteria to grow, and the food will deteriorate more rapidly. For a large piece of meat, a suboptimal temperature distribution can lead a greater weight loss. As a result, temperature uniformity remains a key problem that needs to be solved.
To optimize the vacuum cooling techniques for meat, multiple studies have been carried out. For example, Self et al. found that changing the pump rate can achieve a better cooling effect for the homogenous batch of samples (Self et al., 1990). Because the evaporative area is the most important factor for the cooling rate, most of these reports focused on the pore size and distribution in the cooked meat (Wang and Sun, 2003; McDonald et al., 2000; McDonald and Sun, 2001a; McDonald and Sun, 2001b; McDonald et al., 2001; McDonald and Sun, 2001c). The conclusions of these studies indicated that the cooling rate is proportional to the porosity of the sample and that the porosity is the sum of the holes and void space (McDonald and Sun, 2001c).
However, some research groups failed to reach an agreement on certain problems. For example, the locations of the maximum and minimum temperatures varied among papers. T.X. Jin stated that the lowest temperature occurred at the surface and the highest temperature should be at the core of the cooked meat during the vacuum cooling process (Jin, 2007). However, Wang and Sun found that the highest temperature was not always observed at the core (Wang and Sun, 2002). It is possible that these differences were affected by the material characteristics used in the respective experiments (Webb, 1954; McDonald and Sun, 2001c; Landfeld et al., 2002). However, the refrigeration effect of pores and void space in meat were mostly similar, which has been accurately verified by mathematic simulation. In addition to the refrigeration effect of pores, the development trend of pores has also been reported (McDonald et al., 2000; McDonald and Sun, 2001b).
However, few studies have described the real-time temperature distribution across a portion of food because the internal temperature distribution is difficult to detect intuitively over time (McDonald and Sun, 2001c). To explore the mechanisms of heat and mass transfer of ham during the vacuum cooling process, an infrared imaging system was used to capture the real-time temperature distribution. The concept of varying volumetric displacement with pressure was introduced into this experiment with the goal of achieving a more uniform temperature. The obtained results are inspiring because varying the volumetric displacement not only reduce the temperature difference but also is beneficial to reducing weight loss and saving energy.
Sample preparation Hams (moisture content, 70%; fat content, 4.9%; protein content, 17.7%; carbohydrate content, 2.4%; content of inorganic salt, 1%; content of other components, 5%) were bought from a supermarket. The hams were shaved by a knife to the desired dimensions for experiments: 82.5 mm in diameter and 20 mm in height, as shown in Fig.1a. Because one aim of this paper was simulating the inner section temperature of an infinitely long ham, the upper and lower end faces of the samples had to be sealed to prevent vapor from escaping along the axial direction. The upper end face was uniformly covered with vacuum grease (Fig. 1b) to ensure the accurate detecting of the real-time temperature of the upper end on the one hand, and prevent vaporizing of water on the other hand. The lower end face was sealed using plastic wrap. Since 20 ∼ 70°C is the most sensitive temperature range for microbial growth. Temperature change mechanism of ham during this range is very important. So, the hams before vacuum cooling process were heated using a thermostat water bath (45 ± 2°C) until the core temperatures reached 45°C.
Bare ham, ham covered by vacuum grease and the dimensions of the shaped ham. On the covered ham, T1 indicates the core temperature, T2 indicates the temperature at 1/2 center, and T3 indicates the boundary temperature.
Vacuum cooling and temperature measurement To detect the real-time temperature distribution of the upper end face during the vacuum cooling process, a new vacuum cooler equipped with a sensitive infrared thermal imaging system (Testo, 885-2, Germany) was designed, as shown in Fig. 2.
Schematic diagram of the vacuum cooler: (1) vacuum chamber; (2) infrared window fixed at the top of vacuum chamber; (3) online infrared thermal imager; (4) electromagnetic bleeding valve; (5) pressure sensor; (6) computer;(7) temperature sensor;(8) programmable logic controller;(9) frequency converter; (10) vacuum pump; (11) cold trap; and (12) sample.
The vacuum pump (Leybold, D8C, Germany) and the frequency converter (SINAMICS, V10, Germany) were combined to reduce the total pressure of the vacuum chamber (0.045 m3) to 1150 ∼ 1200 Pa (Testo, 435-4-0638 1835, Germany) and to keep the total pressure within this interval until the vacuum was released. The temperature of cold trap was maintained between −12 and −15°C to adsorb the water vapor coming from the vacuum chamber. Both the control task of the vacuum cooler and data collection were performed by a programmable logic controller (SINAMICS, S7-224, Germany) with the help of a programming software (STEP 7-MicroWin V4.0.8.06).
Relationship between the volumetric displacement and the pressure of the vacuum chamber Three schemes for adjusting the volumetric displacement were used in this study : fast group (large volumetric displacement), moderate group (medium volumetric displacement) and slow group (small volumetric displacement).
The actual volumetric displacement (S) of vacuum pump was calculated using the Eq.1, as follows:
 |
Where 0.002 is the maximum volumetric displacement of vacuum pump, m3/s; 50 is the rotational frequency of the vacuum pump when it runs in the volumetric displacement of 0.002 m3/s, Hz; f, which was input into the frequency convertor by the programmable logic controller, is the set rotational frequency value of the vacuum pump, Hz.
In each group, the rotational frequency of the vacuum pump was changed according to the corresponding curve described in Fig. 3.
Frequency conversion schemes
Weight loss rate The weight loss rate of the ham was calculated using the Eq.2, as follows:
 |
Where A is the weight loss rate; W0 is the original weight of the ham before vacuum cooling, g; and W1 is the weight of ham after vacuum cooling, g.
Statistical analysis The weight loss and energy consumption were compared using the t-test at the 5% probability level (P < 0.05) (N = 5) with a statistical program SPSS (Version 18.0).
Results
(1) Temperature distribution of the upper end face in fast group
a) Temperature profile of the feature points
It can be observed from Fig. 4 that both T1 and T2 dropped slowly with the decrease in total pressure. However, when the total pressure reached approximately 100 kPa, they both dropped 3°C rapidly and then suddenly increased 2°C. Then, both points dropped to 10°C steadily. The behavior of T3 was similar, except there was no sudden drop before the continuous cooling process.
Changing curves of T1, T2 and T3 in the fast group
In addition, this graph indicates that the temperature became lower at increasing distances from the section core.
b) Overall temperature distribution of the upper end face
The overall real-time temperature distribution on the upper end face of the ham during the vacuum cooling process at the fast group was described visually in Fig. 5. The following conclusions can be drawn: (1) The boundary temperature decreased most rapidly. It just took 9 minutes to cool the boundary to 9.7°C, but the hottest point cost more than 24 minutes; (2) The max temperature did not always exist at the geometric center of the upper end face and located at about 1/5 center at the third minute, which agreed with the conclusion of reference (Sun and Wang, 2006); (3) The temperature distribution during the first 3 minutes was the most uniform and became increasingly nonuniform over time. Especially at the ninth minute when the max temperature located at about 1/3 center, an obvious hot zone appeared at the offset-center position continually. Certainly, this degree of nonuniformity was weakened by the internal heat conduction.
Infrared pictures of upper end face for the fast group
(2) Temperature on upper end faces during the various groups
a) Temperature profile of the feature points
Figure 6 showed that the slow group required the least amount of time (550 s) to cool the ham core from 40 to 10°C, which was 450 s less than the fast group. Obviously, this phenomenon did not agree with the conventional view that using a larger volumetric displacement of pump will save time.
Comparison of T1 in various groups. For comparing the T1 values obtained from three different schemes visually, 40°C was used as the starting point and the temperature fluctuation between 80s and 100s shown in Fig.4 was ignored.
It can be seen from Figure 7 that the temperature of the half center of the upper end faces decreased faster in moderate group compared to fast group. Even if T2 chosen randomly in moderate group was affected by the uniform temperature distribution of upper end face of ham and thereby decreased faster compared to slow group before 500 s , the slow group saved approximately 50 s compared to the moderate group finally.
Comparison of T2 in various groups. For comparing the T2 values obtained from three different schemes visually, 44°C was used as the starting point and the temperature fluctuation between 80s and 100 s shown in Fig.4 was ignored.
As shown in Figs. 6, 7 and 8, the boundary of ham was colder than the other parts when the pump was started because it was exposed to the chamber directly. Figure 8 indicated that as the increasing of the volumetric displacement, the time used to cool the ham boundary to 10°C became longer. Especially in the slow group, this processing time was approximately 170 s, which was 170 s less than the moderate group and 630 s less than the fast group.
Comparison of T3 in various groups. For comparing the T3 values obtained from three different schemes visually, 25°C was used as the starting point and the temperature fluctuation between 80 s and 100 s shown in Fig.4 was ignored.
From the above results, an unanticipated conclusion was summarized: a larger volumetric displacement of pump does not always result in a shorter processing cycle during the vacuum cooling process.
b) Overall temperature distributions of the upper end faces
Figure 9 indicated that using either a medium or small volumetric displacement can achieve a more uniform temperature distribution compared to the large volumetric displacement. In particular, the slow group achieved the best uniformity. For example, the hot zone on the upper end face began to become increasingly irregular from the sixth and ninth minutes in the fast and moderate groups, respectively. However, in the slow group, the hot zone was standard and rounded throughout the process, and its center was located at the geometrical center of the upper end face. Consequently, reducing the volumetric displacement of pump offers a useful strategy for achieving a uniform temperature distribution.
Overall temperature distributions of the upper end face in various groups. Note: Because these three sets of data were obtained from three different videos, the same original color did not indicate the same temperature across the various groups. Therefore, they were convered to black and white pictures firstly and then were adjusted.
(3) Inner structure of cooled ham
The ham portions were cut open in slices that were parallel to the upper end face to observe the inner structure of the cooled ham, as shown in Fig. 10. It can be seen from Fig. 10 that the cooled ham was composed of two zones: a hard boundary crust that agreed with a previous study (Desmond et al., 2000) and a soft inner original texture. The soft zone contained many microscopic holes and a few large holes. These holes were useful for evaporating the free water inside the ham by increasing the porosity. As a result, the distribution of these holes may be a factor that affects the temperature distribution (Wang and Sun, 2002; Sun and Wang, 2006).
Relationship between the structural change of the ham and the mass and heat transfer within it
However, this theory cannot explain the difference in uniformity of temperature for the three groups as shown in Fig. 9. One possible explanation may be that the transportation of inner vapor or water was restrained more easily at larger volumetric displacement by some distortional texture that was located at the junction between the boundary crust and the inner original texture. To verify this assumption, it was necessary to analyze the mechanism of heat and mass transfer in the ham during this experiment.
Discussion Based on Fig. 4, the means of moisture transfer in the ham in the fast group may consist of the following steps. First, the pump was started to produce the vacuum for the chamber at the fastest punp speed. Because the partial pressure of the water vapor in the chamber was reduced simultaneously, the moisture in the ham began to diffuse from the ham boundary to the chamber (Cepeda et al., 2000). This diffusion process is able to cool the ham at a slow rate. During the vacuum cooling process, moisture at the core was pumped out gradually via a junction zone such that the whole ham was cooled simultaneously. Once the pressure dropped to a “flash point” at which the water in the ham boiled, the cooling rates of all parts of the ham increased suddenly. However, due to the effects of the final pressure and the adsorption capacity of the other components on free water, the cooling rates dwindled.
If the volumetric displacement of pump is too large, the holes in the boundary crust will be enlarged, and the cooling rate of the junction zone will increase greatly. When the junction zone was cooled to a certain temperature, it shrank, narrowing the evaporative channel of the inner water vapor and decreasing the permeability of the water vapor. With further narrowing, the temperature on the same section was divided into two zones. Because the materials used to make the ham were mixed together in a disorderly fashion, these two zones became increasingly malformed. If the volumetric displacement was larger, the uniformity would be worse, as shown in Fig. 9. This shrunken zone also can be confirmed by the dried boundary crust, as indicated in Fig. 11 which was captured by CT scan. It also follows from Fig. 11 that the dryness and uniformity were both worst at the boundary crust of the ham processed at a large volumetric displacement.
Comparison of the boundary crusts of hams dried by natural air. Note: In this CT scan (XT H 225 ST, X-Tek Systems Ltd, UK), black indicates that the texture lost too much water, and white indicates that the texture remained similar to the original
Otherwise, as shown in Fig. 4, T2 underwent a short down-and- up course, and T3 underwent a short rising course. This outcome occurred because after starting the pump, the vapor in the holes was forcibly transported away, and the surface of the holes cooled suddenly. However, the vapor-liquid mixing films on the surfaces of these holes were too thin to cool the invisible parts of the ham. Consequently, the ham returned to its normal temperature quickly. At the same time, part of the diffusing water vapor was prevented and adsorbed by the substrate. When the adsorbed moisture transformed from vapor to liquid it released a lot of latent heat resulting in the temperature rise of sample. Because the initial temperature of the boundary crust was lower than that of the other parts, once the hot vapor coming from the inner parts touched it, T3 increased.
In addition to achieving better temperature uniformity, a varying volumetric displacement with pressure seems to be good for reducing weight loss and energy consumption during the vacuum cooling process for ham, which agreed with reference (Feng et al., 2012a). As shown in Table 1, both the weight loss and energy consumption were clearly reduced in turn from fast to moderate to slow group. The small volumetric displacement was able to minimize the weight loss due to the slow phase change and slow flow velocity of moisture in ham. Because evaporative water was transferred in the form of vapor (Jin and Xu, 2006), it had a faster velocity than did liquid water in ham. The vapor can thus easily carry liquid water into the micro channels. If the volumetric displacement is small, the water vapor will have more time to be absorbed by the matrix in meat. Moreover, there often has some uncondensed gas in the cooked foods(Daudin, 1990; McDonald et al., 2002), which has the entrainment effect on the water vapor when it was pump out. If the volumetric displacement was reduced, both the entrainment effect on the water vapor and the expansive action on the food structure (Tara, 1990; Bartos et al., 2002; Cheng and Lin, 2007) will be weakened, which was helpful for water vapor to be absorbed by the matrix too. For the energy consumption, it can be found that the pump frequency decreased from fast to moderate to slow group in the pressure range from 80 kPa to 2 kPa, according to Fig. 3, therefore the average power also decreased from fast to moderate to slow group.
| W0(g) | W1(g) | A(%) | Energy consumption(kJ/kg) | |
|---|---|---|---|---|
| Fast | 123.37 ± 0.04 | 111.87 ± 0.02 | 9.32a | 9.97a |
| Moderate | 126.23 ± 0.05 | 115.09 ± 0.04 | 8.83b | 7.90b |
| slow | 122.69 ± 0.04 | 114.37 ± 0.03 | 6.78c | 7.63c |
Means with different letters in the same column indicate significant difference (P < 0.05)
The mechanism of simultaneous heat and mass transfer of moisture in ham during a vacuum cooling process was studied. The results showed that the boundary always became cold more rapidly than did the other parts, resulting in the production of a crust. Through this crust, the inner water can be exhausted by the pressure difference and thereby cool the whole ham. The volumetric displacement should not be too large. Otherwise, the junction zone will shrink and leave the transfer channel of inner water restrained. Except for the texture difference, the shrinking of the junction zone was the primary reason for the uneven temperature distribution. The uneven degree of temperature distribution across the section became reduced with proper variation of the volumetric displacement. In addition to the temperature distribution, both the weight loss and energy consumption can be reduced by varying the rate with pressure. Consequently, using a varied volumetric displacement scheme to replace the fast rate was reasonable.
Acknowledgements This work was supported by the Innovation Fund Project for Graduate Student of Shanghai (JWCXSL1201) and National Natural Science Foundation of China (51076108).
