Thawing
Thawing
Friday, 10 April 2009
Interestingly, investigations of meat thawing at Langford preceded those on freezing. In the late 1960s, major pioneering studies on freezing had already been carried out in Australia and New Zealand, but little work had been undertaken on thawing. Initial studies on pork legs were rapidly expanded to meet other industrial demands for data on thawing systems.
Factories that produce pies, pasties, burgers, or other canned, frozen or chilled meat products require regular supplies of raw material. To take advantage of cheaper raw material arising from seasonal fluctuations in availability and to provide for scarcities, meat is often purchased, frozen and stored until needed. Thawing such material presents difficulties in obtaining reproducible times, bacteriological condition, physical appearance and loss of weight. Thawing is the reverse of the freezing process, but the three-fold difference in thermal conductivity between frozen and unfrozen material increasingly restricts the rate of heat penetration during thawing. If bacterial numbers are to be maintained at an acceptable level then limitations must also be applied to the maximum surface temperature.
Early surveys at Langford revealed a range of very uncontrolled thawing systems that were being used within the meat industry (Cutting & Malton, 1974). Most were air based, often involving no more than placing product on racks or pallets and relying on ambient air to carry out the thawing process, whilst others used tanks of running water supplied by a hose. In the latter, some water ran out through a hole in the bottom of the tank, whilst the rest overflowed. The flow was often interrupted for indefinite periods when the filling hose was required for some other purpose. In one plant, temperatures in beef quarters were still below -3°C after 6 days of thawing.
The thawing of frozen legs of pork for subsequent curing and ham production was the first topic. Pilot plants were set up at the site to look at three different thawing media: air, water and steam at sub-atmospheric pressure (vacuum thawing).
In the air thawing tunnel an axial flow fan blew air through a diffuser and distributor into a 1200 x 550 x 550 mm working section to produce velocities in the range 0.25 to 6 ms-1. The tunnel was installed in a temperature controlled room that operated over a range from 5 to 35°C (±0.5°C). The water thawing unit consisted of a 1170 x 890 x 890 mm galvanised steel outer tank, containing 900 litres of water, with a 610 x 480 x 890 mm inner working section. Water could be pumped through the working section at speeds between 0.006 and 0.023 ms-1 and its temperature controlled to ±0.5°C over the range 5 to 35°C. A 1-ton capacity vacuum-thawing unit (AVP Parafreeze, Thetford), that could maintain vapour temperatures to ±0.5°C within a 10 to 30°C range, was used for the vacuum thawing studies.
A major factor in air and water thawing is the movement of the thawing medium over the meat surface, which reduces the effective thickness of the boundary layer and increases the surface film heat transfer coefficient. When the thawing medium is steam at below atmospheric pressure, heat is transferred to the surface of the meat by the latent heat of condensation of the vapour. The film coefficient then depends on the volume of non-condensable gases present at the surface and its orientation. In all cases, the rate of heat transfer into the material being thawed will be a function of the temperature difference between the thawing medium and the surface of the food.
Over 200 individual pork legs were thawed in experimental investigations that covered 15 combinations of thawing method and media conditions. Specially constructed multi point temperature probes were used to locate the thermal centre of the legs and provide data on the temperature gradient during thawing. Experimental data were also obtained on changes in the appearance, weight and bacterial condition of the legs during the process. However, legs of pork range in size from 2 to over 9 kg, and only a small subset of possible commercial thawing combinations could be covered experimentally.
Shortly after the start of the experimental programme, Charm et al. (1972) published a paper on the prediction of food freezing which included a complete FORTRAN coding of a computer program, based on a numerical method of Dusinberre (1949). Difficulties were found in adapting Charm’s program to thawing pork legs, principally due to its inability to handle temperature dependent thermal properties. A finite difference coding based directly on the numerical method of Dusinberre was therefore developed that allowed for changes in heat transfer coefficients (h) and temperature and positional dependent thermal properties. This was the start of modelling at Langford.
The approach taken was very simple. As it was impossible to calculate directly the heat flow through the complicated structure of a pork leg, the leg was equated to a geometric solid that could be treated mathematically. The figure below shows a cross section of a large slab of thickness x having a uniform cross-sectional area A; the solid is divided into a number of finite slices of thickness ∆x by temperature reference planes. A heat balance is written on the cross-hatched zone abcd; the slope -dt/dx at plane ad is approximately equal to the chord slope (to - t1)/∆x. Similarly the slope -dt/dx at plane bc is replaced by (t1 - t2)/∆x. The temperature at plane 1 approximates to the average temperature of the cross-hatched zone.
Dusinberre diagram off one-dimensional transients in a slab. x axis depth and y axis temperature in arbitrary units
The resulting heat balance is:
where t1’ is the new temperature at plane 1, after the elapse of a finite time increment ∆t, K is the thermal conductivity, is the density and C the specific heat. Replacing the dimensionless ratio (∆x)2 Cp/K∆T by the modulus M we obtain the equation for heat conduction through the interior of the block:
Using a similar process, equations can be derived for heat flow at the surface and centre of the slab (Bailey et al., 1974).
An infinite slab is not a good geometric analogue to a pork leg in which heat flow is three dimensional. However, it can be argued that the same principles used for the slab apply to a sphere where the heat flow, although in fact three-dimensional, is radial. A code was therefore developed for spheres and thawing times were calculated for spheres of lean pork with a range of radii, media temperatures and surface heat transfer coefficients (h).
The final stage was to relate the radii of the spheres to pork legs of different weights. It was evident from the model that increasing the surface heat transfer coefficient (h) above 500 Wm-2K-1 had little effect in further reducing thawing time. Coefficients as high as 5,000 Wm-2K-1 could be regarded as “infinite”. Published values for heat transfer coefficients in vacuum heat thawing were in the 500 to 5,000 Wm-2K-1 range. The vacuum thawing results were therefore related to the predicted thawing times at 5,000 Wm-2K-1 to provide a relationship between leg weights and the diameter of equivalent spheres. This relationship was then used together with published data on the heat transfer coefficients in different media, to calculate thawing times for a range of different thawing systems. Good agreement was found between the predicted and experimental results. The whole study (Bailey et al., 1974) and a design data chart, which extended the applications (Bailey & James, 1974), were published in 1974.
As confidence grew in the ability to predict thawing times with increased accuracy, the technique was used with a reduced requirement for experimental investigations. Data was produced on the thawing time of beef quarters (James, Creed & Roberts, 1977; James & Creed, 1980), lamb (Creed et al., 1979) and mutton (Creed & James, 1984) carcasses, meat blocks (James & Bailey, 1980; Creed & James, 1981) and cook-freeze catering packs (Brown et al., 2006). In addition to spheres, infinite slabs and cylinders were used as geometric analogues in these investigations. During this time a range of increasingly sophisticated pilot plants were constructed at Langford, with better control and more accurate data collection and analysis equipment. Details of these modelling approaches, and the use of empirical and analytical methods used to produce much of the design data on chilling published by us, have been recently published in a chapter on “Modelling of Heat Transfer during Food Chilling, Freezing, Thawing and Distribution” (James et al., 2009) in “Predictive Modeling and Risk Assessment”. This is part of the ISEKI-Food book series aimed at graduate students and senior level undergraduate students as well as professionals and researchers.
Some small but significant changes to the predictive models and their method of use were undertaken as the methodology developed. Due to the substantial changes in the thermal properties of meat in the temperature range encountered during thawing significant errors were produced in using the simplified equation 2. The codes were therefore changed to solve equation 1, and other similar equations, directly using thermal properties evaluated at the relevant temperature (James, Creed & Roberts, 1977). Data were also sought on the relationship between conditions in the thawing medium and the value of ‘h’. Few data were available on ‘h’ in food processing in general and on thawing in particular. Values from different sources could differ substantially.
Investigations were carried out to measure ‘h’ during air thawing at different relative humidities. These revealed that, in conventional air based thawing systems, the contribution to ‘h’ resulting from the condensation of water vapour was substantial (James & Bailey, 1982a). At 30°C and a dew point of 28°C, ‘h’ at 3 ms-1 varies from 70 to a peak value of 115 Wm-2K-1 over the surface temperature range -20 to +26°C.
The curves resulting from plotting heat transfer coefficient against surface temperature can be divided into specific sections, which are related to the dew point of the air and the orientation of the surface of the meat. At surface temperatures below 0°C water sublimes to ice on the surface of the meat enhancing the surface heat transfer coefficient because of the phase change. At 0°C, this ice layer melts, and a temporary heat requirement causes the inflection in the curve. From 0°C to the dew point, water condenses as liquid on the surface of the meat, further enhancing the surface heat transfer coefficient by the heat of condensation. Above the dew point conditions are reversed; the layer of condensate evaporates and the surface heat transfer coefficient falls because of the evaporative requirement. When evaporation is complete ‘h’ has a substantially constant value dependent on convection and radiation. The duration of the evaporative section is clearly affected by the orientation of the surface since a much larger amount of condensate will remain on a horizontal than an inclined surface.
Even using methods that achieved high rates of surface heat transfer, thawing times were still long e.g. 20 h plus for a 15 cm thick meat carton. This was due to the poor thermal conductivity of the thawed layer that progressively advanced inward from the surface, and the need to limit the surface temperature so that rapid bacterial growth would not occur. Consequently there has always been industrial interest in electro-magnetic systems, i.e. microwave, radio frequency (RF) and ohmic, that are not restricted by thermal conductivity (Bailey, 1975; James & Bailey, 1983). Most efforts had concentrated on the use of microwaves, with successful commercial systems developed for meat and butter tempering (see next section). However, in thawing their application was limited by thermal instability and penetration depth.
Thermal instability results from preferential absorption of energy by warmer sections of the foodstuff and by different ingredients, such as fat. Warmer sections may be present at the start of the process, for example the surface temperature may be warmer than the centre. Alternatively, they may be produced during the process, by energy being absorbed at the surface rather than penetrating all of the product. In extreme cases, such warming can result in some parts of the food being cooked while other sections remain frozen.
Combining vacuum cooling with microwave heating, was a successful method of overcoming the problem of thermal instability (James, 1984). Twenty five kg cartons of meat were thawed in 1 h with final meat temperatures in the range 0 to 20°C. However, the process resulted in weight losses of up to 15%, was very energy inefficient and difficult to scale up to commercial throughputs.
The rate of thawing in meat, and most food for that matter, is clearly limited by its poor conductivity. High conductivity inserts, such as heat pipes, have been studied at Langford to overcome this problem. A “heat pipe” as its name implies is simply a pipe through which heat passes if there is a small temperature difference between its ends. If the pipe is placed in contact with a warm medium heat will be extracted from the medium and transferred to the coldest part of the pipe. It can therefore be used as a means of either extracting heat from a warm medium or introducing heat into a cold medium. The same process could be achieved using a metal rod but the heat pipe has by its construction the advantage of being up to 500 times as effective as copper at transferring heat energy. The main advantage of the heat pipe being that it relies on condensation and evaporation rather than conduction to transfer heat. Heat can therefore be extracted directly from the warmer inner regions of a food. A typical heat pipe consists of an evacuated metal tube, on the inside of which is a wick containing a certain amount of working fluid. Heating one end of the tube causes the liquid to evaporate and the vapour travels along the tube. As the other end of the tube is cooler the vapour condenses, at this stage the latent heat is released and the liquid travels back to the original end along the wick by capillary action, to complete the cycle. The heat pipe is an efficient means of transferring heat because it uses the latent heat energy of the working fluid rather than the heat capacity. Some idea of the advantage gained can be gathered by considering water as a working fluid. The heat capacity of water is roughly 4.19 kJkg-1K-1, while the latent heat of vaporisation is 2257 kJkg-1. So the same amount of heat is released in condensing 1 gram of steam as that released in cooling 538 grams of water by 1°C.
During the mid-1970’s the “thermal cooking pin” or “thermal skewer” was developed using heat pipe technology. This was designed to be inserted into joints of meat during cooking, thus it was claimed achieving a considerable reduction in cooking time, producing a more evenly cooked joint and a large fuel saving. Though the thermal skewer showed great promise it was marketed at the domestic and catering area of the food industry and not adopted to any great extent. Around this time research was also carried out at Langford on the effect of heat pipes on freezing and thawing times for meat joints. This showed that freezing and thawing times of meat joints could be reduced by up to 42% and 55%, respectively, using heat pipes. This work was not published at the time but formed the basis of a paper in the late 1990s (James & James, 1999) when we had the opportunity of re-examining the potential of heat pipes yet again. A number of subsequent studies at FRPERC have shown that using heat pipes; chilling, freezing and thawing times of food products and meat joints can be reduced by between 20 and 45% (James, Ketteringham & James, 1998; Ketteringham & James, 2000; James et al., 2005b).