14 Temperature measurement It is often stated by those in the meat and refrigeration industries that ‘anyone can measure a temperature’. Many millions of measurements are made of both meat and environmental temperatures in the meat industry. However, in many cases the measurements made are an unreliable guide to the effectiveness of the refrigeration process. Even when the correct tem- peratures have been obtained the data are often poorly analysed and rarely acted upon. If a group of people are asked to measure the temperature of a beef carcass, the number of values obtained is often the same as the number of people in the group. Few initially ask the obvious question, ‘what is meant by the temperature of the carcass?’ Is it the average temperature, the highest temperature, the lowest surface temperature, the average surface temperature...? The increase in temperature legislation and the desire of meat produc- ers and retailers to maintain the organoleptic and microbiological quality of meat throughout the chilled and frozen distribution chain has cre- ated an increased demand for equipment and expertise on temperature measurement. The industry needs to measure temperatures accurately, reliably, mean- ingfully, simply and cheaply. It needs to be able to analyse the data and respond when required. It needs the correct instrumentation and the exper- tise to collect and interpret the temperature data. 14.1 Instrumentation The first consideration is the range of temperatures to be measured. For the meat industry, a range from -40 to +150 °C would cope with the tempera- tures found in freezers, chillers, storage rooms, retail display cabinets and in water used for cleaning or scalding tanks in the abattoir. If they produce cooked meat products then the upper temperature may rise to 250–300 °C. As well as the measuring range, the range of ambient temperatures over which the instrument will work needs to be considered. The electronics of many temperature measurement instruments are designed to work to the specified accuracy only within certain ambient temperature ranges, usually 0–40 °C. If temperatures in a cold store are to be measured the instrument itself may need to be kept warm until it is used. 14.1.1 Hand-held digital thermometers Purely from a cost consideration many small producers and retailers rely on spot temperature checks obtained using hand-held thermometers to produce the temperature records they require. The main tasks they carry out with such equipment is the measurement of air temperature, between pack or product temperature, and the temperature of the meat itself. They require thermometers that are accurate, easy to use, react quickly and are robust. Ease of use is a personal judgement and best answered by trying out a range of instruments. Most modern electronic thermometers are reli- able if handled with reasonable care. However, in general, the more robust the sensor the slower the response. There are three types of digital thermometer generally available: ther- mocouple, platinum resistance or semi-conductor (thermistor). The name refers to the type of temperature sensor used. Type T (copper–constantan) thermocouple thermometers with a wide range of interchangeable sensors are the most widely used because of their wide temperature range and rea- sonable accuracy.The accuracy of the temperature measurement of a digital thermometer will depend on how accurate the instrument and the sensor are. It can be seen from Table 14.1 that only thermometers based on ther- mistor or platinum resistance sensors can be guaranteed to provide better than ±0.5 °C accuracy. However, it is possible to calibrate any thermometer at known temperatures and use the calibration curve obtained to correct errors in measured values. In many cases the supplier of the instru- ment can provide a calibration curve for a particular instrument/sensor combination. Sensors do not immediately measure the temperature of the meat or air in which they are positioned. Their response rate depends on the sensor itself and the environment in which it is used. A thin sensor in a wet/solid food will respond rapidly, and a thick sensor in still air very slowly. When 284 Meat refrigeration sensors are shrouded to improve their robustness their response times increase substantially (Table 14.2). Most manufacturers supply a range of cased sensors (probes) normally made of stainless steel suitable for most applications. These include blunt- ended probes for general purpose use, needle-point or hypodermic probes for inserting into solid or semi-solid food, probes with spring-loaded ends for measuring surface temperatures or very robust probes that can be screwed or hammered into frozen meat. Probe length is not important for measuring air and liquid temperatures. However, to check the deep leg temperature of a side of beef the probe needs to be at least 15 cm long to measure the temperature at the deepest point. 14.1.2 Temperature recorders There may often be a requirement to measure the values of temperatures at many different positions at the same time. The simplest solution to this problem is often to attach a multipole switch to a digital thermometer. A number of temperature sensors can then be connected to the switch and their temperatures monitored in succession. This procedure is com- monly used in central plant rooms where an operator can routinely look at and record the temperatures at many different locations. A hand-held digital thermometer will provide information on one temperature at one Temperature measurement 285 Table 14.1 Accuracy of digital instrument, temperature sensor and overall temperature accuracy of combined thermometer Instrument (°C) Sensor (°C) Both (°C) Type K thermocouple ±0.3 ±1.5 ±1.8 Type T thermocouple ±0.3 ±0.5 ±0.8 Platinum ±0.2 ±0.2 ±0.4 Thermistor ±0.2 ±0.1 ±0.3 Table 14.2 Response times (s) of sensors in air Sensor Air condition Still Moving Bare thermocouple 20 5 Bare thermistor 45 20 Shrouded thermocouple 150 40 Shrouded thermistor 260 50 Shrouded platinum 365 65 time. However, in many cases there is a need to measure temperatures over a long time period. In these situations a temperature recorder is required. Historically, the temperature history of a point has been obtained using a temperature sensor connected to a moving chart. In its simplest form this is a stylus on the end of a bimetallic strip that bends in response to tem- perature changes and scratches a continuous trace on a carbon chart moved by a clockwork motor. More sophisticated devices use electrical tempera- ture sensors attached to a small chart recorder. The recorders can be driven from batteries or direct from the mains and the chart can be circular or rec- tangular and mounted on a drum or on continuous rolls. Typically instru- ments will provide a continuous trace for up to a week, but some specially developed for long distance shipboard transportation of meat can operate for 6–8 weeks. Increasingly, solid state electronic devices are being used to obtain the temperature history of a point. In most solid state devices the output from an integral electronic sensor is measured at set time intervals, converted to a temperature measurement and stored in a computer memory chip. In a small number of devices the interval between recordings can be adjusted and recordings started using buttons or switches on the instrument and the temperatures examined on an in-built display. A development is the use of small printers that can either be used to print out the temperatures as they are measured or attached after data collection is finished and the whole temperature history printed out. However, with the majority of instruments a small computer is required to set up start times, logging intervals and so on and recover the temperature recordings. With many systems it is difficult to look at the temperature while it is being recorded or even check that the required information has been obtained before leaving the recording site. Some of the newest instruments are totally encapsulated in waterproof plastic and can be placed in direct contact with solid or even within liquid foods. Solid state devices can be very small, effectively tamper proof, and the value of the temperature at set times easily obtained, but the requirement for an associated processing facility substantially increases their cost. In many cases moving chart instruments may still provide the most economic and convenient solution to a monitoring problem. If precise temperature values at certain times are not required, then a quick examination of the chart may be sufficient to show that the temperature of the display cabinet, store room or transport vehicle has kept within the prescribed limits. However, obtaining tempera- ture values from a small chart can be time consuming and inaccurate. In some situations the actual or relative position of the sensing points is important, whilst in others the position of maximum or minimum temper- ature is required. There are few, if any, commercial sources of multi-point temperature probes and most have to be specially constructed. The sensors are attached to basic probes constructed from composite fibre or wood of 286 Meat refrigeration the smallest cross-sectional area that can be used, whilst maintaining the required robustness, to minimise heat conduction and achieve a rapid tem- perature response. For hygienic reasons, probes are thinly coated with inert epoxy resin. Currently there is no real alternative to the use of an array of individual temperature sensors if data on temperature distribution are required. For over 50 years temperature sensors attached to multi-point chart recorders have been used to obtain the temperature history of up to 24 positions and these systems are still common in many processing plants. To differentiate between the sensors on the charts, a range of methods including different colours, line types or numbers have been used. As the number of sensors increases it becomes more and more difficult to identify individual sensors and/or temperature values. If the sole purpose of the recordings is to show that all the temperatures remain within upper and lower limits then chart recording systems are more than adequate. In situations where a more detailed analysis of the data is required then they are being increasingly replaced by microprocessor-controlled data logging systems. Data loggers range from multisensor (typically 2–16 temperatures) ver- sions of the solid state instruments already mentioned, to sophisticated processing systems with thousands of measurement points. Two types of portable logging systems are available. The larger type (approximately the size of a large paperback novel) has built in displays and buttons or switches to set start times, time intervals between measure- ments and to scan through, using the display, the temperatures, that have been measured. Instruments can be purchased with between 2 and 16 plug- in temperature sensors, and some will display the maximum, minimum and mean temperature recorded by a particular sensor. For more detailed analy- sis of the temperatures, the recorded data are transferred to a personal com- puter. The PC is required to program the start time, recording interval and so on, and analyse the data with the smaller loggers. These instruments usually have a maximum capacity of 8 temperature sensors, 1 of which is often built into the instrument. Developments in computer storage chips are continually extending the number of temperature values that can be held in both types of logger. Modern instruments would typically be able to take readings of 4 tem- perature sensors at 5 min intervals over a 2–3-week period. Further developments in electronics are extending the temperature range over which the instruments will operate. Some instruments will record accu- rately inside blast and spiral freezing systems whilst others can operate in ambient temperatures up to 70–80 °C. For extended use at sub-zero tem- peratures special batteries are required. Logging systems have been devel- oped which use insulated heat resistant cases to allow operation for several hours at temperatures up to 300 °C. This allows measurement of product and processing temperatures in batch and continuous baking/cooking operations. Temperature measurement 287 Most data logging systems that will measure over 20 temperatures are physically too big to be considered as truly portable even though some can be battery powered. Some have a built in display and keyboard but the majority are operated using a video display unit. A basic system would consist of: ? a number of input cards to which the temperature sensors are connected, ? a card-based voltmeter to measure the output from sensors when instructed, ? a microcomputer to provide the instructions and convert voltages into temperature measurements, ? a storage system that could be floppy or hard disc, and a video display unit. Many systems can be expanded to hundreds and in some cases thousands of temperature sensors by the addition of extra input cards, some of which can be up to a mile away from the central system. The temperature measurement possibilities of large logging systems are only limited by the ingenuity of the programmer/operator. Different com- binations of temperature sensors can be monitored at varying time inter- vals and the data displayed, analysed, used to control processes or set off alarms or be transmitted to central control rooms hundreds of miles away. All three types of temperature sensor – thermocouple, thermistor and platinum resistance – are commonly used for multi-point temperature measurement. Thermocouples are cheap, especially if the wire is purchased in bulk, and very small sensors can be manufactured. Thermistors are more expensive, slightly larger but more accurate over limited temperature ranges. Platinum resistance sensors are typically 2–3 times the cost of ther- mistors, but are capable of better than 0.1 °C accuracy. Thin wire and thin film platinum resistance sensors can be very small. Commercial sensors are often enclosed in stainless steel sheaths, which makes them more robust, but increases their response time. 14.1.3 Time–temperature indicators There are many different types of temperature or time–temperature indi- cators. Almost anything that undergoes a sensibly detectable change with temperature can be used. Liquid crystal devices change colour to indicate the temperature at the time they are observed and time–temperature indi- cators change irreversibly after a time dependent upon the temperature history or when a temperature threshold is exceeded. Temperature indicators are already used as cheap, safe and hygienic ther- mometers in the food chain. Several types have been developed to the point where they have been introduced on some chilled and frozen foods in the USA and on chilled foods in France. 288 Meat refrigeration 14.2 Calibration Any temperature measuring system should be tested over the operating range at regular intervals to ensure accuracy and should also have a current calibration certificate from its manufacturer or official standards laboratory. The system can be checked by means of a calibration instrument, or against a reference thermometer that is known to be accurate. Melting ice (which if made from distilled water should read 0°C, or -0.06 °C if made from tap water with 0.1% salt) may be used to check sensor accuracy. The ice should be broken up into small pieces and placed in a wide-necked vacuum flask with a depth of more than 50 mm. The system should be agitated frequently and the temperature read after a few minutes when stable. If differences of more than 0.5 °C are found, the instrument should either be very carefully adjusted or sent for calibration. Other simple calibration systems are available. These consist of a small stirred tank that can be filled with water or oil. The temperature of the stirred liquid is measured using a standard calibrated platinum resistance thermometer. The temperature sensors to be calibrated are placed in the liquid and compared with the standard measurement. The temperature of the liquid can be raised or lowered to different values by the addition of ice, cold liquid or hot liquid. 14.3 Measuring temperature data Accurately determining the temperature of chilled meat throughout the cold chain is difficult. Training and experience are required to locate posi- tions of maximum and minimum temperature in abattoirs, stores, vehicles and display cabinets.The problem is further exaggerated by changes in posi- tion with time caused by loading patterns and the cycling of the refrigera- tion plants. Obtaining a relationship between environmental temperatures (that can be measured relatively easily) and internal meat temperatures is not a simple process. Relating temperatures obtained in a non-destructive manner with internal meat temperatures again poses problems. Determin- ing the temperature of cuts of meat with regular shapes is quite simple but doing so for irregular cuts of meat is more difficult. All the temperature measurement problems associated with chill foods will equally apply to quick-frozen foods. In addition, there are a number of other problems. Many instruments have sensors that will accurately measure temperatures of -20 °C and below, but the instruments themselves become inaccurate or fail to operate at low temperatures. If frozen foods are removed from their low temperature environment to one suitable for the instrument the surface temperature rises very rapidly. However, the main problem is that of actually inserting a temperature sensor into frozen meat. Temperature measurement 289 14.3.1 Contact non-destructive methods The surface temperature of a food or pack can be measured by placing a temperature sensor (such as those discussed above) in contact with the surface. In practice there are very large temperature gradients on both sides of the surface and the presence of the sensor can influence the temperature being measured. Extending the surface of the sensor to measure the average temperature over a larger surface area is one method used to minimise these problems. This method is recommended for such applications as between-pack measurement. Since it is impossible to measure the temperature of an exposed surface accurately, the next best thing is to take a measurement of the temperature between two food items. As long as good thermal contact is achieved between the temperature sensor and the packs, a between-pack method should provide an accurate measurement of the pack temperature. If the thermal conductivity of the packaging material is high and the food makes a good thermal contact with the pack then the temperature measured will be close to that of the product. With a product such as skin-wrapped chilled sausages the above require- ments are satisfied. A temperature sensor, especially a flat-headed probe, can be sandwiched between two packs. An accurate measurement is obtained owing to the combination of a flexible food and a thin wrapping. With chilled food in cartons or bubble packs the accuracy is much lower. The contact problems are much greater with a frozen product. Since the surface of a frozen product is not flexible, only point contact can be achieved between the surface of the product and that of the pack or probe. Using a flat probe with extended contact surfaces does not necessarily improve the accuracy of temperature measurement. In extreme cases, for example with frozen sausages, the contact surfaces may extend out into the air stream and measure air, not product temperature. With packs of small items such as diced meat the accuracy will be much better. Care must also be taken to precool the probe before temperatures are measured. This is especially important with low heat capacity packaging materials. Alternatively ‘temperature sensitive’ paints can be painted directly onto the surface of interest and will accurately determine its temperature. However, painting foods is not a practical solution. 14.3.2 Non-contact non-destructive methods Non-contact temperature measurement devices measure the amount of energy in an area of the infrared spectrum that is radiated from the surface being measured. Basic instruments measure the average temperature of the area in a small field of view. More complicated systems of thermal imaging provide a temperature picture of all the objects over a much wider area. There are two types of detector currently used in low temperature infrared thermometers, thermopile detectors and pyroelectric detectors. 290 Meat refrigeration Thermopile detectors consist of a collection of rods that act as thermocou- ples to sense emitted thermal radiation. Pyroelectric detectors contain a crystal which exhibits temperature-dependent polarisation and requires the incident radiation to be ‘cut’ by a ‘chopping device’ to prevent currents building up within the crystal that nullify this charge. A certain amount of knowledge is needed in order to interpret the values that such instruments give (Evans et al., 1994; James and Evans, 1994). The first point to bear in mind when using infrared thermometry is that the tem- perature measured is the surface temperature. If the meat has been in surroundings that have not changed in temperature for a long period of time, then it is likely that the surface temperature will be very close to that of the meat beneath the surface. However, if the temperature of the sur- roundings is changing or has changed over the previous 24 h, then it is likely that the surface temperature will not be the same as the temperature deep within the meat. For example, transfer a carton of frozen meat from a refrigerated lorry where the temperature has been maintained at -18 °C to a refrigerated loading bay at 4 °C, and the surface of the carton will warm very rapidly. Therefore, if an infrared thermometer is to be used to check the meat tem- perature the value must be taken before it is removed from the lorry if any degree of accuracy is to be obtained. Even temperature fluctuations of ±2 °C caused by normal control fluctuations will mean that the surface tem- perature will differ significantly from the deep meat temperature. If the meat is still cooling down, its surface temperature will be warmer than the air temperature and, in turn, the interior of the meat will be warmer still. These problems can be overcome if the operator is aware of them. For example, the temperature of the carton of frozen meat could be measured in the lorry, or if removed the carton could be opened and the temperature of an inner surface of one of the packs inside immediately read with the infrared thermometer. However, in doing this two of the principle advan- tages of using infrared have been removed: namely, being quick and totally non-destructive. It is also necessary for the operator to know how much of the surface is ‘seen’ by the infrared instrument, as it will measure the ‘average’ tempera- ture over the whole of this area. The target area can vary significantly from instrument to instrument and with the distance between the instrument and the surface. A further complication in the use of infrared thermometers is reflected radiation. The instruments will ‘see’ the radiation emitted from a surface and also an amount of radiation from the surroundings that is reflected by that surface. The reflected radiation will therefore constitute an error. For warm objects at a temperature greater than their surroundings, the amount of reflected radiation will be small in relation to that from the surface and consequentially the error will be small. With frozen meat, the temperature of the meat is no warmer than, and often colder than, the temperature of Temperature measurement 291 the surroundings. Therefore, the amount of reflected radiation coming from the surface constitutes a significant error. However, the proportion of radiation emitted by a surface relative to that of a perfect black body is the same proportion of incident radiation that would be absorbed by the surface. If the absorbtivity or emissivity of the surface is known and the surface is not transparent, which is true for all packaging materials except some plastics, the reflectance of the surface will also be known by subtracting the absorbtivity from 1. Hence it is possible to calculate the extent of the error. Unfortunately this requires a lot of information about the emissivity and reflectance of the various packaging materials and takes a long time. There- fore, the advantage of taking quick accurate readings will be lost. The Meat Research Corporation in Australia (1995a) recommend that an infrared thermometer should be permanently located in beef chill rooms. It should be positioned to measure the surface temperature of one of the last sides to be loaded. It can then be used to provide a permanently logged output of surface temperature and control the refrigeration system. When the surface temperature has reached ca. 2 °C the fan speed can be auto- matically reduced and the suction pressure raised. This will reduce both weight loss and operating costs. Infrared systems have also been used to monitor the surface tempera- ture of pig carcasses in chill rooms (Metternick-Jones and Skevington, 1992). The thermometer required a minimum stabilisation time of 120min in the chill room. After this time the temperatures measured were within 1 °C of other methods and were repeatable to the same accuracy. 14.3.3 Contact destructive methods Determining the temperature of small cuts of meat with regular shapes is quite simple. Determining the temperature of irregular cuts of meat, par- ticularly large pieces, is more difficult. Possibly the most difficult problem is ascertaining deep leg temperature in beef carcasses. The Meat Research Corporation in Australian (1995b) recommend that the temperature sensor should touch the trochanter major (aitch bone), which is the ‘knob’ of bone on the opposite side of the femur to the hip joint. To locate the sensor in this position it should be inserted through the ‘pope’s eye’ at an angle about 15–20° below the horizontal (Fig 14.1). It should be aimed at an imaginary vertical line approximately one-third of the distance from the Achilles tendon to the last tailbone. Because conduction occurs along the steel shaft of a probe it is impor- tant that the probe is inserted as far as possible into the meat. For example, to take the temperature of a cut of meat it is better practice to insert the probe to its full depth along the long axis of the cut rather than to insert the probe to half its length through the short axis (CSIRO, 1991). For manufacturers to produce accurate data on the freezing of their meat 292 Meat refrigeration or meat products is a straightforward if time-consuming process. Tempera- ture sensors have to be inserted and positioned close to the thermal centre prior to freezing. The samples are then packaged with the sensor leads sealed into the packs. If further information is required then additional sensors can be inserted close to the surface and at other positions of inter- est. With packs of small individual items care must be taken to maintain the position of the sensor within the pack when it is moved. Fixed or portable data loggers are then used to record temperatures during the freezing process, and throughout distribution if required, and the resulting data are analysed. Routine determination of the temperature of frozen foods in the cold chain is a much more difficult process. Making a hole 2.5 cm or 3–4 times the probe diameter deep and then inserting a temperature probe sounds easy, but in practice is much more complicated. There are three basic methods of making the hole: (1) forcing in a sharp- ened pointed instrument, (2) screwing in an auger-type bit or (3) drilling a hole. All cause problems with different types of frozen produce. Forcing a pointed instrument into frozen food at -18 °C requires considerable force and the product has to be secured on a firm base to stop it slipping and to provide the resistance required. Fragile products such as beefburgers, pies, and so on will readily shatter when subjected to such treatment. An auger is better but the product still has to be restrained. Holding down a small Temperature measurement 293 Fig. 14.1 Recommended position for taking deep leg temperatures in a chilled beef carcass (source: Meat Research Corporation, 1995b). food pack by hand introduces a substantial amount of heat into the product. The choice of auger size is also a compromise. If the auger is too big it will shatter fragile products and produce a hole too big for the probe. If it is too small it can bend or even worse the end may break off in the frozen product. Using a portable electric drill is probably the best compromise. However, allowance must be made for the small temperature rise caused by the oper- ation and in some products the food will clog the flutes of the bit and the tip may shear off in the product. The meat swarf produced by the drilling needs to be carefully collected to avoid it becoming a source of contamination. 14.3.4 Storage 14.3.4.1 Cold store Research investigations have shown that it is very difficult to locate the warmest positions within a refrigerated space unless large arrays of tem- perature sensors are positioned within the space. This is not possible in a commercial operation. A UK code of practice (Department of Health, 1991) provides some guidance about the number and position of sensors in different situations. It recommends that the number of sensors should range from 2 for a 500–5000 m 3 cold store to 6 for a cold store of over 8500m 3 . The positions used in descending order of importance should be: at the maximum height of the food at the furthest position away from the cooler fans, or in the air return to the evaporator; along the walls at two thirds the height of the room away from the doors and not directly in the path of the air outlets from the evaporator; positioned 2 m above the floor level directly opposite the evaporator. The code also recommends monitoring the air from and returning to the evaporator coil. The above are good general recommendations but they cannot be guar- anteed to locate the warmest food positions. In poorly designed or badly loaded cold stores, air movement can be effectively non-existent around parts of the load. In a small number of situations heat infiltration through the fabric or doors can cause local warm areas. Infrared radiation thermometers can have a useful role to play in locat- ing areas of high temperature within cold stores. In cold stores the walls, ceilings, floors and the products are at similar temperatures. This minimises the problems of radiant energy from warmer surfaces being reflected by the surface of the body being measured and producing erroneous results. The only warm surface likely to produce problems is that of the person making the measurements. Errors caused by differences in emissivity between packaging materials are also reduced because most bulk packs tend to be wrapped in similar materials. When checking the temperature data from air sensors, considering just the mean and the range of the temperatures is not sufficient. Air tempera- 294 Meat refrigeration tures cycling up and down over a few minutes will have little if any influ- ence on the temperature of the food being stored. Long cycles of the same amplitude are much more likely to reduce product quality and result in tem- peratures above the legal minimum. 14.3.4.2 Chilled store When monitoring air temperature in a chilled store, the best position is not by the door, (where the temperature will go up and down like a yo-yo as the door opens and closes), but in the air returning to the cooler. More information can be obtained on the functioning of the cold store if the air leaving the cooler is also measured (Fig. 14.2). Apart from the normal control cycles, the air temperature rises dramati- cally as the door is opened and when the evaporator is defrosted. If the defrosts occur at reasonable intervals and the door discipline is good, the temperature will fall very quickly after each occasion back to its normal control temperature of 0°C. In most coolers there is also a difference between temperatures during the day, when the store is in use, and tem- peratures at night and at the weekend, when the store is not. If these are substantial it indicates that the meat being put into the chiller is too warm, the amount of time that the door is open is too long or the capacity of the cooler is not large enough. 14.3.5 Distribution Monitoring representative temperatures within a distribution vehicle is a more difficult operation. The vehicles are designed to maintain the tem- perature of precooled loads and cater for a small amount of heat infiltra- tion. In the code of practice it is suggested that the differential between the Temperature measurement 295 Defrost Door openings Defrost Overloaded 0 6 12 18 24 Time (h) –5 Night > 5 15 T emper ature ( ° C) Chilled store Air on Air off Fig. 14.2 Air temperatures in chill store. air entering and leaving the evaporator is indicative of the performance of the refrigeration system. It recommends installing two sensors in the vehicle, one measuring the return air and the other on the ceiling at a posi- tion approximately three quarters of the length down the vehicle. In vehi- cles that are not fitted with forced air systems temperatures should be taken from above and below the load. In practice, products near the rear door are often the warmest owing to heat pick up during loading and infiltration through doors and their seals. Temperature sensors positioned close to the doors will often provide a better indication of any temperature abuse. However, care must be taken when examining the temperature data to allow for rapid temperature rises during loading and unloading periods when the doors are open. 14.3.6 Retail Most legislative requirements state that the temperature must be controlled in the centre of the meat to below a prescribed maximum. However, for quality control the temperature at the surface must also be controlled and this is often much more difficult. It is obviously impractical to measure the temperature of every piece of meat all the time. Therefore a system is needed to monitor how well things are doing from day to day without the need to take a large number of measurements. One way of doing this is to control the use of cabinets and chill rooms and to monitor the temperature of the air in them. If this is done then a large number of meat temperatures only need to be measured once. The temperature of the meat can then be related to the monitored air temperature, and thus monitored air temper- atures in the future can be related to those of the meat. A thermometer is then only needed to check meat temperatures on delivery and if unusual circumstances occur, as they will do from time to time. 14.3.6.1 Cutting areas In many countries there are no requirements for controlling temperatures in cutting areas in butchers’ shops. However, it would be as well for these to be as low as possible and the area to be arranged so that meat can be quickly brought from the chiller, cut, and returned to either the chiller, or the display case, within the shortest possible time. When temperature mon- itoring is initially introduced it is a good idea to measure meat tempera- tures at the centre and at the surface before and after cutting to learn just how far temperatures rise during this process. 14.3.6.2 Retail display (chilled) During retail display air temperatures should be measured entering and leaving the case. In the case of fan-assisted coolers this is where the air enters the display case at the top of the rear of the cabinet and returns back to the coolers through the return-air grill at the front of the cabinet. Air 296 Meat refrigeration temperature will fluctuate within wide limits during normal use of the display case and one-off measurements are of little use. Instead, the sensor of the thermometer should be inserted in the air stream and the tempera- ture observed over a period of time when its rise and fall can both be observed.The maximum and minimum temperatures can then be recorded, or an average temperature if that facility is included in the thermometer. An alternative way of obtaining an average temperature is to use a ther- mometer that is large and heavy and which consequently has a very slow thermal response. This can be inserted into the air stream and left to equalise over a long period of time, possibly over an hour. The temperature can then be recorded and will give a reasonable indication of the average air temperature at that point. The monitored air temperatures would ideally be recorded at intervals automatically on a data logging system, but the cost of this can only be jus- tified in larger shops. The alternative is to keep records measuring the tem- perature at the same time each day and preferably twice during the day and recording these on a data sheet. These data can be used to check whether the equipment is adequate for the uses to which it is put. 14.3.6.3 Retail display (frozen) Display cases are much more difficult to maintain at frozen temperatures because of the much greater temperature difference between the display and store temperatures. The conditions that apply to chilled display cases remain applicable to frozen display cases but with greater effect. Retail display cabinets for frozen food are likely to present the weakest link in the temperature chain. Individual retail packs of food will react more quickly to temperature changes than similar products in bulk packs. The packs are of necessity more exposed to outside ambient influences than in any other part of the cold chain. If they cannot be seen and easily handled by the consumer then they will not sell. Retail display cabinets are designed to maintain the temperature of pre- cooled products and have a very limited capacity to extract heat from inad- equately cooled food. Studies of chilled display cabinets have shown that it is very difficult to locate the position of the warmest areas in the cabinets. The same applies to display cabinets for frozen foods. The Department of Health 1991 guideline suggests that monitoring air- on and air-off the evaporator coil will provide an indication of the perfor- mance of the cabinet. Monitoring temperatures at these positions will indicate if there are any changes in the performance of the cabinet. However, FRPERC confidential studies have shown that it is possible for all these temperatures to be below -12 °C and still find warmer product in some areas of the cabinets. Currently using large arrays of temperature sensors to obtain a rela- tionship between the meat temperatures and those in the monitored posi- tions is the only certain option available. Even then, monitoring the array Temperature measurement 297 must be carried out over a complete operational cycle of the cabinet. Reflections and emissivity problems produce large errors in infrared tem- perature measurement in stores. In general, in horizontal cabinets the warmest packs will be those at the surface in the centre of the display area. In forced air circulation cabinets the warmest packs will be in the top layer nearest to the air returning to the coil. However, the position of warm spots will vary with the perfor- mance of the individual cabinet and its position in the retail store. One cost-effective method of locating the position of the warmest packs can be to use thermochromic temperature indicators. However, FRPERC investigations with chill cabinets have shown that obtaining a good thermal contact between the strip and the pack is critical if sensible results are to be obtained. It is too easy to measure the air, not the pack temperature, if a good contact is not obtained. With frozen product good contact is even more important and any ice formation on the pack will exaggerate the problem. 14.4 Interpreting temperature data With modern equipment it is quite easy to gather data on the temperature history of many points both in and around a product. Producing firm con- clusions from those data is not so easy and requires a clear understanding of the process being monitored and the reason for the monitoring. The cer- tainty of the conclusions will depend upon the number of temperatures measured, their position and their accuracy. Each monitoring task has its own particular problems and requirements, but the following examples illustrate a number of common problems and general methods of solution. 14.4.1 Example 1 EU export regulations for red meat require that the maximum temperature within a carcass side is reduced to 7 °C before it is butchered or transported. Locating the maximum temperature in a side of beef is difficult, but as pre- viously discussed it is normally found in the deep leg near the aitch bone (trochanter major). Taking measurements at different depths using a single or multi-point sensor is the best available method. Table 14.3 shows typical results from two different beef sides. In both sides some of the temperatures are above 7 °C, even allowing for an error of ±0.5 °C in the values, so the meat is not chilled sufficiently to meet the regulations. However, far more useful information becomes appar- ent if the data are plotted (Fig. 14.3). In carcass 1 only a small volume of the meat, ca. 6 cm in diameter, is above 7°C and the surface is close to 0 °C. If the beef side remained in the chiller for a short time longer, probably less 298 Meat refrigeration than half an hour, it would achieve the legislative requirements. In the second side, approximately half the meat is above 7 °C and the surface tem- perature has only been reduced to 6°C. Many hours of further cooling would be required to cool the meat fully. Both sides had been in their respective chill rooms for over 30 h when the measurements were taken. A surface temperature of 6°C in the second side points towards a high air temperature in the chill room as the most likely cause of the poor performance. In most cases a logical analysis of such data will provide the clues required to solve refrigeration problems. 14.4.2 Example 2 The importance of interpretation is demonstrated by analysing data obtained from the same relative position in two different chilled food display cabinets. It is difficult to obtain any real impression of cabinet performance from the raw temperature data (Table 14.4). A basic analysis of the data in the Temperature measurement 299 Table 14.3 Temperatures (°C) measured after chilling at different depth in two beef sides Depth (cm) 024681012141618 Carcass 1 0.2 2.2 3.7 5.0 6.1 7.0 7.6 8.0 7.4 6.0 Carcass 2 6.0 6.5 6.9 7.3 7.6 7.9 8.1 8.2 8.3 8.0 0 10 20 Depth from surface (cm) Carcass 2 Carcass 1 0 2 4 6 8 10 T emper ature ( ° C) Fig. 14.3 Temperature at different depths from the surface of 2 beef carcasses. table reduces each set of 66 measurements to the 4 shown in Table 14.5. Statistically there is no difference between temperatures measured in the 2 cabinets because the difference between the 2 means is far smaller than the standard deviation, which is a measure of the scatter within the mea- surements. However, intuitively most people would consider the first cabinet to be slightly better than the second because it has a lower mean and maximum temperature, and the minimum temperature would just avoid surface freezing of most foods. Only when the data are plotted does the true significance of the data begin to appear (Fig. 14.4). The temperature in the first cabinet is regularly 300 Meat refrigeration Table 14.4 Temperatures (°C) measured in the same positions and at the same time intervals in 2 retail display cabinets Cabinet 1 Cabinet 2 Cabinet 1 Cabinet 2 Cabinet 1 Cabinet 2 0.2 4.2 0.0 4.2 0.0 4.3 1.0 4.4 1.0 4.4 1.0 4.4 2.1 6.0 2.1 4.7 2.1 4.7 2.4 13.0 2.6 4.7 2.4 4.7 3.7 14.2 3.6 4.8 3.6 7.2 6.6 7.0 6.6 4.5 6.8 12.3 7.2 0.2 7.2 4.2 7.2 14.5 8.3 3.9 8.3 3.9 8.3 5.0 5.4 4.0 5.4 4.0 5.4 -0.1 2.3 4.1 2.2 4.2 2.2 4.1 -1.2 4.2 -1.2 4.2 -1.4 4.2 0.0 4.2 0.0 4.2 0.0 4.2 1.5 4.4 1.0 4.4 1.0 4.4 2.1 4.7 2.1 4.7 2.1 4.7 2.4 4.7 2.4 4.7 2.4 4.7 3.6 4.8 3.6 4.8 3.6 4.8 6.6 4.5 6.6 4.5 6.6 4.5 6.9 4.2 7.2 4.2 7.2 4.2 8.3 3.9 8.3 3.9 8.3 3.9 5.4 4.0 5.4 4.0 5.4 4.0 10.0 4.1 2.2 4.1 2.4 4.1 5.6 4.2 -1.2 4.2 -1.2 4.2 Table 14.5 Maximum, minimum, mean and standard deviation of temperatures (°C) in two retail display cabinets Maximum Minimum Mean Standard deviation Cabinet 1 10.0 -1.4 3.7 2.99 Cabinet 2 14.5 -0.1 4.9 2.40 cycling by ca. 9 °C from -1 to +8 °C, whilst the second cabinet spends the majority of its time at 4.4 ± 0.5 °C. The two temperature peaks probably indicate defrost periods that could be considerably reduced, if not elimi- nated, by correct adjustment. If further measurements revealed the same temperature pattern and degree of control throughout the cabinet, then the control setting could be adjusted to produce a maximum temperature that would substantially reduce the growth of pathogenic organisms. 14.5 Conclusions 1 While it may be literally true that ‘anyone can measure a temperature’ the meat industry needs to measure temperatures accurately, reliably, meaningfully, simply and cheaply. 2 It needs to be able to analyse the data and respond when required. 3 It needs the correct instrumentation and the expertise to collect and interpret the temperature data. 4 There is an increasing range of both simple and more sophisticated multi-point temperature measurement devices available. Monitoring the temperature of meat and meat products is therefore becoming less of a problem. 5 Deciding which temperatures are critical to the overall safety of the product and process, or will be needed to meet legislative requirements, is a more formidable process and requires training and expertise. 6 Producing firm conclusions from the data obtained is even more diffi- cult and demands a clear understanding of the process and the reason for monitoring. Temperature measurement 301 Cabinet 1 Cabinet 2 0 10 20 30 40 50 60 Time (h) –2.5 2.5 7.5 12.5 T emper ature ( ° C) Fig. 14.4 Temperature history at points in same relative positions in 2 cabinets. 14.6 References CSIRO (1991), Thermometers, Meat Research News Letter, 91/2, CSIRO Division of Food Processing, Meat Research Laboratory. Department of Health (1991), Guidelines on the Food Hygiene (amendment) Reg- ulations 1990, HMSO. evans j a, russell s l and james s j (1994), An evaluation of infrared non-contact thermometers for food use, Developments in Food Science 36, Oxford, Elsevier Science, 43–50. james s j and evans j a (1994), The accuracy of non-contact temperature measure- ment of chilled and frozen food, IChemE Food Engineering Symposium, Uni- versity of Bath 19–21/9/94. Meat Research Corporation (1995a), Measurement of surface temperatures of chilled carcasses and sides, Meat Technology Update, 95/6, Australian Meat Tech- nology Meat Research Newsletter. Meat Research Corporation (1995b), Measurement of temperatures in fresh and processed meats, Meat Technology Update, 95/1, Australian Meat Technology Meat Research Newsletter. metternick-jones s l and skevington s g (1992), Evaluation of non-contact infrared thermometry for measuring the temperature of pig carcasses in chillers, Meat Sci, 32 1–9. 302 Meat refrigeration