Food Safety System for Food Items in Cooled Environments

Abstract

A food safety system for food items in cooled environments includes a temperature sensor unit having a temperature sensor, a power supply and a data transmission element. The temperature sensor unit is positioned in a cooler, wherein the cooler has a plurality of predefined food item positions. A control center unit having a computer processor and a memory is adapted to execute a deterministic mode function to predict the core temperature change of such a food item on a predefined food item position in said cooler. The deterministic mode function depends on heat transfer parameters related to the predefined food item position of the cooler used, food specific coefficients related to the kind of food item taken from a group of food types, the environment temperature measured by the temperature sensor, and the predicted current core temperature of the food item.

Claims

1. A food safety system for food items in cooled environments in a cooler, the food safety system comprising: a temperature sensor unit comprising: a temperature sensor, a power supply, a data transmission element and a control unit, and a control center unit having: a computer processor and a memory, wherein the temperature sensor unit is configured to be positioned at a predetermined position in the cooler having a plurality of predefined food item positions, wherein the control center unit is adapted to execute a deterministic mode function predicting a core temperature change of a food item on a predefined food item position in the cooler according to a following equation: $\frac{dTc\left(t\right)}{dt}=f\left(h,Q,\mathrm{Te},Tc\right)$ wherein Tc is the predicted current core temperature of the food item, f is the deterministic mode function, h are cooler specific heat transfer parameters, Q are food specific coefficients related to a kind of food item taken from a group of food types, and Te is an environment temperature measured by the temperature sensor, wherein the deterministic mode function depends on cooler specific heat transfer parameters related to the predefined food item position in the cooler used, food specific coefficients related to the kind of food item taken from a group of food types, the environment temperature measured by the temperature sensor and the predicted current core temperature of the food item.

2. The food safety system according to claim 1, wherein white noise is added to the deterministic mode function.

3. The food safety system according to claim 1, wherein the environment temperature is measured in predetermined time intervals.

4. The food safety system according to claim 3, wherein the predetermined time interval is a regular interval.

5. The food safety system according to claim 4, wherein the predetermined time interval is between one and ten minutes.

6. The food safety system according to claim 1, wherein the temperature sensor unit is connected with the control center via a Long Range Wide Power Network.

7. The food safety system according to claim 1, wherein the control center is connected to an alarm server to automatically call a predefined electronic communication device being present at premises of a monitored cooler transmitting information of an incident at the monitored cooler in question identified by the related temperature sensor unit.

8. The food safety system according to claim 1, wherein memory data in the control center is related to at least one type of cooler from the group encompassing one or more types of an open cooler, a closed cooler having a lid, a closed cold storage room, a cooler of a refrigerated car, a tray, or a coolbox.

9. The food safety system according to claim 1, wherein the food types are taken from a group including meat, fish, fluid dairy products, solid dairy products, canned food, and solid convenience products.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Preferred embodiments of the invention are described in the following with reference to the drawings, which are for the purpose of illustrating the present preferred embodiments of the invention and not for the purpose of limiting the same. In the drawings,

[0035] FIG. 1 shows a schematic perspective view of a first vertical cooler type to be monitored with a safety system according to the invention,

[0036] FIG. 2 shows a schematic perspective view of a closable second horizontal cooler type to be monitored with a safety system according to the invention,

[0037] FIG. 3 shows predicted temperature curves compared to manual temperature checks provided in the framework of validation of the safety system according to the invention;

[0038] FIG. 4 shows a schematic view of a temperature sensor unit to be used in connection with the invention;

[0039] FIG. 5 shows a schematic view of a food safety system according to the invention together with a plurality of coolers to be monitored.

DESCRIPTION OF THE INVENTION

[0040] FIG. 1 shows a schematic perspective view of a first vertical cooler type 10 to be monitored with a safety system according to the invention. The first vertical cooler type 10 comprises a mainly vertically oriented open front access surface 11 with a bottom access surface 12 to be reached by a customer from above. The drawing shows as an example specimen of perishable cooled food items 13 as positioned in trays in the bottom part of the cooler 10, on shelves within the open room of the cooler or which are positioned on hangers. The first vertical cooler type 10 has an inner cooled space portion 15 which is virtually separated from the environment of the selling space 16, i.e. the environment by the plane defined by the front edges 17 of the cooler 10. Usually air convection is generated to maintain the cold air within the storage part of the cooler and thus inside of the edges 17 of the cooler According to the invention, one temperature sensor unit 20 is positioned in the cooled space portion 15 at a predetermined position which position will be explained in the following description. It is preferred that this temperature sensor unit 20 has a similar maximum distance to any food item 13 to be positioned in the cooler 10, so it is often put in the middle (left to right) of the cooler 10 and preferred in the upper part, e.g. under the top cover plate 21 of the cooler unit 10.

[0041] In contrast to FIG. 1, FIG. 2 shows a schematic perspective view of a closable second horizontal cooler type 10′. Here the front edges 17 of FIG. 1, defining the plane enclosing the cooler air behind it, are replaced by horizontal abutments and guides 17′ around the upper surface of the cooler 10′, wherein the separation between the inner cooled space portion 15 from the selling environment 16 is provided by a slidable door 18 which moves essentially horizontally in the direction of the double arrow 19. The products 13 as well as the temperature sensor unit 20 are positioned inside the cooler 10′, wherein the temperature sensor unit 20 is preferably positioned in the middle (left to right) at the inside of the back wall of the cooler unit 10′. There may also be two sliding doors 18 covering each 50% of the edge surface to close the cooler 10′ if no food item 13 has to selected; or there might only one door 18 so that one half of the opening between the edges 17 is always open.

[0042] According to the invention one temperature sensor unit 20 is sufficient to be positioned at a predetermined position in any one cooler 10 or 10′. The temperature sensor unit 20 is shown in detail in connection with FIG. 4. The temperature sensor 25 built in the temperature sensor unit 20 allows real-time monitoring of core temperatures of various food products according to the following method of using the food safety system.

[0043] Before the use of the food safety system on site, e.g. in a structure as shown in FIG. 5, within the cooling process the core temperature is mathematically modeled, followed by a data-driven model parameter calibration which allows the model validation through comparison of the model prediction with measured real-world temperature observations through running tests and check measured core temperatures.

[0044] The model is simulated bottom up with known material properties and convection influence, in other words, the model is adapted to every different cooler type 10, 10′ in use. The model is based on a lumped parameter model, also called lumper element model, which simplifies the description of the behavior of spatially distributed physical systems into a topology consisting of discrete entities that approximate the behavior of the distributed system under certain assumptions which are within the parameter of the product groups and the usual temperature(s) of the environment 16. The temperature field is calculated based on ambient temperature only. The model is calibrated on measurement data. The temperature prediction relies only on ambient temperature, i.e. temperature of air in the cooler.

[0045] The model is based on a selection of a few state variables and parameters, especially One set of variables is determined based on the cooler 10, 10′; such variables determine the dependency of the environment temperature at any storage point of the cooler 10, 10′, e.g. if the food item to be remotely monitored is near the temperature sensor unit; at the bottom of the fridge or at a hanger in the above left corner. The second set of variables is determined based on the food to be stored; a liquid dairy product has a different temperature changing curve compared to a loose salad under a cellophane cover. All other processes are included in the model as noise.

[0046] It is known in the art to use stochastic differential equation (SDE) to model various phenomena such as unstable stock prices or—as it is done here—physical systems subject to thermal fluctuations. A SDE is a differential equation in which one or more of the terms is a stochastic process, resulting in a solution which is also a stochastic process.

[0047] Although it is preferred to use SDEs, it is possible to apply simple differential equations with higher temperature error estimation.

[0048] The model, e.g. based on SDEs, is then checked with calibrated available data as data-driven parameter inference. Based on a time series analysis with prediction the core temperature of product groups at specific places in the cooler is predicted. The forward model is used to make probabilistic predictions which can then allow the model validation and calibration. Model validation as such is not necessary for the function of the safety system at stake, but proves that the predicted core temperature of any food item from the list of predictable food items at any predetermined place in the cooler is in line with a manually measured core temperature which is usually the official check of compliance with regulations and would be applied in case an official of footstock control would appear and make his own check.

[0049] This model validation occurs for all types of coolers 10, 10′ (and 130) separately, so as do use the specific lumped sum parameter. Beside cooler types 10, 10′ the model also depends on different type of food products, which are divided within the model calibration in a predetermined number of food groups, currently taken from the group encompassing meat, fish, fluid dairy products (e.g., milk, yoghurt), solid dairy products (e.g., cheese), canned food and finally solid convenience products (e.g., sandwiches). It is possible to split them into further groups, e.g. if solid dairy products are transported form a logistics center 120 to a retail store 110 with a lorry 130, then “the” product is usually a packaged group of e.g. 4 times 4 times 5 solid dairy product packages, so that the core of this packaged group is different to the display in a retail store.

[0050] The stochastic model is based on the addition of a deterministic model function f, depending on heat transfer:

[00003]$\frac{dTc\left(t\right)}{dt}=f\left(h,Q,\mathrm{Te},\mathrm{Tc}\right)+\eta \left(t\right)$

[0051] Therein is dTc/dt the rate of core temperature change, f the deterministic model function based on cooler specific heat transfer h, food specific coefficient Q, the environment temperature Te and the core temperature Tc. η is the stochastic noise describing unpredictable random events that can perturb the system. It is noted that only Q is related to the food item as part of one of the above mentioned groups and that only h is related to the cooler type (i.e. 10, 10′ or 130) beside of course the environment temperature Te to be taken inside the cooler at a predetermined place. But the environment temperature Te is a measured value determined by the temperature sensor unit and not a predetermined value depending on the cooler. The heat transfer is influenced by the environment temperature Te if different the parameters for one cooler are established based on different placements of the temperature sensor unit. The dependencies of the different parameters of function f on the cooler to be monitored (=cl), food position in such a cooler (=fp) and food type provided at this position (=ft) are shown in the general description part of the present specification. Here, the formula is applied based on data relating to one specific cooler and arrangement of food and food types as shown in FIG. 1 or FIG. 2.

[0052] The model calibration is performed with the assumption of constant white noise and a normal distribution of core temperatures.

[0053] Parameter inference can be carried out using the Markov Chain Monte Carlo (MCMC) method called EMCEE proposed by Goodman and Weare published in Comm. App. Math. And Comp. Sci., 5, 65 (2010).

[0054] As mentioned above, the model separates the parameter for the cooler from the parameter of the food groups. As a consequence: The method works with a single temperature sensor in the temperature sensor unit 20 at the center of the cooler. Center of the cooler is usually the point of the cooler having fastest impact on disturbances of the system and having the least greatest distance to food items stored in the cooler. In other words; a sensor positioned in the center of the cooler has equal distance to the left side of the cooler as to the right side and usually, if it is positioned at middle height, also a similar distance to a right high corner as to a lower left corner. Usually a sensor unit 20 can be used to predict the core temperatures in a distance of up to 3 meter, preferably used for a radius of up to 2 meter.

[0055] The model then allows to estimate core temperatures Tc at any location within the cooler 10, 10′, 130. The food-dependent parameters are associated with specific food groups and will work in any cooler with an controlled environment. Such an environment can also be an outside environment 16, if there is no influence from direct or reflected light from the sun or extensive convection, i.e. without adding heat via additional radiation or convection. Cooler-dependent parameters will fully describe cooler 10, 10′ properties and do not depend in any way on the type of food inside the cooler, only from the known disposition of food packages in the cooler.

[0056] The quality control of the predictive model is performed by integrating the above mentioned formula of function f(h, Q, Te, Tc). The temperature sensor unit 20 records every 2 minutes the environment temperature Te in the cooler 10 and transfers it to a remote control unit. The predicted values are then compared with actual core temperature measurements performed in the shop premises.

[0057] FIG. 3 shows three excerpts of predicted curves of temperature 94 (Te) against time 93 (t) for a perishable food 13 from the food group “diary” in the back row of a specific cooler type 10. Each excerpt shows the predicted temperature 90 at one thousand measurement points in time, i.e. over a time interval of 2000 minutes each=about 33 hours for each curve. The curves 91 reflect the upper uncertainty of the model +1 degree Celsius and the curves 92 the lower uncertainty of the model −1 degree Celsius. Three manual measurements were made, having received the reference numerals 95 with an instrumental error of +−0.5 degree Celsius which clearly show the correct prediction of the core temperature.

[0058] FIG. 4 shows a schematic view of a temperature sensor unit 20 to be used in connection with the invention. The temperature sensor unit 20 comprises a case with openings 26 to allow air from outside to circulate inside over a temperature sensor 25, to determine the air temperature Te. The temperature sensor 25 is connected to a control unit 27 which is connected to a battery or power supply 28 to control taking measurements with the temperature sensor 25 in predefined intervals (here 2 minutes) and to send them vie antenna unit 29 to the remote control center 300, preferably via LoRaWan 200 as shown in FIG. 5 but also other especially wireless transmission means are possible. The temperature sensor unit 20 can be attached to metallic surfaces with the magnet 31.

[0059] FIG. 5 shows a schematic view of a food safety system for an integrated company according to the invention. The above mentioned temperature sensor unit 20 is provided at and in every cooler 10 in the retail stores 110 of the company. Beside retail stores 110, such place can also be simple cooler 10 at distribution points of partner companies or at distribution points with clients, when perishable goods are sold at other premises. Furthermore, these temperature sensor units 20 can be used in distribution centers 120; there usually different packaging (groups of perishable food items) are stored in larger cooler units (not shown here). Finally, in order to monitor the entire cold chain the refrigerated lorries 130 can also receive temperature sensor units 20; here inside the vehicle as such. It is possible beside a data transfer via LoRaWan 200 and the internet 205 also to use the GSM/mobile phone infrastructure 210.

[0060] All measurement data arrive at the control center 300 and are used to determine the predictive temperatures for the different positions in the cooler 10 or 10′. Preferably, the knowledge, if a temperature of one food category at a specific position in the cooler 10 would raise and perhaps rise beyond the authorized threshold for the product base on the predicted temperature, is transmitted via an alarm server 310 to a user 320 at the premises 110 via e.g. a smartphone or an alarm computer in the unit 110. The user 320, usually an employee of the company can then check the reasons for this finding and take appropriate measures, e.g. replace a faulty cooler, check the isolation of doors 18 etc. In fact, the safety system comprises the control center 300 and at least one of here a plurality of the temperature sensor units, wherein the control center 300 comprises the database with the parameters for each of the cooler systems to be monitored. If a new cooler has to be added to the group of available coolers to the company, it is sufficient to update the database of the control center with the data relating to this new cooler.

[0061] The deterministic model function f based on cooler specific heat transfer h and food specific coefficient Q is developed based on real measurements based on how the different product groups in the core react to temperature change, day and night, in different cooler types. The cooler type parameter also takes into account the position in the cooler. These measurements leading to the model ar eperformed with possible variations of the four product groups: Meat, fish, dairy, convenience foods, such as tranched, shredded, etc., effectively in a real world environment. At least, the cooler and food specificities lead two a two dimensional parameter set. In addition, depending on the cooler type, the topping up effect is taken into account as well as e.g. if several yoghurt cups are on top of each other. The model uses a percentage of filling level of the cooler which is usually observed.