WO2025144365A2

WO2025144365A2 - Temperature control system and method for refrigerated display cabinets

Info

Publication number: WO2025144365A2
Application number: PCT/TR2024/051818
Authority: WO
Inventors: İhsan Gökhan GÖKYOL; Muhammed EROĞLU; Ece BİNİCİER
Original assignee: Kaplanlar Sogutma Sanayi Ve Ticaret AS
Current assignee: Kaplanlar Sogutma Sanayi Ve Ticaret AS
Priority date: 2023-12-29
Filing date: 2024-12-27
Publication date: 2025-07-03
Anticipated expiration: 2026-06-29
Also published as: TR2023019619A2; WO2025144365A3

Abstract

The invention relates to the temperature control system and method for the cooling system created with the machine learning method and fuzzy logic of the cooling systems of the refrigerator cabinets, which offer a highly effective use for businesses that need product display or storage, such as supermarkets, buffets, patisseries, cafes. The elements used in the invention are; refrigerated display cabinet (10), blowing temperature sensor (S4), average product temperature value (EPT) suction temperature sensor (S3), door sensor (Dn), humidity sensor (1), temperature sensor (2), camera (3), fuzzy cooling class coefficient (4), number of customers (5), database (6), artificial intelligence model (7), density class (8), new temperature value (9).

Description

TEMPERATURE CONTROL SYSTEM AND METHOD FOR REFRIGERATED DISPLAY CABINETS

Technical Field

The invention relates to a control system and method created with machine learning method and fuzzy logic for the cooling systems of refrigerated cabinets that offer a highly effective use for businesses that need product display or storage such as supermarkets, buffets, patisseries, cafes.

Prior Art

Today, refrigerated display cabinets used for product display; In order to keep the cold air in it for a longer period of time, the frequency of compressor start / stop frequency per unit cooling, the decrease in efficiency until the heat balance is achieved again due to the disturbance of the refrigerant distribution in each cooling cycle is a long-standing issue in the technique. Some solutions to this issue are available in the prior art. The effective solution to this problem is important in many respects. In this respect, this problem is still an important problem waiting to be solved in many areas, especially in supermarkets.

Another problem mentioned in the prior art is that the temperature of the displayed product is maintained at a certain temperature and humidity conditions in order to protect the physiological structures and to display them to the end user. Since the temperature of the displayed product is not taken as a reference to environmental conditions, each cycle operates at a certain interval and causes extra energy consumption and icing. If there is a small air intake into the refrigerated display cabinets due to continuous opening and closing, this hot and dense humid air can cause excessive icing. For such reasons, the desired efficiency cannot be obtained in the preservation of the products.

Abstract of the application numbered 2018/07660 resulting from the technical researches; ‘A refrigerated display cabinet comprising a display cabinet body, at least one top-accessible refrigerating chamber and a cover cap covering the upper face of the display cabinet body, and two mutually accessible faces in which each refrigerating chamber can be grasped in the open cover cap, wherein the cover cap has two three-way rail profiles arranged parallel to each other, at least one 3-sliding cover arrangement comprising two transparent sliding covers arranged on both access faces in the closing position and a transparent middle sliding cover arranged in the middle between the two edge sliding covers in the center position, wherein the rail profile is arranged in the transverse direction of the access faces and each has a lower, a middle and an upper track, and wherein the middle sliding cover is guided in the lower rail profile track.” is in the form.

As can be seen, the system is related to the cooler display cabinet, but does not mention a structure that can provide a solution to the disadvantages mentioned above.

As a result, due to the above-mentioned negativities and the inadequacy of the existing solutions on the subject, it has become necessary to make a development in the relevant technical field.

Purpose of the Invention

The invention aims to provide a structure having different technical features which brings a new opening in this field, different from the structures used in the present art.

The primary purpose of the invention is to provide a control system and method for the cooling system of refrigerated cabinets, which offer a highly effective use for businesses that need product display or storage such as supermarkets, buffets, patisseries, cafes, etc., with machine learning method and fuzzy logic.

In the algorithm developed to ensure the protection of food safety and minimise energy losses, the door sensors and camera module in the refrigerated display cabinets calculate the density of customers interacting with the retail cabinet by machine learning with the decision-making algorithm. With the calculation of customer density, the operating temperature of the retail cabinet is determined as shown in Figure 1 as a result of fuzzy inference.

In addition, the appropriate temperature of the fluid cold air in the refrigerated display cabinet is dynamically regulated at the exit of each cooling cycle (setpoint of the cabinet) and ensures a balance in energy use.

Today air ducts direct the air flow inside the refrigerated display cabinet. Generally, it works in two ways. The first one carries the cold air pressed by the fan to the blowing holes for air distribution in the display cabinet, where it is measured by the blowing temperature sensor (probe). For this reason, the blowing temperature sensor is the part that determines the flow in the cabinet. In the second working principle, the heated air in the cabinet is sucked and directed back to the evaporator.

A database is a systematic collection of data resulting from the organised electronic storage of data. It can contain all kinds of data, including words, numbers, images, videos and files. Edge computing was used for a high-performance database. Instead of hosting the processing power in the cloud or centralised data system, edge computing performs the data in a large number of small data centres located at or near the source. Edge computing is about bringing the data as close as possible to the actual device (usually 30 metres or less). In some cases, this provides accurate results, reduces latency and uses less network. Cloud computing is a completely different model and provides remote access to users' data, applications and infrastructure. However, latency and performance due to storage between these remote servers becomes a problem.

It offers advantages over cloud computing in many use cases, especially in terms of efficiency in network and data storage. These advantages can be listed as follows;

1 . Faster performance: Although the cloud can process data quickly, the latency issues mentioned above can result in slower response times for internet-connected devices. On the contrary, the Edge process increases the data transfer rate, allowing the same devices to benefit from significantly better performance.

2. Cost savings: The upfront costs associated with storage and processing power are a costly investment for any organisation. With edge computing, much less data is transferred between users and data centres. Because data can be processed locally, companies can choose what they want to run locally and what they want to run in the cloud. It is a more cost-effective solution.

3. Improved confidentiality: To date, privacy remains one of the biggest barriers to cloud adoption. Edge computing not only reduces cloud traffic, but also filters and stores sensitive data locally. Organisations therefore build infrastructure around privacy and compliance requirements.

In line with this information, the refrigerated display cabinet stores EPT (average product temperature value), humidity and temperature inputs and density class data in the database with Edge informatics. Due to the fact that there is a different situation other than the optimum conditions in the database, the output of the artificial intelligence model trained with feedback artificial intelligence training is obtained. Feedback artificial intelligence training is tested in person and temperature conditions by considering and animating many scenario alternatives during the research and development phase of the refrigerated display cabinet. Adaptive network based fuzzy inference system (ANFIS) was used for the untestable cases. When there is a critical situation, machine learning is performed through predictive fuzzy logic with ANFIS.

The fuzzy cooling class coefficient controls the values in the temperature evaluation table as to whether the internal temperature of the refrigerated display cabinet should change or not. In the known state of the art, retail refrigerated display cabinets have certain standards of operating temperatures according to the product stored therein. For example; Although the operating temperature of the cabinets used for storing milk may vary seasonally and depending on the environment, it is considered appropriate to be between +4 and +6 degrees. Considering the temperature evaluation table, the temperature value suitable for the product is decided whether the temperature value of the refrigerated display cabinet will change with machine learning and fuzzy logic. Thus, the loop is completed.

The invention is also of great benefit in terms of energy efficiency. It controls the operation of all kinds of refrigerated display cabinets used for industrial or individual purposes in a hierarchy. It learns the optimum operating conditions and when it goes out of these operating conditions, it changes the optimum set value and notifies the user via appropriate communication terminals.

The feedback artificial intelligence training of the invention performs machine learning for the product in the refrigerated display cabinet. By learning the events, it makes logical decisions in the face of similar possible situations and events. Previously, it can produce information about unforeseen events. During artificial intelligence training, the refrigerated display cabinet is able to generate information about new samples in different condition samples given to it.

In order to fulfil the purposes described above, the invention is a temperature control system for the cooling systems of refrigerating cabinets used in enterprises such as supermarkets, buffets, patisseries, cafes, etc. which need product display or storage, and its feature;

• refrigerated display cabinet on which the system and algorithm work,

• blowing temperature sensor, which is located on the refrigerated display cabinet and used in the calculation of the average product temperature value, • suction temperature sensor, which is located on the refrigerated display cabinet and used in the calculation of the average product temperature value,

• door sensor, which informs the system about the open / closed status of the door in case the refrigerated display cabinet is a cabinet with a door,

• average product temperature value, which is calculated with the data received from the blowing and suction temperature sensor of the product displayed in the refrigerated display cabinet,

• humidity sensor for measuring humidity integrated in the refrigerated display cabinet,

• temperature measuring temperature sensor, which is integrated in the refrigerated display cabinet,

• camera, which is fixed to the refrigerated display cabinet, allows the image to be taken for image processing,

• fuzzy cooling class coefficient, which is an integer obtained after the fuzzy inference process from the density class and plays a role in whether the refrigerated display cabinet temperature value changes or not,

• number of customers, which indicates the number of people for whom the outputs obtained as a result of image processing with the camera are generated,

• database where data from sensors are stored,

• trained artificial intelligence model, which is created with fuzzy logic and artificial intelligence algorithms as a result of feedback artificial intelligence training,

• number of customers in the environment and the density class in which the data from the door sensor is classified,

• new temperature value, which emerges in deciding whether to change the temperature value by looking at the temperature evaluation table, after the fuzzy inference process.

The structural and characteristic features and all advantages of the invention will be more clearly understood by means of the figures given below and the detailed description written by making references to these figures, and therefore, the evaluation should be made by taking these figures and detailed description into consideration.

Figures to Help Understanding of the Invention

Figure 1 is a schematic representation of the system of the invention. The drawings are not necessarily to scale and may omit details that are not necessary to understand the present invention. Furthermore, elements that are at least substantially identical or have at least substantially identical functions are indicated by the same number.

Description of Part References

S4 Blowing Temperature Sensor

S3 Suction Temperature Sensor

Dn Door Sensor

EPT Average Product Temperature Value

1 . Humidity Sensor

2. Temperature Sensor

3. Camera

4. Fuzzy Cooling Class Coefficient

5. Number of Customers

6. Database

7. Trained Artificial Intelligence Model

8. Density Class

9. New Temperature Value

10. Display Cabinet with Cooler

11 . Temperature evaluation table

A. EPT calculation

B. Is there a different condition?

C. Feedback artificial intelligence training

D. Interaction intensity classification

E. Fuzzy inference process

F. Should the temperature value change?

G. Temperature value changes

H. Image processing

X. Yes

A. No

Detailed Description of the Invention In this detailed description, preferred embodiments of the invention are explained only for the purpose of better understanding the subject and in a way that does not create any limiting effect.

The invention relates to the temperature control system and method for the cooling system created with the machine learning method and fuzzy logic of the cooling systems of the refrigerator cabinets, which offer a highly effective use for businesses that need product display or storage, such as supermarkets, buffets, patisseries, cafes.

The elements and their functions used in the system which is the subject of the invention are as follows;

The blowing temperature sensor (S4) is located on the refrigerated display cabinet (10) and is used in the calculation of the average product temperature value (EPT).

The suction temperature sensor (S3) is located on the refrigerated display cabinet (10) and is used in the calculation of the average product temperature value (EPT).

The door sensor (Dn) informs the system about whether the door is open / closed if the refrigerated display cabinet (10) has a door.

The average product temperature value (EPT) is the value calculated from the data received from the blowing and suction temperature sensors (S4, S3) of the product displayed in the refrigerated display cabinet (10).

The humidity sensor (1 ) is a humidity measuring sensor integrated into the refrigerated display cabinet (10).

The temperature sensor (2) is a temperature measuring sensor integrated into the refrigerated display cabinet (10).

The camera (3) provides image acquisition for image processing and is fixed to the refrigerated display cabinet (10). The fuzzy cooling class coefficient (4) is the integer obtained after performing the fuzzy inference process from the density class (8). The refrigerated display cabinet (10) plays a role in whether the temperature value changes or not.

The number of customers (5) indicating the number of people for whom the outputs obtained as a result of image processing with the camera (3) are generated,

The database (6) is the electronic medium where the data received from the sensors are stored.

The trained artificial intelligence model (7) is a model created with fuzzy logic and artificial intelligence algorithms as a result of feedback artificial intelligence training.

The density class (8) is where the number of customers in the environment (5) and the data coming from the door sensor (Dn) are classified.

After the fuzzy inference process, the new temperature value (9) that emerges in deciding whether to change the temperature value by looking at the temperature evaluation table (11 ).

The refrigerated display cabinet (10) is the retail cabinet on which the system and algorithm operate.

The working principle of the system which is the subject of the invention is as follows;

When looking at the algorithm diagram of the invention in Figure 1 , there is a blowing temperature sensor (S4) and a suction temperature sensor (S3) integrated into the refrigerated display cabinet (10). These sensors are used to calculate the average product temperature (EPT). The calculation of average product temperature (EPT) is mentioned in patent file number 2022-01738.

If the refrigerated display cabinet in question has (10) doors, door sensors (Dn) are installed as many times as the number of doors. Customer interaction data received from the door sensors (Dn) plays a role in determining the number of customers (5) interacting with the refrigerated display cabinet (10) in classifying the interaction intensity. It uses machine learning in its algorithm to decide on interaction intensity. In the said refrigerated display cabinet (10), there is at least one humidity sensor (1 ) and at least one temperature sensor (2). The humidity sensor (1 ) and the temperature sensor (2) make measurements due to the continuous opening and closing of the doors of the refrigerated display cabinet (10). The measured temperature (2) and humidity sensor (1 ) data are stored in the database (6) of the controller to be used in the refrigerated display cabinet (10). In addition to these, the mentioned sensor data will be used in the fuzzy inference process which will be created with fuzzy logic.

Another module camera (3) located on the refrigerated display cabinet (10) is responsible for image processing. Using image processing, it calculates the number of customers (5) within the viewing angle of the camera (3) positioned on the refrigerated display cabinet (10). Whether or not the customer interacts with the refrigerated cabinet (10) is calculated separately. As a result, the customer number (5) data obtained is transmitted to the interaction intensity class (8). The interaction intensity classification process is initiated with the customer number (5) data obtained from the door sensors (Dn) and image processing.

After the density class (8) interaction density classification process is completed, it moves on to the fuzzy inference process stage to create the fuzzy logic algorithm.

The fuzzy inference system / ANFIS process is carried out by transmitting the average product temperature (EPT), humidity sensor (1 ), temperature sensor (2) and density class (8) data to the fuzzy inference process. Thus, an algorithm for estimating the cooling system method created with fuzzy logic is created. In the production of refrigerated display cabinets (10), an adaptive network-based fuzzy inference system is implemented for situations that cannot be tested in test rooms. In critical situations that cannot be tested for refrigerated display cabinets (10) in stores, machine learning is performed using ANFIS and predictive fuzzy logic.

The temperature sensor (2), humidity sensor (1 ), average product temperature (EPT) and density class (8) transmitted to the database (6) query whether there is a different condition. In case of a condition different from the current situation, feedback artificial intelligence training is performed.

The trained artificial intelligence model (7) is a model created with fuzzy logic and artificial intelligence algorithms as a result of feedback artificial intelligence training. The obtained trained artificial intelligence model (7) is transferred to the fuzzy inference process. The Fuzzy cooling class coefficient (4) is the integer obtained after performing the fuzzy inference process from the density class (8). The Fuzzy cooling coefficient (4) is not a fixed number according to the data from the sensors, but a number that varies according to the state of the system. In the state of the art, refrigerated display cabinets (10) are adjusted during the installation phase to operate within certain temperature ranges. However, with the cooling coefficient (4) value obtained in our invention, a decision-making algorithm has been created as to whether the temperature value of the refrigerated display cabinet (10) should change or not. When deciding on the temperature of the refrigerated display cabinet (10), the decision-making stage is carried out by looking at the refrigerated display cabinet (10) temperature evaluation table (11 ) known in the literature.

In the case where the temperature value should not change by referring to the temperature evaluation table (11 ), the system completes the cycle by giving the answer "NO". If the temperature value needs to be changed, the system responds with "YES" and the appropriate temperature value is selected.

The appropriate temperature value for the refrigerated display cabinet (10) is the new temperature value (9). This new temperature value (9) changes the current temperature value of the refrigerated display cabinet (10). The system in question is related to the refrigerated display cabinet (10) that decides its temperature itself with fuzzy logic / ANFIS, artificial intelligence training and machine learning in order to preserve the products displayed in the most appropriate way at the end of the algorithm and to extend the shelf life of the products.

In addition to the features mentioned, EDGE computing is used to store so much sensor data and querying processes in the database (6) of the refrigerated display cabinet (10) controller. Instead of hosting processing power in the cloud or a central data system, EDGE computing stores data in many small data centers located at or near the source.

EDGE computing is about getting the stored data as close to the device as possible (Usually 30 meters or less.). Due to this, accurate results are provided in the data transfer between the database (6) and the system, data transmission delay is low and less network data is used.

The process steps carried out with the system which is the subject of the invention are as follows;

Integrating at least one suction temperature sensor (S3), blowing temperature sensor (S4), door sensor (Dn), humidity sensor (1 ), temperature sensor (2) and camera (3) into the refrigerated display cabinet (10), • Obtaining the average product temperature (EPT) by calculating the average product temperature (EPT) from the suction temperature sensor (S3) and the blowing temperature sensor (s4),

• Transmitting the humidity sensor (1 ) and temperature sensor (2) outputs to the database (6) of the refrigerated display cabinet (10) controller,

• Transmitting the calculated average product temperature (EPT) to the database (6),

• Performing the image processing by at least one camera (3) fixed on the refrigerated display cabinet (10),

• Calculating the demographic data and number of customers (5) interacting with the refrigerated display cabinet (10) using image processing,

• Using the number of customers coming from the door sensor (Dn) and image processing (5) in classifying the interaction intensity,

• Forming the intensity class (8) as a result of the interaction intensity classification process,

• Transmitting the resulting density class (8) data to the database (6),

• Performing a query to see if there is a different condition in the database (6) according to the data stored in Edge computing,

• Performing feedback-based artificial intelligence training if the answer to a different condition is “YES”, o Forming the trained artificial intelligence model (7) as a result of the training of the trained artificial intelligence model, o Using the trained artificial intelligence model (7) to perform mathematical operations on the average product temperature (EPT), density class (8), humidity sensor (1 ) and temperature sensor (2) values in the fuzzy inference process, o Obtaining the fuzzy cooling class coefficient (4) is obtained after the mathematical operations performed in the fuzzy inference process, (The fuzzy cooling class coefficient (4) may vary depending on the sensor data. The Fuzzy cooling class coefficient (4) can be a decimal or integer value.) o Making a logical decision on whether the internal temperature value of the refrigerated display cabinet (10) should change or not by looking at the values in the temperature evaluation table (11 ) and obtaining the new temperature value (9), o Updating the current temperature value of the refrigerated display cabinet (10) according to the new temperature value (9) obtained, • Current temperature values of the refrigerated display cabinet (10) remaining constant, if the answer is “NO”.

Claims

1. A temperature control system for the cooling systems of refrigerators used in businesses that need product display or storage such as supermarkets, buffets, patisseries, cafes, characterized by comprising;

• refrigerated display cabinet (10) on which the system and algorithm works,

• blowing temperature sensor (S4), which is located on the refrigerated display cabinet (10) and used in the calculation of the average product temperature value (EPT),

• suction temperature sensor (S3), which is located on the refrigerated display cabinet (10) and is used in the calculation of the average product temperature value (EPT),

• door sensor (Dn), which informs the system about whether the door is open / closed if the refrigerated display cabinet (10) has a door,

• average product temperature value (EPT), which is the value calculated from the data received from the blowing and suction temperature sensors (S4, S3) of the product displayed in the refrigerated display cabinet (10),

• humidity sensor (1), which is a humidity measuring sensor integrated into the refrigerated display cabinet (10),

• temperature sensor (2), which is a temperature measuring sensor integrated into the refrigerated display cabinet (10),

• camera (3), which is fixed to the refrigerated display cabinet (10), allows the image to be taken for image processing,

• fuzzy cooling class coefficient (4), which is an integer obtained after the fuzzy inference process from the density class (8) and plays a role in whether the refrigerated display cabinet (10) temperature value changes or not,

• number of customers (5), which indicates the number of people for whom the outputs obtained as a result of image processing with the camera (3) are generated,

• database (6) where data from sensors are stored,

• trained artificial intelligence model (7) created with fuzzy logic and artificial intelligence algorithms as a result of feedback artificial intelligence training,

• density class (8) is where the number of customers in the environment (5) and the data coming from the door sensor (Dn) are classified,

• new temperature value (9) that emerges in deciding whether to change the temperature value by looking at the temperature evaluation table after the fuzzy inference process.

2. A temperature control method for the cooling systems of refrigerators used in businesses that need product display or storage such as supermarkets, buffets, patisseries, cafes, characterized by comprising the process steps of;

• Integrating at least one suction temperature sensor (S3), blowing temperature sensor (S4), door sensor (Dn), humidity sensor (1), temperature sensor (2) and camera (3) into the refrigerated display cabinet (10),

• Obtaining the average product temperature (EPT) by calculating the average product temperature (EPT) from the suction temperature sensor (S3) and the blowing temperature sensor (S4),

• Performing image processing by at least one camera (3) fixed on the refrigerated display cabinet (10),

• Transmitting the resulting density class (8) data to the database (6),

• Performing feedback-based artificial intelligence training, if the answer to a different condition is “YES”, o Forming the trained artificial intelligence model (7) as a result of the training of the trained artificial intelligence model o Using the trained artificial intelligence model (7) to perform mathematical operations on the average product temperature (EPT), density class (8), humidity sensor (1) and temperature sensor (2) values in the fuzzy inference process, o Obtaining the fuzzy cooling class coefficient (4) is obtained after the mathematical operations performed in the fuzzy inference process, o Making a logical decision on whether the internal temperature value of the refrigerated display cabinet (10) should change or not by looking at the values in the temperature evaluation table (11 ) and obtaining the new temperature value (9), o Updating the current temperature value of the refrigerated display cabinet (10) according to the new temperature value (9) obtained,

• Current temperature values of the refrigerated display cabinet (10) remaining constant, if the answer is “NO”.