Method for achieving refrigeration house control through edge calculation

Abstract

The invention relates to a cold storage system control method, in particular to a method for achieving cold storage control through edge calculation. The control method comprises the following steps that the edge computing terminal outputs a corresponding control instruction according to whether the detected temperature of the refrigeration storage reaches a preset temperature range or not, and basic refrigeration control is achieved for the refrigeration storage system; after the edge computing terminal carries out primary processing on the data, uploading the temperature, the superheat degree, the valve opening degree and other data with a time sequence to a cloud server side in a wireless communication mode, and storing the data in a database; at the cloud server side, deep analysis is carried out on the temperature data by using an RNN recurrent neural network algorithm, and corresponding valve opening parameters of the electronic expansion valve are output; the control parameters are transmitted to an edge computing terminal in a wireless communication mode, the refrigeration house electronic expansion valve is accurately controlled, and self-adaptive control over the refrigeration house system is achieved.

Description

Technical Field

[0001] The present invention belongs to the field of cold storage control technology and relates to a cold storage system control method, and more specifically, to a method for implementing cold storage control through edge computing. Background Art

[0002] As an important part of the storage and transportation of agricultural products and medicines, the cold storage system requires a certain stability in temperature control, especially for the storage of medicines, which has strict temperature requirements. Traditional cold storage systems usually use simple control logic to turn off or on the refrigeration system by judging whether the temperature is within the set range. Traditional control methods can ensure the stable operation of the system, but the temperature control has a certain lag and the control accuracy is relatively low. Especially when people enter and leave the cold storage, the temperature will fluctuate greatly and often exceed the preset range. The control program of the traditional cold storage control system is usually stored in the PLC programmable logic controller, which is extremely unfavorable for protecting the confidentiality of the control program.

[0003] In summary, the existing technology has technical problems to varying degrees in terms of temperature control accuracy, stability, response speed, anti-interference ability, control program security, system scalability and flexibility. These problems limit the application effect and development potential of traditional cold storage systems in the field of high-precision temperature control of agricultural products, medicines, etc. Summary of the invention

[0004] In order to solve the above-mentioned technical problems existing in the prior art, the present invention provides a method based on edge computing technology, which can further accurately control the temperature of a cold storage system while ensuring the normal operation of the cold storage, thereby improving the temperature control accuracy and stability of the cold storage system, and enhancing the security, scalability and intelligence level of the system.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for implementing cold storage control by edge computing, the steps are as follows:

[0007] Step S1: collect various data during the operation of the cold storage through the temperature and humidity sensors and pressure sensors in the cold storage, and connect the various data to the edge computing terminal;

[0008] Step S2: The edge computing terminal compares the cold storage temperature, the preset temperature and the error range through if logic, outputs and sends the corresponding control signal to the control terminal, and starts and stops the compressor, electronic expansion valve and air cooler in turn to realize automatic control of the cold storage system;

[0009] Step S3: Perform preliminary processing on the evaporator outlet temperature and evaporation pressure at the edge computing terminal, where the superheat calculation formula is:

[0010] Superheat = evaporator outlet temperature - saturation temperature at current evaporation pressure

[0011] The saturation temperature is the highest temperature at which the refrigerant can exist without boiling under the current pressure;

[0012] Step S4: The edge computing terminal uses the MQTT protocol to transmit the cold storage temperature data with time series to the cloud server, and automatically stores it in the database on the cloud server;

[0013] Step S5: The cloud server performs in-depth analysis of the temperature data through a recurrent neural network algorithm to predict future temperature changes;

[0014] Step S6: Using a PID control algorithm to control the opening of the electronic expansion valve, input the predicted future temperature into the PID, calculate the temperature difference from the target temperature and the current superheat, and output a more accurate electronic expansion valve opening parameter;

[0015] Step S7: Use the MQTT protocol to transmit the output electronic expansion valve opening parameters to the edge computing terminal, and then transmit them to the control end through the wired network to achieve adaptive control of the cold storage system.

[0016] Furthermore, the start and stop of the refrigeration system is determined by the if control logic according to whether the cold storage temperature is within a preset range.

[0017] Furthermore, step S4 specifically includes:

[0018] Step S401: Use the edge computing terminal as the MQTT data publishing terminal and set the cold storage temperature data with time series to be published at regular intervals;

[0019] Step S402: Run Python related programs to build an MQTT client on the cloud server;

[0020] Step S403: The cloud server connects to the MQTT Broker and subscribes to the Topic that needs to receive data;

[0021] Step S404: Run the python related program to automatically store the received data into the MySQL database;

[0022] Step S405: Performing algorithm analysis on the temperature on the cloud server side and outputting valve opening parameters;

[0023] Step S406: The cloud server acts as the MQTT data publishing end, and the edge computing terminal acts as the MQTT data receiving end to accept control parameters and transmit the parameters to the control terminal via a wired network to precisely control the cold storage system.

[0024] Further, step S5, using a recurrent neural network algorithm on the cloud server to perform in-depth analysis on the data and predict future temperature changes, includes the following steps:

[0025] Step S501: Establish a three-layer RNN recurrent neural network model, including an input layer, a hidden layer, and an output layer, and the number of neurons is 3-10-1;

[0026] Step S502: The cold storage temperature, superheat, and valve opening are used as the original input data, and the original input data is normalized using minimum-maximum scaling. For each feature x, the conversion formula is:

[0027]

[0028] After the transformation, all eigenvalues ​​will fall within the range of [0,1];

[0029] Step S503: The activation function is the sigmoid function, and the error is calculated using MSE. The MSE calculation formula is:

[0030]

[0031] Among them, N represents the number of samples, oi represents the output result of the RNN prediction model, and yi represents the true value;

[0032] Step S504: The first 80% of the processed data is used as training data, and the last 20% is used as verification data. When the error of the RNN prediction model is less than 0.1°C, the training of the RNN model is completed;

[0033] Step S505: Calculate the predicted temperature of the cold storage in the next time period.

[0034] Further, in step S4, the user views historical data or real-time data monitoring by using the database.

[0035] Furthermore, by deploying several groups of temperature and humidity sensors, the temperature of different areas in the cold storage is zoned to monitor and control the temperature of each area.

[0036] Furthermore, step S1: connecting various data to the edge computing terminal through a wired network connection.

[0037] Furthermore, the edge computing terminal may be FBOX or IVP09A.

[0038] Beneficial effects of the present invention:

[0039] Compared with the prior art, the method for implementing cold storage control by edge computing described in the present invention has the following technical characteristics and beneficial effects:

[0040] (1) Improve control accuracy and response speed: The cold storage operation data is directly processed through the edge computing terminal, and the electronic expansion valve opening is accurately controlled in combination with the PID control algorithm, which effectively reduces the lag of temperature control and improves the system's response speed and temperature control accuracy. In particular, when there are external interferences such as personnel entering and exiting or cargo loading and unloading, the temperature can be restored to the preset temperature range more quickly.

[0041] (2) Enhance the intelligence and adaptability of the system: Use the recurrent neural network algorithm on the cloud server to conduct in-depth analysis of temperature data, predict future temperature changes, and adjust the control strategy accordingly, so that the cold storage system can adaptively adjust its working state according to real-time data and prediction results, thereby improving the intelligence level and operating efficiency of the system.

[0042] (3) Improve data security and confidentiality: Data is securely transmitted to the cloud server through the MQTT protocol and stored and analyzed in the cloud, avoiding the potential security risks of directly storing the control program in the PLC in the traditional way, and enhancing data confidentiality and system security.

[0043] (4) Facilitate data management and analysis: The database storage function of the cloud server allows users to view historical data or conduct real-time data monitoring at any time, which facilitates the operation and maintenance management of the cold storage and provides data support for further optimization of control strategies.

[0044] (5) Enhanced system scalability and flexibility: The technical solution of the present invention adopts a modular design concept, and each component is relatively independent and easy to maintain and upgrade. At the same time, through algorithm analysis and control strategy adjustment on the cloud server side, it can easily adapt to cold storage systems of different scales and types, thereby improving the scalability and flexibility of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the present invention will be described in detail below in combination with the accompanying drawings and detailed embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0046] Figure 1 is a schematic diagram of a control method of an embodiment disclosed in the present application;

[0047] Figure 2 It is an application schematic diagram of the embodiment disclosed in this application;

[0048] Figure 3It is a scheme diagram of using an RNN neural network model to adjust the valve opening of an electronic expansion valve provided in a disclosed embodiment of the application. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Figure 1-3 For further explanation, the technical solutions in the embodiments of the present application are clearly and completely described; obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments, and all other embodiments obtained by ordinary technicians in this field without making creative work based on the embodiments in the present application are within the scope of protection of the present application.

[0050] Example 1

[0051] A cold storage system control method based on edge computing includes the following steps:

[0052] Step S1: various data during the operation of the cold storage are collected through sensors such as temperature, humidity, and pressure in the cold storage, and the various data are connected to the edge computing terminal through a wired network connection;

[0053] Step S2: The edge computing terminal compares the cold storage temperature with the preset temperature and the error range through if logic, outputs and sends the corresponding control signal to the control terminal, and performs basic start and stop control on the compressor, electronic expansion valve, air cooler and other equipment in turn to realize automatic control of the cold storage system;

[0054] Step S3: Preliminary processing of the evaporator outlet temperature, evaporation pressure and other data is performed on the edge computing terminal, where the superheat calculation formula is:

[0055] Superheat = evaporator outlet temperature - saturation temperature at current evaporation pressure

[0056] The saturation temperature is the highest temperature at which the refrigerant can exist without boiling under the current pressure;

[0057] Step S4: The edge computing terminal uses the MQTT protocol to transmit the cold storage temperature data with time series to the cloud server, and automatically stores it in the database on the cloud server. Users can also use the database to view historical data or real-time data monitoring;

[0058] Wherein, step S4 uses the MQTT protocol to complete the method of wireless data transmission, which specifically includes:

[0059] Step S401: Use the edge computing terminal as the MQTT data publishing terminal, and set the cold storage temperature, superheat, and valve opening data with time series to be published every certain period of time (for example, 1 minute);

[0060] Step S402: Run the python related program, through client = mqtt.Client (client_id = "") and

[0061] client.username_pw_set(username="",password="")Set the MQTT client ID, username and password, and build the MQTT client on the cloud server;

[0062] Step S403: Set the MQTT address and port number through the client.connect("",) program to connect the cloud server to the MQTT Broker, and subscribe to the Topic that needs to receive data through client.subscribe("Topic");

[0063] Step S404: Set the address, user name, password and database name of the MySQL database through mysql_host="", mysql_user="", mysql_password="mysql_database="" and other programs and connect to the MySQL database. Automatically store the received data into the MySQL database through sql INSERT INTO, VALUES statements and cursor.execute(sql,values), connection.commit() and other programs;

[0064] Step S405: Perform recurrent neural network algorithm analysis on the current parameters on the cloud server side to predict the temperature change in the next time period and use it as feedback input for the PID control algorithm, combined with the preset temperature and the current output valve opening parameters;

[0065] Step S406: The cloud server acts as the MQTT data publishing end, and the edge computing terminal acts as the MQTT data receiving end to receive the control parameters, and transmits the parameters to the control terminal via a wired network to precisely control the cold storage system;

[0066] Step S5: The cloud server performs in-depth analysis on the temperature data through the RNN recurrent neural network algorithm to predict future temperature changes;

[0067] Among them, step S5, using a recurrent neural network algorithm on the cloud server to deeply analyze the data and predict future temperature changes, includes the following steps:

[0068] Step S501: Establish a three-layer RNN recurrent neural network model, including an input layer, a hidden layer, and an output layer, and the number of neurons is 3-10-1;

[0069] Step S502: The cold storage temperature, superheat, and valve opening are used as the original input data, and the original input data is normalized using Min-Max Scaling. For each feature x, the conversion formula is:

[0070]

[0071] Among them, x' i (t) represents the eigenvalue of the i-th feature after normalization, x i (t) represents the original value of the i-th feature, min(x i ) and max(x i ) represent the minimum and maximum original data of the i-th feature respectively. After conversion, all eigenvalues ​​will fall within the range of [0,1];

[0072] Step S503: The activation function is the sigmoid function, and the error is calculated using MSE. The MSE calculation formula is:

[0073]

[0074] Where N represents the number of samples, i Represents the output result of the RNN prediction model, y i represents the true value;

[0075] Step S504: The first 80% of the processed data is used as training data, and the last 20% is used as verification data. When the error of the RNN prediction model is less than 0.1°C, the training of the RNN model is completed; Step S505: Calculate the predicted temperature of the cold storage in the next time period;

[0076] Step S6: Using a PID control algorithm to control the opening of the electronic expansion valve, input the predicted future temperature into the PID, calculate the temperature difference from the target temperature and the current superheat, and output a more accurate electronic expansion valve opening parameter;

[0077] Step S7: Use the MQTT protocol to transmit the output electronic expansion valve opening parameters to the edge computing terminal, and then transmit them to the control end through the wired network to achieve adaptive control of the cold storage system.

[0078] Example 2

[0079] This embodiment is a new embodiment or a supplement to Embodiment 1.

[0080] This embodiment discloses a cold storage system control method. Figure 1 , including cold storage equipment 1, sensor equipment 2, control terminal 3, human-computer interaction equipment 4, edge computing terminal 5, cloud server 6, and cloud configuration interface 7.

[0081] The sensor device 2 is installed in the cold storage device 1 , and its main function is to collect various data during the operation of the cold storage. The sensor device 2 communicates with the control terminal 3 through the wired network 101 .

[0082] The control terminal 3 can be an electronic device that supports communication with the human-computer interaction device 4 and the edge computing terminal 5, including but not limited to a single-chip microcomputer, a PLC programmable logic control, etc. The control terminal 3 is connected to the edge computing terminal 5 via Ethernet 102 and is connected to the human-computer interaction device 4 via Ethernet 105.

[0083] The human-computer interaction device 4 can be a display screen with communication function, etc., through which the preset temperature and error range, start-stop delay and other parameters of the cold storage system can be set. The edge computing terminal 5 is deployed near the control terminal 3. For example, the edge computing terminal can be a device such as FBOX, IVP09A, etc.

[0084] The cloud server 6 is a virtualized computing resource provided based on cloud computing technology, which is accessed and used through the Internet. The edge computing terminal 5 is connected to the cloud server 6 through wireless communication.

[0085] Figure 2 The schematic diagram of the implementation scheme is an example. The sensor device 2 includes a temperature sensor 21 installed inside the cold storage, a temperature sensor 22 for detecting the evaporation temperature of the evaporator, a temperature sensor 23 for detecting the outlet temperature of the air cooler, and a pressure sensor 24 for detecting the evaporation pressure of the evaporator; the control terminal 3 uses PLC programmable logic control; the human-computer interaction device 4 uses a Siemens display screen; and the edge computing terminal 5 uses FBOX.

[0086] The specific implementation steps are as follows:

[0087] Step S1: During the operation of the cold storage, all sensor data are transmitted to the edge computing terminal 5 via a wired network;

[0088] Step S2: The edge computing terminal 5 compares the cold storage temperature with the preset temperature and the error range through if logic, outputs and sends the corresponding control signal to the control terminal 3, and performs basic start and stop control on the compressor, electronic expansion valve, air cooler and other equipment in turn;

[0089] Step S3: Preliminary processing of the data such as the evaporator outlet temperature and evaporation pressure is performed at the edge computing terminal 5, wherein the superheat calculation formula is:

[0090] Superheat = evaporator outlet temperature - saturation temperature at current evaporation pressure

[0091] The saturation temperature is the highest temperature at which the refrigerant can exist without boiling under the current pressure;

[0092] Step S4: the edge computing terminal 5 sends the cold storage temperature data with time series to the cloud server 6 via the wireless network using the MQTT protocol, and automatically stores it in the database at the cloud server 6;

[0093] In step S4, the edge computing terminal 5 wirelessly communicates with the cloud server 6 and uses the MQTT protocol to transmit data. The specific method includes:

[0094] Step S401: Using the edge computing terminal 5 as the MQTT data publishing terminal, the cold storage temperature data with time series is published every 1 minute;

[0095] Step S402: the cloud server 6 runs a python program with a paho-mqtt library and a pymysql library, and builds an MQTT client on the cloud server;

[0096] Step S403: The cloud server connects to the MQTT Broker and subscribes to the Topic that needs to receive data;

[0097] Step S404: Run the python program with the pymysql library to automatically store the received data into the MySQL database;

[0098] Step S405: performing algorithm analysis on the temperature at the cloud server 6 and outputting valve opening parameters;

[0099] Step S406: The cloud server acts as the MQTT data publishing end, and the edge computing terminal 5 acts as the MQTT data receiving end to receive the control parameters and transmit the data to the control terminal 3 to precisely control the cold storage system;

[0100] Step S5: The cloud server uses a recurrent neural network algorithm to perform in-depth analysis on the temperature data, predict future temperature changes, and output more accurate electronic expansion valve opening parameters; combined with Figure 3 The RNN neural network model is used to adjust the electronic expansion valve opening scheme diagram, and the recurrent neural network algorithm is used on the cloud server to deeply analyze the data and predict future temperature changes. Step S4 specifically includes the following steps:

[0101] Step S501: Establish a three-layer RNN recurrent neural network model, including an input layer, a hidden layer, and an output layer, and the number of neurons is 3-10-1;

[0102] Step S502: The cold storage temperature, superheat, and valve opening are used as the original input data, and the original input data is normalized using Min-Max Scaling. For each feature x, the conversion formula is:

[0103]

[0104] Among them, x' i (t) represents the eigenvalue of the i-th feature after normalization, x i (t) represents the original value of the i-th feature, min(x i ) and max(x i ) represent the minimum and maximum original data of the ith feature respectively. After conversion, all feature values ​​will fall within the range of [0,1]. Step S503: The activation function is the sigmoid function, and the error is calculated using MSE. The MSE calculation formula is:

[0105]

[0106] Where N represents the number of samples, i Represents the output result of the RNN prediction model, y i represents the true value;

[0107] Step S504: The first 80% of the processed data is used as training data, and the last 20% is used as verification data. When the error of the RNN prediction model is less than 0.1°C, the training of the RNN model is completed;

[0108] Step S505: Calculate the predicted temperature of the cold storage in the next time period;

[0109] Step S6: Use PID control algorithm to control the valve opening of the electronic expansion valve, input the predicted future temperature into PID, calculate the temperature difference with the target temperature and the current superheat value, and output a more accurate valve opening parameter of the electronic expansion valve. In this embodiment, the PID control parameters are set as proportional gain P=15, integral time I=150s, and differential time D=5s;

[0110] Step S7: Use the MQTT protocol to transmit the output electronic expansion valve opening parameters to the edge computing end, and transmit them to the control end through the wired network to achieve adaptive control of the cold storage system.

[0111] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for implementing cold storage control by edge computing, characterized in that: Here are the steps: Step S1: collect various data during the operation of the cold storage through the temperature and humidity sensors and pressure sensors in the cold storage, and connect the various data to the edge computing terminal; Step S2: The edge computing terminal compares the cold storage temperature, the preset temperature and the error range through if logic, outputs and sends the corresponding control signal to the control terminal, and starts and stops the compressor, electronic expansion valve and air cooler in turn to realize automatic control of the cold storage system; Step S3: Perform preliminary processing on the evaporator outlet temperature and evaporation pressure at the edge computing terminal, where the superheat calculation formula is: Superheat = evaporator outlet temperature - saturation temperature at current evaporation pressure The saturation temperature is the highest temperature at which the refrigerant can exist without boiling under the current pressure; Step S4: The edge computing terminal uses the MQTT protocol to transmit the cold storage temperature data with time series to the cloud server, and automatically stores it in the database on the cloud server; Step S5: The cloud server performs in-depth analysis of the temperature data through a recurrent neural network algorithm to predict future temperature changes; Step S6: Using a PID control algorithm to control the opening of the electronic expansion valve, input the predicted future temperature into the PID, calculate the temperature difference from the target temperature and the current superheat, and output a more accurate electronic expansion valve opening parameter; Step S7: Use the MQTT protocol to transmit the output electronic expansion valve opening parameters to the edge computing terminal, and then transmit them to the control end through the wired network to achieve adaptive control of the cold storage system.

2. The method for implementing cold storage control by edge computing according to claim 1 is characterized in that: Depending on whether the cold storage temperature is within the preset range, the refrigeration system is started and stopped using the if control logic.

3. The method for implementing cold storage control by edge computing according to claim 1 is characterized in that: Step S4 specifically includes: Step S401: Use the edge computing terminal as the MQTT data publishing terminal and set the cold storage temperature data with time series to be published at regular intervals; Step S402: Run Python related programs to build an MQTT client on the cloud server; Step S403: The cloud server connects to the MQTT Broker and subscribes to the Topic that needs to receive data; Step S404: Run the python related program to automatically store the received data into the MySQL database; Step S405: Performing algorithm analysis on the temperature on the cloud server side and outputting valve opening parameters; Step S406: The cloud server acts as the MQTT data publishing end, and the edge computing terminal acts as the MQTT data receiving end to accept control parameters and transmit the parameters to the control terminal via a wired network to precisely control the cold storage system.

4. The method for implementing cold storage control by edge computing according to claim 1 is characterized in that: Step S5, using a recurrent neural network algorithm on the cloud server to perform in-depth analysis on the data and predict future temperature changes, includes the following steps: Step S501: Establish a three-layer RNN recurrent neural network model, including an input layer, a hidden layer, and an output layer, and the number of neurons is 3-10-1; Step S502: The cold storage temperature, superheat, and valve opening are used as the original input data, and the original input data is normalized using minimum-maximum scaling. For each feature x, the conversion formula is: Among them, x' i (t) represents the eigenvalue of the i-th feature after normalization, x i (t) represents the original value of the i-th feature, min(x i ) and max(x i ) represent the minimum and maximum original data of the i-th feature respectively. After conversion, all eigenvalues ​​will fall within the range of [0,1]; Step S503: The activation function is the sigmoid function, and the error is calculated using MSE. The MSE calculation formula is: Among them, N represents the number of samples, oi represents the output result of the RNN prediction model, and yi represents the true value; Step S504: The first 80% of the processed data is used as training data, and the last 20% is used as verification data. When the error of the RNN prediction model is less than 0.1°C, the training of the RNN model is completed; Step S505: Calculate the predicted temperature of the cold storage in the next time period.

5. The method for implementing cold storage control by edge computing according to claim 1 is characterized in that: In step S4, the user checks historical data or monitors real-time data by using the database.

6. The method for implementing cold storage control by edge computing according to claim 1 is characterized in that: In step S2, several groups of temperature and humidity sensors are deployed to control the temperature of different areas in the cold storage to monitor and control the temperature of each area.

7. The method for implementing cold storage control by edge computing according to claim 1 is characterized in that: In step S2, step S1: connect various data to the edge computing terminal through a wired network connection.

8. The method for implementing cold storage control by edge computing according to any one of claims 1 to 7, characterized in that: The edge computing terminal may be FBOX or IVP09A.

Images