Tobacco leaf loosening and moisture regaining roller on-line temperature measuring system based on RFID passive technology

Abstract

The invention relates to the technical field of online temperature measurement in a loosening and moisture regaining roller, in particular to an online temperature measurement system for a tobacco leaf loosening and moisture regaining roller based on an RFID passive technology. According to the technical scheme, the system comprises an RFID passive temperature measurement label module, an RFID reader-writer module, an edge calculation and intelligent regulation and control module and a visualization and early warning module. According to the invention, cable-free temperature measurement in the roller is realized through the RFID passive temperature measurement technology, and the high-temperature-resistant magnetic adsorption tag and the intelligent reader-writer are combined, so that the problem of accurate temperature sensing under complex working conditions is solved; the edge calculation module is fused with a multi-dimensional compensation algorithm and predictive control, so that the temperature and humidity regulation and control precision is remarkably improved; the self-cleaning and energy recovery system guarantees long-acting and stable operation of equipment, the block chain evidence storage technology realizes credible tracing of whole-process data, and the whole scheme improves the process stability, reduces the operation and maintenance cost and enhances the production safety.

Description

Technical Field

[0001] The present invention relates to the technical field of on-line temperature measurement inside a loose tobacco re-drying drum, and particularly to an on-line temperature measurement system for a loose tobacco re-drying drum based on RFID passive technology. Background Art

[0002] In the cigarette production process, the loose tobacco re-drying process has a crucial impact on the quality of tobacco slices. In this process, the cut tobacco slices are usually placed in a rotating drum, and the tobacco slices are loosened and fully moistened under the action of a plow plate. Temperature control is one of the key parameters to ensure the quality of tobacco slices. However, due to the high-temperature and high-humidity semi-closed environment inside the loose tobacco re-drying drum, traditional temperature measurement methods (such as thermocouples or infrared thermometers) are difficult to work stably. Therefore, there is an urgent need for a wireless temperature measurement system that can work reliably inside the loose tobacco re-drying drum.

[0003] Therefore, this application proposes an on-line temperature measurement system for a loose tobacco re-drying drum based on RFID passive technology. Summary of the Invention

[0004] The object of the present invention is to address the problem in the background art that due to the high-temperature and high-humidity semi-closed environment inside the loose tobacco re-drying drum, traditional temperature measurement methods such as thermocouples or infrared thermometers are difficult to work stably, and to propose an on-line temperature measurement system for a loose tobacco re-drying drum based on RFID passive technology.

[0005] The technical solution of the present invention: An on-line temperature measurement system for a loose tobacco re-drying drum based on RFID passive technology, comprising:

[0006] An RFID passive temperature measurement tag module, uniformly fixed on the inner wall of the drum or the surface of the plow plate, including a high-temperature resistant encapsulation housing, a temperature sensor for real-time temperature perception, and a passive RFID chip for storing temperature data and transmitting it through radio frequency signal reflection modulation;

[0007] An RFID reader / writer module, deployed outside the drum outlet, including:

[0008] A spiral antenna array for multi-tag group reading in a rotating state;

[0009] An adaptive frequency modulation unit for dynamically adjusting the communication frequency band according to the drum rotation speed;

[0010] An edge computing and intelligent control module for processing temperature data and adjusting the steam valve and hot air temperature in real time, and triggering an alarm based on a temperature threshold;

[0011] A visualization and warning module for providing real-time temperature display and over-limit alarm functions.

[0012] Optionally, the RFID passive temperature measurement tag module is fixed to the inner wall of the drum through a high-temperature resistant magnetic adsorption structure, specifically including:

[0013] A flexible ceramic substrate, made of silicon nitride ceramic, with a thickness of 2 mm, and a bending radius adapted to the curvature of the inner wall of the drum (R = 500 - 800 mm);

[0014] A rare earth permanent magnet, embedded inside the substrate, made of neodymium iron boron material (N52 grade), with a magnetic suction force ≥ 20 N, a mechanical vibration frequency tolerance ≤ 200 HZ, and a magnetic force attenuation rate < 1% / year;

[0015] A high-temperature resistant adhesive layer, coated on the contact surface between the substrate and the drum, using aluminosilicate fiber glue (temperature resistance ≥ 300 °C), and the shear strength after curing ≥ 5 MPa.

[0016] Optionally, the RFID reader module further includes:

[0017] A multi-channel diversity receiving unit, using 2×2 multiple-input multiple-output (MIMO) space-time coding technology, and improving the signal reception signal-to-noise ratio through the maximum ratio combining algorithm (gain ≥ 6 dB);

[0018] A self-calibration unit, with a built-in reference tag (accuracy ±0.1 °C), automatically triggering the calibration process every 30 minutes: the reader reads the temperature value of the reference tag, compares it with the preset reference value, and if the deviation > 0.5 °C, corrects the readings of other temperature measurement tags through the linear regression algorithm.

[0019] Optionally, the edge computing and intelligent control module specifically includes:

[0020] A data preprocessing unit, using a sliding window filtering algorithm to denoise the original temperature data, and eliminating the data time series deviation caused by the rotation of the drum through the timestamp alignment technology;

[0021] A dynamic PID controller, based on the temperature distribution data of 8 sectors divided inside the drum (3 temperature measurement tags are installed in each sector), calculating the steam valve opening adjustment amount in real time (accuracy ±1%), and optimizing the hot air temperature setting value through the fuzzy logic algorithm;

[0022] A machine learning model, using a long short-term memory network (LSTM), training historical temperature data (time span ≥ 6 months) to predict the temperature trend within the next 10 minutes, and triggering the process parameter adjustment instruction in advance when the prediction deviation > 5%;

[0023] A distributed temperature compensation algorithm, based on the measurement data of the drum rotation angle sensor and the material thickness radar, establishing a compensation model:

[0024] T 修正 =T 原始 +k1·θ+k2·d

[0025] Among them, k1 is the angle compensation coefficient, k2 is the material thickness compensation coefficient, θ is the roller rotation angle, and d is the material stacking thickness;

[0026] The process parameter optimization database stores the optimal temperature and humidity curves for different tobacco varieties (such as flue-cured tobacco and air-dried tobacco), and recommends control strategies in real time through fuzzy logic matching. The matching rules include:

[0027] The correlation mapping between tobacco leaf moisture content (17%-21%) and temperature setting value;

[0028] Drum speed (5-15rpm) and steam flow (0.5-2.0m 3 / h) dynamic relationship table.

[0029] Optionally, the visualization and early warning module includes the following submodules:

[0030] The three-dimensional thermal map display interface reconstructs the temperature field distribution inside the drum based on the finite element analysis algorithm and supports dynamic rotation of the viewing angle;

[0031] Sound and light alarm unit, when the temperature exceeds the predicted deviation by more than 5%, it triggers the buzzer (frequency 2-4kHz) and LED flashing (red warning), and links the PLC to control the drum to stop for protection;

[0032] The blockchain evidence storage unit uploads temperature data and operation records to the chain through the Hyperledger Fabric framework, generates an unalterable hash value, and realizes the full life cycle traceability of the process

[0033] The blockchain evidence storage unit is used to perform the following operations:

[0034] Receive temperature data from the RFID reader module, pre-process the data, and use the SHA-256 hash algorithm to generate a 256-bit hash value;

[0035] Bind the hash value with the PLC operation record and timestamp (accuracy 1ms), and write it into the private chain block. The block header includes the hash value of the previous block and the Merkle tree root.

[0036] Provides a Representational State Transfer Application Programming Interface (RESTful API) to support third-party audit platforms to access blockchain data through digital certificate authentication;

[0037] A cross-platform data interaction unit, constructed based on the OPC Unified Architecture (OPC UA) protocol, and specifically implemented as follows:

[0038] Define the following standardized data nodes:

[0039] Real-time temperature node: including label ID (16-bit encoding), temperature value (floating-point type, resolution 0.1 °C), timestamp (ISO8601 format);

[0040] Alarm status node: including alarm code (4-bit hexadecimal), alarm level (enumeration type: warning / serious / urgent), trigger timestamp;

[0041] Process parameter node: including steam valve opening degree (percentage, 0 - 100%), hot air temperature set value (integer, unit °C).

[0042] Optionally, the system further includes:

[0043] A self-cleaning module, configured on the surface of the reader antenna, including:

[0044] A high-pressure air nozzle, with a spraying pressure ≥ 0.5 MPa, spraying once every 10 minutes (duration 2 s), and the air temperature ≥ 80 °C to prevent condensation;

[0045] A micro camera, detecting the pollution degree of the antenna surface through image recognition, and triggering a cleaning instruction when the stain coverage rate > 10%;

[0046] An energy recovery unit, including:

[0047] A permanent magnet generator installed at the end of the roller shaft, with an output power ≥ 50 W;

[0048] A supercapacitor energy storage module (capacity ≥ 100 F), converting kinetic energy into electrical energy and providing auxiliary power supply for the reader module (endurance time ≥ 8 h).

[0049] Optionally, the error correction formula of the linear regression algorithm in the self-calibration unit is:

[0050]

[0051] where: ΔT is the correction amount, α is the reference temperature weight coefficient, β is the label temperature distribution weight coefficient, T 基准 is the preset reference label reference temperature value, T 参考 is the measured temperature of the current reference label, T 标签，i is the original temperature value of the i-th temperature measurement label, and n is the number of effective temperature measurement labels.

[0052] Optionally, the dynamic calculation formula of the material thickness compensation coefficient k2 in the distributed temperature compensation algorithm is:

[0053]

[0054] where: ΔT max is the maximum allowable temperature deviation, δ max is the maximum threshold of the material thickness, γ is the attenuation coefficient, d is the real-time measured material accumulation thickness; when d > δ max k2 is calculated according to the saturation value.

[0055] Optionally, the generation formula of the hash value in the blockchain evidence storage unit is:

[0056] H 区块 = SHA-256(T 均值 ||σ T ||t 时间戳 || PLC operation code)

[0057] where: H 区块 is the block hash value, T 均值 is the average temperature of all temperature measurement tags in the drum; σ T is the temperature standard deviation, t 时间戳 is the data acquisition time (accuracy 1ms); || represents the data splicing operation; SHA-256 is the hash algorithm, which outputs a 256-bit fixed-length string.

[0058] Optionally, the cleaning efficiency η of the self-cleaning module is evaluated by the following formula:

[0059]

[0060] where: A 残留 is the area of residual stains on the antenna surface after cleaning, A 初始 is the area covered by stains before cleaning, P is the high-pressure air flow injection pressure, t is the duration of a single injection, μ is the adhesion coefficient of the air flow and the pollutant, ρ is the air density, v is the air flow velocity, P0 is the reference pressure, α is the environmental temperature correction coefficient; T 环境 is the external environmental temperature of the drum. When η < 85%, the repeated cleaning process is triggered until η ≥ 90%.

[0061] Compared with the prior art, the present application includes at least one of the following beneficial technical effects:

[0062] By adopting the RFID backscatter modulation technology, the temperature measurement tags in the drum are powered passively, breaking through the wiring limitations of traditional wired temperature measurement and adapting to complex industrial environments.

[0063] Integrate the angle compensation and material thickness compensation algorithms, combine with the LSTM prediction model to achieve precise temperature control and early warning, and optimize the steam valve adjustment strategy.

[0064] Ensure stable communication in complex electromagnetic environments through spiral antenna arrays, frequency hopping spread spectrum technology, and dynamic calibration of reference tags; the self-cleaning system and shaft-end power generation and energy storage technology enhance the continuous operation ability of the equipment.

[0065] Ensure production safety through dual-redundancy monitoring of temperature thresholds and a rapid alarm response mechanism; blockchain evidence storage technology enables data immutability and supports full-life cycle traceability.

[0066] Build a standardized data interaction interface based on the OPC UA protocol to support seamless integration with third-party systems and optimize the matching of process parameters.

[0067] Simplify the tag replacement process through magneto-adhesive flexible installation design, and the self-cleaning system reduces the need for manual maintenance, significantly reducing operation and maintenance costs.

[0068] This invention realizes cable-free temperature measurement inside the drum through RFID passive temperature measurement technology. Combining high-temperature resistant magneto-adhesive tags and intelligent readers / writers, it breaks through the problem of precise temperature perception under complex working conditions; the edge computing module integrates multi-dimensional compensation algorithms and predictive control to significantly improve the accuracy of temperature and humidity regulation; the self-cleaning and energy recovery system ensures the long-term stable operation of the equipment, and blockchain evidence storage technology realizes trustworthy traceability of the whole process data. The overall solution improves process stability, reduces operation and maintenance costs, and enhances production safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 It is a schematic diagram of the principle of an on-line temperature measurement system for a tobacco leaf loosening and rewetting drum based on RFID passive technology. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0070] The following further describes the technical solutions of the present invention in conjunction with the drawings and specific embodiments.

[0071] Embodiment 1

[0072] As Figure 1 shown, an on-line temperature measurement system for a tobacco leaf loosening and rewetting drum based on RFID passive technology proposed by the present invention includes an RFID passive temperature measurement tag module, an RFID reader / writer module, an edge computing and intelligent control module, and a visualization and warning module. Each module will be described in detail below.

[0073] I. The RFID passive temperature measurement tag module is evenly fixed on the inner wall of the loosening and rewetting drum or the surface of the platen. Each temperature measurement tag includes:

[0074] A high-temperature resistant ceramic encapsulation housing, made of alumina ceramic material, with a polytetrafluoroethylene anti-metal interference coating with a thickness of 50 - 100 μm on the surface, a temperature resistance grade ≥ 150 °C, and a mechanical shock resistance strength ≥ 50 G;

[0075] An ultra-low power consumption temperature sensor, using a MEMS thermopile structure, with a measurement accuracy of ±0.3 °C, a sampling frequency that can be dynamically adjusted (1 - 10 Hz), and supporting a sleep mode to reduce power consumption;

[0076] A passive RFID chip, integrating an EEPROM memory (capacity ≥ 2 KB), storing temperature data and tag ID information, and using backscatter modulation technology to encode the temperature data into a radio frequency signal;

[0077] The RFID passive temperature measurement tag module is fixed to the inner wall of the drum through a high-temperature resistant magnetic adsorption structure, specifically including:

[0078] A flexible ceramic substrate, made of silicon nitride ceramic, with a thickness of 2 mm, and a bending radius adapted to the curvature of the inner wall of the drum (R = 500 - 800 mm);

[0079] A rare earth permanent magnet, embedded inside the substrate, using neodymium iron boron material (N52 grade), with a magnetic suction force ≥ 20 N, a mechanical vibration resistance frequency ≤ 200 Hz, and a magnetic force attenuation rate < 1% / year;

[0080] A high-temperature resistant adhesive layer, coated on the contact surface between the substrate and the drum, using aluminosilicate fiber glue (temperature resistance > 300 °C), and having a shear strength ≥ 5 MPa after curing.

[0081] In the RFID passive temperature measurement tag module, alumina ceramic + polytetrafluoroethylene coating is used to achieve a wide temperature range operation of -40 °C - 300 °C, meeting the requirements of the 150 °C high-temperature environment for tobacco leaf processing. The MEMS thermopile sensor supports 1 - 10 Hz dynamic sampling, and with the sleep mode, the single tag life is extended to more than 5 years. The flexible silicon nitride substrate is adapted to the R500 - R800 mm curved surface, the N52 grade neodymium iron boron magnet provides a 20 N adsorption force, and the aluminosilicate glue achieves a 5 MPa shear strength at 300 °C.

[0082] Second, the RFID reader module, deployed outside the drum outlet, includes:

[0083] A spiral antenna array, composed of copper spiral coils, with a signal gain ≥ 8 dBi, a penetration intensity ≥ 10 dB, supporting multi-tag group reading (maximum concurrent tag number ≥ 200), and a coverage radius ≥ 8 m;

[0084] The adaptive frequency modulation unit dynamically adjusts the communication frequency band (860 - 960 MHz) based on the real-time rotation speed signal fed back by the roller rotation speed sensor, and adopts the frequency hopping spread spectrum technology (FHSS) to suppress electromagnetic interference, with a frequency hopping interval ≤ 1 ms;

[0085] The multi-channel diversity receiving unit adopts the 2×2 MIMO space-time coding technology and improves the signal reception signal-to-noise ratio through the maximum ratio combining algorithm (gain ≥ 6 dB);

[0086] The self-calibration unit has a built-in reference tag (accuracy ±0.1 °C) and automatically triggers the calibration process every 30 minutes: the reader reads the temperature value of the reference tag and compares it with the preset reference value. If the deviation > 0.5 °C, the readings of other temperature measurement tags are corrected through the linear regression algorithm. The error correction formula of the linear regression algorithm in the self-calibration unit is:

[0087]

[0088] where: ΔT is the correction amount, α is the reference temperature weight coefficient, β is the tag temperature distribution weight coefficient,

[0089] T 基准 is the preset reference temperature value of the reference tag, T 参考 is the measured temperature of the current reference tag, T 标签，i is the original temperature value of the i-th temperature measurement tag, and n is the number of effective temperature measurement tags.

[0090] The RFID reader module, through the spiral antenna array + 8 dBi gain + frequency hopping spread spectrum technology, achieves an 8-meter coverage radius at a 10 dB penetration intensity. It has a built-in ±0.1 °C reference tag and realizes dynamic correction at the 0.5 °C level through the linear regression algorithm to eliminate the influence of environmental temperature drift. It supports concurrent reading of 200 tags through the 2×2 MIMO technology to meet the fast data acquisition requirements at a roller rotation speed of 15 rpm

[0091] III. Edge computing and intelligent control module, including:

[0092] The data preprocessing unit uses the sliding window filtering algorithm to denoise the original temperature data and eliminates the data time series deviation caused by the roller rotation through the timestamp alignment technology;

[0093] The dynamic PID controller, based on the temperature distribution data of 8 sectors divided in the roller (3 temperature measurement tags are installed in each sector), calculates the steam valve opening adjustment amount in real time (accuracy ±1%) and optimizes the hot air temperature set value through the fuzzy logic algorithm;

[0094] The machine learning model uses an LSTM neural network to train historical temperature data (time span ≥ 6 months) to predict the temperature trend within the next 10 minutes. When the prediction deviation is greater than 5%, the process parameter adjustment instruction is triggered in advance;

[0095] The edge computing and intelligent control module further includes:

[0096] Distributed temperature compensation algorithm, based on the drum rotation angle sensor and material thickness radar measurement data, establishes a compensation model:

[0097] T 修正 =T 原始 +k1·θ+k2·d

[0098] Wherein, k1 is the angle compensation coefficient (0.05°C / °), k2 is the material thickness compensation coefficient (0.1°C / mm), θ is the drum rotation angle, and d is the material stacking thickness; the dynamic calculation formula of the material thickness compensation coefficient k2 in the distributed temperature compensation algorithm is:

[0099]

[0100] Where: ΔT max is the maximum allowable temperature deviation, δ max is the maximum threshold of material thickness, γ is the attenuation coefficient, and d is the material accumulation thickness measured in real time; when d>δ max When k2 is calculated according to the saturation value.

[0101] In the edge computing and intelligent control module, the multimodal compensation algorithm integrates the angle sensor (0.05℃ / ° compensation) and the radar thickness measurement (0.1℃ / mm compensation) to establish a three-dimensional temperature compensation model. The LSTM neural network is used to predict the temperature trend within 10 minutes, and the fuzzy PID controller is used to adjust the steam valve with an accuracy of ±1%. Through parallel processing of 8-sector temperature data, the response time is less than 50ms, and a data throughput of 200 tags / second is supported.

[0102] 4. Visualization and early warning module, including:

[0103] The three-dimensional thermal map display interface reconstructs the temperature field distribution inside the drum based on the finite element analysis algorithm and supports dynamic rotation of the viewing angle;

[0104] The sound and light alarm unit triggers a buzzer (frequency 2-4kHz) and LED flashes (red warning) when the temperature exceeds the preset process range (65-95℃) or the predicted deviation is greater than 5%, and the PLC is linked to control the drum to stop for protection:

[0105] The blockchain evidence storage unit uploads temperature data and operation records to the blockchain through the Hyperledger Fabric framework, generates an immutable hash value, and realizes the full life cycle traceability of the process.

[0106] The visualization and early warning module includes the following sub-modules:

[0107] (a) The blockchain evidence storage unit is used to perform the following operations:

[0108] Receive temperature data from the RFID reader module, preprocess the data, and generate a 256-bit hash value using the SHA-256 hash algorithm; the generation formula of the hash value in the blockchain evidence storage unit is:

[0109] H 区块 =SHA-256(T 均值 ||σ T ||t 时间戳 ||PLC operation code)

[0110] Where: H 区块 is the block hash value, T 均值 is the average temperature of all temperature measurement tags in the drum;

[0111] σ T is the temperature standard deviation, t 时间戳 is the data acquisition time (accuracy 1ms); || represents the data concatenation operation; SHA-256 is the hash algorithm, and the output is a 256-bit fixed-length string.

[0112] The process parameter optimization database stores the best temperature and humidity curves for different tobacco leaf varieties (such as flue-cured tobacco and sun-cured tobacco), and recommends real-time control strategies through fuzzy logic matching. The matching rules include:

[0113] The correlation mapping between the moisture content of tobacco leaves (17%-21%) and the temperature set value;

[0114] The dynamic relationship table between the drum rotation speed (5-15rpm) and the steam flow rate (0.5-2.0m 3 / h);

[0115] Bind the hash value with the PLC operation record and the timestamp (accuracy 1ms), and write it into the private chain block. The block header includes the previous block hash value and the Merkle root;

[0116] Provide a RESTful API interface to support third-party audit platforms to access blockchain data through digital certificate authentication;

[0117] (b) The cross-platform data interaction unit is built based on the OPC UA protocol and specifically realizes:

[0118] Define the following standardized data nodes:

[0119] Real-time temperature node: including label ID (16-bit encoding), temperature value (floating-point type, resolution 0.1 °C), timestamp (ISO 8601 format);

[0120] Alarm status node: including alarm code (4-bit hexadecimal), alarm level (enumeration type: warning / serious / urgent), trigger timestamp;

[0121] Process parameter node: including steam valve opening (percentage, 0-100%), hot air temperature set value (integer, unit °C)

[0122] The visualization and early warning module has the following benefits:

[0123] Industrial-grade safety design: Dual-redundancy monitoring of temperature thresholds from 65 to 95 °C, sound and light alarm response time of 2-4 kHz < 100 ms;

[0124] Blockchain traceability system: SHA-256 hash algorithm + Hyperledger Fabric framework, realizing second-level on-chain storage and certification of temperature data;

[0125] Intelligent decision support: Based on a fuzzy matching model with a moisture content of 12-18%, providing an optimal temperature and humidity curve recommendation strategy.

[0126] Embodiment 2

[0127] Based on Embodiment 1, the system further includes:

[0128] Self-cleaning module, configured on the surface of the reader antenna, including:

[0129] High-pressure air nozzle, with a spraying pressure ≥ 0.5 MPa, spraying once every 10 minutes (duration 2 s), and air temperature ≥ 80 °C to prevent condensation;

[0130] Miniature camera, detecting the pollution degree of the antenna surface through image recognition, and triggering a cleaning instruction when the stain coverage rate > 10%;

[0131] Energy recovery unit, including:

[0132] Permanent magnet generator installed at the end of the roller shaft, with an output power ≥ 50 W;

[0133] Supercapacitor energy storage module (capacity ≥ 100 F), converting kinetic energy into electrical energy and providing auxiliary power supply for the reader module (endurance time ≥ 8 h).

[0134] The cleaning efficiency η of the self-cleaning module is evaluated by the following formula:

[0135]

[0136] Wherein: A 残留 is the area of residual stains on the antenna surface after cleaning, A 初始 is the area covered by stains before cleaning, P is the injection pressure of the high-pressure air flow, t is the duration of a single injection, μ is the adhesion coefficient between the air flow and the pollutants, ρ is the air density, v is the air flow velocity, P0 is the reference pressure, and α is the environmental temperature correction coefficient; T 环境 is the external environmental temperature of the drum. When η < 85%, the repeated cleaning process is triggered until η ≥ 90%.

[0137] In this embodiment, for high-pressure air flow cleaning: a 0.5 MPa @ 80 °C air flow realizes a 10-minute cycle cleaning. Combined with image recognition (> 10% pollution trigger), the antenna cleaning efficiency ≥ 90%; the nozzle design is optimized using the Bernoulli equation, and through the cleaning efficiency formula η = 1 - Aresidual / Ainitial, the dynamic balance between energy consumption and cleaning effect is achieved;

[0138] In the energy recovery system, the 50 W output of the permanent magnet generator is combined with a 100 F supercapacitor to achieve a kinetic - electrical energy conversion efficiency ≥ 85%; the 8-hour energy storage support system can complete data preservation and safe shutdown in case of sudden power failure, meeting the continuous production requirements of the tobacco industry.

[0139] The system of this embodiment solves the wiring problem of traditional wired temperature measurement through passive temperature measurement technology. The self-cleaning and energy recovery technologies improve the operation reliability of the equipment, and the blockchain evidence storage meets the traceability requirements of the tobacco industry. The measured data shows that the temperature control accuracy of the system is improved and the energy consumption is low.

[0140] The above specific embodiments are merely several alternative embodiments of the present invention. Based on the technical solution of the present invention and the relevant revelations of the above embodiments, those skilled in the art can make various alternative improvements and combinations to the above specific embodiments.

Claims

1. An on-line temperature measurement system for a tobacco leaf loosening and tempering drum based on RFID passive technology, characterized in that, include: The RFID passive temperature measurement tag module is evenly fixed on the inner wall of the drum or the surface of the copy board, including a high-temperature resistant packaging shell, a temperature sensor for real-time temperature sensing, and a passive RFID chip that stores temperature data and transmits it through radio frequency signal reflection modulation; The RFID reader-writer module is deployed outside the drum outlet and includes: Helical antenna array, used for group reading of multiple tags in a rotating state; Adaptive frequency modulation unit, dynamically adjusts the communication frequency band according to the drum speed; Edge computing and intelligent control module, used to process temperature data and adjust steam valves and hot air temperature in real time, and trigger warnings based on temperature thresholds; The visualization and early warning module provides real-time temperature display and over-limit alarm functions.

2. The on-line temperature measurement system for the tobacco leaf loosening and tempering drum based on RFID passive technology according to claim 1, wherein The RFID passive temperature measurement tag module is fixed to the inner wall of the drum through a high temperature resistant magnetic adsorption structure, specifically comprising: Flexible ceramic substrate, made of silicon nitride ceramic, with a bending radius adapted to the curvature of the inner wall of the drum; Rare earth permanent magnets are embedded inside the substrate; The high temperature resistant adhesive layer is applied on the contact surface between the substrate and the roller.

3. The on-line temperature measurement system for the tobacco leaf loosening and tempering drum based on RFID passive technology according to claim 1, characterized in that, The RFID reader-writer module also includes: Multi-channel diversity receiving unit, using 2×2 multiple-input multiple-output space-time coding technology; The self-calibration unit has a built-in reference tag, which automatically triggers the calibration process every 30 minutes: the reader reads the temperature value of the reference tag and compares it with the preset benchmark value. If the deviation is greater than 0.5°C, the readings of other temperature measurement tags are corrected through a linear regression algorithm.

4. The on-line temperature measuring system for the tobacco leaf loosening and tempering drum based on the RFID passive technology according to claim 1, characterized in that, The edge computing and intelligent control module specifically includes: The data preprocessing unit uses a sliding window filtering algorithm to denoise the original temperature data and eliminates the data timing deviation caused by the drum rotation through the timestamp alignment technology; Dynamic PID controller, based on the temperature distribution data of the drum divided into 8 sectors, calculates the steam valve opening adjustment in real time and optimizes the hot air temperature setting value through fuzzy logic algorithm; The machine learning model uses a long short-term memory network to train historical temperature data to predict the temperature trend within the next 10 minutes. When the prediction deviation is greater than 5%, the process parameter adjustment instruction is triggered in advance. Distributed temperature compensation algorithm, based on the drum rotation angle sensor and material thickness radar measurement data, establishes a compensation model: T 修正 = T 原始 + k1·θ + k2·d Among them, k1 is the angle compensation coefficient, k2 is the material thickness compensation coefficient, θ is the roller rotation angle, and d is the material stacking thickness; The process parameter optimization database stores the best temperature and humidity curves for different tobacco varieties and recommends control strategies in real time through fuzzy logic matching. The matching rules include: The correlation mapping between tobacco leaf moisture content and temperature setting value; Dynamic relationship table between drum speed and steam flow rate.

5. The on-line temperature measurement system for the tobacco leaf loosening and tempering drum based on the RFID passive technology according to claim 1, wherein, The visualization and early warning module includes the following submodules: The three-dimensional thermal map display interface reconstructs the temperature field distribution inside the drum based on the finite element analysis algorithm and supports dynamic rotation of the viewing angle; Sound and light alarm unit, when the temperature exceeds the predicted deviation by more than 5%, it triggers the buzzer and LED flashing, and links the PLC to control the drum to stop protection; The blockchain evidence storage unit uploads temperature data and operation records to the chain through the Hyperledger Fabric framework, generates an unalterable hash value, and traces the entire life cycle of the process; the blockchain evidence storage unit is used to perform the following operations: Receive temperature data from the RFID reader module, preprocess the data, and generate a 256-bit hash value using the SHA-256 hash algorithm; Bind the hash value to the PLC operation record and timestamp, and write it into the private chain block. The block header includes the previous block hash value and the Merkle root; Provide a Representational State Transfer (REST) Application Programming Interface (API) to support third-party audit platforms to access blockchain data through digital certificate authentication; A cross-platform data interaction unit, built based on the OPC Unified Architecture protocol, is specifically used to define the following standardized data nodes: Real-time temperature node: including tag ID, temperature value, and timestamp; Alarm status node: including alarm code, alarm level, and trigger timestamp; Process parameter node: including steam valve opening and hot air temperature set value.

6. The online temperature measurement system for the tobacco leaf loosening and tempering drum based on the RFID passive technology according to claim 1, wherein The system also includes: A self-cleaning module, configured on the surface of the reader antenna, including: A high-pressure air nozzle with a spraying pressure ≥ 0.5 MPa, spraying once every 10 minutes, and the air temperature ≥ 80 °C to prevent condensation; A micro camera, which detects the pollution degree of the antenna surface through image recognition and triggers a cleaning instruction when the stain coverage rate > 10%; An energy recovery unit, including: A permanent magnet generator installed at the end of the roller shaft; A supercapacitor energy storage module, which converts kinetic energy into electrical energy and provides auxiliary power supply for the reader module.

7. An on-line temperature measurement system for a tobacco leaf loosening and tempering drum based on RFID passive technology according to claim 3, characterized in that The error correction formula of the linear regression algorithm in the self-calibration unit is: Where: ΔT is the correction amount, α is the reference temperature weight coefficient, β is the tag temperature distribution weight coefficient, T 基准 is the preset reference tag reference temperature value, T 参考 is the measured temperature of the current reference tag, T 标签，i is the original temperature value of the i-th temperature measurement tag, and n is the number of effective temperature measurement tags.

8. The online temperature measurement system for the tobacco leaf loosening and tempering drum based on the RFID passive technology according to claim 1, wherein, The dynamic calculation formula of the material thickness compensation coefficient k2 in the distributed temperature compensation algorithm is: Where: ΔT max is the maximum allowable temperature deviation, δ max is the maximum threshold value of the material thickness, γ is the attenuation coefficient, and d is the real-time measured material accumulation thickness; when d > δ max k2 is calculated according to the saturation value.

9. The on-line temperature measuring system for the tobacco leaf loosening and tempering drum based on RFID passive technology according to claim 5, characterized in that, The generation formula of the hash value in the blockchain evidence storage unit is: H 区块 = SHA-256(T 均值 || σ T || t 时间戳 || PLC operation code) Where: H 区块 is the block hash value, T 均值 is the average temperature of all temperature measurement tags in the drum; σ T is the standard deviation of temperature, t 时间戳 is the data acquisition time; || represents the data concatenation operation; SHA-256 is the hash algorithm, which outputs a 256-bit fixed-length string.

10. An on-line temperature measuring system for a tobacco leaf loosening and tempering drum based on RFID passive technology according to claim 6, characterized in that, The cleaning efficiency η of the self-cleaning module is evaluated by the following formula: Where: A 残留 is the area of residual stains on the antenna surface after cleaning, A 初始 is the area covered by stains before cleaning, P is the injection pressure of the high-pressure air flow, t is the duration of a single injection, μ is the adhesion coefficient between the air flow and the pollutants, ρ is the air density, v is the air flow velocity, P0 is the reference pressure, α is the environmental temperature correction coefficient; T 环境 is the external environmental temperature of the drum. When η < 85%, the repeated cleaning process is triggered until η ≥ 90%.

Images