Traditional Chinese medicine decoction piece roasting end point on-line intelligent near-infrared detection device suitable for high-temperature environment and method thereof

Abstract

The invention provides a traditional Chinese medicine decoction piece roasting end point online intelligent near-infrared detection device suitable for a high-temperature environment and a method thereof, and relates to the technical field of traditional Chinese medicine processing process monitoring. Comprising a GSA104 type high-temperature-resistant online near infrared spectrum probe, a thermoelectric refrigeration TEC module, a composite heat dissipation system, a PLC intelligent control system, a data preprocessing module, a model analysis module, an abnormity early warning module, a data storage module and a remote monitoring module. The GSA104 type high-temperature-resistant online near-infrared spectrum probe is integrally mounted in processing equipment and is used for acquiring near-infrared diffuse reflection spectrum data of materials in real time; the thermoelectric refrigeration active temperature control module is combined with the collaborative design of an air cooling and water cooling composite heat dissipation system, a multi-stage precise temperature control system is constructed, heat radiation and heat conduction interference of a 300 DEG C high-temperature environment can be effectively isolated, a stable working microenvironment is created for a GSA104 type high-temperature-resistant online near infrared spectrum probe, and continuous and stable operation of spectrum acquisition equipment is ensured.

Description

Technical Field

[0001] This invention relates to the field of monitoring technology for the processing of traditional Chinese medicine, and in particular to an online intelligent near-infrared detection device and method for the endpoint of roasting traditional Chinese medicine slices in high-temperature environments. Background Technology

[0002] In the processing of traditional Chinese medicine decoction pieces, the accurate determination of the processing endpoint has a decisive impact on the quality of the finished product and its clinical efficacy. Traditional methods mainly rely on the sensory experience of operators, which has significant drawbacks such as large subjective errors, poor reproducibility, low production efficiency, and difficulty in achieving automated control.

[0003] Existing near-infrared spectroscopy detection technology is susceptible to thermal radiation interference when applied to high-temperature environments, leading to decreased stability of spectral acquisition equipment and distorted spectral data, which cannot meet the stringent requirements for real-time online monitoring and precise control of the processing process.

[0004] Therefore, there is an urgent need to develop an online near-infrared detection technology with excellent high-temperature resistance and stability to support the modernization, intelligent and precise processing of traditional Chinese medicine decoction pieces. Summary of the Invention

[0005] In view of the above-mentioned shortcomings of the prior art, the present invention provides an online intelligent near-infrared detection device and method for the endpoint of roasting Chinese herbal medicine slices in high-temperature environments.

[0006] This invention provides an online intelligent near-infrared detection device for the endpoint of roasting Chinese herbal medicine slices in high-temperature environments. Specifically, it includes: a GSA104 type high-temperature resistant online near-infrared spectral probe, a thermoelectric cooling (TEC) module, a composite heat dissipation system, a PLC intelligent control system, a data preprocessing module, a model analysis module, an anomaly early warning module, a data storage module, and a remote monitoring module. The GSA104 type high-temperature resistant online near-infrared spectral probe is integrated and installed inside the processing equipment for real-time acquisition of near-infrared diffuse reflectance spectral data of the materials. The thermoelectric cooling (TEC) module is tightly fitted to the heat source area of ​​the main control circuit board of the GSA104 type high-temperature resistant online near-infrared spectral probe. The composite heat dissipation system includes an air-cooling subsystem and a water-cooling subsystem. The air-cooling subsystem is connected to the heat dissipation area of ​​the thermoelectric cooling (TEC) module, the spectral sensor of the GSA104 type high-temperature resistant online near-infrared spectral probe, and the main control circuit board area. The system is domain-adaptive, with a water-cooling subsystem covering the key thermally sensitive areas of the GSA104 high-temperature resistant online near-infrared spectroscopy probe. The data preprocessing module, model analysis module, and data storage module are all installed in the server room. The data preprocessing module is connected to the PLC intelligent control system, and the model analysis module is connected to the data preprocessing module. Anomaly warning modules are partially installed at the processing equipment site and partially associated with remote terminals, while also connecting to the model analysis module. The data storage module is connected to the data preprocessing module, model analysis module, and anomaly warning module respectively. The remote monitoring module includes a web client and a mobile client, which are connected to the PLC intelligent control system, model analysis module, and data storage module respectively. The PLC intelligent control system communicates bidirectionally with the GSA104 high-temperature resistant online near-infrared spectroscopy probe, the thermoelectric cooling (TEC) module, and the composite heat dissipation system to achieve data transmission and command issuance.

[0007] Optionally, the core components of the GSA104 high-temperature resistant online near-infrared spectral probe are made of high-temperature resistant materials and optical elements. The spectral acquisition range is 780-2500nm, the acquisition frequency is 1 time / 2s, the data transmission delay is ≤50ms, it supports non-contact non-destructive testing, and it can work continuously and stably in an environment of ≤300℃. The optical window covers the representative material area inside the processing equipment to avoid mechanical damage.

[0008] Optionally, the thermoelectric cooling TEC module uses pulse width modulation (PWM) technology to control the working state of the cold / hot ends of the semiconductor thermoelectric element, has a cold / hot end polarity switching function, a cooling temperature range of -10℃ to room temperature, a temperature control accuracy of ±0.1℃, and a response time of ≤20ms, and is used to construct a local active cooling zone to ensure the stable performance of core electronic components in high-temperature environments.

[0009] Optionally, the air-cooled subsystem of the composite heat dissipation system consists of an air compressor, air ducts, and a heat dissipation fin array. The air compressor provides clean compressed air flow, which flows through the heat dissipation fins of the thermoelectric refrigeration (TEC) module, the spectral sensor of the GSA104 high-temperature resistant online near-infrared spectral probe, and the main control circuit board area. The air ducts are optimized to achieve forced convection heat dissipation, and the surface area of ​​the heat dissipation fins is increased by 30% compared with the traditional structure. During operation, a positive pressure difference is formed inside to prevent external hot air from penetrating. The water-cooled subsystem adopts a stainless steel jacket structure, which tightly wraps the key heat-sensitive area of ​​the GSA104 high-temperature resistant online near-infrared spectral probe to form a physical thermal barrier. It connects a circulating water pump, a water tank, and a cooling device to form a closed-loop cooling water circuit. The cooling water flow rate is adjustable from 5-15L / min, and the inlet water temperature is ≤25℃. By continuously absorbing and discharging externally transferred heat, it blocks the heat source from penetrating into the internal cavity of the device.

[0010] Optionally, the PLC intelligent control system has a built-in high-performance processor with a computing speed of ≥1GHz, supports multi-protocol communication, communicates with the GSA104 high-temperature resistant online near-infrared spectral probe in real time to receive raw data, outputs temperature control commands to the composite heat dissipation system, and, based on the judgment results of the model analysis module, links and adjusts key process parameters such as the fire intensity and material turning speed of the processing equipment, with an adjustment accuracy of ≤±5%, to achieve closed-loop control.

[0011] Optionally, the data preprocessing module sequentially performs data cleaning, standardization, and noise reduction: During data cleaning, missing values, outliers, and duplicate values ​​in the spectral data are removed. When the missing value ratio is <5%, linear interpolation is used to fill the missing value; when the missing value ratio is ≥5%, the data is discarded and the discard status is recorded. Outliers are identified using the Grubbs test and processed using the nearest neighbor replacement method or data smoothing method according to the degree of deviation. Duplicate values ​​are directly deleted, retaining only one valid data. Data standardization uses the Standard Normal Variable Transform (SNV) method to unify the data units and eliminate analytical errors caused by data differences under different detection conditions. Data noise reduction uses wavelet transform noise reduction. Wavelet decomposition is performed on the preprocessed spectral data. The db4 wavelet basis function and the number of 3 decomposition levels are selected to obtain multi-scale wavelet coefficients. The wavelet coefficients are processed according to a preset threshold rule to suppress coefficients corresponding to noise. The processed wavelet coefficients are then reconstructed to obtain high-quality spectral data.

[0012] Optionally, the model analysis module uses a chemometrics algorithm combined with a BP neural network model to construct a processing endpoint analysis model. The input layer has four neurons, corresponding to the spectral feature values, material temperature, particle size parameters, and collection timestamp data processed by the data preprocessing module. The hidden layer has two layers, with 12 neurons per layer by default. The output layer has three neurons, corresponding to the three categories of judgment results: "insufficient processing", "moderate processing", and "over-processing". The model training uses the gradient descent method, using standard sample spectra of known processing degree and their corresponding reference values ​​of key component content and processing degree level as samples, adjusting the network weights and thresholds to make the prediction error ≤2%.

[0013] Optionally, the anomaly warning module integrates a signal acquisition chip, a warning drive circuit, and a wireless communication module, and is equipped with anomaly signal recognition software. This module establishes a data connection with the model analysis module through the signal acquisition chip to receive processing status judgment signals and equipment operation status feedback signals. It connects to the audible and visual warning actuator through the warning drive circuit and establishes a bidirectional data connection with the remote monitoring module through the wireless communication module to realize the transmission of anomaly warning information and event records. When equipment failures occur, such as the failure of the composite heat dissipation system or the interruption of data acquisition by the GSA104 high-temperature resistant online near-infrared spectral probe, an early warning is triggered.

[0014] Optionally, the data storage module is internally configured with a storage controller, a solid-state drive array, and a data cache chip, and is equipped with dedicated data management software and redundant backup programs. This module establishes multiple data input connections with core modules such as spectral acquisition, data preprocessing, model analysis, and anomaly early warning through the storage controller, and establishes a one-way data output connection with the remote monitoring module through the data cache chip. With the help of the solid-state drive array and redundant backup programs, it realizes secure archiving and rapid retrieval of data throughout the entire process.

[0015] Optionally, a detection method for an online intelligent near-infrared detection device for the endpoint of roasting Chinese herbal medicine slices in high-temperature environments includes the following steps: S1. Equipment Deployment: Integrate and install the GSA104 high-temperature resistant online near-infrared spectroscopy probe in a suitable location inside the processing equipment, such as a wok, ensuring that its optical window can effectively cover the representative material area and avoid mechanical damage; complete the assembly and debugging of the thermoelectric cooling TEC module, composite heat dissipation system and GSA104 high-temperature resistant online near-infrared spectroscopy probe, and reliably connect the cables of each module. S2. Parameter initialization: Set the equipment operating parameters through the remote monitoring module, including the spectral acquisition frequency of the GSA104 high-temperature resistant online near-infrared spectral probe, the temperature threshold of the composite heat dissipation system, the criteria for determining the processing endpoint, and the contact information of the early warning receivers. S3. Spectral Acquisition: During the processing, the GSA104 high-temperature resistant online near-infrared spectral probe is used to collect the near-infrared diffuse reflectance spectrum of the material sample in real time, obtain the raw spectral data, record the acquisition timestamp synchronously, and transmit the raw spectral data to the PLC intelligent control system. S4. Heat dissipation control: The PLC intelligent control system monitors the internal temperature T of the processing equipment in real time and dynamically activates the corresponding heat dissipation mechanism: when T < 100℃, only the water cooling subsystem of the composite heat dissipation system is activated; when 100℃ ≤ T < 150℃, both the air cooling subsystem and the water cooling subsystem of the composite heat dissipation system are activated; when T ≥ 150℃ up to 300℃, the thermoelectric refrigeration TEC module is activated for active cooling, and both the air cooling subsystem and the water cooling subsystem of the composite heat dissipation system are activated. S5. Spectral Preprocessing: The PLC intelligent control system transmits the raw spectral data to the data preprocessing module, which performs preprocessing on the raw spectral data, including multivariate scattering correction (MSC), standard normal variable transformation (SNV), Savitzky-Golay smoothing, derivative processing, etc., and sequentially completes data cleaning, standardization, and noise reduction to obtain high-quality spectral data. S6. Model Establishment and Calibration: Based on chemometric methods such as principal component analysis (PCA), partial least squares regression (PLSR), and support vector machine (SVM), a robust processing endpoint prediction or classification analysis model is established using standard sample spectra with known processing degrees and their corresponding reference values ​​such as the content of key components and the processing degree level. The model is then calibrated and validated using validation set data. S7. Online Prediction and Judgment: The real-time acquired spectrum processed by the data preprocessing module is input into the calibrated analysis model. The model analysis module calculates the predicted value or classification result of the current material's processing status and outputs three judgment results: "insufficient processing", "moderate processing" and "over-processing". S8. Endpoint Decision and Control: Based on the output results of the model analysis module, such as the predicted key indicator values ​​and the category of the processing stage, the system intelligently determines whether the preset processing endpoint has been reached. If it is determined that the processing is insufficient, the processing continues and the equipment maintains the current process parameters. If it is determined that the processing is appropriate, the PLC intelligent control system immediately outputs a control signal to terminate the processing and control the material to enter the next process. If it is determined that the processing is excessive, the PLC intelligent control system immediately terminates the processing, the abnormal warning module triggers the corresponding warning, and the abnormal data is recorded for subsequent traceability. S9. Data storage and monitoring: All raw data, processed data, judgment results, and equipment operation logs are stored in the data storage module in real time. Managers can view the real-time status, historical data, and reports through the web client or mobile client of the remote monitoring module.

[0016] The beneficial effects are as follows: By using a thermoelectric cooling active temperature control module, combined with the synergistic design of a combined air-cooling and water-cooling heat dissipation system, a multi-level precise temperature control system is constructed. This effectively isolates the interference of heat radiation and heat conduction in a 300-degree Celsius high-temperature environment, creating a stable working microenvironment for the GSA104 high-temperature resistant online near-infrared spectral probe. This ensures that the spectral acquisition equipment can operate stably and continuously in a wide temperature range from -20 degrees Celsius to 300 degrees Celsius, fundamentally solving the core problems of poor high-temperature adaptability and spectral data distortion in traditional near-infrared detection equipment.

[0017] With high detection accuracy and non-destructive real-time monitoring, relying on the high-performance spectral acquisition capabilities of the customized GSA104 high-temperature resistant online near-infrared spectral probe, combined with preprocessing techniques such as multivariate scattering correction and standard normal variable transformation, as well as chemometric models such as principal component analysis and partial least squares regression, it can achieve real-time, non-destructive, and accurate determination of the processing endpoint of Chinese herbal medicine pieces. The detection error is low, which is significantly better than traditional human sensory experience judgment, ensuring the uniformity of the quality of the herbal medicine pieces and the stability of clinical efficacy.

[0018] With a high degree of intelligence, it promotes the upgrading of production automation. Through the PLC intelligent control system, it realizes the closed-loop linkage of the entire process of spectrum acquisition, data processing, model analysis, endpoint determination and equipment control. It can automatically adjust key parameters such as the heat intensity of the stir-frying equipment and the material turning speed according to the endpoint determination results, without the need for manual intervention. At the same time, it is equipped with a remote monitoring module to support real-time status viewing and historical data traceability, which greatly improves production efficiency and meets the core requirements of the intelligent manufacturing development plan for the digital and intelligent upgrading of traditional Chinese medicine manufacturing.

[0019] With a wide range of applications, it meets the production needs of high-value-added medicinal slices. The technical solution can be directly applied to continuous and automated Chinese medicine processing production lines, especially suitable for the stringent requirements of precise control of the processing endpoint for high-value-added medicinal slices such as ginseng and deer antler. It does not require major equipment modifications for different types of medicinal slices, reducing enterprise upgrade costs and helping the Chinese medicine processing industry to develop in a standardized and large-scale manner. Attached Figure Description

[0020] Figure 1 A schematic diagram of the overall process structure of an embodiment of the present invention is shown; Figure 2 This diagram illustrates the connection structure of the GSA104 high-temperature resistant online near-infrared spectroscopy probe in an embodiment of the present invention. Figure 3 A schematic diagram of the connection structure of the thermoelectric refrigeration (TEC) module in an embodiment of the present invention is shown; Figure 4 A schematic diagram of the connection structure of the composite heat dissipation system in an embodiment of the present invention is shown; Figure 5A schematic diagram of the connection structure of the PLC intelligent control system in an embodiment of the present invention is shown; Figure 6 A schematic diagram of the connection structure of the data preprocessing module in an embodiment of the present invention is shown; Figure 7 A schematic diagram of the connection structure of the model analysis module in an embodiment of the present invention is shown; Figure 8 A schematic diagram of the connection structure of the anomaly warning module in an embodiment of the present invention is shown; Figure 9 A schematic diagram of the connection structure of the data storage module in an embodiment of the present invention is shown.

[0021] List of reference numerals in the attached diagram: 1. GSA104 type high temperature resistant online near-infrared spectroscopy probe; 2. Thermoelectric cooling TEC module; 3. Composite heat dissipation system; 301. Air cooling subsystem; 302. Water cooling subsystem; 4. PLC intelligent control system; 5. Data preprocessing module; 6. Model analysis module; 7. Anomaly early warning module; 8. Data storage module; 9. Remote monitoring module. Detailed Implementation

[0022] To make the objectives, solutions, and advantages of the technical solutions of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings of specific embodiments of the present invention.

[0023] Example 1: Please refer to the accompanying drawings in the instruction manual. Figures 1 to 9 As shown: This invention proposes an online intelligent near-infrared detection device and method for the endpoint of roasting Chinese herbal medicine slices in high-temperature environments. The device includes: a GSA104 type high-temperature resistant online near-infrared spectral probe 1, a thermoelectric cooling TEC module 2, a composite heat dissipation system 3, a PLC intelligent control system 4, a data preprocessing module 5, a model analysis module 6, an anomaly early warning module 7, a data storage module 8, and a remote monitoring module 9. The GSA104 type high-temperature resistant online near-infrared spectral probe 1 is integrated and installed inside the processing equipment for real-time acquisition of near-infrared diffuse reflectance spectral data of the materials. The thermoelectric cooling TEC module 2 is tightly fitted and installed in the heat source area of ​​the main control circuit board of the GSA104 type high-temperature resistant online near-infrared spectral probe 1. The composite heat dissipation system 3 includes an air-cooling subsystem 301 and a water-cooling subsystem 302. The air-cooling subsystem 301 is connected to the heat dissipation area of ​​the thermoelectric cooling TEC module 2, the spectral sensor area of ​​the GSA104 type high-temperature resistant online near-infrared spectral probe 1, and the main control circuit board area. Domain adaptation: The water-cooled subsystem 302 encloses the key thermally sensitive area of ​​the GSA104 high-temperature resistant online near-infrared spectroscopy probe 1; the data preprocessing module 5, model analysis module 6, and data storage module 8 are all installed in the server room. The data preprocessing module 5 is connected to the PLC intelligent control system 4, and the model analysis module 6 is connected to the data preprocessing module 5. The anomaly early warning module 7 is partially installed at the processing equipment site and partially associated with the remote terminal, and is also connected to the model analysis module 6; the data storage module 8 is connected to the data preprocessing module 5, model analysis module 6, and anomaly early warning module 7 respectively; the remote monitoring module 9 includes a web client and a mobile client, which are connected to the PLC intelligent control system 4, model analysis module 6, and data storage module 8 respectively; the PLC intelligent control system 4 communicates bidirectionally with the GSA104 high-temperature resistant online near-infrared spectroscopy probe 1, the thermoelectric cooling TEC module 2, and the composite heat dissipation system 3 to realize data transmission and command issuance.

[0024] Among them, the core components of the GSA104 high-temperature resistant online near-infrared spectral probe are made of high-temperature resistant materials and optical elements. The spectral acquisition range is 780-2500nm, the acquisition frequency is 1 time / 2s, the data transmission delay is ≤50ms, it supports non-contact non-destructive testing, and can work continuously and stably in an environment of ≤300℃. The optical window covers the representative material area inside the processing equipment to avoid mechanical damage.

[0025] Among them, the thermoelectric cooling TEC module 2 uses pulse width modulation (PWM) technology to control the working state of the cold / hot ends of the semiconductor thermoelectric element. It has the function of switching the polarity of the cold and hot ends, the cooling temperature range is -10℃ to room temperature, the temperature control accuracy is ±0.1℃, and the response time is ≤20ms. It is used to construct a local active cooling zone to ensure the stable performance of core electronic components in high-temperature environments.

[0026] Among them, the air-cooled subsystem 301 of the composite heat dissipation system 3 consists of an air compressor, air ducts and heat dissipation fin array. The air compressor provides clean compressed air flow, which flows through the heat dissipation fins of the thermoelectric refrigeration TEC module 2, the spectral sensor of the GSA104 high-temperature resistant online near-infrared spectral probe 1 and the main control circuit board area. The air duct is optimized to achieve forced convection heat dissipation. The surface area of ​​the heat dissipation fins is increased by 30% compared with the traditional structure. During operation, a positive pressure difference is formed inside to prevent external hot air from penetrating. The water-cooled subsystem 302 adopts a stainless steel jacket structure, which tightly wraps the key heat-sensitive area of ​​the GSA104 high-temperature resistant online near-infrared spectral probe 1 to form a physical thermal barrier. It connects the circulating water pump, water tank and cooling device to form a closed-loop cooling water circuit. The cooling water flow rate can be adjusted from 5-15L / min and the inlet water temperature is ≤25℃. It continuously absorbs and discharges external heat, blocking the heat source from penetrating into the internal cavity of the device.

[0027] Among them, the PLC intelligent control system 4 has a built-in high-performance processor with a computing speed of ≥1GHz, supports multi-protocol communication, communicates with the GSA104 high-temperature resistant online near-infrared spectral probe 1 in real time to receive raw data, outputs temperature control commands to the composite heat dissipation system 3, and, based on the judgment results of the model analysis module 6, links and adjusts key process parameters such as the fire intensity of the processing equipment and the material turning speed, with an adjustment accuracy of ≤±5%, realizing closed-loop control.

[0028] The data preprocessing module 5 sequentially performs data cleaning, standardization, and noise reduction: During data cleaning, missing values, outliers, and duplicate values ​​in the spectral data are removed. When the missing value ratio is <5%, linear interpolation is used to fill the missing value; when the missing value ratio is ≥5%, the data is discarded and the discard status is recorded. Outliers are identified using the Grubbs test and processed using the nearest neighbor replacement method or data smoothing method according to the degree of deviation. Duplicate values ​​are directly deleted, retaining only one valid data. Data standardization uses the Standard Normal Variable Transform (SNV) method to unify the data units and eliminate analytical errors caused by differences in data under different detection conditions. Data noise reduction uses wavelet transform noise reduction. Wavelet decomposition is performed on the preprocessed spectral data. The db4 wavelet basis function and the 3-level decomposition layer are selected to obtain multi-scale wavelet coefficients. The wavelet coefficients are processed according to the preset threshold rules to suppress the coefficients corresponding to noise. Then, the processed wavelet coefficients are reconstructed to obtain high-quality spectral data.

[0029] Among them, the model analysis module 6 uses a chemometric algorithm combined with a BP neural network model to construct a processing endpoint analysis model. The input layer has four neurons, corresponding to the spectral feature values, material temperature, particle size parameters, and collection timestamp data processed by the data preprocessing module 5. The hidden layer is set to two layers, with 12 neurons in each layer by default. The output layer has three neurons, corresponding to the three judgment results of "insufficient processing", "moderate processing" and "over-processing". The model training adopts the gradient descent method, using the standard sample spectrum of known processing degree and its corresponding reference value key component content and processing degree level as samples, adjusting the network weights and thresholds to make the prediction error ≤2%.

[0030] The abnormal early warning module 7 integrates a signal acquisition chip, an early warning drive circuit, and a wireless communication module, and is equipped with abnormal signal recognition software. This module establishes a data connection with the model analysis module 6 through the signal acquisition chip to receive the processing status judgment signal and the equipment operation status feedback signal. It connects to the audible and visual early warning actuator through the early warning drive circuit and establishes a two-way data connection with the remote monitoring module 9 through the wireless communication module to realize the transmission of abnormal early warning information and event records. When equipment failures occur, such as the failure of the composite heat dissipation system 3 or the interruption of data acquisition by the GSA104 high-temperature resistant online near-infrared spectral probe 1, an early warning is triggered.

[0031] The data storage module 8 is equipped with a storage controller, solid-state drive array, and data cache chip, and features dedicated data management software and redundant backup programs. This module establishes multiple data input connections with core modules such as spectral acquisition, data preprocessing, model analysis, and anomaly warning through the storage controller, and establishes a one-way data output connection with the remote monitoring module 9 through the data cache chip. With the help of the solid-state drive array and redundant backup programs, it achieves secure archiving and rapid retrieval of data throughout the entire process.

[0032] One method for detecting the endpoint of roasting Chinese herbal medicine slices in high-temperature environments using an online intelligent near-infrared detection device includes the following steps: S1. Equipment Deployment: Integrate and install the GSA104 high-temperature resistant online near-infrared spectral probe 1 in a suitable position inside the processing equipment such as a wok, ensuring that its optical window can effectively cover the representative material area and avoid mechanical damage; complete the assembly and debugging of the thermoelectric cooling TEC module 2, the composite heat dissipation system 3 and the GSA104 high-temperature resistant online near-infrared spectral probe 1, and reliably connect the cables of each module. S2. Parameter initialization: Set the equipment operating parameters through the remote monitoring module 9, including the spectral acquisition frequency of the GSA104 high-temperature resistant online near-infrared spectral probe 1, the temperature threshold of the composite heat dissipation system 3, the criteria for determining the processing endpoint, and the contact information of the early warning receivers. S3. Spectral Acquisition: During the processing, the GSA104 high-temperature resistant online near-infrared spectral probe 1 is used to collect the near-infrared diffuse reflectance spectrum of the material sample in real time, obtain the raw spectral data, record the acquisition timestamp synchronously, and transmit the raw spectral data to the PLC intelligent control system 4. S4. Heat dissipation control: The PLC intelligent control system 4 monitors the internal temperature T of the processing equipment in real time and dynamically activates the corresponding heat dissipation mechanism: when T < 100℃, only the water cooling subsystem 302 of the composite heat dissipation system 3 is activated; when 100℃ ≤ T < 150℃, both the air cooling subsystem 301 and the water cooling subsystem 302 of the composite heat dissipation system 3 are activated; when T ≥ 150℃ and reaches a maximum of 300℃, the thermoelectric cooling TEC module 2 is activated for active cooling, and both the air cooling subsystem 301 and the water cooling subsystem 302 of the composite heat dissipation system 3 are activated. S5. Spectral preprocessing: The PLC intelligent control system 4 transmits the raw spectral data to the data preprocessing module 5, which performs preprocessing on the raw spectral data, including multivariate scattering correction (MSC), standard normal variable transformation (SNV), Savitzky-Golay smoothing, derivative processing, etc., and sequentially completes data cleaning, standardization, and noise reduction to obtain high-quality spectral data. S6. Model Establishment and Calibration: Based on chemometric methods such as principal component analysis (PCA), partial least squares regression (PLSR), and support vector machine (SVM), a robust processing endpoint prediction or classification analysis model is established using standard sample spectra with known processing degrees and their corresponding reference values ​​such as the content of key components and the processing degree level. The model is then calibrated and validated using validation set data. S7. Online Prediction and Judgment: The real-time acquired spectrum processed by the data preprocessing module 5 is input into the calibrated analysis model. The model analysis module 6 calculates the predicted value or classification result of the current processing status of the material and outputs three judgment results: "insufficient processing", "moderate processing" and "over-processing". S8. Endpoint Decision and Control: Based on the output results of the model analysis module 6, such as the predicted key indicator values ​​and the category of the processing stage, the system intelligently determines whether the preset processing endpoint has been reached. If the processing is deemed insufficient, the processing continues, and the equipment maintains the current process parameters. If the processing is deemed adequate, the PLC intelligent control system 4 immediately outputs a control signal to terminate the processing and control the material to enter the next process. If the processing is deemed excessive, the PLC intelligent control system 4 immediately terminates the processing, the abnormal warning module 7 triggers the corresponding warning, and the abnormal data is recorded for subsequent traceability. S9. Data storage and monitoring: All raw data, processed data, judgment results, and equipment operation logs are stored in the data storage module 8 in real time. Managers can view the real-time status, historical data, and reports through the Web client or mobile client of the remote monitoring module 9.

Claims

1. A traditional Chinese medicine decoction piece roasting end point online intelligent near-infrared detection device and method suitable for high temperature environment, characterized in that, The detection device comprises a GSA104 high-temperature-resistant online near-infrared spectrum probe (1), a thermoelectric refrigeration TEC module (2), a composite heat dissipation system (3), a PLC intelligent control system (4), a data preprocessing module (5), a model analysis module (6), an abnormal early warning module (7), a data storage module (8) and a remote monitoring module (9); the GSA104 high-temperature-resistant online near-infrared spectrum probe (1) is integrally installed in the processing equipment and is used for collecting near-infrared diffuse reflectance spectrum data of materials in real time; the thermoelectric refrigeration TEC module (2) is closely attached to the heat source area of the main control circuit board of the GSA104 high-temperature-resistant online near-infrared spectrum probe (1); the composite heat dissipation system (3) comprises an air cooling subsystem (301) and a water cooling subsystem (302), the air cooling subsystem (301) is adapted to the heat dissipation area of the thermoelectric refrigeration TEC module (2), the spectrum sensor of the GSA104 high-temperature-resistant online near-infrared spectrum probe (1) and the main control circuit board area, and the water cooling subsystem (302) wraps the key heat-sensitive area of the GSA104 high-temperature-resistant online near-infrared spectrum probe (1); the data preprocessing module (5), the model analysis module (6) and the data storage module (8) are all installed in a server room, the data preprocessing module (5) is connected with the PLC intelligent control system (4), the model analysis module (6) is connected with the data preprocessing module (5), the abnormal early warning module (7) is partially installed on the processing equipment site and is partially associated with a remote terminal, and is connected with the model analysis module (6) at the same time; the data storage module (8) is connected with the data preprocessing module (5), the model analysis module (6) and the abnormal early warning module (7) respectively, the remote monitoring module (9) comprises a Web client and a mobile client and is connected with the PLC intelligent control system (4), the model analysis module (6) and the data storage module (8) respectively; the PLC intelligent control system (4) is bidirectionally communicated with the GSA104 high-temperature-resistant online near-infrared spectrum probe (1), the thermoelectric refrigeration TEC module (2) and the composite heat dissipation system (3) to realize data transmission and instruction issuing. 2.The online intelligent near-infrared detection device for the end point of baking of traditional Chinese medicine decoction pieces suitable for high temperature environment according to claim 1, wherein, The core components of the GSA104 high-temperature-resistant online near-infrared spectrum probe (1) are made of high-temperature-resistant materials and optical elements, the spectrum acquisition range is 780-2500nm, the acquisition frequency is 1 / 2s, the data transmission delay is less than or equal to 50ms, and non-contact nondestructive detection is supported. 3.The online intelligent near-infrared detection device for the end point of baking of traditional Chinese medicine decoction pieces suitable for high temperature environment according to claim 1, wherein, The thermoelectric refrigeration TEC module (2) adopts pulse width modulation (PWM) technology to control the cold / heat end working state of the semiconductor thermoelectric element, has the cold / heat end polarity switching function, the refrigeration temperature range is-10℃ to room temperature, the temperature control precision is ±0.1℃, and the response time is less than or equal to 20ms. 4.The online intelligent near-infrared detection device for the end point of baking of traditional Chinese medicine decoction pieces suitable for high-temperature environment according to claim 1, wherein, The air-cooled subsystem (301) of the composite heat dissipation system (3) is composed of an air compressor, a wind guide pipe and a heat dissipation fin array. The air compressor provides clean compressed air flow, which flows through the heat dissipation fins of the thermoelectric refrigeration TEC module (2), the spectral sensor of the GSA104 high-temperature-resistant online near-infrared spectral probe (1) and the main control circuit board area. The wind guide pipe is optimized to achieve forced convection heat dissipation. The surface area of the heat dissipation fins is increased by 30% compared with the traditional structure. A positive pressure difference is formed inside during operation to prevent external hot air from penetrating in. The water-cooled subsystem (302) adopts a stainless steel jacket structure, which tightly wraps the key heat-sensitive areas of the GSA104 high-temperature-resistant online near-infrared spectral probe (1) to form a physical heat barrier. It is connected with a circulating water pump, a water tank and a cooling device to form a closed cooling water circuit. The cooling water flow can be adjusted in the range of 5-15 L / min, and the water inlet temperature is ≤25℃. 5.The online intelligent near-infrared detection device for the end point of baking of traditional Chinese medicine decoction pieces suitable for high-temperature environment according to claim 1, wherein, The PLC intelligent control system (4) is built-in with a high-performance processor with an operating speed ≥1GHz. It supports multi-protocol communication and communicates with the GSA104 high-temperature-resistant online near-infrared spectral probe (1) in real time to receive raw data and output temperature control instructions to the composite heat dissipation system (3). 6.The online intelligent near-infrared detection device for the end point of baking of traditional Chinese medicine decoction pieces suitable for high-temperature environment according to claim 1, characterized in that, The data preprocessing module (5) sequentially performs data cleaning, standardization and noise reduction processing. During data cleaning, missing values, abnormal values and repeated values in the spectral data are removed. When the missing proportion is <5%, linear interpolation method is used for filling. When the missing proportion is ≥5%, the part of data is discarded and the discarding situation is recorded. Abnormal values are identified by Grubbs test method and processed by adjacent data replacement method or data smoothing method according to the deviation degree. Wavelet transform denoising method is used for data denoising. 7.The online intelligent near-infrared detection device for the end point of baking of traditional Chinese medicine decoction pieces suitable for high-temperature environment according to claim 1, wherein, The model analysis module (6) uses chemometrics algorithm to fuse BP neural network model to construct processing endpoint analysis model. The input layer has four neurons, corresponding to the spectral characteristic values, material temperature, particle size parameters and collection time stamp data processed by the data preprocessing module (5). 8.The online intelligent near-infrared detection device for the end point of baking of traditional Chinese medicine decoction pieces suitable for high-temperature environment according to claim 1, wherein, The abnormal early warning module (7) is internally integrated with a signal acquisition chip, a warning driving circuit and a wireless communication module, and carries an abnormal signal identification software. This module establishes data connection with the model analysis module (6) through the signal acquisition chip to receive processing state judgment signals and equipment running state feedback signals, connects the sound-light warning execution part through the warning driving circuit, establishes bidirectional data connection with the remote monitoring module (9) through the wireless communication module, and realizes transmission of abnormal early warning information and event record. 9.The online intelligent near-infrared detection device for the end point of baking of traditional Chinese medicine decoction pieces suitable for high-temperature environment according to claim 1, wherein, The data storage module (8) is internally configured with a storage controller, a solid state disk array and a data cache chip, and carries a special data management software and a redundancy backup program. This module establishes multi-channel data input connection with the core modules such as spectral acquisition, data preprocessing, model analysis and abnormal early warning through the storage controller, establishes one-way data output connection with the remote monitoring module (9) through the data cache chip, and realizes safe archiving and rapid retrieval of full-process data with the help of the solid state disk array and the redundancy backup program. 10.The detection method of the online intelligent near-infrared detection device for Chinese medicine decoction pieces roasted end point suitable for high temperature environment according to any one of claims 1-9, characterized in that, The method comprises the following steps: S1, device deployment: the GSA104 high-temperature-resistant online near-infrared spectrum probe (1) is integrated and installed in the internal appropriate position of the processing equipment such as a frying pan, ensuring that its optical window can effectively cover the representative material area and avoid mechanical damage; the assembly and debugging of the thermoelectric refrigeration TEC module (2), the composite heat dissipation system (3) and the GSA104 high-temperature-resistant online near-infrared spectrum probe (1) are completed, and the cables of each module are reliably connected; S2, parameter initialization: set the device operation parameters through the remote monitoring module (9), including the spectrum acquisition frequency of the GSA104 high-temperature-resistant online near-infrared spectrum probe (1), the temperature threshold of the composite heat dissipation system (3), the processing endpoint determination standard, the contact information of the early warning receiving personnel, etc.; S3, spectrum acquisition: during the processing, the GSA104 high-temperature-resistant online near-infrared spectrum probe (1) is used to collect the near-infrared diffuse reflectance spectrum of the material sample in real time, obtain the original spectrum data, record the acquisition time stamp synchronously, and transmit the original spectrum data to the PLC intelligent control system (4); S4, heat dissipation control: the PLC intelligent control system (4) monitors the internal temperature T of the processing equipment in real time, dynamically activates the corresponding heat dissipation mechanism: when T<100℃, only the water cooling subsystem (302) of the composite heat dissipation system (3) is started; when 100℃≤T<150℃, the air cooling subsystem (301) and the water cooling subsystem (302) of the composite heat dissipation system (3) are started simultaneously; when T≥150℃ up to 300℃, the thermoelectric refrigeration TEC module (2) is started for active refrigeration, and the air cooling subsystem (301) and the water cooling subsystem (302) of the composite heat dissipation system (3) are started simultaneously; S5, spectrum pretreatment: the PLC intelligent control system (4) transmits the original spectrum data to the data pretreatment module (5) for pretreatment, including multivariate scatter correction MSC, standard normal variable transformation SNV, Savitzky-Golay smoothing, derivative processing, etc., sequentially completing data cleaning, standardization, and noise reduction processing to obtain high-quality spectrum data; S6, model establishment and calibration: based on chemometrics methods such as principal component analysis PCA, partial least squares regression PLSR, support vector machine SVM, etc., using standard sample spectra of known processing degree and their corresponding reference values such as key ingredient content, processing degree grade, etc., a robust processing endpoint prediction or classification analysis model is established, and the model is calibrated and verified through validation set data; S7, online prediction and determination: the real-time acquisition spectrum processed by the data pretreatment module (5) is input into the calibrated analysis model, and the model analysis module (6) calculates the processing state prediction value or classification result of the current material, and outputs three kinds of determination results of "insufficient processing", "moderate processing" and "excessive processing". S8, endpoint decision and control: according to the model analysis module (6) output results such as predicted key indicator values, the category of the processing stage, the intelligent judgment is made whether the preset processing endpoint is reached: if it is judged that the processing is insufficient, the processing is continued, and the equipment keeps running at the current process parameters; if it is judged that the processing is moderate, the PLC intelligent control system (4) immediately outputs a control signal to terminate the processing, and controls the material to enter the next process; if it is judged that the processing is excessive, the PLC intelligent control system (4) immediately terminates the processing, and the abnormal early warning module (7) triggers the corresponding early warning, while recording the abnormal data for subsequent tracing; S9, data storage and monitoring: all the raw data, processed data, judgment results and equipment operation logs are stored in the data storage module (8) in real time, and the management personnel can check the real-time state, historical data and report through the Web client or mobile client of the remote monitoring module (9).

Images