A performance monitoring method for mid-infrared sensors

Abstract

The present invention belongs to the field of sensor monitoring technology and discloses a performance monitoring method for a mid-infrared sensor. The method includes the following steps: S100 integrated calibration: integrating FBG and auxiliary sensors into a mid-infrared system, completing connection and calibration; S200 acquisition and analysis: acquiring data, analyzing FBG wavelength changes to infer temperature / deformation, and determining status; S300 fusion evaluation: fusing multi-source data, algorithmically processing, and calculating performance indicators; S400 threshold grading: setting dynamic thresholds and dividing performance levels; S500 response adjustment: executing emergency response, adjusting strategies after failures, and restarting monitoring. By integrating multiple sensors and combining data fusion with intelligent analysis, the present invention achieves comprehensive and accurate monitoring of mid-infrared sensor performance, improving monitoring accuracy and real-time performance, effectively eliminating interference from environmental factors, and more accurately reflecting the actual operating status of the mid-infrared sensor.

Description

Technical Field

[0001] The present invention relates to the technical field of sensor monitoring, and in particular to a performance monitoring method for a mid-infrared sensor. Background Art

[0002] Mid-infrared sensors are widely used in many fields due to their unique performance. Their operating stability and accuracy are crucial to ensuring proper system operation. However, mid-infrared sensors are affected by a variety of environmental factors during operation, such as temperature, humidity, air pressure, electromagnetic interference, and chemical corrosion. These factors can cause performance degradation or even failure. Therefore, monitoring mid-infrared sensor performance and promptly detecting and addressing abnormalities are crucial to ensuring system stability and safety.

[0003] Existing methods for monitoring mid-infrared sensor performance typically rely primarily on single sensor data, such as temperature changes or deformation alone to determine the sensor's operating status. This approach has significant limitations, as a single data point often fails to fully reflect the sensor's actual performance. Furthermore, existing methods fail to adequately consider and compensate for environmental factors, severely impacting the accuracy and reliability of mid-infrared sensor performance monitoring under complex operating conditions. Furthermore, existing methods often employ static threshold settings, unable to dynamically adjust to actual application scenarios and changes in mid-infrared sensor performance, thus limiting their adaptability and flexibility.

[0004] Therefore, it is necessary to develop a new performance monitoring method for mid-infrared sensors to overcome the shortcomings of existing technologies. Summary of the Invention

[0005] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a performance monitoring method for mid-infrared sensors. This method realizes multi-parameter data acquisition and fusion analysis by integrating fiber Bragg gratings (FBGs), humidity and pressure sensors, and can more comprehensively and accurately evaluate sensor performance, and perform dynamic early warning and adjustment.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] A method for monitoring the performance of a mid-infrared sensor, the method comprising the following steps:

[0008] S100, Sensor Integration and Calibration: Investigate the working principle, usage scenarios, and past fault feedback of the mid-infrared sensor, select a fiber Bragg grating (FBG) suitable for the mid-infrared sensor's operating environment, choose a packaging material based on the mid-infrared sensor's structure and operating characteristics, attach the FBG to the mid-infrared sensor, and simultaneously install a humidity sensor and air pressure sensor near the mid-infrared sensor. Connect all sensors to the central control system via shielded cables, complete line connections according to standard communication protocols, and calibrate the zero point and range of each sensor unit.

[0009] S200, basic data acquisition and analysis: The spectrum analyzer is set to acquire the reflection spectrum of the FBG at a preset frequency. The humidity and air pressure monitoring units respectively acquire environmental data at their respective preset frequencies. The central control system synchronously activates all sensors. After verification, the acquired data is transmitted to the central control system. The data processing software built into the central control system extracts the central wavelength change of the FBG and analyzes the current temperature change and deformation of the mid-infrared sensor to determine whether the mid-infrared sensor is in a normal physical state. If it exceeds the normal range, an early warning is triggered.

[0010] S300, Data Fusion and Performance Evaluation: The central control system stores the real-time data collected by the humidity and air pressure monitoring units, and the temperature and strain data converted by the FBG in a dedicated database; collects historical data, and uses the big data platform and machine learning algorithm to train the basic correlation function. and , combined with the determined cross-coupling coefficient , substitute into the preset formula to perform cross compensation of each parameter, the formula is: ,in, Represents the final quantitative value of the mid-infrared sensor performance index, It represents the preliminary performance correlation value, and the calculation formula is: , represents the temperature change of the mid-infrared sensor, represents the deformation amount of the mid-infrared sensor, Indicates humidity, Indicates air pressure, Represents each parameter variable, referring to temperature ,strain ,humidity , air pressure One of Represents The corresponding basic correlation function finally outputs the quantitative value of the mid-infrared sensor performance index ;

[0011] S400, Dynamic Threshold Setting and Level Classification: Collect data on application scenarios and failure cases of the mid-infrared sensor, calculate historical quantified values ​​of mid-infrared sensor performance indicators, determine boundary values ​​for four performance levels: safe, mild abnormality, moderate risk, and severe failure based on the historical quantified values ​​of mid-infrared sensor performance indicators, calculate the risk level based on the real-time collected data, and output the risk level;

[0012] S500, emergency response and strategy adjustment: The central control system receives early warning reports in real time and adopts different early warning processing for different early warning reports: in case of mild abnormality, the background records details and pushes low-priority prompts and troubleshooting guides; in case of moderate risk, the collection frequency is increased and a pop-up window alert is displayed on the operation and maintenance interface; in case of serious fault, the preset decibel sound and light alarm is immediately triggered, an emergency text message is sent, and the surrounding equipment is linked to emergency shutdown; after the fault is repaired, the threshold is dynamically adjusted according to the situation and the monitoring process is restarted.

[0013] Preferably, in the step S200, the central wavelength variation is extracted from the FBG reflection spectrum by the following parameter conversion formula in basic data collection and analysis, and analyzed in combination with the physical parameters of the mid-infrared sensor: Assume that the central wavelength variation of the fiber Bragg grating FBG reflected light is , the temperature change of the mid-infrared sensor is , the shape variable is , introducing the temperature sensitivity coefficient and strain sensitivity coefficient , the formula is: , wherein the temperature sensitivity coefficient The value range is 0.01-0.02nm / ℃, and the strain sensitivity coefficient The value range is 0.78-1.2pm / με, It is the comprehensive environmental interference, which includes factors that are difficult to measure directly, such as electromagnetic interference and weak chemical corrosion. is the environmental interference correction factor, and its value range is 0.01-0.1.

[0014] Preferably, in the step S200, the method for determining the normal range threshold of the physical state of the mid-infrared sensor in the basic data collection and analysis is: assuming that the operating temperature range of the mid-infrared sensor is arrive , the deformation tolerance range is arrive , the working condition correction coefficient is and , its value range is 0.1-0.5, and the calculation formula of the normal range threshold is: , , , ,in, 、 It is the amount of additional adjustment required for the lower and upper temperature limits due to operating conditions, calculated based on historical fault data. is the amount by which the deformation needs to be adjusted.

[0015] Preferably, in the step S300, the cross-coupling coefficient in the data fusion and performance evaluation The determination method is: using an iterative optimization algorithm based on gradient descent, combined with historical monitoring data sets , , Represents the temperature change of the mid-infrared sensor in the mth monitoring , Representative shape variables , Represents humidity H, represents the air pressure P, For the corresponding mid-infrared sensor performance index, a loss function is constructed, and the formula of the loss function is: ,in Based on the current The performance index of the mid-infrared sensor predicted by the value is calculated by the following formula: , during the iteration process, update according to the following formula : ; η is the learning rate, which ranges from 0.001 to 0.1, is the loss function for The partial derivative of Represents the basic correlation function; after each iterative calculation, the performance of the model is verified until the loss function value converges to the preset minimum value, and the optimal cross-coupling coefficient is determined .

[0016] Preferably, in the step S300, the basic correlation function in the data fusion and performance evaluation and The calculation method is: use the polynomial fitting method to determine 、 The example formula is: , ,in represents the temperature change of the mid-infrared sensor, represents the deformation amount of the mid-infrared sensor, Indicates humidity, represents the air pressure, the correlation function 、 The specific coefficients of are determined by fitting the experimental data using the least squares method.

[0017] Preferably, in the step S300, the method for training the correlation function by introducing a combination of high-order polynomial fitting and Gaussian process regression in data fusion and performance evaluation is as follows: collecting experimental data, covering relevant parameters of temperature, strain, humidity and air pressure; cleaning the data and applying minimum-maximum normalization processing; then performing high-order polynomial fitting, for the basic correlation function and , the formula of the high-order polynomial is: , ,in represents the temperature change of the mid-infrared sensor, represents the deformation of the mid-infrared sensor, H represents humidity, and P represents air pressure; the coefficient is solved using the least squares method and ; Perform Gaussian process regression and select radial basis function RBF kernel as kernel function. The formula of the kernel function is: ; Estimate hyperparameters by maximizing the marginal likelihood function and , the formula of the marginal likelihood function is: ; Use the trained model to predict the new data point, and the predicted mean is: , the variance is: ; The results of high-order polynomial fitting and Gaussian process regression are weightedly combined to obtain the final basic correlation function.

[0018] Preferably, in the step S400, the warning priority corresponding to the quantitative value of the mid-infrared sensor performance index is calculated by the following warning priority determination formula in the dynamic setting of the threshold and the level division: Let the mid-infrared sensor performance index S be relative to the standard value in the safe state. The degree of deviation is , the warning priority is , introducing the risk perception index , the calculation formula is: , .

[0019] Preferably, in the step S400, the risk perception index r is calculated according to the following risk perception index calculation formula in the dynamic setting of the threshold and the level division: in is the weight coefficient, ranging from 0.1 to 1.0, and satisfies , Indicates temperature deviation, Indicates humidity deviation, represents the pressure deviation, ρΔL represents the strain change rate, which is the strain of the mid-infrared sensor Relative to normal strain range The degree of change is calculated as follows: , is the fault frequency weight, which is calculated as: ,in is the preset maximum fault frequency, is the scenario risk level.

[0020] Preferably, in the step S400, the threshold value is dynamically set and the level division is determined according to the value of the warning priority W to determine the boundary values ​​of the four performance levels of safety, mild abnormality, moderate risk, and severe failure:

[0021] When 0≤ When ≤24, it means the mid-infrared sensor is in a safe state and is set to a safety level;

[0022] When 25≤ When ≤49, it means that there are some slight abnormalities in the mid-infrared sensor and it is set to the mild abnormality level;

[0023] When 50≤ When ≤74, it means that the mid-infrared sensor has a moderate risk and is set to the moderate risk level;

[0024] When 75≤ When ≤100, it indicates that a serious fault has occurred in the mid-infrared sensor and is set to the serious fault level.

[0025] Compared with the prior art, the present invention has the following beneficial effects:

[0026] 1. This invention integrates fiber Bragg gratings (FBGs), humidity sensors, and air pressure sensors, combined with the data processing capabilities of a central control system, to achieve comprehensive monitoring of mid-infrared sensor performance. This not only improves the accuracy and real-time nature of monitoring, but also effectively eliminates interference from environmental factors such as electromagnetic interference and weak chemical corrosion through the fusion and cross-compensation of multi-parameter data, thereby more accurately reflecting the actual operating status of the mid-infrared sensor. This comprehensive monitoring approach provides powerful data support for fault warning, emergency response, and strategy adjustment for mid-infrared sensors, helping to improve their reliability and service life.

[0027] 2. By collecting application scenarios and failure case data of mid-infrared sensors, the present invention can dynamically adjust thresholds based on actual conditions. It also divides the operating status of mid-infrared sensors into four levels: safe, mildly abnormal, moderately risky, and severely faulty, based on the quantitative values ​​of mid-infrared sensor performance indicators and risk assessment results. This dynamic and flexible threshold setting method makes early warning, emergency response, and strategy adjustment more accurate and efficient, enabling appropriate measures to be taken according to different situations, avoiding losses caused by false alarms or missed alarms. At the same time, this performance level classification also provides clearer guidance for the maintenance and management of mid-infrared sensors, helping to improve the maintenance efficiency and management level of mid-infrared sensors. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort.

[0029] Figure 1 The present invention is a flow chart of a method for monitoring the performance of a mid-infrared sensor. DETAILED DESCRIPTION

[0030] The following will provide a clear and complete description of the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0031] Reference Figure 1 , the present invention provides a performance monitoring method for a mid-infrared sensor.

[0032] Example 1: Industrial furnace temperature monitoring scenario

[0033] In industrial furnace temperature monitoring scenarios, mid-infrared sensors are used to monitor temperature changes in the furnace in real time. The method of the present invention is applied to this scenario, and the specific steps are as follows:

[0034] S100, sensor integration and calibration: After investigation, it was determined that the mid-infrared sensor operates in this high-temperature, complex electromagnetic environment. Fiber Bragg gratings (FBGs) that are resistant to high temperatures and electromagnetic interference were selected, and ceramics were used as the packaging material to bond the FBGs to the sensor. A humidity sensor and an air pressure sensor were installed near the mid-infrared sensor. Each sensor was connected to the central control system via a shielded cable according to the standard communication protocol, and the zero point and range were calibrated.

[0035] S200, basic data collection and analysis: According to the characteristics of the sensor and environment, the spectrum analyzer is set to collect FBG reflection spectrum 10 times per second, and the humidity and air pressure monitoring units are set to collect data at a frequency of 3 minutes and 5 minutes respectively. The central control system starts the work of each sensor, and after the collected data is transmitted, the central wavelength change is extracted from it. , according to the parameter conversion formula to analyze the temperature change of the mid-infrared sensor , shape variable , the formula is: , is the temperature sensitivity coefficient, is the strain sensitivity coefficient, is the comprehensive environmental interference, is the environmental interference correction factor; and based on the working condition correction factor 、 and operating temperature range arrive , deformation tolerance range arrive The normal range threshold is determined and judged, and the threshold formula is:

[0036] , , , ,in, is the temperature adjustment amount, The adjustment amount is the deformation variable; and the physical state of the sensor is judged, and an early warning is triggered if it exceeds the range.

[0037] S300, Data Fusion and Performance Evaluation: The central control system has a dedicated database that integrates humidity and pressure data with temperature and strain data converted by FBG; an iterative optimization algorithm based on gradient descent is used to determine the cross-coupling coefficient , construct the loss function , is the actual performance indicator, To predict the performance index, the coefficients are iteratively updated according to the formula: , is the learning rate, until the loss function converges to the optimal ; Use polynomial fitting method to determine the correlation function: , ; Substitute into the compensation formula to calculate the quantitative value of the performance index , the formula is: in: .

[0038] S400, dynamic threshold setting and classification: collects industrial furnace temperature monitoring scenario failure case data, calculates historical S values, and calculates the degree of deviation according to the formula ,in is the safety standard value; combined with the risk perception index , the formula is: , is the weight coefficient, is the humidity deviation, is the air pressure deviation, is the strain change rate, is the fault frequency weight, The risk level of the scenario is determined by the warning priority formula to determine the performance boundary value: , classification level: 0≤ ≤24 is safe, 25≤ ≤49 is mildly abnormal, 50≤ ≤74 is moderate risk, 75≤ ≤100 indicates a serious fault.

[0039] S500, emergency response and strategy adjustment: The central control system receives warnings and handles them according to the level: minor anomalies are recorded in the background and prompts are pushed for investigation; moderate risks increase the collection frequency and pop-up warnings; serious faults issue a 120-decibel sound and light alarm, SMS notification, and linked shutdown of surrounding equipment; after repairs, monitoring is restarted according to the mood threshold.

[0040] Through the above steps, this embodiment constructs a mid-infrared sensor performance monitoring system for industrial furnace temperature monitoring. In the sensor integration stage, FBG and packaging materials are selected to adapt to high-temperature and strong interference conditions, and multiple sensors are integrated into the central control calibration. In the data acquisition and analysis stage, the frequency is set according to the characteristics, and the temperature and deformation are analyzed by formula to determine anomalies. In the fusion compensation stage, the coefficients are determined by the algorithm and the fitting function is used to obtain the quantitative value. In the dynamic threshold setting stage, four levels of boundaries are set based on case data to evaluate performance. In the emergency response stage, the warning level is progressively processed, and the threshold is repaired and restarted. This ensures that furnace temperature monitoring is accurate and reliable, production is safe, and system stability and operation and maintenance efficiency are improved, consolidating the foundation of industrial temperature control.

[0041] Example 2: Mid-infrared sensor performance monitoring method in industrial furnace monitoring scenario

[0042] In industrial kiln temperature and environment monitoring scenarios, mid-infrared sensors are used to accurately measure the temperature of key parts inside the kiln, which is crucial for quality control of the production process. The method of the present invention is applied to this scenario. The specific steps are similar to those in Example 1, but adaptive adjustments are made to the characteristics of the kiln environment:

[0043] S100 Sensor Integration and Calibration: Considering the kiln's high temperature, complex chemical atmosphere, and strong electromagnetic interference operating environment, in-depth research into the operating principles of mid-infrared sensors, root causes of past failures, and feedback from various operating conditions led to the selection of fiber Bragg gratings (FBGs) with high-temperature resistance, chemical corrosion resistance, and high electromagnetic interference shielding effectiveness. Based on the kiln's structural characteristics, thermal stress distribution, and vibration conditions, appropriate ceramic packaging materials were selected. The FBGs were securely bonded to the critical sensing area of ​​the mid-infrared sensor, ensuring close contact and efficient heat and strain transfer. High-precision humidity and pressure sensors were precisely installed near the mid-infrared sensors within the kiln. Each sensor was connected to the central control system via high-temperature, interference-resistant shielded cables using standard communication protocols. Using professional calibration equipment and standard temperature, humidity, and pressure sources, each sensor's zero point and span were meticulously calibrated according to established procedures to ensure an accurate measurement benchmark, laying a solid foundation for subsequent precise monitoring.

[0044] S200, basic data collection and analysis: In view of the complex and changeable working conditions of the kiln, the spectrum analyzer is set to collect the FBG reflection spectrum at an appropriate frequency. At the same time, the humidity and air pressure monitoring units determine their respective collection frequencies according to the fluctuation characteristics of the kiln environmental parameters. Start the central control system, drive the various sensors to work together, and verify the integrity and accuracy of data transmission in real time. After verification, transmit it to the central control system. With built-in advanced data processing software, the central control system extracts the center wavelength change through a strict parameter conversion formula based on the intrinsic relationship between the FBG reflection spectrum and the physical parameters of the mid-infrared sensor. Assume that the center wavelength change of the FBG reflected light is Δλ, the temperature change of the mid-infrared sensor is ΔT, and the deformation is ΔL. The precise temperature sensitivity coefficient is introduced. and strain sensitivity coefficient , calculated by the formula: , α is the environmental interference correction factor, and ε is the combined interference of electromagnetic and chemical corrosion. The operating temperature range of the mid-infrared sensor is determined by combining the kiln design specifications and process requirements. to , deformation tolerance range to , and the operating condition correction coefficient obtained based on in-depth analysis of historical fault big data 、 , the normal range threshold is accurately calculated by the formula: 、 、 in: is the temperature adjustment amount, The adjustment value for the deformation variable is analyzed in real time. The current physical state of the mid-infrared sensor is immediately triggered when the threshold is exceeded, ensuring safe and stable operation of the kiln.

[0045] S300, Data Fusion and Performance Evaluation: The central control system has developed a large-capacity dedicated database, accurately integrating real-time humidity and pressure data with temperature and strain data converted from FBG, and strictly aligning and matching them according to timestamps. It also collects extensive long-term historical data on kiln operation, covering the changes in various parameters under multiple working conditions. It uses the powerful computing power of the big data platform and advanced machine learning algorithms to train basic correlation functions based on massive data. and For example, after multiple iterations of optimization and verification, the form is determined (a combination of high-order polynomial fitting and Gaussian process regression can be used). The example formula is: 、 The loss function is constructed by combining the historical data set with the gradient descent iterative optimization algorithm: ,in is the actual performance indicator, As a prediction indicator, the cross-coupling coefficient is iteratively updated according to the formula , η is the learning rate, and the optimal coefficient is obtained when the loss function converges to the minimum value. Substitute it into the formula to calculate and output the quantitative value of the reliable sensor performance index : in: .

[0046] S400, dynamic threshold setting and classification: Comprehensively collect kiln production scene information and rich fault case data, and deeply explore the change pattern of sensor performance index quantitative value. Based on the sensor performance index S and the safety status standard value Degree of deviation (The formula is: ) and the risk perception index r to comprehensively determine the warning priority W (the calculation formula is: ), where r is calculated according to the formula: , is the weight coefficient, is the temperature deviation, is the humidity deviation, is the air pressure deviation, is the strain change rate, is the fault frequency weight, is the risk level of the scene). Based on this, safety is accurately determined (0≤ ≤24), mild abnormality (25≤ ≤49), moderate risk (50≤ ≤74), serious failure (75≤ ≤100) four-level performance boundary value, realize accurate graded assessment of kiln sensor status, and efficiently guide operation and maintenance decision-making.

[0047] S500, emergency response and strategy adjustment: The central control system receives early warning reports in real time and implements intelligent policies based on the risk level. In the event of a minor anomaly, the background records the details in detail, accurately pushes low-priority prompts and targeted troubleshooting guides, and assists operation and maintenance personnel in accurately locating potential hidden dangers. In the event of a moderate risk, the frequency of data collection is adaptively increased, and a prominent pop-up window alert appears on the operation and maintenance interface to help personnel closely monitor key parameters. In the event of a serious fault, a powerful 120-decibel sound and light alarm is immediately triggered to deter the entire site, and an emergency text message is delivered to the person in charge in seconds. At the same time, the surrounding supporting equipment of the kiln is shut down urgently to curb the spread of the fault and ensure the safety of production facilities and personnel. After the fault is repaired, the fault characteristics and overall working conditions are deeply reviewed, the threshold is finely adjusted according to the situation, the monitoring process is restarted, and the monitoring efficiency is continuously optimized to ensure the efficient and stable operation of the kiln throughout the process.

[0048] This embodiment has achieved significant success in industrial furnace monitoring scenarios. During the sensor integration and calibration phase, FBGs and packaging materials are selected based on the furnace's harsh operating conditions. Each sensor is properly positioned and precisely calibrated, laying the foundation for accurate monitoring. During basic data collection and analysis, the acquisition frequency is set based on the furnace's characteristics. The central control system verifies and processes the data, using a specific formula to calculate the center wavelength change and threshold. If the threshold is exceeded, an alert is issued to ensure furnace safety. During data fusion and evaluation, the central control system integrates multiple data sets, uses algorithms to train correlation functions and determine coupling coefficients, and then substitutes these into the formula to obtain performance indicators, providing critical support for furnace evaluation. During the dynamic threshold setting and classification phase, indicator deviations and risk indices are calculated based on a large amount of data to determine boundary values, resulting in a four-level, precise assessment. During emergency response strategy adjustment, intelligent responses are implemented based on early warnings, and fault repairs dynamically optimize the threshold restart process. This ensures the stable and efficient operation of the furnace in all aspects, significantly improving production quality and safety while reducing operational and maintenance risks and costs.

[0049] The foregoing description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Those skilled in the art will readily appreciate that various modifications and variations of the present invention are possible. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention. Any reference numerals in the claims should not be construed as limiting the claims to which they relate.

Claims

1. A performance monitoring method for a mid-infrared sensor, characterized in that: The specific steps of this monitoring method are: S100, Sensor Integration and Calibration: Research the operating principles, usage scenarios, and past fault feedback of mid-infrared sensors, select fiber Bragg gratings (FBGs) suitable for the mid-infrared sensor's operating environment, choose packaging materials based on the mid-infrared sensor's structure and operating characteristics, laminate the FBGs to the mid-infrared sensors, and simultaneously install humidity and pressure sensors near the mid-infrared sensors. All sensors are connected to the central control system via shielded cables, connected according to standard communication protocols, and the zero point and range of each unit are calibrated. S200, Basic Data Collection and Analysis: The spectrum analyzer is set to collect FBG reflectance spectra 10 times per second, and the humidity and air pressure monitoring units are set to collect data every 3 minutes and every 5 minutes, respectively. The central control system starts all sensors simultaneously and transmits the collected data to the central control system after verification. The central control system's built-in data processing software extracts the central wavelength change and analyzes the current temperature change and deformation of the mid-infrared sensor to determine whether the mid-infrared sensor is in a normal physical state. If it exceeds the normal range, an alarm will be triggered. In the above S200, the central wavelength variation is extracted from the FBG reflection spectrum and the physical parameters of the mid-infrared sensor through the parameter conversion formula in the basic data collection and analysis. Assume that the central wavelength variation of the fiber Bragg grating FBG reflected light is , the temperature change of the mid-infrared sensor is , the shape variable is , introducing the temperature sensitivity coefficient and strain sensitivity coefficient , the formula is: , where the temperature sensitivity coefficient The value range is 0.01-0.02nm / ℃, and the strain sensitivity coefficient The value range is 0.78-1.2pm / με, It is the comprehensive environmental interference, including electromagnetic interference and weak chemical corrosion, which are difficult to measure directly. is the environmental interference correction factor, with a value range of 0.01-0.1; In the above S200, the normal range threshold of the infrared sensor physical state is determined in the basic data collection and analysis, and the operating temperature range of the infrared sensor is set to arrive , deformation tolerance range arrive , the working condition correction coefficient is and , its value range is 0.1-0.5, and the calculation formula is: , , , ,in, 、 It is the amount of additional adjustment required for the lower and upper temperature limits due to operating conditions, calculated based on historical fault data. is the amount by which the deformation variable needs to be adjusted; S300, Data Fusion and Performance Evaluation: The central control system develops a dedicated database to store the real-time data collected by the humidity and air pressure monitoring units, as well as the temperature and strain data converted by FBG. It also collects historical data and uses the big data platform and machine learning algorithms to train basic correlation functions. and , combined with the determined cross-coupling coefficient , substitute into the formula to perform cross compensation of each parameter, the formula is: ,in, It represents the preliminary performance correlation value, and the calculation formula is: , Indicates the temperature change of the mid-infrared sensor, represents the deformation of the mid-infrared sensor, Indicates humidity, Indicates air pressure, Represents each parameter variable, referring to temperature ,strain ,humidity , air pressure The physical quantities involved in the calculation, Represents the basic correlation function, and finally outputs the quantitative value of the mid-infrared sensor performance index ; S300, cross-coupling coefficient in data fusion and performance evaluation To determine the , , Represents the temperature change of the mid-infrared sensor at a certain time , and so on, To correspond to the performance indicators of the mid-infrared sensor, a loss function is constructed, and the formula is: ,in Based on the current The performance index of the mid-infrared sensor predicted by the value is calculated by the formula: , during the iteration process, update according to the following formula , the formula is: , is the learning rate, ranging from 0.001 to 0.1, is the loss function for The partial derivative of Represents the basic correlation function. After each iterative calculation, the performance of the model is verified until the loss function value converges to the minimum value and the optimal cross-coupling coefficient is determined. ; In the S300, data fusion and performance evaluation, the correlation function is trained by introducing a combination of high-order polynomial fitting and Gaussian process regression, collecting experimental data covering relevant parameters such as temperature, strain, humidity and air pressure, cleaning the data and applying minimum-maximum normalization, and then performing high-order polynomial fitting. For the basic correlation function and , the formula is: , ,in Indicates the temperature change of the mid-infrared sensor, represents the deformation of the mid-infrared sensor, Indicates humidity, Represents air pressure, and the coefficients are solved using the least squares method and , perform Gaussian process regression, select radial basis function RBF kernel as kernel function, the formula is: , the hyperparameters are estimated by maximizing the marginal likelihood function and , the formula is: , use the trained model to predict the new data point, and the predicted mean is: , the variance is: , the results of high-order polynomial fitting and Gaussian process regression are weightedly combined to obtain the final basic correlation function; S400, Dynamic Threshold Setting and Level Classification: This process collects data on mid-infrared sensor application scenarios and failure cases, calculates the quantitative values ​​of mid-infrared sensor performance indicators, and determines four performance boundary values ​​(safe, mild abnormality, moderate risk, and severe failure) based on the quantitative values ​​of mid-infrared sensor performance indicators. The system then calculates and outputs the risk level based on the real-time collected data. S500, emergency response and strategy adjustment: The central control system receives early warning reports in real time and takes different early warning measures for different reports: in case of mild anomalies, the background records details and pushes low-priority prompts and troubleshooting guides; in case of moderate risks, the collection frequency is increased and a pop-up window alert is displayed on the operation and maintenance interface; in case of serious faults, a 120-decibel sound and light alarm is immediately triggered, an emergency text message is sent, and the surrounding equipment is linked to emergency shutdown; after the fault is repaired, the threshold is dynamically adjusted according to the situation and the monitoring process is restarted.

2. The performance monitoring method of a mid-infrared sensor according to claim 1, characterized in that: S300, data fusion and performance evaluation based on correlation function and The calculation of 、 The form is: , ,in Indicates the temperature change of the mid-infrared sensor, represents the deformation of the mid-infrared sensor, Indicates humidity, represents air pressure, and the correlation function 、 The specific range of is determined by fitting the experimental data using the least squares method.

3. The performance monitoring method of a mid-infrared sensor according to claim 1, characterized in that: In the S400, the threshold value is dynamically set and the level is divided, and the quantitative value of the mid-infrared sensor performance index is calculated by the warning priority determination formula. Relative to the standard value in safe state The degree of deviation is , the warning priority is , introducing the risk perception index , the calculation formula is: , .

4. The performance monitoring method of a mid-infrared sensor according to claim 3, characterized in that: In step S400, the threshold is dynamically set and the level is divided into levels. The risk perception index is calculated according to the risk perception index calculation formula. , the formula is: in is the weight coefficient, ranging from 0.1 to 1.0, and satisfies , Indicates temperature deviation, Indicates humidity deviation, Indicates the air pressure deviation, Indicates the strain change rate, which is the strain of the mid-infrared sensor Relative to normal strain range The degree of change is calculated as follows: , is the fault frequency weight, which is calculated as: ,in is the preset maximum fault frequency, is the scenario risk level.

5. The performance monitoring method of a mid-infrared sensor according to claim 1, characterized in that: In step S400, the dynamic setting of thresholds and the classification of levels are performed to determine four performance boundary values ​​of safety, mild abnormality, moderate risk, and severe failure based on the quantitative values ​​of the mid-infrared sensor performance indicators: when When , it means the mid-infrared sensor is in a safe state and is set to a safe level; when When , it means that there are some slight abnormalities in the mid-infrared sensor and it is set to the slight abnormality level; when When , it means that the mid-infrared sensor has a moderate risk and is set to the medium risk level; when , it indicates that a serious fault has occurred in the mid-infrared sensor and is set to a serious fault level.

Images