An Infrared Method for Monitoring and Measuring Flue Gas Carbon with Self-Calibration Function

Abstract

This invention discloses an infrared method for monitoring and measuring carbon in flue gas with self-calibration function, comprising: S1: synchronously acquiring the original response spectral signal generated by the infrared gas analyzer when measuring flue gas, as well as the current operating parameters of the flue gas and the key state parameters of the infrared gas analyzer; S2: pre-training the instrument's full-condition response model, inputting the acquired operating parameters and key component state parameters into the instrument's full-condition response model to generate a standard response spectrum matching the full-condition operation, comparing the standard response spectrum with the original response spectral signal through integration, and calculating a dynamic calibration factor for the current measurement conditions; S3: using the dynamic calibration factor to correct the carbon component molar absorption coefficient and baseline parameters of the concentration inversion algorithm built into the infrared gas analyzer, and performing real-time calculation on the original response spectral signal to obtain the self-calibrated flue gas carbon component concentration.

Description

Technical Field

[0001] This invention relates to the field of next-generation information technology, and in particular to an infrared method for monitoring and measuring carbon in flue gas with self-calibration function. Background Technology

[0002] Currently, traditional methods for monitoring and measuring carbon emissions from flue gas utilize infrared gas analyzers. However, existing infrared gas analyzers generally employ a periodic manual calibration mode, requiring shutdown and the introduction of physical standard gases for calibration. This not only interrupts the monitoring process and increases maintenance costs but also suffers from calibration lag. Furthermore, the temperature, pressure, and humidity of the flue gas at the site are constantly changing, and the instrument's light source, detector, and optical lens assembly will age and degrade over time. Periodic calibration cannot compensate for the systematic errors caused by these dynamic changes in real time, leading to monitoring results gradually deviating from the true values ​​after long-term operation, failing to meet the accuracy and continuity requirements for carbon emission measurement.

[0003] Furthermore, traditional concentration inversion algorithms are based on standard operating conditions (25℃, 101.325kPa, 0% RH) and ideal instrument conditions. They only use fixed molar absorption coefficients and baseline parameters for concentration calculation, without fully considering the combined effects of flue gas condition fluctuations, such as changes in temperature, pressure, and humidity, as well as the degradation of instrument components, decrease in light source intensity, detector sensitivity drift, and reduction in mirror transmittance. This results in the inversion algorithm failing to accurately match the field measurement conditions when the actual operating conditions deviate from the ideal conditions, leading to systematic deviations in the concentration calculation results and making it difficult to reflect the true content of carbon components in the flue gas. Therefore, an infrared method for monitoring and measuring carbon in flue gas with self-calibration function is proposed to solve the above problems. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art and to achieve the above objectives, the present invention proposes the following technical solution: An infrared method for monitoring and measuring carbon in flue gas with self-calibration function includes: S1: Synchronously acquire the raw response spectrum signal generated by the infrared gas analyzer when measuring flue gas, as well as the current operating parameters of the flue gas and the key status parameters of the infrared gas analyzer; S2: Pre-trained instrument full-condition response model. The collected operating parameters and key component status parameters are input into the instrument full-condition response model to generate a standard response spectrum that matches the full-condition operation. The standard response spectrum is compared with the original response spectrum signal by integral operation to calculate the dynamic calibration factor for the current measurement conditions. The instrument's full-condition response model includes an input layer, a physical mechanism constraint layer, a data fitting correction layer, and an output layer. S3: The concentration inversion algorithm built into the infrared gas analyzer is corrected for carbon component molar absorption coefficient and baseline parameters using a dynamic calibration factor, and the original response spectral signal is calculated in real time to obtain the self-calibrated flue gas carbon component concentration. S4: Based on the self-calibrated flue gas carbon component concentration and combined with flue gas velocity data, the carbon emission measurement results are calculated.

[0005] The operating parameters G of the flue gas include the flue gas temperature. Flue gas pressure Flue gas humidity ; The key state parameter H of the infrared gas analyzer includes the infrared light source emission intensity. Detector response sensitivity optical lens group transmittance .

[0006] The process of pre-training the instrument's full-condition response model includes: Obtain the model training dataset, and under each operating condition, introduce standard gas of corresponding concentration into the infrared gas analyzer, simultaneously acquiring the measured response spectrum of the standard gas. The corresponding operating condition parameter G, key state parameter H, and standard gas concentration value are labeled. Construct a labeled dataset that corresponds one-to-one between input parameters and measured spectra; The collected labeled dataset was divided into training, validation, and test sets in a 7:2:1 ratio. The input layer received the operating condition parameters G and key state parameters H, which were then input into the physical mechanism constraint layer to obtain the theoretical response spectrum. The theoretical response spectrum The standard response spectrum is obtained by inputting the data into the data fitting correction layer. And output it through the output layer.

[0007] The implementation process of the physical mechanism constraint layer includes: Based on the Lambert-Beer law, a theoretical response spectrum relationship for a standard gas under ideal operating conditions and ideal instrument conditions is constructed. The molar absorption coefficient of the molar carbon component in the theoretical response spectrum relationship is then fitted using the least squares method. ; The theoretical response spectral relationship is extended to the entire operating condition range, and the emission intensity of the mid-infrared source in the operating condition parameter H is established. Detector response sensitivity optical lens group transmittance Molar absorption coefficient of carbon component The correction relationship is used to output the theoretical response spectrum under any operating condition. .

[0008] The implementation process of the data fitting correction layer includes: Theoretical response spectrum of the output of the computational physical mechanism constraint layer Compared with the measured response spectrum Spectral deviation value ; Radial basis function neural network As a nonlinear fitting process, the operating condition parameter G and the key state parameter H are used as inputs, and the spectral deviation value is used as the input. For output, the network is trained using the gradient descent algorithm; After training, the theoretical response spectrum of the physical mechanism constraint layer will be obtained. Spectral deviation value The standard response spectrum was obtained by fusion. .

[0009] The process of obtaining the dynamic calibration factor for the current measurement conditions includes: Determine the effective measurement wavelength range for the target carbon component. , for standard response spectrum Within the effective measurement wavelength range The total infrared light energy that the standard gas should receive is obtained by performing definite integral calculations within the range. For the original response spectral signal Within the effective measurement wavelength range Perform definite integral calculations to obtain all infrared light energy received under the current actual measurement conditions; The dynamic calibration factor is obtained by comparing the total infrared light energy that the standard gas should receive with all infrared light energy received under the current actual measurement conditions. ; in, This indicates the total infrared energy that the standard gas should receive. This represents all infrared light energy received under the current actual measurement conditions.

[0010] The process of correcting the molar absorption coefficient of carbon components includes: The original carbon component molar absorptivity of the original concentration inversion algorithm is obtained. Through dynamic calibration factor right Real-time correction is performed to obtain the corrected molar absorption coefficient of the carbon component: ; The corrected molar absorption coefficient of the carbon component Real-time writing to the concentration inversion algorithm, replacing the original... The molar absorption coefficient of the carbon component was corrected.

[0011] The process of baseline parameter correction includes: Obtain the initial baseline parameters for the original concentration inversion algorithm. ,by For the correction coefficient, the original baseline parameters are... Perform full wavelength correction for the effective measurement wavelength range. Each wavelength point within Perform correction calculations to obtain the corrected real-time baseline parameters. ; The corrected real-time baseline parameters Real-time writing to the concentration inversion algorithm replaces the original baseline parameters. This completes the dynamic correction of baseline parameters.

[0012] The process of obtaining the self-calibrated concentration of carbon components in flue gas includes: Based on the corrected real-time baseline parameters For the original response spectrum By performing baseline subtraction across the entire wavelength range, a net absorption spectrum containing only the true absorption information of the carbon component is obtained. ; Based on the Lambert-Beer law, the modified molar absorption coefficient of the carbon component is substituted... Net absorption spectrum And the optical path length L, within the effective measurement wavelength range The uncorrected original concentration is obtained through full-wavelength integral inversion. ; With dynamic calibration factor This is a global correction factor for the original concentration. Real-time correction is performed to obtain the self-calibrated flue gas carbon component concentration. , .

[0013] The present invention has the following beneficial effects: By using a pre-trained instrument full-condition response model, a virtual standard gas response spectrum that perfectly matches the current measurement conditions can be generated based on the flue gas condition parameters and instrument component status parameters collected in real time on site. This allows for in-situ online self-calibration without stopping the instrument and introducing physical standard gas.

[0014] By comparing the full wavelength integration of the standard response spectrum and the original response spectrum in real time to generate a dynamic calibration factor, it can immediately compensate for system errors caused by operating condition fluctuations and component aging, get rid of the lag and limitations of periodic calibration, ensure the continuity of the monitoring process and the real-time nature of calibration, and significantly reduce operation and maintenance costs.

[0015] The revised concentration inversion algorithm perfectly matches the actual on-site measurement conditions, eliminating systematic deviations caused by operating condition drift and component aging, and improving the long-term accuracy and stability of flue gas carbon component concentration inversion. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of an infrared method for monitoring and measuring carbon in flue gas with self-calibration functionality, as proposed in this invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1: As Figure 1 As shown, the present invention proposes an infrared method for monitoring and measuring carbon in flue gas with self-calibration function, comprising: S1: Synchronously acquire the raw response spectrum signal generated by the infrared gas analyzer when measuring flue gas, as well as the current operating parameters of the flue gas and the key status parameters of the infrared gas analyzer; The raw response spectral signal acquisition process includes: After initiating the flue gas carbon component measurement process using an infrared gas analyzer, its infrared light source emits an infrared beam within a specific wavelength range. This beam penetrates the flue gas being measured, and the transmitted light signal, absorbed by the flue gas, is received by a photodetector. This optical signal is converted into an electrical signal, which, after analog-to-digital conversion, generates the original response spectral signal, denoted as [symbol missing]. ; in, As a wavelength variable in the infrared spectrum, it directly reflects the absorption characteristics of infrared light by carbon components in the current flue gas, such as carbon monoxide and carbon dioxide. The current flue gas operating parameter acquisition process includes: The operating condition parameters G of the flue gas are synchronously collected by the operating condition sensor group of the flue gas monitoring system, including the flue gas temperature. Flue gas pressure Flue gas humidity Three types of core parameters; Furthermore, flue gas temperature The temperature data collected by the platinum resistance temperature sensor reflects the thermodynamic state of the flue gas, which directly affects the propagation speed of infrared light in the flue gas and the molar absorption coefficient of the carbon component. flue gas pressure The pressure is collected by a pressure transmitter, reflecting the pressure state of the flue gas, which affects the concentration distribution of gas molecules in the flue gas and the infrared absorption intensity. flue gas humidity The humidity is collected by a capacitive humidity sensor to reflect the water vapor content in the flue gas, thus eliminating the interference of water vapor on the characteristic absorption peaks of infrared carbon components. Specifically, this set of parameters directly affects the propagation characteristics of infrared light in flue gas and the molar absorption coefficient of carbon components. It is a key input variable for the instrument's full-condition response model and is used to match the actual flue gas environmental conditions being measured. The key status parameter acquisition process of an infrared gas analyzer includes: The infrared gas analyzer's built-in condition monitoring unit synchronously collects key condition parameters H of critical components, including the infrared light source emission intensity. Detector response sensitivity optical lens group transmittance Three types of core parameters; Furthermore, the infrared light source emits strong light. The light intensity is collected by a light intensity monitoring sensor at the output end of the light source, reflecting the emission power of the infrared light source and eliminating measurement deviations caused by light source aging and attenuation. Detector response sensitivity The photoelectric conversion efficiency of the photodetector is reflected by the calibration circuit built into the detector, thus eliminating system errors caused by detector sensitivity drift. optical lens group transmittance The light intensity is collected by the light intensity detection module of the optical lens group, reflecting the light transmission performance of the optical lens group and eliminating the light intensity attenuation caused by lens group contamination and aging. Specifically, the key state parameter H directly reflects the working state of the analyzer itself and is another core input variable of the instrument's full-condition response model in step S2, used to match the actual working state of the instrument at present. By ensuring that the acquisition timestamps of the original response spectral signal, operating condition parameter G, and key state parameter H are consistent, it is guaranteed that all data correspond to the measurement conditions at the same time, avoiding model matching inaccuracies caused by time differences, and providing time-synchronized input data for the generation of virtual standard gas response.

[0019] S2: Pre-trained instrument full-condition response model. The collected operating parameters and key component status parameters are input into the instrument full-condition response model to generate a standard response spectrum that matches the full-condition operation. The standard response spectrum is compared with the original response spectrum signal by integral operation to calculate the dynamic calibration factor for the current measurement conditions. The process of pre-training the instrument's full-condition response model to output the standard response spectrum includes: First, the model training dataset was obtained. Under each operating condition, a standard gas of the corresponding concentration was introduced into the infrared gas analyzer, and the measured response spectrum of the standard gas was collected simultaneously. The corresponding operating condition parameter G, key state parameter H, and standard gas concentration value are labeled. Construct a labeled dataset that corresponds one-to-one between input parameters and measured spectra; Then, the dataset was partitioned and augmented. The collected labeled dataset was divided into training set, validation set and test set in a ratio of 7:2:1. At the same time, the dataset size was expanded by interpolating the working parameters through data augmentation technology to improve the generalization ability of the model and avoid overfitting. The model architecture of the instrument's full-condition response model includes an input layer, a physical mechanism constraint layer, a data fitting and correction layer, and an output layer, wherein: The input layer receives six parameters across two categories, including flue gas temperature. Flue gas pressure Flue gas humidity Infrared light source emission intensity Detector response sensitivity optical lens group transmittance ; The physical mechanism constraint layer is based on the Lambert-Beer law to construct the basic mechanism of infrared absorption of standard gases, describing the standard response spectrum of standard gases under ideal working conditions and ideal instrument conditions. As the basic constraint of the model, it ensures that the model output conforms to the physical laws of infrared absorption and avoids physical distortion of the model driven by pure data. The data fitting correction layer is based on nonlinear fitting trained on measured data. It is used to correct the deviation of the physical mechanism constraint layer under actual working conditions and actual instrument conditions, and to compensate for the influence of non-ideal factors such as working condition interference and component aging, so as to achieve accurate fitting of the actual response spectrum. The output dimension of the output layer is the standard response spectrum of the standard gas under the corresponding input parameter conditions, denoted as... ,in, The wavelength variable in the infrared spectrum is compared with the original response spectrum acquired in step S1. The wavelength ranges are consistent; Furthermore, the implementation process of the physical mechanism constraint layer includes: Based on the Lambert-Beer law, the theoretical response spectrum relationship of a standard gas under ideal operating conditions and ideal instrument conditions is constructed, in the form of:

[0020] in, For ideal infrared light source emission intensity, For ideal optical lens groups, the transmittance is... The molar absorption coefficient of the carbon component of the standard gas at wavelength λ is given. Where L is the standard gas concentration and L is the infrared path length. For the ideal detector response sensitivity, This is the theoretical response spectrum; Perform mechanism parameter calibration under ideal operating conditions, such as , , Standard gases of different concentrations were introduced, and measured spectral data were collected. The molar absorption coefficient of the carbon component in the theoretical response spectral relationship formula was determined by fitting the data using the least squares method. This ensures that the output of the physical mechanism constraint layer under ideal conditions is highly consistent with the measured spectrum, thus completing the initialization of the mechanism constraint layer. The physical mechanism constraint layer is extended to cover the entire operating condition range, extending the theoretical response spectrum relationship under ideal conditions to the whole operating condition range, and establishing the infrared light source emission intensity in the operating condition parameter H. Detector response sensitivity optical lens group transmittance Molar absorption coefficient of carbon component The correction relationship allows the physical mechanism constraint layer to output the theoretical response spectrum under any operating condition. As the baseline input for the data fitting layer; The implementation process of the data fitting correction layer includes: For each set of data in the pre-defined training set, calculate the theoretical response spectrum of the physical mechanism constraint layer output. Compared with the measured response spectrum Spectral deviation value The deviation value includes the combined effects of non-ideal factors such as operating condition interference and component aging, and serves as the training target for the fitting layer. Then, the fitting model is trained using a radial basis function neural network. As a nonlinear fitting process, the operating condition parameter G and the key state parameter H are used as inputs, and the spectral deviation value is used as the input. To output, execute the training process: The center and width of the radial basis function are initialized, and the initial values ​​are determined based on the distribution of the input parameters in the training set; Optimize using gradient descent algorithm The weights of the network minimize the mean square error between the fitted output bias and the measured bias, thereby achieving accurate learning of the bias characteristics. The training process is monitored in real time using a validation set, and an early stopping mechanism is used to avoid overfitting and ensure the network's generalization ability. After training, the theoretical response spectrum of the physical mechanism constraint layer will be finally obtained. Spectral deviation values ​​of the data fitting correction layer The standard response spectrum of the final full-condition response is obtained by fusion. Output through the output layer; Then, the standard response spectrum is integrated and compared with the original response spectrum signal: Retrieve the built-in hardware configuration parameters of the infrared gas analyzer to determine the effective measurement wavelength range for the target carbon component. ; Specifically, the effective measurement wavelength range It covers the characteristic absorption peaks of carbon component molecules and is the core band for effective concentration inversion and calibration, while excluding irrelevant background noise bands; Then, integration calculations are performed for comparison to determine the upper and lower limits of the integration calculation as the effective measurement wavelength range. Within this range, the wavelength variation of the infrared spectrum is related to the original response spectrum acquired in step S1. The wavelength ranges are consistent; Then obtain the standard response spectrum. Total energy, relative to the standard response spectrum Within the effective measurement wavelength range Perform definite integral operations to obtain the total infrared light energy that the standard gas should receive, i.e., the sum of the standard response spectrum. ; Next, the total spectral energy is determined for the original response spectral signal. In the same effective interval The definite integral operation is performed within the field, and its physical meaning is the total spectral energy actually measured on site, which represents all infrared light energy received under the current actual measurement conditions; The dynamic calibration factor under the current measurement conditions is quantified by the ratio of the sum of the standard response spectral energies to the total energy of the original response spectral energies. The system deviation coefficient reflects the proportional relationship of the amplitudes and the comprehensive deviation between the operating conditions and the state. The formula is expressed as: ; Furthermore, the input to the instrument's full-condition response model comes entirely from the data acquired in step S1. The generation of the virtual standard gas response spectrum depends on the operating conditions and state parameters of step S1. The calculation of the dynamic calibration factor depends on the original response spectrum of step S1 and the virtual standard gas response spectrum generated in this step. The final dynamic calibration factor is obtained from this process. This is the core input for step S3.

[0021] S3: The concentration inversion algorithm built into the infrared gas analyzer is corrected for carbon component molar absorption coefficient and baseline parameters using a dynamic calibration factor, and the original response spectral signal is calculated in real time to obtain the self-calibrated flue gas carbon component concentration. The core of the raw concentration inversion algorithm built into the infrared gas analyzer is based on the Lambert-Beer law, and its expression calibrated under standard operating conditions and standard instrument conditions is as follows:

[0022] in, The molar absorption coefficient of the original carbon component calibrated under standard operating conditions and standard instrument conditions. For the ideal detector response sensitivity, For ideal optical lens groups, the transmittance is... Where L is the standard gas concentration and L is the infrared path length. This represents the intensity of the transmitted light received by the detector. The ideal infrared light source emits light intensity; The process of correcting the molar absorption coefficient of carbon components includes: Molar absorption coefficient of the original carbon component It was calibrated under standard operating conditions (25℃, 101.325kPa, 0%RH) and standard instrument conditions. When measured in the field: flue gas temperature Flue gas pressure Flue gas humidity Deviations from the standard value will cause collisional broadening of gas molecules and shift in the position of absorption peaks, resulting in a difference between the actual molar absorptivity of the carbon component and the standard value. Deviation occurs; Instrument light source attenuation, detector sensitivity drift, and mirror contamination can cause an overall shift in the spectral response amplitude, indirectly amplifying the calculation deviation of the molar absorption coefficient of the carbon component; Therefore, through dynamic calibration factors right Real-time correction is performed to obtain the corrected molar absorption coefficient of the carbon component: ; Furthermore, if This indicates that the actual response amplitude of the instrument is lower than the standard value (e.g., due to light source attenuation or lens contamination). After correction... To compensate for the concentration underestimation caused by the decay of the compensation response amplitude; like This indicates that the actual response amplitude of the instrument is higher than the standard value (e.g., detector gain drift), and should be corrected. The overestimation of concentration is caused by an artificially high compensation response amplitude; like This indicates that the current measurement conditions are consistent with the standard conditions. No corrections are needed; The corrected molar absorption coefficient of the carbon component Real-time writing to the concentration inversion algorithm, replacing the original... This enables dynamic correction of the molar absorption coefficient of the carbon component; Specifically, the corrected molar absorption coefficient of the carbon component It perfectly matches the actual working conditions and instrument status at the current site. After substituting Lambert-Beer Law, it can accurately restore the true absorption characteristics of carbon components in flue gas, eliminate the deviation of carbon component molar absorption coefficient caused by working condition drift and component aging, and provide accurate core parameters for subsequent concentration inversion. The process of baseline parameter correction includes: The original baseline parameters for the original concentration inversion algorithm are the zero-gas response spectra under standard operating conditions. This refers to the response spectrum measured by the instrument when zero gas containing no carbon components is introduced. This spectrum is used to subtract baseline interference such as instrument background noise and stray light from the optical system. Its ideal form is:

[0023] in, For ideal infrared light source emission intensity, For the ideal detector response sensitivity, For the ideal optical lens group, the transmittance is ; The baseline parameter is essentially the zero-air response amplitude of the instrument, and its amplitude variation is related to... There is a linear correspondence: The ratio of the actual zero-air response amplitude to the standard zero-air response amplitude is equal to The reciprocal of, therefore, with For the correction coefficient, the original baseline parameters are... Perform full wavelength correction, that is, correct the effective measurement wavelength range. Each wavelength point within Perform the above correction calculation to obtain the corrected real-time baseline parameters. , ; The corrected real-time baseline parameters Real-time writing to the concentration inversion algorithm replaces the original baseline parameters. This completes the dynamic correction of baseline parameters; Specifically, the corrected real-time baseline parameters To fully match the actual operating status of the current instrument and the field conditions, the original response spectrum was used before concentration inversion. Subtract the corrected baseline It can accurately remove interference such as instrument background noise, optical stray light, and baseline drift, and retain only the true absorption signal of flue gas carbon components, providing pure input data for subsequent concentration inversion; The original response spectrum Input the corrected concentration inversion algorithm and retrieve the dynamic calibration factor. and through Corrected molar absorption coefficient of carbon component With real-time baseline parameters Synchronously retrieve the fixed parameter optical path length L of the instrument to ensure that all parameters have consistent wavelength ranges and timestamps; Based on the corrected real-time baseline parameters For the original response spectrum By performing baseline subtraction across the entire wavelength range, a net absorption spectrum containing only the true absorption information of the carbon component is obtained. ; Based on the Lambert-Beer law, the modified molar absorption coefficient of the carbon component is substituted... Net absorption spectrum And the optical path length L, within the effective measurement wavelength range The uncorrected original concentration is obtained through full-wavelength integral inversion. ; With dynamic calibration factor This is a global correction factor for the original concentration. Real-time correction is performed to obtain the self-calibrated flue gas carbon component concentration. The calculation formula is: ; Specifically, the final concentration of carbon components in the flue gas Used for subsequent carbon emission measurement, it proves that the infrared gas analyzer can still output the true flue gas concentration that meets national metrological standards after experiencing changes in operating conditions and component aging, thus solving the concentration drift problem caused by long-term operation of traditional infrared monitoring.

[0024] S4: Based on the self-calibrated flue gas carbon component concentration and combined with flue gas velocity data, the carbon emission measurement results are calculated. The real-time flow velocity data v of the flue gas is collected by the flow velocity sensor of the flue gas monitoring system. The real-time flue gas flow rate is obtained based on the flue gas velocity v and the cross-sectional area A of the flue gas duct. , ; Specifically, this flow data reflects the volume of flue gas passing through the monitoring section per unit time, serving as the flow basis for carbon emission measurement. The self-calibrated concentration of carbon components in the flue gas With the calculated real-time flue gas flow rate Coupled calculations are performed, and the molar mass of the carbon components and the carbon oxidation rate are fixed parameters to calculate the carbon emission measurement result per unit time, denoted as . ; Where k is the unit conversion factor, which includes the molar mass of carbon components, carbon oxidation rate, and standard condition conversion factor. It is used to convert volume concentration and volume flow rate into mass emission, which meets the national standard requirements for carbon emission measurement. The calculated carbon emission measurement results It performs real-time storage and outputs data in a format that meets environmental monitoring requirements, completing the entire infrared flue gas carbon monitoring and measurement process with self-calibration function.

[0025] In the application, several formulas are calculated by removing dimensions and taking their numerical values. The formulas are established by collecting a large amount of data and simulating the most recent real situation. Some coefficients or weights in the formulas are set by those skilled in the art according to the actual situation, so they will not be elaborated here.

[0026] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0027] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A self-calibrating infrared method for monitoring and measuring carbon in flue gas, characterized in that, include: S1: Synchronously acquire the raw response spectrum signal generated by the infrared gas analyzer when measuring flue gas, as well as the current operating parameters of the flue gas and the key status parameters of the infrared gas analyzer; S2: Pre-trained instrument full-condition response model. The collected operating parameters and key component status parameters are input into the instrument full-condition response model to generate a standard response spectrum that matches the full-condition operation. The standard response spectrum is compared with the original response spectrum signal by integral operation to calculate the dynamic calibration factor for the current measurement conditions. The instrument's full-condition response model includes an input layer, a physical mechanism constraint layer, a data fitting correction layer, and an output layer. S3: The concentration inversion algorithm built into the infrared gas analyzer is corrected for carbon component molar absorption coefficient and baseline parameters using a dynamic calibration factor, and the original response spectral signal is calculated in real time to obtain the self-calibrated flue gas carbon component concentration. S4: Based on the self-calibrated flue gas carbon component concentration and combined with flue gas velocity data, the carbon emission measurement results are calculated.

2. The infrared flue gas carbon monitoring and measurement method with self-calibration function according to claim 1, characterized in that, The operating parameters G of the flue gas include the flue gas temperature. Flue gas pressure Flue gas humidity ; The key state parameter H of the infrared gas analyzer includes the infrared light source emission intensity. Detector response sensitivity optical lens group transmittance .

3. The infrared flue gas carbon monitoring and measurement method with self-calibration function according to claim 1, characterized in that, The process of pre-training the instrument's full-condition response model includes: Obtain the model training dataset, and under each operating condition, introduce standard gas of corresponding concentration into the infrared gas analyzer, simultaneously acquiring the measured response spectrum of the standard gas. The corresponding operating condition parameter G, key state parameter H, and standard gas concentration value are labeled. Construct a labeled dataset that corresponds one-to-one between input parameters and measured spectra; The collected labeled dataset was divided into training, validation, and test sets in a 7:2:1 ratio. The input layer received the operating condition parameters G and key state parameters H, which were then input into the physical mechanism constraint layer to obtain the theoretical response spectrum. The theoretical response spectrum The standard response spectrum is obtained by inputting the data into the data fitting correction layer. And output it through the output layer.

4. The infrared flue gas carbon monitoring and measurement method with self-calibration function according to claim 3, characterized in that, The implementation process of the physical mechanism constraint layer includes: Based on the Lambert-Beer law, a theoretical response spectrum relationship for a standard gas under ideal operating conditions and ideal instrument conditions is constructed. The molar absorption coefficient of the molar carbon component in the theoretical response spectrum relationship is then fitted using the least squares method. ; The theoretical response spectral relationship is extended to the entire operating condition range, and the emission intensity of the mid-infrared source in the operating condition parameter H is established. Detector response sensitivity optical lens group transmittance Molar absorption coefficient of carbon component The correction relationship is used to output the theoretical response spectrum under any operating condition. .

5. The infrared flue gas carbon monitoring and measurement method with self-calibration function according to claim 4, characterized in that, The implementation process of the data fitting correction layer includes: Theoretical response spectrum of the output of the computational physical mechanism constraint layer Compared with the measured response spectrum Spectral deviation value ; Radial basis function neural network As a nonlinear fitting process, the operating condition parameter G and the key state parameter H are used as inputs, and the spectral deviation value is used as the input. For output, the network is trained using the gradient descent algorithm; After training, the theoretical response spectrum of the physical mechanism constraint layer will be obtained. Spectral deviation value The standard response spectrum was obtained by fusion. .

6. The infrared method for monitoring and measuring carbon in flue gas with self-calibration function according to claim 5, characterized in that, The process of obtaining the dynamic calibration factor for the current measurement conditions includes: Determine the effective measurement wavelength range for the target carbon component. , for standard response spectrum Within the effective measurement wavelength range The total infrared light energy that the standard gas should receive is obtained by performing definite integral calculations within the range. For the original response spectral signal Within the effective measurement wavelength range Perform definite integral calculations to obtain all infrared light energy received under the current actual measurement conditions; The dynamic calibration factor is obtained by comparing the total infrared light energy that the standard gas should receive with all infrared light energy received under the current actual measurement conditions. ； in, This indicates the total infrared energy that the standard gas should receive. This represents all infrared light energy received under the current actual measurement conditions.

7. The infrared flue gas carbon monitoring and measurement method with self-calibration function according to claim 6, characterized in that, The process of correcting the molar absorption coefficient of carbon components includes: The original carbon component molar absorptivity of the original concentration inversion algorithm is obtained. Through dynamic calibration factor right Real-time correction is performed to obtain the corrected molar absorption coefficient of the carbon component: ; The corrected molar absorption coefficient of the carbon component Real-time writing to the concentration inversion algorithm, replacing the original... The molar absorption coefficient of the carbon component was corrected.

8. The infrared method for monitoring and measuring carbon in flue gas with self-calibration function according to claim 7, characterized in that, The process of baseline parameter correction includes: Obtain the initial baseline parameters for the original concentration inversion algorithm. ,by For the correction coefficient, the original baseline parameters are... Perform full wavelength correction for the effective measurement wavelength range. Each wavelength point within Perform correction calculations to obtain the corrected real-time baseline parameters. ; The corrected real-time baseline parameters Real-time writing to the concentration inversion algorithm replaces the original baseline parameters. This completes the dynamic correction of baseline parameters.

9. The infrared method for monitoring and measuring carbon in flue gas with self-calibration function according to claim 8, characterized in that, The process of obtaining the self-calibrated concentration of carbon components in flue gas includes: Based on the corrected real-time baseline parameters For the original response spectrum By performing baseline subtraction across the entire wavelength range, a net absorption spectrum containing only the true absorption information of the carbon component is obtained. ; Based on the Lambert-Beer law, the modified molar absorption coefficient of the carbon component is substituted... Net absorption spectrum And the optical path length L, within the effective measurement wavelength range The uncorrected original concentration is obtained through full-wavelength integral inversion. ; With dynamic calibration factor This is a global correction factor for the original concentration. Real-time correction is performed to obtain the self-calibrated flue gas carbon component concentration. , .

Images