Adaptive parameter correction method for laser gas measurement error compensation
By using an adaptive parameter correction method, spectral and environmental data are collected, a multi-source error model is constructed, and the error compensation factor is updated in real time. This solves the error accumulation problem in laser gas measurement under dynamic conditions, and achieves high-precision and stable gas concentration measurement, which is suitable for industrial and environmental monitoring.
Patent Information
- Application Number
- CN202611579733.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-13
- Publication Date
- 2026-11-10
AI Technical Summary
Existing laser gas measurement technology cannot adapt to dynamic operating conditions, resulting in the accumulation of measurement errors, low compensation accuracy, poor versatility, inability to iterate in real time, and inability to meet the requirements for high-precision and real-time gas detection.
By using an adaptive parameter correction method, spectral absorption signals and environmental data are collected to construct a time-series correlated dataset. A multi-source error model is built based on the Lambert-Beer law. The sensitivity coefficient is solved using multivariate continuously differentiable functions and partial derivatives. The total error offset is calculated through definite integrals, and the error compensation factor is updated in real time to achieve adaptive adjustment of the compensation parameters.
It achieves high-precision gas concentration measurement, improves measurement stability and accuracy, adapts to complex working conditions, reduces equipment maintenance costs, and is suitable for industrial and environmental monitoring.
Smart Images

Figure CN123066010A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to gas measurement, and more specifically to an error compensation method for laser gas measurement with adaptive parameter correction. Background Technology
[0002] Currently, laser gas measurement technology mainly relies on the principle of spectral absorption to invert gas concentration and is widely used in scenarios such as industrial flue gas monitoring, ambient atmospheric detection, and chemical gas leak monitoring. However, in actual measurement processes, fluctuations in ambient temperature and air pressure, laser output wavelength shifts, light intensity attenuation, and interference from multiple sources such as spectral signal noise and baseline drift can cause traditional fixed-parameter laser gas measurement error compensation methods to be unable to adapt to dynamic operating conditions. This results in shortcomings such as low compensation accuracy, poor versatility, and the inability to iterate static parameters in real time.
[0003] Existing compensation techniques mostly employ preset fixed parameters and single error model fitting, which cannot adaptively adjust compensation parameters according to real-time operating conditions. Especially in complex variable operating scenarios, measurement errors accumulate significantly, making it difficult to meet the requirements of high-precision, real-time gas detection. Therefore, this invention proposes an adaptive parameter correction method for laser gas measurement error compensation. Through multi-dimensional error modeling and iterative dynamic parameters using calculus, it achieves accurate compensation of measurement errors and possesses outstanding practical value. Summary of the Invention
[0004] (a) Technical problems to be solved
[0005] In view of the above-mentioned shortcomings of the existing technology, the present invention provides an adaptive parameter correction laser gas measurement error compensation method, which can effectively overcome the defects of the existing technology, such as inability to adapt to dynamic working condition changes, low compensation accuracy, poor versatility, and inability to iterate static parameters in real time.
[0006] (II) Technical Solution
[0007] To achieve the above objectives, the present invention provides the following technical solution:
[0008] An adaptive parameter-corrected laser gas measurement error compensation method includes the following steps:
[0009] S1. For the gas to be measured, collect the original spectral absorption signal, laser incident light intensity and laser transmitted light intensity, simultaneously collect environmental data of the measurement area, and construct a time-series correlation dataset.
[0010] S2. Construct a standard gas concentration measurement model based on Lambert-Beer's law, and combine it with environmental data to construct a multi-source error model;
[0011] S3. Establish a multivariate continuously differentiable function for temperature, air pressure and light intensity attenuation coefficients, solve for the sensitivity coefficients of each variable by first-order partial derivatives, and calculate the total error offset by definite integral accumulation.
[0012] S4. The error compensation factor is updated in real time based on the total error offset, so as to realize the adaptive adjustment of the compensation parameters according to the working conditions and equipment status.
[0013] S5. Based on the error compensation factor, the measured gas concentration is compensated to eliminate multi-source coupling error and output high-precision gas concentration.
[0014] Preferably, in S1, for the gas to be measured, the original spectral absorption signal, laser incident light intensity, and laser transmitted light intensity are collected, environmental data of the measurement area are collected simultaneously, and a time-series correlation dataset is constructed, including:
[0015] S11. Adjust the working wavelength of the laser according to the characteristic absorption spectrum of the gas to be measured. After the laser penetrates the measurement area, collect the original spectral absorption signal, the laser incident light intensity and the laser transmitted light intensity, and simultaneously collect environmental data including temperature, air pressure and gas flow rate of the measurement area.
[0016] S12. Preprocess the original spectral absorption signal by using a moving average filtering algorithm to remove high-frequency random noise and retain the effective spectral absorption signal.
[0017] S13. Bind the effective spectral absorption signal and environmental data according to the timestamp to construct a time-series associated dataset {I0(t),I(t),T(t),P(t),V(t)};
[0018] Where I0(t) is the incident laser intensity at time t, I(t) is the transmitted laser intensity at time t, T(t) is the temperature at time t, P(t) is the gas pressure at time t, and V(t) is the gas velocity at time t.
[0019] Preferably, in S2, a standard gas concentration measurement model is constructed based on Lambert-Beer's law, and a multi-source error model is constructed by combining environmental data, including:
[0020] S21. Constructing an error-free standard gas concentration measurement model based on Lambert-Beer's law:
[0021] ;
[0022] Where C0(t) is the standard gas concentration at time t, L is the laser path length, and α is the characteristic absorption coefficient of the gas to be measured.
[0023] S22. Combining environmental data including temperature, air pressure, and gas flow rate in the measurement area, we decompose three core errors: temperature disturbance error, air pressure disturbance error, and equipment light intensity drift and signal distortion error, and construct a multi-source error model:
[0024] ;
[0025] Among them, C real (t) represents the measured gas concentration at time t, and ΔC T (t) represents the temperature disturbance error at time t, ΔC P (t) represents the pressure disturbance error at time t, ΔC S (t) represents the device light intensity drift and signal distortion error at time t.
[0026] Preferably, in S3, a multivariate continuously differentiable function is established for the attenuation coefficients of temperature, air pressure, and light intensity. The sensitivity coefficients of each variable are solved by first-order partial derivatives, and the total error offset is calculated by definite integral accumulation, including:
[0027] S31. Establish a multivariate continuously differentiable function for temperature, air pressure, and light intensity attenuation coefficient, and perform total differential of the multivariate function;
[0028] S32. Using the standard working conditions of the laboratory as a benchmark, the partial derivatives of the real-time collected environmental data are solved to obtain the temperature sensitivity coefficient, air pressure sensitivity coefficient, and equipment sensitivity coefficient at each sampling time.
[0029] S33. Since environmental data is a continuous time-series change, the instantaneous micro-change error needs to be accumulated through time-domain integration to calculate the total error offset within the measurement period.
[0030] Preferably, in S31, a multivariate continuously differentiable function is established for temperature, air pressure, and light intensity attenuation coefficients, and the total differential of the multivariate function is performed, including:
[0031] Establish a multivariate continuously differentiable function C(t) = f(T(t), P(t), k(t)) with respect to temperature, air pressure, and light intensity attenuation coefficients, and perform total differential of the multivariate function:
[0032] ;
[0033] Where k(t) is the light intensity attenuation coefficient at time t, which is calculated from the incident laser intensity I0(t) and the transmitted laser intensity I(t) at time t. The temperature sensitivity coefficient at time t is... Let be the pressure sensitivity coefficient at time t. Let be the device sensitivity coefficient at time t, and d be the differential operator.
[0034] Preferably, in step S32, based on standard laboratory conditions, partial derivatives are calculated on the real-time acquired environmental data to obtain the temperature sensitivity coefficient, air pressure sensitivity coefficient, and equipment sensitivity coefficient at each sampling time, including:
[0035] The temperature sensitivity coefficient, air pressure sensitivity coefficient, and equipment sensitivity coefficient are calculated using the following formulas:
[0036] ;
[0037] ;
[0038] ;
[0039] Among them, S T (t) is the temperature sensitivity coefficient at time t, S P (t) is the pressure sensitivity coefficient at time t, S k (t) represents the device sensitivity coefficient at time t.
[0040] Preferably, in S33, since the environmental data is a continuously changing quantity over time, the instantaneous micro-variable error needs to be accumulated through time-domain integration to calculate the total error offset within the measurement period, including:
[0041] The total error offset within the measurement period is calculated by performing time-domain integration and accumulation using the following formula:
[0042] ;
[0043] Among them, △C total The total error offset within the measurement period, t0, t n These are the start and end times of a single measurement cycle, t0 ≤ t ≤ t n .
[0044] Preferably, in S4, the error compensation factor is updated in real time based on the total error offset to achieve adaptive adjustment of the compensation parameters according to the operating conditions and equipment status, including:
[0045] S41. Update the error compensation factor in real time based on the total error offset within the measurement period:
[0046] ;
[0047] Where η(t) is the error compensation factor at time t;
[0048] S42. Introduce an iterative convergence threshold ε and iteratively optimize the error compensation factor:
[0049] When the absolute value of the difference between the error compensation factor of the current measurement cycle and the error compensation factor of the previous measurement cycle is greater than the iteration convergence threshold ε, the original spectral absorption signal, laser incident light intensity and laser transmitted light intensity are continuously collected, environmental data of the measurement area are collected simultaneously, and a time-series correlation dataset is constructed. S2, S3 and S41 are repeated to update the measured gas concentration and error compensation factor.
[0050] When the absolute value of the difference between the error compensation factor of the current measurement cycle and the error compensation factor of the previous measurement cycle is not greater than the iteration convergence threshold ε, the error compensation factor is determined to have converged, and the error compensation factor of the current measurement cycle is taken as the current optimal error compensation factor η. * This enables the compensation parameters to be adaptively adjusted according to operating conditions and equipment status.
[0051] The iterative convergence threshold ε is preset to 0.001 and can be adjusted according to the measurement accuracy requirements.
[0052] Preferably, in step S5, the measured gas concentration is compensated based on an error compensation factor to eliminate multi-source coupling errors and output a high-precision gas concentration, including:
[0053] S51. Based on the current optimal error compensation factor, compensate for the measured gas concentration and output a high-precision gas concentration:
[0054] ;
[0055] Among them, C final (t) represents the high-precision gas concentration at time t;
[0056] S52. Outlier removal is performed on high-precision gas concentration data. For instantaneous abnormal correction data caused by sudden changes in operating conditions, the neighborhood mean verification method is used to screen valid measurement values, and finally, stable and accurate gas concentration detection results are output.
[0057] Preferably, after compensating for the measured gas concentration based on the error compensation factor in S5 to eliminate multi-source coupling errors and output a high-precision gas concentration, the process includes:
[0058] S61. Store the environmental data, sensitivity coefficient, current optimal error compensation factor, and gas concentration error value before and after compensation obtained in each measurement to construct an adaptive compensation parameter database.
[0059] S62. In subsequent measurement processes, the current optimal error compensation factor under similar working conditions in the adaptive compensation parameter database is first matched as the initial iteration parameter to improve the convergence speed of the compensation parameter. At the same time, the adaptive compensation parameter database is updated to achieve self-learning optimization.
[0060] (III) Beneficial Effects
[0061] Compared with the prior art, the adaptive parameter correction laser gas measurement error compensation method provided by the present invention has the following beneficial effects:
[0062] 1) Breaking through the technical limitations of traditional fixed parameter compensation, through innovative modeling using total differential, partial derivative, and definite integral, the dynamic decomposition and time-domain accumulation of multi-source error weights are realized, accurately quantifying the total measurement error under dynamic working conditions. The error solution accuracy is far higher than that of traditional fixed parameter compensation methods.
[0063] 2) An adaptive compensation parameter iterative correction mechanism is adopted, which can update the compensation parameters in real time according to the working conditions and equipment status, solving the problems of lag and insufficient accuracy of error compensation in variable working conditions, and greatly improving the stability and accuracy of laser gas measurement.
[0064] 3) Through self-learning optimization of the adaptive compensation parameter database, it can quickly adapt to multiple working conditions without the need for manual calibration of compensation parameters, reducing equipment operation and maintenance costs. It can be widely adapted to complex scenarios such as industry and environmental monitoring, and has strong engineering practicality. Attached Figure Description
[0065] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.
[0066] Figure 1 This is a schematic diagram of the process of the present invention;
[0067] Figure 2 This is a detailed flowchart illustrating the calculation of the total error offset in this invention. Detailed Implementation
[0068] To make the objectives, technical solutions, and advantages of the embodiments 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. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.
[0069] The following describes the specific process of the adaptive parameter correction laser gas measurement error compensation method provided by this invention, using a concrete example (e.g.) Figure 1 (as shown) and technical effects.
[0070] S1. For the gas to be measured, acquire the original spectral absorption signal, laser incident light intensity, and laser transmitted light intensity, simultaneously acquire environmental data of the measurement area, and construct a time-series correlated dataset, such as... Figure 2 As shown, it includes:
[0071] S11. Adjust the working wavelength of the laser according to the characteristic absorption spectrum of the gas to be measured. After the laser penetrates the measurement area, collect the original spectral absorption signal, the laser incident light intensity and the laser transmitted light intensity, and simultaneously collect environmental data including temperature, air pressure and gas flow rate of the measurement area.
[0072] S12. Preprocess the original spectral absorption signal by using a moving average filtering algorithm to remove high-frequency random noise and retain the effective spectral absorption signal.
[0073] S13. Bind the effective spectral absorption signal and environmental data according to the timestamp to construct a time-series associated dataset {I0(t),I(t),T(t),P(t),V(t)};
[0074] Where I0(t) is the incident laser intensity at time t, I(t) is the transmitted laser intensity at time t, T(t) is the temperature at time t, P(t) is the gas pressure at time t, and V(t) is the gas velocity at time t.
[0075] S2. Construct a standard gas concentration measurement model based on Lambert-Beer's law, and combine it with environmental data to construct a multi-source error model, such as... Figure 2 As shown, it includes:
[0076] S21. Constructing an error-free standard gas concentration measurement model based on Lambert-Beer's law:
[0077] ;
[0078] Where C0(t) is the standard gas concentration at time t, L is the laser path length, and α is the characteristic absorption coefficient of the gas to be measured.
[0079] S22. Combining environmental data including temperature, air pressure, and gas flow rate in the measurement area, we decompose three core errors: temperature disturbance error, air pressure disturbance error, and equipment light intensity drift and signal distortion error, and construct a multi-source error model:
[0080] ;
[0081] Among them, C real (t) represents the measured gas concentration at time t, and ΔC T (t) represents the temperature disturbance error at time t, ΔC P (t) represents the pressure disturbance error at time t, ΔC S(t) represents the device light intensity drift and signal distortion error at time t.
[0082] S3. Establish a multivariate continuously differentiable function for the attenuation coefficients of temperature, air pressure, and light intensity. Solve for the sensitivity coefficients of each variable using first-order partial derivatives, and calculate the total error offset by accumulating definite integrals. Figure 2 As shown, it includes:
[0083] S31. Establish a multivariate continuously differentiable function for the attenuation coefficients of temperature, air pressure, and light intensity, and perform total differential of the multivariate function, specifically including:
[0084] Establish a multivariate continuously differentiable function C(t) = f(T(t), P(t), k(t)) with respect to temperature, air pressure, and light intensity attenuation coefficients, and perform total differential of the multivariate function:
[0085] ;
[0086] Where k(t) is the light intensity attenuation coefficient at time t, which is calculated from the incident laser intensity I0(t) and the transmitted laser intensity I(t) at time t. The temperature sensitivity coefficient at time t is... Let be the pressure sensitivity coefficient at time t. Let be the device sensitivity coefficient at time t, and d be the differential operator.
[0087] S32. Using standard laboratory conditions (temperature 25℃, air pressure 101.325kPa, light intensity attenuation coefficient 1) as a benchmark, the partial derivatives of the real-time acquired environmental data are calculated to obtain the temperature sensitivity coefficient, air pressure sensitivity coefficient, and equipment sensitivity coefficient at each sampling time. Specifically, this includes:
[0088] The temperature sensitivity coefficient, air pressure sensitivity coefficient, and equipment sensitivity coefficient are calculated using the following formulas:
[0089] ;
[0090] ;
[0091] ;
[0092] Among them, S T (t) is the temperature sensitivity coefficient at time t, S P (t) is the pressure sensitivity coefficient at time t, S k (t) represents the device sensitivity coefficient at time t.
[0093] S33. Since environmental data are continuously changing over time, instantaneous micro-variable errors need to be accumulated through time-domain integration to calculate the total error offset within the measurement period, specifically including:
[0094] The total error offset within the measurement period is calculated by performing time-domain integration and accumulation using the following formula:
[0095] ;
[0096] Among them, △C total The total error offset within the measurement period, t0, t n These are the start and end times of a single measurement cycle, t0 ≤ t ≤ t n .
[0097] The above technical solution breaks through the limitations of traditional fixed parameter compensation. Through innovative modeling using total differentials, partial derivatives, and definite integrals, it achieves dynamic decomposition and time-domain accumulation of multi-source error weights, accurately quantifies the total measurement error under dynamic working conditions, and the error solution accuracy is far higher than that of traditional fixed parameter compensation methods.
[0098] S4. Based on the total error offset, the error compensation factor is updated in real time to achieve adaptive adjustment of the compensation parameters according to the operating conditions and equipment status, including:
[0099] S41. Update the error compensation factor in real time based on the total error offset within the measurement period:
[0100] ;
[0101] Where η(t) is the error compensation factor at time t;
[0102] S42. Introduce an iterative convergence threshold ε and iteratively optimize the error compensation factor:
[0103] When the absolute value of the difference between the error compensation factor of the current measurement cycle and the error compensation factor of the previous measurement cycle is greater than the iteration convergence threshold ε, the original spectral absorption signal, laser incident light intensity and laser transmitted light intensity are continuously collected, environmental data of the measurement area are collected simultaneously, and a time-series correlation dataset is constructed. S2, S3 and S41 are repeated to update the measured gas concentration and error compensation factor.
[0104] When the absolute value of the difference between the error compensation factor of the current measurement cycle and the error compensation factor of the previous measurement cycle is not greater than the iteration convergence threshold ε, the error compensation factor is determined to have converged, and the error compensation factor of the current measurement cycle is taken as the current optimal error compensation factor η. * This enables the compensation parameters to be adaptively adjusted according to operating conditions and equipment status.
[0105] The iterative convergence threshold ε is preset to 0.001 and can be adjusted according to the measurement accuracy requirements.
[0106] The above technical solution adopts an adaptive compensation parameter iterative correction mechanism, in which the compensation parameters can be updated in real time according to the working conditions and equipment status. This solves the problems of lag in error compensation and insufficient accuracy in variable working conditions, and greatly improves the stability and accuracy of laser gas measurement.
[0107] S5. Compensate for the measured gas concentration based on the error compensation factor to eliminate multi-source coupling errors and output high-precision gas concentration, including:
[0108] S51. Based on the current optimal error compensation factor, compensate for the measured gas concentration and output a high-precision gas concentration:
[0109] ;
[0110] Among them, C final (t) represents the high-precision gas concentration at time t;
[0111] S52. Outlier removal is performed on high-precision gas concentration data. For instantaneous abnormal correction data caused by sudden changes in operating conditions, the neighborhood mean verification method is used to screen valid measurement values, and finally, stable and accurate gas concentration detection results are output.
[0112] In the technical solution of this application, after compensating for the measured gas concentration based on the error compensation factor in S5, eliminating multi-source coupling errors, and outputting a high-precision gas concentration, it includes:
[0113] S61. Store the environmental data, sensitivity coefficient, current optimal error compensation factor, and gas concentration error value before and after compensation obtained in each measurement to construct an adaptive compensation parameter database.
[0114] S62. In subsequent measurement processes, the current optimal error compensation factor under similar working conditions in the adaptive compensation parameter database is first matched as the initial iteration parameter to improve the convergence speed of the compensation parameter. At the same time, the adaptive compensation parameter database is updated to achieve self-learning optimization.
[0115] The above technical solution achieves rapid adaptation to multiple operating scenarios through self-learning optimization of the adaptive compensation parameter database. It eliminates the need for manual calibration of compensation parameters, reduces equipment operation and maintenance costs, and can be widely adapted to complex scenarios such as industry and environmental monitoring, making it highly practical for engineering applications.
[0116] To better illustrate the technical solution of this application, a detailed example is provided below.
[0117] This embodiment applies to an online laser monitoring scenario for ammonia slip concentration in an SCR denitrification system of a coal-fired power plant, representing a mainstream high-precision laser gas measurement condition in industry. The gas to be measured in this scenario is ammonia (NH3), and the conventional ammonia slip control index is 0~3ppm, which falls under the category of high-precision detection of trace gases.
[0118] The on-site operating conditions present typical dynamic disturbances: the temperature fluctuation range in the flue is 200℃~380℃, the air pressure fluctuates slightly negatively by ±500Pa, and the flue gas velocity changes in real time. At the same time, long-term high-temperature operating conditions will cause the TDLAS laser light intensity to decrease and the spectral baseline to drift. The maximum measurement error of the traditional fixed parameter compensation method can reach ±0.35ppm, which cannot meet the requirements for precise control of ultra-low emissions. The adaptive parameter correction and multi-source error dynamic compensation technical solution of this invention is very suitable.
[0119] Step S1: Collect spectral signals and environmental data to construct a time-series correlation dataset.
[0120] S11 operating condition adaptation data acquisition: Lock onto the 1.5μm characteristic absorption spectrum of ammonia, adjust the laser operating wavelength to the matching band, and allow the laser to penetrate the measurement area of the power plant's SCR denitrification flue; continuously acquire the original spectral absorption signal, laser incident light intensity I0(t), and laser transmitted light intensity I(t) in real time, and simultaneously acquire the real-time temperature T(t), gas pressure P(t), and flue gas velocity V(t) inside the flue.
[0121] S12 signal preprocessing: A moving average filtering algorithm with a window length of 5 is used to filter out high-frequency random noise caused by flue dust and electromagnetic interference, remove spectral glitch signals, and retain complete and effective ammonia spectral absorption characteristic signals.
[0122] S13 Dataset Construction: The filtered effective spectral signal and the synchronization environment parameters are bound one by one according to the millisecond-level timestamp to construct the time-series association dataset of this embodiment: {I0(t),I(t),T(t),P(t),V(t)}. The dataset completely covers all the original measurement information and operating condition information within a single measurement cycle, providing accurate data support for subsequent error modeling.
[0123] Step S2: Constructing a standard concentration measurement model and a multi-source error model
[0124] S21 Standard Concentration Modeling: Based on Lambert-Beer's Law, an error-free ideal operating condition ammonia standard concentration measurement model is established.
[0125] ;
[0126] Where: L = 1.2m (fixed laser path length), α is the characteristic absorption coefficient of ammonia, and C0(t) is the theoretical standard ammonia concentration at time t;
[0127] S22 Multi-source Error Coupling Modeling: Combining the on-site characteristics of power plant flue temperature and pressure changes, and equipment light intensity drift, three types of core errors are decomposed to construct a multi-source error model adapted to this scenario:
[0128] ;
[0129] In the formula: C real (t) represents the original measured ammonia concentration; ΔC T (t) represents the temperature disturbance error caused by the high temperature fluctuation in the flue; △C P (t) represents the pressure disturbance error caused by the slight negative pressure fluctuation in the flue; △C S (t) represents the light intensity drift and spectral distortion errors caused by the laser's long-term high-temperature operation, completely covering all major sources of measurement error in this scenario.
[0130] Step S3: Solving for dynamic error based on partial derivatives and definite integrals
[0131] S31 Multivariate Differentiable Function and Total Differential Construction: The ammonia concentration measurement is defined as a multivariate continuously differentiable function of temperature, pressure, and light intensity attenuation coefficient, adapted to the continuous dynamic operation of the flue: C(t) = f(T(t), P(t), k(t)). Total differential operations are then performed on the multivariate function.
[0132] ;
[0133] The total differential formula can accurately characterize the independent influence of small changes in a single operating condition parameter on the measurement concentration error, and achieve fine decomposition of the error;
[0134] S32 Dynamic Sensitivity Coefficient Calculation: Based on laboratory standard operating conditions, the three types of error sensitivity coefficients are calculated in real time for each sampling moment using the first-order partial derivative numerical solution method.
[0135] ;
[0136] ;
[0137] ;
[0138] In this embodiment, the temperature sensitivity coefficient S under high-temperature conditions T (t) has the largest value and is the dominant error factor in this measurement. It is highly consistent with the actual working conditions of power plant flue gas measurement and solves the defect that traditional fixed coefficients cannot distinguish between primary and secondary errors.
[0139] S33 definite integral cumulative calculation of total error offset: Since flue temperature, air pressure, and light intensity are all continuous time-series changes, instantaneous micro-errors cannot characterize the overall measurement cycle error. By accumulating all micro-variable errors within a 1-second measurement cycle through time-domain definite integral, the accurate total error offset is obtained:
[0140] ;
[0141] In the formula: t0=0s, t n =1s;
[0142] By using integral calculations to achieve full-domain accumulation of dynamic operating condition errors, the accuracy is significantly improved compared to traditional single-point error calculation, which is the core innovative implementation method of this embodiment.
[0143] Step S4: Iterative optimization of adaptive compensation factor
[0144] S41 Compensation Factor Real-Time Update: Based on the Total Error Offset ΔC Obtained from the Solution total Calculate the adaptive error compensation factor in real time:
[0145] ;
[0146] S42 Iteration Convergence Determination: Preset convergence threshold ε=0.001, perform closed-loop iterative optimization:
[0147] If |η(t)-η(t-1)|>0.001: the parameter is determined to be non-convergent. Continue to collect on-site working conditions and spectral data, repeat the error modeling and dynamic solution process of S2 and S3, and continuously update the compensation factor.
[0148] If |η(t)-η(t-1)|≤0.001: the compensation parameter is considered to have converged, and the current optimal compensation factor η is locked. * Complete the adaptive correction of parameters for this cycle;
[0149] In this embodiment, the convergence time of a single iteration is ≤3s, and the response speed fully meets the requirements of industrial online real-time monitoring, without the need for manual parameter calibration.
[0150] Step S5: Error Compensation and High-Precision Concentration Output
[0151] S51 concentration precise compensation: using the optimal compensation factor η after convergence. * Error compensation was applied to the original measured concentration to obtain a high-precision ammonia concentration value:
[0152] ;
[0153] S52 Outlier Removal Optimization: For sudden changes in flue gas operating conditions and abnormal spectral data caused by sudden changes in power plant load, a 5-point neighborhood mean verification method is used to remove abnormal measurement values and retain stable and valid data, ultimately outputting continuous, stable, and high-precision ammonia escape concentration detection results.
[0154] Step S6: Database self-learning and iterative optimization of operating conditions
[0155] S61 Parameter Storage and Database Construction: This database stores the flue gas temperature, pressure, and flow velocity data measured during the current cycle, along with the sensitivity coefficient and optimal compensation factor η. *The concentration error data before and after compensation are stored one-to-one to build a database of adaptive compensation parameters for power plant denitrification conditions.
[0156] S62 Rapid Adaptation and Optimization: When the unit switches between high and low loads or flue gas conditions fluctuate significantly, the system prioritizes matching the optimal compensation factor of similar operating conditions in the database as the initial iteration parameter, which greatly shortens the parameter convergence time. At the same time, the database parameters are continuously updated to enable the equipment to learn and optimize itself in the long term and adapt to all operating conditions.
[0157] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. An adaptive parameter correction method for laser gas measurement error compensation, characterized in that: Includes the following steps: S1. For the gas to be measured, collect the original spectral absorption signal, laser incident light intensity and laser transmitted light intensity, simultaneously collect environmental data of the measurement area, and construct a time-series correlation dataset; S2. Based on Lambert-Beer's law, a standard gas concentration measurement model is constructed. Combined with environmental data, temperature disturbance error, gas pressure disturbance error, equipment light intensity drift and signal distortion error are decomposed to construct a multi-source error model. S3. Establish a multivariate continuously differentiable function for temperature, air pressure and light intensity attenuation coefficients, solve for the sensitivity coefficients of each variable by first-order partial derivatives, and calculate the total error offset within the measurement period by time-domain integration based on the sensitivity coefficients. S4. Update the error compensation factor in real time using the total error offset, and iteratively optimize the error compensation factor based on the relationship between the error compensation factor of two adjacent measurement cycles and the iterative convergence threshold. S5. Based on the error compensation factor, the measured gas concentration is compensated to eliminate multi-source coupling error and output high-precision gas concentration.
2. The adaptive parameter correction laser gas measurement error compensation method according to claim 1, characterized in that: In S1, for the gas to be measured, the raw spectral absorption signal, incident laser intensity, and transmitted laser intensity are collected. Simultaneously, environmental data of the measurement area are acquired, and a time-series correlated dataset is constructed, including: S11. Adjust the working wavelength of the laser according to the characteristic absorption spectrum of the gas to be measured. After the laser penetrates the measurement area, collect the original spectral absorption signal, the laser incident light intensity and the laser transmitted light intensity, and simultaneously collect environmental data including temperature, air pressure and gas flow rate of the measurement area. S12. Preprocess the original spectral absorption signal by using a moving average filtering algorithm to remove high-frequency random noise and retain the effective spectral absorption signal. S13. Bind the effective spectral absorption signal and environmental data according to the timestamp to construct a time-series associated dataset {I0(t),I(t),T(t),P(t),V(t)}; Where I0(t) is the incident laser intensity at time t, I(t) is the transmitted laser intensity at time t, T(t) is the temperature at time t, P(t) is the gas pressure at time t, and V(t) is the gas velocity at time t.
3. The adaptive parameter correction laser gas measurement error compensation method according to claim 2, characterized in that: S2 constructs a standard gas concentration measurement model based on Lambert-Beer's law, and combines environmental data to decompose temperature disturbance error, gas pressure disturbance error, equipment light intensity drift and signal distortion error, thus constructing a multi-source error model, including: S21. Constructing an error-free standard gas concentration measurement model based on Lambert-Beer's law: ; Where C0(t) is the standard gas concentration at time t, L is the laser path length, and α is the characteristic absorption coefficient of the gas to be measured. S22. Combining environmental data including temperature, air pressure, and gas flow rate in the measurement area, we decompose three core errors: temperature disturbance error, air pressure disturbance error, and equipment light intensity drift and signal distortion error, and construct a multi-source error model: ; Among them, C real (t) represents the measured gas concentration at time t, and ΔC T (t) represents the temperature disturbance error at time t, ΔC P (t) represents the pressure disturbance error at time t, ΔC S (t) represents the device light intensity drift and signal distortion error at time t.
4. The adaptive parameter correction laser gas measurement error compensation method according to claim 3, characterized in that: In S3, a multivariate continuously differentiable function is established for the attenuation coefficients of temperature, air pressure, and light intensity. The sensitivity coefficients of each variable are solved using first-order partial derivatives. Based on the sensitivity coefficients, the total error offset within the measurement period is calculated by time-domain integration, including: S31. Establish a multivariate continuously differentiable function for temperature, air pressure, and light intensity attenuation coefficient, and perform total differential of the multivariate function; S32. Using the standard working conditions of the laboratory as a benchmark, the partial derivatives of the real-time collected environmental data are solved to obtain the temperature sensitivity coefficient, air pressure sensitivity coefficient, and equipment sensitivity coefficient at each sampling time. S33. Since environmental data is a continuous time-series change, the instantaneous micro-change error needs to be accumulated through time-domain integration to calculate the total error offset within the measurement period.
5. The adaptive parameter correction laser gas measurement error compensation method according to claim 4, characterized in that: In S31, a multivariate continuously differentiable function is established for temperature, air pressure, and light intensity attenuation coefficients, and the total differential of the multivariate function is performed, including: Establish a multivariate continuously differentiable function C(t) = f(T(t), P(t), k(t)) with respect to temperature, air pressure, and light intensity attenuation coefficients, and perform total differential of the multivariate function: ; Where k(t) is the light intensity attenuation coefficient at time t, which is calculated from the incident laser intensity I0(t) and the transmitted laser intensity I(t) at time t. The temperature sensitivity coefficient at time t is... Let be the pressure sensitivity coefficient at time t. Let be the device sensitivity coefficient at time t, and d be the differential operator.
6. The adaptive parameter correction laser gas measurement error compensation method according to claim 5, characterized in that: S32 uses standard laboratory conditions as a benchmark to solve for partial derivatives of the real-time acquired environmental data, obtaining the temperature sensitivity coefficient, air pressure sensitivity coefficient, and equipment sensitivity coefficient at each sampling time, including: The temperature sensitivity coefficient, air pressure sensitivity coefficient, and equipment sensitivity coefficient are calculated using the following formulas: ; ; ; Among them, S T (t) is the temperature sensitivity coefficient at time t, S P (t) is the pressure sensitivity coefficient at time t, S k (t) represents the device sensitivity coefficient at time t.
7. The adaptive parameter correction laser gas measurement error compensation method according to claim 6, characterized in that: In S33, since the environmental data is a continuously changing quantity over time, the instantaneous micro-variable error needs to be accumulated through time-domain integration to calculate the total error offset within the measurement period, including: The total error offset within the measurement period is calculated by performing time-domain integration and accumulation using the following formula: ; Among them, △C total The total error offset within the measurement period, t0, t n These are the start and end times of a single measurement cycle, t0 ≤ t ≤ t n .
8. The adaptive parameter correction laser gas measurement error compensation method according to claim 4, characterized in that: In S4, the error compensation factor is updated in real time using the total error offset, and iterative optimization of the error compensation factor is performed based on the relationship between the error compensation factor of two adjacent measurement cycles and the iteration convergence threshold, including: S41. Update the error compensation factor in real time based on the total error offset within the measurement period: ; Where η(t) is the error compensation factor at time t; S42. Introduce an iterative convergence threshold ε and iteratively optimize the error compensation factor: When the absolute value of the difference between the error compensation factor of the current measurement cycle and the error compensation factor of the previous measurement cycle is greater than the iteration convergence threshold ε, the original spectral absorption signal, laser incident light intensity and laser transmitted light intensity are continuously collected, environmental data of the measurement area are collected simultaneously, and a time-series correlation dataset is constructed. S2, S3 and S41 are repeated to update the measured gas concentration and error compensation factor. When the absolute value of the difference between the error compensation factor of the current measurement cycle and the error compensation factor of the previous measurement cycle is not greater than the iteration convergence threshold ε, the error compensation factor is determined to have converged, and the error compensation factor of the current measurement cycle is taken as the current optimal error compensation factor η. * This enables the compensation parameters to be adaptively adjusted according to operating conditions and equipment status. The iterative convergence threshold ε is preset to 0.001 and can be adjusted according to the measurement accuracy requirements.
9. The adaptive parameter correction laser gas measurement error compensation method according to claim 8, characterized in that: S5 compensates for measured gas concentrations based on an error compensation factor, eliminating multi-source coupling errors and outputting high-precision gas concentrations, including: S51. Based on the current optimal error compensation factor, compensate for the measured gas concentration and output a high-precision gas concentration: ; Among them, C final (t) represents the high-precision gas concentration at time t; S52. Outlier removal is performed on high-precision gas concentration data. For instantaneous abnormal correction data caused by sudden changes in operating conditions, the neighborhood mean verification method is used to screen valid measurement values, and finally, stable and accurate gas concentration detection results are output.
10. The adaptive parameter correction laser gas measurement error compensation method according to claim 9, characterized in that: S5 compensates for the measured gas concentration based on an error compensation factor, eliminates multi-source coupling errors, and outputs a high-precision gas concentration, including: S61. Store the environmental data, sensitivity coefficient, current optimal error compensation factor, and gas concentration error value before and after compensation obtained in each measurement to construct an adaptive compensation parameter database. S62. In subsequent measurement processes, the current optimal error compensation factor under similar working conditions in the adaptive compensation parameter database is first matched as the initial iteration parameter to improve the convergence speed of the compensation parameter. At the same time, the adaptive compensation parameter database is updated to achieve self-learning optimization.
