Industrial gas concentration analysis system based on infrared light

Abstract

The invention relates to the technical field of industrial gas analysis, and particularly discloses an industrial gas concentration analysis system based on infrared light, which is used for solving the problems of low wavelength drift control precision, poor disturbance adaptability, insufficient multi-gas regulation compatibility and lack of an active feedback mechanism in the existing industrial gas detection. Comprising an environmental parameter acquisition module and an infrared light gas analysis regulation and control module, temperature, humidity and pressure disturbance information is acquired in real time and subjected to standardization processing, and man-machine collaborative embedding characteristics are constructed by combining physical characteristics of CO, N2O and H2O, state characteristics of a heating furnace and regulation and control behavior characteristics of a laser; according to the method, through embedding fusion representation of the word vectors and the knowledge graph, disturbance scene modeling and adjustment strategy prediction are achieved, the precision, responsiveness and system generalization ability of wavelength regulation and control are improved, and the method is suitable for high-reliability wavelength control in a multi-gas-concentration cooperative detection scene.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial gas analysis, and more specifically, to an industrial gas concentration analysis system based on infrared light. Background Art

[0002] Industrial gas concentration analysis is widely used in many key areas such as safety monitoring, environmental governance and process control. Tunable Diode Laser Absorption Spectroscopy (TDLAS) has been widely used in various gas detection scenarios due to its high selectivity, high sensitivity and fast responsiveness. Existing industrial systems generally use mid-infrared quantum cascade lasers as the core device for gas detection. By controlling the temperature and current of the laser to adjust its emission wavelength, it covers the specific absorption spectrum of the gas to be measured, thereby achieving high-precision inversion of gas concentration. However, the actual industrial environment is often accompanied by complex dynamic interferences such as temperature, humidity and pressure disturbances, and heating furnace start-stop fluctuations. These factors can cause nonlinear drift of the laser wavelength, causing photoelectric signal fluctuations and concentration inversion errors.

[0003] Although the existing system has introduced a dynamic environmental compensation module, and uses the environmental sensor module to collect temperature, humidity, pressure and other data for normalization, construct environmental mapping items and time integral items, and correct the signal based on the compensation factor, this type of compensation method still belongs to a passive response mechanism, lacking deep modeling of the disturbance structure and semantic interpretation of the behavior pattern. Especially in scenarios where multiple gases are measured together, the control history is complex, and the operation behavior has time series characteristics, the traditional PID or decoupling control scheme faces the problems of insufficient control and weak generalization ability. It lacks structured modeling and deep correlation analysis of the environmental state, and fails to fully utilize the historical control behavior of the laser and the physical response law of the gas to form a general control experience model. Especially in the scenario of simultaneous detection of multiple gases, different gases have different sensitivities to disturbances, and the laser control strategy cannot be integrated and the response accuracy is poor. In addition, the feedback path of the existing system is mostly limited to one-way closed-loop regulation, lacks active intervention channels from engineering experts or user experience, and cannot form a high-responsiveness, high-precision and high-generalization control system under complex working conditions. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides an industrial gas concentration analysis system based on infrared light, which introduces a human-machine collaborative feedback control mechanism, integrates gas characteristics, heating furnace status and historical control behavior, and combines environmental disturbance modeling to achieve CO, N 2 O, H 2 High-precision wavelength adjustment and concentration detection of O3 gas under dynamic conditions.

[0005] To achieve the above object, the present invention provides the following technical solutions: An industrial gas concentration analysis system based on infrared light, which can make the laser pass through the measured gas twice to enhance CO, N 2 O, H 2 The signal intensity of the concentration of three gases is detected simultaneously, and a dynamic environment compensation correction signal is introduced, including a transceiver and a reflector. The transceiver includes a laser. The laser is an infrared quantum cascade laser with an output wavelength range of 2202.8 to 2205.6 cm -1 An off-axis parabolic mirror with a hole in the middle is provided on one side of the laser, a window piece 1 is provided on the side of the off-axis parabolic mirror away from the laser, and the window piece 1 is used for the first emission of the infrared laser and the re-incidence after reflection, a filter is provided above the off-axis parabolic mirror, a photoelectric detector is provided above the filter, a window piece 2 and a full-angle reflector are provided at the reflecting end, a red light generator is provided in the laser, and the red light generator is used for the combined beam of red light and infrared light as an emission path indication, the photoelectric detector is connected to a data inversion analysis module, the data inversion analysis module is connected to a dynamic environment compensation module, and the dynamic environment compensation module is connected to an environmental parameter The infrared light gas analysis and control module is connected to the infrared light gas data acquisition module. The environmental parameter acquisition module collects the temperature, humidity and pressure within 0.5 cm of the laser in real time and performs standardization to construct a normalized environmental state vector. The infrared light gas analysis and control module is connected to the laser. Based on the characteristics of the measured gas, the state characteristics of the heating furnace, and the characteristics of the laser control behavior, a collaborative feature representation for simultaneous detection of the measured gas concentration is constructed to predict the infrared light adjustment parameters of the tunable laser output based on artificial collaborative feedback. The measured gases include CO, N 2 O, H 2 O.

[0006] As a further solution of the present invention, the infrared light gas analysis and control module includes a human-machine collaborative feature construction submodule, a human-machine collaborative feature embedding submodule, and an infrared light control parameter prediction module. The human-machine collaborative feature construction submodule is connected to the human-machine collaborative feature embedding submodule, the human-machine collaborative feature embedding submodule is connected to the infrared light control parameter prediction module, and the infrared light control parameter prediction module is connected to the environmental parameter acquisition module and the laser.

[0007] In order to extract the measured gas characteristics, heating furnace state characteristics, and laser control behavior characteristics from the existing data: As a further solution of the present invention, in the human-machine collaborative feature construction submodule, the method of obtaining the measured gas characteristics is: first, according to CO, N 2 O, H 2O The labels are initialized based on the physical properties of the three gases and the external knowledge graph. After the laser enters the operation stage, the control response behavior labels are extracted according to the historical wavelength shift performance and control response of the three gases under different environmental conditions. The control response behavior labels are weighted by time decay according to the occurrence time to obtain the weighted control response behavior labels, which are used as the characteristic labels of the measured gases. The external knowledge graph is a structured target detection gas physical response knowledge graph, including the semantic relationship between the target detection gas type, environmental state, laser control method and control behavior results. The construction form is a triple, which is used to provide an initial label set based on external experience and experimental data. The external knowledge graph covers CO, N 2 O, H 2 The typical absorption band, interference sensitivity, response law under environmental disturbance and recommended control strategy of O gas are analyzed. In the initialization stage, the external knowledge graph is called to assign a label vocabulary set to each gas as the characteristic label of the measured gas.

[0008] As a further scheme of the present invention, in the human-machine collaborative feature construction submodule, the method of obtaining the state characteristics of the heating furnace is: using the structural parameters and control mode of the heating furnace to initialize the label list, constructing the initialization label, and during the operation of the heating furnace, according to the dynamic changes of the temperature, pressure and airflow monitoring data in the furnace, identifying the current operating behavior state of the heating furnace, constructing a dynamic behavior label, and assigning different weights to the operating behavior labels in different time periods according to a preset time attenuation function, forming a set of operating behavior labels that reflect the current operating trend of the heating furnace, which is used as the heating furnace state feature label.

[0009] As a further solution of the present invention, the time decay function formula is: ; Where: To run the behavior tag index, For the The timestamp when the running behavior tag is identified, is the memory retention attenuation coefficient, For the current moment, For the The weight of a running behavior label.

[0010] As a further solution of the present invention, in the human-machine collaborative feature construction submodule, the method of obtaining the laser control behavior characteristics is: the control instruction log recorded during the operation of the laser is used as a human-machine collaborative behavior text set, each log includes the control action, adjustment parameters, and feedback response words of the laser, and all words in the behavior text are segmented and TF-TDF weighted, and the weighted key behavior labels with TF-TDF weights within the set threshold range are extracted to obtain a list of key behavior weighted labels for each log, which are used as laser control behavior feature labels.

[0011] In order to embed the measured gas characteristics, heating furnace state characteristics, and laser control behavior characteristics as labels to obtain human-machine collaborative embedding features: As a further solution of the present invention, in the human-machine collaborative feature embedding submodule, all labels are embedded and weighted summed using a pre-trained word vector model to obtain the WE feature of the label group. ; The TransE model is used to quantify each node of the external knowledge graph, and the objective function is: ; Where: is the embedding objective function value, , are respectively the positive sample set in the external knowledge graph and the randomly sampled negative sample set. , , are the head entity, relation, and tail entity in the triple respectively. , They are used to generate the head entity and tail entity that are randomly replaced by negative samples, , , They are the KGE (label embedding based on external knowledge graph) representations of the head entity, relationship, and tail entity, respectively. , are the KGE representations of the head entity and tail entity of the negative sample, respectively. is the interval coefficient, where The measured gas feature labels, heating furnace state feature labels and laser control behavior feature labels are input into the TransE model and then weighted summed to obtain the KGE feature of the label group ; The WE features and KGE features of the label group are concatenated to obtain the human-machine collaborative embedding features: ; in, To embed features for human-machine collaboration, , Dimensions for embedding features for human-machine collaboration.

[0012] In order to combine the normalized environment state vector with the human-machine collaborative embedding representation: As a further solution of the present invention, the infrared light control parameter prediction module includes a feature recall submodule and an infrared light adjustment amount sorting submodule.

[0013] As a further solution of the present invention, the feature recall submodule utilizes the current human-machine collaborative embedded features and the normalized environmental state vector to jointly construct the current disturbance scene vector, calculates the cosine similarity between the current disturbance scene vector and the historical disturbance scene vector in the training set, selects the disturbance scenes whose cosine similarity is within the set threshold range as the recall results, and extracts the corresponding laser output infrared light control response behavior as the candidate adjustment strategy set.

[0014] As a further solution of the present invention, the infrared light adjustment amount sorting submodule adopts a regression model to score each group of historical adjustment strategies in the recall candidate adjustment strategy set. The input features include the current disturbance scene vector and the recall strategy features. The predicted score of each group of adjustment strategies in the current scene is output, and the predicted scores are arranged in descending order. The one with the highest score is selected as the target adjustment parameter of the current laser.

[0015] In order to deal with the problem of no historical response data reference in the initial operation state of the laser: As a further solution of the present invention, the infrared light control parameter prediction module also includes an initial parameter analysis submodule. In the first operation stage of the laser, the initial parameter analysis submodule constructs an initial input feature vector based on the normalized environmental state vector, the measured gas feature label and the embedded features of the heating furnace state label, and calls the pre-trained initial regression model to predict the infrared light target adjustment parameters of the laser. First, three target gases (CO, N 2 O, H 2 O) Initial laser wavelength control strategy under the current environmental parameters (temperature, humidity, pressure). After these strategy labels are input into TransE embedding, they are combined with the current normalized environmental state vector to form an initial input feature vector. The initial input feature vector is inferred using a pre-trained regression model to output the initial laser target adjustment parameters. The regression model is built based on existing experimental data before system deployment. When the laser enters stable operation, the system gradually transitions to a tunable laser wavelength control method based on human-machine collaborative feedback. The laser target adjustment parameters are predicted based on the measured gas characteristics, the heating furnace operating status, and the laser historical control behavior.

[0016] The technical effects of the infrared light-based industrial gas concentration analysis system of the present invention are as follows: The present invention combines infrared quantum cascade lasers with a double-pass absorption path to achieve CO, N 2O, H 2 O gas is detected simultaneously. By collecting the temperature, humidity and pressure disturbances around the laser, a normalized environmental state vector is constructed to achieve standardized modeling of the laser operating environment. A human-machine collaborative feedback control mechanism is introduced to integrate the physical properties of the measured gas, the operating status of the heating furnace and the historical control behavior of the laser. A human-machine collaborative feature representation is constructed. The infrared light adjustment parameters of the laser are further predicted based on the human-machine collaborative features to achieve a priori recognition of the laser wavelength deviation trend and improve the active control capability of the laser output infrared light. 2 O, H 2 In the common measurement of O concentration and the dynamic and frequent gas concentration detection in the heating furnace, the accuracy, response speed and control stability of the laser wavelength adjustment are improved, the inversion error of the measured gas concentration is effectively reduced, and the CO and N 2 O, H 2 Robustness and intelligent adaptive capabilities of O concentration co-measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the optical path of the prior art of the present invention; Figure 2 is a system block diagram of the present invention; In the figure: 01, laser; 02, off-axis parabolic mirror; 03, window one; 04, filter; 05, photodetector; 06, window two; 07, full-angle reflector; 101, transceiver; 102, reflector. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] Example 1

[0020] like Figure 1 As shown in the figure, the present invention proposes an industrial gas concentration analysis system based on infrared light, and the existing system established by the present invention can make the laser pass through the measured gas twice to enhance the CO and N 2 O, H 2The signal intensity of the concentration of three gases is detected simultaneously, and a dynamic environment compensation correction signal is introduced, including a transceiver end 101 and a reflector end 102. A laser 01 is arranged in the transceiver end 101, and a red light generator is installed in the laser 01. An off-axis parabolic mirror 02 with a hole in the middle is arranged on one side of the laser 01. A filter 04 is arranged above the off-axis parabolic mirror 02, and a photodetector 05 is arranged above the filter 04. A window 1 03 is arranged on the side of the off-axis parabolic mirror 02 away from the laser 01. A window 2 06 is arranged in the reflector end 102, and a side of the window 2 06 is provided. The full-angle reflector 07 and the photoelectric detector 05 are connected to a data inversion analysis module, which is connected to a dynamic environment compensation module, which is connected to an environment sensing module. The dynamic environment compensation module uses the environmental interference data to construct a normalized vector and a normalized deviation of the environmental variables, obtains environmental parameter mapping items, weightedly accumulates the historical gas concentration inversion error within the set time window, obtains the environmental parameter time integral item, and inputs the two into the dynamic environment compensation formula to dynamically compensate for environmental interference. The optical path between the transceiver 101 and the reflector 102 includes: The red indicator light emitted by the red light generator is combined with the infrared light emitted by the laser 01 as the output light beam, which first passes through the middle opening of the off-axis parabolic mirror 02 and the window piece 1 03, enters the gas area to be measured, and completes the first absorption; then the laser beam passes through the window piece 2 06 and shoots to the full-angle reflector 07, and after being reflected along the original path, it passes through the window piece 2 06 again to enter the gas area to be measured, and is absorbed for the second time. The reflected light beam passes through the window piece 1 03 and then enters the parabola of the off-axis parabolic mirror 02. Its reflected light passes through the filter 04 and is focused onto the light sensitive surface of the photodetector 05. The light signal collected by the photodetector 05 is converted into an electrical signal through photoelectric conversion, and then the data inversion analysis module performs detection electrical signal processing and gas concentration inversion calculation, performs dynamic environment correction based on the dynamic environment compensation algorithm, and feedback controls the laser 01.

[0021] It should be further explained that in order to further solve the impact of the current scene disturbance on the wavelength output of Laser 01, improve the accuracy of data inversion, and conduct structured modeling and deep correlation analysis of environmental states, make full use of the historical control behavior of Laser 01 and the physical response law of gases to form a general control experience model. Especially in the scenario of simultaneous detection of multiple gases, different gases have different sensitivities to disturbances. The control strategy of Laser 01 is integrated and compatible, and the response accuracy is improved. Combined with the active intervention channel from engineering experts or user experience, a control system with high responsiveness, high precision and high generalization ability is formed under complex working conditions, such as Figure 2As shown, the improvement of the present invention lies in that the analysis module is connected to a dynamic environment compensation module, which is connected to an environment parameter acquisition module and an infrared light gas analysis and control module. The environment parameter acquisition module collects the temperature, humidity and pressure within a range of 0.5 cm from the laser 01 in real time and performs standardization processing to construct a normalized environment state vector. The infrared light gas analysis and control module is connected to the laser 01. Based on the characteristics of the measured gas, the characteristics of the heating furnace state, and the characteristics of the laser 01 control behavior, a collaborative feature representation for simultaneous detection of the measured gas concentration is constructed to predict the infrared light adjustment parameters output by the tunable laser 01 based on artificial collaborative feedback. The measured gases include CO, N 2 O, H 2 O.

[0022] For example, laser 01 was first used in an industrial heating furnace at a temperature of 38°C, a humidity of 70%, and a pressure of 101.2 kPa to detect N 2 O concentration, by calling the external knowledge graph, find N 2 O is most sensitive to the control of the laser 01 wavelength range of 1280-1320nm in an environment with humidity greater than 65%. According to historical data, N 2 O response wavelength has a +1.3nm shift. The system uses this shift behavior as a label for the wavelength to shift to long waves, and assigns a weight of the set threshold (because it is close to the current state). The final output feature label is long-wave shift, high humidity interference sensitivity, and medium absorption intensity. The heating furnace runs continuously for 60 minutes, the temperature rises from 400°C to 600°C, and the pressure and airflow rise steadily. The operating data is detected to identify the heating stage and continuous heating behavior. Multiple heating labels are recorded in the past 40 minutes, but the earlier the time, the lower the weight. The label weight 40 minutes ago is reduced to 0.13, and the label weight 10 minutes ago is 0.61. The current output The characteristic labels of the heating furnace status are significant heating trend and high thermal disturbance risk warning; by inputting the control instruction "adjust gain +1.5, observe the wavelength drift", the control instruction text is recorded in the instruction log, and the keywords "adjustment", "gain", "wavelength", and "drift" are counted after word segmentation. Combined with TF-TDF weight calculation, "gain" and "wavelength" have significant weights, and they are extracted as key behavior labels. The "gain + adjustment" behavior that frequently appears in multiple consecutive instructions is further weighted. By extracting the high-frequency operation behavior in the control instructions, the human-machine control intention can be quickly understood, providing a data basis for behavior prediction and adjustment optimization.

[0023] It should be noted that the infrared light gas analysis and control module includes a human-machine collaborative feature construction submodule, a human-machine collaborative feature embedding submodule, and an infrared light control parameter prediction module. The human-machine collaborative feature construction submodule is connected to the human-machine collaborative feature embedding submodule, the human-machine collaborative feature embedding submodule is connected to the infrared light control parameter prediction module, and the infrared light control parameter prediction module is connected to the environmental parameter acquisition module and laser 01.

[0024] In order to extract the measured gas characteristics, heating furnace state characteristics, and laser 01 control behavior characteristics from the existing data: Specifically, the human-machine collaborative feature construction submodule obtains the characteristics of the measured gas. More specifically, first, according to CO, N 2 O, H 2 O The labels are initialized based on the physical properties of the three gases and the external knowledge graph. After laser 01 enters the operation stage, the control response behavior labels are extracted according to the historical wavelength shift performance and control response of the three gases under different environmental conditions. The control response behavior labels are weighted by time decay according to the occurrence time to obtain the weighted control response behavior labels, which are used as the characteristic labels of the measured gases. The external knowledge graph is a structured target detection gas physical response knowledge graph, including the semantic relationship between the target detection gas type, environmental state, laser 01 control method and control behavior results. The construction form is a triple, which is used to provide an initial label set based on external experience and experimental data. The external knowledge graph covers CO, N 2 O, H 2 The typical absorption band, interference sensitivity, response law under environmental disturbance and recommended control strategy of O gas are analyzed. In the initialization stage, the external knowledge graph is called to assign a label vocabulary set to each gas as the characteristic label of the measured gas; Specifically, the human-machine collaborative feature construction submodule obtains the state characteristics of the heating furnace. More specifically, the label list is initialized using the structural parameters and control mode of the heating furnace to construct the initialization label. During the operation of the heating furnace, the current operation behavior state of the heating furnace is identified according to the dynamic changes of the temperature, pressure and airflow monitoring data in the furnace, and the dynamic behavior label is constructed. Different weights are assigned to the operation behavior labels in different time periods according to the preset time decay function to form an operation behavior label set that reflects the current operation trend of the heating furnace, which is used as the state feature label of the heating furnace. The time decay function formula is: ; Where: To run the behavior tag index, For the The timestamp when the running behavior tag is identified, is the memory retention attenuation coefficient, For the current moment, For the The weight of the running behavior label; Specifically, the human-machine collaborative feature construction submodule obtains the control behavior characteristics of the laser 01. More specifically, the control instruction log recorded during the operation of the laser 01 is used as the human-machine collaborative behavior text set. Each log includes the control actions, adjustment parameters, and feedback response words of the laser 01. By performing word segmentation and TF-TDF weight calculation on all words in the behavior text, the weighted key behavior labels with TF-TDF weights within the set threshold range are extracted to obtain a list of key behavior weighted labels for each log, which are used as the control behavior feature labels of the laser 01.

[0025] In order to clearly illustrate the extraction of the above three features, an example is given below: For example, when a system deployed inside an industrial heating furnace starts laser 01 for the first time, the current detection object is a mixed gas of CO, N 2 The environmental parameters collected by the system are temperature 35°C, humidity 68%, and pressure 100.8 kPa. The goal is to determine the CO and N 2 O's laser 01 wavelength control feature is used to assist in subsequent wavelength adjustment. First, the structured target detection gas physical response knowledge graph is loaded, which contains three-dimensional group information (CO, high humidity, negative wavelength shift), (N 2 O, high temperature, increased response bandwidth), (CO, high pressure, recommended strategy: fine-step adjustment), (N 2 O, strong interference, recommended strategy: heating furnace temperature increase), according to CO and N 2 O gas type, extract its typical labels under high humidity and high temperature conditions, such as the label vocabulary set assigned to CO ("absorption center wavelength 1560nm", "high humidity offset risk increases", "recommended adjustment strategy: reduce adjustment step size"), 2 The label vocabulary set assigned to O ("wide absorption range, band coverage 1285–1325 nm", "nonlinear response enhancement at high temperature", "suggested strategy: adapt to the temperature rise state") was used to retrieve the past N 2 The control records of O were retrieved and its historical offset was found to be +1.4nm. The control behavior of CO in a high humidity environment was retrieved and its drift was found to be -0.6nm. These control response behaviors were extracted and given time decay weights. For example, the weight of data within a week is 0.92, the weight of data one month ago is 0.41, and the weight of data half a year ago is 0.07. The initialization label is fused with the historical control behavior label to output the weighted measured gas feature label set, as shown in Table 1: Table 1 Weighted characteristic label set of the measured gas

[0026] Even if the system is running for the first time and there is no current on-site feedback behavior, it can reasonably construct CO and N based on external experience (knowledge graph) and historical response records. 2 The current state characteristics of O provide high-value feature input for the recall and sorting of subsequent adjustment strategies, improve the accuracy and stability of initial adjustment, and integrate knowledge labels and experience labels to make the system have scalable, explainable and generalizable initial regulation capabilities.

[0027] By introducing mechanisms based on external knowledge graph initialization, historical behavior label extraction, time weighting and semantic text modeling in the human-machine collaborative construction submodule, the characteristics of the measured gas, the state characteristics of the heating furnace and the control behavior characteristics of the laser 01 are extracted respectively. The physical properties of the measured gas and the external knowledge graph can be used to construct the characteristic labels of the measured gas, so that in the initial stage where there is a lack of real-time data, the interference type and the wavelength response law of the laser 01 can be quickly identified with the help of known experience, forming a universal control experience mode. Especially in the scenario of simultaneous detection of multiple gases, the laser 01 has different sensitivities to disturbances, and the laser 01 can be controlled by the state characteristics of the heating furnace. 1. The control strategy achieves integrated compatibility and improved response accuracy. Combined with the active intervention channel from engineering experts or user experience external knowledge, a control system with high responsiveness, high accuracy and high generalization ability is formed under complex working conditions. By weighting the dynamic labels of the heating furnace operation behavior with time decay, a behavioral trend representation is constructed to improve the perception and response capabilities of the complex dynamic changes of the heating furnace. The control instruction log is used as semantic text input, and the key control behaviors are extracted through TF-TDF weights. It is possible to extract significantly representative behavior labels from unstructured control behaviors to facilitate the understanding and extraction of control strategies.

[0028] In order to embed the measured gas characteristics, heating furnace state characteristics, and laser 01 control behavior characteristics as labels to obtain human-machine collaborative embedding features: It should be noted that in the human-machine collaborative feature embedding submodule, all labels are embedded using the pre-trained word vector model and then weighted summed to obtain the WE feature of the label group. ; The TransE model is used to quantify each node of the external knowledge graph, and the objective function is: ; Where: is the embedding objective function value, , are respectively the positive sample set in the external knowledge graph and the randomly sampled negative sample set. , , are the head entity, relation, and tail entity in the triple respectively. , They are used to generate the head entity and tail entity that are randomly replaced by negative samples, , , They are the KGE (label embedding based on external knowledge graph) representations of the head entity, relationship, and tail entity, respectively. , are the KGE representations of the head entity and tail entity of the negative sample, respectively. is the interval coefficient, where The measured gas feature label, heating furnace state feature label and laser 01 control behavior feature label are input into the TransE model and then weighted summed to obtain the KGE feature of the label group ; The WE features and KGE features of the label group are concatenated to obtain the human-machine collaborative embedding features: ; in, To embed features for human-machine collaboration, , Dimensions for embedding features for human-machine collaboration.

[0029] For example, when laser 01 is currently running, the status detected by the heating furnace is the current temperature of 620°C, the current pressure of 0.8MPa, the current airflow is rising steadily, the continuous operation time of the heating furnace is 70 minutes, and the status labels in the last 30 minutes (temperature rise, high pressure fluctuation, heat flow disturbance). In the human-machine collaborative feature construction submodule, the weights are calculated based on the acquired monitoring data + time decay function, and the following labels are extracted and quantified as shown in Table 2: Table 2 Heating furnace status label display table

[0030] Use the pre-trained word vector model to encode the above label words, and the warming state is , and the thermal disturbance trend is , pressure instability , sum each word vector with the weight to get the WE feature of the heating furnace , and then read the triples (heating furnace, state is, heating), (heating, leads to, thermal disturbance), (thermal disturbance, association, wavelength drift), (high pressure, recommended strategy, steady-state regulation) from the structured external knowledge graph, and map the entities and relations in these triples to the corresponding TransE embedding model , , , calculated by the trained TransE embedding model , get the positive sample scores of all triplets, randomly sample to generate negative samples, and optimize the total loss function , output the corresponding entity vector embedding, and obtain the weighted sum of the representations of these labels in the TransE embedding space Similarly, the embedded feature representations of the historical control behavior of the measured gas machine laser 01 are , ,final , , and then concatenate to get the human-machine collaborative embedding feature vector , this vector is the core input feature of the subsequent recall and adjustment strategy ranking model.

[0031] In order to combine the normalized environment state vector with the human-machine collaborative embedding representation: It should be noted that the infrared light control parameter prediction module includes a feature recall submodule and an infrared light adjustment amount sorting submodule.

[0032] It should be further explained that the feature recall submodule uses the current human-machine collaborative embedded features and the normalized environmental state vector to jointly construct the current disturbance scene vector, calculates the cosine similarity between the current disturbance scene vector and the historical disturbance scene vector in the training set, selects the disturbance scene whose cosine similarity is within the set threshold range as the recall result, and extracts the corresponding laser 01 output wavelength control response behavior as the candidate adjustment strategy set.

[0033] It should be further explained that the infrared light adjustment amount sorting submodule uses a regression model to score each group of historical adjustment strategies in the recall candidate adjustment strategy set. The input features include the current disturbance scene vector and the recall strategy features. The predicted score of each group of adjustment strategies in the current scene is output, and they are arranged in descending order according to the predicted score. The one with the highest score is selected as the target adjustment parameter of the current laser 01.

[0034] The current system not only considers the control behavior and knowledge labels of laser 01, but also simultaneously integrates the environmental status (temperature, humidity, pressure) to construct a high-dimensional disturbance scene vector; this scene vector has both environmental characteristics and semantic embedding features, and can effectively express the combined impact of current external disturbances and equipment status; by matching the cosine similarity with historical disturbance scenes, it can recall successful strategies in highly similar scenes from the "experience library", making up for the defect that traditional adjustment algorithms only make local optimal predictions based on current data, so that the system has scene recognition and memory capabilities.

[0035] Example 2

[0036] The difference between Example 2 of the present invention and Example 1 is that this example introduces the use of the tunable laser 01 wavelength control system based on human-machine collaborative feedback when it is used for the first time without the historical control behavior data of the laser 01.

[0037] In order to deal with the problem of no historical response data reference in the initial operation state of laser 01, the infrared light control parameter prediction module also includes an initial parameter analysis submodule. In the first operation stage of laser 01, the initial parameter analysis submodule constructs an initial input feature vector based on the normalized environmental state vector, the measured gas feature label and the embedded features of the heating furnace state label, and calls the pre-trained initial regression model to predict the target adjustment parameters of laser 01. First, three target gases (CO, N 2 O, H 2 O) The initial laser 01 wavelength control strategy under the current environmental parameters (temperature, humidity, pressure). After these strategy labels are input into TransE embedding, they are combined with the current normalized environmental state vector to form an initial input feature vector. The initial input feature vector is inferred using a pre-trained regression model to output the initial laser 01 target adjustment parameters. The regression model is built based on existing experimental data before system deployment. When laser 01 enters stable operation, the system gradually transitions to a tunable laser 01 wavelength control method based on human-machine collaborative feedback. The laser 01 target control parameters are predicted based on the measured gas characteristics, the heating furnace operating status, and the laser 01 historical control behavior.

[0038] Specifically, in the industrial heat treatment scenario, the system was deployed in an industrial heating furnace emission monitoring device and was used for the first time to detect CO and N in a high temperature environment. 2 O concentrations of the two target gases. When laser 01 is initially started, the normalized environmental state vector collected by the environmental parameter acquisition module is: [0.92, 0.30, 0.88], corresponding to the furnace temperature of 480°C, humidity of 35%, and pressure of 0.95MPa. The target detection gases are CO and N 2 O, the heating furnace is in the initial constant temperature operation state.

[0039] The initial parameter analysis submodule first calls the structured target detection gas physical response knowledge graph to extract the label triple information that matches the current detection gas and environmental state, including: "CO is recommended to fine-tune the wavelength under high temperature and high pressure conditions, and the adjustment step is halved", "N 2 O response band moves up in dry atmosphere", "there is a moderate risk of thermal interference in the initial stage of constant temperature", etc. Based on this, the system generates a set of labels, including "high temperature and high pressure fine-tuning", "band moving up", "moderate interference", and "adjustment step size halved".

[0040] Subsequently, the human-machine collaborative feature embedding submodule inputs the above labels into the pre-trained word vector model and knowledge graph embedding model respectively. The word vector model is used to extract the semantic embedding of the label, and the WE feature vector is obtained after the weighted summation; the knowledge graph embedding model (using the TransE method) is used to quantify the relevance of the label in the structured triples to obtain the weighted KGE feature vector. The normalized environment state vector is concatenated with the above two label embedding vectors to construct the initial input feature vector.

[0041] The initial input feature vector is input into the initial regression model for inference. The model is trained based on a large amount of historical industrial experimental data collected before the system is deployed, and has the ability to predict the initial adjustment strategy. Finally, the system outputs the initial target wavelength adjustment parameters of laser 01 under the current environmental conditions, and the predicted value is 1573.6nm.

[0042] This adjustment parameter is used for wavelength setting in the first operation of laser 01, and realizes the automatic generation of a more accurate adjustment strategy based on the target gas characteristics, heating furnace status and environmental disturbance conditions without the participation of historical behavior data. As the feedback data during the operation of laser 01 gradually accumulates, the system can smoothly transition to a wavelength control mode based on human-machine collaborative feedback, feature recall and strategy sorting mechanism. This implementation method effectively realizes the rapid cold start capability of laser 01, improves the stability and prediction accuracy of the system's first operation, and has good engineering adaptability and scalability.

[0043] In summary, in combination with Example 1 and Example 2, the present invention combines the infrared quantum cascade laser 01 with the double-pass absorption path to achieve the absorption of CO and N 2 O, H 2 O gas is detected simultaneously. By collecting the temperature, humidity and pressure disturbances around the laser 01, a normalized environmental state vector is constructed to realize the standardized modeling of the operating environment of the laser 01. The human-machine collaborative feedback control mechanism is introduced to integrate the physical properties of the measured gas, the operating status of the heating furnace and the historical control behavior of the laser 01. The human-machine collaborative feature representation is constructed, and the wavelength adjustment parameters of the laser 01 are further predicted based on the human-machine collaborative features to realize the prior recognition of the wavelength deviation trend of the laser 01, improve the active control capability of the output wavelength of the laser 01, and in CO, N 2 O, H 2 In the common measurement of O concentration and the dynamic and frequent gas concentration detection in the heating furnace, the accuracy, response speed and control stability of the laser 01 wavelength adjustment are improved, the inversion error of the measured gas concentration is effectively reduced, and the CO and N 2 O, H 2 Robustness and intelligent adaptive capabilities of O concentration co-measurement.

[0044] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0045] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An industrial gas concentration analysis system based on infrared light, comprising a transceiver end (101) and a reflector end (102), wherein the transceiver end (101) comprises a laser (01), a photoelectric detector (05) on one side of the laser (01), the photoelectric detector (05) is connected to a data inversion analysis module, and the data inversion analysis module is connected to a dynamic environment compensation module, characterized in that: The dynamic environment compensation module is connected to the environment parameter acquisition module, the environment parameter acquisition module is connected to the infrared light gas analysis and control module, the infrared light gas analysis and control module is connected to the laser, the environment parameter acquisition module real-time acquires the temperature, humidity and pressure within the range of 0.5 cm from the laser, and performs standardization processing to construct a normalized environment state vector, the infrared light gas analysis and control module constructs a collaborative feature representation for simultaneous detection of the concentrations of three gases under test, namely, CO, N2O and H2O, based on the characteristics of the measured gas, the characteristics of the heating furnace state and the characteristics of the laser control behavior, so as to predict the infrared light adjustment parameters output by the tunable laser (01) based on human-machine collaborative feedback.

2. The infrared light-based industrial gas concentration analysis system according to claim 1, characterized in that: The infrared light gas analysis and control module includes a human-machine collaborative feature construction submodule, a human-machine collaborative feature embedding submodule, and an infrared light control parameter prediction module. The human-machine collaborative feature construction submodule is connected to the human-machine collaborative feature embedding submodule, the human-machine collaborative feature embedding submodule is connected to the infrared light control parameter prediction module, and the infrared light control parameter prediction module is connected to the environmental parameter acquisition module and the laser (01).

3. The infrared light-based industrial gas concentration analysis system according to claim 2, characterized in that: In the human-machine collaborative feature construction submodule, the method of obtaining the characteristics of the measured gas is as follows: first, the label is initialized according to the physical properties of the three gases CO, N2O, and H2O and the external knowledge graph. After the laser (01) enters the operation stage, the control response behavior label is extracted according to the historical wavelength shift performance and control response of the three gases under different environmental conditions, and the control response behavior label is weighted by time decay according to the occurrence time to obtain the weighted control response behavior label, which is used as the characteristic label of the measured gas. The external knowledge graph is a structured target detection gas physical response knowledge graph, including the semantic relationship between the target detection gas type, environmental state, laser (01) control method and control behavior result. The construction form is a triple, which is used to provide an initial label set based on external experience and experimental data. The external knowledge graph covers the typical absorption bands, interference sensitivity, response laws under environmental disturbances and recommended control strategies of CO, N2O, and H2O gases. In the initialization stage, the external knowledge graph is called to assign a label vocabulary set to each gas, which is used as the characteristic label of the measured gas.

4. The infrared light-based industrial gas concentration analysis system according to claim 3, characterized in that: In the human-machine collaborative feature construction submodule, the method of obtaining the state characteristics of the heating furnace is as follows: the label list is initialized using the structural parameters and control mode of the heating furnace to construct the initialization label. During the operation of the heating furnace, the current operating behavior state of the heating furnace is identified according to the dynamic changes of the temperature, pressure and airflow monitoring data in the furnace, and a dynamic behavior label is constructed. Different weights are assigned to the operating behavior labels in different time periods according to the preset time attenuation function to form an operating behavior label set that reflects the current operating trend of the heating furnace, which is used as the heating furnace state feature label.

5. The infrared light-based industrial gas concentration analysis system according to claim 4, characterized in that: In the human-machine collaborative feature construction submodule, the method of obtaining the laser control behavior feature is as follows: the control instruction log recorded during the operation of the laser (01) is used as a human-machine collaborative behavior text set, each log includes the control action, adjustment parameters, and feedback response words of the laser (01), and by performing word segmentation and TF-TDF weight calculation on all words in the behavior text, the weighted key behavior labels with TF-TDF weights within a set threshold range are extracted to obtain a key behavior weighted label list for each log, which is used as the laser control behavior feature label.

6. The infrared light-based industrial gas concentration analysis system according to claim 5, characterized in that: In the human-machine collaborative feature embedding submodule, all labels are embedded using the pre-trained word vector model and then weighted summed to obtain the WE feature of the label group. ; The TransE model is used to quantify each node of the external knowledge graph, and the objective function is: ； Where: is the embedding objective function value, , are respectively the positive sample set in the external knowledge graph and the randomly sampled negative sample set. , , are the head entity, relation, and tail entity in the triple respectively. , They are used to generate the head entity and tail entity that are randomly replaced by negative samples, , , are the KGE representations of the head entity, relationship, and tail entity, respectively. , are the KGE representations of the head entity and tail entity of the negative sample, respectively. is the interval coefficient, where The measured gas feature labels, heating furnace state feature labels and laser control behavior feature labels are input into the TransE model and then weighted summed to obtain the KGE feature of the label group ; The WE features and KGE features of the label group are concatenated to obtain the human-machine collaborative embedding features: ； in, To embed features for human-machine collaboration, , Dimensions for embedding features for human-machine collaboration.

7. The infrared light-based industrial gas concentration analysis system according to claim 6, characterized in that: The infrared light control parameter prediction module includes a feature recall submodule and an infrared light adjustment amount sorting submodule.

8. The infrared light-based industrial gas concentration analysis system according to claim 7, characterized in that: The feature recall submodule uses the current human-machine collaborative embedded features and the normalized environment state vector to jointly construct the current disturbance scene vector, calculates the cosine similarity between the current disturbance scene vector and the historical disturbance scene vector in the training set, selects the disturbance scene whose cosine similarity is within a set threshold range as the recall result, and extracts the corresponding laser (01) output infrared light control response behavior as a candidate adjustment strategy set.

9. The infrared light-based industrial gas concentration analysis system according to claim 8, characterized in that: The infrared light adjustment amount sorting submodule uses a regression model to score each group of historical adjustment strategies in the recall candidate adjustment strategy set. The input features include the current disturbance scene vector and the recall strategy features. The predicted score of each group of adjustment strategies in the current scene is output and arranged in descending order according to the predicted score. The one with the highest score is selected as the target adjustment parameter of the current laser (01).

10. The infrared light-based industrial gas concentration analysis system according to claim 9, characterized in that: The infrared light control parameter prediction module also includes an initial parameter analysis submodule. During the first operation phase of the laser (01), the initial parameter analysis submodule constructs an initial input feature vector based on the normalized environment state vector, the measured gas feature label and the embedded features of the heating furnace state label, and calls a pre-trained initial regression model to predict the infrared light target control parameters of the laser (01).

Images