A flame temperature field and OH concentration field absorption tomography system and method

Through the combination of LED light source, fiber bundle and miniaturization spectrometer, the problem of flame temperature field and OH concentration field imaging is solved, low-cost and efficient absorption tomography is achieved, and measurement accuracy and speed are improved.

CN118817632BActive Publication Date: 2024-11-15BEIJING CHANGCHENG INST OF METROLOGY & MEASUREMENT AVIATION IND CORP OF CHINA
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
CN202411287711.1
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-14
Publication Date
2024-11-15
Estimated Expiration
2044-09-14

AI Technical Summary

Technical Problem

The prior art is difficult to achieve accurate imaging measurements of temperature fields and OH concentration fields in flames, especially in the problems of high equipment costs, large volumes, long time and lack of time resolution capabilities.

Method used

Using multiple LED light sources, fiber bundles, optical machine structures, multi-channel miniaturization spectrometers and neural networks, the ultraviolet spectrum is measured simultaneously by tomography optical path arrays, and the neural network is used to perform absorption tomography of flame temperature field and OH concentration field. Imaging.

Benefits of technology

The tomographic optical path array spectral measurement with relatively low equipment cost and volume is realized, which significantly improves the measurement accuracy and speed and has time resolution capabilities.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118817632B_ABST
    Figure CN118817632B_ABST
Patent Text Reader

Abstract

The present invention discloses a flame temperature field and OH concentration field absorption tomography system and method, the system includes multiple LED light sources, multiple optical fiber bundles, an optomechanical structure, multiple optical fiber arrays, multiple multi-channel miniaturized spectrometers and a host computer, multiple optical fiber bundles connect multiple ultraviolet LED light sources and the optomechanical structure, multiple optical fiber arrays connect the receiving end collimator array of the optomechanical structure and the multi-channel miniaturized spectrometer, each optical fiber in the optical fiber array connects an optical path and a spectrometer channel, the multi-channel miniaturized spectrometer receives optical signals from the multiple optical fiber arrays, measures experimental absorption spectra and transmits them to the host computer, and the host computer uses a neural network to perform flame temperature field and OH concentration field absorption tomography according to the experimental absorption spectra measured by the multi-channel miniaturized spectrometer. The present invention can realize the simultaneous measurement of the spectra of each optical path in the tomography optical path array with relatively low equipment cost and volume, and greatly improve the accuracy and speed of tomography calculation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of combustion diagnosis and absorption spectrum measurement, and in particular relates to a flame temperature field and OH concentration field absorption tomography system and method. Background Art

[0002] In recent years, with the increasing demand for clean and efficient use of energy and the development of power fields such as aerospace, it is crucial to develop advanced optical testing methods to achieve accurate measurement of physical parameters such as temperature and concentration in the combustion field. Flame temperature is one of the most important measurement parameters in combustion diagnosis. It reflects the energy exchange process between the flame and the outside world and also restricts the rate of the combustion reaction. OH radicals are important carriers of chain branching reactions in hydrogen-containing fuel flames and are important trace radicals that maintain combustion and indicate the combustion state. Realizing imaging measurement of temperature fields and OH radical concentration fields in flames is of great significance for studying and optimizing combustion states, improving power system performance and improving safety. It has become an important development direction of combustion diagnosis that is currently of concern to the academic and engineering communities.

[0003] In existing research, the measurement research involving temperature and OH concentration imaging in flames mainly uses the following technologies: (1) two-color planar laser induced fluorescence technology to obtain high spatial resolution imaging of OH concentration field and temperature field; (2) laser absorption tomography technology to obtain imaging of temperature field and concentration field; (3) ultraviolet broadband absorption spectroscopy to obtain temperature field and OH concentration field imaging under the assumption of axisymmetry and steady state. However, the above three types of measurement technologies all have a series of problems: planar laser induced fluorescence technology is difficult to measure the cross section perpendicular to the flow direction, it is difficult to measure the absolute concentration, and the temperature measurement accuracy is poor; infrared laser absorption tomography technology has weak absorption and is difficult to achieve concentration field imaging of low-concentration free radicals such as OH; ultraviolet absorption tomography system can only measure steady-state flames and has no time resolution capability. In addition, the equipment is bulky, expensive, and the imaging measurement takes a long time, so it is not possible to achieve absorption tomography measurement of temperature field and OH concentration field in flames. Summary of the invention

[0004] The purpose of the present invention is to provide a flame temperature field and OH concentration field absorption tomography system and method, which can achieve simultaneous measurement of the spectra of each light path in the tomographic light path array with relatively low equipment cost and volume, and greatly improve the accuracy and speed of tomographic calculation.

[0005] In order to achieve the above-mentioned object, one aspect of the present invention provides a flame temperature field and OH concentration field absorption tomography system, comprising a plurality of LED light sources, a plurality of optical fiber bundles, an optical-mechanical structure, a plurality of optical fiber arrays, a plurality of multi-channel miniaturized spectrometers and a host computer.

[0006] The multiple LED light sources are used to provide ultraviolet light sources with multiple light paths;

[0007] The plurality of optical fiber bundles are used to connect the plurality of LED light sources with the optomechanical structure, and transmit the optical signals of the plurality of LED light sources to the optomechanical structure;

[0008] The optical-mechanical structure is equipped with a plurality of transmitting end collimators and a receiving end collimator array, wherein the plurality of transmitting end collimators are connected to the plurality of optical fiber bundles to convert the light sources provided by the plurality of LED light sources into a tomographic optical path array, and the plurality of receiving end collimator arrays are used to receive the optical signals of the tomographic optical path array and couple the optical signals into the plurality of optical fiber arrays;

[0009] The plurality of optical fiber arrays are used to connect the receiving end collimator array of the optomechanical structure and the multi-channel miniaturized spectrometer, and each optical fiber in the optical fiber array is connected to an optical path and a spectrometer channel;

[0010] The multi-channel miniaturized spectrometer is used to receive optical signals from the multiple optical fiber arrays, measure experimental absorption spectra and transmit them to the host computer;

[0011] The host computer is used to perform absorption tomography of the flame temperature field and the OH concentration field using a neural network according to the experimental absorption spectrum measured by the multi-channel miniaturized spectrometer.

[0012] Preferably, the multiple optical fiber bundles are circular-to-linear optical fiber bundles, the circular end of each optical fiber bundle is connected to the LED light source, and the linear end is connected to the emission end collimator, so as to convert the light source provided by the LED light source into a fan-shaped light beam.

[0013] Preferably, the optomechanical structure is a regular dodecagon with five emitting edges and five receiving edges, forming a five-angle fan-shaped tomography optical path structure, the five emitting edges are equipped with emitting end collimators, and the five receiving edges are equipped with receiving end collimator arrays.

[0014] Preferably, the central wavelength of the multiple LED light sources is 310 nm, and each channel of the multi-channel miniaturized spectrometer adopts a high grating line number and a low slit width, which can realize the simultaneous measurement of multiple spectral lines in a wide wavelength range around 310 nm.

[0015] Another aspect of the present invention provides a flame temperature field and OH concentration field absorption tomography method, which uses the above system to perform flame temperature field and OH concentration field absorption tomography, comprising:

[0016] Step S1: Generate a large number of random multi-peak distributions as samples through simulation, and pre-train each optical path l Temperature and concentration A neural network tomography model for mapping temperature concentration field;

[0017] Step S2: Using the system to l Measured experimental absorption spectrum And the absorption coefficient theoretical model , the temperature is obtained by least squares fitting and concentration :

[0018]

[0019] Step S3: l Temperature and concentration Input the pre-trained neural network tomography model and output each spatial position Temperature and concentration , complete the tomographic imaging of temperature field and OH radical concentration field.

[0020] Preferably, the step S1 comprises:

[0021] Step S11: Generate two sets of random multimodal distributions, each containing n There are random distributions, each of which is formed by superposition of at least three Gaussian functions. The position and half-height full width of each Gaussian function are random within the reconstruction area, and the Gaussian functions are independent of each other. n The random distributions are also independent of each other, and the two sets of multimodal distributions are also independent of each other;

[0022] Step S12: Using a set of random multi-peak distributions as the temperature distribution , and perform overall random scaling on the values ​​so that the maximum value of each distribution is randomly selected within the required temperature range: , and let the minimum value of the distribution be randomly selected in the corresponding temperature range: , the random scaling of each distribution is independent of each other;

[0023] Step S13: Using another set of random multi-peak distribution as the concentration distribution , and perform overall random scaling on the values ​​so that the maximum value of each distribution is randomly selected within the required concentration range: , and let the minimum value of the distribution be randomly selected in the corresponding concentration range: , the random scaling of each distribution is independent of each other;

[0024] Step S14: Pair the two groups of distributions to obtain n Temperature distribution and concentration distribution , for each pair and , according to the absorption spectrum principle, the absorption spectrum on each light path is calculated by forward simulation:

[0025]

[0026] Then use the least squares fitting to calculate each light path l Temperature and concentration

[0027]

[0028] Then each pair and The corresponding and , forming the training samples of the neural network, the input is n right and , the output is n right and ;

[0029] Step S15: Training The neural network is created and saved.

[0030] The flame temperature field and OH concentration field absorption tomography system and method according to the above aspects of the present invention can achieve simultaneous measurement of the spectra of each light path in the tomographic light path array with relatively low equipment cost and volume, and greatly improve the accuracy and speed of tomographic calculations. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings used in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work:

[0032] Figure 1 It is a schematic diagram of the overall structure of a flame temperature field and OH concentration field absorption tomography system according to an embodiment of the present invention;

[0033] Figure 2 A schematic diagram of an optical-mechanical structure according to an embodiment of the present invention;

[0034] Figure 31 is a comparison diagram of reconstruction errors between a tomographic optical path array according to an embodiment of the present invention and a conventional parallel or fan-shaped optical path array, wherein: (a) multi-peak temperature and OH concentration distribution used in the simulation comparison; (b) schematic diagram of a conventional parallel or fan-shaped optical path array; (c) temperature reconstruction error comparison;

[0035] Figure 4 It is a schematic diagram of the principle of a fan-shaped tomography optical path array according to an embodiment of the present invention;

[0036] Figure 5 1 is a comparison diagram of a neural network tomography method according to an embodiment of the present invention and a traditional tomography method based on an ART algorithm, wherein: (a) statistical distribution of reconstruction error; and (b) statistical distribution of computational time. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. 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.

[0038] An embodiment of the present invention provides a flame temperature field and OH concentration field absorption tomography system, comprising a plurality of LED light sources, a plurality of optical fiber bundles, an optomechanical structure, an optical fiber array and a multi-channel miniaturized spectrometer.

[0039] exist Figure 1 In the illustrated embodiment, the flame temperature field and OH concentration field absorption tomography system of the embodiment of the present invention includes five LED light sources 1 for providing ultraviolet light sources, five circular-to-linear optical fiber bundles 2 connecting the LED light sources 1 and the optomechanical structure 3 and used to generate a fan-shaped optical path array, an optomechanical structure 3 maintaining the tomography optical path array and equipped with a collimator, five optical fiber arrays 4 for connecting the optomechanical structure 3 and a spectrometer 5 to transmit optical signals, five multi-channel miniaturized spectrometers 5 for detecting spectral signals on each optical path, and a host computer 6 connected to the spectrometer 5 to complete data storage and processing.

[0040] The multiple LED light sources use low-cost, compact LEDs as ultraviolet light sources with a wavelength of approximately 310 nm, and are used to provide a light source for ultraviolet absorption spectrum measurement. Each LED provides ultraviolet light in one direction, and has a smaller size and lower cost than the light source used in existing ultraviolet absorption tomography systems.

[0041] The five LED light sources 1 are used to provide 310nm ultraviolet light, and have the advantages of small size, low cost, high stability, etc. In this embodiment, the central wavelength is 310nm, the half-width at half maximum is 30nm, and the optical power is 3W.

[0042] The optical fiber bundle is used to connect the LED and the collimator at the emission end of the optomechanical structure, and convert the light source provided by the LED into a tomographic light path array. Figure 1 In the embodiment shown, there are 5 circular-to-linear optical fiber bundles 2, the circular end of each optical fiber bundle is connected to the LED light source 1, and the linear end is connected to the emitting end collimator, so as to convert the light source provided by the LED light source 1 into a fan-shaped light beam; in this embodiment, 10 light paths are arranged in each direction, and the optical fiber bundle is selected to be 10 ultraviolet quartz optical fibers with a core diameter of 200 μm.

[0043] The optical-mechanical structure 3 is used to maintain the tomographic optical path array, and is equipped with a transmitting end collimator for cooperating with the optical fiber bundle to generate the tomographic optical path array, and is equipped with a receiving end collimator array for receiving the optical signal and coupling it into the optical fiber array 4, and finally inputting it into the spectrometer to complete the spectrum acquisition. In this embodiment, the tomographic optical path array and the optical-mechanical structure arrangement are optimized by using the simulation method combined with the ART algorithm. Figure 2 The optimized tomographic optical path array and optomechanical structure layout are shown in the figure, which uses a regular dodecagon, but only arranges 5 emitting edges and 5 receiving edges to form a five-angle fan-shaped tomographic optical path structure. Figure 3 The reconstruction error of the tomographic optical path array of the embodiment of the present invention compared with the traditional parallel or fan-shaped optical path array is demonstrated. It can be seen that the reconstruction error of the traditional scheme decreases with the increase of the number of angles. When the number of projection angles is 2-5, the error is always above 4%. When the number of projection angles is 6-7, the error drops to 2.5%, but the equipment cost and system complexity also increase accordingly. Therefore, after optimal comparison, the layout scheme of this embodiment is selected, and an error level similar to that of 6-7 angles is achieved at 5 angles. It can be considered as the preferred solution under the trade-off between accuracy and equipment cost.

[0044] In this embodiment, each side of the optical-mechanical structure 3 is about 12 cm long, and the five sides are equipped with a transmitting end collimator, which is combined with the end face of the linear optical fiber bundle to form a fan-shaped optical path array. The basic principle is as follows Figure 4As shown: in the optical fiber end face 31, 10 optical fiber bundles are arranged in a line, the total length of the end face is about 2.2mm, the focal length of the collimating lens 32 is selected to be 20mm, and the distance between the optical fiber end face 31 and the collimating lens 32 is adjusted so that the optical fiber end face imaging 33 is located on the opposite beam collimator array. Then, an optical path is formed between each optical fiber end face in the optical fiber bundle and its imaging, forming a fan-shaped optical path array with an optical path length of about 90cm and an end face imaging size of about 9mm in diameter; ordinary ultraviolet quartz collimator arrays are arranged on the opposite 5 sides of the optomechanical structure 3. In this embodiment, the diameter is 1cm, the clear aperture is 6mm, and the number in each direction is 10. The angle is adjusted to face the collimator at the emitting end through assembly.

[0045] In this embodiment, each of the five optical fiber arrays 4 is composed of 10 ultraviolet quartz optical fibers, which connect the receiving end collimator array and the corresponding miniaturized spectrometer channel to complete the optical signal transmission.

[0046] The multi-channel miniaturized spectrometer 5 is used to receive optical signals from multiple optical fiber arrays 4 and perform spectral measurements to obtain experimental absorption spectra. It is composed of a series of miniaturized high-resolution spectrometers. Each spectrometer channel corresponds to an optical path. Each channel uses a high grating line number and a low slit width to ensure that each channel has a small volume and high spectral resolution. It can achieve high-resolution simultaneous measurement of multiple spectral lines in a wide wavelength range near 310nm, and obtain higher temperature measurement sensitivity and temperature field tomography accuracy; the multi-channel spectrometer is connected to the host computer to transmit the measured experimental absorption spectrum to the host computer. In one embodiment, a 10-channel spectrometer is used in each direction to measure 10 optical paths. Each spectrometer channel is built with a 3600 line / mm grating and a 10μm slit. The detectable wavelength range is 290nm~330nm, the spectral resolution is 0.06nm full width at half maximum, and the maximum sampling frame rate is 1kHz, which can achieve time-resolved spectral measurement.

[0047] The host computer 6 is used to store and process experimental data, and uses a neural network to perform absorption tomography of the flame temperature field and OH concentration field according to the experimental absorption spectrum measured by the multi-channel miniaturized spectrometer 5. In this embodiment, it is connected to the multi-channel miniaturized spectrometer 5 through a network port and a data switch to ensure the stability of high-speed data transmission.

[0048] An embodiment of the present invention further provides a flame temperature field and OH concentration field absorption tomography method, which uses the flame temperature field and OH concentration field absorption tomography system of the above embodiment of the present invention to perform flame temperature field and OH concentration field absorption tomography, including:

[0049] Step S1: Generate a large number of random multi-peak distributions as samples through simulation, and pre-train each optical path l Temperature and concentration A neural network tomography model for mapping temperature concentration field;

[0050] Step S2: Using the system to l Measured experimental absorption spectrum , and the absorption coefficient theoretical model , the temperature is obtained by least squares fitting and concentration :

[0051]

[0052] Step S3: l Temperature and concentration Input the pre-trained neural network tomography model and output each spatial position Temperature and concentration , that is, tomographic imaging of the temperature field and OH radical concentration field is completed.

[0053] The neural network training in step S1 includes:

[0054] Step S11: Generate two sets of random multimodal distributions as temperature distribution and OH concentration distribution, each set contains n There are random distributions, each of which is formed by superposition of at least three Gaussian functions. The position and half-height full width of each Gaussian function are random within the reconstruction area, and the Gaussian functions are independent of each other. n The random distributions are also independent of each other, and the two sets of multimodal distributions are also independent of each other;

[0055] Step S12: Using a set of random multi-peak distributions as the temperature distribution , and perform overall random scaling on the values ​​so that the maximum value of each distribution is randomly selected within the required temperature range: , and let the minimum value of the distribution be randomly selected in the corresponding temperature range: , the random scaling of each distribution is independent of each other;

[0056] Step S13: Using another set of random multi-peak distribution as the concentration distribution , and perform overall random scaling on the values ​​so that the maximum value of each distribution is randomly selected within the required concentration range: , and let the minimum value of the distribution be randomly selected in the corresponding concentration range: , the random scaling of each distribution is independent of each other;

[0057] Step S14: Pair the two groups of distributions to obtain nTemperature distribution and concentration distribution , for each pair and , according to the absorption spectrum principle, the absorption spectrum on each light path is calculated by forward simulation:

[0058]

[0059] Then use the least squares fitting to calculate each light path l Temperature and concentration

[0060]

[0061] Then each pair and The corresponding and , forming the training samples of the neural network, the input is n right and , the output is n right and .

[0062] Step S15: Training BP neural network and save it.

[0063] In one embodiment, in step S11, two groups of random multimodal distributions are generated, each group containing 200,000 random distributions;

[0064] In step S12, a set of random multi-peak distributions is used as the temperature distribution , let the maximum value of each distribution be randomly selected in the required interval: , and let the minimum value of the distribution be randomly selected in the corresponding interval: ;

[0065] In step S13, another set of random multi-peak distribution is used as the concentration distribution , let the maximum value of each distribution be randomly selected in the required interval: , and let the minimum value of the distribution be randomly selected in the corresponding interval: ;

[0066] In step S14, the two groups of distributions are paired to obtain 200,000 pairs of temperature distributions. and concentration distribution , get each pair and The corresponding and , forming the training samples of the neural network, with 200,000 pairs of input and , the output is 200,000 pairs and .

[0067] In step S15, training The BP neural network is trained and saved. In this embodiment, the grid uses a 5-layer BP neural network, each layer has 500 neural network nodes, and the training takes about 100 hours.

[0068] In this embodiment, Monte Carlo simulation verification is performed to compare the computational time and reconstruction accuracy of the trained neural network and the ultraviolet absorption spectrum tomography method based on the neural network with the application of the traditional tomography algorithm. A large number of random multi-peak distributions are simulated, and the number, position, height, and half-height width of the peaks are all random between 1 and 5. The reconstruction error and computational time of the traditional tomography algorithm and the neural network tomography method of the embodiment of the present invention are compared. Figure 5 As shown, compared with the traditional method, the reconstruction error level of the method of the present invention is improved from an average of 3.3% to an average of 2.4%, which significantly improves the accuracy of tomography, and the calculation time is increased from an average of 0.5s to an average of 0.88ms, which is nearly three orders of magnitude, greatly improving the speed of tomography. The simulation comparison of the embodiments of the present invention verifies the effectiveness of the measurement method of the present invention and its excellent performance in reconstruction accuracy and calculation speed.

[0069] In summary, the absorption tomography system of the flame temperature field and OH concentration field based on the ultraviolet absorption spectrum in the embodiment of the present invention, through array LED and multi-channel miniaturized spectrometer, can realize the simultaneous measurement of the spectra of each light path in the tomography light path array with relatively low equipment cost and volume. Compared with the existing measurement technology, it can realize the absorption tomography of the flame temperature field and OH concentration field with time resolution capability, providing new measurement technology support for combustion related research; the absorption tomography method of the flame temperature field and OH concentration field based on the ultraviolet absorption spectrum in the embodiment of the present invention, based on the absorption tomography system of the flame temperature field and OH concentration field in the embodiment of the present invention, uses a large number of random multi-peak distributions generated by simulation as samples, pre-trains the neural network for mapping the temperature concentration of each light path to the temperature concentration field, and then uses the neural network tomography algorithm to perform tomography, which can greatly improve the accuracy and speed of tomography calculation, and provides a new measurement method reference for absorption spectrum tomography technology.

[0070] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A flame temperature field and OH concentration field absorption tomography system, characterized in that: It includes multiple LED light sources, multiple fiber bundles, optomechanical structures, multiple fiber arrays, multiple multi-channel miniaturized spectrometers and a host computer. The multiple LED light sources are used to provide ultraviolet light sources with multiple light paths; The plurality of optical fiber bundles are used to connect the plurality of LED light sources with the optomechanical structure, and transmit the optical signals of the plurality of LED light sources to the optomechanical structure; The optical-mechanical structure is equipped with a plurality of transmitting end collimators and a receiving end collimator array, wherein the plurality of transmitting end collimators are connected to the plurality of optical fiber bundles to convert the light sources provided by the plurality of LED light sources into a tomographic optical path array, and the plurality of receiving end collimator arrays are used to receive the optical signals of the tomographic optical path array and couple the optical signals into the plurality of optical fiber arrays; The plurality of optical fiber arrays are used to connect the receiving end collimator array of the optomechanical structure and the multi-channel miniaturized spectrometer, and each optical fiber in the optical fiber array is connected to an optical path and a spectrometer channel; The multi-channel miniaturized spectrometer is used to receive optical signals from the multiple optical fiber arrays, measure experimental absorption spectra and transmit them to the host computer; The host computer is used to perform absorption tomography of the flame temperature field and the OH concentration field using a neural network according to the experimental absorption spectrum measured by the multi-channel miniaturized spectrometer.

2. The system according to claim 1, characterized in that The multiple optical fiber bundles are circular-to-linear optical fiber bundles, the circular end of each optical fiber bundle is connected to the LED light source, and the linear end is connected to the emission end collimator, so as to convert the light source provided by the LED light source into a fan-shaped light beam.

3. The system according to claim 1 or 2, characterized in that: The optical-mechanical structure is a regular dodecagon, having 5 emitting edges and 5 receiving edges, forming a five-angle fan-shaped tomography optical path structure, wherein the 5 emitting edges are equipped with emitting end collimators, and the 5 receiving edges are equipped with receiving end collimator arrays.

4. The system according to claim 1 or 2, characterized in that: The central wavelength of the multiple LED light sources is 310 nm. Each channel of the multi-channel miniaturized spectrometer adopts a high grating line number and a low slit width, which can realize the simultaneous measurement of multiple spectral lines in a wide wavelength range around 310 nm.

5. A flame temperature field and OH concentration field absorption tomography method, characterized in that: Using the system according to any one of claims 1 to 4 to perform flame temperature field and OH concentration field absorption tomography, comprising: Step S1: Generate a large number of random multi-peak distributions as samples through simulation, and pre-train each optical path l Temperature and concentration A neural network tomography model for mapping temperature concentration field; Step S2: Using the system to l Measured experimental absorption spectrum And the absorption coefficient theoretical model , the temperature is obtained by least squares fitting and concentration : Step S3: l Temperature and concentration Input the pre-trained neural network tomography model and output each spatial position Temperature and concentration , complete the tomographic imaging of temperature field and OH radical concentration field.

6. The method according to claim 5, characterized in that The step S1 comprises: Step S11: Generate two sets of random multimodal distributions, each containing n There are random distributions, each of which is formed by superposition of at least three Gaussian functions. The position and half-height full width of each Gaussian function are random within the reconstruction area, and the Gaussian functions are independent of each other. n The random distributions are also independent of each other, and the two sets of multimodal distributions are also independent of each other; Step S12: Using a set of random multi-peak distributions as the temperature distribution , and perform overall random scaling on the values ​​so that the maximum value of each distribution is randomly selected within the required temperature range: , and let the minimum value of the distribution be randomly selected in the corresponding temperature range: , the random scaling of each distribution is independent of each other; Step S13: Using another set of random multi-peak distribution as the concentration distribution , and perform overall random scaling on the values ​​so that the maximum value of each distribution is randomly selected within the required concentration range: , and let the minimum value of the distribution be randomly selected in the corresponding concentration range: , the random scaling of each distribution is independent of each other; Step S14: Pair the two groups of distributions to obtain n Temperature distribution and concentration distribution , for each pair and , according to the absorption spectrum principle, the absorption spectrum on each light path is calculated by forward simulation: Then use the least squares fitting to calculate each light path l Temperature and concentration Then each pair and The corresponding and , forming the training samples of the neural network, the input is n right and , the output is n right and ; Step S15: Training The neural network is created and saved.

Citation Information

Patent Citations

  • Quantitative measurement method and device for transient combustion field

    CN110823849A

  • Temperature field and concentration field measurement system and method based on double-optical-comb spectrum

    CN111077110A

  • Flame multi-parameter field cooperative measurement method based on active and passive optical tomography fusion detection

    CN111795746A

  • Wavelength modulation absorption spectrum chromatography reconstruction system based on deep learning

    CN114034653A

  • Combustion field multi-physical-quantity field chromatography measurement method and device based on multi-source light cooperation

    CN117825328A