A method for determining the fuel-air ratio of methane combustion based on image processing technology

Abstract

The present invention relates to a methane combustion fuel-air ratio determination method based on image processing technology. According to the color of flame chemiluminescence, which is mainly generated by CH* and C2* ion luminescence, the blue-green chemiluminescence flame located in the hue hue of 180°~252° in the flame combustion area is extracted, and then the brightness of the blue and green channels of the pixel points and the sensitivity of CH* and C2* and the camera hue are respectively listed according to the camera's sensitivity to chromaticity. The brightness ratio of CH* and C2* presented in the camera is obtained by solving the equation and the corresponding relationship between the ratio and the fuel-air ratio is established, which is used to determine the fuel-air ratio of the flame to be tested. Compared with the prior art, the present invention can realize instantaneous automatic monitoring of the flame state in the burner through image processing technology, and provides convenient and effective technical means for the flame distribution in the burner, the performance test of the burner, and the burner design.

Description

Technical Field

[0001] The invention belongs to the field of flame detection, and in particular relates to a method for determining the fuel-air ratio of methane combustion based on image processing technology. Background Art

[0002] Effective use of fuel and reduction of emissions are the main goals of combustion detection research, and optical visualization of flames plays an important role in gaining a deeper understanding of complex combustion phenomena. Taking engine internal combustion as an example, by recording images such as ignition status, gas dynamic characteristics, and flame oscillation, information such as flame color, temperature, edge detection, and regional division can be analyzed. Combined with actual combustion knowledge, the combustion status in the engine can be diagnosed without dismantling it, and the damage and material defects that occur can also be detected, identified, and quantitatively evaluated. Combustion diagnosis technology based on high-speed cameras is not only highly instantaneous, such as the ignition moment, extinguishing moment, collision moment, and disturbance moment can be accurately captured, but also the image information is intuitive and rich, and the visibility is high, which provides the necessary technical means for improving the scientific research of transient flames, and at the same time plays an extremely important role in the design, development, and performance evaluation of burners.

[0003] There are three main types of flame combustion forms: premixed, partially premixed, and diffused. The specific situation depends on the degree of mixing of the fuel and the oxidant and the mixing level before the reaction. Among them, the diffusion flame is bright yellow and is the main flame form that is usually studied by patents, such as fire protection (CN104504382), explosion flame (CN103617635), and furnace flame (CN103914709). The premixed flame is faintly blue, easily disturbed and difficult to capture, and is often ignored as noise or background in flame image processing. With the increasingly stringent requirements for carbon reduction and emission reduction, more and more low-carbon and high-efficiency combustion energy sources are being developed and applied, such as liquid natural gas, which has replaced oil and is widely used in buses and heavy trucks. Methane, the main component of natural gas, appears faintly blue when it is fully and efficiently burned after premixing. However, the existing combustion diagnosis method is not sufficient to support the effective detection and analysis of the flame combustion state in this state. Therefore, how to effectively capture the flame in this state and how to analyze the combustion state in combination with the luminescence principle of the premixed flame have become key issues in the detection of natural gas premixed combustion flames. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method for determining the fuel-air ratio of methane combustion based on image processing technology.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A method for determining the fuel-air ratio of methane combustion based on image processing technology comprises the following steps:

[0007] Step 1: Calibrate the color sensitivity of the experimental camera to obtain the sensitivity spectrum conversion rate of light of different wavelengths in different channels of the RGB space;

[0008] Step 2: Obtain flame images of methane combustion at different fuel-air ratios in a dark room;

[0009] Step 3: Convert the flame image into a grayscale image, and use the grayscale image to obtain the flame area map;

[0010] Step 4: Multiply the flame image and the flame area map and convert from RGB color space to HSV color space to obtain the HSV image. According to the combustion luminescence hue interval of CH* and C2*, extract the pixel points whose H channel pixel values ​​are in the corresponding range in the HSV image, and set the area where they are located as the flame effective area;

[0011] Step 5: The flame image and the flame effective area are superimposed and the intersection is taken. The obtained RGB image is the ion combustion luminescence image of CH* and C2*. The wavelengths of the light emitted when CH* and C2* burn are 430nm and 516nm respectively. The image model of the ion combustion luminescence image is expressed as:

[0012]

[0013] Among them, I R ,I G ,I B They represent the R channel, G channel, and B channel of the ion combustion luminescence image respectively. They represent the theoretical brightness values ​​of the R, G, and B channels of light with a wavelength of 430nm, respectively. Respectively represent the theoretical brightness values ​​of the R channel, G channel, and B channel for light with a wavelength of 516nm; the theoretical brightness values ​​of the R channel, G channel, and B channel for light with a wavelength of λ The calculation formula is:

[0014]

[0015] Among them, I λ The value represents the brightness value of light with wavelength λ. Respectively represent the sensitivity spectrum conversion rate of light with wavelength λ of R channel, G channel and B channel;

[0016] Step 6: Substitute the light with wavelengths of 430nm and 516nm and their corresponding ions into formula (2) to obtain:

[0017]

[0018] in, and Respectively represent the brightness values ​​of CH* and C2* when burning;

[0019] Combining formula (1) with formula (3), we get

[0020]

[0021] Solving formula (4) we can get

[0022]

[0023] Step 7: Calculate from formula (5) The non-zero elements in the ratio matrix correspond to the flame effective area. The flame effective area is divided into an outer oxidation layer and an inner layer. The means of the ratio matrix elements corresponding to the outer oxidation layer, the inner layer and the entire flame effective area are calculated respectively, and corresponded to the corresponding fuel-air ratio to construct a linear relationship.

[0024] Step 8: For the flame with the fuel-air ratio to be measured, obtain the mean values ​​of the ratio matrix elements corresponding to the outer oxidation layer, the inner layer and the entire flame effective area according to steps 2-7, and determine its fuel-air ratio according to the linear relationship established in step 7.

[0025] Furthermore, the specific steps of step 1 include:

[0026] S1.1. The full-band light emitted by the calibration light source is converged into parallel light by a lens, and then filtered into monochromatic light by a monochromator;

[0027] S1.2. In a dark room, project monochromatic light directly onto the photosensitive element of the experimental camera;

[0028] S1.3, adjust the wavelength of monochromatic light by using a monochromator, take multiple photos and record them at certain wavelength intervals within the wavelength range of 350-750nm, so as to obtain several monochromatic light RGB images at corresponding wavelengths;

[0029] S1.4, respectively extract the intensity value of each channel of several monochromatic light RGB images at corresponding wavelengths, calculate the average intensity and then draw a wavelength-sensitivity intensity graph to obtain a hue sensitivity response curve;

[0030] S1.5. Obtain the sensitivity spectrum conversion ratio of the G channel and the B channel at wavelengths of 430 nm and 516 nm through the hue sensitivity response curve.

[0031] Furthermore, the photosensitivity in S1.4 is the ratio of the average intensity to 255, and the sensitivity spectrum conversion rate of the G channel and the B channel at the wavelength of 430nm and 516nm in S1.5 is the photosensitivity at the corresponding wavelength.

[0032] Furthermore, the method for obtaining the flame region map in step 3 is: firstly, the pixel value of the pixel point with a pixel value of 255 in the grayscale image is set to 0, and then an appropriate threshold is taken to binarize the grayscale image, and the obtained binary image is the flame region map.

[0033] Furthermore, in step 4, the combustion luminescence hue range of CH* and C2* is 180° to 252°.

[0034] Furthermore, step 4 also includes performing noise reduction processing on the flame effective area. For a certain position point in the flame effective area, if three points in its eight neighborhoods are determined to be effective areas, the position point is retained as a effective area, otherwise it is deleted.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] (1) The use of high-speed cameras to capture instantaneous images of flames can fully record the transient state of flames, such as the instantaneous fuel-air ratio and degree of combustion completeness in the states of ignition, flameout, and disturbance, which is convenient for scientific research and application on transient flames;

[0037] (2) The present invention can realize instantaneous automatic monitoring of the flame state in the burner through image processing technology, providing a convenient and effective technical means for the flame distribution in the burner, the performance test of the burner and the burner design;

[0038] (3) The methods of image denoising, image segmentation, color extraction, etc. involved in the image processing process of the present invention lay a foundation for the study of transient flame combustion state. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 Schematic diagram of the experimental scenario of the embodiment;

[0040] Figure 2 : is a wavelength-photosensitive intensity diagram obtained in the embodiment;

[0041] Figure 3 This is a schematic diagram of the flame layers;

[0042] Figure 4 Different flame regions established in the embodiment Linear plot of mean value versus fuel-air ratio. DETAILED DESCRIPTION

[0043] The present invention will now be described in further detail with reference to the accompanying drawings.

[0044] The present invention mainly utilizes the research method combining the flame chemiluminescence principle and image processing technology, uses a high-speed camera to take transient flame photos, applies image processing technology to extract the effective area of ​​flame chemiluminescence, analyzes and calculates the CH* and C2* The ion luminescence ratio is used to analyze and determine the degree of current flame combustion, providing effective technical support for performance tests such as transient combustion state monitoring of the flame in the burner and combustion efficiency judgment.

[0045] The present invention provides a methane combustion fuel-air ratio determination method based on image processing technology, which mainly includes the following steps:

[0046] Step 1: Calibrate the color sensitivity of the high-speed camera Photron SA-4 used in the experiment to obtain the sensitivity spectrum conversion rate of light of different wavelengths in different channels of the RGB space, specifically:

[0047] A fiber optic illuminator (MI-150 150w) that can evenly emit equal-density, full-band light is used as the calibration light source. The light emitted by the calibration light source is focused by a convex lens and filtered by a monochromatic light instrument to emit monochromatic light.

[0048] Use the camera used in the experiment to capture and record monochromatic light with a wavelength range from 350nm to 750nm in a dark room. To eliminate interference from other light, the camera must capture monochromatic light in a completely dark room, and the camera should not be equipped with a lens, so that the light can be projected on the camera's photosensitive element to avoid filtering by the lens. The layout of the experimental scene is as follows: Figure 1 As shown;

[0049] The intensity values ​​of each channel of several monochromatic RGB images under the corresponding wavelength are extracted respectively, and the wavelength-sensitivity intensity graph is drawn after the average intensity is calculated to obtain the hue sensitivity response curve. The sensitivity intensity is the ratio of the average intensity to 255, such as Figure 2 As shown;

[0050] The sensitivity spectral light conversion rate of the camera CCD to the CH* and C2* luminescence bands (i.e. wavelengths of 430nm and 516nm) in the G and B layers is obtained through response curve calibration, which is the photosensitivity at the corresponding wavelength.

[0051] Step 2: Build a test bench and obtain flame images of methane combustion at different fuel-air ratios in a dark room. Premix methane and air at a certain volume. The premixing range is 0.8-1.2 fuel-air ratio with an accuracy interval of 0.05. The detailed fuel and air flow rates are shown in Table 1.

[0052] Step 3: Convert the flame image into a grayscale image, and use the grayscale image to obtain the flame area map. Overexposure of the camera will make some areas of the image appear bright, so first set the pixel value of the pixel point with a pixel value of 255 caused by overexposure in the grayscale image to 0, and then take an appropriate threshold to binarize the grayscale image, and the resulting binary image is the flame area map.

[0053] Step 4: Multiply the flame image and the flame area map and convert from RGB color space to HSV color space to obtain the HSV image. According to the combustion luminescence hue interval of CH* and C2*, extract the pixels with H channel pixel values ​​in the range of 180° to 252° in the HSV image, and set the area where they are located as the flame effective area. For a certain position point in the flame effective area, if 3 points in its eight neighborhoods are determined to be effective areas, then keep the position point as the effective area, otherwise delete it.

[0054] Table 1 Flow rates of methane and air at different fuel-air ratios

[0055] Fuel-air ratio Methane flow rate (L / min) Air flow (L / min) 0.80 0.3 3.581 0.85 0.3 3.371 0.90 0.3 3.183 0.95 0.3 3.016 1.00 0.3 2.865 1.05 0.3 2.729 1.10 0.3 2.605

[0056] Step 5: The flame image is superimposed with the flame effective area and the intersection is obtained. The resulting RGB image is the ion combustion luminescence image of CH* and C2*. The wavelengths of the light emitted when CH* and C2* are burned are 430nm and 516nm respectively. The image model of the ion combustion luminescence image can be expressed as shown in formula (1), where the theoretical brightness value of the light with wavelength λ in the R channel, G channel, and B channel is The calculation formula is as shown in formula (2).

[0057] Step 6: Substitute the light with wavelengths of 430nm and 516nm and their corresponding ions into formula (2) to obtain formula (3). Then combine formula (1) with formula (3) to obtain formula (4). The solution of formula (4) is

[0058]

[0059] Step 7: From the above formula, we can get The non-zero elements in the ratio matrix correspond to the flame effective area, which divides the flame effective area into the outer oxidation layer and the inner layer, such as Figure 3 As shown, the mean values ​​of the ratio matrix elements corresponding to the outer oxidation layer, the inner layer and the entire flame effective area are calculated respectively, and corresponded with the corresponding fuel-air ratio to construct a linear relationship, such as Figure 4 shown.

[0060] Step 8: For the flame with the fuel-air ratio to be measured, obtain the mean values ​​of the ratio matrix elements corresponding to the outer oxidation layer, the inner layer and the entire flame effective area according to steps 2-7, and determine its fuel-air ratio according to the linear relationship established in step 7.

[0061] The above are only preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should be regarded as the protection scope of the present invention.

Claims

1. A method for determining the fuel-air ratio of methane combustion based on image processing technology, characterized in that: The following steps are involved: Step 1: Calibrate the color sensitivity of the experimental camera to obtain the sensitivity spectrum conversion rate of light of different wavelengths in different channels of the RGB space; Step 2: Obtain flame images of methane combustion at different fuel-air ratios in a dark room; Step 3: Convert the flame image into a grayscale image, and use the grayscale image to obtain the flame area map; Step 4: Multiply the flame image and the flame area map and convert from RGB color space to HSV color space to obtain the HSV image. According to the combustion luminescence hue interval of CH* and C2*, extract the pixel points whose H channel pixel values ​​are in the corresponding range in the HSV image, and set the area where they are located as the flame effective area; Step 5: The flame image and the flame effective area are superimposed and the intersection is taken. The obtained RGB image is the ion combustion luminescence image of CH* and C2*. The wavelengths of the light emitted when CH* and C2* burn are 430nm and 516nm respectively. The image model of the ion combustion luminescence image is expressed as: Among them, I R ,I G ,I B They represent the R channel, G channel, and B channel of the ion combustion luminescence image respectively. They represent the theoretical brightness values ​​of the R, G, and B channels of light with a wavelength of 430nm, respectively. Respectively represent the theoretical brightness values ​​of the R channel, G channel, and B channel for light with a wavelength of 516nm; the theoretical brightness values ​​of the R channel, G channel, and B channel for light with a wavelength of λ The calculation formula is: Among them, I λ The value represents the brightness value of light with wavelength λ. Respectively represent the sensitivity spectrum conversion rate of light with wavelength λ of R channel, G channel and B channel; Step 6: Substitute the light with wavelengths of 430nm and 516nm and their corresponding ions into formula (2) to obtain: in, and Respectively represent the brightness values ​​of CH* and C2* when burning; Combining formula (1) with formula (3), we get Solving formula (4) we can get Step 7: Calculate from formula (5) The non-zero elements in the ratio matrix correspond to the flame effective area. The flame effective area is divided into an outer oxidation layer and an inner layer. The means of the ratio matrix elements corresponding to the outer oxidation layer, the inner layer and the entire flame effective area are calculated respectively, and corresponded to the corresponding fuel-air ratio to construct a linear relationship. Step 8: For the flame with the fuel-air ratio to be measured, obtain the mean values ​​of the ratio matrix elements corresponding to the outer oxidation layer, the inner layer and the entire flame effective area according to steps 2-7, and determine its fuel-air ratio according to the linear relationship established in step 7.

2. The method for determining the fuel-air ratio of methane combustion based on image processing technology according to claim 1, characterized in that: The specific steps of step 1 include: S1.

1. The full-band light emitted by the calibration light source is converged into parallel light by a lens, and then filtered into monochromatic light by a monochromator; S1.

2. In a dark room, project monochromatic light directly onto the photosensitive element of the experimental camera; S1.3, adjust the wavelength of monochromatic light by using a monochromator, take multiple photos and record them at certain wavelength intervals within the wavelength range of 350-750nm, so as to obtain several monochromatic light RGB images at corresponding wavelengths; S1.4, respectively extract the intensity value of each channel of several monochromatic light RGB images at corresponding wavelengths, calculate the average intensity and then draw a wavelength-sensitivity intensity graph to obtain a hue sensitivity response curve; S1.

5. Obtain the sensitivity spectrum conversion ratio of the G channel and the B channel at wavelengths of 430 nm and 516 nm through the hue sensitivity response curve.

3. The method for determining the fuel-air ratio of methane combustion based on image processing technology as claimed in claim 2, characterized in that: The photosensitivity in S1.4 is the ratio of the average intensity to 255, and the sensitivity spectrum conversion rate of the G channel and the B channel at wavelengths of 430nm and 516nm in S1.5 is the photosensitivity at the corresponding wavelength.

4. The method for determining the fuel-air ratio of methane combustion based on image processing technology according to claim 1, characterized in that: The method for obtaining the flame region map in step 3 is: firstly, the pixel value of the pixel point with a pixel value of 255 in the grayscale image is set to 0, and then an appropriate threshold is taken to binarize the grayscale image, and the obtained binary image is the flame region map.

5. The method for determining the fuel-air ratio of methane combustion based on image processing technology according to claim 1, characterized in that: In step 4, the combustion luminescence hue range of CH* and C2* is 180° to 252°.

6. The method for determining the fuel-air ratio of methane combustion based on image processing technology according to claim 1, characterized in that: The step 4 also includes performing noise reduction processing on the flame effective area. For a certain position point in the flame effective area, if three points in its eight neighborhoods are determined to be effective areas, the position point is retained as the effective area, otherwise it is deleted.

Images