Method for determining flue gas behavior in a radiation section of a combustion plant

Abstract

A method for determining the flue gas behavior in a radiation section (1) of a combustion plant is described, wherein the flue gas contains particles (8) that move at the speed of the flue gas in the direction of flow of the flue gas. The method comprises the following steps: a) Taking images of a plane (7) in the radiation train (1) for a predetermined exposure time, wherein the particles (8) moving in the flue gas stream are depicted as a line (12) in the respective taken image (11a, 11b); b) Determining the flue gas behavior from the length (14) of the linear particle images (12) which depends on the predetermined exposure time and the flow behavior of the flue gas and / or the angle (15) of the linear particle images (12) which depends on the flow behavior of the flue gas.

Description

[0001] The present invention relates to a method for determining the flue gas behavior in a radiation section of a combustion plant, wherein the flue gas contains particles that move at the velocity of the flue gas in the direction of flow. Furthermore, a device for determining the flue gas behavior in a radiation section of a combustion plant is described. The present invention also relates to a computer program comprising commands that cause the device according to the invention to execute the method according to the invention. State of the art

[0002] In the thermal treatment of waste in waste incineration plants, the composition of the fuel varies greatly depending on the nature of the waste. This could lead to uneven and therefore incomplete combustion processes and consequently high pollutant concentrations.

[0003] To reduce high pollutant concentrations, a combustion process that is as complete as possible must be ensured. Therefore, the boiler of the plant must be operated at above-average temperatures and heavily air-trimmed. This results in high energy consumption for the fans and the radiant flue. The downstream flue gas cleaning system also operates disproportionately. For example, in selective non-catalytic reduction (SNCR), a disproportionately large amount of ammonia is injected into the radiant flue for denitrification.

[0004] During the combustion process, the harsh conditions within the boiler and radiant flue make it difficult to collect data on the uniformity of the combustion process at the boiler bottom. Consequently, important input parameters for combustion control are missing, particularly for secondary air trim and SNCR control. One possibility is to obtain data on the flue gas behavior, as the flue gas is, at least partially, influenced by the combustion process.

[0005] Methods for determining the flow behavior of a fluid based on particles dispersed in the fluid are known in the prior art.

[0006] DE 20 2014 004 129 U1 describes a device for counting particles suspended in gases using laser-induced thermionic emission. Particles in the focus of a pulsed optical laser are excited to thermionic emission. The thermionic emission is detected by an optical detector. The particle number concentration can be derived from the ratio of the number of optical pulses to the number of detected thermionic emissions. In a preferred embodiment, the thermionic emission is recorded by a camera in such a way that the particle's thermionic emission is visible as a trace, from whose length and the known flow velocity of a gas sample the decay time of the thermionic emission can be determined.

[0007] EP 3 663 727 B1 relates to a flow measurement method for the optical measurement of flow parameters of a fluid flowing through a cross-section, wherein particles contained in the fluid are illuminated with a light source and imaged onto a sensor unit. The signals from the sensor unit are converted into a velocity profile in real time using spatial filtering technology. Flow parameters are continuously determined from the velocity profile in real time.

[0008] It has become apparent that state-of-the-art methods are difficult to adapt to the specific conditions in a radiation chamber of a combustion plant. The use of two precisely matched components requires accurate calibration, which is difficult to implement and maintain in a radiation chamber. Furthermore, these methods require significant equipment and complex algorithms for determining fluid flow behavior.

[0009] The object of the present invention is to provide a method for determining flue gas behavior in a radiation chamber of a combustion plant, with which the flue gas behavior can be determined with low equipment and computational effort and under the special conditions in a radiation chamber. Description of the invention

[0010] The problem is solved by the method with the features of claim 1, the device with the features of claim 14, and the computer program with the features of claim 16. Advantageous embodiments are described in the dependent claims.

[0011] It is proposed that the determination of flue gas behavior in a radiation section of a combustion plant, wherein the flue gas contains particles that move at the speed of the flue gas in the direction of flow of the flue gas, is carried out using the following steps: a) Taking images of a plane in the radiation train for a predetermined exposure time, wherein the particles moving in the flue gas stream are depicted as a line in the respective taken image; b) Determining the flue gas behavior from the length of the linear particle images, which depends on the predetermined exposure time and the flow behavior of the flue gas, and / or the angle of the linear particle images, which depends on the flow behavior of the flue gas.

[0012] "Flue gas behavior" refers to the behavior of the flue gas in a radiant flue. The flue gas is produced by the combustion of waste as fuel on the combustion grates at the bottom of the boiler. The flue gas is drawn through the radiant flue and, after extensive post-treatment, released into the atmosphere via the chimney. The flue gas behavior is influenced by factors including the combustion process, the extraction through the radiant flue, and thermal currents within the flue. The flue gas behavior can be uniform throughout the entire radiant flue. It is also possible for different flue gas behaviors to occur at different points within the radiant flue.

[0013] The term "particle" refers to a particle dispersed in the flue gas. The particles contained in the flue gas typically exhibit a wide particle size distribution, ranging from a few micrometers to several centimeters. The particles analyzed are preferably smaller than 1 cm. This allows for sufficiently sharp imaging of the particles.

[0014] It has been shown that the flue gas behavior in a combustion chamber of a combustion plant can be determined very easily and without contact using the method according to the invention. The properties of the particles in the combustion chamber are used to draw conclusions about the flue gas behavior and thus about the quality of the combustion process, without requiring complex equipment or computational effort. The particles moving in the flue gas stream can be imaged with high contrast to the surrounding flue gas stream. Due to this easy distinguishability from the surrounding flue gas, the particles can be imaged by a very simple image acquisition unit, in particular by a camera optic, such as a camera optic with a low frame rate of 10 fps (frames per second) to 50 fps, 20 fps to 40 fps, or in particular 25 fps to 35 fps.This method does not require two coordinated sensor components, as in the prior art; rather, one sensor component is sufficient. Accordingly, calibration of both sensor components is also unnecessary for implementing the method according to the invention.

[0015] It was discovered that the flue gas behavior can be determined from the length and / or angle of the linear particle images and the predetermined exposure time. This requires only a very simple algorithm with minimal computational effort.

[0016] The exposure time can be, for example, 1 / 500 s to 1 / 2 s, 1 / 20 s to 1 / 120 s, or 1 / 30 s to 1 / 60 s. This allows for sufficiently long imaging of individual particle trajectories that remain distinguishable from the trajectories of other particles.

[0017] The flue gas behavior can be precisely determined from the length and / or angle of the linear particle image. For this to be possible, the distance between the recording location and the particle moving in the flue gas stream, which is imaged as a linear particle image, must be known, preferably predetermined, or at least estimable. If the linear particle image is created by recording images of a plane within the radiation path, it can be ensured that only those particles moving within the flue gas stream at a predetermined distance from the recording location are recorded as linear particle images. A predetermined distance is known by setting up the camera and, if necessary, calibrating it.

[0018] The imaging of a plane in the radiation path can have a depth of field of less than 0.3 m, particularly in the range of 0.2 m to 0.1 m. For this purpose, a wide aperture can be used in a camera lens to capture images with a shallow depth of field. This ensures that only one plane is in focus, and only those particles moving within that plane in the flue gas stream are imaged as line-like particle images.

[0019] To avoid overexposure or excessively long exposure, a neutral density filter can be used in front of the camera, preferably with a strength of between 3 and 10 f-stops.

[0020] The distance between the plane imaged by an image acquisition unit with a camera lens in the radiation path and the camera lens itself can be constant. Maintaining a constant distance between the image acquisition unit and the imaged plane in the radiation path makes determining the flue gas behavior particularly easy, as the image of the plane only needs to be set up once, and subsequent calibration is either unnecessary or only required to a minimal extent.

[0021] The particles can be solids, especially incandescent solids. Solids are particularly easy to image as linear particle patterns. Due to their high optical density, solids are easily distinguishable from the surrounding flue gas in the image. The composition of the solid is irrelevant. The solid can be unburned fuel, partially burned fuel, and / or ash, such as fly ash. The solid can be generated during the combustion process of the fuel, particularly in the area of ​​the bottom of the combustion plant's boiler.

[0022] It is advantageous if the particle is a glowing solid. A glowing solid is particularly well suited for imaging as a linear particle because it emits light and thus stands out strongly from the surrounding flue gas. The glowing solid can be any solid produced during the combustion process of the fuel that has a high temperature, such as an incandescent temperature, which causes it to glow. The glowing solid can be glowing fuel and / or glowing ash, such as glowing fly ash.

[0023] The procedure may further include the step of determining the flow velocity, flow direction and / or frequency of combustion events from the linear particle images.

[0024] "Combustion events" are defined as events in which the combustion process in the boiler becomes particularly uneven, limited in time and space. Combustion events can occur when the boiler is loaded with highly combustible or poorly combustible fuel. A combustion event can also occur if highly combustible or poorly combustible fuel ignites during the combustion process. During a combustion event, the combustion process can be particularly intense or particularly weak.In other words, the combustion process can be particularly intense when the boiler is loaded with highly combustible fuel or when the highly combustible fuel ignites, and the combustion process can be particularly inefficient when the boiler is loaded with poorly combustible fuel or when the poorly combustible fuel ignites. A combustion process proceeds uniformly when no or at least very few combustion events occur. Those skilled in the art are aware of the detrimental effect such combustion events have on the uniformity of the combustion process. Furthermore, those skilled in the art are able to counteract the occurrence of combustion events, for example, by controlling the air distribution and selectively feeding the boiler with fuel.

[0025] The flow velocity, flow direction, and / or frequency of combustion events can be determined from the linear particle images. These images contain information about the flow velocity, flow direction, and / or frequency of combustion events. From this information, particularly the length and angle of the depicted linear path of a particle, as well as the statistical number of lines in a spatial region over time, the flue gas flow velocity, the flue gas flow angle, and / or the frequency of combustion events can be determined very easily. This can be implemented automatically using a very simple algorithm with minimal computational effort.

[0026] The method can further include the step of determining the flue gas flow velocity from the predetermined exposure time and the respective length of the recorded linear particle images. The computer-aided, automated determination of the pixel count in the digitally recorded linear particle image can be performed using conventional digital image processing methods. By recording an image of a plane in the radiation path, the ratio between the pixel count in the recorded linear particle image and the actual path length of the particle moving in the flue gas stream is known. This allows the actual path length of the particle in the flue gas stream to be determined. Using the predetermined exposure time, the velocity of the particle moving in the flue gas stream can be easily determined from the particle's actual path length.Since the particle is contained in the flue gas and moves at the speed of the flue gas, the flow velocity of the flue gas can be determined from the velocity of the particle. Due to this simple determination of the flue gas flow velocity from the number of pixels in the recorded linear particle image, the method according to the invention can be automatically implemented using a computer with a very simple algorithm and minimal computational effort.

[0027] The method can further include the step of determining the flow direction of the flue gas from the respective angle of the recorded linear particle images. A person skilled in the art is able to determine the angle of the recorded linear particle images relative to the horizontal. Since the angle of the linear particle image correlates with the movement of the particles in the flow direction of the flue gas, the flow direction of the flue gas can be easily determined from the angle of the linear particle image. This can be implemented automatically using a computer with a very simple algorithm and minimal computational effort.

[0028] The procedure can further include the step of detecting an overexposed image if it exhibits a brightness exceeding a predefined threshold. The brightness of an image can be easily determined using established digital image processing methods, particularly if a brightness threshold is appropriately set for the measurement environment. Overexposed images are easily identified by exceeding this threshold. This can be automatically implemented using a very simple algorithm with minimal computational effort. Images that are not detected as overexposed are considered normally exposed.

[0029] Overexposed images used to determine flue gas behavior can be disregarded when evaluating linear particle images for this purpose. It has been recognized that flue gas behavior cannot be determined, or at least not reliably, from overexposed images, as these images contain no usable information about the flue gas behavior. It is advantageous to disregard overexposed images for determining flue gas behavior because, firstly, they would distort the determination, and secondly, fewer images need to be evaluated, thus reducing the computational effort required for automated computer-aided analysis.

[0030] The procedure may further include the step of determining the ratio of the number of overexposed exposures to the number of normally exposed exposures or to the total number of exposures in a shooting interval. This ratio can be determined by dividing the number of overexposed exposures, the number of normally exposed exposures, or the total number of exposures in a shooting interval.

[0031] The method can further include the step of determining the frequency of combustion events in the combustion plant, particularly flames, from the ratio determined by the number of overexposed photographs. It has been recognized that the ratio of overexposed photographs to the total number of photographs in a given time period contains information about the frequency of combustion events, particularly flames, and that the frequency of combustion events, particularly flames, can be easily determined from this ratio. The advantage of this is that no additional photographs need to be taken, since the photographs already identified as overexposed, the normally exposed photographs, or the total number of photographs are used to determine the frequency of combustion events, particularly flames.During a flame burst, the combustion process is particularly intense, limited in both space and time. This results in a particularly uneven combustion process. A flame burst can occur when the boiler is loaded with highly flammable fuel or when highly flammable fuel ignites, especially with a burst of flame. Determining the frequency of combustion events, particularly flame bursts, can be automated using a very simple algorithm with minimal computational effort.

[0032] The procedure can further include the step of acquiring images from multiple planes along the flue gas stream in the combustion chamber and determining the flue gas behavior at more than one position within the flue gas stream. The flue gas behavior can be differentiated at various positions within the flue gas stream. Thus, a first flue gas behavior may prevail at one position, and a second flue gas behavior may prevail at a second position, which differs from the first. Determining the flue gas behavior at more than one position in the flue gas stream is advantageous because it provides a comprehensive overall picture of the flue gas behavior in the combustion chamber. This allows the uniformity of the combustion process across the entire boiler surface to be estimated and precisely controlled.

[0033] The method can further include the step of capturing the images in color and converting them to grayscale, with the flue gas behavior then being determined based on the converted grayscale images. It is advantageous to capture the images in color and convert them to grayscale, and to determine the flue gas behavior based on these grayscale images, as this further increases the distinguishability between the particle and the surrounding flue gas behavior in the grayscale images. This can be implemented automatically using a very simple algorithm with minimal computational effort.

[0034] The method can be implemented using a device for determining flue gas behavior in a radiation path of a combustion plant. The device comprises an image acquisition unit with a camera optic directed at a plane within the radiation path, and an image processing unit. It is proposed that the image acquisition unit be configured to capture images of a plane within the radiation path for a predetermined exposure time, whereby the particles moving in the flue gas stream are depicted as lines in each captured image. The image processing unit is configured to determine the flue gas behavior from the length and / or angle of the line-shaped particle images, which depend on the predetermined exposure time and the flow behavior of the flue gas.

[0035] The image evaluation unit can be configured to perform the steps of the procedure described above for determining the flue gas behavior.

[0036] The invention is explained in more detail below with reference to the accompanying drawings. These show: Fig. 1 - a schematic representation of a radiation train with a device for determining the flue gas behavior with two image acquisition units and an image evaluation unit; Fig. 2 - a schematic representation of an image acquisition plane with line-shaped particle images.

[0037] The in Fig. The radiation chamber 1 shown in Figure 1 has a front wall 3, a left side wall 4, a rear wall 5, and a right side wall 6. For illustrative purposes, the rear wall 5 and the right side wall 6 are shown as transparent. At the bottom of the radiation chamber 1 is the grate 2 on which the fuel is burned. This process produces particles 8 that move with the flue gas stream. The image acquisition units 9a and 9b are mounted on the right side wall 6 and focused on a plane 7 within the radiation chamber 1. The particles 8 move through this plane 7 with the flue gas stream. The image acquisition units 9a and 9b capture images of plane 11a and 11b. Based on the images of plane 11a and 11b, an image evaluation unit 10 determines the flue gas behavior within the radiation chamber 1.

[0038] For example, several image acquisition units 9a, 9b, ... 9n can be arranged in a horizontal row and focused on the common plane 7. Optionally, further image acquisition units can be arranged vertically spaced in the direction of flow in at least one further horizontal row. It is also conceivable to arrange image acquisition units opposite each other, with which the same plane 7 can be captured from both sides, thus mirroring the plane 7. Alternatively, it is conceivable that the opposing image acquisition units each capture images of spaced-apart (virtual) planes 7 of the image acquisition, these planes 7 preferably being aligned parallel to each other.

[0039] In the Fig. 2 is a schematic representation of a Fig. The image is visible on level 7. A photograph is taken on level 7 with sharp focus set to this area. Due to the long exposure time during the recording of the image... Fig. In plane 7, the particles 8 are imaged as linear particle images 12. The known image height 13 is converted into a pixel count. From this, the distance traveled 14 of the particle 8 can be determined. Using the known exposure time and the distance traveled 14 of the particle 8, the velocity of the particle 8 can be calculated, which correlates directly with the flow velocity of the flue gas. Furthermore, the angle 15 of the linear particle image 12 can be determined, which correlates directly with the flow direction of the flue gas.

[0040] An image acquisition unit 9a, 9, comprising a camera optic and a lens, is directed into the radiation path 1 of a waste incineration plant. A transparent protective screen may be located between the camera optic and the radiation path 1 to protect the camera optic from the heat and flue gases in the radiation path 1.

[0041] The camera lens is focused on a plane 7 located in front of the back wall of the radiation path 1. To minimize the depth of field of the images of this plane 7, the aperture of the camera lens is opened as wide as possible. To prevent overexposure of the images, an ND4 neutral density filter, for example, can be used. In this way, a plane 7 with a depth of field of 0.3 m can be captured. The plane 7 can be positioned approximately 4 m to 7 m from the camera lens, and the image diagonal of the images of plane 7 can be approximately 1 m.

[0042] The camera optics capture the image of plane 7 in radiation path 1, for example, with an exposure time of 10 s. Particles 8, which are produced by the combustion process at the bottom of the boiler and rise in the radiation path with the flue gas, move within plane 7. Due to the long exposure time, the camera optics image the particles 8 as line shapes at a specific angle. If the particles 8 are glowing, their brightness and luminous color allow them to be distinguished from the background with even greater contrast than non-glowing particles 8.

[0043] The determination of the flue gas behavior, including the flow velocity and flow direction of the flue gas, as well as the frequency of flame bursts, is carried out on the basis of the linear particle image 12 by an image evaluation unit 10. This can be done computer-aided with algorithms for the automatic image evaluation of the recorded images of an image sequence.

[0044] First, a brightness threshold is determined. An image of plane 7 whose brightness exceeds this threshold is considered overexposed and is therefore excluded from the determination of the flue gas behavior. A normally exposed image is used and processed to save computing power and reduce artifacts: The image is converted from color to grayscale and its edges are cropped. Finally, the image is sharpened.

[0045] For initial calibration, the image height is determined and converted into a pixel count. The pixel count of the linear particle image 12 is detected using an edge detection algorithm and converted into the distance traveled by particle 8 in the radiation path 1 during the known exposure time. Based on the distance traveled and the exposure time, the velocity of particle 8 in the flue gas and the directly correlated flow velocity of the flue gas can be determined. The angle of the linear particle image 12 is also determined. Thus, the flow direction of the flue gas is also known.

[0046] To reduce the statistical error, a variety of Fig. The respective (image acquisition) plane 7 is recorded over a time interval, and the flue gas behavior is determined from a multitude of linear particle images 12. In addition to the average flow velocity and the average flow direction of the flue gas, the scatter of the flow velocity and the flow direction can also be calculated and taken into account when determining the flue gas behavior.

[0047] The frequency of flame strikes is determined from the ratio of overexposed images to the total number of images.

[0048] The flue gas behavior parameters determined in this way are sent to the combustion power control and used for the control of the secondary air trim and for SNCR control.

[0049] In one embodiment, several, e.g., two, three, four or more, image acquisition units 9a, 9b, ... are directed into the radiation path of a waste incineration plant and focused on a first plane 7. The image acquisition units 9a, 9b, ... are arranged side by side in a row, such that the images captured by the image acquisition units 9a, 9b, ... Fig. , ... the levels 7 do not overlap.

[0050] On the opposite side of radiation path 1, a corresponding number of additional image acquisition units are positioned side by side and focused on a second plane. These image acquisition units are also arranged side by side in a row, so that the images of the planes captured by the image acquisition units do not overlap.

[0051] Linear particle images 12 are recorded from two planes 7, 16 at three different positions in the radiation path 1, and the flow velocity and flow direction of the flue gas, as well as the frequency of flame bursts, are determined at these positions.

[0052] In one embodiment, three image acquisition units 9a, 9b, ... are directed into the radiation path 1 of a waste incineration plant and focused on a first plane 7. The image acquisition units 9a, 9b, ... are arranged side by side in a row, so that the images of the planes captured by the image acquisition units 9a, 9b, ... do not overlap.

[0053] Furthermore, in the radiation path, further groups of image acquisition units can be focused on the same plane 7 in the direction of flow downwards towards boiler grate 2 or in the direction of flow upwards. These groups of image acquisition units are also arranged side by side in a row, so that the respective images of plane 7 do not overlap.

[0054] Line-shaped particle images 12 are recorded from a plane 7 at six different positions in the radiation path 1, and the flow velocity and flow direction of the flue gas, as well as the frequency of flame bursts, are determined at each of these positions.

[0055] The image acquisition units 9a, 9b, ..., 9a', 9b', ... can be configured to record a film containing a sequence of consecutive images, with the individual images of the film being evaluated. Only those images of the film sequence that are not too bright can be considered, whereby the ratio of the number of these used images to the total number of individual images in the film sequence within a given period, or the ratio of the number of overly bright images not considered to the total number of individual images in the film sequence, can be evaluated as a measure of flame flashes.

[0056] To reduce processing time, images preferably captured in color with image acquisition units 9a, 9b, ... can be converted into grayscale images. This process utilizes the higher information density initially achieved with color sensors, which is then reduced to grayscale information. Reference symbol list 1 radiation train 2 boiler grate 3 Front wall 4 left side wall 5 Back panel 6 right side wall Level 7 8 particles 9a, 9b Image acquisition unit 10 Image evaluation unit 11a, 11b Mapping of the plane 12 linear particle imaging 13 Image height 14 Pathway of the particle 15 angles of linear particle imaging QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] DE 20 2014 004 129 U1

[0006] EP 3 663 727 B1

[0007]

Claims

[1] Method for determining flue gas behavior in a radiation train (1) of a combustion plant, wherein the flue gas contains particles (8) which move at the speed of the flue gas in the direction of flow of the flue gas, characterized by the steps: a) Taking images of a plane (7) in the radiation train (1) for a predetermined exposure time, wherein the particles (8) moving in the flue gas stream are depicted as a line (12) in the respective taken image (11a, 11b); b) Determining the flue gas behavior from the length (14) of the linear particle images (12) which depends on the predetermined exposure time and the flow behavior of the flue gas and / or the angle (15) of the linear particle images (12) which depends on the flow behavior of the flue gas. [2] Method according to claim 1, characterized byDetermining the flue gas flow velocity, flue gas flow direction and / or the frequency of combustion events from the linear particle images (12) in the recorded image (11a, 11b). [3] Method according to claim 2, characterized by Determining the flow velocity of the flue gas from the predetermined exposure time and the respective length (14) of the recorded linear particle images (12). [4] Method according to claim 2 or 3, characterized by Determining the flow direction of the flue gas from the respective angle (15) of the recorded linear particle images (12). [5] Method according to any one of the preceding claims, characterized by Detecting an overexposed image when an image (11a, 11b) has a brightness that exceeds a specified threshold. [6] Method according to claim 5, characterized by, that overexposed images are disregarded when determining the flue gas behavior when evaluating line-shaped images (12) of particles (8) in the images for determining the flue gas behavior. [7] Method according to any one of the preceding claims, characterized by Determining the ratio of the number of overexposed shots to the number of normally exposed shots or to the total number of shots in a shooting interval. [8] Method according to any one of the preceding claims, characterized by Determining the frequency of combustion events in the combustion plant, especially of flame bursts, from the ratio determined from the number of overexposed photographs. [9] Method according to any one of the preceding claims, characterized by , that the particle (8) is a solid, in particular an incandescent solid. [10] Method according to any one of the preceding claims, characterized by, that the image (11a, 11b) of the plane (7) in the radiation path has a depth of field of less than 0.3 m, in particular in the range of 0.2 m to 0.1 m. [11] Method according to any one of the preceding claims, characterized by Recording images (11a, 11b) of several planes (7) along the flue gas stream in the radiation train (1) and determining the flue gas behavior at more than one position in the flue gas stream. [12] Method according to any one of the preceding claims, characterized by , that the distance between the plane (7) imaged by an image acquisition unit (9a, 9b) with a camera optic in the radiation path (1) and the camera optic is constant. [13] Method according to any one of the preceding claims, characterized by Taking the images (11a, 11b) in step a) with color images and converting the color images into grayscale images, with step b) being carried out using the converted grayscale images. [14] Device for determining flue gas behavior in a radiation path (1) of a combustion plant, comprising an image acquisition unit (9a, 9b) having a camera optics aligned to a plane (7) in the radiation path (1), and an image evaluation unit (10), characterized by , that the image acquisition unit (9a, 9b) is configured to capture images (11a, 11b) of a plane (7) in the radiation train (1) for a predetermined exposure time, wherein the particles (8) moving in the flue gas stream are depicted as a line in the respective captured image (11a, 11b), and that the image evaluation unit (10) is configured to determine the flue gas behavior from the length of the linear particle images (12) and / or angle (15) of the linear particle images (12), which depend on the predetermined exposure time and the flow behavior of the flue gas. [15] Device according to claim 14, characterized by, that the image evaluation unit (10) is designed to carry out the steps of the method according to one of claims 1 to 13 for determining the flue gas behavior. [16] Device according to claim 14 or 15, characterized by , that the image acquisition unit (9a, 9b) and the process steps for image evaluation take place in a common evaluation unit or at two locally separate locations. [17] Device according to any one of claims 14 to 16, characterized by that the evaluation unit is set up to transmit the evaluation results, in particular flow velocity, angle and number of images, via hard wiring or bus systems. [18] Computer program comprising instructions that cause the device according to one of claims 14 to 17 to perform the method steps according to one of claims 1 to 13.

Images