METHOD FOR DETECTING TEST GAS ESCAPING FROM A TEST ITEM USING AN OPTICAL SENSOR

Description

[0001] The invention relates to a method for detecting test gas escaping from a test object.

[0002] The detection of test gas escaping from a test object is used to identify a leak in the test object. Stationary or particularly large test objects are not examined in a test chamber, but typically with the help of a hand-held sniffer probe. The sniffer probe is brought by the operator to the test point on the test object to be examined. The sniffer probe continuously sucks in air through an inlet opening of the sniffer probe. The sucked-in air is directed to a gas detector, which can selectively detect the leak gas. The leak gas, i.e. the gas escaping from a leak in the test object, is typically a known test gas with which the test object is filled or which is already present in the test object. If test gas escapes from a leak at the test point, the leak gas is sucked in together with the air in the environment of the test point, so that a gas mixture of air and test gas is directed to the detector.The concentration of the test gas in the drawn-in sniffer gas stream depends on both the leak rate and the size of the continuously drawn-in air flow. The lower the leak rate and the larger the drawn-in air flow, the lower the test gas concentration in the drawn-in gas stream.

[0003] It is also known to use thermographic infrared cameras to detect gas clouds containing an infrared-active gas, i.e., a gas whose absorption spectrum includes infrared wavelengths. The wavelength range reaching the camera's sensor field is restricted using an optical filter. The passband of the filter contains the absorption spectrum or an absorption band of the gas to be detected and blocks other wavelength ranges. If the camera is then pointed at a corresponding gas cloud, the radiation components transmitted through the gas cloud appear darker in the infrared absorption spectrum than the radiation reflected by the background. As a result, the gas cloud appears as a darkened area in the infrared camera image.

[0004] For example, the FLIR GF320 camera is known for subtracting the individual pixel amplitudes of successive images, thereby amplifying the representation of movements in a gas cloud.

[0005] US 2003 / 0025081 A describes a method for quantitatively displaying gas emissions using an infrared camera.

[0006] The imaging of gas cloud movements in infrared images is described in WO 2018 / 45107 A1, EP 3 392 635 A1 and EP 3 351 916 A1.

[0007] US 2019 / 0078966 A1 describes a method for estimating the position of a gas leak in an optical image.

[0008] US 5,523,569 A describes a device for automated gas leak detection using optical images.

[0009] US 10,031,040 B1 describes a system for gas leak analysis using machine learning.

[0010] The invention is based on the object of providing an improved automated method for detecting test gas escaping from a test object.

[0011] The method according to the invention is defined by the features of claim 1.

[0012] Accordingly, optical radiation reflected or emitted by the test object to be examined is first recorded using an optical sensor, for example a digital sensor or a CCD chip. The sensor is designed to detect at least one wavelength of the optical absorption spectrum of the test gas. Preferably, a test gas with an absorption spectrum in the infrared wavelength range is used. The sensor can be designed to detect at least one wavelength of the absorption spectrum using suitable optics, for example by using a suitable optical filter in the beam path between the test object and the sensor in order to filter out wavelengths outside the absorption spectrum. For example, the passband of such an optical filter can contain the absorption spectrum or an absorption band, while the stopband of the filter covers the ranges of the adjacent remaining wavelengths.

[0013] The idea behind this is that optical radiation that is reflected or emitted by the test object to be examined is recorded and evaluated in order to determine, based on the recorded radiation spectrum, whether the radiation is transmitted by test gas in order to conclude that test gas is present.

[0014] The optical radiation is recorded at a first point in time and then again at a second point in time following the first point in time. Two digital images are generated from the optical radiation recorded at both points in time, the pixels of which have signal amplitudes that correspond to the amplitude of at least one absorption wavelength of the test gas at the location in question. The pixel at a location where no test gas is present therefore has a greater signal amplitude than a pixel that corresponds to a location with test gas. At locations with test gas, radiation of the absorption wavelength is absorbed, i.e. the amplitude of the radiation transmitted through the test gas is lower than the amplitude of radiation that is not transmitted through the test gas.The pixels depicting a location with test gas therefore have a lower signal amplitude under homogeneous illumination than pixels of a location without test gas.

[0015] The special feature of the invention is that between the times the two images are taken, a possible test gas cloud of the test gas escaping from a leak is actively moved. In other words, this means that a gas blast, for example, is emitted to the location where a gas leak is present or suspected, and where a test gas cloud is therefore present or suspected, to blow away any possible cloud of test gas. This can be done as a compressed air pulse or with the help of a fan. It is important that the gas used to move the test gas cloud is different from the test gas and does not have the same absorption bands as the test gas.

[0016] In other words, the location where a gas leak is present or suspected and where a test gas cloud is present or suspected is the location where the recorded optical radiation is reflected or emitted or from which the image section is recorded.

[0017] According to the invention, the first image is compared with the second image of the reflected or emitted optical radiation, wherein the signal amplitude of the pixels of the first and second images corresponds to the amplitude of at least one absorption wavelength range of the test gas. One or more dynamically consecutive images are compared with one or more continuously acquired images.

[0018] Preferably, both the camera and the object are fixed so that the image section of the consecutively taken images is identical.

[0019] According to the invention, a leak is automatically considered detected if at least the difference between the signal amplitude of at least one first pixel of the first image and the signal amplitude of at least one second pixel of the second image exceeds a threshold value. While known methods for thermographic imaging of gas merely capture and display images of the gas, the method according to the invention automatically evaluates the pixel amplitudes to detect a leak. This allows a leak to be detected regardless of the user and the distance.

[0020] The method according to the invention performs an integral assessment of the tightness of a test specimen by calculating the sum of the differences between corresponding pixels of the first image and the second image. This means that the amplitude of a pixel x ij with i=1...n and j=1...m, where n, m are natural numbers, of the first image is subtracted from the amplitude of the pixel corresponding to the first pixel. x ij of the second image. To this difference, the difference in the amplitudes of another pixel, e.g., x i+1,j or xi, j+1 of both images, is added for multiple pixels.

[0021] This summation can, for example, be performed for all pixels within a selected area, for all pixels of the entire image, or for every nth pixel, where n is a natural number. If the sum exceeds a certain threshold, a leak is considered detected.

[0022] If there is no test gas cloud at the location where the recorded optical radiation is reflected or transmitted, emitting the gas burst at this location does not cause any movement of a test gas cloud, so that the amplitudes of the pixels in both images do not show any significant differences. The difference between the pixels is then below a suitable threshold value. However, as soon as a test gas cloud is present at that location, the gas burst causes this cloud to shift, so that the test gas cloud is depicted at a different position in the first image than in the second image. After subtracting the amplitudes of the pixels in the two images, significant amplitude values ​​above the threshold value are still present due to the shifted test gas cloud. This means that the difference between the pixel amplitudes can make it possible to detect a test gas cloud.

[0023] The amplitude component resulting from background radiation, background noise, or reflected radiation not reflected by the test gas is reduced by subtracting the respective pixels, while the amplitude components of the absorption spectra of those pixels corresponding to a location with test gas remain. As soon as the sum of these amplitudes exceeds a certain value, a leak can be considered automatically detected. According to the invention, an automatic comparison with the respective threshold value is thus performed. As soon as this threshold is exceeded, a signal can be generated and / or transmitted containing the information "a leak is present."

[0024] Alternatively or additionally, automated localization of a leak in the test object can also be carried out, namely by calculating the difference between the amplitude of at least a first pixel x ij of the first image and the amplitude of at least a second pixel of the first image that is different from the first. The difference is compared with a threshold value, whereby a leak is deemed to be present at the location of the first pixel if the difference exceeds a threshold value. In this case, the sums of the amplitudes of several pixels in a first area of ​​the first image can also be compared with the sums of the amplitudes of the pixels in a second area of ​​the first image that is different from the first. If the difference between the sums of the amplitudes of the pixels from the two areas exceeds a predetermined threshold value, a leak at the location of the first area is deemed to have been detected.In this case, too, a signal can be sent and / or generated automatically, containing the information that a leak has been detected or is considered to have been detected.

[0025] The test specimen is preferably irradiated with optical radiation whose spectrum includes the absorption spectrum of the test gas. If the absorption spectrum of the test gas includes absorption wavelengths in the infrared wavelength range, the test specimen is irradiated with infrared radiation.

[0026] When carrying out the method according to the invention, it may be advantageous if the test object and / or the location of the measurement on the test object is shielded from the external environment, e.g. by protective walls, in such a way that air movements of the external environment are kept away from the test object or from the measurement location.

[0027] To irradiate the test object, a radiation source can be used whose emission spectrum covers a large portion of the infrared thermal radiation, such as more than 50 nm, more than 100 nm, or several hundred nanometers, i.e., whose emission spectrum is a thermal broadband spectrum. Alternatively, the radiation source can be a narrowband radiation source whose emission spectrum covers only a small portion of the thermal radiation, ranging from a few nanometers to up to 50 nm, such as a laser or an LED.

[0028] An embodiment of the invention is explained in more detail below with reference to the figures. They show: Fig. 1 is a schematic representation of the embodiment and Fig. 2 is a schematic representation of the recorded images.

[0029] Fig. 1 shows a test specimen 12 in the form of a pipeline that transports a gas or fluid, e.g., a refrigerant, that contains a test gas or can itself be used as a test gas. Test gas 16 flows out through a leak 14 in the test specimen 12 and forms a cloud in the area of ​​the leak 14.

[0030] A radiation source 18 emits infrared radiation 20 toward the test object 12. The radiation 20 is reflected by the test object 12 and by the background of the test object. The reflected radiation 20 is recorded by a sensor 22, which may be the sensor of a thermal imaging camera, e.g., in the form of a CCD chip. An optical filter 24 is positioned in front of the sensor 22 in the beam path of the reflected thermal radiation 20.

[0031] In Fig. 2 a first image 30 and a second image 32 of the radiation 20 recorded by the sensor 22 are shown. Both images 30, 32 have the same number of pixels x ij , where i=1...n with n being a natural number and j=1...m with m being a natural number. Each of the two images 30, 32 therefore consists of n columns and m rows. When comparing the two images, the pixels in a first area 34 of the first image 30 can be compared with the pixels in an area 34 of the second image 32 corresponding to the first area 34. Alternatively or additionally, the pixels in the first area 34 of one of the images can also be compared with the pixels in a second area 36 different from the first area 34. This area can be used to localize a leak.

[0032] In particular, the comparison of the pixels x ij , x ij of the two images 30, 32 using the term ∑ i , j = 1 n , m x ij − x ij ¯ or using the term ∑ i , j = 1 n , m x ij − x ij ¯ . Here, x ij is a pixel of the first image at the location of column i and row j, while x ij is a pixel of the second image corresponding to the location of the first pixel, ie a pixel of the second image at the location of column i and row j.

[0033] If this term exceeds a certain threshold, a leak is considered detected. A signal can be generated and / or output indicating that a leak is present or has been detected.

[0034] Compared to the prior art, the method according to the invention offers the advantage of automated leak detection of a test object by capturing and evaluating digital images of the test object, without requiring a human observer to perform the evaluation. In particular, the method according to the invention, or at least the comparison of the captured images and pixels and the evaluation of the pixels, can be carried out under computer control or by a microprocessor.

Claims

1. A method for detecting a test gas cloud (16) escaping from a leak in a test specimen (12), comprising the following steps: receiving optical radiation (20) reflected or emitted from the test specimen (12) or its background at a first point in time with an optical sensor (22) configured to detect at least one wavelength or a wavelength range of the optical absorption spectrum of the test gas (16), creating a first digital image (30) from the optical radiation (20) received at the first point in time such that the signal amplitudes of the image points xij correspond to the amplitude of at least one absorption wavelength range of the test gas (16), receiving optical radiation (20) reflected or emitted from the test specimen (12) or its background at the second point in time by means of the optical sensor (22), creating a second digital image (32) from the optical radiation (20) received at the second point in time such that the signal amplitudes of the image points xij correspond to the amplitude of at least one absorption wavelength range of the test gas (16), comparing the first image (30) with at least one second digital image (32) of the reflected optical radiation, which is different from the first image, wherein a leak (14) is considered detected when at least the difference of the amplitude of at least a first image point xij of the first image (30) and the amplitude of at least a second image point xij of the second image (32) exceeds a threshold value, wherein i, j are each natural numbers xij is an image point at the location of column i and row j of the first image (30), and xij is an image point at the location of column i and row j of the second image (32) characterized in that for integral evaluation of the tightness of the test specimen (12), the sum is formed from the differences of the amplitudes of mutually corresponding image points xij of the first image (30) and xij of the second image (32), a leak (14) being regarded as detected if the sum exceeds the threshold value.

2. The method according to claim 1, characterized in that a gas shock is emitted after capturing an image at the first point in time and before capturing a second image at a second point in time different from the first point in time, in the direction of the location from which the image section is captured.

3. The method according to claim 1 or 2, characterized in that the optical sensor (22) and the test specimen (12) are spatially fixed such that the image sections of the captured images are almost identical.

4. The method according to any one of the preceding claims, characterized in that, for localizing a leak (14) in the test specimen (12), the difference is formed from the amplitude of at least a first image point xij of the first image (30) and the amplitude of at least a second image point of the first image (30), a leak (14) at the location of the first image point xij being regarded as detected if the difference exceeds a threshold value.

5. The method according to any one of the preceding claims, characterized in that the test specimen (12) is irradiated with optical radiation (20) whose spectrum includes at least a part of the absorption spectrum of the test gas (16).

6. The method according to any one of the preceding claims, characterized in that the optical radiation reflected by the test specimen (12) and to be received is filtered with an optical filter (24) whose passband includes at least one absorption wavelength of the optical absorption spectrum of the test gas (16).

7. The method according to any one of the preceding claims, characterized in that during the measurement the test specimen (12) is shielded from the environment in such a way that air movements in the environment are kept away from the test specimen (12).

8. The method according to any one of the preceding claims, characterized in that the test specimen (12) is irradiated with a radiation source (18) whose emission spectrum covers a large part of the infrared thermal radiation, such as a halogen lamp, an incandescent lamp, a radiant heater or a flash lamp.

9. The method according to any one of the preceding claims, characterized in that the test specimen (12) is irradiated with a narrow-band radiation source (18) whose emission spectrum covers only a small part of the thermal radiation, such as a laser or an LED.

10. The method according to any one of the preceding claims, characterized in that the image points xij of the two images (30, 32) are compared with each other using the term ∑ i , j = 1 n , m x ij − x ij ¯ wherein i, j, n, m are each natural numbers, xij is an image point at the location of column i and row j of the first image (30), and xij is an image point at the location of column i and row j of the second image (32).