Method and computer program product for automatically detecting defects while borescoping an engine

Abstract

In the method, a video borescope (2) is introduced into an engine (80) in such a way that when an engine shaft (83, 86) is rotated, the engine blades (90) of an engine stage (81, 84, 85, 87) secured thereto are successively moved through the image region of the video borescope (2), wherein the engine blades (90) currently situated in the image region of the video borescope (2) can be identified. The method is characterized by the following steps: - identifying possible defects (11) on an engine blade (90) by means of an image recognition process based on a plurality of successive individual images (10) of the video borescope (2) such that a plurality of images are present for a possible defect (11); - determining a characteristic variable for each of the individual possible defects (11) via the individual images (10) of the video borescope (2); - sorting the individual possible defects (11) according to the characteristic variables thereof; and - providing a list of the individual possible defects (11) in a sorted sequence.

Description

[0001] 16.03.2026 / BR

[0002] Method and computer program product for automated defect detection during engine boroscopy

[0003]

[0001] The invention relates to a method and computer program product for supporting defect detection during engine boroscopy.

[0004]

[0002] Engines, especially jet engines of aircraft, must be inspected regularly to verify compliance with technical safety requirements and to detect any damage at an early stage. Particularly for inspections while the jet engine is mounted on the aircraft (on-wing), side panels on the jet engine are opened and / or individual components are removed so that an inspector can look directly into the interior of the jet engine or with the aid of a borescope to inspect the engine blades. The inspector's task is to examine the engine blades directly at the eyepiece of the borescope or – in the case of a video borescope – on the connected video monitor, and to reliably identify defects in the engine blades, such as notches or dents in the tenths of a millimeter range.

[0005]

[0003] It is known to inspect the engine blades of a single rotatable engine stage of the jet engine by rotating the engine stage in such a way as to pass a viewing opening of the jet engine or a borescope inserted into the jet engine, so that all engine blades of the engine stage of the jet engine can be inspected successively in the viewing opening or the field of view of the borescope.

[0006]

[0004] Inspecting all engine stages of a jet engine takes several hours; the inspection of the nine stages of the high-pressure compressor of a CFM56 engine alone requires four hours. The result is an evaluation by the inspector that reflects the inspector's visual impression, but regularly does not meet the requirements of a standardized inspection report.

[0007]

[0005] In particular, the classification of the individual defects with regard to their significance and the repair work that may result from them is carried out by the inspector and is therefore subjective.

[0008]

[0006] It is an object of the present invention to provide a method by which the inspection of the engine blades of rotatable engine stages of a jet engine can be improved.

[0009]

[0007] This problem is solved by a method according to claim 1 and a computer program product according to claim 11. Advantageous further developments are the subject of the dependent claims.

[0010]

[0008] Accordingly, the invention relates to methods for assisted defect detection during engine boroscopy, in which a video boroscope is inserted into an engine such that when an engine shaft is rotated, the attached engine blades of an engine stage are moved successively through the image area of ​​the video boroscope, wherein the engine blades currently arranged in the image area of ​​the video boroscope are identifiable, comprising the steps:

[0011] - Identifying possible defects on an engine blade by image recognition based on several consecutive individual images of the video borescope, so that multiple images are available for a possible defect;

[0012] - Determination of the average size of the individual possible defects via the individual images of the video boroscope;

[0013] - Sorting the individual possible defects according to their average size; and

[0014] - Provide a list of each possible defect in descending order.

[0015]

[0009] The invention further relates to a computer program product comprising program parts which, when loaded into a computer, are designed to carry out the method according to the invention.

[0016]

[0010] First, some terms used in connection with the invention are explained.

[0017]

[0011] A “video boroscope” is a boroscope that provides a continuous analog or digital video image of the boroscope's field of view for further processing. The image captured by optics at the free end of the boroscope can be transmitted via an optical line to a camera at the other end of the boroscope and converted there into the desired video image. Alternatively, the image captured directly at the free end of the boroscope can be converted into a video image by a suitable chip (e.g., a CCD or CMOS chip), which is then transmitted via a data line to the other end of the boroscope and is available there for further processing.

[0012] Although, for the sake of clarity, it is assumed below that the image area of ​​the video borescope is generally large enough to capture the engine blades along their entire blade length, the method according to the invention is also suitable for inspecting individual longitudinal sections of the engine blades. By sufficiently repeating the method for the individual longitudinal sections, a complete inspection of the engine blades of an engine stage can then be carried out along their entire blade length.

[0018]

[0013] The “video image” is a continuous sequence of “individual images” in which the time interval between the recording of the individual images is usually constant, which is why the frame rate can also be used as a measure for the number of individual images per given time period.

[0019]

[0014] “Two consecutive still images” refers to two still images that occur consecutively in time. These do not necessarily have to be two immediately consecutive images of the video image. It is also possible that the video image includes intermediate images between the two consecutive still images, which, however, are not taken into account, at least in the determination of the movement of the engine blades in the image area of ​​the video borescope according to the invention.

[0020]

[0015] The invention provides a method by which the demanding but also monotonous task of individually inspecting the multitude of turbine blades of an aircraft engine can be supported, at least partially, in an automated manner. For this (partial) automation of engine inspection to be considered at all, it is necessary not only to achieve high reliability in defect detection, but also to sort the detected potential defects in a way that is comprehensible and meaningful to the inspector, so that the (additional) effort required by the inspector when applying the method according to the invention is minimized. This also includes ensuring that the actual process of inspecting all engine stages of a jet engine by the inspector is affected as little as possible.

[0021]

[0016] The method according to the invention is based on the engine boroscopy with video borescope known for engine inspection. The actual process of engine boroscopy remains unchanged, so that an inspector can carry out the engine inspection as usual. For this purpose, a video borescope is inserted into an aircraft engine through a suitable opening and aligned so that at least one turbine blade of the engine stage to be examined is located within the field of view of the video borescope. Subsequently, by rotating the engine shaft to which the turbine blades of the engine stage to be examined are attached, the individual turbine blades of the engine stage can be moved one after the other through the field of view of the video borescope.Based on a direct reproduction of the image area captured by the video borescope on a suitable display unit, the inspector can directly examine the engine blades for defects.

[0022]

[0017] In parallel, the video image or its individual frames are processed using the method according to the invention.

[0023]

[0018] In a first step, potential defects on an engine blade are identified by image recognition based on several consecutive individual frames of the video borescope. By identifying defects based on several consecutive individual frames of the video borescope—in particular, by ensuring that a potential defect is only identified as such if it is visible in several consecutive individual frames—false results—e.g., due to light reflections, shadows, dirt, dust, etc., which are only visible in individual frames—can be avoided. A similar result can also be achieved, in particular, by applying the method according to European Patent EP 4 022 290 Bl.

[0024]

[0019] Because the identification of a possible defect is based on several consecutive individual images, multiple images of each possible defect are available. Based on these images, a characteristic value for the respective defect is then determined. The characteristic value should preferably provide an indication of the significance of the respective defect, i.e., for example, allow a statement as to whether a defect is potentially critical or does not initially require any maintenance measures.

[0025]

[0020] The characteristic size preferably comprises the average size of the potential defect. The average size can be determined, for example, via the average size of a minimal bounding box around a potential defect in successive individual images and / or via the average area of ​​a potential defect in successive individual images. When determining the average size of the potential defect, the change in the position of the defect in the different individual images can be taken into account: If a defect on a turbine blade moves away from the video borescope due to the rotation of the turbine shaft, the defect will generally appear smaller in the image area in later acquired individual images than in earlier acquired individual images. The same applies if a defect moves towards the video borescope due to the rotation of the turbine shaft.Such a movement of a possible defect and the resulting change in the size of its representation in the individual frames of the video image can be taken into account when determining the average size.

[0026]

[0021] The characteristic value can also include an assessment of the type of potential defect. If a potential defect is not only recognized as such, but its type is also determined, this can be incorporated into the characteristic value. For example, if a potential defect is identified as an abrasion on the lower edge (tip rub) or as a dirt deposit on the turbine blades (deposit), this can be incorporated into the characteristic value in such a way that corresponding potential defects are ranked higher or lower in the subsequent sorting process than would be the case based solely on other factors influencing the characteristic value, such as the average size of the defect.

[0027]

[0022] Furthermore, the location of a potential defect on a turbine blade, in particular the distance of the potential defect from the root of the turbine blade, can also be included in the characteristic size. Thus, a defect with a smaller average size near the root of the turbine blade may be more critical than a defect with a larger average size near the blade tip, which can be reflected accordingly in the characteristic size with regard to the subsequent sorting. For example,A notch with a depth of 0.13 mm in the area of ​​the transition from the root of the turbine blade to the actual turbine blade may be more critical due to the centrifugal forces occurring during engine operation than a notch with a depth of 1 mm in the area of ​​the blade tip, which can be represented in the characteristic sizes of the two defects in such a way that the root-adjacent defect is classified as more critical in the subsequent sorting of defects based on the characteristic size.

[0028]

[0023] The preceding process steps of identifying potential defects and determining the characteristic size are preferably performed, at least in part, by or at least supported by artificial intelligence. A suitably trained artificial intelligence can, in principle, recognize a potential defect in a single image and, if necessary, identify its type. When considering several consecutive images, a suitably trained artificial intelligence can also determine the average size of the potential defects and, based on a single image, identify what might be recognized as a potential defect as something else, such as shadows or dust, when considering several images together, thus eliminating false results. The artificial intelligence can also determine the characteristic size or at least some of the factors influencing this size, such as...Determine the average size and / or type of a possible defect.

[0029]

[0024] The identified potential defects are then sorted according to their characteristic size. The sorting is carried out such that the defects are sorted in descending order of their respective criticality, regardless of whether a high or low value of the characteristic size indicates high criticality. If a potential defect is identified as an actually relevant defect and not, for example, as dirt, the sorting can be based on the average size; the potential defects identified as dirt are excluded from the list or sorted at the end.

[0025] The list of potential defects thus prepared is then made available.The inspector, who has carried out his inspection of the engine according to the prior art in parallel with the method according to the invention, can compare his findings with the list of possible defects generated by the method according to the invention and, if necessary, re-examine individual possible defects – in particular those whose position in the provided sorted list does not correspond to his own classification. The method according to the invention thus offers the inspector an additional means of checking his own findings.

[0030]

[0026] It is preferred if the list is shortened to include only those possible defects whose characteristic size exceeds or falls below a predetermined threshold. In other words, the provided list should only include those possible defects that cannot already be classified as rather non-critical based on their determined characteristic size.

[0031]

[0027] In particular, if the method is carried out in parallel with an inspection by an inspector, the inspector incurs practically no additional effort and, in particular, practically no relevant loss of time. The additional control option created more than compensates for the described comparison of one's own findings with the list generated by the method according to the invention.

[0032]

[0028] As an alternative to the parallel execution, it is also possible to subject the video image of the video borescope, which is usually recorded during the inspection by an inspector, to the method according to the invention after the actual inspection. Since the method can, for example, be implemented as a computer program product and thus run completely automatically, the actual additional effort is also comparatively manageable here.

[0033]

[0029] It is preferred if the identification of possible defects on an engine blade includes the identification of the engine blade itself. If the engine blade on which a possible defect is located is suitably identified, any necessary repair measures can be specifically directed to the engine blade in question.

[0034]

[0030] Particularly in the area of ​​the high-pressure turbine of an engine, it is rare for all engine blades to have unique identifying features that would allow for direct identification. It is therefore preferred if the identification of the engine blades is carried out starting from an initial position of the engine shaft, taking into account the rotational movement. The initial position is preferably determined by image recognition of previously known damage and / or markings on a component attached to the engine shaft. The component can, in particular, also be an engine blade. The selection of an initial position implies that at least one specific initial engine blade is located within the field of view of the video borescope.Starting from the initial turbine blade, the relative positions of the other turbine blades can be unambiguously determined by appropriately considering the movement detected by the video borescope. For example, it is possible to "number" the individual turbine blades starting from the initial turbine blade.

[0035]

[0031] It is preferred if the list of possible defects provided by the method includes the identification of the respective turbine blade and / or a pictorial representation of the possible defects. By specifying the identification of the turbine blade and / or by providing a pictorial representation of the possible defects, the inspector can quickly review the possible defects in the order determined by the method according to the invention and compare them with his own findings.

[0036]

[0032] It is preferred if the rotation of the engine shaft during the process is carried out at a rotational speed of no more than two revolutions per minute (720° / min). It has been found that at such a maximum rotational speed, a sufficiently good recording quality for the process can regularly be ensured with conventional video borescopes.

[0037]

[0033] For an explanation of the computer program product according to the invention, reference is made to the preceding statements.

[0038]

[0034] The invention will now be described by way of example with reference to an advantageous embodiment and the accompanying drawings. These show:

[0039] Figure 1: an arrangement designed to carry out the method according to the invention;

[0040] Figure 2: a series of successive still images of the video image taken from the video boroscope of the arrangement according to Figure 1;

[0041]

[0035] Figure 1 shows an exemplary arrangement 1 designed for carrying out the method according to the invention during its use for borescope inspection of an aircraft engine 80, which is only indicated in the figure.

[0036] The aircraft engine 80 is a turbofan engine with a low-pressure compressor stage 81 comprising the fan 82, which is connected to a low-pressure turbine 84 via a first shaft 83, and a high-pressure compressor stage 85, which is connected to a high-pressure turbine stage 87 via a second shaft 86. The combustion chamber arranged between the high-pressure compressor stage 85 and the high-pressure turbine stage 87 is not shown in Figure 1 for the sake of clarity.

[0042]

[0037] For the inspection of the engine blades 90 of the individual stages 81, 84, 85, and 87, the engine cover 88 is open (and therefore shown with dashed lines) so that access to the core engine 89 is free. The core engine 89 has various borescope openings 91 into which a video borescope 2 can be selectively inserted. Figure 1 shows two borescope openings 91 as examples, into each of which a borescope 2 can be inserted to inspect the engine blades 90 of a specific turbine stage—in the illustrated example, either the high-pressure compressor stage 85 or the high-pressure turbine stage 87—through the borescope 2. By rotating the shaft 83, 86 to which the engine blades 90 to be inspected are attached - in the example, the second shaft 86 - all engine blades 90 of the engine stage located in the image area of ​​the borescope 2 can be inspected one after the other.A direction of rotation is specified for this purpose, indicated by arrow 95. The actual rotation of the second shaft 86 is effected by a drive unit (not shown) integrated into the aircraft engine 80.

[0043]

[0038] The video borescope 2 is connected to a computer unit 3, which is designed for processing the video image from the video borescope 2 as described below and, in particular, also incorporates the artificial intelligence described below. Furthermore, the computer unit 3 includes an output 4 for connection to the control unit of the aircraft engine 80 in order to control the drive unit for the second shaft 86. It has been found that, as a rule, the rotational speed for the second shaft 86 should be a maximum of two revolutions per minute (720° / min). This ensures that a sufficiently good recording quality for the procedure can generally be guaranteed with conventional video borescopes. A terminal 5 is also connected to the computer unit 3, via which the video image or individual frames thereof can be inspected by an inspector.

[0044]

[0039] The inspector can carry out the inspection of the aircraft engine 80 in the usual way via Terminal 5.

[0045]

[0040] Figure 2 shows, by way of example, three temporally successive individual frames 10 of the video image of a single stage of the high-pressure turbine stage 87, recorded by the video-boroscope 2 according to Figure 1, from (a) to (c). Between the individual frames shown are intermediate frames, which are not necessary for explaining the method according to the invention and have therefore been omitted from the illustration in Figure 2 for the sake of clarity.

[0046]

[0041] Due to a rotation of the second shaft 86 in the direction of rotation 95 shown in Figure 1, initiated by the computer unit 3 via the output 4, the engine blades 90 of the engine stage 87 move through the field of view of the video borescope 2. As a result, the engine blade 90 shown moves from left to right in the individual frames 10 of Figure 2.

[0047]

[0042] For each of the individual images 10 according to Figure 2, image recognition is performed to detect possible defects. For this purpose, a suitably trained artificial intelligence is used, which is specifically trained to detect notches and dents in the engine blades 90. In this way, four possible defects 11, 11' are detected in the individual image 10 of Figure 2a, and three possible defects 11, 11' are detected in the individual images 10 of Figures 2b and 2c. Since the one possible defect 11' is only found in the individual image 10 according to Figure 2a, although the engine blade 90 with the supposed defect 11' is also depicted in the other individual images 10 (cf. Figures 2b and 2c), the possible defect 11' is considered an image artifact and ignored in the further course of the process.

[0048]

[0043] For each of the remaining possible defects 11, an average size is determined. For this purpose, in each of the individual images 10, a minimal bounding rectangle 12 is determined for each possible defect 11 using the appropriately trained artificial intelligence. The areas of the rectangles 12 for the respective possible defects 11 are averaged over the individual images 10, and these average sizes are taken as the respective average sizes of the individual defects 11.

[0049]

[0044] Furthermore, the artificial intelligence is trained to recognize the type of defect. Thus, the artificial intelligence can, among other things, distinguish between damage or dents on the admission ticket and the engine blade surfaces 90, as well as non-critical soiling or scratches.

[0050]

[0045] The average size of the defects 11 and their type are combined to form a characteristic value for each defect 11. For example, the average size of each defect 11 can be multiplied by a criticality factor defined for the respective type, whereby critical defects such as dents at the leading edge have a significantly higher criticality factor than, for example, contamination, whose criticality factor can be close to zero.

[0051]

[0046] The defects 11 are then sorted in descending order according to their characteristic size.

[0052]

[0047] For each possible defect 11, the identification of the turbine blade 90 on which it is located is also determined. For this purpose, an initial position of the turbine shaft 86 is determined based on a marking 20 visible in one of the individual images 10, from which, among other things, the turbine blade 90 located immediately adjacent to the marking 20 can then be clearly identified. The other turbine blades 90 can then be identified starting from this one clearly identifiable turbine blade 90 by observing the movement of the turbine shaft 86 in the video image of the video borescope 2 relative to the clearly identifiable turbine blade 90. For example, the turbine blades 90 of a turbine stage can be numbered starting from the clearly identifiable turbine blade 90.

[0053]

[0048] The list of possible defects 11, sorted according to their characteristic size, is provided including the respective identification of the engine blade 90 on which the possible defect 11 is located, a section of an image from one of the individual images 10 on which the respective defect 11 is shown, and, if applicable, further information such as the type of the respective defect 11 determined by the artificial intelligence.

[0054]

[0049] The inspector can compare the list provided in this way with his own findings and thus verify them.

Claims

Patent claims 1. Method for assisted defect detection in engine boroscopy, in which a video boroscope (2) is inserted into an engine (80) such that when an engine shaft (83, 86) is rotated, the attached engine blades (90) of an engine stage (81, 84, 85, 87) are moved successively through the image area of ​​the video boroscope (2), whereby the engine blades (90) currently arranged in the image area of ​​the video boroscope (2) are identifiable. characterized by the following steps: - Identifying possible defects ( 11 ) on an engine blade ( 90 ) by image recognition based on several successive individual images ( 10 ) of the video borescope ( 2 ), so that for a possible defect ( 11 ) several images are available; - Determination of a characteristic size for each of the individual possible defects ( 11 ) via the individual images ( 10 ) of the video boroscope ( 2 ); - Sorting the individual possible defects ( 11 ) according to their characteristic size; and - Provide a list of each possible defect ( 11 ) in sorted order .

2. Method according to claim 1 , characterized by the fact that the characteristic size comprises the mean size of the individual possible defects ( 11 ) across the individual images ( 10 ).

3. Method according to claim 2, characterized by the fact that the mean size comprises the average size of a minimal surrounding rectangle ( 12 ) around a possible defect ( 11 ) on the successive frames ( 10 ) and / or the average area of ​​a possible defect ( 11 ) on successive frames ( 10 ).

4. Method according to one of the preceding claims, characterized in that the characteristic size includes an assessment of the type of possible defect ( 11 ).

5. Method according to one of the preceding claims, characterized in that the list is reduced to possible defects ( 11 ) whose characteristic size exceeds or falls below a specified threshold.

6. Method according to one of the preceding claims, characterized in that the identification of possible defects ( 11 ) on an engine blade ( 90) includes the identification of the engine blade ( 90) itself .

7. Method according to claim 2, characterized by the fact that The identification of the engine blades (90) is carried out starting from an initial position of the engine shaft (83, 86) taking into account the rotational movement, wherein the initial position is preferably determined by image recognition of a mark on a component connected to the engine shaft (83, 86).

8. Method according to one of the preceding claims, characterized in that the provided list of possible defects ( 11 ) includes the identification of the respective engine blade ( 90 ) and / or a pictorial representation of the possible defects ( 11 ).

9. Method according to one of the preceding claims, characterized in that The rotation of the engine shaft ( 83, 86) takes place at a rotational speed of a maximum of two revolutions per minute.

10. Method according to one of the preceding claims, characterized in that at least some of the steps of the process are supported and / or carried out by artificial intelligence.

11. Computer program product comprising program parts which, when loaded into a computer, are designed to carry out the method according to any one of claims 1 to 9.

Images