Gap information extraction method and extraction system for metal cabin cover

Abstract

The present application belongs to the field of information extraction technology, and discloses a method and system for extracting gap information of a metal cabin cover. A method and system for extracting gap information of a metal cabin cover, comprising the following steps: Step 1: A plurality of first acoustic wave detectors are arranged on the surface of the metal cabin cover to detect the crack sound waves generated inside the metal cabin cover in real time to obtain sound wave information. Step 2: The information processing device determines the location of the crack according to each sound wave information, and obtains a set of initial crack positions. In the technical solution provided by the present application, only the first acoustic wave detectors of the same group receive sound wave signals within the same time period, and they will be judged as valid sound wave signals, so that many noise signals can be eliminated. In addition, the grating optical fiber is connected to the surface of the metal cabin cover, and then the sound wave signal in the metal cabin cover can be diffused into the grating optical fiber to capture the changes of smaller sound sources.

Description

Technical Field

[0001] The present application relates to the technical field of information extraction, and in particular to a method and system for extracting gap information of a metal cabin cover. Background Art

[0002] The metal cabin cover is used to protect the various devices in the cabin. When inspecting the cabin, the staff will observe with the naked eye whether there are cracks on the surface of the cabin and whether the anti-rust coating is damaged. For metal cabins with high requirements, sonic detection instruments will be used to detect the metal cabin and find the gaps inside the metal cover in time to avoid safety accidents in the air.

[0003] At present, the method for detecting gaps inside metal cabins generally uses ultrasonic detectors to detect metal. The ultrasonic detector transmits sound wave signals to the metal cabin, and then determines the distance between the reflecting surface and the ultrasonic detector based on the time and intensity of the returned sound wave signals. When there is a crack, the sound wave signal returns because it touches the boundary of the crack, so the sound wave detector can determine whether there is a crack based on the return time of the sound wave.

[0004] In this solution, the position of the ultrasonic detector needs to be constantly changed to complete the detection of the inside of the metal cabin. In order to increase the detection efficiency, an array of ultrasonic detectors is used to detect the metal cabin. In this way, the echo signal received by each ultrasonic detector is full of a large amount of redundant information.

[0005] At present, ultrasonic detectors all determine whether the ultrasonic wave they receive is the ultrasonic wave they send out based on the strength of the ultrasonic echo information. Because of the interference problem of ultrasonic signals generated by each detector, adopting this solution will result in the elimination of a lot of valid ultrasonic information. Summary of the invention

[0006] The content of this application is used to introduce concepts in a brief form, which will be described in detail in the detailed implementation section below. The content of this application is not intended to identify the key features or essential features of the technical solution claimed for protection, nor is it intended to limit the scope of the technical solution claimed for protection.

[0007] As a first aspect of the present application, in order to solve the technical problem that it is difficult to accurately extract ultrasonic information, the present application provides a method for extracting gap information of a metal nacelle cover, comprising the following steps:

[0008] Step 1: a plurality of first acoustic wave detectors are arranged on the surface of the metal engine cabin cover to detect in real time the acoustic wave signals generated by cracks inside the metal engine cabin cover to obtain acoustic wave information;

[0009] Step 2: The information processing device determines the location of the crack according to each acoustic wave information, and obtains a set of initial crack locations;

[0010] Step 3: When the number of cracks is greater than a preset threshold, the information processing device calculates the arrangement positions and arrangement angles of a plurality of second acoustic wave detectors according to the initial position set of the cracks, and obtains the detector position information;

[0011] When the number of cracks is not greater than the preset threshold:

[0012] The first acoustic wave detector continues to detect the acoustic wave signals generated by the cracks inside the metal cabin cover;

[0013] Step 4: Arrange a plurality of second acoustic wave detectors according to the detector position information, and each second acoustic wave detector releases an ultrasonic signal to the generating position;

[0014] Step 5: Receive the echo signals received by each second acoustic wave detector to obtain echo information, send the echo information and detector position information to the neural network model, filter out interference information in the echo information of each second acoustic wave detector, and obtain the echo signals of each second acoustic wave detector.

[0015] The technical solution provided by the present application aims to solve the problem that the echo information is easily disturbed when multiple acoustic wave detectors are used to form an array for gap positioning. First, multiple first acoustic wave detectors are used to determine the approximate location of the crack. Then, the second acoustic wave detectors are arranged in a targeted manner according to the position information, thereby reducing the number of second acoustic wave detectors and reducing the interference of echo information in the metal cabin cover. Finally, a neural network model is used to screen the echo information to obtain the effective echo information of each second acoustic wave detector, which effectively reduces the amount of effective data eliminated in the echo information. Furthermore, the echo information extracted in the present application is used to analyze the location and size of the crack, achieving higher accuracy.

[0016] Furthermore, step 1 includes the following steps:

[0017] Step 11: Obtain at least three fixed detection positions on the surface of the metal nacelle cover that are not located in the same plane;

[0018] Step 12: Fixing a first acoustic wave detector at each fixed detection position;

[0019] Step 13: Each first acoustic wave detector detects in real time the acoustic wave signal generated by the crack inside the metal cabin cover to obtain acoustic wave information.

[0020] In the technical solution provided in the present application, by selecting three fixed detection positions that are not located in the same plane on the surface of the metal cover, a first acoustic wave detector is used to detect the position of the sound source in the metal cabin cover, thereby obtaining the position information of the suspected cracks.

[0021] In step 1, the acquired sound wave signal is actually a weak vibration signal generated by a crack inside the metal cabin cover. In order to improve the detection capability of such weak signals, it is necessary to increase the detection accuracy of the first sound wave detector. However, while improving the detection accuracy, it will inevitably be affected by the noise signal in the circuit. To address this problem, the present application provides the following technical solutions:

[0022] There are n first acoustic wave detectors arranged, where n is an even number greater than zero, and the first acoustic wave detectors are divided into n / 2 detection groups, and the first acoustic wave detectors in each detection group are connected via two grating optical fibers;

[0023] Among them, in any detection group, the grating optical fiber has one and only one end connected to the surface of the metal cabin cover, and the ends of the two grating optical fibers connected to the metal cabin cover are different.

[0024] Further, step 13 includes the following steps:

[0025] Step 131: When the first acoustic wave detector k receives the acoustic wave signal k i When the first acoustic wave detector k sends a standard optical signal to the first acoustic wave detector k' through the grating optical fiber connected to the metal cabin cover at the end of the first acoustic wave detector k, until the acoustic wave information disappears, and the first acoustic wave detector k' detects the wavelength change information ω of the grating optical fiber. k,i ;

[0026] Wherein, k is the serial number of the first acoustic wave detector, k' is the serial number of the first acoustic wave detector in the same detection group as the kth first acoustic wave detector; k and k' are both integers; k i The acoustic wave signal i detected by the first acoustic wave detector k is the index of the acoustic wave signal, and i is an integer;

[0027] Step 132: When the first acoustic wave detector k' receives the acoustic wave signal k i When the first acoustic wave detector k' sends a standard optical signal to the first acoustic wave detector k' through the grating optical fiber connected to the metal cabin cover at the end of the first acoustic wave detector k', until the sound source information disappears, and the first acoustic wave detector k' detects the wavelength change information ω of the grating optical fiber. k,i `;

[0028] Step 133: If the first acoustic wave detector k and the first acoustic wave detector k' receive the acoustic wave signal k iIf the time interval of k is within the preset time interval, i ,ω k,i ,ω k,i `, k` i As the sound wave information k I , k I represents the Ith acoustic wave information received by the kth first acoustic wave detector, and I represents the index of the acoustic wave information received by the kth first acoustic wave detector.

[0029] In the technical solution provided by the present application, only when the first acoustic wave detectors of the same group receive the acoustic wave signal in the same time period, it is determined as a valid acoustic wave signal. This mechanism helps to eliminate a large number of noise signals. In addition, by connecting the grating optical fiber to the surface of the metal cabin cover, the acoustic wave signal in the metal cabin cover can be diffused into the grating optical fiber, thereby capturing the changes of smaller sound sources, further improving the sensitivity and accuracy of detection.

[0030] The gap in the metal cabin cover is usually very small. In order to accurately obtain the size and position of the gap, it is necessary to try to find the largest opening surface of the gap. In this way, when arranging the second acoustic wave detector, the sound wave can be emitted perpendicular to the largest opening surface, thereby increasing the detection accuracy. However, the current sound source localization technology can only locate the approximate position of the sound source, but cannot use the subtle sound generated by the crack to find a relatively large opening surface. To address this problem, this application provides the following technical solutions:

[0031] Furthermore, step 2 includes the following steps:

[0032] Step 21: The information collection module collects each sound wave information k I , the sound wave information k that is within the preset time threshold I Merge into one event set;

[0033] Step 22: For each event set, extract all sound wave information k I k` i , k i ; Calculate the number K of the first acoustic wave detectors that receive acoustic wave information in the event set;

[0034] Step 23: All the sound wave signals k in the event set i , sound wave signal k i Use Fourier transform to transform into time domain information V (t, f);

[0035] ;

[0036] Among them, V(t,f) is a negative vector with dimension K, which represents the spectrogram at time t and frequency f; t represents time, f represents frequency index, and M l (t, f) represents the short-time Fourier transform result of the lth source signal at time t and frequency f; U(f, ϕ l ,θ l )a: frequency domain steering vector representing the direction of the lth source signal at frequency f, φ l represents the direction angle of the lth source signal, θ l represents the elevation angle of the lth source signal, a represents the magnification factor of the direction angle and elevation angle, a>10; C represents an M-dimensional vector space in the complex domain, L represents the number of source signals, the source signal is the signal emitted by the position where the acoustic wave signal is generated when the crack is generated, and T ra represents the transpose of the vector;

[0037] Step 24: The first acoustic wave detectors related to all event sets are formed into corresponding signal receiving arrays, from k` i , k i Extract the steering vector in, Indicates k` i , k i The steering vector of the lth source signal in;

[0038] Step 25: Draw the guide vector , and time domain information V (t, f) are used as input features and input into the positioning neural network model to obtain the initial crack position set.

[0039] Since the size of metal cracks is very small, the distance of sound signals at the edge of the cracks is also very small. In such a small range, it is difficult to distinguish the sound source signals that appear simultaneously using existing transducers. To address this problem, this application proposes the following technical solutions:

[0040] Furthermore, in step 23, when distinguishing the number of source signals, the following scheme is adopted:

[0041] Step 231: For each event set, all sound wave information k I ω in k,i ,ω k,i `Align with each other at the initial end of the time node;

[0042] Step 232: Extract all ω k,i ,ω k,i There are time nodes that change in `;

[0043] Step 233: Set all ω k,i ,ω k,i `For comparison, if there is a time node PkI , when ω exceeds a% k,i ,ω k,i ` exists in the same position; then the time node P kI As a sound source change point;

[0044] Step 234: For each sound source change point, use it as the access point of the source signal or the exit point of the source signal.

[0045] In the technical solution provided in the present application, the change node of the wavelength change information is used as the access point and the exit point of the sound source, so the source signal can be separated from the source data more accurately, thereby facilitating the accurate location of the source signal.

[0046] Further, for the first acoustic wave detectors of the acoustic wave signals that can be collected in one event set, one first acoustic wave detector is selected as the calibrated first acoustic wave detector, and the remaining first acoustic wave detectors are arranged in order;

[0047] In step 24:

[0048] ;

[0049] ;

[0050] ;

[0051] Where t represents time, f represents frequency, represents the time difference between the source signal l reaching the calibrated first acoustic wave detector and the time difference between the source signal l reaching the nth first acoustic wave detector; φ l represents the direction angle of the source signal l, θ l represents the elevation angle of the source signal l, c represents the speed of sound in the metal cabin cover;

[0052] represents the arrival distance of the source signal l between the nth first acoustic wave detector and the calibrated first acoustic wave detector, ε1 represents the Cartesian coordinates of the calibrated first acoustic wave detector, and ε n The Cartesian coordinate n representing the nth first acoustic wave detector is the index of the first acoustic wave detector in the same event set.

[0053] In the solution provided in the present application, for each event set, all first acoustic wave detectors that can receive sound signals are used as a receiving array, and then information extraction is performed based on this, so as to accurately extract the sound source information that characterizes the location of the source signal.

[0054] Furthermore, step 3 includes the following steps:

[0055] Step 31: monitor the number of event sets in real time. When the number of event sets is greater than a preset threshold, the information processing device regards each event set as a corresponding crack.

[0056] Step 32: For each event set, all source signal positions are obtained, and the source signal positions are used as the crack edges corresponding to the event set;

[0057] Step 33: For each crack, take the largest surface of the crack edge as the projection surface, take the point where the perpendicular line from the center of the projection surface intersects with the metal cabin cover surface as a detection position, and take the angle between the perpendicular line from the detection position to the center of the projection surface and the metal cabin surface as the layout angle.

[0058] In the technical solution provided by the present application, effective detection of the maximum diameter of the crack is achieved by carefully determining the detection position and the layout angle. Specifically, the point where the perpendicular line from the center of the projection surface intersects with the surface of the metal cabin cover is determined as the detection position, and the angle between the perpendicular line from the detection position to the center of the projection surface and the surface of the metal cabin is used as the layout angle. This design enables the second acoustic wave detector to more accurately capture the maximum diameter information of the crack, providing a more reliable data basis for subsequent analysis and processing, thereby improving the accuracy and efficiency of detection.

[0059] Furthermore, the neural network model is an LSTM neural network model.

[0060] The beneficial effect of the present application is that the technical solution provided by the present application uses a neural network model to screen the echo information to obtain the effective echo information of each second acoustic wave detector, which effectively reduces the amount of effective data removed from the echo information. Furthermore, the echo information extracted in the present application is used to analyze the location and size of the crack, achieving higher accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The drawings constituting a part of this application are used to provide a further understanding of this application, so that other features, purposes and advantages of this application become more obvious. The illustrative embodiment drawings and their descriptions of this application are used to explain this application and do not constitute an improper limitation on this application.

[0062] In addition, throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that the components and elements are not necessarily drawn to scale.

[0063] In the attached picture:

[0064] Figure 1 The figure is a flow chart of a method for extracting gap information of a metal engine compartment cover.

[0065] Figure 2Schematic diagram of the arrangement of the first acoustic wave detector and grating optical fiber.

[0066] Figure 3 ω k,i ,ω k,i `Schematic diagram after alignment. DETAILED DESCRIPTION

[0067] Embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms, and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.

[0068] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other.

[0069] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments.

[0070] Reference Figure 1 The present application provides a method for extracting acoustic wave information from an acoustic wave detector array, which can accurately filter out clutter information in a metal cabin cover, so that the echo information collected by the acoustic wave detector array can accurately restore the position and number of gaps in the metal cabin cover. In order to achieve this purpose, the technical solution of the present application will be described in detail below.

[0071] Embodiment 1: A method for extracting gap information of a metal nacelle cover comprises the following steps:

[0072] Step 1: a plurality of first acoustic wave detectors are arranged on the surface of the metal engine cabin cover to detect in real time the acoustic wave signals generated by cracks inside the metal engine cabin cover to obtain acoustic wave information.

[0073] Step 1 includes the following steps:

[0074] Step 11: Obtain at least three fixed detection positions on the surface of the metal nacelle cover that are not located in the same plane.

[0075] The surface of the metal cabin cover is not flat, so it is only necessary to find three locations with different thicknesses on its surface as the corresponding fixed detection locations. When selecting fixed detection locations, try to ensure that they are evenly distributed on the surface of the metal cabin cover to avoid gathering too many fixed detection locations in a certain area and ignoring other areas.

[0076] Step 12: Fix a first acoustic wave detector at each fixed detection position.

[0077] The first acoustic wave detector uses an ultrasonic transducer and has the function of converting the received acoustic wave signal into an electrical signal. In practical applications, the receiving end of the first acoustic wave detector fits tightly against the metal cabin cover. When gaps and vibrations occur inside the metal cabin cover, the first acoustic wave detector can sensitively receive these weak signals.

[0078] Step 13: Each first acoustic wave detector detects in real time the acoustic wave signal generated by the crack inside the metal cabin cover to obtain acoustic wave information.

[0079] The acoustic signal generated inside the metal cabin cover is very weak. For each first acoustic wave detector, various interferences or abnormalities in the internal circuit may cause erroneous acoustic information. To address this problem, the following solutions are provided:

[0080] There are n first acoustic wave detectors arranged, where n is an even number greater than zero, and the first acoustic wave detectors are divided into n / 2 detection groups, and the first acoustic wave detectors in each detection group are connected via two grating optical fibers;

[0081] In any detection group, one and only one end of the grating optical fiber is connected to the surface of the metal cabin cover, and the ends of the two grating optical fibers connected to the metal cabin cover are different from each other.

[0082] refer to Figure 2 For example, when n=4, that is, four first acoustic wave detectors are arranged on the metal cabin cover, and these four detectors can be divided into two detection groups. In each detection group, two first acoustic wave detectors are connected by two grating optical fibers. Specifically, one end of these two grating optical fibers will be connected to the metal cabin cover. Taking detection group 1 as an example, it includes two first acoustic wave detectors A1 and A2, and A1 and A2 are connected by two grating optical fibers B1 and B2. One end of the grating optical fiber B1 connected to the first acoustic wave detector A1 is connected to the metal cabin cover, and the other end is suspended on the metal cabin cover; at the same time, one end of the grating optical fiber B2 connected to the first acoustic wave detector A2 is also connected to the metal cabin cover, and the other end is also suspended.

[0083] This setting ensures that only one end of the grating fiber can access the acoustic signal in the metal cabin cover, avoiding the acoustic signal entering from both ends of the grating fiber at the same time, causing the grating fiber wavelength to be affected by too many factors, thereby improving the accuracy and stability of detection.

[0084] The aforementioned suspension mainly refers to the difficulty of directly transmitting the acoustic wave signal on the metal cabin cover to the grating optical fiber. For the grating optical fiber and the first acoustic wave detector, they both have an access point for receiving the acoustic source signal. In this solution, the two first acoustic wave detectors are connected by the metal cabin cover. For each individual grating optical fiber, only one access point for the acoustic source signal is connected.

[0085] The above configuration allows only one end of the grating fiber to access the acoustic signal in the metal cabin cover, thereby preventing the acoustic signal from entering the grating from both ends of the grating fiber, causing the wavelength of the grating fiber to be affected by too many factors.

[0086] Thus, after adopting the above-mentioned arrangement mode of the first acoustic wave detector, step 13 specifically includes the following steps:

[0087] Step 131: When the first acoustic wave detector k receives the acoustic wave signal k i When the first acoustic wave detector k sends a standard optical signal to the first acoustic wave detector k' through the grating optical fiber connected to the metal cabin cover at the end of the first acoustic wave detector k, until the acoustic wave information disappears, and the first acoustic wave detector k' detects the wavelength change information ω of the grating optical fiber. k,i ;

[0088] Wherein, k is the serial number of the first acoustic wave detector, k' is the serial number of the first acoustic wave detector in the same detection group as the kth first acoustic wave detector; k and k' are both integers; k i represents the index of the acoustic wave signal detected by the first acoustic wave detector k, where i is an integer;

[0089] Step 132: When the first acoustic wave detector k' receives the acoustic wave signal k i When the first acoustic wave detector k' sends a standard optical signal to the first acoustic wave detector k' through the grating optical fiber connected to the metal cabin cover at the end of the first acoustic wave detector k', until the sound source information disappears, and the first acoustic wave detector k' detects the wavelength change information ω of the grating optical fiber. k,i `;

[0090] Step 133: If the first acoustic wave detector k and the first acoustic wave detector k' receive the acoustic wave signal k i If the time interval of k is within the preset time interval, i ,ω k,i ,ω k,i `, k` i As the sound wave information k I , I represents the index of the sound wave information received by the kth first sound wave detector.

[0091] The wavelength change information is: wavelength related to time. In the two-dimensional graph, the horizontal axis is time and the vertical axis is wavelength. When the wavelength changes, it means that the sound wave entering the grating fiber has changed.

[0092] In this solution, the preset time interval is related to the distance between the two first acoustic wave detectors and the sound speed. For example, if the distance between the two first acoustic wave detectors of the same detection group is 100 meters and the sound speed is 10 meters / s (this is an assumed sound speed), the preset time interval will be slightly greater than 10 seconds.

[0093] In a further embodiment, the first acoustic wave detector in the same detection group needs to be able to receive the acoustic wave signal within a preset time interval in order to record it as a valid acoustic wave signal.

[0094] Step 2: The information processing device determines the location of the crack based on the information of each sound wave and obtains a set of initial crack positions.

[0095] Step 2 includes the following steps:

[0096] Step 21: The information collection module collects each sound wave information k I , the sound wave information k that is within the preset time threshold I Merge into one event collection.

[0097] The time threshold in step 21 is designed in the same manner as the time interval in step 133, and is set according to the propagation speed of the sound wave. The preset time threshold is the distance between the two farthest first sound wave detectors divided by the sound speed, plus a fixed error.

[0098] In step 21, for one event set, it is actually an event that one crack appears in the metal nacelle cover.

[0099] Step 22: For each event set, extract all sound wave information k I k` i , k i ; Calculate the number K of the first acoustic wave detectors that receive acoustic wave information in the event set;

[0100] In step 22, the core task is to accurately locate the source signal generation position of each event set. The method adopted in this scheme is to form an array of all first acoustic wave detectors and use these detectors for positioning. Since the position information of the first acoustic wave detectors has been recorded in advance, the position of each acoustic wave signal can be roughly determined by measuring and analyzing the time difference of the acoustic wave signals received by each first acoustic wave detector. This method makes full use of the characteristics of acoustic wave signals propagating in space and the spatial distribution of the detector array to achieve accurate estimation of the source signal position.

[0101] Step 23: All the sound wave signals k in the event set i , sound wave signal k i Use Fourier transform to transform into time domain information V (t, f);

[0102] ;

[0103] Among them, V(t,f) is a negative vector with dimension K, which represents the spectrogram at time t and frequency f; t represents time, f represents frequency index, and M l (t, f) represents the short-time Fourier transform result of the lth source signal at time t and frequency f; U(f, aϕ l ,aθ l ): represents the frequency domain steering vector of the direction of the lth source signal at frequency f, φ l represents the direction angle of the lth source signal, θ l represents the elevation angle of the lth source signal, a represents the amplification factor of the direction angle and elevation angle, a>10; C represents an M-dimensional vector space in the complex domain, L represents the number of source signals, and the source signal is the signal emitted by the position where the acoustic wave signal is generated when the crack occurs.

[0104] It is challenging to accurately determine whether there is a crack in the metal cabin cover and the size of the crack. However, by analyzing the received sound wave information, the existence and location of the crack can be roughly determined. For a single crack, it is necessary not only to know its location, but also to locate the boundary of the crack as accurately as possible, so as to find the best placement position of the second sound wave detector according to the boundary of the crack.

[0105] The essence of locating the crack boundary is to distinguish the location of the source signal in an event set. When a crack occurs, its edge will vibrate, and then several source signals will be generated at the edge of the crack. Therefore, the location of the source signal corresponds to the boundary position of the crack. When determining the source signal, conventional schemes are difficult to capture the weak vibration of the edge when the crack is torn. For this reason, this scheme uses grating optical fiber to capture this weak signal. At present, grating optical fiber has been developed in acoustic wave detection technology, especially acoustic wave detectors based on fiber Bragg grating (FBG). However, the existing technology is mainly used to calculate the frequency, wavelength and other information of the acoustic wave signal. In fact, if a fiber Bragg grating (FBG) acoustic wave detector is used, it is also difficult to detect the various source signals when the crack is torn.

[0106] Therefore, based on the existing grating fiber technology, this scheme will use the wavelength change rate in the grating fiber to determine the access point of the source signal in order to separate the source signal from the sound signal. The specific scheme is as follows:

[0107] In step 23, when distinguishing the number of source signals, the following scheme is adopted:

[0108] Step 231: For each event set, all sound wave information k I ω in k,i ,ω k,i `Align with each other at the initial end of the time node;

[0109] Step 232: Extract all ω k,i ,ω k,i There are time nodes that change in `;

[0110] Step 233: Set all ω k,i ,ω k,i `For comparison, if there is a time node P kI , when ω exceeds a% k,i ,ω k,i ` exists in the same position; then the time node P kI As a sound source change point;

[0111] Step 234: For each sound source change point, use it as the access point of the source signal or the exit point of the source signal.

[0112] refer to Figure 3 , the key point of step 23 is: because ω k,i ,ω k,i `They all appear with the appearance of the sound source and disappear with the disappearance of the sound source, so k,i ,ω k,i The first ends of the gratings are aligned with each other, and the effect of the sound signal on each grating fiber can be known. k,i ,ω k,i ` shows that the wavelength has changed, it can be roughly inferred that a new sound wave signal is connected to the grating optical fiber at this time.

[0113] In step 244, after the sound source change points are determined, the first half of the sound source change points are used as the access points of the source signal, and the second half of the sound source signal is used as the exit points of the source signal.

[0114] In this solution, in view of the weak crack signal, the grating fiber is not used to determine the wavelength, size and position of each source signal. Instead, the grating fiber is used to determine whether the source signal exists. In this way, the weak source signal can be separated.

[0115] Step 24: The first acoustic wave detectors related to all event sets are formed into corresponding signal receiving arrays, from k` i , ki Extract the steering vector in, Indicates k` i , k i The steering vector of the lth source signal in .

[0116] In step 24:

[0117] ;

[0118] ;

[0119] ;

[0120] Where t represents time, f represents frequency, represents the time difference between the source signal l reaching the calibrated first acoustic wave detector and the time difference between the source signal l reaching the nth first acoustic wave detector; φ l represents the direction angle of the source signal l, θ l Represents the elevation angle of the source signal l, c represents the speed of sound in the metal cabin cover; Tra represents the transpose of the vector;

[0121] represents the arrival distance of the source signal l between the nth first acoustic wave detector and the calibrated first acoustic wave detector, ε1 represents the Cartesian coordinates of the calibrated first acoustic wave detector, and ε n The Cartesian coordinate n representing the nth first acoustic wave detector is the index of the first acoustic wave detector in the same event set.

[0122] Step 25: Draw the guide vector The time domain information V (t, f) is used as input features and input into the positioning neural network model to obtain the location of the source signal, and the initial position set of the cracks is obtained according to the location of the source signal.

[0123] In this scheme, the steering vector contains the information between each source signal and each first acoustic wave detector, and the time domain information contains the short-time Fourier transform results of all source signals, so the location of the source signal can be located using the positioning neural network model.

[0124] In this solution, the positioning neural network model is a common convolutional neural network (CNN). CNN simulates the visual processing mechanism of the human brain, performs convolution operations on the input sound signal, extracts local features, and reduces the feature dimension through the pooling layer to reduce the amount of calculation.

[0125] In this scheme, the input is the steering vector And the time domain information V, the output is the position of each source signal. The neural network model is an existing neural network model, so the structure and training method of the positioning neural network model are not described here.

[0126] Step 3: When the number of cracks is greater than a preset threshold, the information processing device calculates the layout positions and layout angles of a plurality of second acoustic wave detectors according to the initial position set of the cracks to obtain the detector position information.

[0127] When the number of cracks is not greater than the preset threshold:

[0128] The first acoustic wave detector continues to detect acoustic wave signals generated by cracks inside the metal nacelle cover.

[0129] Step 3 includes the following steps:

[0130] Step 31: Monitor the number of event sets in real time. When the number of event sets is greater than a preset threshold, the information processing device regards each event set as a corresponding crack.

[0131] When the number of event sets is not greater than the preset threshold: the first acoustic wave detector continues to detect the acoustic wave signals generated by the cracks inside the metal cabin cover.

[0132] That is, only when the number of event sets exceeds a certain threshold, indicating that the number of cracks in the metal cabin cover is too large, will the subsequent steps be performed to detect the location and size of the cracks with a second acoustic wave detector.

[0133] Step 32: For each event set, all source signal positions are obtained, and the source signal positions are used as the crack edges corresponding to the event set;

[0134] According to the description above, in each event set, there is actually a crack. Therefore, the source signal position in each event set is the edge position of the crack.

[0135] Step 33: For each crack, take the largest surface of the crack edge as the projection surface, take the point where the perpendicular line from the center of the projection surface intersects with the metal cabin cover surface as a detection position, and take the angle between the perpendicular line from the detection position to the center of the projection surface and the metal cabin surface as the layout angle.

[0136] Using the solution in step 2, for each crack, about 2 to 5 source signals or positions can be basically distinguished. Connecting these two points as fixed points can roughly obtain a projection surface, which is the maximum possible position of the crack. Then use the point where the perpendicular line from the center of the projection surface intersects with the surface of the metal cabin cover as a detection position, and then use the angle between the perpendicular line from the detection position to the center of the projection surface and the surface of the metal cabin as the layout angle. At this time, the second acoustic wave detector will be able to shoot at the gap at the best angle.

[0137] Step 4: Arrange a plurality of second acoustic wave detectors according to the detector position information, and each second acoustic wave detector releases an ultrasonic signal to the generating position;

[0138] Step 5: Receive the echo signals received by each second acoustic wave detector to obtain echo information, send the echo information and detector position information to the neural network model, filter out interference information in the echo information of each second acoustic wave detector, and obtain the echo signals of each second acoustic wave detector.

[0139] Furthermore, the neural network model in step 5 is an LSTM neural network model.

[0140] Specifically, in step 5, because the positions of the second acoustic wave detectors and the emission angles of the acoustic wave signals have been determined, the influence of the echo signals received by each second acoustic wave detector is relatively related to the distance.

[0141] In this scheme, for any second acoustic wave detector, the acoustic wave intensity, the position of the second acoustic wave detector and the emission angle of the acoustic wave signal are input into the LSTM neural network model, and the LSTM neural network model outputs the intensity range and frequency range of the acoustic wave signal received by the second acoustic wave detector.

[0142] The LSTM neural network model is a mature network model, and the specific construction method will not be discussed here. The key point of this solution is that the number of second acoustic wave detectors is reduced, so there will not be a large number of ultrasonic signals reflected back and forth in the metal plate, and then the long-term memory ability of the LSTM neural network model can be used to find the interference relationship between ultrasonic signals of different positional relationships.

[0143] Embodiment 2: A gap information extraction system for a metal engine nacelle cover, characterized in that it comprises an information processor, a first acoustic wave detector and a second acoustic wave detector, wherein the first acoustic wave detector and the second acoustic wave detector are respectively connected to the information processor by signal;

[0144] The extraction system uses the gap information extraction method of the metal cabin cover to extract the echo signal.

[0145] The above description is only some preferred embodiments of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present application is not limited to the technical solutions formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above invention concept. For example, the above features are replaced with (but not limited to) technical features with similar functions disclosed in the embodiments of the present application to form a technical solution.

Claims

1. A method for extracting gap information of a metal nacelle cover, characterized in that: The steps include: Step 1: a plurality of first acoustic wave detectors are arranged on the surface of the metal engine cabin cover to detect in real time the acoustic wave signals generated by cracks inside the metal engine cabin cover to obtain acoustic wave information; Step 2: The information processing device determines the location of the crack according to each acoustic wave information, and obtains a set of initial crack locations; Step 3: When the number of cracks is greater than the preset threshold: The information processing device calculates the arrangement positions and arrangement angles of the plurality of second acoustic wave detectors according to the initial position set of the gaps, and obtains the detector position information; When the number of cracks is not greater than the preset threshold: The first acoustic wave detector continues to detect the acoustic wave signals generated by the cracks inside the metal cabin cover; Step 4: Arrange a plurality of second acoustic wave detectors according to the detector position information, and each second acoustic wave detector releases an ultrasonic signal to the generating position; Step 5: receiving the echo signals received by each second acoustic wave detector to obtain echo information, sending the echo information and the detector position information to the neural network model, filtering out the interference information in the echo information of each second acoustic wave detector, and obtaining the echo signals of each second acoustic wave detector; Step 2 includes the following steps: Step 21: The information collection module collects each sound wave information k I , the sound wave information k that is within the preset time threshold I Merge into an event set, where I represents the index of the sound wave information received by the kth first sound wave detector; Step 22: For each event set, extract all sound wave information k I k` i , k i ; Calculate the number K of the first acoustic wave detectors that receive acoustic wave information in the event set; Step 23: All the sound wave signals k in the event set i , sound wave signal k i Use Fourier transform to transform into time domain information V (t, f); ； Among them, V(t,f) is a complex vector with dimension K, which represents the spectrogram at time t and frequency f; t represents time, f represents frequency index, and M l (t, f) represents the l The short-time Fourier transform result of a source signal at time t and frequency f; U(f,ϕ l ,θ l )a: indicates the l The direction of the source signal at frequency f is the frequency domain steering vector, φ l Indicates l The direction angle of the source signal, θ l Indicates l The elevation angle of the source signal, a represents the magnification factor of the direction angle and elevation angle, a>10; C represents an M-dimensional vector space in the complex domain, L represents the number of source signals, and the source signal is the signal emitted by the position where the acoustic wave signal is generated when the crack is generated; Step 24: The first acoustic wave detectors related to all event sets are formed into corresponding signal receiving arrays, from k` i , k i Extract the steering vector ,in, Indicates k` i , k i Middle l A steering vector of a source signal; Step 25: Draw the guide vector The time domain information V (t, f) is used as input features and input into the positioning neural network model to obtain the location of the source signal, and the initial position set of the crack is obtained according to the location of the source signal; In step 23, when distinguishing the number of source signals, the following scheme is adopted: Step 231: For each event set, all sound wave information k I ω in k,i ,ω k,i `Align with each other at the initial end of the time node; Step 232: Extract all ω k,i ,ω k,i There are time nodes that change in `; Step 233: Set all ω k,i ,ω k,i `For comparison, if there is a time node P kI , when the preset ratio is exceeded k,i ,ω k,i `There is the same time node at the same position in the kI As a sound source change point; Step 234: for each sound source change point, use it as a source signal access point or a source signal exit point; For the first acoustic wave detectors of the acoustic wave signals that can be collected in one event set, one first acoustic wave detector is selected as the calibrated first acoustic wave detector, and the remaining first acoustic wave detectors are arranged in order; In step 24: ； ； ； Where t represents time, f represents frequency, Indicates source signal l The time difference between reaching the first calibrated acoustic wave detector and reaching the nth first acoustic wave detector; φ l Indicates source signal l The direction angle, θ l Indicates source signal l The elevation angle, c represents the speed of sound in the metal cabin cover, and Tra represents the transpose of the vector; Indicates source signal l The reaching distance between the nth first acoustic wave detector and the calibration first acoustic wave detector, ε1 represents the Cartesian coordinates of the calibration first acoustic wave detector, ε n represents the Cartesian coordinates of the nth first acoustic wave detector, where n is the index of the first acoustic wave detector in the same event set; Step 3 includes the following steps: Step 31: monitor the number of event sets in real time. When the number of event sets is greater than a preset threshold, the information processing device regards each event set as a corresponding crack. Step 32: For each event set, all source signal positions are obtained, and the source signal positions are used as the crack edges corresponding to the event set; Step 33: For each crack, take the largest surface of the crack edge as the projection surface, take the point where the perpendicular line from the center of the projection surface intersects with the metal cabin cover surface as a detection position, and take the angle between the perpendicular line from the detection position to the center of the projection surface and the metal cabin surface as the layout angle.

2. The method for extracting gap information of a metal nacelle cover according to claim 1, characterized in that: Step 1 includes the following steps: Step 11: Obtain at least three fixed detection positions on the surface of the metal nacelle cover that are not located in the same plane; Step 12: Fixing a first acoustic wave detector at each fixed detection position; Step 13: Each first acoustic wave detector detects in real time the acoustic wave signal generated by the crack inside the metal cabin cover to obtain acoustic wave information.

3. The method for extracting gap information of a metal nacelle cover according to claim 2, characterized in that: There are n first acoustic wave detectors arranged, where n is an even number greater than zero; The first acoustic wave detector is divided into n / 2 detection groups, and the first acoustic wave detectors in each detection group are connected via two grating optical fibers; In any detection group, one and only one end of the grating optical fiber is connected to the surface of the metal cabin cover, and the ends of the two grating optical fibers connected to the metal cabin cover are different from each other.

4. The method for extracting gap information of a metal nacelle cover according to claim 3, characterized in that: Step 13 includes the following steps: Step 131: When the first acoustic wave detector k receives the acoustic wave signal k i When the first acoustic wave detector k sends a standard optical signal to the first acoustic wave detector k' through the grating optical fiber connected to the metal cabin cover at the end of the first acoustic wave detector k, until the acoustic wave signal disappears, the first acoustic wave detector k' detects the wavelength change information ω of the grating optical fiber. k,i ; Wherein, k is the serial number of the first acoustic wave detector, k' is the serial number of the first acoustic wave detector in the same detection group as the kth first acoustic wave detector; k and k' are both integers; k i represents the acoustic wave signal detected by the first acoustic wave detector k, i is the index of the acoustic wave signal, and i is an integer; Step 132: When the first acoustic wave detector k' receives the acoustic wave signal k i When the first acoustic wave detector k' sends a standard optical signal to the first acoustic wave detector k' through the grating optical fiber connected to the metal cabin cover at the end of the first acoustic wave detector k', until the acoustic wave signal disappears, and the first acoustic wave detector k' detects the wavelength change information ω of the grating optical fiber. k,i `; Step 133: If the first acoustic wave detector k and the first acoustic wave detector k' receive the acoustic wave signal k i If the time interval of k is within the preset time interval, i ,ω k,i ,ω k,i `, k` i As the sound wave information k I , k I Represents the Ith sound wave information received by the kth first sound wave detector.

5. The method for extracting gap information of a metal nacelle cover according to claim 1, characterized in that: The neural network model in step 5 is an LSTM neural network model.

6. A gap information extraction system for a metal nacelle cover, characterized in that: It includes an information processor, a first acoustic wave detector and a second acoustic wave detector, wherein the first acoustic wave detector and the second acoustic wave detector are respectively connected to the information processor by signal; The extraction system uses the gap information extraction method of the metal nacelle cover according to any one of claims 1 to 5 to extract the echo signal.

Images