Method for identifying concrete cracks of hydropower house

Abstract

The invention discloses a hydropower station plant concrete crack identification method, which comprises the steps of applying physical property sensitive excitation and reference excitation to the same target point of a concrete crack in a time-sharing manner, and synchronously collecting a first original time domain data set and a second original time domain data set which are respectively generated by the two kinds of excitation; respectively and independently executing reverse time migration imaging on the two groups of data sets so as to reconstruct corresponding first and second three-dimensional optical sound source intensity images; and based on the two three-dimensional images, calculating a photoacoustic physical index through difference and normalization operation, and judging whether the interior of the crack is in a water-containing or dry physical state or other physical states according to a preset criterion. According to the invention, through bimodal excitation and differential analysis, quantitative diagnosis of the internal physical state of the crack is realized; reverse time migration imaging is adopted, so that high-precision reconstruction of a three-dimensional form is ensured; and in combination with non-contact measurement, the safety and convenience of detection are improved.

Description

Technical Field

[0001] This invention belongs to the field of non-destructive testing technology, specifically relating to a method for identifying concrete cracks in hydropower station powerhouses. Background Technology

[0002] As the core structural material of large-scale infrastructure such as hydropower plant buildings, the structural health of concrete during its long-term service life directly affects the safe and stable operation of the entire project. Due to the complex water pressure, temperature changes, and vibration loads it withstands, concrete structures inevitably develop and crack. These cracks not only weaken the structure's load-bearing capacity and integrity, but more importantly, when they act as channels for external corrosive media (such as water and chloride ions), they significantly accelerate the corrosion of reinforcing steel and freeze-thaw damage to the concrete, thus severely shortening the structure's service life. Therefore, timely and accurate detection and diagnosis of concrete cracks, especially the assessment of their three-dimensional spatial distribution and key physical properties such as the presence of water inside, are crucial for ensuring the safe operation of hydropower plants.

[0003] Currently, methods for detecting concrete cracks are mainly divided into two categories: surface inspection and internal non-destructive testing. Traditional surface inspection methods, such as manual visual inspection, crack width measurement, and image processing-based surface crack identification, are simple to operate, but they can only obtain two-dimensional information about the cracks on the structural surface, such as length and width. These methods cannot detect key three-dimensional geometric parameters such as crack depth, internal direction, and spatial connectivity, nor can they determine whether the inside of the crack is dry or filled with water. Therefore, the diagnostic information they provide is extremely limited and cannot support accurate structural safety assessments.

[0004] To investigate the internal structure of cracks, various non-destructive testing (NDT) techniques have been developed. For example, ultrasonic testing is currently a widely used method for detecting internal defects. This technique infers the presence and approximate location of defects by analyzing the propagation time and attenuation of ultrasonic waves in concrete. However, conventional ultrasonic testing typically requires the use of a coupling agent to achieve acoustic contact between the sensor and the concrete surface, which is inconvenient on rough, damp, or vertical structural surfaces and makes rapid automated scanning difficult. More importantly, while ultrasonic signals are sensitive to cracks (i.e., interfaces with discontinuous acoustic impedance), their ability to distinguish the filling medium (such as air and water) within cracks is very limited, hindering effective physical property diagnosis. Other techniques, such as ground-penetrating radar (GPR), are sensitive to moisture, but their spatial resolution is often insufficient to accurately depict the complex morphology of minute cracks; while radiation imaging techniques such as X-rays, although high-resolution, are unsuitable for large-scale on-site inspections of large structures such as hydroelectric powerhouses due to their complex equipment, high cost, and radiation safety concerns.

[0005] In summary, existing technologies have significant limitations in the comprehensive diagnosis of concrete cracks. There is an urgent need for an identification method that can simultaneously achieve high-precision three-dimensional geometric imaging and quantitative identification of internal physical properties, while also possessing non-contact and high-efficiency operation capabilities, to meet the pressing needs of refined and preventative maintenance of modern large-scale infrastructure. Summary of the Invention

[0006] The purpose of this invention is to provide a method for identifying concrete cracks in hydropower plant buildings, which solves the problem that existing technologies cannot perform non-contact, high-precision three-dimensional geometric imaging of concrete cracks and simultaneously quantitatively determine the physical state of the internal filling medium.

[0007] The technical solution adopted in this invention is a method for identifying concrete cracks in hydropower plant buildings. This method combines spectrally selective dual-modal photoacoustic excitation with reverse time migration imaging technology, which can accurately handle wave propagation in complex media, and differential photoacoustic analysis for quantitative property inversion, ultimately generating a diagnostic model that integrates the three-dimensional geometric information and internal physical state information of the cracks.

[0008] The method includes the following steps: S1: Apply both property-sensitive and reference excitations to the same target point of the concrete crack in a time-division manner, and simultaneously acquire the first and second original time-domain datasets generated by the two excitations respectively. The principle of this step is that by selecting two laser wavelengths with significant absorption differences for specific media (such as water) that may exist inside the crack, the physical property information of the medium is encoded into two sets of photoacoustic signals with different intensities.

[0009] S2: Perform reverse time migration imaging independently on the first and second original time-domain datasets to reconstruct the corresponding first and second three-dimensional photoacoustic source intensity images. This step is based on the time invariance of the acoustic wave equation, using the signal acquired by the sensor as a time-reversed sound source, and retracing the wave field by numerically solving the wave equation.

[0010] S3: Based on the first and second three-dimensional photoacoustic source intensity images, calculate the photoacoustic property index at each location within the crack region; and determine the physical state of the filling medium inside the crack according to a preset property discrimination criterion. The principle of this step is to eliminate common-mode measurement noise and highlight the signal changes caused by differences in the spectral absorption characteristics of the medium through specific normalized differential operations.

[0011] S4: Extract the three-dimensional geometric model of the crack and map the physical state determined in S3 onto the three-dimensional geometric model to generate a three-dimensional physical property diagnostic model that integrates geometric and physical property information for crack identification.

[0012] The invention is further characterized by: In a preferred embodiment, the laser wavelength used for property-sensitive excitation is the strong absorption wavelength of water, and the laser wavelength used for reference excitation is the weak absorption wavelength of water.

[0013] Specifically, the strong absorption wavelength of water can be 1450nm or 1940nm, and the weak absorption wavelength of water can be 1064nm or 532nm.

[0014] In a preferred embodiment, synchronous acquisition is achieved through a non-contact acoustic sensing unit, which is one or more laser Doppler vibrometers, to enable completely non-contact measurement.

[0015] In a preferred embodiment, reverse time migration imaging specifically includes: using the time-reversed original time-domain dataset as an artificial source term, substituting it into the time-reversed acoustic wave equation for numerical solution to obtain the backpropagating wave field; and applying a zero-time imaging condition at the last time step, using the wave field value at that time as the intensity value of the three-dimensional photoacoustic source intensity image. The numerical solution of this wave equation can be achieved using the finite difference time-domain method.

[0016] In a preferred embodiment, the property discrimination criteria include: setting a high-level discrimination threshold and a low-level discrimination threshold; when the photoacoustic property index value is greater than the high-level discrimination threshold, the physical state of the corresponding position is judged as a water-containing state; when the absolute value of the photoacoustic property index value is less than or equal to the low-level discrimination threshold, the physical state of the corresponding position is judged as a dry state.

[0017] In a preferred embodiment, the step of generating a three-dimensional physical property diagnostic model specifically includes: extracting a three-dimensional geometric model of the crack based on a first three-dimensional photoacoustic source intensity image by threshold segmentation or region growing; and assigning the physical state of the medium filling the crack to the corresponding position in the three-dimensional geometric model by color mapping.

[0018] In a preferred embodiment, the reverse time-shift imaging is based on a preset sound velocity field model, which is either a uniform sound velocity model or a non-uniform sound velocity model set according to the known distribution of objects inside the concrete structure, in order to improve the accuracy of the imaging.

[0019] Compared with the prior art, the method for identifying concrete cracks in hydropower plant buildings of the present invention has the following beneficial effects: 1) This invention uses dual-modal photoacoustic excitation with physical property sensitive excitation and reference excitation, and performs differential photoacoustic analysis on the reconstructed three-dimensional photoacoustic image to achieve quantitative diagnosis of the physical state inside the crack. It can effectively determine whether the inside of the crack is water-containing or dry, and provides a direct quantitative basis for assessing the risks of steel corrosion, freeze-thaw damage and other damage caused by cracks. 2) This invention uses a wave equation-based inverse time migration imaging method to process the acquired acoustic signals, which can accurately simulate the propagation and scattering process of sound waves in the non-uniform concrete medium, thereby obtaining a high-resolution three-dimensional geometric image of the crack. Compared with traditional imaging algorithms, it can more clearly restore the spatial direction, depth and complex shape of the crack. 3) This invention uses a non-contact acoustic sensing unit to collect signals and finally fuses and visualizes the geometric model and physical property state, realizing completely non-contact non-destructive testing, avoiding damage to the structure and facilitating automated scanning. At the same time, the generated three-dimensional physical property diagnostic model can intuitively and comprehensively display crack information, providing convenience for the evaluation and decision-making of engineering technicians. Attached Figure Description

[0020] Figure 1 This is a flowchart of the method for identifying concrete cracks in a hydropower station powerhouse according to the present invention; Figure 2 This is a schematic diagram of the three-dimensional physical property diagnostic model in the method for identifying concrete cracks in hydropower plant buildings according to the present invention; Figure 3 This is a schematic diagram of the dual-modal excitation principle in the identification method of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1: Please see the appendix Figure 1 - Appendix Figure 3 The method for identifying concrete cracks in a hydropower station powerhouse according to the present invention includes the following steps: S1 uses a detection system containing a dual-modal photoacoustic excitation unit and a non-contact acoustic sensing unit to emit two specific wavelength laser pulses at the same target point on the surface of a concrete structure in a time-division manner, and simultaneously collects the acoustic signals excited by the two laser pulses and propagating from the inside of the concrete to the surface, thereby obtaining two independent sets of raw time-domain datasets.

[0023] S2, the data processing and imaging unit receives the above two sets of original time-domain datasets, and based on the preset concrete sound velocity model, independently executes the reverse time migration imaging algorithm on each set of data. By solving the acoustic wave equation of time inversion, the surface acoustic signal is reverse propagated and focused, thereby reconstructing two three-dimensional intensity images that correspond to two excitation modes and characterize the spatial distribution of the photoacoustic source.

[0024] S3, the data processing and imaging unit performs pixel-level difference and normalization operations on the two three-dimensional intensity images generated in S2, and calculates a photoacoustic property index for each voxel in the three-dimensional image. This index is used to quantitatively characterize the physical properties of the filling medium inside the crack.

[0025] S4, the data processing and imaging unit extracts the three-dimensional geometric model of the crack from the three-dimensional intensity image generated in S2, and accurately maps the photoacoustic property index calculated in S3 onto the three-dimensional geometric model in the form of color coding. Finally, it generates and outputs a fusion diagnostic model that integrates the three-dimensional spatial morphology and internal property distribution information of the crack.

[0026] Example 2: Based on Example 1, in S1, the synchronous excitation and acquisition process of the dual-mode photoacoustic signal is as follows: Figure 3 As shown, the details are as follows: The scanning and positioning system, along with the synchronous scanning control unit, drives the detection system to move to the preset scanning point on the concrete surface to be tested. This point is the common target point for the subsequent two laser excitations.

[0027] The first physical property-sensitive excitation and acquisition, synchronously scanning and control unit triggers the first laser in the dual-modal photoacoustic excitation unit, directing it towards the target point. emitting a wavelength of A short-pulse laser. The energy of this laser pulse is absorbed by the concrete surface and the near-surface medium, and converted into initial sound pressure according to the photoacoustic effect. The initial sound pressure generated by the photoacoustic effect. It can be described by the following formula: ; in: It is a spatial position vector.

[0028] It is the Green's Eisen coefficient, which is related to the thermodynamic properties of the medium.

[0029] It is the medium at wavelength The light absorption coefficient at that point.

[0030] It is the laser energy density at that point.

[0031] Due to the physical properties and sensitive wavelength The wavelength selected as the strong absorption wavelength for water is used when the laser irradiates a water-containing crack area; water molecules react with... The high absorption coefficient of the laser will result in a significantly enhanced initial sound pressure level in the region, thus encoding information about the presence of moisture in the acoustic signal originating from the crack geometry. Simultaneously with laser emission, the sensor array of the non-contact acoustic sensing unit begins recording its location under the control of a synchronous trigger signal. The surface vibration time-domain signal. Data acquisition lasts for a preset duration. Thus, the first set of original time-domain datasets was obtained. .

[0032] The second reference excitation and acquisition: After the first acquisition is completed, a short time interval of milliseconds is elapsed before the synchronous scan control unit scans at the same scan point. Trigger the second laser to emit a wavelength of Laser pulses. Due to water at wavelengths... absorption coefficient at Far lower than its absorption coefficient at If the crack contains water, the initial sound pressure generated at the same location by this excitation will be... This will be significantly reduced. If the crack is dry, the photoacoustic signals generated by the two excitations mainly originate from the concrete matrix, and their strength difference is not significant. The non-contact acoustic sensing unit also acquires and records the second set of original time-domain datasets under synchronous triggering. .

[0033] Scan iteration. Complete at scan point. After two excitations and acquisitions, the synchronous scanning control unit drives the detection system to move to the next preset scanning point. This process is repeated until the entire area to be measured is scanned and covered.

[0034] The dual-modal excitation process proposed in this invention actively utilizes the absorption differences of different media to specific spectra to encode and convert the originally invisible internal physical property information of cracks into two sets of measurable acoustic signals with significant differences. and This design allows the method to move beyond passively probing the geometric discontinuities of cracks, and instead actively and selectively examine the physical state inside the cracks, thus providing direct and reliable raw data input for the subsequent quantitative property inversion in S3.

[0035] Example 3: Building upon Example 2, before performing the three-dimensional sound source reconstruction based on reverse time migration in S2, a computational model corresponding to the concrete structure under test needs to be established for subsequent numerical simulations. This model establishment process includes discretizing the physical space and assigning acoustic properties to the discretized space.

[0036] A three-dimensional numerical computation grid is established. In the data processing and imaging unit, a three-dimensional Cartesian coordinate system is first defined, with a spatial extent sufficient to encompass the entire area under test and the positions of all sensors. Subsequently, this three-dimensional space is discretized along the X, Y, and Z directions, forming an orthogonal grid composed of a large number of cubic voxels. The position of any node in the grid... It can be represented as: ; in: These are the integer indices of the node in the X, Y, and Z directions, respectively.

[0037] , , It is the grid spacing along the three coordinate axes.

[0038] The choice of grid spacing needs to meet the stability and accuracy requirements of numerical simulation. It should generally be less than one-tenth of the wavelength corresponding to the highest frequency of the acoustic signal involved in the simulation. This three-dimensional numerical grid provides a discretized spatial basis for subsequent wave field numerical simulation.

[0039] Define the sound velocity field model for each node in the mesh. By assigning a sound speed value, a sound speed field model for the entire computational domain is constructed. .

[0040] Example 4: Based on Example 3, a uniform sound velocity model can be established. That is, it is assumed that the sound velocity throughout the entire concrete structure is uniform, and the sound velocity field is set as a constant. : ; This constant value The ultrasonic pulse velocity can be pre-determined by testing concrete samples from the same batch, or set according to the standard values ​​in relevant engineering materials handbooks. The method for obtaining this sound velocity value is a conventional technique in this field and will not be elaborated upon here.

[0041] Example 5: Building upon Example 3, a non-uniform sound velocity model can be established to improve imaging accuracy. If it is known that the concrete structure contains objects with different acoustic properties, such as reinforcing bars or embedded pipes, different sound velocity values ​​can be assigned to corresponding locations on the three-dimensional numerical grid based on the design drawings. For example, a higher sound velocity value for steel can be set at the grid node representing the reinforcing bars, while a background sound velocity value for concrete can be set in other areas. This non-uniform sound velocity model can more realistically reflect the propagation path of sound waves within complex structures, providing more accurate input for the inverse time migration algorithm.

[0042] Once the computational model and sound velocity field are established, the data processing and imaging unit will have the necessary conditions to perform wave field inverse time extrapolation calculations.

[0043] Example 6: Based on Example 5 or Example 4, after establishing the computational model and sound velocity field, this invention uses the reverse time migration method to process the two sets of original time-domain datasets collected in S1. and Wavefield inversion imaging is performed. This method utilizes the time invariance of the acoustic wave equation. By simulating the time reversal of the wavefield in a numerical model, the signals recorded by sensors distributed on the surface of the structure are accurately traced back and focused on their physical source, thereby obtaining a high-precision three-dimensional sound source image.

[0044] Wave field reverse-time extrapolation: This step aims to numerically solve a time-reversed acoustic wave equation driven by a surface-recorded signal, for any set of datasets. Define a backpropagating wave field The equation it satisfies is: ; in: In position Reverse time The reverse propagation sound pressure field.

[0045] It is a model of the sound velocity field.

[0046] It is the Laplace operator.

[0047] It is an artificial seismic source term loaded at the sensor location, defined as a time-reversed acquisition signal: ; At the sensor location The time-reversal signal injected at the location, This represents the total number of sensors.

[0048] For Dirac function.

[0049] The solution to this equation is from =0 starting from, until = End, initial condition is zero, that is and .

[0050] Example 7: Based on Example 6, the wave equation is solved using the finite difference time domain method, which transforms the sound pressure field on the three-dimensional computational grid. Discretize into Its second-order precision time and space difference scheme can be expressed as: ; in It is a discrete Laplace operator. It is a node The speed of sound, It is the time step. It is a discrete artificial source term. The specific implementation of the FDTD method is well known to those skilled in the art and will not be described in detail here.

[0051] Apply imaging conditions when the wave field from =0 propagates back to = Time (this moment corresponds to the initial moment of the physical world) =0), all wavefronts tracing back from different paths will coherently superimpose at their actual physical sound source locations, forming a maximum energy value. This invention uses zero-time imaging conditions to construct the final three-dimensional photoacoustic source image. Its mathematical expression is: ; In a discrete computational grid, the wavefield value at the last time step is taken as the image intensity value at that point, and this applies to all nodes within the computational domain. Perform this operation to obtain a complete 3D digital image.

[0052] By analyzing the original dataset and The processes are executed independently, ultimately generating two three-dimensional photoacoustic source intensity images. and Compared with conventional imaging algorithms such as synthetic aperture focusing, the reverse time migration method used in this invention can more accurately handle complex propagation phenomena such as multiple scattering and diffraction of sound waves in non-uniform media such as concrete. Therefore, it has a stronger imaging capability for irregular and complex crack networks, providing a more accurate spatial location and clearer morphology base image for subsequent physical property inversion.

[0053] Example 8: Based on Example 7, in S3, two three-dimensional photoacoustic source intensity images generated from S2 are... and This invention proposes a method for constructing and calculating a photoacoustic property index to extract the physical property information inside the crack. The purpose of this index is to eliminate common-mode interference in the measurement process through specific mathematical operations and to highlight the signal changes caused by the difference in the absorption rate of different wavelength lasers by the medium filling the crack.

[0054] This inversion process is achieved through the following steps: Calculate photoacoustic property indices. Data processing and imaging unit analyzes 3D images. and Perform voxel-to-voxel operations for each voxel position within the computation domain. Calculate its corresponding PPI value. In a specific embodiment, the photoacoustic property index... The calculation formula is as follows: ; in: At the voxel position The photoacoustic property index is calculated at [location].

[0055] At wavelength Under the incentive, in position The reconstructed photoacoustic source intensity value.

[0056] At wavelength Under the incentive, in position The reconstructed photoacoustic source intensity value.

[0057] It is a small regularization quantity introduced to ensure the stability of numerical calculations. It is an extremely small positive number used to prevent the denominator from being zero in weak signal regions.

[0058] In this formula, the numerator term Used to extract the difference signal generated by the difference in the spectral absorption characteristics of the medium. Denominator term As a normalization factor, its function is to eliminate or suppress interference caused by common-mode factors that simultaneously affect two measurements. These common-mode factors include, but are not limited to: successive fluctuations in laser pulse energy, changes in the distance between the probe and the concrete surface, and differences in local light absorption and acoustic coupling efficiency caused by variations in surface finish at different points.

[0059] Through the design of this ratio-type index, the calculated The value can be effectively decoupled from the changes in the above measurement conditions, thus making it primarily dependent on the measured point. The physical properties of the medium itself, specifically the difference in absorption coefficients for the two wavelengths of laser light, constitute the basis for establishing a stable and reliable quantitative index that reflects the physical properties of the filling material within the crack, providing a solid data foundation for subsequent property determination.

[0060] The photoacoustic property index was calculated. After obtaining the three-dimensional distribution map, a clear set of physical property discrimination criteria needs to be established to correspond the value with the specific physical state of the filling medium inside the crack.

[0061] The specific process for determining the physical property is as follows: The physical property discrimination criterion is applied. This criterion is based on the differences in the spectral absorption characteristics of different media. This is achieved by calculating... The value is compared with a preset threshold to classify the physical properties of each voxel.

[0062] In a specific embodiment, the discrimination criteria are as follows: For determining the presence of water: because water has a wavelength of... It has strong absorption at a certain wavelength, while at a certain wavelength... The absorption is relatively weak, and for water-bearing crack regions, the intensity of the reconstructed photoacoustic image meets the requirements. According to the formula for calculating the photoacoustic property index, this will lead to its The value approaches +1. Therefore, a high-order discrimination threshold can be set. (For example, =0.5). When the exponent value of a voxel satisfies > At that time, the region represented by the voxel was identified as being in a hydrated state.

[0063] Determining the dryness state: For dry, air-filled cracks, or intact concrete matrix, it is important to consider the following: and The difference in light absorption coefficients between the two wavelengths is very small, resulting in a small difference in the intensity of the photoacoustic signal. and Approximately equal, in this case, the numerator of the photoacoustic property index approaches zero, thus making The value also approaches 0. Therefore, a low-order discrimination threshold can be set. When the exponent value of a voxel satisfies At that time, the region represented by the voxel was identified as a dry state.

[0064] For the identification of other filling media: If the crack is filled with substances other than water and air (such as oil, chemical precipitates, etc.), as long as the substance is present... and It exhibits unique absorption spectral characteristics at wavelengths different from those of water. The value will then exhibit a characteristic value different from +1 or 0. For example, if a substance... Stronger absorption than ,That The value may be negative. Through prior calibration experiments, a richer database of correspondences between PPI values ​​and material types can be established, thereby enabling the identification of various filling media.

[0065] By applying the above discrimination criteria, the data processing and imaging unit can convert quantified data into numerical data. The three-dimensional distribution map is converted into a classification label map that characterizes the physical state of different regions inside the crack, providing clear physical property partitioning information for the final three-dimensional physical property diagnostic model generation.

[0066] Example 9: Based on Example 8, in S4, in order to generate the final three-dimensional physical property diagnostic model, it is first necessary to extract the three-dimensional geometric model of the crack from the three-dimensional photoacoustic source intensity image obtained in S2.

[0067] The extraction process is as follows: Extracting the three-dimensional geometric model of the crack, a process preferably performed at wavelengths sensitive to physical properties. Intensity image generated under excitation This is because the image typically has the highest crack signal-to-noise ratio and contrast, making it most suitable for accurate geometric shape recognition. The essence of the extraction process is to segment the three-dimensional intensity image into crack regions and background regions.

[0068] In one specific embodiment, the geometric model can be extracted using a threshold segmentation method. This method sets an intensity threshold. This identifies voxels in the image with an intensity higher than a certain threshold as part of a crack. Threshold The threshold can be determined manually based on observation of the image intensity histogram, or an automatic thresholding algorithm, such as Otsu's method, can be used. This algorithm automatically determines the optimal segmentation threshold by maximizing the inter-class variance. The implementation of this algorithm is well-known in the field and will not be described in detail here.

[0069] In another embodiment, a region growing method can be used for segmentation. This method first manually or automatically selects one or more seed points in the region identified as inside the crack in the image. Then, voxels adjacent to the seed points and meeting preset growth criteria are gradually incorporated into the crack region until no more voxels that meet the conditions can be added. This method has advantages for extracting crack structures with good connectivity.

[0070] To improve the accuracy of the geometric model, three-dimensional morphological filtering operations can be performed on the resulting binarized image after initial segmentation. For example, morphological opening operations can eliminate isolated bright spots caused by noise, and morphological closing operations can fill tiny holes inside cracks caused by weak signals.

[0071] After the above steps, a set of voxels labeled as cracks is finally obtained. This set constitutes a discretized three-dimensional geometric model of the crack, providing a geometric carrier for subsequent geometric-physical property information fusion.

[0072] Geometric property information fusion and color mapping: This step combines the calculated three-dimensional photoacoustic property index field map with the extracted three-dimensional geometric model of the crack. Perform precise spatial registration and fusion. Specifically, this involves: for each voxel constituting the crack geometry model... (in ∈ ), and its corresponding value in the three-dimensional PPI field map Assign it to the element as its physical property.

[0073] To transform abstract numerical attributes into intuitive visual information, this invention employs a color mapping method. This method establishes a color lookup table that defines photoacoustic property indices. The mapping relationship between the numerical range of values ​​and the color space. For example... Figure 2 As shown, in a specific embodiment, the mapping relationship can be set as follows: High-order discrimination threshold The PPI values ​​above are mapped to a specific color, such as blue, to visually represent water-bearing areas.

[0074] Low-order discrimination threshold The following PPI values ​​are mapped to a second specific color, such as red, to visually represent dry areas.

[0075] Will be between and The PPI values ​​between these values ​​are mapped to a transitional color from red to blue to represent intermediate states of different water saturation levels.

[0076] Through this color mapping process, each voxel in the three-dimensional geometric model of the crack is assigned a color value that reflects its internal physical state.

[0077] Visualization, rendering, and output of 3D physical property diagnostic models. The colored 3D geometric model is rendered to generate a final diagnostic model that can be used for human-computer interaction. This rendering process can be implemented using standard techniques in 3D computer graphics, the specific implementation of which is well-known in the field and will not be elaborated upon here.

[0078] In one embodiment, a volume rendering technique can be used, which can directly render the entire colored voxel dataset to generate an image with a semi-transparent effect, thereby enabling simultaneous observation of the surface morphology and internal physical property distribution of the crack.

[0079] In another embodiment, surface rendering technology can also be used, that is, firstly, an algorithm such as moving cube is used to extract isosurfaces from the voxel model of the crack as its surface mesh, and then the obtained color is used as a texture and applied to the surface mesh for rendering.

[0080] The final output 3D physical property diagnostic model can be rotated, scaled, and sectioned on a display device. The model's geometry intuitively reproduces the spatial orientation, depth, and connectivity of the crack, while the color distribution on its surface or inside quantitatively and clearly indicates the physical property state at different locations of the crack, achieving integrated diagnosis of the crack's geometry and internal physical properties.

[0081] The method for identifying concrete cracks in hydropower plant buildings of the present invention achieves quantitative diagnosis of the internal physical state of cracks through dual-modal excitation and differential analysis; it uses reverse time-shift imaging to ensure high-precision reconstruction of the three-dimensional morphology; and it combines non-contact measurement to improve the safety and convenience of detection.

Claims

1. A method for identifying concrete cracks in a hydropower station powerhouse, characterized in that, Includes the following steps: S1: Apply physical property sensitive excitation and reference excitation to the same target point of the concrete crack in a time-division manner, and simultaneously collect the first original time domain dataset and the second original time domain dataset generated by the two excitations respectively; S2: Perform reverse time-shift imaging independently on the first original time-domain dataset and the second original time-domain dataset to reconstruct the corresponding first three-dimensional photoacoustic source intensity image and the second three-dimensional photoacoustic source intensity image, respectively; S3: Based on the first three-dimensional photoacoustic source intensity image and the second three-dimensional photoacoustic source intensity image, calculate the photoacoustic property index at each location within the crack region; And based on the preset physical property discrimination criteria, the physical state of the filling medium inside the crack is determined; S4: Extract the three-dimensional geometric model of the crack, and map the physical state determined in S3 onto the three-dimensional geometric model to generate a three-dimensional physical property diagnostic model that integrates geometric and physical property information for crack identification.

2. The method for identifying concrete cracks in a hydropower station powerhouse according to claim 1, characterized in that, The laser wavelength used for the property-sensitive excitation is the strong absorption wavelength of water, while the laser wavelength used for the reference excitation is the weak absorption wavelength of water.

3. The method for identifying concrete cracks in a hydropower station powerhouse according to claim 2, characterized in that, The strong absorption wavelength of the water is 1450nm or 1940nm, and the weak absorption wavelength of the water is 1064nm or 532nm.

4. The method for identifying concrete cracks in a hydropower station powerhouse according to claim 1, characterized in that, The synchronous acquisition is achieved through a non-contact acoustic sensing unit, which is one or more laser Doppler vibrometers.

5. The method for identifying concrete cracks in a hydropower station powerhouse according to claim 1, characterized in that, The reverse time migration imaging includes: using the time-reversed original time-domain dataset as an artificial source term, substituting it into the time-reversed acoustic wave equation for numerical solution, in order to obtain the wave field propagating in the reverse direction; Then, at the last time step, the zero-time imaging condition is applied, and the wave field value at that time is used as the intensity value of the three-dimensional photoacoustic source intensity image.

6. The method for identifying concrete cracks in a hydropower station powerhouse according to claim 5, characterized in that, The numerical solution of the time-reversed acoustic wave equation is achieved through the finite difference time domain method.

7. The method for identifying concrete cracks in a hydropower station powerhouse according to claim 1, characterized in that, The photoacoustic property index is calculated as follows: Divide the difference between the intensity values ​​of the first three-dimensional photoacoustic source intensity image and the second three-dimensional photoacoustic source intensity image at the corresponding positions by the sum of their intensity values ​​at those positions.

8. The method for identifying concrete cracks in a hydropower station powerhouse according to claim 7, characterized in that, The physical property discrimination criteria include: setting a high-level discrimination threshold and a low-level discrimination threshold; When the photoacoustic property index value is greater than the high-level discrimination threshold, the physical state of the corresponding position is judged as a water-containing state; When the absolute value of the photoacoustic property index is less than or equal to the low-order discrimination threshold, the physical state of the corresponding position is judged as a dry state.

9. The method for identifying concrete cracks in a hydropower station powerhouse according to claim 1, characterized in that, The steps for generating the three-dimensional physical property diagnostic model specifically include: Based on the first three-dimensional photoacoustic source intensity image, the three-dimensional geometric model of the crack is extracted by threshold segmentation or region growing method; The physical state of the medium filling the crack is assigned to the corresponding position in the three-dimensional geometric model through color mapping.

10. The method for identifying concrete cracks in a hydropower station powerhouse according to claim 1, characterized in that, The reverse time migration imaging is based on a preset sound velocity field model, which is either a uniform sound velocity model or a non-uniform sound velocity model set according to the known distribution of objects inside the concrete structure.

Images