Acoustic monitoring signal processing method and related device for laser powder bed melting process

Abstract

This invention discloses a method and related apparatus for processing acoustic monitoring signals during laser powder bed melting. The method involves enhancing the high-frequency physical information data of the airborne acoustic emission signal using a pre-emphasis filter; performing a discrete Fourier transform on the enhanced airborne acoustic emission signal to obtain a spectral sequence signal; extracting key low-frequency and key high-frequency sequence signals from the spectral sequence signal and reconstructing them according to the original spectral sequence order to obtain a one-dimensional key spectral sequence signal; converting the one-dimensional key spectral sequence signal into a two-dimensional key spectral sequence grayscale image using an improved grayscale mapping method; and enhancing the grayscale mapping feature of the two-dimensional key spectral sequence grayscale image to obtain a key spectral sequence grayscale enhanced mapping image. This invention enables a richer and clearer expression of physically reasonable acoustic signal defect information during laser powder bed melting, thereby improving the reliability of defect information expression and the accuracy of defect feature description, and contributing to the accuracy of acoustic monitoring of defects in the laser powder bed melting process.

Description

Technical Field

[0001] This invention belongs to the field of metal additive manufacturing technology, specifically relating to an acoustic monitoring signal processing method and related device for laser powder bed melting process. Background Technology

[0002] Laser powder bed fusion (LPBF) additive manufacturing (AM), also known as selective laser melting (SLM), uses a high-power laser beam to selectively melt metal powder materials, building up layers to construct metal components designed in computer models. In recent years, this process has provided a creative paradigm for the design and production of complex and precision metal parts. Due to its unique advantages in manufacturing flexibility, material savings, development cycle time, complex structure forming, and microstructure properties, this technology has gained widespread attention and application in industries such as aerospace, defense, automotive, and biomedicine. Despite its significant advantages, this technology still faces serious challenges in terms of quality consistency and process repeatability. The rapid melting and solidification processes during the intense interaction between the laser and powder result in extremely complex molten pool dynamics. Unsuitable process parameters can lead to unstable molten pool flow, ultimately resulting in defects. In particular, the process is highly prone to generating unpredictable porosity defects, which severely limits its application range and hinders its wider industrial application. Real-time on-site process monitoring is a promising solution and a key to ensuring the success of additive manufacturing. Therefore, it is necessary to monitor the process on-site.

[0003] Acoustic monitoring technology has proven to be a viable solution and has attracted widespread attention due to the high sensitivity, low cost, and excellent equipment compatibility of its sensors. This technology can not only monitor the molten pool dynamics and volume dynamics of the laser powder bed melting process in real time, but also obtain information on complex physical processes such as melting, solidification, crack propagation, and porosity formation. It has been widely applied to quality monitoring in laser welding processes. Currently, research on acoustic monitoring methods in the field of laser additive manufacturing is limited. Because the laser additive manufacturing process is a layer-by-layer accumulation process with complex geometries far more complex than welding, the acoustic signals of the processes and defects involved are more complex, making related research more challenging. The acoustic signal generation mechanism of the laser powder bed melting process is very complex. Extracting features reflecting target attributes from the signals through signal processing methods is the most important part of LPBF process monitoring based on acoustic monitoring technology. Achieving reliable feature extraction through effective acoustic signal processing is key to establishing the correlation between air-propagated acoustic emission and laser powder bed melting defects, and a prerequisite for accurate monitoring of laser powder bed melting defects. Currently, acoustic monitoring of the LPBF process still faces the following problems: On the one hand, the LPBF acoustic monitoring scenario is very complex, making it difficult to directly obtain high-quality defect features from the original acoustic signals to achieve reliable defect differentiation; on the other hand, traditional time-frequency analysis methods cannot take into account both information features and the correlation mechanism between the LPBF process, defects, and signals, making it difficult to achieve an interpretable and accurate expression of defect feature information. Summary of the Invention

[0004] To address the problems existing in the prior art, this invention provides a method and related apparatus for acoustic monitoring signal processing in the laser powder bed melting process. This method can make the expression of physically reasonable acoustic signal defect information in the laser powder bed melting process richer and clearer, thereby improving the reliability of defect information expression and the accuracy of defect feature description, and contributing to the accuracy of acoustic monitoring of defects in the laser powder bed melting process.

[0005] To solve the above-mentioned technical problems, the present invention is achieved through the following technical solution:

[0006] According to a first aspect of the present invention, an acoustic monitoring signal processing method for a laser powder bed melting process is provided, comprising:

[0007] Acquire airborne acoustic emission signals generated during the laser powder bed fusion additive manufacturing process;

[0008] The airborne acoustic emission signal is enhanced with high-frequency physical information data by passing it through a pre-emphasis filter;

[0009] The airborne acoustic emission signal enhanced by the high-frequency physical information data is subjected to discrete Fourier transform to obtain the spectral sequence signal of the airborne acoustic emission signal;

[0010] The key low-frequency sequence signal and the key high-frequency sequence signal are extracted from the spectrum sequence signal, and the key low-frequency sequence signal and the key high-frequency sequence signal are reconstructed according to the original spectrum sequence order to obtain a one-dimensional key spectrum sequence signal.

[0011] The one-dimensional key spectrum sequence signal is converted into a two-dimensional key spectrum sequence grayscale image using an improved grayscale mapping method;

[0012] The grayscale mapping image of the two-dimensional key spectrum sequence is enhanced by grayscale mapping feature contrast to obtain a grayscale enhanced mapping image of the key spectrum sequence.

[0013] In one possible implementation of the first aspect, the extraction of key low-frequency sequence signals and key high-frequency sequence signals from the spectral sequence signal specifically includes:

[0014]

[0015] The process of reconstructing the key low-frequency sequence signal and the key high-frequency sequence signal according to the original spectral sequence order to obtain a one-dimensional key spectral sequence signal is as follows:

[0016]

[0017] In the formula, k start and k end The starting frequency f of the key low-frequency sequence signal or the key high-frequency sequence signal, respectively. start and termination frequency f end The corresponding sequence values; L is the key low-frequency sequence signal; H is the key high-frequency sequence signal; f s Z[n] is the sampling frequency; Z[n] is the one-dimensional key spectrum sequence signal with a time window length of n; Q is the data length of the key low-frequency sequence signal; M is the data length of the key high-frequency sequence signal.

[0018] In one possible implementation of the first aspect, converting the one-dimensional key spectrum sequence signal into a two-dimensional key spectrum sequence grayscale image using an improved grayscale mapping method includes:

[0019] The one-dimensional key spectral sequence signal is used to sequentially fill the rows of the converted grayscale matrix through a sliding window R and a spectral sequence scaling factor δ. Zero-padding is used to ensure that the matrix size is R×R, and normalization is performed to keep the pixel values ​​between 0 and 255. Specifically:

[0020]

[0021] In the formula, P(x,y) is the pixel value of the grayscale image coordinates (x,y) of the two-dimensional key spectrum sequence; round{·} is the rounding function.

[0022] In one possible implementation of the first aspect, the pre-emphasis filter is a first-order digital high-pass filter, specifically:

[0023] H[n] = X[n] - αX[n-1]

[0024] In the formula, X[n] is the airborne acoustic emission signal with a time window length of n; α is the pre-emphasis coefficient; H[n] is the airborne acoustic emission signal after high-frequency physical information data enhancement with a time window length of n.

[0025] In one possible implementation of the first aspect, the step of performing a discrete Fourier transform on the airborne acoustic emission signal enhanced with the high-frequency physical information data specifically involves:

[0026]

[0027] In the formula, k is the sequence value of the spectrum sequence signal X[k] of the airborne acoustic emission signal after high-frequency physical information data enhancement; H[n] is the airborne acoustic emission signal after high-frequency physical information data enhancement with a time window length of n; N is the length of H[n].

[0028] In one possible implementation of the first aspect, the frequency range corresponding to the key low-frequency sequence signal is 5kHz to 20kHz, and the frequency range corresponding to the key high-frequency sequence signal is 35kHz to 50kHz.

[0029] In one possible implementation of the first aspect, the grayscale mapping feature contrast enhancement of the two-dimensional key spectral sequence grayscale mapping image includes:

[0030] The Gamma algorithm is used to enhance the grayscale mapping features of the two-dimensional key spectral sequence grayscale mapping image, specifically as follows:

[0031]

[0032] In the formula, I in It is the pixel value of the grayscale mapping image of the two-dimensional key spectral sequence; I out γ is the grayscale pixel value of the enhanced image after grayscale mapping of the key spectral sequence; γ is the Gamma correction coefficient.

[0033] According to a second aspect of the present invention, an acoustic monitoring signal processing device for a laser powder bed melting process is provided, comprising:

[0034] The acquisition module is used to acquire the airborne acoustic emission signals generated during the laser powder bed fusion additive manufacturing process;

[0035] The high-frequency enhancement module is used to enhance the airborne acoustic emission signal with high-frequency physical information data through a pre-emphasis filter;

[0036] The transformation module is used to perform discrete Fourier transform on the enhanced airborne acoustic emission signal of the high-frequency physical information data to obtain the spectral sequence signal of the airborne acoustic emission signal.

[0037] The reconstruction module is used to extract key low-frequency sequence signals and key high-frequency sequence signals from the spectrum sequence signal, and reconstruct the key low-frequency sequence signals and key high-frequency sequence signals according to the original spectrum sequence order to obtain a one-dimensional key spectrum sequence signal.

[0038] The conversion module is used to convert the one-dimensional key spectrum sequence signal into a two-dimensional key spectrum sequence grayscale mapping image through an improved grayscale mapping method.

[0039] The contrast enhancement module is used to perform grayscale mapping feature contrast enhancement on the two-dimensional key spectrum sequence grayscale mapping image to obtain a key spectrum sequence grayscale enhanced mapping image.

[0040] According to a third aspect of the present invention, an apparatus is provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the aforementioned acoustic monitoring signal processing method for a laser powder bed melting process.

[0041] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the acoustic monitoring signal processing method for a laser powder bed melting process.

[0042] Compared with the prior art, the present invention has at least the following beneficial effects:

[0043] This invention provides a signal processing method for acoustic monitoring of the laser powder bed melting process. It combines acoustic information features with the LPBF process-defect-signal correlation mechanism to enhance high-frequency physical information data and reconstruct one-dimensional key spectral sequence signals. On the one hand, the high-frequency physical information data enhancement avoids the loss of key high-frequency information strongly correlated with defects in LPBF acoustic signals and the influence of noise. On the other hand, the one-dimensional key spectral sequence signal reconstruction driven by the physical mechanism achieves data simplification of the air-propagating acoustic emission signal spectral sequence, reducing information redundancy while enhancing the reliable representation of strongly correlated defect information, thus contributing to the interpretability and accurate expression of defect features. Simultaneously, this invention employs an improved grayscale enhancement mapping strategy to convert the one-dimensional key spectral sequence signal into a two-dimensional key spectral sequence grayscale enhancement mapping image that is highly efficient, possesses rich and physically reasonable defect information expression, and maps key defect strongly correlated features. This not only achieves prominent feedback of valuable correlation defect information between adjacent pixels in the horizontal and vertical directions through grayscale images based on the intuitive expression of physically reasonable defect features, but also realizes contrast enhancement of grayscale mapping features, which is beneficial to improving the accuracy of acoustic monitoring.

[0044] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in the specific embodiments of the present invention, the drawings used in the description of the specific embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0046] Figure 1 This is a flowchart of an acoustic monitoring signal processing method for a laser powder bed melting process according to an embodiment of the present invention;

[0047] Figure 2 The image shows the air-propagating acoustic emission signal and its spectrum during the laser powder bed melting process, collected by the air-propagating acoustic emission sensor in this embodiment of the invention.

[0048] Figure 3 This is a schematic diagram illustrating the process of reconstructing the key spectral sequence signal in an embodiment of the present invention;

[0049] Figure 4 This is a schematic diagram illustrating the conversion of grayscale images of key spectral sequences in an embodiment of the present invention;

[0050] Figure 5 The images shown are grayscale enhanced mapping images of key spectral sequences of different defect signals in embodiments of the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions 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, 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.

[0052] like Figure 1 As shown, this embodiment of the invention provides an acoustic monitoring signal processing method for a laser powder bed melting process, specifically including the following steps:

[0053] S1. Acquire the airborne acoustic emission signal generated during the laser powder bed fusion additive manufacturing process.

[0054] Specifically, an acoustic sensor array is installed near the laser powder bed fusion additive manufacturing equipment to capture and record the airborne acoustic emission signals generated throughout the additive manufacturing process. These airborne acoustic emission signals contain rich information about the molten pool dynamics and potential defect generation.

[0055] For example, the airborne acoustic emission signals generated during the laser powder bed fusion additive manufacturing process are collected by an AM4I air-coupled acoustic emission sensor manufactured by PAC Corporation of the United States.

[0056] S2. The airborne acoustic emission signal is enhanced with high-frequency physical information data by passing it through a pre-emphasis filter.

[0057] In other words, the collected airborne acoustic emission signal is input into a pre-emphasis filter. The purpose of the pre-emphasis filter is to enhance the high-frequency part of the airborne acoustic emission signal in order to highlight the high-frequency physical information that may be weak in the original signal but is important for defect identification.

[0058] In one embodiment, the pre-emphasis filter is implemented using a first-order digital high-pass filter, and its transfer function is as follows:

[0059] H[n] = X[n] - αX[n-1]

[0060] In the formula, X[n] is the airborne acoustic emission signal with a time window length of n; α is the pre-emphasis coefficient, and the value of α is 0.98; H[n] is the airborne acoustic emission signal after high-frequency physical information data enhancement with a time window length of n.

[0061] S3. Perform a discrete Fourier transform on the enhanced airborne acoustic emission signal based on the high-frequency physical information data to obtain the spectral sequence signal of the airborne acoustic emission signal.

[0062] In other words, the airborne acoustic emission signal after pre-emphasis processing is subjected to discrete Fourier transform to convert it from the time domain to the frequency domain, thus obtaining a spectral sequence signal.

[0063] Specifically, the formula for calculating the Discrete Fourier Transform is as follows:

[0064]

[0065] In the formula, k is the sequence value of the spectral sequence signal X[k] of the airborne acoustic emission signal after high-frequency physical information data enhancement; N is the length of H[n].

[0066] S4. Extract key low-frequency sequence signals and key high-frequency sequence signals from the spectrum sequence signals, and reconstruct the key low-frequency sequence signals and key high-frequency sequence signals according to the original spectrum sequence order to obtain a one-dimensional key spectrum sequence signal.

[0067] In this embodiment, the frequency range corresponding to the key low-frequency sequence signal is 5kHz to 20kHz, and the frequency range corresponding to the key high-frequency sequence signal is 35kHz to 50kHz.

[0068] Specifically, by analyzing the frequency characteristics of the obtained spectral sequence signals, key low-frequency and high-frequency sequence signals that are highly correlated with molten pool stability and defect generation are identified. Subsequently, these key low-frequency and high-frequency sequence signals are reconstructed according to the order of the original spectral sequence to form a one-dimensional key spectral sequence signal, accurately identifying the spectral components that are important for defect identification.

[0069] In one embodiment, key low-frequency sequence signals and key high-frequency sequence signals are extracted from the spectral sequence signal, as follows:

[0070]

[0071]

[0072]

[0073] In the formula, k start and k end The starting frequency f of the key low-frequency sequence signal or the key high-frequency sequence signal, respectively. start and termination frequency f end The corresponding sequence values; L is the key low-frequency sequence signal; H is the key high-frequency sequence signal; f s It is the sampling frequency.

[0074] In one embodiment, the key low-frequency sequence signal and the key high-frequency sequence signal are reconstructed according to the original spectral sequence order to obtain a one-dimensional key spectral sequence signal, specifically as follows:

[0075]

[0076] In the formula, Z[n] is a one-dimensional key spectral sequence signal with a time window length of n; Q is the data length of the key low-frequency sequence signal; and M is the data length of the key high-frequency sequence signal.

[0077] S5. The one-dimensional key spectrum sequence signal is converted into a two-dimensional key spectrum sequence grayscale image using an improved grayscale mapping method. The two-dimensional key spectrum sequence grayscale image not only retains the important information in the original signal, but also displays the spectral characteristics of the signal in an intuitive way, which is beneficial for subsequent defect identification and monitoring.

[0078] Specifically, the improved grayscale mapping method involves filling the rows of the converted grayscale matrix sequentially with the one-dimensional key spectral sequence signal through a sliding window R and a spectral sequence scaling factor δ. Zero-padding is used to ensure the matrix size is R×R, and normalization is performed to keep the pixel values ​​consistently between 0 and 255. Its mathematical expression is as follows:

[0079]

[0080] In the formula, P(x,y) is the pixel value of the grayscale image coordinates (x,y) of the two-dimensional key spectrum sequence; round{·} is the rounding function.

[0081] More specifically, a sliding window of size R is defined. This window slides across the one-dimensional key spectrum sequence signal, selecting signal values ​​within the window to fill the grayscale matrix with each slide. Simultaneously, a spectrum sequence scaling factor δ is defined to adjust the scaling relationship when mapping from the one-dimensional key spectrum sequence signal to the two-dimensional key spectrum sequence grayscale image, ensuring that the signal information is displayed appropriately in the image. The sliding window iterates through the one-dimensional key spectrum sequence signal, sequentially filling the corresponding rows of the grayscale matrix with signal values ​​within the window. If the length of the one-dimensional key spectrum sequence signal is insufficient to completely fill the entire grayscale matrix, zero-padding (adding zero values ​​to the end of the grayscale matrix) is used to ensure that the grayscale matrix reaches the preset size of R×R.

[0082] Since the amplitude range of a one-dimensional key spectral sequence signal may exceed the pixel value range of 0-255, the grayscale matrix is ​​normalized. The purpose of normalization is to map each pixel value in the grayscale matrix to the range of 0-255 to facilitate subsequent image display and processing.

[0083] S6. Perform grayscale mapping feature contrast enhancement on the two-dimensional key spectrum sequence grayscale mapping image to obtain a key spectrum sequence grayscale enhanced mapping image.

[0084] In other words, grayscale mapping feature contrast enhancement processing is performed on the obtained two-dimensional key spectrum sequence grayscale mapping image to enhance the key feature information in the image.

[0085] Preferably, the Gamma algorithm is used to enhance the grayscale mapping features of the two-dimensional key spectrum sequence grayscale mapping image, specifically as follows:

[0086]

[0087] In the formula, I in It is the pixel value of the grayscale mapping image of the two-dimensional key spectral sequence; I out γ is the grayscale pixel value of the enhanced mapped image after grayscale transformation of the key spectral sequence through Gamma grayscale mapping; γ is the Gamma correction coefficient, which determines the degree of nonlinearity of the mapping.

[0088] Example:

[0089] In this embodiment, an airborne acoustic emission sensor was used to collect the airborne acoustic emission signals generated during the laser powder bed melting experiment. 316L stainless steel powder was used as the original metal powder material for printing, and eight sets of pore defect simulation parts with different porosities were set up. The defect simulation parts had dimensions of 10mm × 10mm × 10mm. During the experiment, the data acquisition frequency of the airborne acoustic emission sensor was set to 100kHz, and the sensor was installed in a suitable position for acoustic signal acquisition.

[0090] In this embodiment, according to step S1 of the present invention, an AM4I air-coupled acoustic emission sensor manufactured by PAC Corporation of the United States is used to collect airborne acoustic emission signals generated during the laser powder bed fusion additive manufacturing process. The time-domain waveform and spectrum of the collected airborne acoustic emission signals are shown below. Figure 2 As shown. From Figure 2 It can be found that the spectral information of the defect signal is mainly distributed in the critical low-frequency band of 5-20kHz and the critical high-frequency band of 35-50kHz. The information in the critical high-frequency band can effectively reflect the dynamic changes of the defect.

[0091] Following step S2 above, the airborne acoustic emission signal from the laser powder bed melting process, acquired by a sensor with a window length n of 4096, is enhanced with high-frequency physical information data using a pre-emphasis filter with a pre-emphasis coefficient α = 0.98, resulting in an enhanced airborne acoustic emission signal sample. This high-frequency physical information data enhancement effectively avoids the loss of information on high-frequency components of signals with high attenuation characteristics during airborne propagation and reduces the potential adverse effects of low-frequency noise on defect monitoring.

[0092] Following step S3 above, the high-frequency physical information data with a length of 4096 is enhanced and subjected to discrete Fourier transform to obtain the spectral sequence signal of the airborne acoustic emission sample signal.

[0093] Following step S4 above, as Figure 3 As shown, a key low-frequency sequence signal L (frequency range 5kHz–20kHz) and a key high-frequency sequence signal H (frequency range 35kHz–50kHz) are extracted from the spectral sequence of the airborne acoustic emission (LPBF) sample signal. These are then reconstructed according to the spectral sequence order to obtain a one-dimensional key spectral sequence signal Z. The constructed key spectral sequence signal not only effectively preserves the internal data characteristics and original spectral information structure of the LPBF process airborne acoustic emission signal, but also achieves data simplification of the airborne acoustic emission signal spectral sequence based on physically reasonable LPBF acoustic signal defect information expression characteristics. This reduces information redundancy while enabling interpretable expression of strongly related defect information.

[0094] Following step S5 above, as follows Figure 4 As shown, the one-dimensional key spectrum sequence signal is sequentially filled into the rows of the transformed grayscale matrix under the control of a sliding window R=224 and a spectrum sequence scale factor δ=4. Zero-padding is used to ensure that the size of the grayscale matrix is ​​224×224, and normalization calculation is used to ensure that the grayscale pixel values ​​are always kept between 0-255, finally obtaining a two-dimensional key spectrum sequence grayscale mapping image. This operation realizes the correlation representation of the key spectrum sequence in different gradient directions, which is conducive to a more comprehensive description of the complex changes of different defect signal characteristics. On the one hand, the horizontal gradient direction has a fixed frequency scale, which can effectively reflect the local dynamic change characteristics of the key spectrum sequence within the detail frequency range. On the other hand, the spectrum in the vertical gradient direction realizes the dimensional leap from low frequency to high frequency, which can effectively reflect the key frequency component characteristics most relevant to the molten pool sound. Through the image representation method of data reuse, the loss of defect feature expression caused by signal distortion or insufficient feature information is minimized.

[0095] Following step S6 above, the grayscale mapping image of the two-dimensional key spectrum sequence is enhanced by grayscale mapping feature contrast using the Gamma algorithm to obtain the grayscale enhanced mapping image of the key spectrum sequence, where the Gamma correction coefficient γ is set to 0.6. The nonlinear brightness adjustment method provided by the Gamma algorithm can optimize the visual features of the grayscale image as a carrier of defect information, which helps to improve the recognition accuracy of the deep learning model for defect monitoring. Figure 5 The grayscale enhanced mapping images of key spectral sequences of different defect signals are shown. It can be found that the grayscale enhanced mapping images of key spectral sequences provide a clear and detailed characterization of the information aggregation features of key frequency bands strongly correlated with defects, and realize the expression of the local dynamic differences of key spectral sequence information related to defects.

[0096] In this embodiment, based on the grayscale enhanced mapping images of key spectral sequences of different defect signals, a dataset is constructed containing a training set of 70% key spectral sequence grayscale enhanced mapping images and a test set of 30% key spectral sequence grayscale enhanced mapping images. AlexNet and MS-CNN models are selected as the deep learning models for defect detection. The training set containing the 70% key spectral sequence grayscale enhanced mapping images is used for training the deep learning model, and the test set containing the 30% key spectral sequence grayscale enhanced mapping images is used for validation, resulting in the final deep learning model for defect detection.

[0097] Table 1 shows the defect identification accuracy results using different deep learning models for defect monitoring in this embodiment. It can be seen that the AlexNet and MS-CNN models achieve identification accuracy of over 97% for all eight types of defects, and the single-sample response time of the models is less than 1ms, demonstrating excellent defect monitoring performance.

[0098] Table 1. Defect identification accuracy results of different deep learning models for defect monitoring.

[0099]

[0100] It is evident that, after processing by the acoustic monitoring signal processing method for the laser powder bed melting process proposed in this invention, the acoustic signals of the laser powder bed melting process can be efficiently applied to the reliable differentiation of pore defects in the complex working conditions of the laser powder bed melting process. This achieves the interpretable and accurate expression of the acoustic signal defect feature information of the laser powder bed melting process, and, combined with a deep learning model, completes high-accuracy in-situ acoustic monitoring of defects in the laser powder bed melting additive manufacturing process, demonstrating good engineering applicability.

[0101] In another embodiment of the present invention, an acoustic monitoring signal processing device for laser powder bed melting process is provided, specifically including the following modules:

[0102] The acquisition module is used to acquire the airborne acoustic emission signals generated during the laser powder bed fusion additive manufacturing process.

[0103] The high-frequency enhancement module is used to enhance the airborne acoustic emission signal with high-frequency physical information data through a pre-emphasis filter.

[0104] The transformation module is used to perform discrete Fourier transform on the enhanced airborne acoustic emission signal of the high-frequency physical information data to obtain the spectral sequence signal of the airborne acoustic emission signal.

[0105] The reconstruction module is used to extract key low-frequency sequence signals and key high-frequency sequence signals from the spectrum sequence signal, and reconstruct the key low-frequency sequence signals and key high-frequency sequence signals according to the original spectrum sequence order to obtain a one-dimensional key spectrum sequence signal.

[0106] The conversion module is used to convert the one-dimensional key spectrum sequence signal into a two-dimensional key spectrum sequence grayscale image using an improved grayscale mapping method.

[0107] The contrast enhancement module is used to perform grayscale mapping feature contrast enhancement on the two-dimensional key spectrum sequence grayscale mapping image to obtain a key spectrum sequence grayscale enhanced mapping image.

[0108] All relevant content regarding the steps involved in the aforementioned embodiments of the acoustic monitoring signal processing method for laser powder bed melting process can be referenced to the functional description of the corresponding functional module of the acoustic monitoring signal processing device for laser powder bed melting process in the embodiments of the present invention, and will not be repeated here. The module division in the embodiments of the present invention is illustrative and is merely a logical functional division. In actual implementation, there may be other division methods. Furthermore, the functional modules in the various embodiments of the present invention can be integrated into a processor, exist separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0109] In another embodiment of the present invention, a computer device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions in the computer storage medium to achieve a corresponding method flow or corresponding function. The processor described in this embodiment of the present invention can be used in the operation of an acoustic monitoring signal processing method for a laser powder bed melting process.

[0110] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory), which is a memory device in a computer device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the computer device and extended storage media supported by the computer device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, the storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory or a non-volatile memory, such as at least one disk storage device. The processor can load and execute one or more instructions stored in the computer-readable storage medium to implement the corresponding steps of the acoustic monitoring signal processing method for a laser powder bed melting process in the above embodiments.

[0111] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0112] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0113] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0114] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0115] In this invention, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0116] Finally, it should be noted that the above-described embodiments are merely specific implementations of the present invention, used to illustrate the technical solutions of the present invention, and not to limit it. The scope of protection of the present invention is not limited thereto. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that any person skilled in the art can still modify or easily conceive of changes to the technical solutions described in the foregoing embodiments within the technical scope disclosed in the present invention, or make equivalent substitutions for some of the technical features; and these modifications, changes, or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for acoustic monitoring signal processing in a laser powder bed melting process, characterized in that, include: Acquire airborne acoustic emission signals generated during the laser powder bed fusion additive manufacturing process; The airborne acoustic emission signal is enhanced with high-frequency physical information data by passing it through a pre-emphasis filter; The airborne acoustic emission signal enhanced by the high-frequency physical information data is subjected to discrete Fourier transform to obtain the spectral sequence signal of the airborne acoustic emission signal; The key low-frequency sequence signal and the key high-frequency sequence signal are extracted from the spectrum sequence signal, and the key low-frequency sequence signal and the key high-frequency sequence signal are reconstructed according to the original spectrum sequence order to obtain a one-dimensional key spectrum sequence signal. The one-dimensional key spectrum sequence signal is converted into a two-dimensional key spectrum sequence grayscale image using an improved grayscale mapping method; The grayscale mapping image of the two-dimensional key spectrum sequence is enhanced by grayscale mapping feature contrast to obtain a grayscale enhanced mapping image of the key spectrum sequence.

2. The acoustic monitoring signal processing method for laser powder bed melting process according to claim 1, characterized in that, The extraction of key low-frequency sequence signals and key high-frequency sequence signals from the spectral sequence signal specifically involves: The process of reconstructing the key low-frequency sequence signal and the key high-frequency sequence signal according to the original spectral sequence order to obtain a one-dimensional key spectral sequence signal is as follows: In the formula, k start and k end The starting frequency f of the key low-frequency sequence signal or the key high-frequency sequence signal, respectively. start and termination frequency f end The corresponding sequence values; L is the key low-frequency sequence signal; H is the key high-frequency sequence signal; f s Z[n] is the sampling frequency; Z[n] is the one-dimensional key spectrum sequence signal with a time window length of n; Q is the data length of the key low-frequency sequence signal; M is the data length of the key high-frequency sequence signal.

3. The acoustic monitoring signal processing method for laser powder bed melting process according to claim 2, characterized in that, The step of converting the one-dimensional key spectrum sequence signal into a two-dimensional key spectrum sequence grayscale image using an improved grayscale mapping method includes: The one-dimensional key spectral sequence signal is used to sequentially fill the rows of the converted grayscale matrix through a sliding window R and a spectral sequence scaling factor δ. Zero-padding is used to ensure that the matrix size is R×R, and normalization is performed to keep the pixel values ​​between 0 and 255. Specifically: In the formula, P(x,y) is the pixel value of the grayscale image coordinates (x,y) of the two-dimensional key spectrum sequence; round{·} is the rounding function.

4. The acoustic monitoring signal processing method for laser powder bed melting process according to claim 1, characterized in that, The pre-emphasis filter is a first-order digital high-pass filter, specifically: H[n] = X[n] - αX[n-1] In the formula, X[n] is the airborne acoustic emission signal with a time window length of n; α is the pre-emphasis coefficient; H[n] is the airborne acoustic emission signal after high-frequency physical information data enhancement with a time window length of n.

5. The acoustic monitoring signal processing method for laser powder bed melting process according to claim 1, characterized in that, The step of performing a discrete Fourier transform on the enhanced airborne acoustic emission signal based on the high-frequency physical information data is as follows: In the formula, k is the sequence value of the spectrum sequence signal X[k] of the airborne acoustic emission signal after high-frequency physical information data enhancement; H[n] is the airborne acoustic emission signal after high-frequency physical information data enhancement with a time window length of n; N is the length of H[n].

6. The acoustic monitoring signal processing method for laser powder bed melting process according to claim 1, characterized in that, The key low-frequency sequence signal corresponds to a frequency range of 5kHz to 20kHz, and the key high-frequency sequence signal corresponds to a frequency range of 35kHz to 50kHz.

7. The acoustic monitoring signal processing method for laser powder bed melting process according to claim 1, characterized in that, The step of enhancing the grayscale mapping feature of the two-dimensional key spectrum sequence grayscale mapping image includes: The Gamma algorithm is used to enhance the grayscale mapping features of the two-dimensional key spectral sequence grayscale mapping image, specifically as follows: In the formula, I in It is the pixel value of the grayscale mapping image of the two-dimensional key spectral sequence; I out γ is the grayscale pixel value of the enhanced image after grayscale mapping of the key spectral sequence; γ is the Gamma correction coefficient.

8. An acoustic monitoring signal processing device for laser powder bed melting process, characterized in that, include: The acquisition module is used to acquire the airborne acoustic emission signals generated during the laser powder bed fusion additive manufacturing process; The high-frequency enhancement module is used to enhance the airborne acoustic emission signal with high-frequency physical information data through a pre-emphasis filter; The transformation module is used to perform discrete Fourier transform on the enhanced airborne acoustic emission signal of the high-frequency physical information data to obtain the spectral sequence signal of the airborne acoustic emission signal. The reconstruction module is used to extract key low-frequency sequence signals and key high-frequency sequence signals from the spectrum sequence signal, and reconstruct the key low-frequency sequence signals and key high-frequency sequence signals according to the original spectrum sequence order to obtain a one-dimensional key spectrum sequence signal. The conversion module is used to convert the one-dimensional key spectrum sequence signal into a two-dimensional key spectrum sequence grayscale mapping image through an improved grayscale mapping method. The contrast enhancement module is used to perform grayscale mapping feature contrast enhancement on the two-dimensional key spectrum sequence grayscale mapping image to obtain a key spectrum sequence grayscale enhanced mapping image.

9. An apparatus comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the acoustic monitoring signal processing method for laser powder bed melting process as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the acoustic monitoring signal processing method for laser powder bed melting process as described in any one of claims 1 to 7.

Images