Method and device for real-time imaging monitoring of 3D printing interface quality

Abstract

The invention discloses a method and device for real-time imaging monitoring of the quality of a 3D printing interface, and belongs to the field of additive manufacturing. A printing area is divided into uniform grids, and an ultrasonic sensor array is arranged on the edge of a bottom plate; lamb waves without deposition are collected as reference signals, and a calibration curve containing unit length reference energy attenuation coefficients corresponding to different layer heights and variation of the unit length reference energy attenuation coefficients is established by printing a defect-free standard test piece layer by layer and taking an A0 modal energy value as a standard. Performing full-path scanning at a fixed time interval during printing, reconstructing an energy attenuation coefficient matrix per unit length of each grid, and generating a real-time monitoring image; according to the calibration curve, the energy attenuation coefficient variation per unit length of each grid in the monitoring image is analyzed, the defect type and position are judged, and whether printing is stopped or not is determined according to conditions. According to the method, real-time imaging can be carried out on the interface quality of a printed piece, and different types of defects such as bottom debonding / warping and interlayer stripping with specific layer height can be accurately distinguished.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of additive manufacturing, and particularly relates to a method and device for real-time imaging monitoring of 3D printing interface quality. BACKGROUND

[0002] Additive manufacturing (AM) is a rapidly developing manufacturing technology, which is different from the subtractive process in traditional machining process. AM generates corresponding workpieces by adding materials layer by layer, has high design freedom and no material waste. Fused deposition modeling (FDM) is one of the most widely used AM technologies. Due to its relatively low cost and easy use, it is often used for small batch production and daily use environment. In the deposition process, the 3D printer automatically feeds the thermoplastic material filament into the extruder, and the extruder follows the printing path defined by the slicing algorithm to print layer by layer. However, material breakage, nozzle deviation, unstable temperature control and other problems often occur during FDM printing, resulting in product scrap. Real-time monitoring of the printing process and early detection of defects are crucial to improve product quality and production efficiency.

[0003] Ultrasonic non-destructive testing (NDT) has the ability of internal detection without damage and high sensitivity, and is introduced into the monitoring of 3D printing process. The invention patent with publication number CN112229911A fixes a pair of ultrasonic probes on both sides of the printing base plate. When the "delamination" (i.e. debonding) defect occurs between the printed part and the base plate, some characteristics of the ultrasonic signal will change. The system determines whether the delamination occurs by monitoring this change. Once the delamination occurs, the printing can be stopped immediately, saving materials and improving printing efficiency.

[0004] The above prior art solution has obvious limitations. First, since only a pair of probes is used, its monitoring range covers the entire printing area, and it cannot accurately locate the specific position where the defect occurs. The staff still needs to spend time to investigate the problem. Second, in order to avoid false positives caused by single-point measurement, the technology needs to continue printing and monitor the signals of the subsequent layers to make a comprehensive judgment after initially determining the risk of delamination, which leads to a lag in feedback and cannot respond in the first time when the defect occurs. In addition, this solution is single-functioned and can only detect the debonding between the printed part and the base plate, and cannot identify the more common interlayer peeling defects inside the printed part. Therefore, there is an urgent need in the field for a monitoring method that can accurately locate and classify the printing defects in real time, full field and accurate positioning, in order to overcome the shortcomings of the prior art. SUMMARY

[0005] In view of the above problems, the present application provides a method and device for real-time imaging monitoring of 3D printing interface quality, which can achieve real-time imaging of the interface quality of the printed part and positioning of possible defect positions by arranging an ultrasonic sensor array on the edge of the printing base plate. The same defect will be monitored by multiple pairs of sensors, and the possibility of misjudgment is greatly reduced, so that real-time closed-loop feedback can be achieved.

[0006] The technical solutions adopted by the present application are as follows: In a first aspect, the present application provides a method for real-time imaging monitoring of 3D printing interface quality, comprising the following steps: (1) defining an imaging area on the printing base plate and dividing the imaging area into multiple uniform grids, arranging an ultrasonic sensor array on the edge of the printing base plate according to the grid division of the imaging area and calculating an ultrasonic propagation path length matrix, the elements of the matrix representing the length of each ultrasonic propagation path passing through each grid; (2) collecting a reference ultrasonic signal propagating in the form of Lamb wave through the ultrasonic sensor array when there is no deposition; subsequently, printing a standard pattern without defects on the printing base plate and collecting an ultrasonic signal propagating in the form of Lamb wave again, calculating the reference energy attenuation coefficient of each ultrasonic propagation path based on the A0 mode energy values of the two collected signals; by printing a standard pattern with different layer heights, a calibration curve containing the unit length reference energy attenuation coefficient corresponding to different layer heights and its change amount is established; (3) during the printing process, performing full-path ultrasonic scanning at a fixed time interval through the ultrasonic sensor array, collecting real-time ultrasonic signals and calculating a real-time energy attenuation coefficient matrix, using an image reconstruction algorithm to invert the unit length energy attenuation coefficient matrix of each grid in the imaging area, and generating a real-time monitoring image; (4) determining the type and position of the defect based on the change amount of the unit length energy attenuation coefficient of each grid in the real-time monitoring image and according to the calibration curve, and deciding whether to continue printing according to the comparison result of the number of grids with defects and the preset threshold.

[0007] Further, the ultrasonic sensor array comprises u ultrasonic transmitting probes and v ultrasonic receiving probes, where u and v are both integers greater than 1; the u ultrasonic transmitting probes and the v ultrasonic receiving probes work in a full combination manner to form u×v ultrasonic propagation paths; the arrangement mode of the sensor array enables the u×v ultrasonic propagation paths to cover all grids in the imaging area.

[0008] Further, the calculation of the reference energy attenuation coefficient of each ultrasonic propagation path based on the two collected signals specifically comprises: For each ultrasonic propagation path i, the reference energy attenuation coefficient thereof is calculated according to the following formula : ; wherein, is the center frequency of the ultrasonic sensor, is the ultrasonic signal waveform energy value of path i without deposition, is the ultrasonic signal waveform energy value of path i after printing the standard pattern.

[0009] Further, the waveform energy value is the A0 mode energy value in the received Lamb wave signal.

[0010] Further, the establishment includes a calibration curve of the unit length reference energy attenuation coefficient corresponding to different layer heights and the change amount thereof, which comprises: By printing a k-layer defect-free pattern (k is an integer greater than or equal to 1), the unit length reference energy attenuation coefficient of each path after the k-layer printing is obtained : ; wherein, is the reference energy attenuation coefficient of path i after the k-layer printing, is the length of path i passing through the k-layer printed material; The of all paths is taken to obtain the unit length reference energy attenuation coefficient of the layer ; based on the under different k values, the change amount of the unit length reference energy attenuation coefficient is calculated , wherein, , ( ), thereby forming the calibration curve.

[0011] Further, it further comprises the step of determining the system detection capability threshold: When printing patterns of different layer heights k, record the sum of the energy attenuation coefficients of each path obtained by full-path data acquisition , wherein is the sum of the energy attenuation coefficients of the k-layer; When is satisfied, it is determined that the system loses effective detection capability for the newly deposited material above the p-layer, and the p-layer is taken as the threshold layer for defect determination, and the calibration curve is only the calibration result in the p-layer.

[0012] Further, the unit length energy attenuation coefficient distribution of each grid in the imaging area is inversed by using an image reconstruction algorithm, specifically: The imaging problem is constructed as a linear equation set Wherein, L is the ultrasonic propagation path length matrix, Z is the real-time path energy attenuation coefficient matrix, A is the grid unit length energy attenuation coefficient matrix to be solved. An iterative reconstruction algorithm is used to solve the matrix A, thereby reconstructing the real-time monitoring image of the imaging area.

[0013] Further, in step (4), the process of determining the type and position of the defect comprises: The grid unit length energy attenuation coefficient change amount of each grid in the real-time monitoring image is compared with the change amount of the unit length reference energy attenuation coefficient in the calibration curve Comparison, 1≤k≤q≤p, q is the current printing layer, p is the threshold number of layers that makes the system lose effective detection capability of the newly added deposit: If , the deposit bottom at the jth grid is determined to have debonding and warping; If , 2≤k≤q, the internal kth layer of the deposit at the jth grid is determined to have interlayer peeling; If , further overall determination is made on the jth grid, and when or , it is determined that the deposit at the jth grid has a defect, otherwise the deposit at the jth grid has no defect; Wherein, is the grid side length, is a preset threshold parameter, is the unit length energy attenuation coefficient of the grid j, is the unit length reference energy attenuation coefficient of the kth layer, is the change amount of the unit length reference energy attenuation coefficient of the first layer, and are the change amounts of the unit length reference energy attenuation coefficients of the kth layer and the k-1th layer, respectively.

[0014] Further, in step (4), after determining the type and position of the defect each time, the number of defect grids Ne in the n grids of the imaging area is counted, and if the proportion of the number of defect grids exceeds a preset proportion, the printing is ended.

[0015] In a second aspect, the present application provides a device for real-time imaging monitoring of 3D printing interface quality, applied to an FDM printer, for realizing the above-mentioned method for real-time imaging monitoring of 3D printing interface quality.

[0016] The present application has the beneficial effects that: The present application discretizes the printing area into a fine grid in advance, and uses an ultrasonic sensor array arranged by optimization to form a measurement network. By solving the imaging equation, the energy attenuation coefficient distribution image of the entire printing area can be reconstructed in near real time. Not only can the presence or absence of defects in each grid be detected, but also different types of defects such as bottom debonding / warping and interlayer peeling at a specific layer height can be accurately distinguished by analyzing the difference between the current image and the historical image and relying on the calibration curve established in advance, which contains different layer height attenuation characteristics. The system provides clear fault diagnosis information for the operator, far beyond the limitations of traditional methods that can only alarm.

[0017] In addition, the system obtains reference attenuation data under different printing layers through calibration, and can automatically judge the saturation point of the monitoring capability, which enables the system to adapt to different printing materials and parameters, and to recognize its monitoring limit, thereby avoiding false judgments under invalid monitoring and ensuring the reliability of the judgment. By setting a defect grid proportion threshold, the system can also automatically judge the severity of the defect and make intelligent decisions to continue printing or pause and alarm, greatly reducing the dependence on manpower and improving the automation level and efficiency of the production process. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 is the device structure diagram of the real-time imaging monitoring of the 3D printing interface quality proposed by the present application; Figure 2 is a waveform example collected by the device for real-time imaging monitoring of the 3D printing interface quality proposed by the present application; Figure 3 is a flowchart of the method for real-time imaging monitoring of the 3D printing interface quality proposed by the present application. DETAILED DESCRIPTION

[0019] The present application will be further described and explained with specific embodiments. The embodiments are only exemplary and do not limit the scope of the disclosure. The technical features of each embodiment in the present application can be combined accordingly without conflict.

[0020] In order to realize real-time, full-field imaging monitoring of the printing deposition layer, the present application sets up a device for real-time imaging monitoring of the 3D printing interface quality as shown in Figure 1 The mechanical components of the device mainly include the following: An FDM printer; as the main monitored equipment, it is responsible for performing regular printing tasks, including material melting, path planning and layer-by-layer deposition, and its core components include a printing base plate (build platform), an extruder, a nozzle and a three-dimensional motion control system, etc.

[0021] An ultrasonic sensor array is fixed above the printing plate. As the core sensing unit, it consists of multiple independent transmitting and receiving probes mounted in a specific array on the edge of the printing plate. The array can be arranged linearly, in an L-shape, or around the entire plate, ensuring that the ultrasonic propagation path formed between the probes fully covers the entire effective printing area. The sensors are securely coupled to the upper or side surface of the plate edge using mechanical clamps or a special adhesive, ensuring efficient excitation and reception of the ultrasonic signals.

[0022] The ultrasonic signal excitation and acquisition system integrates signal generation, power amplification, signal acquisition, and preliminary conditioning functions. It consists of a multi-channel ultrasonic signal card or a combination of a signal generator and an acquisition card. It is connected to the ultrasonic sensor array through a customized adapter board. The adapter board is used to adapt the signal generator to the ultrasonic sensor array. It automatically switches the combination of excitation and receiving sensors according to a predetermined timing logic, thereby efficiently traversing all ultrasonic propagation paths and controlling their transmission and reception.

[0023] The data processing and control system consists of a high-performance computer connected to the ultrasonic signal excitation and acquisition system via a high-speed data bus. It is responsible for sending commands to the ultrasonic signal excitation and acquisition system, controlling the timing of data acquisition, channel switching, and parameter settings, and receiving the acquired raw ultrasonic signal data to execute all algorithmic processes. It features a human-machine interface screen that displays the reconstructed image, defect information, printing status, etc., in real time, and provides a user interface for parameter setting and historical data querying.

[0024] This device integrates an ultrasonic sensor array onto a standard FDM printer, and together with a dedicated ultrasonic signal excitation and acquisition system and a data processing and control system, it forms a closed-loop online quality monitoring system. This enables non-contact, full-field, and real-time monitoring of the printing process, providing a foundation for online evaluation and intelligent decision-making regarding print quality.

[0025] The method for real-time imaging monitoring of 3D printing interface quality using the above-mentioned device is as follows: Phase 1: Initial Calibration This is preparation work before monitoring, the purpose of which is to obtain the attenuation data that the ultrasound should have under the "perfect print" condition.

[0026] 1.1 Define the imaging region and mesh generation Determine the monitoring range: Based on the printer's printing plate and actual monitoring needs, clearly define a two-dimensional imaging area to be monitored. This area is typically a rectangle or other regular shape, covering the effective printing area on the printing plate. Subsequently, this imaging area is divided into n uniformly sized fine grids on a digital model. The grids are typically square, with side lengths... The spatial resolution of the final image is determined by this, and each grid is assigned a unique index number j (j = 1, 2, ..., n). Subsequent sensor layout and algorithms are all designed to ensure effective monitoring of this specific imaging area.

[0027] 1.2 Deployment of Ultrasonic Sensor Array Based on the imaging area defined in step 1.1, optimize the design of the number, type, and arrangement of ultrasonic sensors at the edge of the printing substrate, such as single-span, double-span, or fan-shaped arrangements, to ensure that the ultrasonic wave propagation paths formed between all sensor pairs can completely cover and optimally penetrate the defined imaging area. Here, any transmitting probe and any receiving probe constitute a sensor pair.

[0028] According to the above design, the ultrasonic sensor is firmly coupled to a designated position on the edge of the printing base plate, and the sensor array is connected to an ultrasonic signal excitation and acquisition system using an adapter board, and then connected to the data processing and control system.

[0029] 1.3 Calculate the ultrasound propagation path length matrix L Record the straight lines connecting all sensor pairs; these represent the m ultrasonic propagation paths. For the i-th path and the j-th grid, use a ray tracing algorithm to accurately calculate the path length within the grid. If it doesn't pass through, the value is 0. (This applies to all...) The lengths of the ultrasonic propagation paths are combined into an m-row, n-column matrix L, where m represents the number of ultrasonic propagation paths and n represents the number of grids.

[0030] 1.4 Acquiring sediment-free reference signals The system is started when there is no printing material on the printing substrate. Each pair of sensors (one transmitter and one receiver) is excited sequentially. The transmitter probe excites a Lamb wave, and the receiver probe receives the signal. The complete waveform signal of the Lamb wave propagating in the substrate is recorded for each path. For the signal of each path i, the energy value of the A0 mode is extracted and recorded as the waveform energy value when there is no deposition on each path. This value represents the energy reference of the ultrasound wave propagating on the empty plate along path i.

[0031] like Figure 2As shown, the print bed in 3D printing can be regarded as a thin layer structure. When the ultrasonic wave is excited, the ultrasonic wave generates transverse wave and longitudinal wave due to the refraction law. The transverse wave and longitudinal wave are reflected back and forth on the upper and lower surface boundaries of the print bed and propagate in the direction parallel to the print bed surface. After a period of propagation, the transverse wave and longitudinal wave are fully mixed, and the propagating elastic wave is the Lamb wave. In the FDM printing process, the molten wire material is deposited on the print bed and continuously solidified. According to Snell's law, if the Lamb wave velocity in the print bed exceeds the Lamb wave velocity in the deposited material, the energy of the Lamb wave will leak into the deposited layer. This condition is met in most printing cases. In order to avoid the aliasing interference of high-order modes, a smaller frequency-thickness product should be selected, and only A0 and S0 modes are present. Compared with the S0 mode, the A0 mode has a larger out-of-plane displacement and a more significant energy attenuation phenomenon, and is suitable for calculating the path energy attenuation coefficient.

[0032] 1.5 Calibration of reference energy attenuation coefficient Using the material and parameters to be officially printed, a known defect-free and simple standard pattern is printed layer by layer on the print bed. After printing the first layer, the printing is paused and the process of step 1.4 is repeated to collect the waveform signals of all paths at this time, and the waveform energy values of each path with deposition are calculated .

[0033] The transmission attenuation tomography formula is introduced into the Lamb wave propagation process, and the energy leakage along the ultrasonic propagation path can be characterized by the energy attenuation coefficient . Assuming that the Lamb wave propagates along a straight line in the print bed, the formula of the energy attenuation coefficient is as follows: wherein, is the frequency-dependent attenuation coefficient per unit length along the ultrasonic propagation path in the region, is the waveform energy with deposition, is the waveform energy without deposition, is the integral along the projection path, is the center frequency of the sensor.

[0034] Therefore, for each path i, the reference energy attenuation coefficient is calculated according to the formula The larger the reference energy attenuation coefficient

[0035] When the deposited material has no printing defects, the energy attenuation coefficient of a certain ultrasonic propagation path is approximately linearly related to the length of the deposited material on the path, and the reference energy attenuation coefficient on the path is divided by the length of the deposited material on the path to obtain the unit length reference energy attenuation coefficient corresponding to the path when the first layer is printed, that is , wherein is the length of path i in the printed material, determined by the matrix L and the printing area.

[0036] The unit length reference energy attenuation coefficient corresponding to all paths is (i=1,2,…,m) is taken as the unit length reference energy attenuation coefficient of the layer In this embodiment, the trimmed mean is obtained by removing the highest and lowest 10% and then averaging.

[0037] Through this method, the unit length reference energy attenuation coefficients of double-layer, triple-layer and more layers of deposition can also be obtained , , etc. The calibration curve of the unit length reference energy attenuation coefficient is obtained in this way, and the unit length reference energy attenuation coefficient variation amount of different layer heights , , etc., wherein , ( ), and the unit length reference energy attenuation coefficient variation amount gradually decreases as the number of layers increases.

[0038] Therefore, as the printing height increases, the energy attenuation coefficient will gradually tend to be flat or even decrease, losing the detection ability. Considering the diversity of printing materials, printing layer height and deposited shape, the sum of the energy attenuation coefficients of each path is recorded when collecting full path data each time , wherein represents the sum of the path energy attenuation coefficients obtained by the pth full path collection, and when , it is determined that the system has lost the detection ability for the newly deposited material above the pth layer under the printing condition. Generally, at this time, the pth layer is used as the threshold for defect judgment, and the present scheme is applicable to the printing quality monitoring within the pth layer.

[0039] Second stage: real-time monitoring and imaging During the formal printing process, the system starts to work in a cycle.

[0040] Step 2.1: Cycle data collection The system automatically performs a full path scan at a fixed time interval (such as 3 seconds) to collect real-time waveforms of all paths and calculate a real-time path energy attenuation coefficient matrix Z, wherein .

[0041] The energy attenuation coefficient in the above formula (1) The energy attenuation coefficient is most sensitive to the first layer of deposits, and in the case of uniform deposit thickness on the path, the energy attenuation coefficient is linearly related to the length of the material deposition on the path; in the case of uniform deposit length on the path, the energy attenuation coefficient is linearly related to the thickness of the deposit on the path. Combining all the ultrasound propagation paths, the imaging problem can be represented in matrix form: wherein, represents the length matrix of each ultrasound propagation path in each grid, represents the number of grids, and m is the number of paths, represents the energy attenuation coefficient matrix of each projection path, represents the energy attenuation coefficient matrix of each grid.

[0042] Substituting the real-time path energy attenuation coefficient matrix Z and the ultrasound propagation path length matrix L into the above formula (2), the imaging problem is converted into a mathematical formula, and the unit length energy attenuation coefficient matrix is iteratively solved to invert the unit length energy attenuation coefficient of each grid, and the energy attenuation distribution image, i.e., the real-time monitoring image, is reconstructed.

[0043] In this embodiment, the simultaneous iterative reconstruction algorithm (SIRT) is used to iteratively solve the unit length energy attenuation coefficient matrix The formula for image reconstruction by the SIRT algorithm is as follows: wherein, is the estimated energy attenuation coefficient of the jth grid in the kth iteration, is the correction value of the energy attenuation coefficient, is the calculated energy attenuation coefficient of the ith ultrasound propagation path in the kth iteration, is the measured energy attenuation coefficient, and to ensure the stability of the iterative calculation, the correction increment is multiplied by a relaxation factor .

[0044] Third stage: defect diagnosis and printing decision The actual deposited area is known, which is corresponded to the imaging to determine the total number N of grids that need to be judged.

[0045] 3.1 Real-time defect judgment The energy attenuation coefficients of these grids are judged for defects, wherein is a threshold parameter, is a grid length value, is the difference between the energy attenuation coefficient at the end of the current imaging iteration and the energy attenuation coefficient at the end of the previous imaging iteration, i.e. the current value minus the previous frame value.

[0046] The energy attenuation coefficient and its change are used to determine the type of common defects and determine the orientation, including interlayer defect determination based on the energy attenuation coefficient change and overall defect determination based on the energy attenuation coefficient. Both determination methods can locate to a specific grid.

[0047] Specifically, based on the energy attenuation coefficient change and the reference energy attenuation coefficient change per unit length of different layer heights , , For each grid, when it has deposited q layers of deposited material , it is determined according to the following relationship: If , it is determined that the deposition at the jth grid has debonding and warping at the bottom.

[0048] If , 2≤k≤q, it is determined that the deposition at the jth grid has interlayer peeling at the kth layer.

[0049] If , further overall determination is performed on the jth grid. When or , it is determined that the deposition at the jth grid has defects, otherwise the deposition at the jth grid has no defects. Wherein, is the grid length, is a threshold parameter (generally set to 0.9), is the energy attenuation coefficient per unit length of grid j, is the reference energy attenuation coefficient per unit length of the kth layer.

[0050] The above scheme can accurately distinguish different types of defects such as bottom debonding / warping and interlayer peeling at a specific layer height, and provides clear fault diagnosis information for the operator.

[0051] 3.2 Real-time statistics and decision After real-time defect determination for n target grids layer by layer, considering possible misjudgments and the allowable degree for some defects, it is determined whether to end the printing, wherein is the total number of grids with defects, is an error coefficient: When When the defect is within the acceptable range, the system determines that the printing needs to be continued.

[0052] When the defect is within the acceptable range, the system determines that the printing needs to be continued. When the defect is within the acceptable range, the system determines that the printing needs to be continued.

[0053] By setting the defect grid ratio threshold , the system can automatically determine the severity of the defect and make a decision of "continue printing" or "suspend alarm", greatly reducing the dependence on manpower and improving the automation level and efficiency of the production process.

[0054] The above-described embodiments only express several embodiments of the present application, and the description is more specific and detailed, but it cannot be understood as a limitation on the scope of the present application. For ordinary skilled persons in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application.

Claims

1. A method for real-time imaging monitoring of 3D printed interface quality, characterized in that, Includes the following steps: (1) Define the imaging area on the printing base plate and divide the imaging area into multiple uniform grids. According to the grid division of the imaging area, deploy an ultrasonic sensor array on the edge of the printing base plate and calculate the ultrasonic propagation path length matrix. The elements of the matrix represent the length of each ultrasonic propagation path in each grid. (2) Acquire reference ultrasonic signals propagating in the form of Lamb waves without deposition through the ultrasonic sensor array; then, print a defect-free standard pattern on the printing substrate, and acquire ultrasonic signals propagating in the form of Lamb waves again. Calculate the reference energy attenuation coefficient of each ultrasonic propagation path based on the A0 mode energy values ​​of the two acquired signals; establish a calibration curve containing the reference energy attenuation coefficient per unit length corresponding to different layer heights and its variation by printing defect-free patterns of different layer heights. (3) During the printing process, full-path ultrasonic scanning is performed through the ultrasonic sensor array at fixed time intervals to collect real-time ultrasonic signals and calculate the real-time energy attenuation coefficient matrix. The energy attenuation coefficient matrix of each grid unit length in the imaging area is inverted using the image reconstruction algorithm to generate a real-time monitoring image. (4) Based on the change in energy attenuation coefficient per unit length of each grid in the real-time monitoring image, and according to the calibration curve, determine the type and location of the defect, and decide whether to continue printing based on the comparison result of the number of grids with defects and the preset threshold.

2. The method for real-time imaging monitoring of 3D printing interface quality according to claim 1, characterized in that, The ultrasonic sensor array includes u ultrasonic transmitting probes and v ultrasonic receiving probes, where u and v are both integers greater than 1; the u ultrasonic transmitting probes and the v ultrasonic receiving probes work in a fully combined manner to form u×v ultrasonic propagation paths; the arrangement of the sensor array is such that the u×v ultrasonic propagation paths can cover all grids within the imaging area.

3. The method for real-time imaging monitoring of 3D printing interface quality according to claim 1, characterized in that, The calculation of the reference energy attenuation coefficient for each ultrasonic propagation path based on the signals acquired twice specifically includes: For each ultrasound propagation path i, its reference energy attenuation coefficient is calculated according to the following formula. : ； in, The center frequency of the ultrasonic sensor. The energy value of the ultrasonic signal waveform along path i when there is no deposition. The ultrasonic signal waveform energy value of path i after printing the standard pattern.

4. The method for real-time imaging monitoring of 3D printing interface quality according to claim 3, characterized in that, The waveform energy value is the A0 mode energy value in the received Lamb wave signal.

5. The method for real-time imaging monitoring of 3D printing interface quality according to claim 1, characterized in that, The establishment of calibration curves that include the reference energy attenuation coefficient per unit length and its variation for different floor heights includes: By printing k layers of defect-free patterns (k being an integer greater than or equal to 1), the reference energy attenuation coefficient per unit length of each path after printing the kth layer is obtained. : ； in, Let i be the reference energy attenuation coefficient for path i after printing the k-th layer. Let be the length that path i travels through the printed material in the kth layer; For all paths corresponding By calculating the truncated mean, the general reference energy attenuation coefficient per unit length for this layer can be obtained. Based on different k values Calculate the change in the reference energy attenuation coefficient per unit length. ,in, , ( This forms the calibration curve.

6. The method for real-time imaging monitoring of 3D printing interface quality according to claim 5, characterized in that, It also includes the step of determining the system's detection capability threshold: When printing patterns with different layer heights k, record the sum of the energy attenuation coefficients for each path obtained from the full-path data acquisition. ,in It is the sum of the energy attenuation coefficients of the k-th layer; When satisfied When the system loses its ability to effectively detect new deposits above the p-th layer, it uses the p-th layer as the threshold layer for defect determination, and the calibration curve is only the calibration result within the p-th layer.

7. The method for real-time imaging monitoring of 3D printing interface quality according to claim 1, characterized in that, The method of using image reconstruction algorithms to invert the distribution of energy attenuation coefficient per unit length of each grid within the imaging region specifically involves: The imaging problem is constructed as a system of linear equations. , where L is the ultrasonic propagation path length matrix, Z is the real-time path energy attenuation coefficient matrix, and A is the energy attenuation coefficient matrix per unit length of each grid to be determined; An iterative reconstruction algorithm is used to solve matrix A, thereby reconstructing the real-time monitoring image of the imaging area.

8. The method for real-time imaging monitoring of 3D printing interface quality according to claim 1, characterized in that, In step (4), the process of determining the type and location of the defect includes: The change in energy attenuation coefficient per unit length of each grid in the image will be monitored in real time. The change in the reference energy attenuation coefficient per unit length in the calibration curve In contrast, 1≤k≤q≤p, where q is the current printing layer and p is the threshold layer number that causes the system to lose its ability to effectively detect new deposits. For each grid cell, the following relationship applies: like The sediment at the j-th grid point was determined to have debonded and warped at its bottom. like If 2≤k≤q, it will be determined that interlayer delamination has occurred in the k-th layer of the sediment at the j-th grid. like Then, a global determination is further performed on the j-th grid, when... or If the condition is met, the sediment at the j-th grid is considered defective; otherwise, the sediment at the j-th grid is considered defect-free. in, Let be the grid side length. The preset threshold parameter, Let be the energy attenuation coefficient per unit length of grid j. Let be the reference energy attenuation coefficient per unit length of the k-th layer. This represents the change in the reference energy attenuation coefficient per unit length of the first layer. and These represent the changes in the reference energy attenuation coefficient per unit length for the k-th and (k-1)-th layers, respectively.

9. The method for real-time imaging monitoring of 3D printing interface quality according to claim 1, characterized in that, In step (4), after each determination of the type and location of the defect, the number of grids Ne in the n grids of the imaging area with defects is counted. If the proportion of the number of defective grids exceeds the preset proportion, the printing ends.

10. A device for real-time imaging monitoring of 3D printing interface quality, applied to an FDM printer, characterized in that, include: An ultrasonic sensor array is fixed to the edge of the printing plate of the FDM printer; An ultrasonic signal excitation and acquisition system, connected to the ultrasonic sensor array, is used to excite the probes in the ultrasonic sensor array and acquire ultrasonic wave signals. A data processing and control system, communicatively connected to the ultrasonic signal excitation and acquisition system, is used to control signal transmission and reception, and to execute the steps of the method for real-time imaging monitoring of 3D printing interface quality as described in any one of claims 1 to 9.

Images