Method and system for quality monitoring of laser printing

Abstract

The embodiment of the specification provides a quality monitoring method and system of laser printing, and relates to the technical field of laser printing. The method comprises the following steps: acquiring a molten pool optical detection signal corresponding to each laser track segment in a laser printing process; determining a process category to which the laser track segment belongs according to process parameter information corresponding to each laser track segment, and allocating a corresponding line type number to different process categories; performing statistics on the molten pool optical detection signals of at least part of the laser track segments belonging to the same line type number, and establishing a quality monitoring benchmark corresponding to the line type number; and performing abnormality determination on the molten pool optical detection signals of the laser track segments according to the quality monitoring benchmark corresponding to the line type number to which each laser track segment belongs, so that the accuracy of laser printing can be improved.

Description

Technical Field

[0001] The embodiments described in this specification relate to the field of laser printing technology, specifically to a laser printing quality monitoring method and system. Background Technology

[0002] In laser manufacturing or laser printing processes, the forming quality and abnormal defects are typically monitored online by collecting the molten pool light radiation feedback signal. Existing detection solutions assume that the entire printing area uses uniform and stable process parameters, and establish a uniform defect judgment threshold based on a pre-calibrated normal signal mean or peak range. When the actual detection signal deviates from the preset threshold, an abnormal defect is determined to exist.

[0003] However, in actual laser printing, due to the complex structure of the parts and the significant heat accumulation effect, different regions often require different laser processing conditions. For example, thin-walled areas require reduced laser power or increased scanning speed to prevent overheating, while thick-walled or filled areas require increased power or reduced scanning speed to ensure sufficient melting. The molten pool state differs under different laser conditions, resulting in significant differences in the corresponding detection signal amplitude benchmarks—higher signal amplitudes under high power and lower signal amplitudes under low power or high-speed scanning.

[0004] If a uniform threshold is still used for defect determination, the same defect may produce different detection results under different process conditions: normal signals in the high-power region may be misjudged as defects (false alarms) due to their high amplitude, while real defect signals in the low-power region may fail to trigger determination (missed alarms) due to their low amplitude, which seriously affects the accuracy of detection.

[0005] Therefore, how to establish applicable detection benchmarks for different process areas and independently determine anomalies during the printing process where multiple laser processing conditions coexist has become a pressing technical problem in the field of online defect detection. Summary of the Invention

[0006] In view of this, various embodiments of this specification aim to provide a laser printing quality monitoring method and system to improve the accuracy and print quality of laser printing.

[0007] This specification provides a quality monitoring method for laser printing, the method comprising: acquiring the molten pool optical detection signal corresponding to each laser trajectory segment during laser printing; determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning a corresponding thread type number to different process categories; statistically analyzing the molten pool optical detection signals of at least a portion of the laser trajectory segments belonging to the same thread type number to establish a quality monitoring benchmark corresponding to the thread type number; and determining anomalies in the molten pool optical detection signals of the laser trajectory segment based on the quality monitoring benchmark corresponding to the thread type number to which each laser trajectory segment belongs.

[0008] This specification provides a quality monitoring system for laser printing, comprising: a multi-optical detection device for acquiring optical detection signals generated by the molten pool during laser printing; a laser power detection device for acquiring actual laser output power signals in real time; a galvanometer coordinate detection device for acquiring galvanometer coordinate signals in real time; and an industrial control computer, which is communicatively connected to the multi-optical detection device, the laser power detection device, and the galvanometer coordinate detection device, and is used to execute the quality monitoring method described in any of the above embodiments.

[0009] In the various embodiments provided in this specification, the accuracy and quality of laser printing can be improved by acquiring the optical detection signal of the molten pool corresponding to each laser trajectory segment during the laser printing process; determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning corresponding line types to different process categories; statistically analyzing the optical detection signals of the molten pool for at least a portion of the laser trajectory segments belonging to the same line type to establish a quality monitoring benchmark corresponding to the line type; and determining anomalies in the optical detection signals of the molten pool for each laser trajectory segment based on the quality monitoring benchmark corresponding to the line type to which each laser trajectory segment belongs. Attached Figure Description

[0010] Figure 1 This is a scenario example of the laser printing quality monitoring method provided in the embodiments of this specification; Figure 2 This is a schematic diagram of the quality monitoring method for laser printing provided in the embodiments of this specification; Figure 3 This is a schematic diagram of simultaneous processing by multiple laser heads provided in the embodiments of this specification; Figure 4 A schematic diagram of a computer device provided for an embodiment of this specification. Detailed Implementation

[0011] To enable those skilled in the art to better understand the solutions described in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of them. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.

[0012] This specification provides an example application scenario for a laser printing quality monitoring method. Please refer to [link / reference]. Figure 1 This application scenario can be a quality monitoring system for laser printing, which may include multiple optical detection devices 4, a laser power detection device 1, a galvanometer coordinate detection device, and an industrial control computer.

[0013] Multiple optical detection devices 4 are used to collect optical detection signals generated by the molten pool during laser printing; Laser power detection device 1 is used to collect the actual laser output power signal in real time; A galvanometer coordinate detection device is used to acquire galvanometer coordinate signals or galvanometer motion coordinate signals in real time. The industrial control computer is connected to multiple optical inspection devices, a laser power inspection device, and a galvanometer coordinate inspection device to monitor quality during the laser printing process.

[0014] In this scenario example, the multi-optical detection device 4 may include a coaxial detection optical path or a paraxial detection optical path for simultaneously acquiring at least one of visible light signals, infrared light signals, and reflected light signals.

[0015] In this scenario example, the laser power detection device 1 can directly acquire the power of a portion of the laser signal L, L1, which is split off by the beam splitter 51, and thus acquire the actual laser power signal; the portions of the laser signal L, L2 and L3, are used for laser printing. The multi-optical detection device 4 can collect the optical detection signal generated by the molten pool during laser printing, or in other words, it can acquire the optical detection signal generated by the molten pool on the surface of the workpiece (printed part) along the laser L4 / L5 / L6; the optical detection signal enters the multi-optical detection device 4 through the optical path calibration module 3.

[0016] The beam splitter module 5 and the galvanometer 21 are part of the laser head. The galvanometer 21 can adjust the laser emission direction.

[0017] In this scenario example, the quality monitoring system for laser printing can adapt to various complex application scenarios, such as switching between multiple process areas with a single laser head, printing with multiple laser heads simultaneously, accurate classification with printing planning parameters, and clustering recognition without printing planning parameters. It has wide applicability and robustness.

[0018] This specification provides a method for quality monitoring in laser printing. Please refer to [link / reference]. Figure 2 , Figure 2 This is a flowchart illustrating a laser printing quality monitoring method provided in this specification. This embodiment provides the method operation steps as shown in the flowchart, but based on conventional or non-creative labor, more or fewer operation steps may be included. The order of steps listed in the embodiment is merely one possible execution order among many, and does not represent the only possible execution order. In actual system or server product execution, the method can be executed sequentially as shown in the embodiment or in parallel (e.g., in a parallel processor or multi-threaded processing environment). This laser printing quality monitoring method can be applied to an industrial control computer in a laser printing quality monitoring system, specifically as follows... Figure 2 As shown, the quality monitoring method for laser printing may include the following steps.

[0019] Step S110: Obtain the optical detection signal of the molten pool corresponding to each laser trajectory segment during the laser printing process.

[0020] In some cases, it is necessary to automatically adjust the signal benchmarks and judgment rules used in quality monitoring according to changes in the laser process conditions used in the current processing area, so as to achieve quality monitoring or adaptive defect detection. In this way, it is not necessary to manually set different detection thresholds for different process areas. Instead, the system can automatically identify the process category to which the current laser trajectory segment belongs and call the detection benchmark that matches the process category to make defect judgment, thereby achieving adaptive matching between detection strategy and process conditions.

[0021] Laser printing, in this context, can refer to laser 3D printing. Laser printing can include, but is not limited to, laser additive printing and laser powder coating printing. The following explanation will use laser additive printing as an example.

[0022] Laser additive printing is an additive manufacturing process that uses a laser beam to melt metal powder material layer by layer to build three-dimensional solids. For example, laser additive printing can include, but is not limited to, selective laser melting, laser metal deposition, and laser powder bed melting. During laser printing, the laser beam moves along a preset scanning or processing path, forming a molten pool at each processing position along the path. The shape and radiation characteristics of the molten pool reflect the melting quality and forming state at the current processing position.

[0023] A laser trajectory segment is a continuous laser signal or pulsed laser signal that occurs within a single continuous time period, during which the laser operates continuously and moves along the processing path.

[0024] The molten pool optical detection signal can be a light signal reflecting the radiation characteristics of the molten pool, acquired by an optical detection device. For example, under laser heating, the molten pool emits light radiation containing rich state information. The intensity, spectral distribution, and temporal characteristics of this light radiation are closely related to the temperature, size, stability, and presence of defects such as spatter and porosity of the molten pool. For instance, the molten pool optical detection signal can include at least one of visible light, infrared light, and reflected light signals. Visible light signals reflect changes in molten pool brightness, infrared light signals reflect the thermal radiation state of the molten pool, and reflected light signals reflect the laser reflection state and the stability of the molten pool. As an example, the molten pool optical detection signal can serve as basic data for defect determination, subsequently compared with quality monitoring benchmarks to determine the presence of anomalies.

[0025] The optical detection signal of the molten pool corresponding to each laser trajectory segment is the molten pool optical detection signal collected at the same laser trajectory segment's action time or location. Since the laser pulse frequency and data sampling frequency may differ during laser printing, timestamp synchronization and interpolation alignment are required to ensure that the collected molten pool optical detection signal and the corresponding laser trajectory segment are precisely matched in time and space. Therefore, each laser trajectory segment has a unique corresponding molten pool optical detection signal, reflecting the molten pool state under the action of that laser trajectory segment.

[0026] Step S120: Based on the process parameter information corresponding to each laser trajectory segment, determine the process category to which the laser trajectory segment belongs, and assign the corresponding line type number to different process categories.

[0027] Process parameter information refers to data characterizing the laser processing conditions used in the current laser trajectory segment. For example, during laser printing, laser processing conditions affecting the molten pool state and forming quality include parameters such as laser power, scanning rate, duty cycle, scanning spacing, powder feed rate, scanning strategy, or scanning path. Among these, laser power and scanning rate have the most significant impact on the molten pool state. As an example, process parameter information may include at least laser power and scanning rate. Process parameter information can be obtained through various means, such as directly reading layer data as printing planning parameters from the printing equipment's control system, or acquiring actual processing parameters in real time through a laser power detection device and a galvanometer coordinate detection device. Different methods of obtaining process parameter information lead to different methods of subsequently determining the process category, which will be described in detail in subsequent embodiments.

[0028] Process category refers to the classification result obtained by dividing laser trajectory segments according to different process parameter information or laser processing conditions. For example, when multiple laser trajectory segments use the same or similar laser power and scanning rate, these laser trajectory segments have similar melt pool states and signal reference characteristics, and therefore can be classified into the same process category. As an example, the granularity of process category division depends on the number of process condition combinations used in the actual printing process. For instance, if there are three different power-rate combinations in the printing process (such as high power low speed, medium power medium speed, and low power high speed), the number of process categories is 3; if there is only one power-rate combination, the number of process categories is 1.

[0029] A line type number is an identifier used to uniquely identify a process category. For example, a line type number can be understood as a number or code corresponding to different process categories. Assigning a unique line type number to each process category allows for quick indexing and retrieval of the corresponding quality monitoring benchmark in subsequent processing. For instance, a process category of high-power, low-speed filling can be assigned line type number 1, a process category of low-power, high-speed contouring can be assigned line type number 2, and a process category of medium-power, medium-speed overlapping can be assigned line type number 3. The assignment of line type numbers can be pre-set (e.g., reading process categories or process combinations from print planning parameters and mapping them to corresponding line type numbers), or it can be dynamically generated during the printing process (e.g., automatically identifying process categories through cluster analysis and then assigning temporary line type numbers). As an example, a line type number can be a numeric number, an alphanumeric number, or a combination of numbers and letters, as long as it can identify different process categories.

[0030] Step S130: Statistically analyze the optical detection signals of the molten pool for at least a portion of the laser trajectory segments belonging to the same line type, and establish a quality monitoring benchmark corresponding to the line type.

[0031] Since the laser trajectory segments corresponding to the same line type number use the same or similar laser power and scanning rate, their molten pool state should have consistent signal characteristics under normal circumstances. Therefore, the molten pool optical detection signals of at least some laser trajectory segments assigned to the same line type number can be aggregated together, and a benchmark value that can characterize the normal signal level of this category can be extracted from these signals of the same type.

[0032] Generally, a single region can be selected, and the optical inspection signals of the molten pool for the same wire type within that region can be statistically analyzed to form a quality monitoring benchmark for that region. To determine whether a signal of the same wire type within that single region is defective, it can be compared with the quality monitoring benchmark to determine if a defect exists.

[0033] For example, statistical methods may include, but are not limited to, calculating central tendency statistics such as the arithmetic mean, weighted average, median, and mode. For instance, calculating the arithmetic mean can be used as a statistical method. For example, if the number of laser trajectory segments corresponding to a certain line type is N, and the molten pool optical detection signal values ​​of each laser trajectory segment are S1, S2, ..., S... N Then the average value S obtained through statistics mean = (S1+S2+……+S N ) / N.

[0034] A quality monitoring baseline is a reference standard used to determine whether there are abnormal defects in a laser trajectory segment. For example, the quality monitoring baseline is established based on the signal characteristics under normal conditions for the corresponding process category and reflects the signal level of that process category under normal printing conditions. For instance, for line type 1 (high-power fill process), its quality monitoring baseline can be the baseline value of the molten pool optical detection signal for all normal laser trajectory segments of that line type; for line type 2 (low-power contour process), its quality monitoring baseline is the baseline value of the molten pool optical detection signal for all normal laser trajectory segments of that line type. Since the laser process conditions corresponding to different line types are different, the amplitude baseline of their molten pool optical detection signals is naturally different as well; therefore, the quality monitoring baselines for each line type are also different.

[0035] For example, the specific form of the quality monitoring benchmark includes, but is not limited to: the statistical mean of the molten pool optical detection signal, a preset benchmark threshold range, variance, the allowable fluctuation range consisting of the statistical mean and standard deviation, interquartile range, root mean square value, or a dynamic baseline based on time-series characteristics.

[0036] Specifically, the quality monitoring benchmark value of the thread type can be determined based on the statistical results, and the benchmark value can be associated, linked or mapped with the thread type and stored.

[0037] Step S140: Based on the quality monitoring benchmark corresponding to the line type number of each laser trajectory segment, perform anomaly judgment on the molten pool optical detection signal of the laser trajectory segment.

[0038] For example, for the laser trajectory segment to be judged, the quality monitoring benchmark value corresponding to the line type number to which the laser trajectory segment belongs can be obtained first, and then the actual molten pool optical detection signal of the laser trajectory segment can be compared with the quality monitoring benchmark value. Based on the comparison result, it can be determined whether there is a defect at the position of the laser trajectory segment.

[0039] In the above implementation, the process category of each laser trajectory segment is automatically identified and a line type number is assigned based on process parameter information. Then, independent quality monitoring benchmarks are established for all laser trajectory segments with the same line type number. This allows for defect determination using appropriate benchmarks under different laser process conditions. Thus, the quality monitoring benchmark corresponding to each line type number only reflects the signal level under normal conditions for that process category, and signal differences between different process categories do not interfere with each other. Since each laser trajectory segment calls its corresponding quality monitoring benchmark according to its own line type number, even if the signal amplitude jumps due to changes in laser power or scanning rate in the process switching area, the system can accurately distinguish between "signal changes caused by normal process switching" and "signal anomalies caused by defects," thereby effectively reducing false alarms and missed detections in defect detection.

[0040] In some implementations, the process parameter information includes the actual laser power signal and the real-time scanning rate.

[0041] In this embodiment, step S120 involves determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning a corresponding line type number to different process categories. This may include the following steps S210-S240.

[0042] Step S210: If the layer data used as printing planning parameters cannot be obtained, then obtain the actual laser power signal and galvanometer coordinate signal for each laser trajectory segment, and calculate the real-time scanning rate corresponding to each laser trajectory segment based on the galvanometer coordinate signal.

[0043] Layer data is a set of data containing preset process parameters for each printing layer, serving as printing planning parameters. For example, in laser printing, a 3D part model is divided into multiple 2D layers after being processed by slicing software, with each layer corresponding to a printing layer. The printing data, or layer data, for each layer contains the position coordinates of all scanning paths within that layer, as well as corresponding preset laser power, scanning rate, and other process parameter information. For instance, for a given layer, its layer data might include: the outer contour scanning path and its corresponding preset laser power (e.g., 150W) and preset scanning rate (e.g., 1200mm / s), and the inner filling scanning path and its corresponding preset laser power (e.g., 350W) and preset scanning rate (e.g., 600mm / s). However, in actual industrial scenarios, some printing equipment cannot provide layer data due to compatibility issues such as closed interfaces, incompatible communication protocols, or historical system upgrades. For example, the control systems of some imported or older equipment can only receive and execute process documents but cannot output the currently executed preset parameters such as power and rate to external detection systems via data interfaces.

[0044] Without access to layer data, the system cannot directly determine which preset process parameters are used for the current laser trajectory segment. Therefore, it needs to indirectly identify the process category by using the actual acquired power signal and the calculated scanning rate.

[0045] The actual laser power signal, also known as the data signal that reflects the true power value of the laser beam at actual output, is acquired in real time by a laser power detection device. During laser printing, the laser outputs laser light according to a preset power value. However, due to factors such as laser fluctuations, optical path attenuation, galvanometer dynamic response, and changes in material reflectivity, the actual output power of the laser may deviate from the set theoretical power. By acquiring the actual laser power signal, the true laser energy input received by each laser trajectory segment can be more accurately characterized, providing reliable power dimension data for subsequent process category classification. The acquisition of the actual laser power signal can be performed simultaneously with the acquisition of the molten pool optical detection signal to ensure that the power signal and optical signal corresponding to the same laser trajectory segment correspond precisely in time and space.

[0046] Real-time scanning rate is the actual scanning speed of the laser beam on the powder bed surface or the processing plane during laser printing. The scanning motion of the laser beam is controlled by a galvanometer system, and the deflection angle of the galvanometer determines the position of the laser beam on the processing plane. During printing, the laser beam needs to move along a preset scanning path, and its speed, i.e., the scanning rate, directly affects the interaction time between the laser and the material, the heat input density, and the molten pool morphology. For example, when the scanning rate is high, the interaction time between the laser and the material is short, the heat input is low, and the molten pool size is small; when the scanning rate is low, the interaction time between the laser and the material is long, the heat input is high, and the molten pool size is large. Therefore, the scanning rate is one of the key process parameters that determines the molten pool state and the reference for the detection signal.

[0047] The galvanometer coordinate signal is a data signal that reflects the galvanometer deflection angle or the spatial position of the laser beam on the machining plane, acquired in real time by a galvanometer coordinate detection device. The galvanometer coordinate detection device can acquire the galvanometer deflection angle signals in the X and Y axes, or directly acquire the two-dimensional coordinate position signal of the laser beam on the machining plane. The industrial control computer calculates the average speed of the laser beam within that time period based on the changes in the galvanometer coordinate signals at adjacent sampling times and the time interval between adjacent sampling times; this average speed is the real-time scanning rate corresponding to that laser trajectory segment.

[0048] Step S220: Determine the process category of the laser trajectory segment based on the actual laser power signal and real-time scanning rate of each laser trajectory segment, and assign a corresponding line number to each process category.

[0049] In some implementations, when the layer data used as printing planning parameters cannot be obtained, step S120 may include the following steps S310-S330: determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning a corresponding line number to different process categories.

[0050] Step S310: Construct a two-dimensional feature vector using the actual laser power signal and real-time scanning rate for each laser trajectory segment.

[0051] A two-dimensional feature vector is a data vector composed of two dimensions: the actual laser power signal value and the real-time scanning rate value. For example, if the actual laser power of a certain laser trajectory segment is 320W and the real-time scanning rate is 580mm / s, then its two-dimensional feature vector can be represented as (320, 580). The two-dimensional feature vector maps the process state of each laser trajectory segment to a two-dimensional space spanned by power and rate. In this two-dimensional space, laser trajectory segments with the same or similar process conditions will naturally cluster in similar location regions. For example, high-power, low-speed laser trajectory segments cluster in the upper left region of the two-dimensional space (high power, low rate), while low-power, high-speed laser trajectory segments cluster in the lower right region of the two-dimensional space (low power, high rate). By constructing two-dimensional feature vectors, the problem of process category identification can be transformed into a clustering problem of data point distribution in a two-dimensional space.

[0052] Step S320: Perform cluster analysis on the two-dimensional feature vectors of all laser trajectory segments to dynamically form K cluster centers.

[0053] Cluster analysis, an unsupervised learning method, aims to group data objects into groups based on their inherent similarity without pre-labeling. In this implementation, without layered data, since the number of process categories and their corresponding power-rate combinations are unknown beforehand, cluster analysis is needed to discover the process category divisions from the inherent structure of the two-dimensional feature vectors of all laser trajectory segments. In laser printing, when multiple laser process conditions exist in the same printing process, the laser trajectory segments corresponding to each category will naturally form several dense clusters in the actual power-rate two-dimensional space. Cluster analysis can automatically identify the distribution of these dense clusters and assign data points to different clusters, thereby achieving automatic process category division. For example, if there are three different power-rate combinations in the printing process, the two-dimensional feature vectors will form three dense clusters in the power-rate space. Cluster analysis can automatically identify these three clusters and classify the laser trajectory segments into three process categories accordingly.

[0054] Cluster centers are the central location vectors of each cluster obtained from cluster analysis, reflecting the typical process characteristics of that cluster. For example, if the cluster center of a certain cluster is (330, 590), it means that the average power of the laser trajectory segment of that cluster is 330W and the average scanning rate is 590mm / s.

[0055] K is the preset number of cluster categories in cluster analysis, meaning it aims to divide all laser trajectory segments into K different process categories. K can be a positive integer. For example, when using cluster analysis for process category identification, the value of K can be determined in the following ways: When layer data is unavailable, the system can automatically estimate the K value based on the distribution characteristics of the two-dimensional feature vectors during actual processing, for example, using evaluation metrics such as the Elbow Method or Silhouette Coefficient to determine the optimal K value; alternatively, the operator can pre-set a reasonable K value based on process experience, for example, setting K equal to the number of known process combinations for the printing task. As an example, K can depend on the number of different power-rate combinations present in the actual printing process. For example, if only two processes, filling and contouring, exist in the printing process, then K=2; if four processes, including filling, contouring, support, and overlapping, then K=4. Without layer data, the system can automatically discover the existence of these process categories and determine the corresponding cluster centers through cluster analysis.

[0056] Step S330: Based on the distribution of cluster centers, divide different laser trajectory segments into K process categories, and assign a unique line number to each process category.

[0057] For example, if the two-dimensional feature vector of a certain laser trajectory segment is (310, 600), and the three cluster centers are C1=(330, 590), C2=(150, 1100), and C3=(400, 500), then the distance between this laser trajectory segment and C1 is the smallest. Therefore, this laser trajectory segment can be classified into the process category corresponding to cluster center C1. After completing the classification of all laser trajectory segments, the set of laser trajectory segments corresponding to each process category can be obtained, and then a unique line type number can be assigned to each process category.

[0058] In the above embodiments, when the printing device cannot provide layer data as printing planning parameters, a two-dimensional feature vector can be constructed by acquiring the actual laser power signal and real-time scanning rate of each laser trajectory segment. Then, cluster analysis is used to automatically aggregate laser trajectory segments with similar power-rate characteristics into the same process category and assign them line type numbers. Thus, even without prior layer data information, the system can automatically identify different process conditions present during printing based on actual processing data, thereby providing a basis for process category classification. This approach eliminates reliance on open interfaces of the printing device, adaptively discovers natural cluster structures in the data, exhibits good robustness and universality, and is suitable for scenarios with closed equipment, outdated equipment, or where process parameters are unavailable, demonstrating good compatibility and versatility.

[0059] In some implementations, step S320, cluster analysis, may further include: dynamically updating cluster centers according to a preset time period or a preset number of layers to compensate for process parameter drift caused by laser thermal drift or long-term printing.

[0060] The preset time period can be a pre-defined time interval to periodically trigger the update of the cluster centers.

[0061] The preset number of layers can be a pre-defined number of printing layers. The cluster centers are updated after each preset number of layers are printed. For example, the preset number of layers can be set to 1, 3, or 5 layers, and the specific value can depend on the total number of layers in the printed part, the rate at which process parameters change with the number of layers, and the availability of system computing resources.

[0062] In some implementations, the process parameter information includes the laser power and scan rate of each laser trajectory segment in the layer data, or, in other words, the preset laser power and preset scan rate of each laser trajectory segment.

[0063] In this embodiment, step S120 involves determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning a unique line type number to different process categories. This may include the following steps S410-S440.

[0064] Step S410: If layer data containing preset process parameters is obtained, pre-calculate the process categories corresponding to all different combinations of laser power and scanning rate in the layer data, and establish a mapping relationship table between process categories and line types.

[0065] For example, the preset laser power and preset scanning rate of all laser trajectory segments recorded in the layer data can be traversed and analyzed to identify all different power-rate combinations and determine each combination as a process category.

[0066] A mapping table can be a data structure that records a one-to-one correspondence between process categories and wire types. For example, the mapping table can store each process category and its unique corresponding wire type. For instance, after identifying three process categories, wire types can be assigned to them: process category A (350W, 600mm / s) corresponds to wire type 1, process category B (150W, 1200mm / s) corresponds to wire type 2, and process category C (200W, 900mm / s) corresponds to wire type 3, thus forming the mapping table. As an example, the mapping table can be pre-built and stored in the industrial control computer before printing begins, and used for online matching and querying during the printing process. The specific form of the mapping table can be any data structure that enables fast lookup, such as an array, linked list, hash table, or database table; no limitation is made here.

[0067] Step S420: During the laser printing process, online matching is performed in the mapping table based on the actual laser power signal and real-time scanning rate of the current laser trajectory segment obtained in real time.

[0068] Online matching is a process in which, during laser printing, for each current laser trajectory segment, its actual processing parameters are compared and matched in real time with the process categories stored in the mapping table.

[0069] Step S430: If the actual laser power signal and real-time scanning rate of the current laser trajectory segment match any process category in the mapping table, then assign the corresponding line type number to the current laser trajectory segment.

[0070] A successful match means that the process category corresponding to the combination of the actual laser power signal and the real-time scanning rate of the current laser trajectory segment is recorded in the mapping table. For example, a successful match can be an exact match, where the actual power value and actual rate value are exactly the same as the preset power value and preset rate value for a certain process category in the mapping table; or it can be an approximate match, where the deviation between the actual power value and actual rate value and the preset power value and preset rate value for a certain process category in the mapping table is within a preset allowable range.

[0071] When a match is successful, the line type number corresponding to the process category can be read from the mapping table, and that line type number can be assigned to the current laser trajectory segment. Alternatively, the corresponding line type number can be assigned to the process category of the current pulse's actual laser power value and actual scanning rate value.

[0072] Step S440: If the matching fails, a new process category is dynamically created for the current laser trajectory segment, and a temporary line type number is assigned.

[0073] During the online matching process, if the combination of the actual laser power signal and the real-time scanning rate of the current laser trajectory segment fails to match any process category in the mapping table, the matching is considered to have failed. In this case, a new process category can be dynamically created for the current laser trajectory segment, and a temporary line number can be assigned to the new process category.

[0074] In the above implementation, a mapping table is established by pre-statistically analyzing all different combinations of laser power and scanning rate in the layer data. During the printing process, the actual laser power signal and real-time scanning rate of each laser trajectory segment are matched online. If the match is successful, a line type number is directly assigned; if the match fails, a new process category is dynamically created and a temporary line type number is assigned. This approach offers several advantages: First, it eliminates the need for computationally intensive operations such as cluster analysis, allowing process category identification solely through table lookup, resulting in low computational complexity, high real-time performance, and suitability for high-speed online manufacturing and inspection scenarios. Second, by using the actual acquired power signal and scanning rate for matching, rather than directly using preset values, it can withstand the influence of factors such as laser fluctuations, ensuring the authenticity of the matching results. Third, by introducing a mechanism for dynamically creating new process categories and temporary line type numbers, even if new power-rate combinations not planned in the layer data appear during printing, the system can adaptively handle them, avoiding interruptions to the subsequent defect detection process due to matching failures. Fourth, the introduction of temporary line type numbers ensures that the dynamically created new process categories can still establish corresponding quality monitoring benchmarks and perform anomaly judgments in subsequent processing, ensuring the integrity of the inspection process.

[0075] In some implementations, after the current layer of the laser printing process has finished printing, the process may further include merging or mapping dynamically created temporary line types with theoretical line types in a mapping table to maintain consistency in online matching.

[0076] In some implementations, step S120, determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning a corresponding line type number to different process categories, may include the following steps S510-530.

[0077] Step S510: If layer data containing preset process parameters is obtained, determine the number of cluster categories K in advance based on the number of process categories corresponding to the combination of laser power and scanning rate in the layer data; where K is a positive integer.

[0078] Step S520: Based on the number of cluster categories K, perform cluster analysis on the laser trajectory segment in combination with the actual laser power signal and real-time scanning rate to obtain K cluster results.

[0079] Specifically, K can be used as the preset number of cluster categories. A two-dimensional feature vector is constructed based on the actual laser power signal and real-time scanning rate of each laser trajectory segment. Then, cluster analysis is performed on the two-dimensional feature vectors of all laser trajectory segments to divide all laser trajectory segments into K cluster results.

[0080] Step S530: Match the clustering results with the process categories in the layer data one by one, and assign a corresponding line number to each clustering result.

[0081] Specifically, after the cluster analysis is completed, the K cluster results can be paired with the K process categories pre-statistically identified in the layer data to determine which process category in the layer data each cluster result corresponds to.

[0082] Since the K clustering results are obtained based on the actual acquired signals, while the K process categories in the layer data are obtained based on the preset power-rate combination statistics, the two are equal in number, but not naturally consistent in order, and need to be matched to determine the correspondence.

[0083] For example, the Euclidean distance between the cluster center of each clustering result (e.g., the average actual power and average actual rate of the cluster) and the preset power and preset rate of each process category in the layer data can be calculated, and the pair with the smallest distance can be selected as the matching result. Of course, this is only one exemplary implementation.

[0084] Specifically, after matching the clustering results with the process categories in the layer data, the line number corresponding to each process category can be assigned to the matching clustering result. For example, if clustering result 1 successfully matches process category A in the layer data, and process category A was assigned line number 1 during the line number allocation stage, then all laser trajectory segments in clustering result 1 will be assigned line number 1.

[0085] In the above embodiments, when layer data containing preset process parameters can be obtained, a process category identification scheme based on cluster analysis guided by prior information is provided. Specifically, the number of cluster categories K is first determined according to the number of preset power-rate combinations in the layer data. Then, cluster analysis is performed on the actually acquired laser trajectory segment data based on K. Finally, the clustering results are matched one-to-one with the process categories in the layer data, and a line type number is assigned to each clustering result. Thus, firstly, by utilizing prior information from the layer data to determine the number of cluster categories K, the uncertainty of determining the K value without prior information is avoided, making the clustering analysis objective clearer, the clustering convergence speed faster, and the clustering results more stable. Secondly, the clustering analysis is based on the actual acquired laser power signal and scanning rate, which can reflect the actual process state deviations caused by actual processing factors such as laser fluctuations and optical path attenuation. Compared with directly using preset values ​​from the layer data for classification, the clustering results are closer to the actual process state. Thirdly, by matching the clustering results with the process categories in the layer data one-to-one, it is ensured that the clustering results can be accurately mapped to each process category defined in the layer data, thereby assigning the correct line number to each clustering result. Since the K value comes directly from the layer data rather than data-driven estimation, it avoids clustering errors caused by inaccurate K value estimation, resulting in higher accuracy in process category identification.

[0086] In some implementations, step S120, determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning a unique line number to different process categories, may include the following steps S610-620.

[0087] Step S610: Pre-set multiple laser power ranges and multiple scanning rate ranges, and establish the correspondence between the process category and line number corresponding to different combinations of power ranges and rate ranges.

[0088] Multiple laser power ranges can be divided at equal intervals or at non-equal intervals (e.g., using narrower ranges in power-sensitive regions and wider ranges in power-insensitive regions). This discretizes continuous power values ​​into several power levels, allowing the power value of each laser trajectory segment to be categorized into a specific power level, thus facilitating combination with scanning rate ranges to form process categories.

[0089] The division of multiple scanning rate intervals can also be done by equal intervals or non-equal intervals. In this way, continuous rate values ​​can be discretized into several rate levels, so that the rate value of each laser trajectory segment can be classified into a certain rate level, which makes it easier to combine with the laser power range to form a process category.

[0090] Each combination of laser power range and each scan rate range is treated as a process category, and a unique line number is assigned to each combination, thus forming a mapping relationship between range combinations and line numbers.

[0091] Step S620: During the laser printing process, determine the combination of the actual laser power signal and real-time scanning rate that each laser trajectory segment falls into, and assign the corresponding line number.

[0092] During the laser printing process, for each laser trajectory segment currently received, its actual laser power signal value and real-time scanning rate value are obtained. Then, the power value is compared with multiple preset laser power intervals to determine the power interval to which it belongs. At the same time, the rate value is compared with multiple preset scanning rate intervals to determine the rate interval to which it belongs, thereby obtaining the interval combination corresponding to the laser trajectory segment.

[0093] The above embodiments provide a process category identification scheme based on preset interval rules. This scheme eliminates the need for layer data and cluster analysis. It only requires pre-setting power and rate intervals and establishing a correspondence between interval combinations and wire type numbers. During printing, process category identification and wire type number assignment for each laser trajectory segment can be completed through simple interval comparison. Thus, it eliminates the need for iterative calculations required by cluster analysis and the search operations required for online matching. The determination is completed solely through numerical interval comparison, resulting in the simplest calculation logic, minimal computational load, and high real-time performance. It is suitable for scenarios where layer data is unavailable but the approximate range of process parameters is known.

[0094] In some embodiments, the quality monitoring method for laser printing may further include: if multiple laser heads are processing simultaneously in the laser printing process, the process category, line type number, and quality monitoring benchmark of each laser trajectory segment of the laser head are independently determined for each laser head, using the laser head serial number as the isolation dimension. Even if different laser heads have the same line type number, their corresponding quality monitoring benchmarks are also established independently.

[0095] For example, please refer to Figure 3 , Figure 3 A schematic diagram of simultaneous processing by multiple laser heads is shown. Figure 3 In the diagram, purple represents the trajectory of laser head 0, and red represents the trajectory of laser head 1. The trajectory of laser head 0 may have line type 1, and the trajectory of laser head 1 may also have line type 1. However, the quality monitoring benchmarks for the two are established independently. Instead, the quality monitoring benchmarks are established separately based on the laser head serial number as the isolation dimension. The internal line type 1 of laser head 0 is judged as a whole, and the same applies to laser head 1.

[0096] In some embodiments, the laser printing quality monitoring method may further include: mapping the position points of laser trajectory segments belonging to the same line type to a spatial region based on the galvanometer coordinate signal and / or process parameter boundary information in the layer data, so as to aggregate and form at least one connected process region.

[0097] In this embodiment, the optical detection signals of the molten pool for all laser trajectory segments belonging to the same line type are statistically analyzed to establish a quality monitoring benchmark corresponding to the line type. This may include: for each connected process area, statistically analyzing the benchmark value of the optical detection signal of the molten pool for all laser trajectory segments within the connected process area, and using the benchmark value as the quality monitoring benchmark corresponding to that connected process area.

[0098] Accordingly, in this embodiment, step S140, based on the quality monitoring benchmark corresponding to the line type number of each laser trajectory segment, performs anomaly determination on the molten pool optical detection signal of the laser trajectory segment, which may include the following steps S710-S720.

[0099] Step S710: Calculate the deviation between the optical detection signal of the molten pool of a single laser trajectory segment and the reference value of the optical detection signal of the molten pool of the corresponding process area.

[0100] Step S720: If the deviation value exceeds the preset threshold, it is determined that there is a defect at the position point of the laser trajectory segment.

[0101] In some embodiments, the laser printing quality monitoring method may further include: after acquiring the molten pool optical detection signal, actual laser power signal and galvanometer coordinate signal corresponding to each laser trajectory segment during the laser printing process, performing time stamp synchronization and interpolation alignment on the molten pool optical detection signal, actual laser power signal and galvanometer coordinate signal corresponding to the same laser trajectory segment.

[0102] In some implementations, process parameter information includes the actual laser power signal and real-time scan rate, and / or, the laser power and / or scan rate included in the layer data as part of the printing plan.

[0103] In this embodiment, step S210 involves determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning a unique line type number to different process categories. This may include the following steps S810-860.

[0104] Step S810: When the actual scanning rate or the change in the actual scanning rate during laser printing is less than a preset threshold, the actual laser power signal of each laser trajectory segment is used as a one-dimensional feature vector; or, when the laser printing process plan only adjusts the laser power without adjusting the scanning rate, the actual laser power signal of each laser trajectory segment is used as a one-dimensional feature vector.

[0105] Step S820: When the actual laser power or the change in actual laser power during laser printing is less than a preset threshold, the actual scanning rate of each laser trajectory segment is used as a one-dimensional feature vector.

[0106] Step S830: When only the laser power is adjusted and not the scanning rate is adjusted in the laser printing process planning, the actual laser power signal of each laser trajectory segment is used as a one-dimensional feature vector.

[0107] Step S840: When only the scanning rate is adjusted and not the laser power is adjusted in the laser printing process planning, the actual scanning rate of each laser trajectory segment is used as a one-dimensional feature vector.

[0108] Step S850: Perform cluster analysis on the one-dimensional feature vectors of all laser trajectory segments to dynamically form K cluster centers.

[0109] Step S860: Based on the distribution of cluster centers, divide different laser trajectory segments into K process categories, and assign a unique line number to the process category to which each laser trajectory segment belongs.

[0110] In some implementations, the molten pool optical detection signal includes at least one of the following four signals: visible light signal, infrared light signal, reflected light signal, and actual laser power signal.

[0111] In some implementations, step S140, which involves determining anomalies in the molten pool optical detection signal of the laser trajectory segment, may include: extracting at least one of the following from the molten pool brightness, peak energy amplitude, waveform width, and thermal radiation intensity based on the molten pool optical detection signal of the laser trajectory segment, as a detection feature for determining anomalies against a quality monitoring benchmark.

[0112] This specification also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a computer, implements the laser printing quality monitoring method in any of the above embodiments.

[0113] This specification also provides a computer program product containing instructions that, when executed by a computer, cause the computer to implement the laser printing quality monitoring method in any of the above embodiments.

[0114] This specification also provides an industrial control computer, i.e., a computer device, including a memory and a processor. The memory stores a computer program, which, when executed by the processor, implements the laser printing quality monitoring method in any of the above embodiments.

[0115] In some implementations, please refer to Figure 4The computer device can be a terminal, and its internal structure diagram can be as follows: Figure 4 As shown, the computer device includes a processor, memory, and a communication interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for quality monitoring in laser printing.

[0116] It is understood that the specific examples in this document are only intended to help those skilled in the art better understand the embodiments described herein, and are not intended to limit the scope of the invention.

[0117] It is understood that in the various embodiments described in this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments described in this specification.

[0118] It is understood that the various implementation methods described in this specification can be implemented individually or in combination, and the implementation methods in this specification are not limited in this respect.

[0119] Unless otherwise stated, all technical and scientific terms used in the embodiments of this specification have the same meaning as commonly understood by one of ordinary skill in the art. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the scope of this specification. The term "and / or" as used in this specification includes any and all combinations of one or more of the associated listed items. The singular forms "a," "the," and "the" as used in the embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0120] It is understood that the processor in the embodiments of this specification can be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method embodiments can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this specification. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this specification can be directly implemented by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above methods.

[0121] It is understood that the memory in the embodiments of this specification may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. Non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may be random access memory (RAM). It should be noted that the memory in the systems and methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0122] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this specification.

[0123] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the aforementioned method implementations, and will not be repeated here.

[0124] The above description is merely a specific embodiment of this specification, but the scope of protection of this invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this specification should be included within the scope of protection of this specification. Therefore, the scope of protection of this invention should be determined by the scope of the claims.

Claims

1. A method for quality monitoring in laser printing, characterized in that, The method includes: Acquire the optical detection signal of the molten pool corresponding to each laser trajectory segment during the laser printing process; Based on the process parameter information corresponding to each laser trajectory segment, the process category to which the laser trajectory segment belongs is determined, and a corresponding line type number is assigned to different process categories; Statistical analysis is performed on the optical detection signals of the molten pool for at least a portion of the laser trajectory segments belonging to the same wire type number, and a quality monitoring benchmark corresponding to the wire type number is established. Based on the quality monitoring benchmark corresponding to the line type number of each laser trajectory segment, anomaly determination is made on the molten pool optical detection signal of the laser trajectory segment.

2. The quality monitoring method according to claim 1, characterized in that, The process parameter information includes the actual laser power signal and the real-time scanning rate; The step of determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning corresponding line type numbers to different process categories, includes: If the layer data used as printing planning parameters cannot be obtained, the actual laser power signal and galvanometer coordinate signal of each laser trajectory segment are obtained, and the real-time scanning rate corresponding to each laser trajectory segment is calculated based on the galvanometer coordinate signal. The process category of the laser trajectory segment is determined based on the actual laser power signal and the real-time scanning rate of each laser trajectory segment, and a corresponding line number is assigned to each process category.

3. The quality monitoring method according to claim 2, characterized in that, The step of determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning corresponding line type numbers to different process categories, includes: A two-dimensional feature vector is constructed based on the actual laser power signal and the real-time scanning rate for each laser trajectory segment; Cluster analysis is performed on the two-dimensional feature vectors of all laser trajectory segments to dynamically form K cluster centers; Based on the distribution of the cluster centers, different laser trajectory segments are divided into K process categories, and a unique corresponding line number is assigned to each process category, where K is greater than or equal to 1.

4. The quality monitoring method according to claim 3, characterized in that, The cluster analysis also includes: The cluster centers are dynamically updated according to a preset time period or a preset number of layers to compensate for process parameter drift caused by laser thermal drift or long-term printing.

5. The quality monitoring method according to claim 1, characterized in that, The process parameter information includes the laser power and scanning rate of each laser trajectory segment in the layer data; The step of determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning corresponding line type numbers to different process categories, includes: If layer data containing preset process parameters is obtained, the process categories corresponding to all different combinations of laser power and scanning rate in the layer data are statistically analyzed in advance, and a mapping table between process categories and line types is established. During the laser printing process, online matching is performed in the mapping table based on the actual laser power signal and real-time scanning rate of the current laser trajectory segment acquired in real time. If the actual laser power signal and real-time scanning rate of the current laser trajectory segment successfully match any process category in the mapping table, then the corresponding line type number is assigned to the current laser trajectory segment. If a match fails, a new process category is dynamically created for the current laser trajectory segment, and a temporary line number is assigned.

6. The quality monitoring method according to claim 5, characterized in that, After the current layer of the laser printing process is completed, it also includes: The dynamically created temporary line number is merged or mapped with the theoretical line number in the mapping table.

7. The quality monitoring method according to claim 1, characterized in that, The step of determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning corresponding line type numbers to different process categories, includes: If layer data containing preset process parameters is obtained, the number of cluster categories K is determined in advance based on the number of process categories corresponding to the combination of laser power and scanning rate in the layer data; where K is a positive integer. Based on the number of cluster categories K, cluster analysis is performed on the laser trajectory segment by combining the actual laser power signal and real-time scanning rate to obtain K cluster results; The clustering results are matched one-to-one with the process categories in the layer data, and a corresponding line number is assigned to each clustering result.

8. The quality monitoring method according to claim 1, characterized in that, The step of determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning corresponding line type numbers to different process categories, includes: Multiple laser power ranges and multiple scanning rate ranges are preset, and a correspondence is established between the process category corresponding to different combinations of power ranges and rate ranges and the line type number; During the laser printing process, the combination of the actual laser power signal and the real-time scanning rate corresponding to each laser trajectory segment is determined, and the corresponding line number is assigned.

9. The quality monitoring method according to claim 1, characterized in that, The method further includes: If multiple laser heads are used simultaneously in the laser printing process, the laser head serial number is used as the isolation dimension. For each laser head, the process category, line type, and quality monitoring benchmark of each laser trajectory segment of the laser head are independently determined.

10. The quality monitoring method according to claim 1, characterized in that, The method further includes: Based on the process parameter boundary information in the galvanometer coordinate signal and / or layer data, the position points of laser trajectory segments belonging to the same line type are mapped to the spatial region to aggregate and form at least one connected process region.

11. The quality monitoring method according to claim 10, characterized in that, The step of statistically analyzing the molten pool optical detection signals for all laser trajectory segments belonging to the same wire type number and establishing a quality monitoring benchmark corresponding to the wire type number includes: For each of the connected process regions, the baseline value of the molten pool optical detection signal of all laser trajectory segments in the connected process region is statistically analyzed, and the baseline value is used as the quality monitoring baseline corresponding to the connected process region. Accordingly, the step of determining anomalies in the molten pool optical detection signal of each laser trajectory segment based on the quality monitoring benchmark corresponding to the line type number to which each laser trajectory segment belongs includes: Calculate the deviation between the optical detection signal of the molten pool of a single laser trajectory segment and the reference value of the optical detection signal of the molten pool in the corresponding process area; If the deviation value exceeds a preset threshold, it is determined that there is a defect at the position point of the laser trajectory segment.

12. The quality monitoring method according to claim 1, characterized in that, The method further includes: After acquiring the molten pool optical detection signal, actual laser power signal, and galvanometer coordinate signal corresponding to each laser trajectory segment during the laser printing process, the molten pool optical detection signal, actual laser power signal, and galvanometer coordinate signal corresponding to the same laser trajectory segment are time-stamped and interpolated for alignment.

13. The quality monitoring method according to claim 1, characterized in that, The process parameter information includes the actual laser power signal and real-time scanning rate, and / or, the laser power and / or scanning rate used as printing planning in the layer data; The step of determining the process category to which the laser trajectory segment belongs based on the process parameter information corresponding to each laser trajectory segment, and assigning corresponding line type numbers to different process categories, includes: When the actual scanning rate or the amount of change in the actual scanning rate during the laser printing process is less than a preset threshold, the actual laser power signal of each laser trajectory segment is used as a one-dimensional feature vector. Alternatively, when the actual laser power or the amount of change in actual laser power during the laser printing process is less than a preset threshold, the actual scanning rate of each laser trajectory segment is used as a one-dimensional feature vector. Alternatively, when the laser printing process plan only adjusts the laser power without adjusting the scanning rate, the actual laser power signal of each laser trajectory segment is used as a one-dimensional feature vector. Alternatively, when the laser printing process plan only adjusts the scanning rate without adjusting the laser power, the actual scanning rate of each laser trajectory segment is used as a one-dimensional feature vector. Cluster analysis is performed on the one-dimensional feature vectors of all laser trajectory segments to dynamically form K cluster centers; Based on the distribution of the cluster centers, different laser trajectory segments are divided into K process categories, and a corresponding line number is assigned to the process category to which each laser trajectory segment belongs.

14. The quality monitoring method according to claim 1, characterized in that, The molten pool optical detection signal includes at least one of the following four signals: visible light signal, infrared light signal, reflected light signal, and actual laser power signal; The step of determining anomalies in the optical detection signal of the molten pool in the laser trajectory segment includes: extracting at least one of the following from the optical detection signal of the molten pool in the laser trajectory segment: molten pool brightness, peak energy amplitude, waveform width, and thermal radiation intensity, as a detection feature for determining anomalies against the quality monitoring benchmark.

15. A quality monitoring system for laser printing, characterized in that, include: Multiple optical detection devices are used to collect optical detection signals generated by the molten pool during laser printing; A laser power detection device is used to acquire the actual laser output power signal in real time. A galvanometer coordinate detection device is used to acquire galvanometer coordinate signals in real time. An industrial control computer is communicatively connected to the multi-optical detection device, the laser power detection device, and the galvanometer coordinate detection device, respectively, and is used to execute the quality monitoring method as described in any one of claims 1 to 14.

16. The quality monitoring system according to claim 15, characterized in that, The multi-optical detection device includes a coaxial detection optical path or a paraxial detection optical path, used to simultaneously acquire at least one of visible light signals, infrared light signals, and reflected light signals.

Images