A laser drilling process monitoring and regulation combined protection method and system

Abstract

The application provides a laser drilling process monitoring and regulation combined protection method and system, which comprises the following steps: collecting signal evolution information, current hole depth and motion trajectory information in a laser drilling process; performing feature extraction on the obtained signal evolution information to obtain time sequence characteristic values evolving with time, and constructing a penetration time prediction model according to the obtained time sequence characteristic values; combining the obtained current hole depth and motion trajectory information with the time sequence characteristic values to construct a hole drilling stage identification model; comprehensively constructing a state identification model by combining a penetration time prediction model prediction result with a hole drilling stage identification model identification result confidence at different moments; identifying a current hole drilling processing stage according to the obtained state identification model, controlling a laser processing process according to an identification result, and adopting different hole drilling strategies in different hole drilling stages; and the application can ensure efficient, high-precision and no-wall-damage processing of the laser drilling process.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of laser processing manufacturing, and particularly relates to a laser drilling process monitoring and regulation combined protection method and system. BACKGROUND

[0002] Laser processing has the characteristics of high processing precision, high processing quality, almost any material can be processed, and non-contact processing. In laser processing, due to its non-contact processing and Gaussian light transmission characteristics, the processing depth cannot be controlled by the shaft feed, and the laser near the focal point range will still cause damage to the material. At the same time, the transmission of light in the laser processing process is easily disturbed by the structure evolution, plasma eruption and other phenomena in the processing process, resulting in fluctuations in the processing process, and problems in processing repeatability and stability. Therefore, when processing hollow cavity workpieces (such as blades and oil nozzles), there is often a problem of over-ablation, which leads to poor processing quality or reduced service life of the workpiece, which is not tolerated in most cases.

[0003] The current methods to solve the problem of over-ablation mainly include two types: cavity material filling and laser processing process monitoring. The cavity material filling is to fill materials in the hollow cavity to weaken or even eliminate the effect of laser energy, and the laser processing process monitoring is to monitor the drilling state in real time and regulate the processing process according to different drilling states. However, the current cavity material filling cannot feedback the protection state and is difficult to achieve protection for narrow cavities. The signals obtained by the current processing process monitoring are often indirectly related to the processing process, and it is difficult to achieve accurate control of the processing process, which seriously hinders the application of laser processing on workpieces with hollow cavities and limits the development of laser processing technology. SUMMARY

[0004] The present application aims to provide a laser drilling process monitoring and regulation combined protection method and system, which solves the defect of poor processing quality in the current laser processing.

[0005] In order to achieve the above-mentioned purpose, the technical solution adopted by the present application is as follows:

[0006] The present application provides a laser drilling process monitoring and regulation combined protection method, which comprises the following steps:

[0007] Step 1: collecting signal evolution information, current hole depth and motion trajectory information in the laser drilling processing process;

[0008] Step 2: extracting features from the signal evolution information obtained in step 1 to obtain time sequence feature values that evolve with time, and constructing a penetration time prediction model according to the obtained time sequence feature values;

[0009] Step 3, combine the current hole depth and motion trajectory information obtained in step 1 with the time sequence characteristic value to construct a hole forming stage recognition model;

[0010] Step 4, combine the penetration time prediction model prediction result with the hole forming stage recognition model recognition result confidence at different times to construct a state recognition model;

[0011] Step 5, according to the obtained state recognition model to determine the current hole forming stage, and according to the determination result to control the laser processing process, and different hole forming strategies are adopted in different hole forming stages.

[0012] Preferably, in step 2, a penetration time prediction model is constructed according to the obtained time sequence characteristic value, and the specific method is:

[0013] The hole forming process is set to four stages, which are the non-penetration stage, the early penetration stage, the hole forming stage and the hole completion stage;

[0014] According to the evolution process of the set non-penetration stage and early penetration stage, the residual hole penetration time value corresponding to the time sequence characteristic value at different times is obtained;

[0015] The obtained time sequence characteristic value is combined with the laser parameters and material parameters as the training features, and the residual hole penetration time value is used as the training label, and an autoregressive deep learning method is used to construct a penetration time prediction model.

[0016] Preferably, in step 3, the current hole depth and motion trajectory information obtained in step 1 are combined with the time sequence characteristic value to construct a hole forming stage recognition model, and the specific method is:

[0017] The current hole depth and motion trajectory information obtained in step 1 are fused to obtain the depth information of a specific point on the hole forming trajectory;

[0018] Form three-dimensional point cloud information according to the obtained depth information;

[0019] Filter, segment and fit the obtained three-dimensional point cloud information in turn to obtain a three-dimensional point cloud model of the hole;

[0020] Feature extraction is performed on the three-dimensional point cloud model of the hole to obtain the appearance features such as the entrance and exit diameters of the hole;

[0021] Combine the depth features, appearance features and time sequence characteristic values of the hole, and use the depth features and appearance features of the hole corresponding to different hole forming stages as labels to construct a hole forming stage recognition model by using a machine learning method.

[0022] Preferably, in step 4, a state recognition model is constructed according to the combination of the penetration time prediction model and the hole forming stage recognition model, and the specific combination method is:

[0023] Input the temporal feature values ​​and morphological features corresponding to different times into the hole-making stage identification model established in step 3, so as to obtain the confidence level of the hole-making stage identification results at different times in real time.

[0024] The time series feature values ​​corresponding to different times are input into the penetration time prediction model established in step 2 to predict the remaining penetration time corresponding to the current time in real time.

[0025] The confidence scores of the hole-making stage identification results at different times are weighted and fused to obtain the comprehensive confidence score at the current time.

[0026] The remaining penetration time predicted at the current moment is probabilistically fused with the comprehensive confidence level to obtain the state recognition model.

[0027] Preferably, in step 5, the different hole-making strategies are as follows:

[0028] When the current hole machining stage is the non-penetration stage, the machining parameters are adjusted with machining efficiency as the target.

[0029] When the current hole processing stage is the early penetration stage and the hole forming stage, the goal is to control the hole depth by adjusting the light field distribution of the laser processing.

[0030] When the hole-making stage transitions from the hole-forming stage to the hole-completion stage, processing is stopped.

[0031] Preferably, when the current hole machining stage is the non-penetration stage, the machining parameters are adjusted with machining efficiency as the target. The specific method is as follows:

[0032] The current hole depth, laser power, and focal position are used as inputs, and the actual ablation rate change is used as the output. A neural network is used to establish a correlation model between the ablation rate change and the current hole depth, laser power, and focal position. The processing parameters are adjusted using this correlation model.

[0033] Preferably, when the current hole processing stage is the early penetration stage and the hole forming stage, the laser processing light field distribution is adjusted with the goal of controlling the hole depth. The specific method is as follows:

[0034] By using coaxial vision to locate the coordinates of the machining point, the Gaussian machining optical path is switched to a beam shaping optical path;

[0035] A Gaussian beam is shaped into an axially ablation-constrained Bessel beam using a beam shaping device.

[0036] The shaped beam is used to reshape the hole.

[0037] A laser drilling process monitoring and regulation combined protection system comprises a laser, a beam transmission system, a dichroic mirror, a sensing system, a beam switching system, a beam shaping system, a focusing lens and a scanning galvanometer, wherein the laser beam output by the laser is incident to the beam switching system through the dichroic mirror after passing through the beam transmission system; the output light path of the beam switching system is divided into two paths, one of which is incident to the focusing lens for focusing after passing through the beam shaping system, and the other is incident to the scanning galvanometer for focusing; the laser beams output by the focusing lens and the scanning galvanometer act on a target material.

[0038] Preferably, the beam shaping system comprises a beam shaping element and a 4f system, wherein the output light path of the beam switching system is incident to the 4f system to form Bessel light through the beam shaping element.

[0039] Compared with the prior art, the present application has the following advantages:

[0040] The laser drilling process monitoring and regulation combined protection method provided by the present application divides the drilling process into four stages, establishes a state recognition model for distinguishing different drilling stages by extracting the differences in the signal distribution of the processing process and reconstructing and extracting the features of the three-dimensional morphology of the hole, adopts different drilling strategies according to different drilling stages, adjusts the longitudinal energy distribution of the Gaussian beam, combines visual positioning to complete accurate switching of the drilling light path and the shaping light path, and ensures efficient, high-precision and no-wall-damage processing of the laser drilling process.

[0041] The laser drilling process monitoring and regulation combined protection system provided by the present application realizes coaxial coupling of multiple sensors and processing light paths through the combination of a dichroic mirror and a beam splitter, realizes real-time monitoring of the laser drilling process and real-time identification of the laser drilling stage, adjusts the longitudinal energy distribution of the Gaussian beam by using an optical shaping element, controls the laser ablation depth within a certain focal depth range, ensures no-wall-damage processing, combines visual positioning technology to realize accurate switching of the Gaussian beam drilling light path and the beam shaping and shaping light path, meets different drilling strategies required in different drilling stages, and ensures efficient, high-precision and no-wall-damage processing of the laser drilling process. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 is a method flowchart of the present application;

[0043] Figure 2 is a method flowchart of the present application for acquiring laser drilling depth information;

[0044] Figure 3 is a schematic diagram of drilling stage division and calibration of the present application;

[0045] Figure 4 is a state recognition model establishment flowchart of the present application;

[0046] Figure 5 is a flow chart of the method for extracting the three-dimensional morphology and features of holes according to the present application;

[0047] Figure 6 is a flow chart of the method for hole making strategy according to the present application;

[0048] Figure 7 is a flow chart of the method for visual positioning according to the present application;

[0049] Figure 8 is a general schematic diagram of the system according to the present application;

[0050] Figure 9 is a schematic diagram of the sensing system according to the present application;

[0051] Figure 10 is a schematic diagram of the interferometric measurement system according to the present application;

[0052] Figure 11 is a schematic diagram of the beam shaping system according to the present application. DETAILED DESCRIPTION

[0053] In the following description, for purposes of explanation and not limitation, specific details are set forth, such as particular sequences of steps, techniques, architectures, etc. in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

[0054] It is to be understood that the terminology “includes”, “has”, “holds”, “contains” and / or “comprising”, when used in this specification and in the following claims, indicates the presence of the described features, integers, steps, operations, elements, and / or components but does not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0055] It is also to be understood that the terminology “and / or” when used in this specification and in the following claims, refers to at least one of the items, or any combination of one or more of the items, and includes all possible combinations of the items.

[0056] As used in this specification and in the claims, the term “if’ can be construed to mean “when” or “once” or “in response to determining” or “in response to detecting,” depending on the context. Similarly, the phrase “if it is determined” or “if [a described condition or event] is detected” can be construed to mean “once it is determined” or “in response to determining” or “once [the described condition or event] is detected” or “in response to detecting [a described condition or event],” depending on the context.

[0057] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0058] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0059] The present invention will now be described in further detail with reference to the accompanying drawings.

[0060] The purpose of this invention is to provide a combined protection method and system for monitoring and controlling the laser hole-making process, so as to achieve efficient, high-quality and non-damaging processing in the laser hole-making process, enhance the controllability of the laser processing process, and promote the application of laser processing technology in the field of hollow cavity material processing.

[0061] like Figures 1 to 6 As shown, the present invention provides a combined protection method for monitoring and controlling the laser hole-making process, comprising the following steps:

[0062] Step 1: Use a sensing system to collect historical data of the target material during the laser hole-making process. The historical data includes signal evolution information, current hole depth, and motion trajectory information.

[0063] Step 2: The hole-making process is set into four stages: the non-penetration stage, the early penetration stage, the hole forming stage, and the hole completion stage.

[0064] like Figure 3 As shown, the calibration method is based on the photodiode sensor signal and the hole shape. The specific calibration method is as follows:

[0065] When the workpiece is not penetrated by the laser, the laser cannot be emitted from the back of the workpiece, and the difference in the amplitude of the photodiode signal is at its lowest level, which is considered as the non-penetration stage.

[0066] When the workpiece is just penetrated by the laser, the exit point of the workpiece is extremely small, and only a small part of the laser is emitted from the back of the workpiece. The differential amplitude of the photodiode signal is at its highest level, which is regarded as the early penetration stage.

[0067] When the workpiece is pierced by the laser for a period of time, the workpiece outlet becomes larger, the laser stably emits from the back of the workpiece, and the difference in the amplitude of the photodiode signal is in a stable change state, which is regarded as a hole forming stage;

[0068] When the workpiece is pierced by the laser for a period of time, the workpiece outlet becomes larger, the laser stably emits from the back of the workpiece, and the difference in the amplitude of the photodiode signal is in a stable change state, which is regarded as a hole forming stage;

[0069] The signal evolution information obtained in step 1 is subjected to feature extraction to obtain time sequence feature values evolving over time. According to the evolution process set for the non-penetration stage and the early penetration stage, the time sequence feature values corresponding to the residual hole penetration time values at different times are obtained. The obtained time sequence feature values, laser parameters and material parameters are combined as training features, and the residual hole penetration time value is used as a training label. An autoregressive deep learning method is used to construct a penetration time prediction model.

[0070] The time sequence feature values include mean value, peak value, variance and the like.

[0071] The laser parameters include laser power, repetition frequency and the like.

[0072] The material parameters include the light absorption rate, thermal conductivity and thickness of the current material and the like.

[0073] The penetration time prediction model will continuously combine the collected data to predict the current hole penetration time.

[0074] As shown in Figure 4 The current hole depth (Z-axis information) obtained in step 1 and the motion trajectory information (XY plane information) are fused to obtain the depth information of a specific point on the hole forming trajectory, thereby forming a three-dimensional point cloud (X, Y, Z) information. The obtained three-dimensional point cloud information is sequentially subjected to filtering, segmentation and fitting to establish a three-dimensional point cloud model of the hole.

[0075] The depth feature and the topographic feature of the hole are extracted based on the three-dimensional point cloud model, and the topographic feature includes the diameter of the inlet and outlet.

[0076] The depth feature and the topographic feature of the hole are combined with the time sequence feature value, the depth feature and the topographic feature of the hole corresponding to different hole forming stages are used as labels, and a machine learning method is used to establish a hole forming stage recognition model to realize the recognition of the non-penetration stage, the early penetration stage, the hole forming stage and the hole completion stage according to the signal features collected in the machining process.

[0077] According to the historical data corresponding to different time obtained in step 1, time sequence characteristic values and topographic features corresponding to different time are obtained; the time sequence characteristic values and the topographic features corresponding to different time are input into the drilling stage recognition model established in step 3, so as to obtain the drilling stage recognition result confidence of different time in real time;

[0078] The time sequence characteristic values corresponding to different time are input into the penetration time prediction model established in step 2, so as to predict the remaining penetration time corresponding to the current time in real time;

[0079] The drilling stage recognition result confidences of different time are weighted and fused, so as to obtain the comprehensive confidence of the current time;

[0080] The obtained comprehensive confidence and the predicted drilling remaining penetration time output by the penetration time prediction model are probabilistically fused, so as to establish a state recognition model.

[0081] Step 3, laser processing parameters are set and drilling of the target material is carried out, current processing stage parameters are obtained, and the related parameters are input into the state recognition model constructed in step 2, so as to judge the current drilling processing stage in real time, and the laser processing process is regulated according to the judgment result, and different drilling strategies are adopted in different drilling stages.

[0082] As shown in Figure 5 , when the current hole processing stage is the non-penetration stage, a high-speed processing strategy is adopted, and the processing parameters are adjusted with the processing efficiency as the target, and the specific method is as follows:

[0083] The current drilling depth, laser power and focal point position are taken as inputs, and the actual ablation rate change is taken as output; a neural network is used to establish a correlation model between the ablation rate change and the current drilling depth, laser power and focal point position, so as to predict the ablation rate of the next pulse or the next cycle time according to the input ablation rate change, and the laser focal point position is taken as the controlled object, the actual ablation rate change is taken as the output, the drilling depth is measured in real time by an interferometric measurement system and converted into real-time ablation rate change as feedback, and the laser ablation rate change is always positive or remains unchanged by adjusting the focal point position in real time. Based on the above idea, a laser drilling high-speed processing control model is established by combining the model predictive control method, so as to realize high-efficiency drilling.

[0084] When the current hole processing stage is the early penetration stage and the hole forming stage, a quantitative removal processing strategy is adopted, the light field distribution of laser processing is adjusted with the hole depth control as the target, the traditional Gaussian laser distribution with hyperbolic transmission characteristics is adjusted to a Bessel-like beam distribution with limited axial transmission, and the hole is reshaped without protective material and the processing depth is as constant as possible, avoiding damage to the back wall.

[0085] The current hole processing stage is in the hole completion stage, and the processing process is stopped.

[0086] Specifically, as shown in step 1, the current hole depth in the processing process is acquired in real time by using the sensing system, and the specific method flow is as follows: Figure 2

[0087] S101, the interferometer in the interferometric measurement system is used to acquire the interference signal in the laser hole forming process in real time;

[0088] S102, the interference signal is processed to obtain the depth information, wherein:

[0089] The interference signal obtained in S101 is preprocessed by using direct current removal and wave number linearization in sequence to obtain the preprocessed interference signal;

[0090] Then, the preprocessed interference signal is demodulated by using Fourier transform to obtain the depth information of each point in the Z direction;

[0091] S103, combined with the motion trajectory information, considering the influence of the processing process debris, plasma eruption and hole shape evolution on data acquisition, the depth information recovered from the collected interference signal at the same position is denoised, the depth information obtained by scanning multiple time frames is averaged, and further filtering processing is performed on the depth information to reduce noise points, and then the remaining depth points are clustered and segmented to extract the depth information in the processing process. The maximum value of the depth information is extracted as the current hole depth.

[0092] In step 2, according to the acquired depth information and motion trajectory information, the three-dimensional morphology of the hole is reconstructed, and the depth feature and the morphology feature are proposed, and the specific method flow is as follows:

[0093] S601, using the position comparison output function of the motion controller, when the motion shaft or the galvanometer shaft reaches a specific position, the motion controller outputs a signal to trigger the interferometric measurement sensing system, and the current hole depth information is acquired in real time, so as to acquire the processing point plane coordinates (x, y) and the depth information z, that is, the three-dimensional coordinate point cloud data (x, y, z)

[0094] S602, the point cloud acquired in the processing process is processed, the noise points are removed by point cloud filtering, and the point cloud data is segmented according to the similarity of the point cloud, and further the point cloud data is fitted into a plane by a fitting algorithm, and the point cloud is reconstructed into a three-dimensional model by establishing the three-dimensional topological relationship of the point cloud data, so as to realize the real-time reconstruction of the three-dimensional morphology in the hole forming process

[0095] S603, considering the key features of the hole shape in the laser hole forming, such as hole entrance diameter, hole exit diameter, hole roundness, etc., the above morphology features are extracted based on the reconstructed three-dimensional model​

[0096] In step 3, different drilling strategies are adopted according to different drilling stages, and the specific drilling strategy is:

[0097] S701, real-time acquisition of sensor signals in the laser drilling process, and identification of the drilling stage according to the state recognition model;

[0098] S702, when the laser processing process is in the non-penetration stage, a high-speed processing strategy is adopted, that is, the current drilling depth, laser power, focal point position and other parameters are input, and a correlation model between the ablation rate change and the current drilling depth, laser power and focal point position is established by using a neural network, so that the ablation rate of the next pulse or the next cycle time is predicted according to the input ablation rate change, the laser focal point position is taken as the controlled object, and the actual ablation rate change is taken as the output. The drilling depth is measured in real time by an interferometric measurement system and converted into a real-time ablation rate change as feedback, and the focal point position is adjusted in real time to keep the laser ablation rate change always positive or unchanged. Based on the above idea, combined with the model predictive control method, a laser drilling high-speed processing control model is established to realize high-efficiency drilling;

[0099] S703, when the laser processing process is converted from the non-penetration stage to the early penetration stage, the processing strategy is switched to the quantitative removal processing strategy, that is, the coaxial vision positioning processing point coordinates are used to switch the Gaussian processing light path to the beam shaping light path, and the shaped beam is used to shape the hole. At the same time, the processing depth is measured in real time by an interferometric measurement system to ensure no damage to the wall while realizing high-quality processing;

[0100] S704, when the laser processing process is in the hole completion stage, stop processing.

[0101] In step 3, the current processing stage parameters are obtained, and the specific method is:

[0102] S31, obtaining signal evolution information, current hole depth and motion trajectory information in the laser drilling process;

[0103] S32, obtaining time sequence characteristic values according to the signal evolution information; and establishing a three-dimensional point cloud model combined with the current hole depth and the motion trajectory information;

[0104] S33, extracting the hole topography feature by using the three-dimensional point cloud model, and the hole topography feature, hole depth and time sequence characteristic value are the current processing stage parameters.

[0105] As shown in Figure 7 In step 3, the traditional Gaussian laser distribution with hyperbolic transmission characteristics is adjusted to a Bessel-like beam distribution with limited axial transmission, and the specific method is:

[0106] S301, calibrate the laser processing system and the camera, the multi-axis motion table coordinate system in the laser processing system is known, and the position information of different points on the motion table is collected by the camera, so that the relationship model of any pixel on the image in the camera and the processing coordinates is obtained, and the conversion model of the motion table coordinate system and the camera coordinate system is established by solving the relationship model;

[0107] Similarly, in combination with the motion of the motion table, the laser processes at different positions, and the coordinates corresponding to the processing positions are obtained by using the camera, so that the correlation model between the laser pose and the camera coordinate system is established, and the conversion model between the three coordinate systems is obtained.

[0108] S302, the image of the processing area is collected by the camera, the image is preprocessed such as enhancement and filtering, the center coordinates of the hole in the processing area are extracted, and the center coordinates are converted to the multi-axis motion table coordinates;

[0109] S303, the Gaussian beam transmission light path is converted to the beam shaping light path by using the beam switching system, the difference between the laser pose of the Gaussian beam transmission light path and the laser pose of the beam shaping light path is considered, which is converted to the difference between the position coordinates of the multi-axis motion table, the workpiece is moved by using the multi-axis motion table, so that the center of the laser focal point position of the beam shaping light path coincides with the center of the processing area, and the hole is further shaped.

[0110] The specific shaping principle of the beam shaping light path is:

[0111] S3031, the phase distribution of the Gaussian beam shaping into the axial ablation limited Bessel-like beam is calculated by using the analytical method, such as calculating the corresponding phase distribution according to the following formula:

[0112]

[0113] Wherein, R is the right angle of the diffraction plane, Z1 and Z2 are the starting point and the end point of the beam respectively;

[0114] S3032, the phase distribution is input to the spatial light modulator or the corresponding diffractive optical element is made, the axial distribution of the beam is controlled, and the focal depth is limited in a specific area by combining the 4f system to meet the processing requirements.

[0115] In step 1, the target material can be engine blades, oil nozzles and other materials with hollow cavities.

[0116] In step 1, the sensing system mainly includes an interferometric measurement system, a camera and a photodiode, etc., wherein the interferometric measurement system can measure the depth of the hole, the camera can be a black and white camera, a color camera or a thermal imaging camera, etc., and the photodiode can be replaced with a spectrometer or the like.

[0117] In step 1, the motion trajectory information mainly refers to the position coordinates of the current multi-axis motion platform, which can be obtained by reading the grating feedback or the system controller.

[0118] In step S302, the extraction of the center position of the machining area hole is mainly completed by using the edge extraction and template matching method.

[0119] As shown in Figure 7 The laser drilling process monitoring and regulation combined protection system provided by the application, comprising a laser 1, a beam transmission system 2, a dichroic mirror 3, a sensing system 4, a beam switching system 5, a beam shaping system 6, a focusing lens 7, a multi-axis motion platform 10, a scanning galvanometer 11, a system controller 12 and an industrial computer 13, wherein the laser beam output by the laser 1 passes through the beam transmission system 2 and then passes through the dichroic mirror 3 to be incident on the beam switching system 5; the output light path of the beam switching system 5 is divided into two paths, one of which is incident on the focusing lens 7 for focusing after passing through the beam shaping system 6, and the other is incident on the scanning galvanometer 11 for focusing; the laser beams output by the focusing lens 7 and the scanning galvanometer 11 act on the target material 8 on the multi-axis motion platform 10.

[0120] The beam switching system 5 is used to control the switching of the laser transmission light path.

[0121] The beam shaping system 6 is used to adjust the longitudinal energy distribution of the Gaussian beam to control the ablation depth within a certain range.

[0122] The system controller 12 is connected with the sensing system 4, the laser 1, the beam switching system 5, the scanning galvanometer 11 and the multi-axis motion platform 10, and is used to process the signals collected by the sensing system, feedback the processing state and cooperatively control the multi-axis motion platform, the scanning galvanometer, the laser and the like.

[0123] The industrial computer 13 is connected with the system controller 12 and is used to control the command logic of the system controller.

[0124] As shown in Figure 10 The beam shaping system 6 comprises a beam shaping element 25 and a 4f system 26, wherein the beam shaping element 25 is a spatial light modulator, a diffractive optical element, a tapered lens or a metasurface element; the high filament light is incident on the 4f system 26 through the beam shaping element 25 to form a Bessel light.

[0125] As shown in Figure 8As shown, the sensing system 4 includes a photodiode 14, a camera 16 and an interferometric measurement system 17, wherein the photodiode 14 is placed on the back of the target material for calibration of the laser drilling stage, the photodiode 14 is coaxially placed with the camera 16 through a beam splitter 15 for monitoring the timing change process of the plasma radiation signal in the drilling process; the camera is coaxially coupled with the interferometric measurement system 17 through a dichroic mirror 18 for monitoring the spatial distribution and positioning of the surface light radiation signal in the drilling process, and the interferometric measurement system 17 is coupled with the machining light through the dichroic mirror 3 for measuring the hole depth in the drilling process.

[0126] As shown in the figure, Figure 9 As shown, the interferometric measurement system 17 includes a probe light source 20, a collimator 122, a beam splitter 21, a collimator 224 and a signal detector 19, wherein the probe light source 20 emits a probe light beam which is incident on the beam splitter 21, the beam splitter 21 divides the light beam into two light beams, one of which is incident on the reference mirror 23 after passing through the collimator 122, and the other of which is coupled with the machining light beam after passing through the collimator 224, the dichroic mirror 18 and focusing on the target material, the sample light reflected from the target material and the reference light reflected from the reference mirror 23 return to the beam splitter 21 to interfere, and further incident on the signal detector 19; the signal detector 19 is used to acquire the interference spectrum signal generated after the sample light and the reference light interfere.

[0127] The probe light source 20 is a super radiation light emitting diode, a sweep frequency laser light source, a fiber laser or a semiconductor laser.

[0128] The signal detector 19 is a photodiode, a spectrometer or a balanced detector.

[0129] Although the laser, the beam switching system, the beam transmission system, the beam shaping system, the dichroic mirror, the sensing system, the scanning galvanometer, the focusing lens, the target material, the multi-axis motion platform, the system controller, the industrial computer, the interferometric sensing system, the photodiode, the camera, the probe light source, the collimator, the beam splitter, the signal detector, the state recognition model, the high-speed machining strategy, the quantitative removal machining strategy and the like are used more in the specification, the possibility of using other terms is not excluded, and the use of these terms is only for more convenient description of the essence of the application, and any additional limitation by interpreting them is contrary to the spirit of the application.

[0130] Example 1: without filler material

[0131] Taking laser processing of blade film cooling hole as an example, single crystal high-temperature alloy blade as target material, cavity thickness 0.5mm-3mm, mainly including the following steps:

[0132] Step 1, determine the processing parameters according to the hole processing diameter, depth, and inclination angle, process the blade, and the sensing system obtains the current processing stage parameters in real time;

[0133] Step 2, according to the calibrated laser drilling stage, the time sequence characteristic value and laser processing parameters in the obtained processing stage parameters, etc. Information, using autoregressive deep learning method to establish the penetration time prediction model; and, combine the hole depth feature, topographic feature and time sequence characteristic value, use the hole shape feature and time sequence characteristic corresponding to different drilling stages calibrated, use machine learning method to establish drilling stage identification model; The confidence of the identification result output by the drilling stage identification model at different times is weighted to obtain the comprehensive confidence of the current time identification result, and the prediction result of the penetration time prediction model is fused to form a state recognition model of the laser drilling stage.

[0134] Step 3, in the process of laser drilling, the current processing stage parameters are obtained in real time, the related parameters are input into the established state recognition model, and the state recognition model is used to identify the current drilling stage. In the non-penetration stage, a high-speed processing strategy is adopted, and the focal point position is adjusted in real time according to the measured hole depth and ablation rate for high-speed processing. After identifying that the drilling stage enters the early penetration stage, the drilling strategy is adjusted to a quantitative removal processing strategy.

[0135] Step 4, switch the Gaussian processing light path to the beam shaping light path by positioning the processing point coordinates with the coaxial vision, switch the processing light path to the beam shaping system with the beam switching system, identify the drilling center coordinates with the coaxial camera, and move the processing workpiece to the beam shaping processing light path position based on the established camera, motion table and laser position coordinate system conversion model. At this time, the shaped Bessel light is used to modify the hole, and the drilling depth is monitored in real time to ensure that it does not exceed the sum of the drilling depth and the cavity thickness; stop processing after monitoring that the laser drilling stage reaches the hole completion stage.

[0136] Comparative Example 1: with filler material

[0137] Taking laser processing of blade film cooling holes as an example, ceramic particles with a diameter of 50 μm to 200 μm are used as protective materials, single crystal high-temperature alloy blades are used as target materials, and the cavity thickness is 0.5 mm to 3 mm. The main steps include:

[0138] Step 1, according to the cavity thickness, ceramic particles with corresponding diameter are selected and filled into the blade cavity;

[0139] Step 2, determine the processing parameters according to the hole processing diameter, depth, and inclination angle, process the blade, and the sensing system obtains the current processing stage parameters in real time;

[0140] Step 3, according to the calibrated laser drilling phase, and the time sequence characteristic value and laser processing parameters in the acquisition processing phase information, an autoregressive deep learning method is used to establish a penetration time prediction model; and the obtained hole depth feature, topography feature and time sequence characteristic value are combined, the hole shape feature and time sequence characteristic corresponding to different drilling phases in the calibration are used, and a machine learning method is used to establish a drilling phase identification model; the identification result confidence of the drilling phase identification model output at different times is weighted to obtain the comprehensive confidence of the current time identification result, and the prediction result of the penetration time prediction model is fused to form a laser drilling phase state recognition model.

[0141] Step 4, in the laser drilling process, the current processing phase parameters are acquired in real time, the related parameters are input into the established state recognition model, and the current drilling phase is identified by the state recognition model; when in the non-penetration phase, a high-speed processing strategy is used, and the focal point position is adjusted in real time according to the measured hole depth and ablation rate for high-speed processing; after the drilling phase is identified to enter the early penetration phase, the drilling strategy is adjusted to a quantitative removal processing strategy.

[0142] Step 5, the coaxial vision positioning processing point coordinates are used to switch the Gaussian processing light path to the beam shaping light path, the processing light path is switched to the beam shaping system by the beam switching system, the coaxial camera is used to identify the drilling center coordinates, the camera, motion table and laser position coordinate system conversion model are established, and the workpiece is moved to the beam shaping processing light path position; at this time, the shaped Bessel light is used to shape the hole, and the drilling depth is monitored in real time to ensure that it does not exceed the sum of the drilling depth and the cavity thickness; after the laser drilling phase is monitored to reach the hole completion phase, the processing is stopped.

[0143] In combination with the drawings Figure 11 Further description of the beam shaping features of the present patent:

[0144] The processing laser generally presents a Gaussian distribution in the processing process, and its transmission in the axial direction is a hyperbolic transmission, which makes the laser have a non-contact processing characteristic. However, when the laser processes a hollow cavity material in the processing zone, it will continue to transmit to the inner wall of the cavity when penetrating the surface layer material, causing damage to the inner wall of the cavity, such as Figure 11 As shown in the first figure in the background art, without protective material and beam shaping, the inner wall of the cavity is easily damaged;

[0145] In order to solve this problem, filling protective material is a common means currently used in the industry, and the protective material is filled in the cavity to weaken or even eliminate the laser energy, such as Figure 11 As shown in the second figure in the background art. However, since it does not change the ablation characteristics of the laser itself, it will still cause damage to the protective material, and when the laser irradiation time is too long or the cavity is too narrow, it will still cause damage to the inner wall of the cavity;

[0146] Therefore, based on the principle of axial beam shaping regulation, the axial energy distribution of the laser is regulated to change the energy distribution of the axial transmission in space, so as to realize the controllable axial ablation depth of the laser. The traditional Bessel beam has the characteristic of long focal depth, but at the same time, its energy outside the focal depth is also limited, and the axial energy distribution of the Bessel beam is as follows Figure 11 As shown in the third figure, however, its energy utilization is low, and the beam transmission range is long, so it is difficult to realize the non-wall damage processing of the narrow cavity without protection material. Therefore, on this basis, the beam is further shaped, and the axial regulation is made into an axial ablation limited Bessel beam, as shown in the fourth figure Figure 11 As shown in the fourth figure, this makes it possible to realize the modification of different positions by controlling the laser feeding without damaging the back wall, and can ensure the non-wall damage processing in the narrow cavity.

[0147] It should be noted that any non-wall damage processing by regulating the axial beam energy distribution is difficult to deviate from the spirit and scope of the present application.

[0148] The above-described embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A combined protection method for monitoring and controlling the laser hole-making process, characterized in that, Includes the following steps: Step 1: Collect signal evolution information, current hole depth, and motion trajectory information during the laser hole-making process; Step 2: Extract features from the signal evolution information obtained in Step 1 to obtain time-series feature values ​​that evolve over time, and construct a time-through prediction model based on the obtained time-series feature values. Step 3: Combine the current hole depth and motion trajectory information obtained in Step 1 with the temporal feature values ​​to construct the hole-making stage identification model; Step 4: Combine the prediction results of the penetration time prediction model with the confidence levels of the identification results of the hole-making stage identification model at different times to construct the state identification model. Step 5: Determine the current hole-making stage based on the obtained state recognition model, and control the laser processing process according to the determination result. Different hole-making strategies are adopted in different hole-making stages. Specifically: When the laser processing is in the non-penetration stage, the current hole depth, laser power, and focal position are used as inputs. A neural network is used to establish a correlation model between the ablation rate change and the current hole depth, laser power, and focal position. The ablation rate of the next pulse or the next cycle time is predicted based on the input ablation rate change. The laser focal position is the controlled object, and the actual ablation rate change is the output. The hole depth is measured in real time by an interferometric measurement system and converted into a real-time ablation rate change as feedback. By adjusting the focal position in real time, the laser ablation rate change is kept positive or constant. Based on the above ideas, combined with model predictive control methods, a high-speed laser hole-making processing control model is established to achieve high-efficiency hole making. In step 3, the current hole depth and motion trajectory information obtained in step 1 are combined with the temporal feature values ​​to construct a hole-making stage identification model. The specific method is as follows: By fusing the current hole depth and motion trajectory information obtained in step 1, the depth information of a specific point on the hole drilling trajectory is obtained. Three-dimensional point cloud information is generated based on the obtained depth information; The obtained 3D point cloud information is filtered, segmented, and fitted sequentially to obtain the 3D point cloud model of the hole; Feature extraction is performed on the 3D point cloud model of the hole to obtain the inlet and outlet diameter morphology features of the hole; By combining the depth features, morphological features, and temporal features of the holes, and using the depth features and morphological features of the holes corresponding to different hole-making stages as labels, a hole-making stage identification model is constructed using machine learning methods.

2. The combined protection method for monitoring and controlling the laser hole-making process according to claim 1, characterized in that, In step 2, a penetration time prediction model is constructed based on the obtained time series feature values. The specific method is as follows: The hole-making process is set into four stages: the non-penetration stage, the early penetration stage, the hole forming stage, and the hole completion stage. Based on the evolution process of the non-penetration stage and the early penetration stage, the remaining hole-making penetration time value corresponding to the time sequence characteristic value at different times is obtained. The obtained time-series feature values ​​are combined with laser parameters and material parameters as training features, and the remaining hole-making penetration time value is used as training label. An autoregressive deep learning method is used to construct a penetration time prediction model.

3. The combined protection method for monitoring and controlling the laser hole-making process according to claim 1, characterized in that, In step 4, a state recognition model is constructed by combining the penetration time prediction model and the hole-making stage identification model. The specific combination method is as follows: The temporal feature values ​​and morphological features corresponding to different times are input into the hole-making stage identification model established in step 3 to obtain the confidence level of the hole-making stage identification results at different times in real time. The time series feature values ​​corresponding to different times are input into the penetration time prediction model established in step 2 to predict the remaining penetration time corresponding to the current time in real time. The confidence scores of the hole-making stage identification results at different times are weighted and fused to obtain the comprehensive confidence score at the current time. The remaining penetration time predicted at the current moment is probabilistically fused with the comprehensive confidence level to obtain the state recognition model.

4. The combined protection method for monitoring and controlling the laser hole-making process according to claim 1, characterized in that, In step 5, the different hole-making strategies are as follows: When the current hole machining stage is the non-penetration stage, the machining parameters are adjusted with machining efficiency as the target. When the current hole processing stage is the early penetration stage and the hole forming stage, the goal is to control the hole depth by adjusting the light field distribution of the laser processing. When the hole-making stage transitions from the hole-forming stage to the hole-completion stage, processing is stopped.

5. The combined protection method for monitoring and controlling the laser hole-making process according to claim 4, characterized in that, When the current hole processing stage is the early penetration stage and the hole forming stage, the goal is to control the hole depth by adjusting the laser processing light field distribution. The specific method is as follows: By using coaxial vision to locate the coordinates of the machining point, the Gaussian machining optical path is switched to a beam shaping optical path; A Gaussian beam is shaped into an axially ablation-constrained Bessel beam using a beam shaping device. The shaped beam is used to reshape the hole.

6. A combined protection system for monitoring and controlling laser hole-making process, characterized in that, Based on the laser hole-making process monitoring and control combined protection method according to claim 1, the system includes a laser (1), a beam transmission system (2), a dichroic mirror (3), a sensing system (4), a beam switching system (5), a beam shaping system (6), a focusing lens (7), and a scanning galvanometer (11). The laser beam output by the laser (1) passes through the beam transmission system (2) and then through the dichroic mirror (3) to be incident on the beam switching system (5). The output optical path of the beam switching system (5) is divided into two paths, one of which passes through the beam shaping system (6) and is incident on the focusing lens (7) for focusing, and the other path is incident on the scanning galvanometer (11) for focusing. The laser beams output by the focusing lens (7) and the scanning galvanometer (11) act on the target material (8).

7. The laser hole-making process monitoring and control combined protection system according to claim 6, characterized in that, The beam shaping system (6) includes a beam shaping element (25) and a 4f system (26), wherein the output optical path of the beam switching system (5) is incident on the 4f system (26) through the beam shaping element (25) to form Bessel light.

Images