Laser processing method and machine with roughness evaluation

Abstract

A laser processing method for cutting and / or drilling a workpiece (2), preferentially made of a metallic material, is disclosed, comprising at least the steps of: a) directing a laser beam (5) onto the workpiece (2) over a processing zone (7) of the workpiece (2) to perform the laser processing, b) performing a relative movement between the laser beam (5) and the workpiece (2), c) acquiring a plurality of acquired images (9) from the processing zone (7), each of which includes a high intensity region (24), d) determining the variation over time of one or more characteristic parameters from the high intensity region (24) of each of the plurality of processing zones (7), and e) estimating the roughness obtained during the laser processing according to the respective variation over time of the one or more characteristic parameters. During step e), at least one respective statistical parameter is determined from the variation over time of the one or more characteristic parameters, and then a continuous real-time estimate of the roughness is calculated depending on each determined statistical parameter.

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This patent application claims priority to Italian Patent Application No. 102022000022038, filed October 25, 2022, the entire disclosure of which is incorporated herein by reference.

[0002] The present invention relates to a laser processing method, preferentially for performing cutting and / or drilling and / or welding of workpieces and / or for performing additive manufacturing to obtain workpieces. Preferentially, the present invention relates to a laser processing method that makes it possible to estimate, preferentially in real time, the roughness caused by the laser processing.

[0003] Furthermore, the invention relates to a laser processing device configured to carry out a laser processing method that allows the roughness caused by the laser processing to be estimated (preferentially in real time). [Background technology]

[0004] Laser processing devices are known for cutting and / or drilling workpieces, for example. A typical laser processing device includes a laser beam source, a workpiece support, an optical group for controlling the focal position of the laser beam, and a movement device for performing relative movement between the laser beam and the workpiece.

[0005] Laser processing performed in a laser processing device, such as cutting a workpiece, can result in some roughness. The actual roughness achieved during laser processing depends not only on the specific material of the workpiece, but also on one or more process parameters, such as the intensity of the laser beam and / or the speed of relative movement between the laser beam and the workpiece.

[0006] To date, the roughness obtained from laser processing can be determined in a laboratory after the laser processing is completed, so that it is possible to control whether the determined roughness corresponds to the desired value. If the roughness is unacceptable, the operator must update one or more process parameters to obtain a workpiece with the desired roughness. This means that some processed workpieces do not correspond to the desired standards and must be discarded or subsequently reprocessed manually.

[0007] Therefore, a need exists in the art for further improvements in laser processing methods and laser processing apparatus. Summary of the Invention

[0008] The object of the present invention is to provide a laser processing method and a laser processing device that allow improving known solutions to obtain an estimate (preferentially in real time) of the roughness caused by laser processing.

[0009] Preferentially, it is an object of the present invention to provide a laser processing method and a laser processing device that make it possible to obtain an estimate of roughness and to control the method according to the estimated roughness.

[0010] The above object is achieved by the present invention, which relates to a laser processing method as defined in independent claim 1. Preferred alternative embodiments are protected by the respective dependent claims.

[0011] The above object is also achieved according to the invention by a laser processing device as defined in independent claim 15. [Brief explanation of the drawings]

[0012] For a better understanding of the invention, preferred embodiments will now be described, by way of non-limiting example, with reference to the accompanying drawings, in which:

[0013] [Figure 1] 1 is a diagram showing a schematic and partial view of a laser processing apparatus according to the present invention; [Figure 2a] 2A to 2C are diagrams showing examples of images obtained during operation of the processing device of FIG. 1. [Figure 2b] FIG. 2b illustrates steps in the analysis of the acquired image of FIG. 2a. [Figure 2c] FIG. 2b illustrates steps in the analysis of the acquired image of FIG. 2a. [Figure 3] FIG. 1 shows the time course of characteristic parameters obtained from the analysis of multiple captured images. [Figure 4] 2A and 2B are diagrams showing distributions of characteristic parameters obtained from the time-dependent changes of the characteristic parameters during operation of the processing apparatus of FIG. 1 under two different conditions. [Figure 5] FIG. 1 is a diagram illustrating the method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0014] In FIG. 1, reference number 1 generally indicates a laser processing device configured as a whole to perform laser processing on a workpiece 2, preferentially cutting and / or drilling the workpiece 2 and / or performing welding and / or performing additive manufacturing.

[0015] Preferentially, the workpiece 2 may comprise and / or consist of a metallic material, for example, the workpiece 2 may comprise and / or consist of carbon steel, aluminum or other metals.

[0016] It should be noted that the method described below has been tested by the applicant on workpieces 2 having different compositions and different shapes that are typically processed by the applicant's processing machines. It should be noted from the outset that the various materials that can be laser processed and the various shapes will have an impact on the details of the machine learning, but will not change the essence of the invention.

[0017] According to some non-limiting embodiments, the workpiece 2 has a flat and / or cylindrical and / or rod-like shape.

[0018] More specifically, the laser processing apparatus 1 comprises a control unit 3 for controlling the operation of the laser processing apparatus 1, a laser beam emitting source 4 operably connected to the control unit 3 and configured to emit the laser beam 5, an optical group 6 preferentially operably connected to the control unit 3 and configured to control the laser beam 5 and preferentially direct the laser beam 5 along an optical axis A onto the workpiece 2 and to the processing zone 7, and a movement device preferentially operably connected to the control unit 3 for relative movement between the workpiece 2 and the laser beam 5 at a determined speed and / or to define the shape of the cutting and / or drilling and / or welding and / or additive processing.

[0019] According to some non-limiting embodiments, the processing zone 7 may be a zone of the workpiece 2 that is exposed to the laser beam 5 during use and is consequently processed, for example to cut and / or drill and / or weld and / or additively machine. Such processing zone 7 may also be dynamic due to relative movement between the laser beam 5 and the workpiece 2 during use.

[0020] Preferentially, the laser processing apparatus 1 may also comprise a manufacturing device (not shown or known as such) operatively connected to a control unit 3 and configured to generate a gas jet to direct the compounds generated away from the workpiece 2 during laser processing, preferentially during cutting and / or drilling of the workpiece 2.

[0021] According to some preferred non-limiting embodiments, the processing apparatus may also include a suction unit configured to remove fumes and / or auxiliary products and / or compounds generated during laser processing.

[0022] According to some preferred, non-limiting embodiments, the control unit 3 may be configured to control process parameters of the laser processing apparatus 1, preferentially the intensity and / or frequency of the laser beam 5 and / or the duty cycle in the pulsed region of the laser beam 5 and / or the diameter of the laser beam 5 and / or the determined speed of relative movement between the laser beam 5 and the workpiece 2 and / or the position of the gas jet and / or a nozzle configured to emit the gas jet.

[0023] Preferentially, the control unit 3 may also be configured to control the process parameters in feedback mode.

[0024] It should be noted that preferentially, the process parameters may be (substantially) all parameters that determine the operation of the laser processing apparatus 1.

[0025] Advantageously, the laser processing apparatus 1 can also comprise a monitoring device 8 configured to monitor the process and / or result of the laser processing. Preferentially, the monitoring device 8 can also be configured to monitor the cutting and / or drilling and / or welding and / or additive processing of the workpiece 2.

[0026] Preferentially, the monitoring device 8 may also be configured to acquire a signal, preferentially an optical signal.

[0027] By way of specific example, the monitoring device 8 is configured to acquire a number of captured images 9 (an exemplary captured image is shown in FIG. 2 a ) preferentially of the processing zone 7 .

[0028] Preferentially, the monitoring device 8 is configured to acquire optical signals (preferentially acquired images 9) continuously and in real time during the laser processing.

[0029] Alternatively, the monitoring device 8 may comprise an interferometer, for example as described in EP-A-3832251, WO-A-2021111393 or WO-A-2021111399.

[0030] Preferentially, the monitoring device 8 is operably connected to the control unit 3, and the control unit 3 may be configured to control the operation of the laser processing apparatus 1 in response to information extracted and / or obtained at least from the monitoring device 8 (preferentially from the acquired signals, further preferentially from the acquired optical signals, and even more preferentially from the acquired images 9).

[0031] Preferentially, the monitoring device 8 may be configured to acquire a signal (preferentially an optical signal, and even more preferentially an acquired image 9) during operation of the laser processing apparatus 1 (in other words, the monitoring device 8 may be configured to operate online and in real time).

[0032] According to some preferred, non-limiting embodiments, the monitoring device 8 may be configured to acquire process radiation, i.e. thermal radiation, preferentially currently present in the processing zone 7. Alternatively, the monitoring device 8 may be configured to acquire electromagnetic waves (light) provided by an illumination system using a separate light source.

[0033] Preferentially, the radiation source 4 may comprise a laser, such as an ND:YAG laser, a fiber laser, a carbon dioxide laser, a diode laser, etc. The laser may also be, for example, a laser emitting a laser beam 5 having a wavelength λ of 1070 nm and a power of 6 kW.

[0034] More specifically, optical group 6 directs laser beam 5 onto workpiece 2 and determines the focus of laser beam 5 .

[0035] Preferentially, the optical group 6 may be configured to define a path P from the source 4 to the workpiece 2 .

[0036] According to some non-limiting embodiments, the path P may be composed of a first portion P1 transverse to the optical axis A, more specifically perpendicular to the optical axis A, and / or a second portion P2 coaxial with the optical axis A.

[0037] In other words, the laser beam 5 propagates along a portion P1 and a portion P2, respectively, P1 being transverse to P2 and preferentially coinciding with the optical axis A.

[0038] Alternatively, the path P may be coaxial with the optical axis A.

[0039] According to some preferred embodiments, the optical group 6 may comprise at least one focusing lens 14 configured to determine the focus of the laser beam 5, preferentially a focusing lens 14 arranged in the portion P2.

[0040] Preferentially, the optical group 6 may also comprise a collimation lens 15 and a dichroic mirror 16 configured to deflect the laser beam 5 from portion P1 to portion P2. In particular, the collimation lens 15 may also be arranged in portion P1.

[0041] Alternatively, the dichroic mirror 16 may be configured to shift the optical axis A but not to alter the path P of the laser beam 5 .

[0042] More specifically, the moving device may be configured to control the movement of the laser beam 5 relative to the workpiece 2 in a first relative advancement direction D1 and / or in a second relative advancement direction transverse to (preferentially perpendicular to) the first relative advancement direction.

[0043] According to some preferred embodiments, the movement device may comprise a support (not shown or known as such) configured to support the workpiece 2. According to some non-limiting variants, the support may be movable to initiate movement to obtain a relative movement between the laser beam 5 and the workpiece 2.

[0044] Alternatively or additionally, at least one part of the movement device may be incorporated into and / or associated with a support for moving the workpiece 2 in order to obtain relative movement between the laser beam 5 and the workpiece 2.

[0045] Alternatively or additionally, the movement device may comprise a moveable support base carrying the emission source 4 and / or the optical group 6 and / or parts of the optical group 6 for moving the laser beam 5 .

[0046] More particularly, the monitoring device 8 may comprise at least one sensor (preferentially an optical sensor) configured to acquire a signal (preferentially an optical signal, and even more preferentially an acquired image 9).

[0047] Preferentially, the sensor is and / or comprises a video camera 17, for example of the CCD or CMOS type, configured to acquire a signal (preferentially an optical signal, and even more preferentially an acquired image 9).

[0048] More specifically, the video camera 17 may be configured to continuously capture captured images 9 in real time to obtain a time series of captured images 9 .

[0049] More preferentially, the video camera 17 may be configured to acquire captured images 9 at a frequency of at least 1000 frames per second, and even more preferably 1500 frames per second.

[0050] For example, the video camera 17 can be a XiQ MQ013MG-ON from Ximea, Münster, Germany, preferentially associated with an infrared filter.

[0051] For example, video camera 17 may have a resolution of 200x200 pixels and operate at 750 Hz.

[0052] Preferentially, the video camera 17 may be configured to capture the electromagnetic radiation beam 18 emerging from the processing zone 7 .

[0053] According to some preferred, non-limiting embodiments, the electromagnetic radiation beam 18 may correspond to process emissions in the processing zone 7. Such embodiments reduce the need for additional light sources.

[0054] Alternatively, the electromagnetic radiation beam 18 can originate from the light of the processing zone 7, for example using a separate light source.

[0055] Preferentially, the electromagnetic radiation beam 18 is allowed to pass through at least a portion of the optical group 6 , in particular the focusing lens 14 and the dichroic mirror 16 .

[0056] According to some non-limiting embodiments, an optical sensor, preferentially a video camera 17, is arranged coaxially with the optical axis A. More preferentially, the electromagnetic radiation beam 18 may, in use, propagate parallel to at least the portion P2.

[0057] Alternatively, the electromagnetic radiation beam 18 may propagate along a path having at least two portions oriented transversely relative to each other.

[0058] More particularly, the monitoring device 8 may also comprise an optical filtering group 19 configured to ensure that the optical sensor, preferentially the video camera 17, receives electromagnetic radiation in a defined wavelength band during use. Preferentially, the optical filtering group 19 may preferably operate in the near infrared (is a near infrared filter).

[0059] Preferentially, the optical filtering group 19 may be arranged upstream in relation to the third direction of the video camera 17 .

[0060] The control unit 3 may be configured to control the operation of the laser processing apparatus 1, as described in more detail below.

[0061] For example, the laser beam 5 may cut material from the workpiece 2 in the processing zone 7, preferentially by heating, to create a slit extending along the entire thickness of the workpiece 2. Such a slit has and / or is bounded by a surface with a roughness determined by the laser processing.

[0062] It should be noted that roughness can be expressed in various ways. For example, roughness can be expressed in terms of the arithmetic mean roughness (known in the art as Ra), i.e., the arithmetic mean value of the roughness profile determined by the deviation from the center line within the evaluation area. Alternatively, roughness can be expressed relative to the maximum height of the peaks downstream of the roughness profile, and the mean value of the evaluation area of ​​the roughness profile determined (such roughness is known in the art as Ra). However, further methods for determining roughness are also known and will be familiar to those skilled in the art.

[0063] However, it should be noted that for the purposes of the present invention, the particular choice of coarseness is not important, merely that it must be consistently utilized in the machine learning process.

[0064] More specifically, the laser beam 5 may cut into the workpiece 2 from a first surface 20 of the workpiece 2 to a second surface 21 opposite the first surface of the workpiece 2. Even more specifically, during laser machining, a cross section of the first surface 30 and the second surface 21 is formed, the cross section having a roughness determined by the laser machining.

[0065] Preferentially, the control unit 3 may comprise an analysis module 22 configured to analyze the acquired signals (preferentially the acquired optical signals, and even more preferentially the acquired images 9).

[0066] As described in more detail below, analysis module 22 may be configured to estimate the roughness (ie, an estimate of the roughness) caused by the laser processing, eg, cutting operation.

[0067] Preferentially, the analysis module 22 may be configured to associate temporal information (i.e., the time of the estimation) and / or spatial information (i.e., the location of the laser processing associated with the roughness estimation) with each estimated roughness.

[0068] Preferentially, the control unit 3 may be configured to control and / or modify (preferentially in a feedback manner) one or more parameters according to the roughness estimated by the analysis module 22 (preferentially to obtain the desired roughness).

[0069] For example, when the estimated roughness should be less than the desired roughness, the control unit 3 may be configured to increase the determined speed of relative movement between the laser beam 5 and the workpiece 2, or when the estimated roughness should be greater than the desired roughness, the control unit 3 may be configured to decrease the determined speed of relative movement between the laser beam 5 and the workpiece 2.

[0070] However, the control unit 3 may also be configured to modify other parameters.Preferentially, the analysis module 22 may be configured to determine (preferentially in real time) the variation over time of at least one characteristic parameter (preferentially the variation over time of several characteristic parameters) (as can be seen in Figures 2c and 3) obtained starting from the acquired signal (preferentially from the optical signal, even more preferentially from the acquired image 9).

[0071] Preferentially, the analysis module 22 may also be configured to calculate at least one statistical parameter (preferably several statistical parameters) from the time-dependent variations of the characteristic parameters and to estimate the roughness from the statistical parameters. Preferentially, each time-dependent variation may be considered for a defined time period, and in particular this defined time period may be constant.

[0072] More particularly, the analysis module 22 may be configured to convert each acquired image 9 into a transformed image 23 (see FIG. 2b), in particular to obtain each binary image by segmentation. Preferentially, each transformed image 23 (binary image) comprises a first color (for example white) and a second color (for example black).

[0073] It should be noted that each acquired image 9 may contain information related to the acquired intensity, e.g., information corresponding to the intensity of the process emission. In particular, a first color and a second color are associated with each region of the transformed image 23 having an intensity below or above a determined intensity threshold, respectively.

[0074] According to some preferred embodiments, each acquired image 9 and the resulting respective transformed image 23 comprises a respective high intensity region 24 .

[0075] Preferentially, each high strength region 24 comprises a respective main portion 25, which can be approximated and / or described, for example, by a circular or elliptical shape, and one or more respective extension portions 26 extending from the respective main portion 25.

[0076] Preferentially, each high intensity region 24 is defined, and preferentially each main portion 25 and / or each extension portion 26 is defined, based on regions of each acquired image 9 having an intensity above a determined intensity threshold.

[0077] With particular reference to FIG. 2 c , each of the aforementioned characteristic parameters may correspond to a geometric parameter of the high intensity region 24 .

[0078] For example, each geometric parameter may be selected from the group consisting of the surface area of ​​the high intensity region 24, the center of gravity c of the high intensity region 24, the width w of the high intensity region 24, the length l of the high intensity region 24, other shape factors of the other high intensity region 24, and / or combinations thereof.

[0079] More specifically, each high strength region 24 may extend in a longitudinal direction Dl and in a transverse direction Dt transverse to (preferentially perpendicular to) the longitudinal direction Dl.

[0080] For example, each width w may be defined by the maximum elongation of the high strength region 24 in the transverse direction Dt.

[0081] For example, the respective length l of each high strength region 24 may be defined by the maximum elongation of the high strength region 24 in the longitudinal direction Dl.

[0082] Each characteristic parameter may be determined not only by geometrical characteristics but also by a combination of these and / or their respective time derivatives.

[0083] For example, the surface area of ​​the high intensity region 24 may correspond to the number of pixels having an intensity equal to or greater than a determined intensity threshold. Preferentially, the surface area may be determined from the respective binary image.

[0084] It should be noted that, according to some embodiments, more than one binary image can be generated for each acquired image 9, each binary image being generated taking into account a different intensity threshold. In other words, it may be possible to consider multiple surface areas as characteristic parameters that are distinct from the particular intensity threshold applied.

[0085] According to some preferred embodiments, the analysis module 22 may be configured to determine a respective probability distribution (see FIG. 4) from the respective time-course of each characteristic parameter, and to determine a parameter or statistical parameter from each probability distribution.

[0086] Preferentially, the respective statistical parameters are selected from the group consisting of the respective means, the respective variances and higher order statistical moments.

[0087] Preferentially, Fig. 4 shows two schematic examples of probability distributions for two cutting operating conditions distinguished for one variation of at least one processing parameter, such as the speed of the relative movement between the laser beam 5 and the workpiece 2. According to these probability distributions, two different roughnesses can be estimated.

[0088] Preferentially, the analysis module 22 may be configured to operate in continuous mode (preferentially in time and / or quantity continuous mode) so as to determine the evolution of the roughness estimate over time.

[0089] Advantageously, the analysis module 22 may be configured to calculate a continuous real-time roughness estimate as a function of each characteristic parameter (preferentially each determined statistical parameter).

[0090] More particularly, the analysis module 22 may be configured to estimate the roughness by statistical regression according to statistical parameters or parameters determined for estimating the roughness.

[0091] In particular, statistical regression models were trained using standard machine learning methods.

[0092] More specifically, the analysis module 22 can estimate the roughness by a linear or non-linear regression model, a decision tree regression, a random forest regression, an extreme gradient boosting regression, a linear probability model regression, or a multi-layer perceptron regression, depending on the respective statistical parameter function.

[0093] Alternatively or additionally, the analysis module 22 may be configured to perform classification according to statistical parameters or parameters determined to estimate the roughness.

[0094] Applicant found that all tested statistical regression models gave good and very satisfactory roughness predictions. Furthermore, Applicant's testing highlighted that no statistical regression model appears to be superior to the others in estimating roughness (taking into account various factors such as model performance, ease of application, etc.).

[0095] For example, linear regression models can be easily integrated, but have slightly inferior performance to multi-layer perceptron regression models.

[0096] According to some embodiments, the control unit 3 may be configured to receive and / or allow the definition and / or modification of the desired roughness, for example by a human-machine interface of the laser processing machine 1 and / or the control unit 3. For example, the desired roughness may be expressed in terms of an average value of the desired roughness.

[0097] Preferentially, the control unit 3 may be configured to control the process parameters depending on the roughness estimate and the desired roughness.

[0098] In practice, the laser processing device 1 performs laser processing of the workpiece 2, for example for cutting and / or drilling and / or welding and / or adaptive production of the workpiece 2.

[0099] Advantageously, the laser processing method (preferentially carried out by the laser processing device 1) comprises at least the following steps a) and b): a) Directing a laser beam 5, preferentially by means of a source 4 and an optical group 6, onto a workpiece 2, preferentially to a processing zone 7 on the workpiece 2 and / or to perform laser processing (e.g. to perform cutting and / or drilling and / or welding operations and / or additive processing). b) A relative movement between the laser beam 5 and the workpiece 2 is carried out, preferentially by means of a movement device, preferentially at a determined speed, in order to preferentially define the cutting shape and / or the hole shape.

[0100] Advantageously, the method can provide (preferably real-time and continuous) monitoring of the laser processing (preferably the result of the laser processing) to estimate the roughness resulting from the laser processing.

[0101] By running the method in real time and continuously, it is meant that the method runs throughout the entire laser processing (i.e. during the entire time path of the laser processing) and can intervene (if necessary) during the laser processing as well as after the laser processing.

[0102] In other words, and as will be explained in more detail below, according to the present invention, roughness estimates are obtained during and substantially at every instant during laser processing.

[0103] For the above reasons, the method may also include the following steps (see FIG. 5). c) Obtaining a signal (preferentially an optical signal, even more preferentially a plurality of acquired images 9) of the processing zone 7. d) determining the time course of one or more characteristic parameters obtained from a signal (preferably an optical signal, even more preferentially a plurality of acquired images 9) (see for example Figures 3 and 5). e) estimating the roughness obtained during laser processing (i.e., the estimated value of each roughness) according to each time course of one or more characteristic parameters;

[0104] Preferentially, steps c) to e) may be performed by the analysis module 22.

[0105] Preferentially, steps c) to e) are performed in real time and continuously, in other words, optical signals (preferentially acquired images 9) can be acquired and analyzed in order to obtain an estimate of the roughness during the laser processing while the laser processing is being performed.

[0106] According to some preferred embodiments, during step d) the variations over time of at least two characteristic parameters (preferentially at least three characteristic parameters) are determined, and during step e) the roughness is estimated according to each determined variation over time. In particular, during step e) one or more statistical parameters of each variation over time are determined, and during step e) the roughness is estimated according to each statistical parameter.

[0107] Applicant has observed that in this way the quality of the roughness estimate can be improved.

[0108] According to some non-limiting embodiments, during step e) temporal and / or spatial information is associated with each roughness estimate (i.e., each roughness estimate).

[0109] More specifically, each spatial information can be generated by correlation between each estimated roughness temporal information and the relative movement (e.g., characterized by speed and / or direction of advance) between the laser beam 5 and the workpiece 2.

[0110] More particularly, and with particular reference to Figures 3 to 5, in step e) at least one respective statistical parameter is determined from the time-dependent variation of the one or more characteristic parameters, and then a continuous real-time roughness estimate is calculated in response to the at least one determined statistical parameter.

[0111] It should be noted that the term "continuous real-time" means that the roughness estimation is performed during the laser processing.

[0112] According to some variants, during step e), a plurality of statistical parameters can be determined from the variation over time of at least one characteristic parameter and / or of a plurality of characteristic parameters. For example, it is also possible to determine two statistical parameters from the variation over time of a first characteristic parameter, while one statistical parameter is determined from the variation over time of a second characteristic parameter.

[0113] According to some preferred, non-limiting embodiments, during step e) a respective probability distribution can be determined from the variation of at least one characteristic parameter over time.

[0114] Preferentially, each statistical parameter may be determined from at least one respective probability distribution.

[0115] For example, statistical parameters such as respective means, respective variances, and higher order statistical moments (from the respective probability distributions) can be determined.

[0116] More specifically, during step e), a statistical regression is performed according to the determined statistical parameters or parameters for roughness estimation. For example, during step e), a linear or non-linear regression model, a decision tree regression, a random forest regression, an extreme gradient boosting regression, a linear probability model regression or a multi-layer perceptron regression can be adopted depending on the statistical parameters to obtain the roughness estimation.

[0117] Alternatively or additionally, during step e), a classification is carried out according to statistical parameters or parameters determined for estimating the roughness.

[0118] In particular, the above algorithms correspond to standard machine learning algorithms.

[0119] As already mentioned, according to a particular selected roughness (e.g. Ra or Rz), an algorithm is trained according to criteria known to those skilled in the art. According to some preferred embodiments, during step c), multiple acquired images 9 of the processing zone 7 can be acquired, preferentially consecutively and in real time. Preferentially, each acquired image 9 can be acquired at a different moment than the other acquired images 9. In other words, during step c), the evolution of the acquired images 9 over time can be determined.

[0120] Preferentially, the time course of each characteristic parameter can be determined from each high intensity region 24. Furthermore, the time course of each characteristic parameter results from the fact that each image 9 was acquired at a different time. Furthermore, but not necessarily, according to such an embodiment, temporal and / or spatial information of the roughness estimate can be determined.

[0121] Preferentially, each characteristic parameter may correspond to a geometric parameter of the high intensity region 24 .

[0122] For example, the geometric parameter may be selected from the group consisting of the surface area of ​​the high intensity region 24, the center of gravity c of the high intensity region 24, the width of the high intensity region 24, the length of the high intensity region 24, other shape factors of the high intensity region 24, and / or combinations thereof.

[0123] The applicant has observed that choosing to use at least the centroid c of the high intensity region 24 as characteristic parameter allows for an accurate estimation of the roughness. Preferentially, one or more other characteristic parameters, such as in particular the surface area, are selected to obtain a further improvement of the estimation.

[0124] According to some preferred, non-limiting embodiments, the method further comprises a controlling step (f), in which one or more process parameters are controlled in response to the estimated roughness. Preferentially, during step (f), one or more process parameters are controlled to obtain a desired roughness.

[0125] It should be noted that the desired roughness may be expressed in terms of an average desired roughness value. According to some possible variations, the desired roughness (i.e., the respective desired average value) may include a defined range within which the estimated roughness must fall.

[0126] It should be noted that the desired roughness may be described relative to one of the known types of roughness, such as Ra or Rz, provided that it is used consistently during the performance of the method, i.e., the desired roughness may refer to, for example, roughness Ra, and therefore the estimated roughness may also refer to estimated roughness Ra.

[0127] Preferentially, during step (f), the control unit 3 compares the estimated roughness with a desired roughness, and the control unit 3 keeps the process parameters constant if the estimated roughness matches the desired roughness, and modifies one or more process parameters if the estimated roughness does not match the desired roughness.

[0128] According to some preferred non-realistic embodiments, steps a) to e) (preferentially steps a) to f) are performed successively in time, preferentially so as to obtain the desired workpiece 2. In other words, control of the method is obtained during the laser processing. It is even possible to intervene if the estimated roughness value does not match the desired roughness value.

[0129] According to some preferred, non-limiting embodiments, the method further comprises a step (g) during which steps a) to e), preferentially steps a) to (f), are repeated, in particular for carrying out the method continuously and preferentially in real time.

[0130] According to some preferred, non-limiting embodiments, steps c) to e), preferentially steps c) to (f), are performed during the execution of steps a) and b), which means that steps c) to e), preferentially steps c) to (f), are performed during (rather than after) the laser processing.

[0131] According to some preferred, non-limiting embodiments, a storage step may also be performed during which the roughness estimates are stored, for example in the memory of the laser processing apparatus 1. Preferentially, the roughness estimates may be stored with temporal and / or spatial information (i.e., each value may be attributed to a specific moment in time of the respective estimate and / or position relative to the laser processing apparatus 1).

[0132] Preferentially, the storage step may be performed automatically by the analysis module 22 .

[0133] According to some preferred, non-limiting embodiments, a generation step may also be performed during which a document is generated from the roughness estimates, preferentially from the roughness estimates stored during the storage step.

[0134] Preferentially, the documentation can be used for certification purposes of the laser processing, for example to demonstrate that the roughness obtained during the laser processing corresponds to the desired value for a defined quality criterion.

[0135] Preferentially, the generation step may be carried out automatically by the analysis module 22 .

[0136] According to some preferred embodiments, during step d), the time course of each characteristic parameter may be determined for a defined time, in particular for a defined fixed time (in other words, the number of acquired images 9 that may be utilized to determine each characteristic parameter is fixed and predefined).

[0137] Each roughness estimate is therefore based on the analysis of a defined number of acquired images 9 .

[0138] More specifically, each captured image 9 is acquired at a specific time t. Consequently, it is possible to extract selected characteristic parameters and associate each characteristic parameter with the respective time t. It is then possible to obtain the time course of the characteristic parameters as shown in Figure 3. Statistical analysis of the time course allows the respective statistical parameters to be determined.

[0139] More particularly, during step c), the analysis module 22 may be configured to operate in a continuous mode (preferentially temporally and quantitatively) to determine the evolution of the characteristic parameters over time.

[0140] More specifically, during step d), the analysis module 22 analyzes, for each characteristic parameter, a number of acquired images 9 acquired one after the other for a defined time, so that the acquired images 9 make it possible to estimate the roughness at different moments.

[0141] For example, analysis module 22 may analyze a first plurality of acquired images 9 and a second plurality of acquired images 9 to determine a first roughness estimate and a second roughness estimate that is temporally subsequent to the first roughness estimate, respectively. Further, the analysis may require that the first plurality of acquired images 9 and the second plurality of acquired images 9 partially overlap in order to determine the respective roughness estimates.

[0142] Preferentially, the second plurality of acquired images 9 includes the first plurality of acquired images 9 and the same number of acquired images 9. The second plurality of acquired images 9 includes a defined number of additional acquired images 9 that are acquired (over time) after an acquired image 9 of the first plurality of acquired images 9 (in other words, a defined number of acquired images 9 over time follow the last acquired image 9 of the first plurality of acquired images 9).

[0143] In other words, the first plurality of acquired images 9 and the second plurality of acquired images 9 cover the same time period, specifically the same as a defined time. The first plurality of acquired images 9 includes acquired images 9 that were acquired before all other acquired images 9, while the second plurality of acquired images 9 includes a series of acquired images 9, at least one of which is acquired after all of the acquired images 9 of the first plurality of acquired images 9.

[0144] The process is repeated, with the second plurality of captured images 9 serving as the first plurality of captured images 9 .

[0145] Preferentially, during step d), a transformation substep may be performed, during which each acquired image 9 is transformed into a respective transformed image 23. The respective characteristic parameters may then be obtained from the transformed images 23. In particular, the transformation substep defines and / or amounts to a "thresholding" substep, during which each acquired image 9 is segmented in order to obtain a respective binary image (transformed image 23).

[0146] Preferentially, and with particular reference to Figure 2b, during the "thresholding" sub-step, the following associations can be made: a first color (for example white) is associated with each area of ​​the transformed image 23 (binary image) that corresponds to each area of ​​the acquired image 9 that has an intensity greater than or equal to the determined intensity threshold, and a second color (for example black) is associated with each area of ​​the transformed image 23 (binary image) that corresponds to each area of ​​the acquired image 9 that has an intensity less than the determined intensity threshold.

[0147] More specifically, during the "thresholding" sub-step, each pixel of the respective acquired image 9 may be associated with a first color or a second color based on an intensity threshold determined to obtain a respective transformed image 23. A first color may be associated with pixels having an intensity greater than or equal to the determined intensity threshold, and a second color may be associated with pixels having an intensity less than the determined intensity threshold.

[0148] By examining the features of the laser processing apparatus 1 and method according to the present invention, the advantages obtained thereby become clear.

[0149] Specifically, it is possible to obtain an estimate of roughness during laser processing rather than afterwards.

[0150] Furthermore, it is possible to estimate the roughness continuously over time during laser processing.

[0151] The characteristic parameters can be controlled to modify them to obtain the desired roughness, thereby reducing the need to reprocess the workpiece 2 after processing and / or obtaining optimization of the productivity of the laser processing.

[0152] Finally, it is clear that modifications and variations can be made to the laser processing apparatus 1 and method described and illustrated herein without departing from the scope of protection defined by the claims. [Explanation of symbols]

[0153] 1. Laser processing equipment 2 workpieces 3. Control Unit 4. Launch Source 5 Laser Beam 6 optical group 7 Processing Zone 8 Monitoring Equipment 9. Acquired images 14 Focusing Lens 15 Collimation Lens 16 Dichroic mirror 17. Video Camera 18 Electromagnetic Radiation Beam 19 Optical Filtering Group 22 Analysis Module 23 Transformed Images 24 High intensity region 25 Main Section 26 Extension part A optical axis

Claims

1. A method for laser processing a workpiece (2), preferentially made of metal material, comprising at least a) directing a laser beam (5) onto the workpiece (2) in a processing zone (7) of the workpiece (2) to perform laser processing; b) performing a relative movement between the laser beam (5) and the workpiece (2); c) acquiring optical signals, preferentially a plurality of acquired images (9) from said processing zone (7); d) determining the time course of one or more characteristic parameters obtained from said optical signal, more preferentially from said plurality of acquired images (9); e) estimating the roughness obtained during the laser processing according to the respective variations over time of the one or more characteristic parameters; Including, During step e), at least one respective statistical parameter is determined from the time-dependent evolution of the one or more characteristic parameters, and thereafter a continuous real-time estimate of the roughness is calculated as a function of each determined statistical parameter. Laser processing method.

2. During step e), at least one respective probability distribution is determined from one or more time variations of the one or more characteristic parameters. The method of claim 1.

3. During step e), a statistical regression and / or classification is performed according to the determined statistical parameters or parameters for estimating the roughness. The method according to claim 1 or claim 2.

4. During step e), a linear or non-linear regression model, a decision tree regression, a random forest regression, an extreme gradient boosting regression, a linear probability model regression or a multi-layer perceptron regression is employed depending on each statistical parameter to obtain the roughness estimate; The method of claim 3.

5. During step c), a plurality of captured images (9) of the processing zone (7) are acquired, Each acquired image (9) includes a high intensity region (24); Each characteristic parameter corresponds to a geometric parameter of the high intensity region (24).

5. The method according to any one of claims 1 to 4.

6. the geometric parameters are selected from the group consisting of a surface area of ​​the high strength region (24), a center of gravity of the high strength region (24), a width of the high strength region (24), a length of the high strength region (24), other shape factors of the high strength region (24), and / or combinations thereof; The method of claim 5.

7. During step e), at least one respective probability distribution is determined from one or more time-dependent variations of said one or more characteristic parameters, and each statistical parameter is selected from the group consisting of a respective mean, a respective variance, and higher order statistical moments of the respective probability distribution; The method according to claim 6, when dependent on any one of claims 1 to 4.

8. Each high intensity region (24) is defined based on an area of ​​each acquired image (9) having an intensity equal to or greater than a determined intensity threshold.

8. The method according to any one of claims 5 to 7.

9. During step d), the variation over time of at least two characteristic parameters, preferentially the variation over time of at least three characteristic parameters, is determined, During step e), one or more statistical parameters are determined from each determined time course and the roughness is estimated according to each statistical parameter.

9. The method according to any one of claims 1 to 8.

10. Steps a) to e) are performed sequentially; and / or Steps c) to e) are performed during the execution of steps a) and b).

10. The method according to any one of claims 1 to 9.

11. and further comprising a controlling step, during which one or more process parameters are controlled in response to the estimated roughness.

11. The method according to any one of claims 1 to 10.

12. During the controlling step, one or more of the process parameters are controlled to obtain a desired roughness. The method of claim 11.

13. a generation step, during which a document, preferentially a document for proof purposes, is generated from said roughness estimates, 13. The method according to any one of claims 1 to 12.

14. During step d), the time course is determined for a defined time, preferentially for a defined constant time, 14. The method according to any one of claims 1 to 13.

15. A laser processing device (1), A control unit (3) for controlling the operation of the laser processing device (1); a source (4) operatively connected to the control unit (3) and configured to emit a laser beam (5); an optical group (6) for controlling said laser beam (5); a movement device operatively connected to the control unit (3) and configured to effect relative movement between the laser beam (5) and the workpiece (2); The control unit (3) is configured and / or programmed to control the emission source (4) and / or the optical group (6) and / or the movement device to perform the method according to any one of claims 1 to 14. Laser processing equipment (1).

Images