CN116765643B

CN116765643B - Light beam adjustment and light signal processing method and device based on welding detection light path

Info

Publication number: CN116765643B
Application number: CN202310838508.8A
Authority: CN
Inventors: 付寄锋
Original assignee: Pushis Guangzhou Technology Co ltd
Current assignee: Pushis Guangzhou Technology Co ltd
Priority date: 2023-07-10
Filing date: 2023-07-10
Publication date: 2024-03-12
Anticipated expiration: 2043-07-10
Also published as: CN116765643A

Abstract

The invention discloses a beam adjustment and optical signal processing method and device based on a welding detection light path, which are characterized in that a welding focus of a welding beam falling on a welding working surface emitted by laser welding equipment is obtained, and a calibration beam is calibrated based on an electric control calibrator, so that the calibration focus of the calibration beam falling on the welding working surface coincides with the welding focus; after the alignment focus is determined to coincide with the welding focus, reflecting welding light beams emitted by the laser welding equipment into a welding detection light path, so that the welding light beams are transmitted into a beam splitter based on an electric control aligner, and the welding light beams are subjected to beam splitting treatment to obtain a plurality of welding beam splitting light beams with different wave bands; transmitting each welding beam to a corresponding electric control attenuator through an optical fiber, and carrying out optical attenuation adjustment on the welding beam to obtain a welding beam; compared with the prior art, the technical scheme of the invention realizes automatic alignment of the welding beam and automatic control of the light intensity of the welding beam.

Description

Light beam adjustment and light signal processing method and device based on welding detection light path

Technical Field

The invention relates to the technical field of welding optical signal processing, in particular to a method and a device for adjusting and processing a light beam based on a welding detection light path.

Background

In the existing laser welding detection technology, after a production welding detection station is fixed, if the production process needs to be replaced again, calibration and calibration are required to be repeated, structural adjustment is required to be carried out on a beam splitter, an attenuator and the like, and screening calibration and quantification are manually adjusted again so as to meet the requirements of the welding detection station.

When the screening, calibration and quantification are adjusted, manual calibration is adopted, a threshold value is planned, a screening frame is used for determining clutter, the software operation workload is large, staff needs to be trained for convenient use, and the workload of a production line is increased; and in the process, based on manual operation, the efficiency is low, and a larger error exists.

Disclosure of Invention

The invention aims to solve the technical problems that: the method and the device for adjusting the light beam and processing the light signal based on the welding detection light path are provided, so that the automatic alignment of the welding light beam and the automatic control of the light intensity of the welding light beam are realized.

In order to solve the technical problems, the invention provides a welding light beam adjusting method based on a welding detection light path, wherein the welding detection light path comprises a beam splitter, an electric control calibrator and an electric control attenuator, and the welding light beam adjusting method comprises the following steps:

Acquiring a welding focus of a welding beam emitted by laser welding equipment on a welding working surface, and calibrating a calibration beam emitted by a laser generator based on the electric control calibrator so as to enable the calibration focus of the calibration beam on the welding working surface to coincide with the welding focus;

after the alignment focus is determined to coincide with the welding focus, reflecting welding light beams emitted by laser welding equipment into the welding detection light path, so that the welding light beams are transmitted into the beam splitter based on the electric control aligner, and the welding light beams are split based on the beam splitter to obtain welding beam splitters with different wave bands;

and transmitting each welding beam-splitting beam to the corresponding electric control attenuator through an optical fiber so that the electric control attenuator can carry out light attenuation adjustment on the welding beam-splitting beam to obtain a welding light attenuation beam.

In one possible implementation manner, the calibrating light beam emitted by the laser generator is calibrated based on the electronically controlled calibrator, so that the calibration focus of the calibrating light beam falling on the welding working surface coincides with the welding focus, and specifically includes:

Transmitting a calibration light beam emitted by a laser generator into a welding detection light path, so that an electric control calibrator in the welding detection light path reversely reflects the calibration light beam to the welding working surface;

acquiring a first image of the welding working surface, and carrying out graying treatment on the first image to obtain a first gray image;

judging whether a calibration focus of the calibration light beam falling on the welding working surface coincides with the calibration focus or not according to the first gray level image, if so, considering that the electric control calibrator finishes calibrating the calibration light beam, otherwise, adjusting the electric control calibrator to change the position of the calibration focus on the welding working surface until the calibration focus coincides with the welding focus.

Further, an automatic calibration judgment process can be realized through processing and analyzing the image, so that the requirement of manual intervention is reduced, and the efficiency and the accuracy are improved; and through acquiring and processing the image in real time, whether the position of the calibration focus coincides with the welding focus or not can be rapidly determined, and the position is fed back to the electric control calibrator in real time for adjustment, so that time and resources are saved.

In one possible implementation manner, adjusting the electronically controlled calibrator specifically includes:

setting a welding working surface coordinate axis by taking the welding focus on the welding working surface as an origin, and acquiring a first coordinate position of the calibration focus on the welding working surface coordinate axis;

and obtaining a first x-axis coordinate variable and a first y-axis coordinate variable based on the first coordinate position, and adjusting the electric control calibrator based on the first x-axis coordinate variable and the first y-axis coordinate variable.

Further, setting a coordinate axis of the welding working surface by taking a welding focus on the welding working surface as an origin, acquiring a first coordinate position of a calibration focus on the coordinate axis, and taking the welding focus as a calibration reference point; thus, the accuracy of the light beam calibration relative to the welding working surface can be ensured, so that the light beam is correctly positioned to the welding focus position; and based on the first coordinate position and the corresponding coordinate variable, the electric control calibrator can be accurately adjusted, and the accuracy and the controllability of the light beam calibration can be realized by fine adjustment and accurate positioning of the calibrator, so that the light beam is ensured to be accurately aligned to a welding focus, and the welding requirement is met.

In one possible implementation, the electronically controlled calibrator includes a mirror plate, a support rod, an origin lens link gimbal, and a mount;

the support rod is arranged above the base, the support rod is perpendicular to the base, the first end of the origin lens linking universal joint is connected with the support rod, and the second end of the origin lens linking universal joint is connected with the reflecting lens.

Further, the supporting rod is perpendicular to the base and is connected with the reflecting mirror plate and the origin mirror plate to be linked with the universal joint, so that a stable and reliable supporting structure is provided, the positions of the reflecting mirror plate and the universal joint are kept stable, the light beam calibration error caused by vibration or displacement is reduced, and the angle of the reflecting mirror plate can be adjusted through the connection of the origin mirror plate to be linked with the universal joint, so that the calibrator can flexibly adjust the direction and the deflection angle of the light beam to meet different light beam calibration requirements; minute beam deflection, rotation and adjustment can be achieved, thereby achieving fine beam alignment and positioning.

In one possible implementation manner, the electronic control calibrator is adjusted based on the first x-axis coordinate variable and the first y-axis coordinate variable, and specifically includes;

Setting a coordinate axis of the reflector plate by taking the center point of the reflector plate as a reference point;

taking the first x-axis coordinate variable as a y-axis coordinate point and the first y-axis coordinate variable as an x-axis coordinate point, and obtaining a second coordinate position of the first coordinate position on the coordinate axis of the reflector based on the y-axis coordinate point and the x-axis coordinate point;

based on the second coordinate position, a second x-axis coordinate variable and a second y-axis coordinate variable are obtained

And adjusting the movement of the reflecting lens on the x axis of the reflecting lens coordinate axis based on the origin lens link universal joint based on the second x axis coordinate variable, and adjusting the movement of the reflecting lens on the y axis of the reflecting lens coordinate axis based on the origin lens link universal joint based on the second y axis coordinate variable so as to enable the datum point to coincide with the second coordinate position.

Further, by setting the coordinate axis of the reflecting mirror plate, and calculating the second coordinate position on the coordinate axis of the reflecting mirror plate based on the first coordinate position, the coordinate axis can be aligned with the reference direction required by the actual operation, which is helpful for accurately understanding and identifying the change of the relative position in the subsequent operation and adjustment process; meanwhile, when the movement of the reflecting lens is regulated based on the coordinate variable, the accuracy and the repeatability of the operation are ensured, and the regulation accuracy and the operation efficiency are improved.

In one possible implementation, the electrically controlled attenuator includes a stepper motor, a rotary disk, a motor output shaft, and a light outlet;

the stepping motor is connected with the motor output shaft, and the motor output shaft is controlled to rotate based on the stepping motor;

the rotating disc is fixed on the motor output shaft, and the rotation of the rotating disc is controlled based on the motor output shaft;

the rotary disk is provided with a plurality of attenuation sheets, and when the rotary disk rotates, the attenuation sheets sequentially correspond to the light outlet.

Further, by arranging different attenuation sheets on the rotating disc, the light attenuation of the welding beam can be flexibly adjusted so as to adapt to different working scenes and requirements. This flexibility makes the electrically controlled attenuator more customizable and adjustable.

In one possible implementation manner, the electrically controlled attenuator performs optical attenuation adjustment on the welding beam splitter to obtain a welding beam splitter, which specifically includes:

acquiring preset light attenuation power, and selecting a target attenuation sheet on the rotating disk based on the preset light attenuation power;

after the state of the stepping motor is determined to be a starting state, controlling the rotating disc to rotate based on the motor output shaft so as to enable a target attenuation piece on the rotating disc to rotate to the light outlet;

Transmitting the welding beam-splitting beam to the target attenuation sheet so that the target attenuation sheet carries out optical attenuation on the welding beam-splitting beam to obtain a welding beam-attenuation beam, and outputting the welding beam-attenuation beam based on the light outlet.

Further, by selecting the target attenuation sheet on the rotating disc and transmitting the welding light beam to the target attenuation sheet, accurate control of light intensity can be realized, so that light attenuation requirements under different application or demand scenes can be met.

In one possible implementation manner, the rotation of the rotating disc is controlled based on the output shaft of the motor, so that the target attenuation piece on the rotating disc is rotated to the light outlet, and the method specifically includes:

acquiring the rotation direction of the motor output shaft, and setting a rotation angle required by each attenuation piece on the rotating disc to rotate to the light outlet based on the rotation direction;

after a target attenuation sheet on the rotating disc is selected, a target rotation angle corresponding to the target attenuation sheet is obtained, and the rotating disc is controlled to rotate based on the rotation direction and the target rotation angle, so that the target attenuation sheet on the rotating disc is rotated to the light outlet.

Further, by acquiring the rotation direction of the motor output shaft and setting the rotation angle corresponding to each attenuation sheet, the system can calculate the rotation angle of each attenuation sheet based on the rotation direction, so that the rotation angle of each attenuation sheet on the rotating disc can be accurately controlled by accurately positioning the system, the target attenuation sheet is enabled to rotate to the position of the light outlet, the light beam attenuation process is enabled to be more reliable and accurate, and the obtained welding light attenuation light beam is ensured to meet the preset light attenuation power requirement.

In one possible implementation manner, the welding beam is transmitted to the beam splitter, so that the beam splitter splits the welding beam to obtain a plurality of welding beam splitters with different wave bands, and the method specifically includes:

the beam splitter comprises a first beam splitter and a second beam splitter;

transmitting the welding light beam to the first beam splitter, and splitting the welding light beam based on the first beam splitter so as to split the welding light beam into a first band welding beam splitter and a first welding beam;

and transmitting the first welding beam to the second beam splitter, and carrying out beam splitting treatment on the first welding beam based on the second beam splitter so as to split the first welding beam into a second-band welding beam splitter and a third-band welding beam splitter.

Further, the welding beam is split into a plurality of bands of light beams by using a beam splitter. The light beams with different wave bands can be respectively processed, so that independent monitoring and analysis of different light signal components in the welding process are realized, and the fine observation and detection capability of the welding light signals are improved; and signals among different wave bands can not interfere with each other, so that the optical signal characteristics of each wave band can be extracted and understood more accurately.

The invention also provides a welding light signal processing method based on the welding detection light path, which comprises the following steps:

acquiring a historical welding light signal set, acquiring a plurality of positive welding light signals from the historical welding light signal sets, and generating a welding detection reference range when the positive welding light signals meet positive judgment standards;

acquiring a welding light attenuation beam based on a PD sensor to obtain a welding light signal; wherein the welding light attenuated beam is obtained according to any one of the above welding light beam adjustment methods based on a welding detection light path;

and inputting the welding light signal into a pre-trained optimal detection model, so that the optimal detection model detects the welding light signal based on a welding detection reference range to obtain a welding detection result.

Further, by acquiring a historical welding light signal set and collecting a positive welding light signal therefrom, past data can be utilized to establish a reference range for welding detection; the standard range can better reflect the normal range of the optical signal and provide reference for subsequent detection; by collecting the welding light attenuated beam based on the PD sensor and obtaining a corresponding welding light signal, the actual signal data during the welding process can be more directly obtained. The method is favorable for reflecting the process conditions of the welding site more accurately and improving the accuracy of the detection model; the obtained welding light signals are input into a pre-trained optimal detection model, the welding light signals can be detected by using the learning capacity of the model, and the reliability and the efficiency of detection can be improved.

In one possible implementation manner, collecting a plurality of positive welding light signals from the plurality of historical welding light signals, and determining that the plurality of positive welding light signals meet a positive judgment standard specifically includes:

collecting a plurality of positive welding light signals from the historical welding light signal set, and setting a scanning window based on the signal length of each positive welding light signal;

window scanning is carried out on each positive welding light signal based on the scanning window, and when the variance of the distribution of the positive welding light signals in the scanning window is larger than a preset variance, the positive welding light signals are determined to be abnormal welding light signals;

Deleting the abnormal welding light signals in the plurality of positive welding light signals to obtain residual positive welding light signals, and taking the residual positive welding light signals as positive welding light signals meeting positive judgment standards.

Further, during the laser welding processing process, the signals collected by the system are affected by various factors, such as the stability problem of the laser in the tolerance range, the influence of ambient light, the workpiece variability in the tolerance limit, etc., so that the distribution of a few positive data is abnormal; based on the above, the method and the device can also eliminate abnormal data in the positive welding optical signal by detecting the collected positive welding optical signal, so that the accuracy of generating the welding detection reference range can be improved when the welding detection reference range is generated based on the positive welding optical signal later.

In one possible implementation manner, generating the welding detection reference range specifically includes:

determining light emitting areas of the welding light signals based on the plurality of positive welding light signals, and respectively dividing the plurality of positive welding light signals into a plurality of areas based on the light emitting areas so as to divide each positive welding light signal into a plurality of areas of positive welding light signals;

Classifying the positive welding light signals of the multiple regions based on the light-emitting regions to obtain positive welding light signals of all first regions corresponding to each light-emitting region, and obtaining positive welding light signals of all first regions at first moments corresponding to first welding moments;

performing data fitting processing on all positive welding light signals at a first moment to obtain a positive welding light signal distribution range corresponding to the first welding moment, and integrating the positive welding light signal distribution ranges corresponding to all welding moments in the positive welding light signals of the first region to obtain a region positive welding light signal distribution range corresponding to the positive welding light signals of the first region;

and integrating the distribution range of the positive welding light signals of the areas corresponding to each light-emitting area to obtain a welding detection reference range.

Furthermore, in the laser welding processing process, the light-emitting mode of the laser is segmented corresponding to each non-physical contact welding point, so that the acquired welding light signals are divided into areas, and the positive welding light signals of each divided area are analyzed, so that more accurate welding light signal states can be obtained, and the accuracy of a welding detection reference range can be improved.

In one possible implementation manner, performing data fitting processing on all positive welding light signals at the first moment to obtain a distribution range of the positive welding light signals corresponding to the first welding moment, which specifically includes:

when detecting that a user selects a first welding light signal distribution range generation mode from a plurality of preset welding light signal distribution range generation modes, performing data fitting processing on all first moment positive welding light signals based on a hypothesized normal distribution function to obtain a hypothesized normal distribution value, and selecting a region of the hypothesized normal distribution value within a preset standard deviation range as a positive welding light signal distribution range corresponding to the first welding moment;

when detecting that a user selects a second welding light signal distribution range generation mode from a plurality of preset welding light signal distribution range generation modes, arranging all first-moment positive welding light signals according to a sequence from big to small, screening all first-moment positive welding light signals based on the arrangement sequence to obtain first-moment screening positive welding light signals, selecting a signal intensity minimum value and a signal intensity maximum value in all first-moment screening positive welding light signals, and determining a positive welding light signal distribution range corresponding to the first welding moment based on the signal intensity minimum value and the signal intensity maximum value.

Further, by setting a plurality of welding light signal distribution range generation modes, the most suitable mode can be selected according to different conditions to obtain the welding light signal distribution range, so that the accuracy of welding light signal analysis is improved.

In one possible implementation manner, the generating process of the optimal detection model specifically includes:

and acquiring a plurality of welding sample optical signals, inputting the plurality of welding sample optical signals into an initial detection model, so that the initial detection model carries out signal detection on the plurality of welding sample optical signals based on the welding detection reference range, and carrying out super-parameter optimization processing on the initial detection model based on the plurality of negative welding sample optical signals after determining that the plurality of negative welding sample optical signals are detected, thereby obtaining an optimal detection model.

Further, by inputting a plurality of welding sample optical signals into an initial detection model, the detection capability of the initial detection model on different samples can be tested, the robustness and generalization capability of the model can be evaluated, whether the model can effectively cope with various welding conditions and optical signal changes can be determined, and by carrying out signal detection on the plurality of welding sample optical signals based on a welding detection reference range, whether the optical signals of the samples fall in a normal range can be judged, which is helpful for identifying samples possibly having problems, and further optimizing and improving the detection model; when a plurality of negative sample welding sample optical signals are detected, the samples can be analyzed and processed to perform super-parameter optimization, so that the parameters of a detection model can be adjusted according to specific problems and abnormal conditions in the samples, and the performance of the model under the abnormal welding condition can be improved; finally, the initial detection model is subjected to super-parameter optimization to obtain an optimal detection model, so that the optimal performance and adaptability of the model in the process of welding quality detection tasks can be ensured, abnormal conditions can be accurately detected, and the false alarm rate is reduced.

In one possible implementation manner, a plurality of welding sample optical signals are acquired, and the plurality of welding sample optical signals are input into an initial detection model, so that the initial detection model performs signal detection on the plurality of welding sample optical signals based on the welding detection reference range, and specifically includes:

acquiring a plurality of welding sample optical signals;

setting a plurality of initial detection models, wherein the plurality of initial detection models comprise a variance characteristic detection model, an average offset detection model, an area detection model and a time detection model;

arbitrarily selecting two initial detection models, respectively inputting the plurality of welding sample optical signals into the two initial detection models, so that the two initial detection models respectively perform signal detection on the plurality of welding sample optical signals based on the welding detection reference range, and the two initial detection models respectively output welding sample detection results of the plurality of welding sample optical signals;

comparing the welding sample detection result with welding sample results corresponding to the welding sample optical signals, and obtaining accurate values corresponding to the two initial detection models based on the comparison result;

And when the accurate value is determined to meet a preset threshold value, selecting a plurality of negative sample welding sample optical signals based on the welding sample result.

Further, performing model performance evaluation on an initial detection model based on the welding sample optical signal to ensure that the initial detection model can accurately detect a negative sample welding sample optical signal; and model parameter optimization is conveniently carried out on the initial detection model based on the negative sample welding sample optical signal.

In one possible implementation manner, the performing, based on the plurality of negative sample welding sample optical signals, a super-parameter optimization process on the initial detection model to obtain an optimal detection model specifically includes:

setting a model selection controller and a super-parameter controller corresponding to each initial detection model, wherein the model selection controller is used for selecting a target initial detection model for the plurality of initial detection models, and the super-parameter controller is used for performing super-parameter optimization on the target initial detection model;

modeling a process of selecting the target initial detection model based on the model selection controller and the process of performing super-parameter optimization on the target initial detection model based on the super-parameter controller as a Markov decision process;

Automatically adjusting the target initial detection model selected by the model selection controller based on the Markov decision process, automatically adjusting the super parameters of the target initial detection model, and recording all the selected target initial detection models and the super parameters corresponding to the target initial detection models;

and respectively carrying out model training on all the selected target initial detection models based on the plurality of negative sample welding light signals so as to enable the super parameters corresponding to the target initial detection models to obtain optimal detection models according to model training results.

Further, the model selection and parameter optimization process is modeled as a Markov decision process, the parameter selection of the model can be automatically adjusted through super parameter optimization to obtain a better model, the accuracy of the model can be improved, and the super parameter selection and the model selection are unified to optimize the overall performance of the whole model, so that an optimal super parameter combination is obtained, and meanwhile, the workload of manual parameter adjustment is reduced.

In one possible implementation manner, setting a model selection controller and a super parameter controller corresponding to each initial detection model specifically includes:

The model selection controllers are respectively connected with the super parameter controllers corresponding to each initial detection model;

the model selection controller comprises a first input embedded layer, a first decision core layer and a first output embedded layer, wherein the first input embedded layer consists of a plurality of perception layers, the first decision core layer consists of three long and short time memory networks, and the first output embedded layer consists of a plurality of perception layers;

the super-parameter controller comprises a second input embedded layer, a second decision core layer and a second output embedded layer, wherein the second input embedded layer is composed of a plurality of perception layers, the second decision core layer is composed of three layers of long and short time memory networks, and the second output embedded layer is composed of a plurality of perception layers.

Furthermore, the memory-based decision network is arranged in the model selection controller and the super parameter controller, so that the controller can be helped to adapt to various data sets and interact, and the generalization performance of the model is improved.

The invention provides a welding light signal processing method based on a welding detection light path, which further comprises the following steps:

reflecting a welding beam emitted by a laser welding device into a power detection device so that the power detection device obtains a laser light emitting power signal of the welding beam;

And based on the laser light-emitting power signal, performing equipment stability evaluation on the laser welding equipment to obtain a laser welding equipment stability result.

Further, stability and performance of the welding equipment can be evaluated by monitoring and analyzing the laser light-emitting power signals, working states of the equipment can be better known, and risks such as equipment faults and damage are avoided.

In one possible implementation manner, based on the laser output power signal, performing device stability evaluation on the laser welding device to obtain a laser welding device stability result, which specifically includes:

dividing the laser light-emitting power signal into a plurality of area laser light-emitting power signals based on the light-emitting areas, and respectively calculating the area light-emitting power corresponding to each area laser light-emitting power signal;

and counting the power variance of the area light-emitting power corresponding to the laser light-emitting power signals of each area in a preset time period, judging whether the power variance is larger than a preset power variance threshold, if so, outputting a stability result of the laser welding equipment to be maintained, otherwise, outputting a stability result of the laser welding equipment to be maintained.

Furthermore, by carrying out regional division and power variance statistics on the laser light-emitting power signals, the abnormality and the fault of the equipment can be rapidly and accurately found, the maintenance and the repair can be timely carried out, and the equipment operation interruption and the economic loss caused by the equipment operation interruption are avoided.

The invention also provides a welding light beam adjusting device based on a welding detection light path, wherein the welding detection light path comprises a beam splitter, an electric control calibrator and an electric control attenuator, and the welding light beam adjusting device comprises: the device comprises a light beam calibration module, a light beam splitting module and a light attenuation adjustment module;

the welding beam adjusting device is connected with the welding detection light path, wherein the beam calibration module is connected with the electric control calibrator, the beam splitting module is connected with the beam splitter, and the light attenuation adjusting module is connected with the electric control attenuator;

the beam calibration module is used for acquiring a welding focus of a welding beam emitted by the laser welding equipment on a welding working surface, and calibrating a calibration beam emitted by the laser generator based on the electric control calibrator so as to enable the calibration focus of the calibration beam on the welding working surface to coincide with the welding focus;

The beam splitting module is used for reflecting a welding beam emitted by the laser welding equipment into the welding detection light path after the alignment focus is determined to coincide with the welding focus, so that the welding beam is transmitted into the beam splitter based on the electric control aligner, and the welding beam is split based on the beam splitter to obtain a plurality of welding beam splitters with different wave bands;

the optical attenuation adjusting module is used for transmitting each welding beam-splitting beam to the corresponding electric control attenuator through the optical fiber, so that the electric control attenuator can carry out optical attenuation adjustment on the welding beam-splitting beam to obtain a welding light attenuation beam.

In one possible implementation, the beam calibration module includes a calibration beam transmission sub-module, an image processing sub-module, and a focus coincidence sub-module;

the calibration beam transmission sub-module is used for transmitting the calibration beam emitted by the laser generator into a welding detection light path so that an electric control calibrator in the welding detection light path reversely reflects the calibration beam onto the welding working surface;

the image processing sub-module is used for acquiring a first image of the welding working surface, and carrying out graying treatment on the first image to obtain a first gray image;

And the focus overlapping sub-module is used for judging whether the calibration focus of the calibration light beam falling on the welding working surface is overlapped with the calibration focus or not according to the first gray level image, if so, the electric control calibrator is considered to finish calibrating the calibration light beam, otherwise, the electric control calibrator is adjusted to change the position of the calibration focus on the welding working surface until the calibration focus is judged to be overlapped with the welding focus.

In one possible implementation manner, the focus coincidence submodule comprises a coordinate position acquisition unit and an electric control calibrator adjustment unit;

the coordinate position acquisition unit is used for setting a welding working surface coordinate axis by taking the welding focus on the welding working surface as an origin, and acquiring a first coordinate position of the calibration focus on the welding working surface coordinate axis;

the electronic control calibrator adjusting unit is configured to obtain a first x-axis coordinate variable and a first y-axis coordinate variable based on the first coordinate position, and adjust the electronic control calibrator based on the first x-axis coordinate variable and the first y-axis coordinate variable.

In one possible implementation manner, the electronic control calibrator adjusting unit comprises a reflector lens coordinate axis setting subunit, a coordinate position transformation subunit, a coordinate variable obtaining subunit and an origin lens linking universal joint moving subunit;

the mirror plate coordinate axis setting subunit is used for setting a mirror plate coordinate axis by taking the center point of the mirror plate as a reference point;

the coordinate position transformation subunit is configured to obtain, based on the y-axis coordinate point and the x-axis coordinate point, a second coordinate position of the first coordinate position on the coordinate axis of the reflector with the first x-axis coordinate variable as a y-axis coordinate point and the first y-axis coordinate variable as an x-axis coordinate point;

the coordinate variable obtaining subunit is configured to obtain a second x-axis coordinate variable and a second y-axis coordinate variable based on the second coordinate position;

the origin lens link gimbal moving subunit is configured to adjust, based on the second x-axis coordinate variable, movement of the reflection lens on the x-axis of the reflection lens coordinate axis based on the origin lens link gimbal, and adjust, based on the second y-axis coordinate variable, movement of the reflection lens on the y-axis of the reflection lens coordinate axis based on the origin lens link gimbal, so that the reference point coincides with the second coordinate position.

In one possible implementation manner, the optical attenuation adjustment module comprises a target attenuation sheet selecting sub-module, a rotating disc rotating sub-module and an optical attenuation sub-module;

the target attenuation sheet selecting sub-module is used for acquiring preset light attenuation power and selecting a target attenuation sheet on the rotating disc based on the preset light attenuation power;

the rotating disc rotating sub-module is used for controlling the rotating disc to rotate based on the motor output shaft after determining that the state of the stepping motor is a starting state so as to enable a target attenuation piece on the rotating disc to rotate to the light outlet;

The optical attenuation submodule is used for transmitting the welding beam-splitting light beam to the target attenuation sheet so that the target attenuation sheet can carry out optical attenuation on the welding beam-splitting light beam to obtain a welding beam-attenuation light beam, and outputting the welding beam-attenuation light beam based on the light outlet.

In one possible implementation manner, the rotary disk rotation submodule includes a rotation angle setting unit and a target attenuation sheet rotation unit;

the rotating angle setting unit is used for obtaining the rotating direction of the motor output shaft and setting the rotating angle required by each attenuation piece on the rotating disc to rotate to the light outlet based on the rotating direction;

the target attenuation sheet rotating unit is used for acquiring a target rotation angle corresponding to the target attenuation sheet after selecting the target attenuation sheet on the rotating disc, and controlling the rotating disc to rotate based on the rotation direction and the target rotation angle so as to enable the target attenuation sheet on the rotating disc to rotate to the light outlet.

In one possible implementation, the beam splitting module includes a first beam splitting processing sub-module and a second beam splitting processing sub-module;

The light splitter comprises a first light splitting piece and a second light splitting piece;

the first beam splitting processing sub-module is used for transmitting the welding beam to the first beam splitting sheet, and splitting the welding beam based on the first beam splitting sheet so as to split the welding beam into a first band welding beam and a first welding beam;

the second beam splitting processing sub-module is configured to transmit the first welding beam to the second beam splitter, and perform beam splitting processing on the first welding beam based on the second beam splitter, so that the first welding beam is split into a second band welding beam splitter and a third band welding beam splitter.

The invention also provides a welding light signal processing device based on the welding detection light path, which comprises: the welding detection reference range generation module, the welding optical signal acquisition module and the welding optical signal detection module;

the welding detection reference range generation module is used for acquiring a historical welding light signal set, acquiring a plurality of positive welding light signals from the historical welding light signal sets, and generating a welding detection reference range when the positive welding light signals meet positive judgment standards;

The welding light signal acquisition module is used for acquiring a welding light attenuation beam obtained in the welding light beam adjusting device based on a PD sensor to obtain a welding light signal; wherein the welding light attenuated beam is obtained by the welding light beam adjusting device based on the welding detection light path according to any one of the above;

the welding light signal detection module is used for inputting the welding light signal into a pre-trained optimal detection model so that the optimal detection model detects the welding light signal based on a welding detection reference range to obtain a welding detection result.

In one possible implementation manner, the welding detection reference range generation module comprises a scanning window setting sub-module, an abnormal welding light signal confirmation sub-module and a welding light signal screening sub-module;

the scanning window setting submodule is used for collecting a plurality of positive welding light signals from the historical welding light signal set and setting a scanning window based on the signal length of each positive welding light signal;

the abnormal welding light signal confirmation sub-module is used for carrying out window scanning on each positive welding light signal based on the scanning window, and when the variance of the distribution of the positive welding light signals in the scanning window is determined to be greater than a preset variance, the positive welding light signals are determined to be abnormal welding light signals;

The welding light signal screening sub-module is used for deleting the abnormal welding light signals in the plurality of positive welding light signals to obtain residual positive welding light signals, and taking the residual positive welding light signals as positive welding light signals meeting positive judgment standards.

In a possible implementation manner, the welding detection reference range generation module further comprises a light emitting region determination sub-module, a first moment positive welding light signal acquisition sub-module, a region positive welding light signal distribution range determination sub-module and a welding detection reference range determination sub-module;

the light emitting region determining submodule is used for determining light emitting regions of the welding light signals based on the plurality of positive welding light signals, and respectively dividing the plurality of positive welding light signals into a plurality of regional positive welding light signals based on the light emitting regions so as to divide each positive welding light signal into a plurality of regional positive welding light signals;

the first moment positive welding light signal obtaining submodule is used for classifying the positive welding light signals of the plurality of areas based on the light-emitting areas to obtain all positive welding light signals of the first areas corresponding to each light-emitting area, and obtaining first moment positive welding light signals of all positive welding light signals of the first areas corresponding to the first welding moment;

The region positive welding light signal distribution range determining submodule is used for carrying out data fitting processing on all positive welding light signals at a first moment to obtain a positive welding light signal distribution range corresponding to the first welding moment, integrating the positive welding light signal distribution ranges corresponding to all welding moments in the first region positive welding light signal, and obtaining a region positive welding light signal distribution range corresponding to the first region positive welding light signal;

and the welding detection reference range determination submodule is used for integrating the area positive welding light signal distribution range corresponding to each light-emitting area to obtain a welding detection reference range.

In one possible implementation manner, the regional positive welding light signal distribution range determining submodule includes a first positive welding light signal distribution range generating unit and a second positive welding light signal distribution range generating unit;

the first normal welding light signal distribution range generation unit is used for carrying out data fitting processing on all first moment normal welding light signals based on an assumed normal distribution function to obtain an assumed normal distribution value when detecting that a user selects a first welding light signal distribution range generation mode from a plurality of preset welding light signal distribution range generation modes, and selecting a region of the assumed normal distribution value within a preset standard deviation range as the normal welding light signal distribution range corresponding to the first welding moment;

The second positive welding light signal distribution range generation unit is configured to, when detecting that a user selects a second welding light signal distribution range generation mode from a plurality of preset welding light signal distribution range generation modes, arrange all positive welding light signals at first moments in a sequence from large to small, screen all positive welding light signals at first moments based on the arrangement sequence to obtain first-moment screening positive welding light signals, select signal intensity minimum values and signal intensity maximum values in all first-moment screening positive welding light signals, and determine a positive welding light signal distribution range corresponding to the first welding moment based on the signal intensity minimum values and the signal intensity maximum values.

The invention also provides a welding light signal processing device based on the welding detection light path, which further comprises: an optimal detection model generation module;

the optimal detection model generation module is used for acquiring a plurality of welding sample optical signals, inputting the plurality of welding sample optical signals into an initial detection model, enabling the initial detection model to perform signal detection on the plurality of welding sample optical signals based on the welding detection reference range, and performing super-parameter optimization processing on the initial detection model based on the plurality of negative welding sample optical signals after determining that a plurality of negative welding sample optical signals are detected, so as to obtain an optimal detection model.

In a possible implementation manner, the optimal detection model generation module comprises a welding sample optical signal acquisition sub-module, an initial detection model setting sub-module, an initial detection model detection sub-module, a welding sample result comparison sub-module and a negative sample welding sample optical signal selection sub-module;

the welding sample optical signal acquisition submodule is used for acquiring a plurality of welding sample optical signals;

the initial detection model setting submodule is used for setting a plurality of initial detection models, wherein the plurality of initial detection models comprise a variance characteristic detection model, an average offset detection model, an area detection model and a time detection model;

the initial detection model detection submodule is used for arbitrarily selecting two initial detection models, respectively inputting the plurality of welding sample optical signals into the two initial detection models, respectively detecting the plurality of welding sample optical signals by the two initial detection models based on the welding detection reference range, and respectively outputting welding sample detection results of the plurality of welding sample optical signals by the two initial detection models;

the welding sample result comparison sub-module is used for comparing the welding sample detection result with welding sample results corresponding to the plurality of welding sample optical signals, and obtaining accurate values corresponding to the two initial detection models based on the comparison result;

And the negative sample welding sample optical signal selecting sub-module is used for selecting a plurality of negative sample welding sample optical signals based on the welding sample result when the accurate value is determined to meet the preset threshold value.

In one possible implementation manner, the optimal detection model generation module comprises a controller setting sub-module, a markov decision making sub-module, a model parameter adjustment sub-module and a model training sub-module;

the controller setting submodule is used for setting a model selection controller and a super-parameter controller corresponding to each initial detection model, wherein the model selection controller is used for selecting a target initial detection model for the plurality of initial detection models, and the super-parameter controller is used for performing super-parameter optimization on the target initial detection model;

the Markov decision building module is used for modeling a process of selecting the target initial detection model based on the model selection controller and a process of performing super-parameter optimization on the target initial detection model based on the super-parameter controller into a Markov decision process;

the model parameter adjustment sub-module is used for automatically adjusting the target initial detection model selected by the model selection controller based on the Markov decision process, automatically adjusting the super parameters of the target initial detection model, and recording all the selected target initial detection models and the super parameters corresponding to the target initial detection models;

The model training submodule is used for respectively carrying out model training on all the selected target initial detection models based on the plurality of negative sample welding light signals so as to enable the super parameters corresponding to the target initial detection models to obtain optimal detection models according to model training results.

In one possible implementation manner, the model selection controller is respectively connected with the super parameter controller corresponding to each initial detection model;

The invention also provides a welding light signal processing device based on the welding detection light path, which further comprises: the laser light-emitting power signal acquisition module and the laser welding equipment stability result acquisition module;

The laser light emitting power signal acquisition module is used for reflecting a welding beam emitted by the laser welding equipment into the power detection equipment so as to enable the power detection equipment to acquire a laser light emitting power signal of the welding beam;

the laser welding equipment stability result acquisition module is used for carrying out equipment stability evaluation on the laser welding equipment based on the laser light-emitting power signal to obtain a laser welding equipment stability result.

In one possible implementation manner, the laser welding equipment stability result obtaining module comprises a regional light emitting power calculating sub-module and a laser welding equipment stability result outputting sub-module;

the area light-emitting power calculation sub-module is used for dividing the laser light-emitting power signal into a plurality of area laser light-emitting power signals based on the light-emitting area, and calculating area light-emitting power corresponding to each area laser light-emitting power signal respectively;

the laser welding equipment stability result output submodule is used for counting the power variance of the area light-emitting power corresponding to the laser light-emitting power signals of each area in a preset time period, judging whether the power variance is larger than a preset power variance threshold, if yes, outputting the laser welding equipment stability result to be that the laser welding equipment needs to be maintained, otherwise, outputting the laser welding equipment stability result to be that the laser welding equipment does not need to be maintained.

The invention also provides a terminal device, which comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, wherein the processor realizes the welding light beam adjusting method based on the welding detection light path or the welding light signal processing method based on the welding detection light path when executing the computer program.

The invention also provides a computer readable storage medium, which comprises a stored computer program, wherein the computer program controls equipment where the computer readable storage medium is located to execute the welding light beam adjusting method based on the welding detection light path or the welding light signal processing method based on the welding detection light path when running.

Compared with the prior art, the method and the device for adjusting the light beam and processing the light signal based on the welding detection light path have the following beneficial effects:

acquiring a welding focus of a welding beam emitted by laser welding equipment and falling on a welding working surface, and calibrating a calibration beam emitted by a laser generator based on the electric control calibrator so that the calibration focus of the calibration beam falling on the welding working surface coincides with the welding focus, and realizing automatic adjustment and calibration of the beam by the electric control calibrator, thereby accurately positioning the welding beam, improving welding precision and welding qualification rate, reducing time waste in a process flow, and further improving production efficiency; after the alignment focus is determined to coincide with the welding focus, reflecting welding light beams emitted by laser welding equipment into the welding detection light path, so that the welding light beams are transmitted into the beam splitter based on the electric control aligner, and the welding light beams are split based on the beam splitter to obtain welding beam splitters with different wave bands; each welding beam-splitting beam is transmitted to the corresponding electric control attenuator through an optical fiber, so that the electric control attenuator carries out light attenuation adjustment on the welding beam-splitting beam to obtain welding light attenuation beams, and the electric control attenuator carries out light attenuation adjustment on the welding light attenuation beams with different wave bands to obtain required welding light signals, thereby improving the sensitivity and the detection precision of a welding detection light path; compared with the prior art, the technical scheme of the invention adopts a double-combination mode of the electric control calibrator and the electric control attenuator, and can realize the focal point adjustment and the light intensity control of the light beam by an accurate electric control means, so that the welding process is safer and more efficient; meanwhile, the optical fiber transmission mode is adopted, so that the layout of a detection light path is more flexible and convenient, the interference of the external environment on the welding process is reduced, and the whole detection system is conveniently monitored and regulated.

Drawings

FIG. 1 is a flow chart of an embodiment of a welding beam adjustment method based on a welding detection light path provided by the present invention;

FIG. 2 is a schematic diagram of an embodiment of a welding light signal processing device based on a welding detection light path according to the present invention;

FIG. 3 is a flow chart of an embodiment of a welding beam adjustment method based on a welding detection light path provided by the present invention;

FIG. 4 is a schematic diagram of an embodiment of a welding light signal processing device based on a welding detection light path according to the present invention;

FIG. 5 is a schematic diagram of the optical path of a collimated beam in accordance with one embodiment of the present invention;

FIG. 6 is a schematic diagram of an electronically controlled calibrator in accordance with an embodiment of the present invention;

FIG. 7 is a schematic diagram of another embodiment of an electronically controlled calibrator in accordance with the present invention;

FIG. 8 is a schematic rear view of an electronically controlled calibrator structure in accordance with an embodiment of the present invention;

FIG. 9 is a schematic diagram of a welding detection light path according to an embodiment of the present invention;

FIG. 10 is a schematic view of a beam splitter according to an embodiment of the present invention;

FIG. 11 is a schematic view of a welding detection light path according to an embodiment of the present invention;

FIG. 12 is a schematic diagram of an electrically controlled attenuator according to one embodiment of the present invention;

FIG. 13 is a schematic view of a rotating disk structure according to an embodiment of the present invention;

FIG. 14 is a schematic diagram of the hardware frame structure of a laser welding in-weld detector system according to one embodiment of the present invention;

FIG. 15 is a schematic diagram of a top-level module logic design of a control system based on SOC (ARM+programmable logic) according to one embodiment of the present invention;

FIG. 16 is a schematic diagram of a data receiving process according to an embodiment of the present invention;

FIG. 17 is a schematic diagram of a temperature control flow for one embodiment provided by the present invention;

FIG. 18 is a schematic diagram of a light calibration control flow according to an embodiment of the present invention;

FIG. 19 is a schematic diagram of an automatic control flow of light attenuation according to an embodiment of the present invention;

FIG. 20 is a schematic diagram of the general form of a welding light signal of one embodiment provided by the present invention;

FIG. 21 is a schematic view of the internal structure of an intelligent agent according to an embodiment of the present invention;

FIG. 22 is a schematic diagram of a model selection controller according to one embodiment of the present invention;

FIG. 23 is a schematic diagram of a super parameter controller architecture and decision making process according to one embodiment of the present invention;

FIG. 24 is a schematic diagram of a welding light signal processing device based on a welding detection light path according to an embodiment of the present invention;

fig. 25 is a schematic diagram of still another structure of a welding light signal processing device based on a welding detection light path according to an embodiment of the present invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Embodiment 1, referring to fig. 1, fig. 1 is a flow chart of an embodiment of a welding beam adjustment method based on a welding detection light path according to the present invention, as shown in fig. 1, and the method includes steps 101 to 103, specifically as follows:

step 101, obtaining a welding focus of a welding beam emitted by a laser welding device on a welding working surface, and calibrating a calibration beam emitted by a laser generator based on the electric control calibrator so as to enable the calibration focus of the calibration beam on the welding working surface to coincide with the welding focus.

In an embodiment, a laser welding apparatus is arranged above a welding face, to which a welding beam is emitted based on the laser welding apparatus, such that the welding beam forms a welding focus on the welding face.

In an embodiment, the welding detection light path includes a laser generator, the generator is disposed in the welding detection light path, and based on the laser generator, a calibration beam is emitted, so that the calibration beam is reversely transmitted in the welding detection light path through an optical fiber, is reflected by the electronic control calibrator, is output from an inlet end of the welding detection light path, and based on a 45-degree beam splitting sheet, the calibration beam is reflected to the welding working surface, so that the calibration beam forms a calibration focus on the welding working surface; as shown in fig. 5, fig. 5 is a schematic diagram of the optical path of the calibration beam.

In one embodiment, when the welding quality of the laser welding device is analyzed, the welding beam emitted by the laser welding device needs to be reflected to the 45-degree beam splitter by the welding workbench, so that the 45-degree beam splitter reflects the welding beam to a welding detection light path, the welding detection light path transmits the welding beam to a beam detection system, and the welding quality of the laser welding device is obtained based on the analysis of the welding beam by the beam detection system; because the welding detection light path needs to receive the reflected welding light beam, in order to ensure that the welding light beam reflected into the welding detection light path and the welding light beam emitted by the laser welding equipment have the same focal position and shape, the accuracy of subsequent welding quality detection is improved; in this embodiment, the calibration beam emitted by the laser generator is further calibrated based on the electronic control calibrator, so that the calibration focal point of the calibration beam falling on the welding working surface coincides with the welding focal point, and deviation between the welding beam input into the welding detection light path and the actual welding beam is avoided, thereby causing inaccuracy or misjudgment of the subsequent detection result.

In one embodiment, the calibrating light beam emitted by the laser generator is calibrated based on the electric control calibrator so that the calibrating focal point of the calibrating light beam falling on the welding working surface coincides with the welding focal point, and specifically, the calibrating light beam emitted by the laser generator is emitted into a welding detection light path so that the electric control calibrator in the welding detection light path reversely reflects the calibrating light beam to the welding working surface; acquiring a first image of the welding working surface, and carrying out graying treatment on the first image to obtain a first gray image; judging whether a calibration focus of the calibration light beam falling on the welding working surface coincides with the calibration focus or not according to the first gray level image, if so, considering that the electric control calibrator finishes calibrating the calibration light beam, otherwise, adjusting the electric control calibrator to change the position of the calibration focus on the welding working surface until the calibration focus coincides with the welding focus.

In one embodiment, a camera is arranged on the welding working surface, and a first image on the welding working surface is acquired based on the camera; and based on a transmission module arranged in the camera, transmitting the first image to upper computer software of a data processing terminal, so that the upper computer software carries out gray processing on the first image, and judging whether the calibration focus is overlapped with the calibration focus or not based on the first gray image.

In an embodiment, when the calibration focus is considered to be misaligned with the calibration focus based on the first gray scale image and the electric control calibrator is adjusted, a welding working surface coordinate axis is set by taking the welding focus on the welding working surface as an origin, a first coordinate position of the calibration focus on the welding working surface coordinate axis is obtained, a first x-axis coordinate variable and a first y-axis coordinate variable are obtained based on the first coordinate position, and the electric control calibrator is adjusted based on the first x-axis coordinate variable and the first y-axis coordinate variable.

Specifically, since the obtained first x-axis coordinate variable and the first y-axis coordinate variable represent the offset of the calibration focus relative to the welding focus on the welding working surface coordinate axis, at this time, the parameters of the electronically controlled calibrator are adjusted based on the first x-axis coordinate variable and the first y-axis coordinate variable, so that the position of the calibration beam can be changed, and the calibration focus is moved to a position overlapping with the welding focus.

In one embodiment, the electronic control calibrator comprises a reflector 1, a support rod 2, an origin lens linking universal joint 3 and a base 4; the support rod 2 is disposed above the base 4, the support rod 2 is perpendicular to the base 4, a first end of the origin lens link universal joint 3 is connected with the support rod 2, and a second end of the origin lens link universal joint 3 is connected with the reflection lens 1.

Preferably, the origin lens-linked gimbal 3 includes an inner ball and an outer ball, the inner ball is fixed on a connecting shaft connected with the reflection lens 1, the outer ball is fixed on the support rod 2, and the inner ball and the outer ball can rotate and twist relatively, and still maintain connection with the reflection lens 1 and the support rod 2 when rotating and twisting.

Preferably, a rotating motor is arranged on the supporting rod 2, and the rotating motor provides power for the origin lens linking universal joint 3, so that the origin lens linking universal joint 3 drives the reflecting lens 1 to move based on the power.

Preferably, the initial angle of the reflection mirror plate 1 is set to 45 degrees.

In an embodiment, when the electronic control calibrator is adjusted based on the first x-axis coordinate variable and the first y-axis coordinate variable, a coordinate axis of the reflecting mirror 1 is set by taking a center point of the reflecting mirror 1 as a reference point; taking the first x-axis coordinate variable as a y-axis coordinate point and the first y-axis coordinate variable as an x-axis coordinate point, and obtaining a second coordinate position of the first coordinate position on the coordinate axis of the reflector based on the y-axis coordinate point and the x-axis coordinate point; obtaining a second x-axis coordinate variable and a second y-axis coordinate variable based on the second coordinate position; adjusting the movement of the mirror plate 1 on the x-axis of the mirror plate 1 coordinate axis based on the origin mirror plate link gimbal 3 based on the second x-axis coordinate variable, and adjusting the movement of the mirror plate 1 on the y-axis of the mirror plate coordinate axis based on the origin mirror plate link gimbal 3 based on the second y-axis coordinate variable; so that the reference point coincides with the second coordinate position.

As a preferred embodiment in this embodiment, when the electronic control calibrator includes the reflecting mirror 1, the supporting rod 2, the origin lens linking gimbal 3 and the base 4, the electronic control calibrator may further include a Y-axis moving arm and an X-axis moving arm, where the Y-axis moving arm includes a Y-axis and lens linking gimbal 5, a Y-axis and lens push rod motor 6, a Y-axis and lens spindle linking gimbal 7, and the X-axis moving arm includes an X-axis and lens spindle linking gimbal 8, an X-axis and lens push rod motor 9, and an X-axis and lens linking gimbal 10.

In one embodiment, the Y-axis and lens spindle link universal joint 7 is disposed on the first side of the support rod 2, the Y-axis and lens push rod motor 6 is connected with the Y-axis and lens spindle link universal joint 7, and the Y-axis and lens link universal joint 5 is connected with the Y-axis and lens push rod motor 6; the X-axis and lens spindle linking universal joint 8 is arranged on the second side of the supporting rod 2, the X-axis and lens push rod motor 9 is connected with the X-axis and lens spindle linking universal joint 8, and the X-axis and lens linking universal joint 10 is connected with the X-axis and lens push rod motor 9; the X-axis and lens link gimbal 10 and the Y-axis and lens link gimbal 5 are connected to the reflecting lens 1, respectively.

Preferably, the Y-axis and lens link universal joint 5 is connected with the Y-axis and lens push rod motor 6 through a Y-axis telescopic shaft, and the X-axis and lens link universal joint 10 is connected with the X-axis and lens push rod motor 9 through an X-axis telescopic shaft. Fig. 7 is a schematic diagram of still another structure of the electronically controlled calibrator as shown in fig. 7, and fig. 8 is a schematic diagram of a rear view of the electronically controlled calibrator structure as shown in fig. 8.

In an embodiment, when the electronic control calibrator is adjusted based on the first x-axis coordinate variable and the first y-axis coordinate variable, a coordinate axis of the reflecting mirror is set by taking a center point of the reflecting mirror 1 as a reference point; taking the first x-axis coordinate variable as a y-axis coordinate point and the first y-axis coordinate variable as an x-axis coordinate point, and obtaining a second coordinate position of the first coordinate position on the coordinate axis of the reflector based on the y-axis coordinate point and the x-axis coordinate point; obtaining a second x-axis coordinate variable and a second y-axis coordinate variable based on the second coordinate position; adjusting the X-axis and lens pusher motor 9 to perform telescopic movement based on the second X-axis coordinate variable so as to control the X-axis and lens link universal joint 10 to adjust the movement of the reflecting lens 1 on the X-axis of the reflecting lens coordinate axis based on the X-axis telescopic axis, and adjusting the Y-axis and lens pusher motor 6 to perform telescopic movement based on the second Y-axis coordinate variable so as to adjust the movement of the reflecting lens 1 on the Y-axis of the reflecting lens coordinate axis based on the Y-axis and lens link universal joint 5 based on the Y-axis telescopic axis control; so that the reference point coincides with the second coordinate position.

In an embodiment, when the calibration light is calibrated based on the electric control calibrator, based on adjustment through a software function on an operation interface, when the data processing terminal detects that the calibration focus is not coincident with the calibration focus, a corresponding calibration variable is output, so that automatic adjustment of the electric control calibrator is realized, the efficiency of a production line can be improved, and the method is different from the conventional design, the reflection lens is manually adjusted generally based on manual operation, and errors caused by manual adjustment are avoided.

102, after the alignment focus is determined to coincide with the welding focus, reflecting welding light beams emitted by a laser welding device into the welding detection light path, so that the welding light beams are transmitted into the beam splitter based on the electric control aligner, and the welding light beams are split based on the beam splitter to obtain welding beam splitters in a plurality of different wave bands;

in an embodiment, a welding beam emitted by a laser welding device is emitted to a welding workbench, the welding beam is reflected based on the welding workbench so as to reflect the welding beam to a preset reflection sheet, and the welding beam is reflected to an entrance end of a welding detection light path based on the reflection sheet so as to reflect the welding beam to the welding detection light path through the entrance end.

In one embodiment, the welding detection light path includes the electrically controlled calibrator 100, the beam splitter 200, and the electrically controlled attenuator 300, and the welding detection light path further includes the collimator 400. As shown in fig. 9, fig. 9 is a schematic structural view of the welding detection light path.

In an embodiment, the beam collimator is configured based on a propagation direction of the welding beam after passing through the electric control calibrator, and the beam collimator is configured to perform light receiving processing on the welding beam after passing through the electric control calibrator, and transmit the light-received welding beam after light receiving into the beam splitter.

In an embodiment, the beam splitter 200 includes a first beam splitter F1 and a second beam splitter F2, where the first beam splitter F1 and the second beam splitter F2 are disposed inside the beam splitter, and the first beam splitter F1 and the second beam splitter F2 are sequentially disposed according to a preset beam splitter disposition distance based on a propagation direction of a welding beam in the beam splitter, so that the welding beam sequentially passes through the first beam splitter F1 and the second beam splitter F2 in the beam splitter, and the preset beam splitter disposition distance is set based on a user requirement, as shown in fig. 10, and fig. 10 is a schematic structural diagram of the beam splitter.

In one embodiment, the first beam splitter F1 and the second beam splitter F2 are 45 ° beam splitters.

In an embodiment, a first optical fiber G1, a second optical fiber G2, and a third optical fiber G3 are disposed in the welding detection optical path, the optical splitter 200 is connected to the first optical fiber G1, the second optical fiber G2, and the third optical fiber G3, and a first optical fiber coupling end surface of the first optical fiber G1, a second optical fiber coupling end surface of the second optical fiber G2, and a third optical fiber coupling end surface of the third optical fiber G3 are disposed inside the optical splitter.

In an embodiment, the first optical fiber coupling end surface is configured based on a first light splitting direction of the first light splitting sheet F1, so that the first welding light splitting beam subjected to the light splitting treatment by the first light splitting sheet F1 is received based on the first optical fiber coupling end surface, and the first welding light splitting beam is transmitted based on the first optical fiber G1.

In an embodiment, the second optical fiber coupling end surface is configured based on the first splitting direction of the second splitting piece F2, so that the second welding splitting beam split by the second splitting piece F2 is received based on the second optical fiber coupling end surface, and the second welding splitting beam is transmitted based on the second optical fiber G2.

In an embodiment, the third optical fiber coupling end face is configured based on a propagation direction of the welding beam in the beam splitter, and the third optical fiber coupling end face is configured to receive the third welding beam split after the second beam splitter F2 splits the welding beam, and transmit the third welding beam split based on the third optical fiber G3.

In an embodiment, when the welding beam is reflected to the welding detection light path through the inlet end, the welding beam is sent to the beam collimating mirror through the reflecting mirror 1 in the electric control collimator, and then the relative parallel light of the welding beam is received based on the beam collimating mirror, so that the diameter of the welding beam is reduced to an ideal state, the spot diameter of the welding beam is extremely small, and the received welding beam is sent to the beam splitter for beam splitting.

Preferably, a fourth optical fiber G4 may be further disposed in the welding detection optical path, the optical splitter 200 is connected to the fourth optical fiber G4, and a fourth optical fiber coupling end surface of the fourth optical fiber G4 is disposed inside the optical splitter 200. As shown in fig. 11, fig. 11 is a schematic view of still another structure of the welding detection light path.

Preferably, the fourth optical fiber coupling end face is configured based on a propagation direction of the welding beam in the optical splitter, and the fourth optical fiber coupling end face is configured to emit the welding beam to the interior of the optical splitter, so that the welding beam is emitted onto the first beam splitter F1.

Preferably, the beam collimator is connected to the fourth optical fiber G4, and a fifth optical fiber coupling end surface of the fourth optical fiber G4 is configured to collect the received welding beam after being received by the beam collimator, transmit the received welding beam based on the fourth optical fiber G4, and input the received welding beam to the beam splitter based on the fourth optical fiber coupling end surface of the fourth optical fiber G4.

Preferably, when the fourth optical fiber G4 is disposed in the welding detection optical path and the welding beam is reflected to the welding detection optical path through the inlet end, the welding beam is sent to the beam collimating mirror by the reflecting mirror 1 in the electronic control collimator, the parallel light of the welding beam is received based on the beam collimating mirror, and the received welding beam is sent to the fifth optical fiber coupling end surface, so that the welding beam is transmitted based on the fifth optical fiber, and the welding beam is emitted to the first beam splitter F1 based on the fourth optical fiber coupling end surface.

In an embodiment, the welding light beam calibrated by the electric control calibrator is received by the light beam collimator through the optical fiber and then transmitted into the beam splitter, so that the split design of the beam splitter and the electric control calibrator in the optical fiber detection light path is realized, and the split design is different from the existing detection light path which is of a fixed structure and is not beneficial to installation and debugging.

In an embodiment, the beam splitter splits the welding beam to obtain a plurality of welding beam splitters with different wavebands, specifically, the welding beam is transmitted to the first beam splitter F1, and the welding beam is split based on the first beam splitter F1, so that the welding beam is split into a welding beam with a first waveband and a first welding beam.

In one embodiment, the first welding beam is transmitted to the second beam splitter F2, and the first welding beam is split based on the second beam splitter F2, so that the first welding beam is split into a second band welding beam splitter and a third band welding beam splitter.

In one embodiment, the first beam splitter F1 is a cut-off filter between 370 and 750nm, and the first beam splitter F1 is configured to cut off a welding beam with a wavelength between 370 and 750nm, so as to split a first band welding beam with a wavelength between 370 and 750nm from the welding beam, reflect the first band beam into a first fiber coupling end surface, so as to send the first band beam into the first fiber G1, set a beam with another wavelength that is not cut off as a first welding beam, and send the first welding beam into the second beam splitter F2.

In one embodiment, the second beam splitter F2 is a cut-off filter with 1070nm±25nm, the first beam splitter F1 is used for cutting off the welding beam with the wavelength of 1070nm±25nm, so that the welding beam with the wavelength of 1070nm±25nm is split from the first welding beam, and the second band split beam is reflected to the coupling end surface of the second optical fiber, so that the second band split beam is sent to the second optical fiber G2, and since the first welding beam after being split by the second beam splitter F2 only leaves the welding beam with the wavelength of 1150-1700nm, no cut-off filter is needed, but the beam with other wavelength which is not cut off is set as the third welding split beam, and the third welding split beam is vertically aligned to the coupling end surface of the third optical fiber to be sent to the third optical fiber G3.

Step 103: and transmitting each welding beam-splitting beam to the corresponding electric control attenuator through an optical fiber so that the electric control attenuator can carry out light attenuation adjustment on the welding beam-splitting beam to obtain a welding light attenuation beam.

In an embodiment, the electrically controlled attenuator in the welding detection light path includes a first electrically controlled attenuator, a second electrically controlled attenuator, and a third electrically controlled attenuator.

In one embodiment, the first electrically controlled attenuator is connected to the first optical fiber G1, and the sixth coupling end face of the first optical fiber G1 is disposed inside the first electrically controlled attenuator; the second electric control attenuator is connected with the second optical fiber G2, and a seventh coupling end face of the second optical fiber G2 is arranged in the second electric control attenuator; the third electric control attenuator is connected with the third optical fiber G3, and an eighth coupling end face of the third optical fiber G3 is arranged in the third electric control attenuator.

In an embodiment, the welding beam splitter output from the beam splitter is transmitted to the electric control attenuator through the optical fiber 15, so as to realize the split design of the beam splitter and the electric control attenuator in the optical fiber detection optical path, which is different from the existing detection optical path which is of a fixed structure, and is not beneficial to installation and debugging.

In one embodiment, each electric control attenuator comprises a stepper motor 11, a rotary disk 12, a motor output shaft 13 and a light outlet 14, wherein the stepper motor 11 is connected with the motor output shaft 13, and the rotation of the motor output shaft 13 is controlled based on the stepper motor 11; the rotary disk 12 is fixed on the motor output shaft 13, and the rotary disk 12 is controlled to rotate based on the motor output shaft 13; the rotary disk 12 is provided with a plurality of attenuation sheets 16, and when the rotary disk 12 rotates, the attenuation sheets 16 sequentially correspond to the light outlet 14.

Preferably, the electric control attenuators are in a box structure, the stepper motor 11, the rotary disk 12, the motor output shaft 13 and the light outlet 14 are arranged in the box, and the optical fibers 15 corresponding to each electric control attenuator are embedded in the box; as shown in fig. 12, fig. 12 is a schematic structural diagram of an electrically controlled attenuator, and 16 in the drawing is an attenuation sheet on a rotating disk.

Specifically, the first electric control attenuator comprises a first stepping motor, a first rotating disc, a first motor output shaft and a first light outlet, wherein the first stepping motor is connected with the first motor output shaft, and the rotation of the first motor output shaft is controlled based on the first stepping motor; the first rotating disk is fixed on the first motor output shaft, and the rotation of the first rotating disk is controlled based on the first motor output shaft; the first rotary disk is provided with a plurality of attenuation sheets, and when the first rotary disk rotates, the attenuation sheets sequentially correspond to the first light outlet.

Specifically, the second electric control attenuator comprises a second stepping motor, a second rotating disc, a second motor output shaft and a second light outlet; the second stepping motor is connected with the second motor output shaft, and the second motor output shaft is controlled to rotate based on the second stepping motor; the second rotating disk is fixed on the second motor output shaft, and the rotation of the second rotating disk is controlled based on the second motor output shaft; and a plurality of attenuation sheets are arranged on the second rotating disk, and when the second rotating disk rotates, the attenuation sheets sequentially correspond to the second light outlet.

Specifically, the third electric control attenuator comprises a third stepping motor, a third rotating disc, a third motor output shaft and a third light outlet, wherein the third stepping motor is connected with the third motor output shaft, and the rotation of the third motor output shaft is controlled based on the third stepping motor; the third rotary disk is fixed on the third motor output shaft, and the rotation of the third rotary disk is controlled based on the third motor output shaft; and a plurality of attenuation sheets are arranged on the third rotary disk, and when the third rotary disk rotates, the attenuation sheets sequentially correspond to the third light outlet.

In an embodiment, the sixth optical fiber coupling end surface is configured based on the first light outlet of the first electrically controlled attenuator, so that a first welding split beam is output based on the sixth optical fiber coupling end surface facing the first light outlet.

In an embodiment, the seventh optical fiber coupling end surface is configured based on the second light outlet of the second electrically controlled attenuator, so that a second welding split beam is output based on the seventh optical fiber coupling end surface facing the second light outlet.

In an embodiment, the eighth optical fiber coupling end surface is configured based on a third light outlet of the third electrically controlled attenuator, so that a third welding split beam is output based on the eighth optical fiber coupling end surface facing the third light outlet.

In an embodiment, the first rotating disc is disposed between the sixth optical fiber coupling end face and the first light outlet, so that after the first welding beam is emitted by the sixth optical fiber coupling end face, the first welding beam is subjected to light attenuation adjustment based on an attenuation piece on the first rotating disc, so as to obtain a first welding light attenuation beam, and the first welding light attenuation beam is output through the first light outlet; the second rotating disk is arranged between the seventh optical fiber coupling end face and the second light outlet, so that after the second welding light splitting beam is emitted by the seventh optical fiber coupling end face, the second welding light splitting beam is subjected to light attenuation adjustment based on an attenuation sheet on the second rotating disk, a second welding light attenuation beam is obtained, and the second welding light attenuation beam is output through the second light outlet; the third rotating disk is arranged between the eighth optical fiber coupling end face and the third light outlet, so that after the third welding light splitting beam is emitted by the eighth optical fiber coupling end face, the third welding light splitting beam is subjected to light attenuation adjustment based on an attenuation sheet on the third rotating disk, a third welding light attenuation beam is obtained, and the third welding light attenuation beam is output through the third light outlet.

In an embodiment, the number of the attenuation sheets set on the first, second and third rotating discs may be set as required, preferably, five attenuation sheets are respectively built in each rotating disc 12, where the five attenuation sheets include attenuation sheets of 1%, 10%, 25%, 50% and the like and all-pass lenses, as shown in fig. 13, and fig. 13 is a schematic diagram of the rotating disc structure.

Preferably, the attenuation sheet is a polarizer, and can be used for selectively transmitting or blocking light with a specific polarization direction; the working principle of the polarizer is adopted to realize the adjustment and calibration of the laser optical signal intensity, the stepless light attenuation adjustment capability is realized, and meanwhile, the full-pass optical attenuation capability of the polarizer cannot be realized, so that the full-pass structural design is added in the design of the electric control attenuator, and the industrial design of the optical fiber welding laser detection optical path standard product which still can meet the requirements under multiple environments is realized.

In an embodiment, when the electric control attenuator performs optical attenuation adjustment on the welding beam splitter to obtain a welding beam splitter, a target attenuation sheet on the rotating disc 12 is selected based on a preset optical attenuation power by acquiring the preset optical attenuation power; after determining that the state of the stepping motor 11 is a starting state, controlling the rotating disc 12 to rotate based on the motor output shaft 13 so as to rotate a target attenuation piece on the rotating disc 12 to the position of the light outlet 14; and transmitting the welding beam-splitting beam to the target attenuation sheet so that the target attenuation sheet can carry out optical attenuation on the welding beam-splitting beam to obtain a welding beam-attenuation beam, and outputting the welding beam-attenuation beam based on the light outlet 14.

In one embodiment, for each electric control attenuator, the rotation of the rotary disk 12 is controlled based on the motor output shaft 13, so that when the target attenuation piece on the rotary disk 12 is rotated to the light outlet 14, a rotation angle required by each attenuation piece on the rotary disk 12 to be rotated to the light outlet 14 is set based on the rotation direction by acquiring the rotation direction of the motor output shaft 13; after selecting the target attenuation sheet on the rotating disc 12, acquiring a target rotation angle corresponding to the target attenuation sheet, and controlling the rotating disc 12 to rotate based on the rotation direction and the target rotation angle, so as to rotate the target attenuation sheet on the rotating disc 12 to the light outlet 14.

In one embodiment, the electric control attenuator is controlled by adopting the full software operation interface, so that the rotation angle of the motor in the electric control attenuator is controlled, thereby realizing a stepless light passing rate adjusting mechanism, replacing the traditional manual disassembly and installation attenuator adjusting process, greatly improving the working efficiency and reducing the operation usability of staff.

In an embodiment, the welding detection light path is provided with an SOC-based in-laser welding detector system, as shown in fig. 14, and fig. 14 is a schematic diagram of a hardware frame structure of the in-laser welding detector system; the laser welding in-process detector system mainly comprises an analog circuit for acquiring PD weak signals, an SOC (system on chip) is used as a main control, the amplification factor of the analog circuit is controlled, the acquisition of multiple ADC (analog-to-digital converter) is realized, an electric control attenuator is controlled, an electric control calibrator and a sensor temperature are controlled, and data interaction is carried out through a PHY (physical layer) chip and upper computer software, so that relevant data detected in a welding detection light path are timely transmitted to the upper computer software based on the PHY chip.

In an embodiment, the SOC includes two hardware parts including a programmable logic resource PL and an ARM A9 processor, as shown in fig. 15, and fig. 15 is a logic design schematic diagram of a top module of a control system based on SOC (arm+programmable logic); the programmable logic resource PLPL and ARM A9 processor can realize high-speed data interaction, specifically 4 high-performance DMA (Direct memory access) channels, the transmission speed of each channel is 1.2 Gbit/s, 2 physical address mapping PAM (Physical address mapping channel) buses are universal, and the transmission speed is 600 Mbit/s.

The logic design part mainly comprises the following logic IP core (Intellectual Property core: refer to a circuit functional module design designed in advance in an FPGA by using verilog hardware description language):

(1) And the multichannel DAC parallel data receiving logic IP controls the multichannel DAC (digital-to-analog converter) to realize digital signal acquisition of 8 paths of analog signals.

(2) The high-speed data interaction logic IP is used for realizing data interaction with other detection modules and endowing the system with an external rapid data interaction function; the main functional modules are as follows: serial data parsing and serial data parallelization.

(3) Two-way DMA+single-way PAM channel logic IP-realizes the rapid data transmission from logic PL to ARM, and one-way DMA is used for the parallel data rapid hand transmission of spectrum data to ARM; and (3) parallelizing external synchronous serial data, and using another DMA to realize rapid data transmission to the ARM.

(4) And the temperature control PID logic IP acquires temperature values through a plurality of ADC (analog-to-digital converter), and the temperature control PID logic IP realizes the temperature control PID of the sensor and ensures the stability of acquired signals by using the temperature values as temperature settings.

(5) Light calibration double-motor control logic IP-control of the rotation direction, rotation speed and total rotation angle of the two motors is realized, and automatic alignment of welding light beams in a measurement area is realized.

(6) The light attenuation automatic control logic IP realizes the control of the rotation direction, the rotation speed and the total rotation angle of the motor and realizes the automatic control of the optimal signal intensity.

(7) Each logic IP parameter setting and pipeline control logic IP-realizing each logic IP parameter setting and the shutdown control of the whole logic pipeline.

In one embodiment, for the multi-channel DAC parallel data receiving logic IP, as shown in fig. 16, fig. 16 is a schematic diagram of a data receiving flow; the data reception logic is designed to: firstly, setting DMA parameters through a parameter control module: set DMA burst length, DAC reset initialization, buffer FIFO (First Input First Output, first-in first-out data buffer circuit) reset initialization. The parallel data of the multi-path DAC and the parallel serial data firstly enter the FIFO to be buffered, which is to collect the data with the minimum burst length required by the DMA transmission and prevent the DMA transmission error. If the FIFO buffer data is detected to be larger than the DMA burst length, starting DMA data transmission, and after the transmission is finished, increasing the destination address of the DMA to prevent the data from being covered. And detecting whether to terminate data collection, resetting all parameters of the logic module if the data collection is terminated, and continuing to collect the data if the data collection is not terminated, wherein the data receiving pipeline is not shut down.

In one embodiment, as shown in fig. 17, for the temperature control PID logic IP, fig. 17 is a schematic diagram of a temperature control flow, and the sensor temperature control logic is designed to: if the temperature control function is enabled, the temperature control logic IP is started. The first step: and acquiring the current temperature of each sensor through a plurality of ADC. And a second step of: if the current temperature of the sensor is different from the preset value, the temperature of the sensor is set in a gradual approximation mode, and the interference of the excessive refrigeration current on the system can be prevented. And a third step of: judging whether to stop the temperature control process, if not, returning to reacquire the sensor temperature to continue the process. If the temperature control is stopped, the logic returns to the initial state waiting for an enable signal.

In one embodiment, as shown in fig. 18, fig. 18 is a schematic diagram of a light calibration control flow for a light calibration dual motor control logic IP, where the light calibration control logic IP is designed to: and the power-on initialization realizes the return to zero of the motor position, the initialization of the rotation angular velocity and the command buffer reset. If the motor drive module is enabled, it is monitored whether a control command arrives in the motion control command FIFO buffer. After the motion control command is analyzed, the current state of the motor is checked, if the motor is busy, the motor is waited, and if the motor is idle, the motor driving logic module is enabled to drive the motor to move. Once the light calibration is completed, the logic module needs to be dormant to prevent accidental driving of the motor, resulting in detection failure.

In one embodiment, as shown in fig. 19, fig. 19 is a schematic diagram of an optical attenuation automatic control logic IP, where the optical attenuation automatic control logic IP is designed to: and the power-on initialization realizes the return to zero of the motor position, the initialization of the rotation angular velocity and the reset of the motion control command buffer FIFO. If the automatic control module of light attenuation is enabled, the signal intensity of the current welding process is collected, and if the signal intensity is within the preset value intensity range, the logic module is forbidden. If the signal intensity is abnormal: and if the supersaturation and the intensity are too high or too low, sending a command to the motion control command FIFO buffer memory to change the clearance quantity of the brake motor, and adjusting the signal intensity to be within a preset value range. Once the optical attenuation control process is completed, the logic module needs to be dormant to prevent signal interference.

In summary, the welding light beam adjusting method based on the welding detection light path provided by the invention adopts a double-combination mode of the electric control calibrator and the electric control attenuator, and can realize the focal point adjustment and the light intensity control of the light beam by an accurate full-automatic electric control means, thereby greatly improving the working efficiency and ensuring the welding process to be safer and more efficient; meanwhile, the blocking design of the welding detection light path can be realized by adopting an optical fiber transmission mode, and the device structure interference corresponding to the laser welding stations under different conditions can be adapted, so that the layout of the welding detection light path is more flexible and convenient, the interference of the external environment on the welding process is reduced, and the monitoring and the adjustment of the whole detection system are facilitated; in the prior art, need artifical scale, manual regulation light path, manual work set up optical signal intensity etc. lead to unable industrialization and intelligent of realizing fast, reduce the practicality of product.

Embodiment 2, referring to fig. 2, fig. 2 is a flow chart of an embodiment of a welding beam adjustment method based on a welding detection light path according to the present invention, as shown in fig. 2, and the method includes steps 201 to 203, specifically as follows:

step 201: and acquiring a historical welding light signal set, acquiring a plurality of positive welding light signals from the historical welding light signal sets, and generating a welding detection reference range when the positive welding light signals meet positive judgment standards.

In one embodiment, during the laser welding process, the welding light signal collected by the system may be affected by various factors, such as laser stability problem in the tolerance range, influence of ambient light, workpiece variability in the tolerance limit, etc., which all cause abnormality in the distribution of a few positive welding light signals; therefore, when the welding detection reference range is generated by using the positive welding light signal, in order to improve the accuracy of the welding detection reference range, it is also necessary to determine the obtained positive welding light signal so as to eliminate the abnormal welding light signal possibly accumulated in the positive welding light signal.

In one embodiment, the positive welding light signal is a signal acquired by a detection system under the normal condition of welding after the welding process is debugged; the negative welding light signal is defined as a signal acquired by a detection system when welding fails.

In an embodiment, when a positive sample judgment standard is detected on a plurality of collected positive sample welding light signals, a plurality of positive sample welding light signals are collected from the historical welding light signal set, and a scanning window is set based on the signal length of each positive sample welding light signal; window scanning is carried out on each positive welding light signal based on the scanning window, and when the variance of the distribution of the positive welding light signals in the scanning window is larger than a preset variance, the positive welding light signals are determined to be abnormal welding light signals; deleting the abnormal welding light signals in the plurality of positive welding light signals to obtain residual positive welding light signals, and taking the residual positive welding light signals as positive welding light signals meeting positive judgment standards.

Preferably, the size of the scanning window is set to be 1/10 of the total length of the positive welding light signal, and if the variance of signal distribution in the window is greater than 2 times of the preset reference boundary variance, the positive welding light signal is determined to be an abnormal welding light signal; otherwise, the plurality of positive sample welding light signals are considered to meet the positive sample judgment standard.

In an embodiment, in the laser welding process, corresponding to each welding point of non-physical contact, the light emitting form of the laser is segmented, so that the continuously acquired welding light signals also need to be detected in a partitioned manner, as shown in fig. 20, fig. 20 is a schematic diagram of the general form of the welding light signals, a convex area of the welding light signals in fig. 20 is the welding light signals acquired in the light emitting stage, a horizontal axis X is the acquisition time, and a vertical axis Y is the sampling value intensity; since only the welding light signal of the light stage needs to be detected when the welding light signal is detected, it is also necessary to partition the welding light signal.

In one embodiment, the light emitting stage of the whole welding light signal is determined by a plurality of positive welding light signals, and one light emitting stage is a partition; setting a partition reference as acquiring data of all positive welding light signals at a certain moment, and fitting a distribution range of the positive welding light signals at a certain moment through the data of all positive welding light signals, wherein the fitted range is a welding detection reference range at the moment.

Specifically, determining light emitting areas of the welding light signals based on the plurality of positive welding light signals, and respectively dividing the plurality of positive welding light signals into a plurality of area positive welding light signals based on the light emitting areas so as to divide each positive welding light signal into a plurality of area positive welding light signals; and classifying the positive welding light signals of the multiple regions based on the light emitting regions to obtain all positive welding light signals of the first regions corresponding to each light emitting region, and obtaining the positive welding light signals of the first regions at the first moment corresponding to the first welding moment.

Specifically, performing data fitting processing on all positive welding light signals at a first moment to obtain a positive welding light signal distribution range corresponding to the first welding moment, and integrating the positive welding light signal distribution ranges corresponding to all welding moments in the positive welding light signals of the first region to obtain a region positive welding light signal distribution range corresponding to the positive welding light signals of the first region; and integrating the distribution range of the positive welding light signals of the areas corresponding to each light-emitting area to obtain a welding detection reference range.

In an embodiment, a plurality of welding light signal distribution range generation manners are further provided, where the plurality of welding light signal distribution range generation manners include a first welding light signal distribution range generation manner based on an assumed normal distribution method and a second welding light signal distribution range generation manner based on a percentage method.

Specifically, for the assumed normal distribution method:

wherein mu is the average value of positive sample welding optical signals, and rho is the standard deviation; assuming that the positive welding light signal data distribution on the Y axis is normal distribution, a Gaussian distribution function is fitted through the positive welding light signal data, and a region within 3 standard deviation ranges of the distribution function is taken as a reference range.

Specifically, for the percentage method:

reference rangeThe sample values off the data distribution center were subtracted in percent.

In the raw data distribution, off-center portions of the data are subtracted as a percentage of the total collection number. The treatment process comprises the following steps: assuming that the total number of data samples is 100 at a time, the percentage of subtraction is set to 5%; the treatment process comprises the following steps: in a first step, 100 sample data are size ordered. Second, the minimum 5 data are removed, the maximum 5 data are removed, and the signal intensity distribution range [ y ] is obtained from the rest 90 data ₁ ,y ₂ ]。

In an embodiment, when the data fitting process is performed on all the positive welding light signals at the first moment to obtain the positive welding light signal distribution range corresponding to the first welding moment, a target welding light signal distribution range generation mode selected by a user from a plurality of preset welding light signal distribution range generation modes is detected.

In an embodiment, when it is detected that a user selects a first welding light signal distribution range generation mode from a plurality of preset welding light signal distribution range generation modes, performing data fitting processing on all positive welding light signals at a first moment based on a assumed normal distribution function to obtain an assumed normal distribution value, and selecting an area of the assumed normal distribution value within a preset standard deviation range as a positive welding light signal distribution range corresponding to the first welding moment.

In an embodiment, when it is detected that a user selects a second welding light signal distribution range generation mode from a plurality of preset welding light signal distribution range generation modes, all first-time positive welding light signals are arranged according to a sequence from big to small, all first-time positive welding light signals are screened based on the arrangement sequence, first-time screening positive welding light signals are obtained, a signal intensity minimum value and a signal intensity maximum value in all first-time screening positive welding light signals are selected, and a positive welding light signal distribution range corresponding to the first welding moment is determined based on the signal intensity minimum value and the signal intensity maximum value.

Step 202: acquiring a welding light attenuation beam based on a PD sensor to obtain a welding light signal; wherein the welding light attenuated beam is obtained according to the welding light beam adjustment method based on the welding detection light path described in the above embodiment 1.

In an embodiment, the welding detection light path further includes a PD sensor, where the PD sensor includes a first PD sensor, a second PD sensor, and a third PD sensor; the first PD sensor is used for receiving light beams with the wavelength of 370-750nm, the second PD sensor is used for receiving light beams with the wavelength of 1070nm plus or minus 25nm, and the third PD sensor is used for receiving light beams with the wavelength of 1150-1700 nm.

In an embodiment, before the welding light attenuation beam is collected based on the sensor, the method further comprises initializing collection parameters, wherein the initialization of the collection parameters mainly realizes that the sensor is in an unsaturated state through adjustment of light intensity and electronic circuit amplification factor, and the signal intensity is located in a system range, so that the adaptation of various welding processes is realized.

In an embodiment, the first PD sensor is configured to receive a first welding light attenuated beam output from the first light outlet in the first electronically controlled attenuator; the second PD sensor is used for receiving a second welding light attenuation beam output by the second light outlet in the second electric control attenuator; the third PD sensor is used for receiving a third welding light attenuation beam output by the third light outlet in the third electric control attenuator; because the welding light attenuation light beams received by each PD sensor are subjected to light reduction adjustment with different intensities through the electric control attenuator, the risk of failure of the PD sensor due to saturation can be reduced, and the stability in the welding detection process is improved.

In one embodiment, since the PD is a photoelectric conversion device, when illuminated, it generates a current proportional to the light intensity, so after the PD sensor collects the welding light attenuated beam, the PD sensor generates a corresponding current, and the current detection circuit or the measuring instrument is used to measure the current generated by the PD sensor, so as to obtain the welding light signal corresponding to the welding light attenuated beam.

Step 203: and inputting the welding light signal into a pre-trained optimal detection model, so that the optimal detection model detects the welding light signal based on a welding detection reference range to obtain a welding detection result.

In one embodiment, a plurality of initial detection models are provided, wherein the plurality of initial detection models includes, but is not limited to, a variance feature detection model, an average offset detection model, an area detection model, and a time detection model.

Specifically, for the variance feature detection model: for signal intensity distribution stability assessment.

The data processing method comprises the following steps: and counting the variances of the one-dimensional signals of the acquired welding light signals.

Defect definition: the welding light signal variance value is greater than a given determination value.

Model superparameter: it is determined whether it is a limit variance value of the defect.

Specifically, for the average offset detection model: for signal overall strength offset assessment.

The data processing method comprises the following steps: calculating the average offset of the one-dimensional signal relative to the reference center line, namely:

wherein N is the signal length, y _t Is the signal value of the signal at time t. s is(s) _t Is the signal strength of the reference centerline at time t.

Defect definition: the signal average offset exceeds a given decision value.

Model superparameter: it is determined whether it is an average offset limit value for the defect.

Specifically, for the area detection model: for signal local in-range anomaly assessment-severely beyond the baseline range and for a certain period of time.

The data processing method comprises the following steps: for the area beyond the reference range, the area beyond the boundary of the reference range is calculated:

wherein t is ₁ For the start time of signal out of range, t ₂ For signal out-of-range end time, y _t For the signal value, l, of the signal at time t _t The signal strength of the reference boundary line at time t.

Defect definition: the actual signal exceeds a certain area of the reference range, namely the area between the signal and the reference boundary, the unit is volt-second, the duration exceeds the specified time, the limit time exceeding the reference range is considered to be defined as the tolerance time, if the limit time is smaller than the tolerance time, the signal is judged to be normal, and if the limit time exceeds the tolerance time, the signal is judged to be defective.

Model superparameter: upper/lower area limit values and a tolerance time for determining whether or not it is a defect.

Specifically, for the time detection model: for signal local in-range anomaly assessment-continuous long time exceeding the reference range.

The data processing method comprises the following steps: the judgment mode beyond the reference range is as follows:

signal intensity > reference boundary value + limit value considered to be given;

for the area beyond the reference range, the time beyond the reference range boundary is calculated:

overrun time=t ₂ ―t ₁ ；

Wherein t is ₁ For the start time of signal out of range, t ₂ Is the end time of the signal out of range.

Defect definition: the actual signal exceeds the limit value given according to the reference range for a certain time.

Model superparameter: upper/lower limit values and a tolerance time for determining whether or not it is a defect.

In an embodiment, a plurality of welding sample optical signals are obtained, the plurality of welding sample optical signals are input into an initial detection model, so that the initial detection model performs signal detection on the plurality of welding sample optical signals based on the welding detection reference range, and after determining that a plurality of negative sample welding sample optical signals are detected, super-parameter optimization processing is performed on the initial detection model based on the plurality of negative sample welding sample optical signals, so as to obtain an optimal detection model.

In an embodiment, since the number of the preset initial detection models is multiple, in the initial stage, the multiple preset initial detection models are not all enabled, and according to the laboratory sample data, the preset number of initial detection models, for example, two initial detection models in the preset multiple initial detection models are enabled, so that the sample data manufactured in the laboratory can be completely and correctly detected; and the initial detection model selection and the super-parameter initialization are given by laboratory experiments, and fine adjustment is performed during use so as to ensure that excessive misjudgment and missed judgment are avoided in the starting stage.

In one embodiment, since the initial detection model is initially selected and parameter set based on laboratory experience sample data, a small amount of negative sample welding sample optical signals need to be obtained, model selection and model parameters can be further optimized based on the negative sample welding sample optical signals, and when enough negative sample welding sample optical signals are acquired, the initial detection model is selected through reinforcement learning, and the corresponding model parameters are optimized; preferably, the number of optical signals of the obtained negative sample welding samples is more than 20.

In an embodiment, after a plurality of welding sample optical signals are acquired, two initial detection models are arbitrarily selected, the plurality of welding sample optical signals are respectively input into the two initial detection models, so that the two initial detection models respectively perform signal detection on the plurality of welding sample optical signals based on the welding detection reference range, and the two initial detection models respectively output welding sample detection results of the plurality of welding sample optical signals; comparing the welding sample detection result with welding sample results corresponding to the welding sample optical signals, and obtaining accurate values corresponding to the two initial detection models based on the comparison result; and when the accurate value is determined to meet a preset threshold value, selecting a plurality of negative sample welding sample optical signals based on the welding sample result.

In an embodiment, when the initial detection model is subjected to super-parameter optimization based on the plurality of negative sample welding sample optical signals to obtain an optimal detection model, a model selection controller and a super-parameter controller corresponding to each initial detection model are firstly set, wherein the model selection controller is used for selecting a target initial detection model for the plurality of initial detection models, and the super-parameter controller is used for performing super-parameter optimization for the target initial detection model.

In one embodiment, the model selection controller is respectively connected with the super parameter controller corresponding to each initial detection model; selecting a controller and a plurality of super parameter controllers to form an intelligent agent based on the model; as shown in fig. 21, fig. 21 is a schematic view of the internal structure of the intelligent agent.

In one embodiment, the process of selecting the target initial detection model based on the model selection controller and the process of performing the super-parameter optimization on the target initial detection model based on the super-parameter controller are modeled as a markov decision process.

Specifically, the problems of initial detection model selection by the model selection controller and super parameter optimization by the super parameter controller on the initial detection model are expressed as a Markov decision process, and 5-tuple is used Representing a reinforcement learning process.

Wherein,is the set of all active actions, +.>Representing the selected algorithm ∈>Representing the hyper-parameters corresponding to the selected algorithm +.>m is the number of super parameters of the initial detection model; the super parameter is the distribution ++given by the super parameter controller of the intelligent agent according to the initial detection model characteristics> And (5) sampling.

Wherein,for a set of active states, the agent selects the current hyper-parameters based on previous decisions, so there are:

wherein,is a reward function; because the agent is updated with the highest accuracy given by the training as the reward signal, the undiminished rewards will only be obtained after the final decision; thus, when t.epsilon.0, m) is ∈0, -, a->Whileaccuracy represents the selected algorithm +.>And selected superparameter->The highest accuracy obtained is trained on the data.

Wherein,for state transition probability functions, for model selectionAnd the hyper-parameter optimization problem are unknown.

Where γ is the discount factor, where γ=1.

In one embodiment, for the model selection controller decision process: the model selection controller comprises a first input embedded layer, a first decision core layer and a first output embedded layer, wherein the first input embedded layer consists of a plurality of perception layers, the first decision core layer consists of three long and short time memory networks, and the first output embedded layer consists of a plurality of perception layers; the model selection controller further comprises a fully-connected network layer, and the dimension output by the fully-connected network layer is the number N of models; the initial input is a standard normal distribution As shown in fig. 22, fig. 22 is a schematic structural view of the model selection controller.

In one embodiment, for the super parameter controller decision process: the super-parameter controller comprises a second input embedded layer, a second decision core layer and a second output embedded layer, wherein the second input embedded layer consists of a plurality of perception layers, the second decision core layer consists of three long and short time memory networks, and the second output embedded layer consists of a plurality of perception layers; the initial input of the parameter optimization control is the output of the model selection controller, and the output at the moment t is the normal distribution of the actionWhich is used as input to the next time controller, i.e. +.>As shown in fig. 23, fig. 23 is a schematic diagram of the super parameter controller structure and decision process.

In one embodiment, for the hyper-parametric sampling process: will be through hyperbolic tangent function tanh Mean>Scaled to a range (-1, 1), the scaled value is +.>From the new distribution->Obtain intermediate value h ⁱ The method comprises the steps of carrying out a first treatment on the surface of the Will be the intermediate value h ⁱ Conversion to the range corresponding to the hyper-parameters +.>

In the method, in the process of the invention,for the lower bound of the super-parametric search range, +.>For the upper bound of the super-parametric search range, +.>Is the lower bound of the super parameter searching range; the clip_and_cover function ensures that the sample value is within the model hyper-parameter value range.

In an embodiment, based on the markov decision process, the target initial detection model selected by the model selection controller is automatically adjusted, and super parameters of the target initial detection model are automatically adjusted, and all selected target initial detection models and super parameters corresponding to the target initial detection models are recorded.

Specifically, for training of an agent: for a decision trajectory: decision a of model selection controller ₀ And each quiltDecision process for selection model:the jackpot function is: />

Training of the agent is to obtain the maximum rewards by continuously updating the internal parameters theta, namely:

updating the parameters of the intelligent agent by a gradient ascent method:

alpha is the gradient update learning rate,the method comprises the following steps:

in an embodiment, based on the plurality of negative welding light signals, all the selected target initial detection models are respectively subjected to model training so as to enable the super parameters corresponding to the target initial detection models to obtain optimal detection models according to model training results.

Preferably, after the initial optimization, if detection errors are found again, such as false alarms or missing alarms; based on the online learning function provided by the software, marking the signal with errors, and optimizing the model selection and parameters again; based on the training of the algorithm, an optimal detection model is quickly established, the accuracy of the detection algorithm is accelerated, the omission ratio and the false alarm rate are reduced, and an effective data basis is provided for laser welding processes such as battery welding and the like.

In an embodiment, a polarizer is disposed at a light outlet of the laser welding device, so that a welding beam emitted by the laser welding device is reflected to a power detection device based on the polarizer, so that the power detection device obtains a laser light-emitting power signal of the welding beam.

In one embodiment, the power detection device is connected to the welding detection light path based on an optical fiber.

In an embodiment, based on the laser output power signal, device stability evaluation is performed on the laser welding device, so as to obtain a laser welding device stability result.

Specifically, based on the light emitting area, dividing the laser light emitting power signal into a plurality of area laser light emitting power signals, respectively calculating area light emitting power corresponding to each area laser light emitting power signal, counting power variance of the area light emitting power corresponding to each area laser light emitting power signal in a preset time period, judging whether the power variance is larger than a preset power variance threshold, if yes, outputting a stability result of the laser welding equipment to be maintained, otherwise, outputting a stability result of the laser welding equipment to be that the laser welding equipment is not required to be maintained.

In one embodiment, the area light output power corresponding to each area laser light output power signal is calculated as follows:

wherein y is _i Power value t of light-emitting area ₁ To start time of light-emitting area, t ₂ For the end time of the light-emitting region, y _t Is the signal value of the signal at time t.

In an embodiment, the power variance of the area light output power corresponding to each area laser light output power signal in a preset time period is counted, wherein the preset time period is 24 hours, and a statistical formula is shown as follows:

in sigma ² And if the variance value is larger than the set specified value, judging that the equipment needs to be overhauled.

In one embodiment, based on the training of the algorithm, an optimal detection model is quickly established, the accuracy of the detection algorithm is accelerated, the omission rate and the false alarm rate are reduced, and an effective data basis is provided for the laser welding process such as battery welding.

In summary, according to the welding light signal processing method based on the welding detection light path, the welding detection reference range is established through the positive welding light signals concentrated by the historical welding light signals and is used as the reference range of the normal welding light signals, so that the detection efficiency of abnormal welding light signals in the welding process can be improved; and the model selection and the super-parameter optimization are carried out on the initial detection model based on reinforcement learning to obtain an optimal detection model, manual intervention is not needed in the process, the influence of human factors on an optimization result can be reduced, the stability and the reliability of optimization are improved, the most suitable detection model can be selected in a personalized manner according to the requirements of specific scenes, the super-parameter is adjusted, so that better detection performance is obtained, in the optimization process, the performance of the detection model can be improved by adopting an iterative mode in the optimization process, the model selection and the super-parameter are continuously improved through interaction with the environment, the performance of the detection model is adapted to the continuously changing welding quality detection requirements, the detection model is enabled to be more accurate and robust, the generalization capability is better, the welding light signal is input into the pre-trained optimal detection model, and the accuracy of the output welding detection result is improved when the optimal detection model detects the welding light signal based on the welding detection reference range, and meanwhile the detection efficiency is improved.

Embodiment 3, referring to fig. 3, fig. 3 is a schematic structural diagram of an embodiment of a welding beam adjusting device based on a welding detection light path, as shown in fig. 3, where the device includes a beam calibration module 301, a beam splitting module 302, and a light attenuation adjusting module 303, and specifically includes the following steps:

in one embodiment, the welding detection light path includes a beam splitter, an electrically controlled calibrator, and an electrically controlled attenuator.

In an embodiment, the welding beam adjusting device is connected to the welding detection light path, where the beam calibration module 301 is connected to the electronic control calibrator, the beam splitter module 302 is connected to the beam splitter, and the optical attenuation adjusting module 303 is connected to the electronic control attenuator.

In an embodiment, the beam calibration module 301 is configured to obtain a welding focus on a welding working surface, where a welding beam emitted by a laser welding device falls on the welding working surface, and calibrate, based on the electronically controlled calibrator, a calibration beam emitted by a laser generator, so that the calibration focus on the welding working surface where the calibration beam falls coincides with the welding focus.

In an embodiment, the beam splitting module 302 is configured to reflect, after determining that the calibration focal point coincides with the welding focal point, a welding beam emitted by the laser welding device to the welding detection optical path, so that the welding beam is transmitted to the beam splitter based on the electronic control calibrator, and the welding beam is split based on the beam splitter to obtain a plurality of welding beam splitters with different wavebands.

In an embodiment, the optical attenuation adjustment module 303 is configured to transmit each welding beam to the corresponding electrically controlled attenuator through an optical fiber, so that the electrically controlled attenuator performs optical attenuation adjustment on the welding beam to obtain a welding beam.

In one embodiment, the beam calibration module 301 includes a calibration beam transmission sub-module 3011, an image processing sub-module 3012, and a focus coincidence sub-module 3013.

In an embodiment, the calibration beam transmission submodule 3011 is configured to transmit a calibration beam emitted by the laser generator into a welding detection light path, so that an electrically controlled calibrator in the welding detection light path retroreflects the calibration beam onto the welding working surface.

In an embodiment, the image processing submodule 3012 is configured to obtain a first image of the welding working surface, and perform graying processing on the first image to obtain a first gray-scale image.

In an embodiment, the focus coincidence submodule 3013 is configured to determine, according to the first gray scale image, whether a calibration focus of the calibration beam falling on the welding working surface coincides with the calibration focus, if yes, consider that the electronic control calibrator finishes calibrating the calibration beam, otherwise, adjust the electronic control calibrator to change a position of the calibration focus on the welding working surface until it is determined that the calibration focus coincides with the welding focus.

In an embodiment, the focus coincidence sub-module 3013 includes a coordinate position acquisition unit 30131 and an electronically controlled calibrator adjustment unit 30132.

In an embodiment, the coordinate position obtaining unit 30131 is configured to set a welding surface coordinate axis with the welding focus on the welding surface as an origin, and obtain a first coordinate position of the calibration focus on the welding surface coordinate axis.

In an embodiment, the electronic control calibrator adjusting unit 30132 is configured to obtain a first x-axis coordinate variable and a first y-axis coordinate variable based on the first coordinate position, and adjust the electronic control calibrator based on the first x-axis coordinate variable and the first y-axis coordinate variable.

In one embodiment, the electronic control calibrator comprises a reflecting lens, a supporting rod, an origin lens linking universal joint and a base; the support rod is arranged above the base, the support rod is perpendicular to the base, the first end of the origin lens linking universal joint is connected with the support rod, and the second end of the origin lens linking universal joint is connected with the reflecting lens.

In one embodiment, the electronically controlled calibrator adjustment unit 30132 includes a mirror plate coordinate axis setting subunit 301321, a coordinate position transformation subunit 301322, a coordinate variable acquisition subunit 301323, and an origin mirror link gimbal movement subunit 301324.

In one embodiment, the mirror plate coordinate axis setting subunit 301321 is configured to set a mirror plate coordinate axis with a center point of the mirror plate as a reference point.

In an embodiment, the coordinate position transforming subunit 301322 is configured to obtain, based on the y-axis coordinate point and the x-axis coordinate point, a second coordinate position of the first coordinate position on the coordinate axis of the reflector with the first x-axis coordinate variable as a y-axis coordinate point and the first y-axis coordinate variable as an x-axis coordinate point.

In an embodiment, the coordinate variable obtaining subunit 301323 is configured to obtain a second x-axis coordinate variable and a second y-axis coordinate variable based on the second coordinate position.

In an embodiment, the origin lens-linked gimbal moving subunit 301324 is configured to adjust, based on the second x-axis coordinate variable, movement of the mirror plate on the x-axis of the mirror plate coordinate axis based on the origin lens-linked gimbal, and adjust, based on the second y-axis coordinate variable, movement of the mirror plate on the y-axis of the mirror plate coordinate axis based on the origin lens-linked gimbal so that the reference point coincides with the second coordinate position.

In one embodiment, the electric control attenuator comprises a stepping motor, a rotating disc, a motor output shaft and a light outlet; the stepping motor is connected with the motor output shaft, and the motor output shaft is controlled to rotate based on the stepping motor; the rotating disc is fixed on the motor output shaft, and the rotation of the rotating disc is controlled based on the motor output shaft; the rotary disk is provided with a plurality of attenuation sheets, and when the rotary disk rotates, the attenuation sheets sequentially correspond to the light outlet.

In one embodiment, the optical attenuation adjustment module 303 includes a target attenuation sheet selection sub-module 3031, a rotating disk rotation sub-module 3032, and an optical attenuation sub-module 3033.

In an embodiment, the target attenuation sheet selecting sub-module 3031 is configured to obtain a preset light attenuation power, and select the target attenuation sheet on the rotating disc based on the preset light attenuation power.

In an embodiment, the rotary disk rotation sub-module 3032 is configured to control, based on the motor output shaft, the rotary disk to rotate after determining that the state of the stepper motor is a start state, so as to rotate the target attenuation sheet on the rotary disk to the light outlet.

In an embodiment, the optical attenuation submodule 3033 is configured to transmit the welding beam to the target attenuation sheet, so that the target attenuation sheet performs optical attenuation on the welding beam to obtain a welding beam, and output the welding beam based on the light outlet.

In one embodiment, the rotating disc rotating sub-module 3032 includes a rotation angle setting unit 30321 and a target damping plate rotating unit 303322.

In an embodiment, the rotation angle setting unit 30321 is configured to obtain a rotation direction of the output shaft of the motor, and set, based on the rotation direction, a rotation angle required by each attenuation piece on the rotating disc to rotate to the light outlet.

In an embodiment, the target attenuator rotating unit 30322 is configured to obtain a target rotation angle corresponding to the target attenuator after selecting the target attenuator on the rotating disc, and control the rotating disc to rotate based on the rotation direction and the target rotation angle, so as to rotate the target attenuator on the rotating disc to the light outlet.

In one embodiment, the beam splitting module 302 includes a first beam splitting processing sub-module 3021 and a second beam splitting processing sub-module 3022.

In one embodiment, the beam splitter includes a first beam splitter and a second beam splitter;

in an embodiment, the first beam splitting sub-module 3021 is configured to transmit the welding beam to the first beam splitting module, and split the welding beam based on the first beam splitting module so as to split the welding beam into a first band welding beam and a first welding beam.

In an embodiment, the second beam splitting sub-module 3022 is configured to transmit the first welding beam to the second beam splitter, and split the first welding beam based on the second beam splitter, so as to split the first welding beam into a second band welding beam splitter and a third band welding beam splitter.

It will be clear to those skilled in the art that, for convenience and brevity of description, reference may be made to the corresponding process in the foregoing method embodiment for the specific working process of the above-described apparatus, which is not described in detail herein.

It should be noted that the embodiments of the welding beam adjusting device based on the welding detection light path described above are only illustrative, where the modules described as separate components may or may not be physically separated, and the components displayed as the modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

On the basis of the embodiment of the welding beam adjustment method based on the welding detection light path, another embodiment of the present invention provides a welding beam adjustment terminal device based on the welding detection light path, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, where the processor executes the computer program to implement the welding beam adjustment method based on the welding detection light path according to any one of the embodiments of the present invention.

Illustratively, in this embodiment the computer program may be partitioned into one or more modules, which are stored in the memory and executed by the processor to perform the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function for describing the execution of the computer program in the welding beam adjustment terminal device based on the welding detection light path.

The welding beam adjusting terminal equipment based on the welding detection light path can be computing equipment such as a desktop computer, a notebook computer, a palm computer and a cloud server. The welding beam adjustment terminal device based on the welding detection light path may include, but is not limited to, a processor, a memory.

The processor may be a central processing unit (Central Processing Unit, CPU), other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. The general purpose processor may be a microprocessor or the processor may be any conventional processor or the like, which is a control center of the welding beam adjustment terminal device based on the welding detection light path, and connects various parts of the whole welding beam adjustment terminal device based on the welding detection light path by using various interfaces and lines.

The memory may be used to store the computer program and/or the module, and the processor may implement various functions of the welding beam adjustment terminal device based on the welding detection light path by running or executing the computer program and/or the module stored in the memory and invoking data stored in the memory. The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the cellular phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart Media Card (SMC), secure Digital (SD) Card, flash Card (Flash Card), at least one disk storage device, flash memory device, or other volatile solid-state storage device.

On the basis of the embodiment of the welding light beam adjusting method based on the welding detection light path, another embodiment of the invention provides a storage medium, which comprises a stored computer program, wherein when the computer program runs, equipment where the storage medium is controlled to execute the welding light beam adjusting method based on the welding detection light path according to any embodiment of the invention.

In this embodiment, the storage medium is a computer-readable storage medium, and the computer program includes computer program code, where the computer program code may be in a source code form, an object code form, an executable file, or some intermediate form, and so on. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth. It should be noted that the computer readable medium contains content that can be appropriately scaled according to the requirements of jurisdictions in which such content is subject to legislation and patent practice, such as in certain jurisdictions in which such content is subject to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.

Embodiment 4, referring to fig. 4, fig. 4 is a schematic structural diagram of an embodiment of a welding optical signal processing apparatus based on a welding detection optical path according to the present invention, and as shown in fig. 4, the apparatus includes a welding detection reference range generating module 401, a welding optical signal obtaining module 402, and a welding optical signal detecting module 403, specifically as follows:

the welding detection reference range generating module 401 is configured to acquire a historical welding light signal set, collect a plurality of positive welding light signals from the historical welding light signal sets, and generate a welding detection reference range when determining that the positive welding light signals meet a positive judgment standard.

The welding light signal obtaining module 402 is configured to collect, based on a PD sensor, a welding light attenuation beam obtained in the welding light beam adjusting device, to obtain a welding light signal; wherein the welding light attenuated beam is obtained by a welding light beam adjustment apparatus based on a welding detection light path according to any one of claims 20-28.

The welding light signal detection module 403 is configured to input the welding light signal into a pre-trained optimal detection model, so that the optimal detection model detects the welding light signal based on a welding detection reference range, and a welding detection result is obtained.

In one embodiment, the welding detection reference range generating module 401 includes a scanning window setting submodule 4011, an abnormal welding light signal confirming submodule 4012 and a welding light signal screening submodule 4013;

in an embodiment, the scanning window setting submodule 4011 is configured to collect a plurality of positive welding light signals from the historical welding light signal set, and set a scanning window based on a signal length of each positive welding light signal.

In an embodiment, the abnormal welding light signal confirmation submodule 4012 is configured to perform window scanning on each positive welding light signal based on the scanning window, and determine that the positive welding light signal is the abnormal welding light signal when it is determined that the variance of the distribution of the positive welding light signal in the scanning window is greater than a preset variance.

In an embodiment, the welding light signal screening submodule 4013 is configured to delete the abnormal welding light signals in the plurality of positive welding light signals to obtain remaining positive welding light signals, and use the remaining positive welding light signals as positive welding light signals that meet a positive judgment standard.

In an embodiment, the welding detection reference range generating module 401 further includes a light emitting region determining submodule 4014, a first moment positive welding light signal obtaining submodule 4015, a region positive welding light signal distribution range determining submodule 4016, and a welding detection reference range determining submodule 4017;

In an embodiment, the light-emitting area determining submodule 4014 is configured to determine light-emitting areas of the welding light signals based on the plurality of positive welding light signals, and divide the plurality of positive welding light signals into a plurality of area positive welding light signals based on the light-emitting areas, respectively.

In an embodiment, the first moment positive welding light signal obtaining submodule 4015 is configured to classify the positive welding light signals of the plurality of regions based on the light emitting regions, obtain all positive welding light signals of the first regions corresponding to each light emitting region, and obtain positive welding light signals of the first moment corresponding to the first welding moment of all positive welding light signals of the first regions.

In an embodiment, the determining submodule 4016 is configured to perform data fitting processing on all positive welding light signals at the first moment to obtain a positive welding light signal distribution range corresponding to the first welding moment, integrate the positive welding light signal distribution ranges corresponding to all welding moments in the positive welding light signals at the first region, and obtain a positive welding light signal distribution range corresponding to the positive welding light signals at the first region.

In an embodiment, the welding detection reference range determining submodule 4017 is configured to integrate the area positive welding light signal distribution range corresponding to each light emitting area to obtain a welding detection reference range.

In one embodiment, the area positive welding light signal distribution range determining submodule 4016 includes a first positive welding light signal distribution range generating unit 40161 and a second positive welding light signal distribution range generating unit 40162;

in an embodiment, the first normal welding light signal distribution range generating unit 40161 is configured to, when detecting that a user selects a first welding light signal distribution range generating mode from a plurality of preset welding light signal distribution range generating modes, perform data fitting processing on all the first moment normal welding light signals based on a hypothetical normal distribution function to obtain a hypothetical normal distribution value, and select an area of the hypothetical normal distribution value within a preset standard deviation range as the normal welding light signal distribution range corresponding to the first welding moment.

In an embodiment, the second positive welding light signal distribution range generating unit 40162 is configured to, when detecting that a user selects a second welding light signal distribution range generating manner from a plurality of preset welding light signal distribution range generating manners, arrange all positive welding light signals at first moments in a sequence from big to small, screen all positive welding light signals at first moments based on the arrangement sequence, obtain first-moment screening positive welding light signals, select a minimum signal intensity value and a maximum signal intensity value in all first-moment screening positive welding light signals, and determine a positive welding light signal distribution range corresponding to the first welding moment based on the minimum signal intensity value and the maximum signal intensity value.

The embodiment provides a welding light signal processing device based on welding detection light path, still includes: the optimal detection model generating module 404 is shown in fig. 24, and fig. 24 is a schematic diagram of still another structure of the welding light signal processing device based on the welding detection light path.

In an embodiment, the optimal detection model generating module 404 is configured to obtain a plurality of welding sample optical signals, input the plurality of welding sample optical signals into an initial detection model, so that the initial detection model performs signal detection on the plurality of welding sample optical signals based on the welding detection reference range, and perform super-parameter optimization processing on the initial detection model based on the plurality of negative welding sample optical signals after determining that a plurality of negative welding sample optical signals are detected, to obtain an optimal detection model.

In an embodiment, the optimal detection model generating module 404 includes a welding sample optical signal acquiring sub-module 4041, an initial detection model setting sub-module 4042, an initial detection model detecting sub-module 4043, a welding sample result comparing sub-module 4044, and a negative sample welding sample optical signal selecting sub-module 4045;

in one embodiment, the welding sample optical signal obtaining submodule 4041 is configured to obtain a plurality of welding sample optical signals.

In an embodiment, the initial detection model setting submodule 4042 is configured to set a plurality of initial detection models, where the plurality of initial detection models includes a variance feature detection model, an average offset detection model, an area detection model, and a time detection model.

In an embodiment, the initial detection model detection submodule 4043 is configured to arbitrarily select two initial detection models, input the plurality of welding sample optical signals into the two initial detection models respectively, so that the two initial detection models perform signal detection on the plurality of welding sample optical signals based on the welding detection reference range respectively, and enable the two initial detection models to output welding sample detection results of the plurality of welding sample optical signals respectively.

In an embodiment, the welding sample result comparison submodule 4044 is configured to compare the welding sample detection result with welding sample results corresponding to the plurality of welding sample optical signals, and obtain accurate values corresponding to the two initial detection models based on the comparison result.

In an embodiment, the negative sample welding sample optical signal selecting submodule 4045 is configured to select a plurality of negative sample welding sample optical signals based on the welding sample result when it is determined that the accurate value meets the preset threshold.

In one embodiment, the optimal detection model generation module 404 includes a controller setup submodule 4046, a markov decision modeling submodule 4047, a model parameter adjustment submodule 4048, and a model training submodule 4049.

In an embodiment, the controller setting submodule 4046 is configured to set a model selection controller and a super parameter controller corresponding to each initial detection model, where the model selection controller is configured to perform target initial detection model selection on the multiple initial detection models, and the super parameter controller is configured to perform super parameter optimization on the target initial detection models.

In one embodiment, the markov decision modeling submodule 4047 is configured to model a process of selecting the target initial detection model based on the model selection controller and a process of performing super-parameter optimization on the target initial detection model based on the super-parameter controller as a markov decision process.

In an embodiment, the model parameter adjustment submodule 4048 is configured to automatically adjust, based on the markov decision process, the target initial detection model selected by the model selection controller, automatically adjust the super parameters of the target initial detection model, and record all selected target initial detection models and the super parameters corresponding to the target initial detection models.

In an embodiment, the model training submodule 4049 is configured to respectively perform model training on all the selected target initial detection models based on the plurality of negative welding light signals, so that the super parameters corresponding to the target initial detection models perform model training, and obtain an optimal detection model according to a model training result.

In one embodiment, the model selection controller is respectively connected with the super parameter controller corresponding to each initial detection model.

In an embodiment, the model selection controller includes a first input embedding layer, a first decision core layer and a first output embedding layer, wherein the first input embedding layer is composed of a plurality of sensing layers, the first decision core layer is composed of three long and short time memory networks, and the first output embedding layer is composed of a plurality of sensing layers.

In an embodiment, the super parameter controller includes a second input embedded layer, a second decision core layer and a second output embedded layer, where the second input embedded layer is composed of multiple sensing layers, the second decision core layer is composed of three long and short time memory networks, and the second output embedded layer is composed of multiple sensing layers.

The welding light signal processing device based on the welding detection light path provided in this embodiment further includes: as shown in fig. 25, fig. 25 is a schematic diagram of still another structure of the welding light signal processing device based on the welding detection light path.

In an embodiment, the laser light emitting power signal obtaining module 405 is configured to reflect a welding beam emitted by a laser welding device to a power detection device, so that the power detection device obtains a laser light emitting power signal of the welding beam.

In an embodiment, the laser welding device stability result obtaining module 406 is configured to perform device stability evaluation on the laser welding device based on the laser output power signal, to obtain a laser welding device stability result.

In one embodiment, the laser welding apparatus stability result obtaining module 406 includes a region light output power calculating submodule 4061 and a laser welding apparatus stability result outputting submodule 4062.

In an embodiment, the area light output power calculating submodule 4061 is configured to divide the laser light output power signal into a plurality of area laser light output power signals based on the light output area, and calculate the area light output power corresponding to each area laser light output power signal respectively.

In an embodiment, the laser welding device stability result output submodule 4062 is configured to count a power variance of the area light output power corresponding to the laser light output power signal of each area in a preset time period, determine whether the power variance is greater than a preset power variance threshold, if yes, output a laser welding device stability result to indicate that the laser welding device needs to be maintained, otherwise, output a laser welding device stability result to indicate that the laser welding device does not need to be maintained.

It should be noted that the above embodiment of the welding optical signal processing device based on the welding detection optical path is merely illustrative, where the modules described as separate components may or may not be physically separated, and components displayed as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

On the basis of the embodiment of the welding light signal processing method based on the welding detection light path, another embodiment of the invention provides a welding light signal processing terminal device based on the welding detection light path, which comprises a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to realize the welding light signal processing method based on the welding detection light path according to any embodiment of the invention.

Illustratively, in this embodiment the computer program may be partitioned into one or more modules, which are stored in the memory and executed by the processor to perform the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function for describing the execution of the computer program in the welding light signal processing terminal device based on the welding detection light path.

The welding light signal processing terminal equipment based on the welding detection light path can be computing equipment such as a desktop computer, a notebook computer, a palm computer and a cloud server. The welding light signal processing terminal device based on the welding detection light path can include, but is not limited to, a processor and a memory.

The processor may be a central processing unit (Central Processing Unit, CPU), other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. The general processor may be a microprocessor or the processor may be any conventional processor, etc., and the processor is a control center of the welding light signal processing terminal device based on the welding detection light path, and connects various parts of the whole welding light signal processing terminal device based on the welding detection light path by using various interfaces and lines.

The memory may be used to store the computer program and/or the module, and the processor may implement various functions of the welding light signal processing terminal device based on the welding detection light path by running or executing the computer program and/or the module stored in the memory and calling the data stored in the memory. The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to the use of the cellular phone, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart Media Card (SMC), secure Digital (SD) Card, flash Card (Flash Card), at least one disk storage device, flash memory device, or other volatile solid-state storage device.

On the basis of the embodiment of the welding light signal processing method based on the welding detection light path, another embodiment of the invention provides a storage medium, which comprises a stored computer program, wherein when the computer program runs, equipment where the storage medium is controlled to execute the welding light signal processing method based on the welding detection light path according to any embodiment of the invention.

In summary, the method and the device for adjusting and processing the light beam based on the welding detection light path, provided by the invention, are used for obtaining the welding focus of the welding light beam emitted by the laser welding equipment on the welding working surface, and calibrating the calibration light beam based on the electric control calibrator, so that the calibration focus of the calibration light beam on the welding working surface coincides with the welding focus; after the alignment focus is determined to coincide with the welding focus, reflecting welding light beams emitted by the laser welding equipment into a welding detection light path, so that the welding light beams are transmitted into a beam splitter based on an electric control aligner, and the welding light beams are subjected to beam splitting treatment to obtain a plurality of welding beam splitting light beams with different wave bands; transmitting each welding beam to a corresponding electric control attenuator through an optical fiber, and carrying out optical attenuation adjustment on the welding beam to obtain a welding beam; compared with the prior art, the technical scheme of the invention realizes automatic alignment of the welding beam and automatic control of the light intensity of the welding beam.

The foregoing is merely a preferred embodiment of the present invention, and it should be noted that modifications and substitutions can be made by those skilled in the art without departing from the technical principles of the present invention, and these modifications and substitutions should also be considered as being within the scope of the present invention.

Claims

1. The welding light beam adjusting method based on the welding detection light path is characterized in that the welding detection light path comprises a beam splitter, an electric control calibrator and an electric control attenuator, and the welding light beam adjusting method comprises the following steps:

acquiring a welding focus of a welding beam emitted by laser welding equipment on a welding working surface, and emitting a calibration beam emitted by a laser generator into a welding detection light path so that an electric control calibrator in the welding detection light path reversely reflects the calibration beam onto the welding working surface; acquiring a first image of the welding working surface, and carrying out graying treatment on the first image to obtain a first gray image; judging whether a calibration focus of the calibration light beam falling on the welding working surface coincides with the welding focus or not according to the first gray level image, if so, considering that the electric control calibrator finishes calibrating the calibration light beam, otherwise, adjusting the electric control calibrator to change the position of the calibration focus on the welding working surface until the calibration focus coincides with the welding focus;

2. The welding beam adjustment method based on a welding detection light path of claim 1, wherein adjusting the electronically controlled calibrator comprises:

3. The welding beam adjustment method based on a welding detection light path according to claim 2, wherein the electrically controlled calibrator comprises a reflecting mirror, a supporting rod, an origin mirror link universal joint and a base;

4. A welding beam adjustment method based on a welding detection light path as defined in claim 3, wherein adjusting said electrically controlled calibrator based on said first x-axis coordinate variable and said first y-axis coordinate variable comprises;

obtaining a second x-axis coordinate variable and a second y-axis coordinate variable based on the second coordinate position;

5. The welding light beam adjusting method based on a welding detection light path according to claim 1, wherein the electric control attenuator comprises a stepping motor, a rotating disk, a motor output shaft and a light outlet;

6. The welding light beam adjustment method based on a welding detection light path according to claim 5, wherein the electrically controlled attenuator performs light attenuation adjustment on the welding split light beam to obtain a welding light attenuation light beam, and specifically comprises:

7. The welding beam adjusting method based on a welding detection light path according to claim 6, wherein the rotating disk is controlled to rotate based on the motor output shaft so as to rotate a target attenuation piece on the rotating disk to the light outlet, specifically comprising:

8. The welding beam adjustment method based on a welding detection light path according to claim 1, wherein the welding beam is transmitted to the beam splitter, so that the beam splitter splits the welding beam to obtain a plurality of welding beam splitters with different wave bands, and the method specifically comprises:

the beam splitter comprises a first beam splitter and a second beam splitter;

9. A welding light signal processing method based on a welding detection light path is characterized by comprising the following steps:

Acquiring a welding light attenuation beam based on a PD sensor to obtain a welding light signal; wherein the welding light attenuated beam is obtained according to the welding light beam adjustment method based on a welding detection light path of any one of claims 1 to 8;

10. The welding light signal processing method based on a welding detection light path according to claim 9, wherein a plurality of positive welding light signals are collected from the plurality of historical welding light signals, and the determining that the plurality of positive welding light signals meet a positive judgment criterion specifically comprises:

11. The welding light signal processing method based on a welding detection light path according to claim 9, wherein generating the welding detection reference range specifically comprises:

12. The welding light signal processing method based on the welding detection light path according to claim 11, wherein the data fitting processing is performed on all positive welding light signals at the first moment to obtain a positive welding light signal distribution range corresponding to the first welding moment, specifically comprising:

13. The welding light signal processing method based on the welding detection light path according to claim 9, wherein the generating process of the optimal detection model specifically comprises:

14. The welding light signal processing method based on a welding detection light path according to claim 13, wherein a plurality of welding sample light signals are acquired, and the plurality of welding sample light signals are input into an initial detection model, so that the initial detection model performs signal detection on the plurality of welding sample light signals based on the welding detection reference range, specifically comprising:

acquiring a plurality of welding sample optical signals;

15. The welding light signal processing method based on a welding detection light path according to claim 14, wherein the performing the super-parametric optimization processing on the initial detection model based on the plurality of negative sample welding sample light signals to obtain an optimal detection model specifically comprises:

and respectively carrying out model training on all the selected target initial detection models and the super parameters corresponding to the target initial detection models based on the plurality of negative sample welding light signals, and obtaining an optimal detection model according to model training results.

16. The welding light signal processing method based on a welding detection light path as defined in claim 15, wherein the setting of the model selection controller and the super parameter controller corresponding to each initial detection model specifically comprises:

17. The welding light signal processing method based on a welding detection light path of claim 11, further comprising:

18. The welding light signal processing method based on the welding detection light path according to claim 17, wherein based on the laser output power signal, performing equipment stability evaluation on the laser welding equipment to obtain a laser welding equipment stability result, specifically comprising:

19. Welding light beam adjusting device based on welding detection light path, its characterized in that, welding detection light path includes beam splitter, automatically controlled calibrator and automatically controlled attenuator, welding light beam adjusting device includes: the device comprises a light beam calibration module, a light beam splitting module and a light attenuation adjustment module;

20. The welding beam adjustment device based on a welding detection light path of claim 19, wherein the beam alignment module comprises an alignment beam transmission sub-module, an image processing sub-module, and a focus coincidence sub-module;

21. The welding beam adjusting device based on a welding detection light path according to claim 20, wherein the focus coincidence submodule includes a coordinate position acquisition unit and an electric control calibrator adjustment unit;

22. The welding beam adjusting device based on a welding detection light path of claim 21, wherein the electrically controlled calibrator comprises a mirror plate, a support bar, an origin mirror plate link gimbal, and a base;

23. The welding beam adjusting device based on a welding detection light path as defined in claim 22, wherein the electrically controlled calibrator adjusting unit comprises a mirror plate coordinate axis setting subunit, a coordinate position conversion subunit, a coordinate variable acquisition subunit, and an origin mirror plate link gimbal moving subunit;

24. The welding beam adjusting device based on a welding detection light path of claim 19, wherein the electrically controlled attenuator comprises a stepper motor, a rotating disk, a motor output shaft, and a light outlet;

25. The welding light beam regulating device based on a welding detection light path of claim 24, wherein the light attenuation regulating module comprises a target attenuation sheet selection sub-module, a rotating disk rotation sub-module and a light attenuation sub-module;

26. The welding beam adjustment device based on a welding detection light path of claim 25, wherein the rotating disk rotor module comprises a rotation angle setting unit and a target attenuation sheet rotation unit;

27. The welding beam adjustment device based on a welding detection light path of claim 19, wherein the beam splitting module comprises a first splitting processing sub-module and a second splitting processing sub-module;

28. A welding light signal processing device based on a welding detection light path, the welding light signal processing device comprising: the welding detection reference range generation module, the welding optical signal acquisition module and the welding optical signal detection module;

the welding light signal acquisition module is used for acquiring a welding light attenuation beam obtained in the welding light beam adjusting device based on a PD sensor to obtain a welding light signal; wherein the welding light attenuated beam is obtained by a welding light beam adjustment apparatus based on a welding detection light path according to any one of claims 19-27;

29. The welding light signal processing device based on a welding detection light path as defined in claim 28, wherein the welding detection reference range generation module comprises a scanning window setting sub-module, an abnormal welding light signal confirmation sub-module and a welding light signal screening sub-module;

30. The welding light signal processing device based on welding detection light path of claim 28, wherein said welding detection reference range generation module further comprises a light-out region determination sub-module, a first moment positive welding light signal acquisition sub-module, a region positive welding light signal distribution range determination sub-module, and a welding detection reference range determination sub-module;

31. The welding light signal processing device based on the welding detection light path as recited in claim 30, wherein said regional positive welding light signal distribution range determination submodule includes a first positive welding light signal distribution range generation unit and a second positive welding light signal distribution range generation unit;

32. The welding light signal processing device based on a welding detection light path of claim 28, further comprising: an optimal detection model generation module;

33. The welding light signal processing device based on the welding detection light path as defined in claim 32, wherein the optimal detection model generation module comprises a welding sample light signal acquisition sub-module, an initial detection model setting sub-module, an initial detection model detection sub-module, a welding sample result comparison sub-module and a negative sample welding sample light signal selection sub-module;

34. The welding light signal processing device based on a welding detection light path of claim 33, wherein the optimal detection model generation module comprises a controller setup sub-module, a markov decision-making sub-module, a model parameter adjustment sub-module, and a model training sub-module;

The model training submodule is used for respectively carrying out model training on all the selected target initial detection models and the super parameters corresponding to the target initial detection models based on the plurality of negative sample welding light signals, and obtaining an optimal detection model according to model training results.

35. The welding light signal processing device based on welding detection light path of claim 34, wherein the model selection controller is respectively connected with a super parameter controller corresponding to each initial detection model;

36. The welding light signal processing device based on a welding detection light path of claim 30, further comprising: the laser light-emitting power signal acquisition module and the laser welding equipment stability result acquisition module;

37. The welding light signal processing device based on the welding detection light path as defined in claim 36, wherein the laser welding equipment stability result obtaining module comprises a regional light output power calculating sub-module and a laser welding equipment stability result outputting sub-module;

38. A terminal device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, the processor implementing the welding light beam adjustment method based on a welding detection light path according to any one of claims 1 to 8 or the welding light signal processing method based on a welding detection light path according to any one of claims 9 to 18 when the computer program is executed.

39. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises a stored computer program, wherein the computer program, when run, controls a device in which the computer-readable storage medium is located to perform the welding light beam adjustment method based on a welding detection light path according to any one of claims 1 to 8, or the welding light signal processing method based on a welding detection light path according to any one of claims 9 to 18.