An online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition

Abstract

This invention relates to an online detection method for the stability of laser welding forming based on array sensing and artificial intelligence recognition. The invention pertains to online welding detection and intelligent control. It selects the thermally excited state signal inside the keyhole as the detection signal to acquire mesoscopic detection signals. These mesoscopic detection signals are preprocessed to obtain computer-recognizable data. Based on the preprocessed data, they are calibrated and divided into training, testing, and validation sets. A neural network model is established, and deep learning is used to train the recognition ability of signals from different regions within the keyhole for weld continuity, surface quality, and welding spatter amount until all three recognition results converge. The model hyperparameters are then adjusted using the validation set, and the accuracy and effectiveness of the prediction results are evaluated using test set data. Finally, the pre-trained optimal model is used to monitor the stability of laser welding forming online.

Description

Technical Field

[0001] The invention relates to the technical field of welding online detection and intelligent control, and is an online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition. Background Art

[0002] Laser and laser composite energy field welding is one of the mainstream technologies in the field of intelligent manufacturing. However, laser welding often has welding instability caused by fluctuations in non-process parameters in engineering applications. Studies have shown that the fluctuations in laser energy absorption by welding plume and keyhole spray and the imbalance in the vaporization process of metal materials are the root causes of laser welding instability. The instability of laser welding will directly lead to many post-welding forming quality problems, such as forming continuity, weld surface quality, spatter, etc. Therefore, reliable online detection of welding stability and closed-loop quality control are extremely important for intelligent manufacturing of laser and laser composite energy field welding.

[0003] Since the vaporization of metal materials is the main reason for the formation of keyholes, and the energy conversion between high-energy beam laser and welded parts during laser and laser composite energy field welding is also completed in the keyhole, the stability of the keyhole is directly related to the stability of laser welding. For example, ① the fluctuation of the inner wall of the keyhole will affect the continuity of weld formation, ② the fluctuation amplitude and frequency of the keyhole opening will have a more direct impact on the surface forming quality after welding, and ③ the abnormal fluctuation of the edge of the keyhole opening is a precursor to the formation of welding spatter. Therefore, the stability of the keyhole has a more direct correlation with the continuity of welding formation, surface quality, welding spatter, etc. At the same time, since the keyhole characteristic changes occur before the weld formation in the time sequence, it is also feasible to adjust the welding process parameters by real-time monitoring of the keyhole changes to achieve stable closed-loop control of the weld formation quality. However, since the signal changes in different areas of the keyhole are very different for identifying different stability characteristics, it is necessary to effectively separate the signals in different areas of the keyhole. The opening diameter of the keyhole itself is less than 1mm, so its internal signal characteristics belong to the mesoscopic field detection category. At the same time, since the types of defects and causes of formation are different, the analysis models of signals in different areas are also very different, so there are extremely high technical requirements for signal optical extraction methods and data analysis methods. However, due to the limitations of macro sampling methods and existing data analysis methods, it is difficult to obtain accurate and reliable welding stability information with existing detection methods, so it is extremely difficult to reliably detect welding stability online. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the present invention provides an online detection method for laser welding stability based on array sensing and artificial intelligence recognition. The present invention proposes an artificial intelligence online detection method that uses array sensing to collect thermally excited state mesoscopic detection signals in the characteristic area inside the keyhole, and establishes a neural network model through machine learning. The trained recognition model is then called to analyze signal characteristics online and extract welding stability information. The method can also have a good predictive recognition effect on weld formation continuity, surface quality, and welding spatter.

[0005] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0006] The present invention provides an online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition. The present invention provides the following technical solutions:

[0007] A welding detection device based on array sensing and artificial intelligence recognition, the device comprising: a housing, a substrate, a sensor, a narrow-band filter and a three-dimensional fine-tuning mechanism;

[0008] The connection between the substrate and the laser welding head serves to fix the entire device. There are mounting holes on the substrate to respectively fix the device housing, the three-dimensional fine-tuning mechanism and the narrow-band filter carrier. The three-dimensional fine-tuning mechanism is connected to the sensor to adjust the three-dimensional spatial position of the sensor sensing chip. The narrow-band filter is located on the optical path between the optical focusing lens group and the array sensor sensing surface. The optical real image after filtering can be projected onto the array sensor sensing surface.

[0009] An online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition, the method is based on a welding detection device based on array sensing and artificial intelligence recognition, and the method comprises the following steps:

[0010] Step 1: Select the thermally excited state signal inside the keyhole as the detection signal to obtain the mesoscopic detection signal;

[0011] Step 2: pre-process the mesoscopic detection signal to obtain computer-recognizable type data;

[0012] Step 3: According to the preprocessed data, the data is calibrated and divided into training set, test set and validation set;

[0013] Step 4: Establish a neural network model and train the recognition ability of signals in different areas of the keyhole for seam continuity, surface quality, and welding spatter through deep learning until all three recognition results converge, and then adjust the model hyperparameters through the validation set;

[0014] Step 5: Test the reliability of the model through the test set;

[0015] Step 6: Call the trained recognition model to analyze signal features online and extract welding stability information.

[0016] Preferably, the thermally excited state mesoscopic signal in the keyhole is selected as the detection object, and the clear real image of the thermally excited state signal inside the keyhole is projected onto the sensing surface of the array sensor by using the optical focusing imaging and spectral transmission principles to obtain the mesoscopic detection signal. The deep method in machine learning is used to automatically identify the characteristics of the mesoscopic signals in different areas, predict the continuity of weld formation, surface quality, and trend of welding spatter, and obtain the current keyhole stability information.

[0017] Preferably, the thermally excited state signal inside the keyhole is a near-infrared signal generated after the laser beam enters the base material, which causes the metal at the bottom of the keyhole to melt and evaporate rapidly through intense energy input, accompanied by high-density energy excitation. Since it is generated on the surface of the keyhole, the characteristic change of the signal is highly consistent with the dynamic behavior of the keyhole.

[0018] Preferably, the method for collecting mesoscopic detection signals is:

[0019] S1. A high-power optical focusing lens group with at least 0.6-1.5m shooting working distance and 10mm shooting depth of field is used to extract a clear real image of the characteristic area at the bottom of the keyhole from the coaxial optical path of the laser welding head. The sufficiently large depth of field can capture a clear real image of the characteristic area in the fluctuating state without changing the focal length;

[0020] S2. In the near-infrared spectrum, the welding plume, laser beam and other large amounts of welding radiation signals above the keyhole are effectively shielded by narrow-band filtering, so that the thermally excited state signals in the characteristic area can be effectively separated, thereby greatly reducing the proportion of invalid signals in the detection signal;

[0021] S3. Project the real image of the thermally excited state signal onto the sensing surface of an array sensor with a large sensing coverage area and high detection accuracy. The sensing area of ​​the array sensor should be ≥ the feature area to be measured, and the resolution accuracy should be ≤10μm, thereby obtaining mesoscopic signals at different positions in the feature area to be measured.

[0022] Preferably, the optimal model of welding stability is obtained as follows:

[0023] The analysis sample data collected by the array sensor or image sensor is converted into computer-recognizable data after certain data processing. Secondly, the analysis samples are calibrated with the actual weld formation continuity, surface quality, and welding spatter, and then divided into training set, test set, and verification set. Then, a neural network model is constructed by computer, and the maximum threshold that can identify different trend changes is trained respectively. Then, when the welding stability is monitored online, the increment of the current trend feature value of different identification objects and the trend feature value of the previous moment or the previous n moments is compared in real time to see whether it exceeds the threshold, and this is used as a basis to judge whether it is stable. Through deep learning, the computer is allowed to simultaneously train the stability identification basis of the signals of different regions in the keyhole for the continuity of seam formation, surface quality, and welding spatter, so that the maximum threshold of trend changes can be identified until all three identification results can converge, and then the model hyperparameters are adjusted through the verification set, and the reliability of the model is tested through the test set.

[0024] Preferably, the collected single-channel / multi-channel / or all mesoscopic detection signal data are converted into computer-recognizable type data after certain data processing, and the trained recognition model is called to compare the current trend characteristic value with the trend characteristic value at the previous moment or the previous n moments in real time to see whether the increment exceeds the threshold. Based on this, it is judged whether the three indicators of weld formation continuity, surface quality, and welding spatter are stable. Finally, the weighted statistical values ​​of the three prediction results are used to give the online diagnosis results of welding stability or provide the basis for the regulation of key process parameters for the welding closed-loop control system, such as welding laser power, welding speed, and defocus.

[0025] A welding detection system based on array sensing and artificial intelligence recognition, the system comprising:

[0026] A data acquisition module, wherein the data acquisition module selects a thermally excited state signal inside the keyhole as a detection signal to obtain a mesoscopic detection signal;

[0027] A preprocessing module, wherein the preprocessing module preprocesses the mesoscopic detection signal to obtain computer-recognizable type data;

[0028] A calibration module, which is used to calibrate the preprocessed data and the actual weld formation continuity, surface quality, and welding spatter conditions, and then divides the data into a training set, a test set, and a validation set;

[0029] A model building module, wherein the model building module builds a neural network model, and trains the stability recognition basis of the signals of different regions in the keyhole for the continuity of seam formation, surface quality, and welding spatter through deep learning, so as to recognize the maximum threshold of trend change until all three recognition results can converge, and then adjust the model hyperparameters through the validation set, and verify the reliability of the model through the test set;

[0030] The online detection module calls the trained recognition model to compare in real time whether the increment of the current trend feature value and the trend feature value at the previous moment or the previous n moments exceeds the threshold, and uses this as a basis to judge whether the three indicators of weld formation continuity, surface quality, and welding spatter are stable.

[0031] A computer-readable storage medium stores a computer program, which is executed by a processor to implement an online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition.

[0032] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, an online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition is implemented.

[0033] The present invention has the following beneficial effects:

[0034] The present invention uses the thermally excited state signal inside the keyhole as the detection signal. First, the signal has good consistency with the keyhole morphology and stability characteristics, so it can be used as a direct detection signal, which can avoid the influence of interference factors such as environmental humidity, temperature, gas flow field, etc. when using indirect detection signals. At the same time, the near-infrared spectrum enhancement characteristics of the signal also support the effective shielding of other harmful signals such as welding plume in this spectrum, increase the effective signal ratio in the detection signal, and reduce the difficulty of signal analysis.

[0035] The present invention proposes a method of projecting the real image of the thermally excited state signal onto the array sensor chip to obtain the mesoscopic signal, which can achieve full coverage recognition of the internal area of ​​the keyhole. Since the laser keyhole is always in a fluctuating state during the laser welding process, the position of the characteristic areas of different stability at the mesoscopic scale will also swing accordingly, so full coverage recognition of the characteristic area at the bottom of the keyhole is very necessary. One of the characteristics of the present invention is that it can adaptively track and identify the characteristic area, and can accurately locate the detection signals of the inner wall area of ​​the hole, the keyhole opening area and the opening edge area. At the same time, the high-resolution characteristics of the array sensor can be used to perform high-resolution recognition of the target mesoscopic area, effectively shielding most of the interference signals and improving the detection reliability.

[0036] The present invention adopts the deep learning method in machine learning, and can automatically identify the mesoscopic signal characteristics of different characteristic areas by detecting signals in thermally excited state signals, thereby completing the effective prediction of weld formation continuity, surface quality, and welding spatter amount, as well as a comprehensive and reliable evaluation of welding stability. In addition, the signal preprocessing method that superimposes and averages data within a period of time can also avoid individual signals, capture the regular characteristics of signals, and effectively avoid data analysis problems caused by complex welding signals, large fluctuations, and many interference signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0038] Figure 1 Schematic diagram of the detection device assembly; 1. Welding detection device housing; 2. Substrate; 3. Array or image sensor; 4. Narrowband filter; 5. Three-dimensional fine-tuning mechanism

[0039] Figure 2 is a flow chart of the method of the present invention;

[0040] Figure 3 This is a flow chart of the artificial intelligence detection method. DETAILED DESCRIPTION

[0041] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0043] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. Specific embodiment one:

[0045] according to Figures 1 to 3 As shown, the specific optimization technical solution adopted by the present invention to solve the above technical problems is: the present invention relates to an online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition.

[0046] A welding detection device based on array sensing and artificial intelligence recognition, the device comprising: a housing, a substrate, a sensor, a narrow-band filter and a three-dimensional fine-tuning mechanism;

[0047] The connection between the substrate and the laser welding head serves to fix the entire device. There are mounting holes on the substrate to respectively fix the device housing, the three-dimensional fine-tuning mechanism and the narrow-band filter carrier. The three-dimensional fine-tuning mechanism is connected to the sensor to adjust the three-dimensional spatial position of the sensor sensing chip. The narrow-band filter is located on the optical path between the optical focusing lens group and the array sensor sensing surface. The optical real image after filtering can be projected onto the array sensor sensing surface.

[0048] An online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition, the method is based on a welding detection device based on array sensing and artificial intelligence recognition, and the method comprises the following steps:

[0049] Step 1: Select the thermally excited state signal inside the keyhole as the detection signal to obtain the mesoscopic detection signal;

[0050] Step 2: pre-process the mesoscopic detection signal to obtain computer-recognizable type data;

[0051] Step 3: According to the preprocessed data, the data is calibrated and divided into training set, test set and validation set;

[0052] Step 4: Establish a neural network model and train the recognition ability of signals in different areas of the keyhole for seam continuity, surface quality, and welding spatter through deep learning until all three recognition results converge, and then adjust the model hyperparameters through the validation set;

[0053] Step 5: Test the reliability of the model through the test set;

[0054] Step 6: Call the trained recognition model to analyze signal features online and extract welding stability information. Specific embodiment 2:

[0056] The difference between the second embodiment of the present application and the first embodiment is that:

[0057] The thermally excited state mesoscopic signal in the keyhole is selected as the detection object, and the clear real image of the thermally excited state signal inside the keyhole is projected onto the sensing surface of the array sensor by using the optical focusing imaging and spectral transmission principle to obtain the mesoscopic detection signal. The deep method in machine learning is used to automatically identify the characteristics of the mesoscopic signals in different areas, predict the continuity of weld formation, surface quality, trend of welding spatter, and obtain the current keyhole stability information.

[0058] The thermally excited state signal inside the keyhole is a near-infrared signal generated when the laser beam enters the base material and the metal at the bottom of the keyhole rapidly melts and evaporates through intense energy input, accompanied by high-density energy excitation. Since it is generated on the surface of the keyhole, the characteristic changes of the signal are highly consistent with the dynamic behavior of the keyhole.

[0059] The acquisition method of mesoscopic detection signal is:

[0060] S1. A high-power optical focusing lens group with at least 0.6-1.5m shooting working distance and 10mm shooting depth of field is used to extract a clear real image of the characteristic area at the bottom of the keyhole from the coaxial optical path of the laser welding head. The sufficiently large depth of field can capture a clear real image of the characteristic area in the fluctuating state without changing the focal length;

[0061] S2. In the near-infrared spectrum, the welding plume, laser beam and other large amounts of welding radiation signals above the keyhole are effectively shielded by narrow-band filtering, so that the thermally excited state signals in the characteristic area can be effectively separated, thereby greatly reducing the proportion of invalid signals in the detection signal;

[0062] S3. Project the real image of the thermally excited state signal onto the sensing surface of an array sensor with a large sensing coverage area and high detection accuracy. The sensing area of ​​the array sensor should be ≥ the feature area to be measured, and the resolution accuracy should be ≤10μm, thereby obtaining mesoscopic signals at different positions in the feature area to be measured.

[0063] The optimal model of welding stability is obtained specifically as follows:

[0064] The analysis sample data collected by the array sensor or image sensor is converted into computer-recognizable type data after certain data processing. Secondly, the analysis samples are calibrated with the actual weld formation continuity, surface quality, and welding spatter amount and then divided into a training set, a test set, and a verification set. Then, a neural network model is constructed by a computer. Through deep learning, the computer is allowed to simultaneously train the stability recognition basis of the weld formation continuity, surface quality, and welding spatter amount of signals in different areas of the keyhole. The maximum threshold of the trend change can be identified until all three recognition results can converge. Then, the model hyperparameters are adjusted through the verification set, and the reliability of the model is tested through the test set.

[0065] The collected single-channel / multi-channel / or all mesoscopic detection signal data are converted into computer-recognizable type data through certain data processing, and the trained recognition model is called to compare the increment of the current trend characteristic value with the trend characteristic value of the previous moment or the previous n moments in real time to see whether it exceeds the threshold. Based on this, it is used to judge whether the three indicators of weld formation continuity, surface quality, and welding spatter are stable. Finally, the weighted statistical values ​​of the three prediction results are used to give the online diagnosis results of welding stability or provide the basis for the regulation of key process parameters for the welding closed-loop control system, such as welding laser power, welding speed, and defocus. Specific embodiment three:

[0067] The difference between the third embodiment of the present application and the second embodiment is that:

[0068] The purpose of the present invention is to provide an online detection method which can effectively detect and identify the welding stability of laser / laser composite energy field.

[0069] The above purpose is achieved through the following technical solutions:

[0070] An online detection method for laser welding stability based on array sensing and artificial intelligence recognition is proposed. This method uses array sensing to collect the thermally excited state mesoscopic detection signal of the characteristic area inside the keyhole, and establishes a neural network model through machine learning. The trained model is then called to analyze the signal characteristics and extract welding stability information. This artificial intelligence online detection method can have a good prediction and recognition effect on the continuity of weld formation, surface quality, and welding spatter. The specific steps are as follows:

[0071] Firstly, the thermally excited state mesoscopic signal in the keyhole is selected as the detection object. Secondly, the clear real image of the thermally excited state signal inside the keyhole is projected onto the sensing surface of the array sensor using the principle of optical focusing imaging and spectral transmission to obtain the mesoscopic detection signal. Thirdly, the actual stability result is associated with the detection data, and a large amount of data is obtained by the same method to form an analysis sample set. Then, a neural network model is constructed by computer. Through deep learning, the computer is allowed to simultaneously train the stability recognition basis of the signals in different regions of the keyhole for the continuity of seam formation, surface quality, and welding spatter. The maximum threshold of trend change can be identified until all three recognition results can converge. Then, the model hyperparameters are adjusted through the validation set, and the reliability of the model is tested through the test set. Finally, the trained recognition model is called to compare the increment of the current trend feature value with the trend feature value of the previous moment or the previous n moments in real time to see whether it exceeds the threshold, and based on this, the three indicators of weld formation continuity, surface quality, and welding spatter are judged to be stable.

[0072] The thermally excited state signal inside the keyhole is a near-infrared signal generated when the laser beam enters the base material and the metal at the bottom of the keyhole is rapidly melted and evaporated through intense energy input, accompanied by high-density energy excitation. Therefore, the thermally excited state signal and the metal vapor recoil pressure are generated at the same time, in the same position, and with the same trend of change. The formation of the keyhole is completed under the recoil pressure of metal evaporation. Therefore, the thermally excited state signal inside the keyhole is highly consistent with the keyhole morphology and stability characteristics.

[0073] The method for collecting mesoscopic detection signals is as follows: first, a high-power optical focusing lens group with a shooting working distance of at least 0.6-1.5m and a shooting depth of field of 10mm is used to extract a clear real image of the characteristic area at the bottom of the keyhole from the coaxial optical path of the laser welding head. A sufficiently large depth of field can capture a clear real image of the characteristic area under a fluctuating state without changing the focal length. Secondly, in the near-infrared spectrum, the welding plume, laser beam and other large amounts of welding radiation signals above the keyhole are effectively shielded by narrow-band filtering, so that the thermally excited state signals in the characteristic area can be effectively separated, thereby greatly reducing the proportion of invalid signals in the detection signal. Then, the real image of the thermally excited state signal is projected onto the sensing surface of an array sensor with a large sensing coverage area and high detection accuracy. The sensing area of ​​the array sensor should be ≥ the characteristic area to be measured, and the resolution accuracy should be ≤10μm. In this way, the mesoscopic signals at different positions in the characteristic area to be measured are obtained. This method can not only obtain the thermal excitation signals at all positions in the area to be measured, and accurately obtain the positions of the characteristic signals of the inner wall of the keyhole, the upper opening and the edge of the opening, but also directly extract the mesoscopic signals at key positions to further increase the proportion of effective signals in the detection data, reduce the amount of data for signal analysis, and provide data guarantee for the next step of detection signal analysis.

[0074] The artificial intelligence detection method utilizes the fact that the thermally excited state signal of the characteristic area inside the keyhole follows the nature of the welding thermal reaction and has regularity in the trend characteristic change of the signal. Through a large amount of data analysis, it avoids individual signal cases and captures regular characteristics, and at the same time improves the accuracy of the analysis through the composite recognition of multi-region signals. The specific method is: the collected single-channel / multi-channel / or all mesoscopic detection signal data is first converted into computer-recognizable type data after certain data processing, and then the trained recognition model is called for calculation to obtain three prediction analysis results of weld formation continuity, surface quality, and welding spatter. Finally, the weighted statistical values ​​of the three prediction results are used to give the online diagnosis results of welding stability or provide the key process parameters for the welding closed-loop control system. The basis for regulation, such as welding laser power, welding speed, defocus, etc.

[0075] The method for establishing the recognition model is to convert the analysis sample data collected by the array sensor or image sensor into computer-recognizable type data after certain data processing, and then calibrate the analysis samples with the actual weld formation continuity, surface quality, and welding spatter amount and divide them into a training set, a test set, and a verification set, and then use a computer to build a neural network model, and train the stability recognition basis of the seam formation continuity, surface quality, and welding spatter amount of signals in different areas of the keyhole through deep learning, so as to identify the maximum threshold of the trend change until the three recognition results can converge, and then adjust the model hyperparameters through the verification set, and verify the reliability of the model through the test set. Specific embodiment four:

[0077] The difference between the fourth embodiment of the present application and the third embodiment is that:

[0078] The present invention provides a welding detection system based on array sensing and artificial intelligence recognition, the system comprising:

[0079] A data acquisition module, wherein the data acquisition module selects a thermally excited state signal inside the keyhole as a detection signal to obtain a mesoscopic detection signal;

[0080] A preprocessing module, wherein the preprocessing module preprocesses the mesoscopic detection signal to obtain computer-recognizable type data;

[0081] A calibration module, wherein the calibration module calibrates the preprocessed data and divides the data into a training set, a test set and a validation set;

[0082] A model building module, wherein the model building module builds a neural network model, and trains the stability recognition basis of the signals of different regions in the keyhole for the continuity of seam formation, surface quality, and welding spatter through deep learning, so as to recognize the maximum threshold of trend change until all three recognition results can converge, and then adjust the model hyperparameters through the validation set, and verify the reliability of the model through the test set;

[0083] The online detection module calls the trained recognition model to compare in real time whether the increment of the current trend feature value and the trend feature value at the previous moment or the previous n moments exceeds the threshold, and uses this as a basis to judge whether the three indicators of weld formation continuity, surface quality, and welding spatter are stable. Specific embodiment five:

[0085] The difference between the fifth embodiment of the present application and the fourth embodiment is that:

[0086] The present invention provides a computer-readable storage medium on which a computer program is stored. The program is executed by a processor to implement an online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition.

[0087] The steps include:

[0088] Firstly, the thermally excited state mesoscopic signal in the keyhole is selected as the detection object. Secondly, the clear real image of the thermally excited state signal inside the keyhole is projected onto the sensing surface of the array sensor using the principle of optical focusing imaging and spectral transmission to obtain the mesoscopic detection signal. Thirdly, the actual stability result is associated with the detection data, and a large amount of data is obtained by the same method to form an analysis sample set. Then, a neural network model is constructed by computer. Through deep learning, the computer is allowed to simultaneously train the stability recognition basis of the signals in different regions of the keyhole for the continuity of seam formation, surface quality, and welding spatter. The maximum threshold of trend change can be identified until all three recognition results can converge. Then, the model hyperparameters are adjusted through the validation set, and the reliability of the model is tested through the test set. Finally, the trained recognition model is called to compare the increment of the current trend feature value with the trend feature value of the previous moment or the previous n moments in real time to see whether it exceeds the threshold, and based on this, the three indicators of weld formation continuity, surface quality, and welding spatter are judged to be stable.

[0089] The thermally excited state signal inside the keyhole is a near-infrared signal generated when the laser beam enters the base material and the metal at the bottom of the keyhole is rapidly melted and evaporated through intense energy input, accompanied by high-density energy excitation. Therefore, the thermally excited state signal and the metal vapor recoil pressure are generated at the same time, in the same position, and with the same trend of change. The formation of the keyhole is completed under the recoil pressure of metal evaporation. Therefore, the thermally excited state signal inside the keyhole is highly consistent with the keyhole morphology and stability characteristics.

[0090] The method for collecting mesoscopic detection signals is as follows: first, a high-power optical focusing lens group with a shooting working distance of at least 0.6-1.5m and a shooting depth of field of 10mm is used to extract a clear real image of the characteristic area at the bottom of the keyhole from the coaxial optical path of the laser welding head. A sufficiently large depth of field can capture a clear real image of the characteristic area under a fluctuating state without changing the focal length. Secondly, in the near-infrared spectrum, the welding plume, laser beam and other large amounts of welding radiation signals above the keyhole are effectively shielded by narrow-band filtering, so that the thermally excited state signals in the characteristic area can be effectively separated, thereby greatly reducing the proportion of invalid signals in the detection signal. Then, the real image of the thermally excited state signal is projected onto the sensing surface of an array sensor with a large sensing coverage area and high detection accuracy. The sensing area of ​​the array sensor should be ≥ the characteristic area to be measured, and the resolution accuracy should be ≤10μm. In this way, the mesoscopic signals at different positions in the characteristic area to be measured are obtained. This method can not only obtain the thermal excitation signals at all positions in the area to be measured, and accurately obtain the positions of the characteristic signals of the inner wall of the keyhole, the upper opening and the edge of the opening, but also directly extract the mesoscopic signals at key positions to further increase the proportion of effective signals in the detection data, reduce the amount of data for signal analysis, and provide data guarantee for the next step of detection signal analysis.

[0091] The artificial intelligence detection method utilizes the fact that the thermally excited state signal of the characteristic area inside the keyhole follows the nature of the welding thermal reaction and has regularity in the trend characteristic change of the signal. Through a large amount of data analysis, it avoids individual signal cases and captures regular characteristics, and at the same time improves the accuracy of the analysis through the composite recognition of multi-region signals. The specific method is: the collected single-channel / multi-channel / or all mesoscopic detection signal data is first converted into computer-recognizable type data after certain data processing, and then the trained recognition model is called for calculation to obtain three prediction analysis results of weld formation continuity, surface quality, and welding spatter. Finally, the weighted statistical values ​​of the three prediction results are used to give the online diagnosis results of welding stability or provide the key process parameters for the welding closed-loop control system. The basis for regulation, such as welding laser power, welding speed, defocus, etc.

[0092] The method for establishing the recognition model is to convert the analysis sample data collected by the array sensor or image sensor into computer-recognizable type data after certain data processing, and then calibrate the analysis samples with the actual weld formation continuity, surface quality, and welding spatter amount and divide them into a training set, a test set, and a verification set, and then use a computer to build a neural network model, and train the stability recognition basis of the seam formation continuity, surface quality, and welding spatter amount of signals in different areas of the keyhole through deep learning, so as to identify the maximum threshold of the trend change until the three recognition results can converge, and then adjust the model hyperparameters through the verification set, and verify the reliability of the model through the test set. Specific embodiment six:

[0094] The difference between the sixth embodiment of the present application and the fifth embodiment is that:

[0095] The present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, an online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition is implemented.

[0096] The steps include:

[0097] Firstly, the thermally excited state mesoscopic signal in the keyhole is selected as the detection object. Secondly, the clear real image of the thermally excited state signal inside the keyhole is projected onto the sensing surface of the array sensor using the principle of optical focusing imaging and spectral transmission to obtain the mesoscopic detection signal. Thirdly, the actual stability result is associated with the detection data, and a large amount of data is obtained by the same method to form an analysis sample set. Then, a neural network model is constructed by computer. Through deep learning, the computer is allowed to simultaneously train the stability recognition basis of the signals in different regions of the keyhole for the continuity of seam formation, surface quality, and welding spatter. The maximum threshold of trend change can be identified until all three recognition results can converge. Then, the model hyperparameters are adjusted through the validation set, and the reliability of the model is tested through the test set. Finally, the trained recognition model is called to compare the increment of the current trend feature value with the trend feature value of the previous moment or the previous n moments in real time to see whether it exceeds the threshold, and based on this, the three indicators of weld formation continuity, surface quality, and welding spatter are judged to be stable.

[0098] The thermally excited state signal inside the keyhole is a near-infrared signal generated when the laser beam enters the base material and the metal at the bottom of the keyhole is rapidly melted and evaporated through intense energy input, accompanied by high-density energy excitation. Therefore, the thermally excited state signal and the metal vapor recoil pressure are generated at the same time, in the same position, and with the same trend of change. The formation of the keyhole is completed under the recoil pressure of metal evaporation. Therefore, the thermally excited state signal inside the keyhole is highly consistent with the keyhole morphology and stability characteristics.

[0099] The method for collecting mesoscopic detection signals is as follows: first, a high-power optical focusing lens group with a shooting working distance of at least 0.6-1.5m and a shooting depth of field of 10mm is used to extract a clear real image of the characteristic area at the bottom of the keyhole from the coaxial optical path of the laser welding head. A sufficiently large depth of field can capture a clear real image of the characteristic area under a fluctuating state without changing the focal length. Secondly, in the near-infrared spectrum, the welding plume, laser beam and other large amounts of welding radiation signals above the keyhole are effectively shielded by narrow-band filtering, so that the thermally excited state signals in the characteristic area can be effectively separated, thereby greatly reducing the proportion of invalid signals in the detection signal. Then, the real image of the thermally excited state signal is projected onto the sensing surface of an array sensor with a large sensing coverage area and high detection accuracy. The sensing area of ​​the array sensor should be ≥ the characteristic area to be measured, and the resolution accuracy should be ≤10μm. In this way, the mesoscopic signals at different positions in the characteristic area to be measured are obtained. This method can not only obtain the thermal excitation signals at all positions in the area to be measured, and accurately obtain the positions of the characteristic signals of the inner wall of the keyhole, the upper opening and the edge of the opening, but also directly extract the mesoscopic signals at key positions to further increase the proportion of effective signals in the detection data, reduce the amount of data for signal analysis, and provide data guarantee for the next step of detection signal analysis.

[0100] The artificial intelligence detection method utilizes the fact that the thermally excited state signal of the characteristic area inside the keyhole follows the nature of the welding thermal reaction and has regularity in the trend characteristic change of the signal. Through a large amount of data analysis, it avoids individual signal cases and captures regular characteristics, and at the same time improves the accuracy of the analysis through the composite recognition of multi-region signals. The specific method is: the collected single-channel / multi-channel / or all mesoscopic detection signal data is first converted into computer-recognizable type data after certain data processing, and then the trained recognition model is called for calculation to obtain three prediction analysis results of weld formation continuity, surface quality, and welding spatter. Finally, the weighted statistical values ​​of the three prediction results are used to give the online diagnosis results of welding stability or provide the key process parameters for the welding closed-loop control system. The basis for regulation, such as welding laser power, welding speed, defocus, etc.

[0101] The method for establishing the recognition model is to convert the analysis sample data collected by the array sensor or image sensor into computer-recognizable type data after certain data processing, and then calibrate the analysis samples with the actual weld formation continuity, surface quality, and welding spatter amount and divide them into a training set, a test set, and a verification set, and then use a computer to build a neural network model, and train the stability recognition basis of the seam formation continuity, surface quality, and welding spatter amount of signals in different areas of the keyhole through deep learning, so as to identify the maximum threshold of the trend change until the three recognition results can converge, and then adjust the model hyperparameters through the verification set, and verify the reliability of the model through the test set.

[0102] In the description of this specification, the description of reference terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples without contradiction. In addition, the terms "first" and "second" are used only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, the features defined as "first" and "second" can explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. Any process or method description in the flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code including one or more executable instructions for implementing the steps of a custom logic function or process, and the scope of the preferred embodiment of the present invention includes other implementations, in which the functions may not be performed in the order shown or discussed, including in a substantially simultaneous manner or in a reverse order according to the functions involved, which should be understood by a person skilled in the art of the art to which the embodiments of the present invention belong. The logic and / or steps represented in the flowchart or otherwise described herein, for example, may be considered as a sequenced list of executable instructions for implementing the logic function, and may be specifically implemented in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" may be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection having one or N wirings (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM).In addition, the computer readable medium can even be paper or other suitable medium on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, then editing, interpreting or processing in other suitable ways as necessary, and then storing it in a computer memory. It should be understood that the various parts of the present invention can be implemented with hardware, software, firmware, or a combination thereof. In the above-mentioned embodiment, N steps or methods can be implemented with software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented with hardware, as in another embodiment, any one of the following technologies known in the art or their combination can be used to implement: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0103] The above is only a preferred implementation of an online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition. The protection scope of an online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition is not limited to the above embodiments. All technical solutions under this idea belong to the protection scope of the present invention. It should be pointed out that for those skilled in the art, several improvements and changes without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. An online detection method for laser welding forming stability based on array sensing and artificial intelligence recognition, the method is based on a welding detection device based on array sensing and artificial intelligence recognition, the device comprising: Housing, substrate, array sensor, narrow-band filter and three-dimensional fine-tuning mechanism; The substrate is connected with the laser welding head to fix the whole device. The substrate is provided with mounting holes to respectively fix the device housing, the three-dimensional fine-tuning mechanism and the narrow-band filter carrier. The three-dimensional fine-tuning mechanism is connected with the array sensor to adjust the three-dimensional spatial position of the array sensor sensing chip. The narrow-band filter is located on the optical path between the optical focusing lens group and the array sensor sensing surface. The optical real image after filtering can be projected on the array sensor sensing surface. The method is characterized in that: the method comprises the following steps: Step 1: Select the thermally excited state signal inside the keyhole as the detection signal to obtain the mesoscopic detection signal; Step 2: pre-process the mesoscopic detection signal to obtain computer-recognizable type data; Step 3: According to the preprocessed data, the data is calibrated and divided into training set, test set and validation set; Step 4: Establish a neural network model, and train the recognition ability of signals in different areas of the keyhole for weld formation continuity, surface quality, and welding spatter through deep learning until all three recognition results converge, and then adjust the model hyperparameters through the validation set; Step 5: Test the reliability of the model through the test set; Step 6: Call the trained recognition model to analyze signal features online and extract welding stability information.

2. The method according to claim 1, characterized in that: The thermally excited state mesoscopic signal in the keyhole is selected as the detection object, and the clear real image of the thermally excited state signal inside the keyhole is projected onto the sensing surface of the array sensor by using the optical focusing imaging and spectral transmission principle to obtain the mesoscopic detection signal. The deep learning method in machine learning is used to automatically identify the characteristics of the mesoscopic signals in different areas, predict the trend of weld formation continuity, surface quality, and welding spatter amount, and obtain the current keyhole stability information.

3. The method according to claim 1, characterized in that: The thermally excited state signal inside the keyhole is a near-infrared signal generated when the laser beam enters the base material and the metal at the bottom of the keyhole is rapidly melted and evaporated through intense energy input, accompanied by high-density energy excitation.

4. The method according to claim 2 or 3, characterized in that: The acquisition method of mesoscopic detection signal is: S1. A clear real image of the characteristic area at the bottom of the keyhole is extracted from the coaxial optical path of the laser welding head through a high-power optical focusing lens group with a shooting working distance of 0.6-1.5m and a shooting depth of field of 10mm. The 10mm shooting depth of field captures a clear real image of the characteristic area in a fluctuating state while the focal length does not change; S2. In the near-infrared spectrum, the welding plume, laser beam and other large amounts of welding radiation signals above the keyhole are effectively shielded by narrow-band filtering, so that the thermally excited state signals in the characteristic area can be effectively separated, thereby greatly reducing the proportion of invalid signals in the detection signal; S3. Project the real image of the thermally excited state signal onto the sensing surface of an array sensor with a large sensing coverage area and high detection accuracy. The sensing area of ​​the array sensor is ≥ the characteristic area to be measured, and the resolution accuracy is ≤10μm, thereby obtaining mesoscopic signals at different positions in the characteristic area to be measured.

5. The method according to claim 4, characterized in that: The optimal model of welding stability is obtained specifically as follows: The analysis sample data collected by the array sensor is converted into computer-recognizable data after certain data processing, and then the analysis samples are calibrated with the actual weld formation continuity, surface quality, and welding spatter amount and divided into a training set, a test set, and a verification set, and then a neural network model is constructed by computer to train a maximum threshold value that can identify trend changes, and then during the online stability monitoring, the increment of the current trend characteristic value and the trend characteristic value at the previous moment or the previous n moments is compared in real time to see whether it exceeds the threshold value, and based on this, whether it is stable; Through deep learning, the computer is allowed to simultaneously train the signals in different areas of the keyhole to identify the stability of weld formation continuity, surface quality, and welding spatter. The maximum threshold of trend changes can be identified until all three identification results converge. The model hyperparameters are then adjusted through the validation set, and the reliability of the model is tested through the test set.

6. The method according to claim 5, characterized in that: The collected single-channel / multi-channel / or all mesoscopic detection signal data are converted into computer-recognizable type data through certain data processing, and the trained recognition model is called to compare the increment of the current trend characteristic value with the trend characteristic value of the previous moment or the previous n moments in real time to see whether it exceeds the threshold. Based on this, it is judged whether the three indicators of weld formation continuity, surface quality, and welding spatter amount are stable. Finally, the weighted statistical values ​​of the three prediction results are used to give the online diagnosis result of welding stability or provide the basis for regulating the key process parameters for the welding closed-loop control system. The key process parameters include welding laser power, welding speed, and defocus.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method according to any one of claims 1 to 6.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

Images