Titanium alloy welding fault monitoring method, device, system, equipment and product

Abstract

The invention discloses a titanium alloy welding fault monitoring method, device, system, equipment and product, and relates to the technical field of welding monitoring. The method comprises the following steps: firstly, receiving visual acquisition data from a visual sensor, spectrum acquisition data from a spectrum sensor, power supply acquisition data from a power supply sensor and temperature acquisition data from an infrared temperature sensor in real time, and then carrying out time synchronization and multi-dimensional feature extraction processing on the collected multi-source data; the method comprises the following steps: obtaining sequence data of a multi-dimensional feature vector currently used for reflecting a titanium alloy welding health instantaneous state, finally importing the sequence data into a titanium alloy welding fault recognition model obtained by pre-training based on a time sequence classification model to obtain a current fault recognition result, and outputting a current titanium alloy welding decision according to the recognition result. In this way, comprehensive perception of the titanium alloy welding process and early-stage accurate prediction of faults such as arc stopping can be achieved, passive detection is changed into active intervention, and the welding stability and the yield are effectively guaranteed.

Description

Technical Field

[0001] This invention belongs to the field of titanium alloy welding monitoring technology, specifically relating to a method, device, system, equipment and product for monitoring titanium alloy welding faults. Background Technology

[0002] Titanium alloys, due to their high strength, low density, and excellent corrosion resistance, have become the preferred material for key structural components in high-end equipment such as aerospace and shipbuilding. However, titanium alloys are chemically highly reactive and extremely sensitive to the purity and effectiveness of the protective atmosphere during welding. They readily react with elements such as oxygen, nitrogen, and hydrogen, leading to metallurgical defects such as embrittlement and porosity in the weld area, severely impairing the mechanical properties of the welded joint.

[0003] Furthermore, the inherent poor thermal conductivity and unique molten pool fluidity of titanium alloys result in an extremely narrow welding process parameter window. This makes them susceptible to a series of process defects during welding, such as incomplete penetration, undercut, and collapse, even from minor parameter fluctuations. These defects and defects accumulate and interact during welding, potentially leading to a serious "arc stoppage" event. This not only scraps parts but also disrupts the production process, resulting in significant economic losses and time costs.

[0004] Currently, the industry mainly relies on post-weld non-destructive testing techniques, such as X-ray inspection or ultrasonic testing, to control the welding quality of titanium alloys. However, these methods are "post-construction verification" means, which can only detect and identify defects after they have formed. They cannot intervene and adjust the welding process in real time, and therefore cannot prevent failures from occurring at the source.

[0005] While some existing welding process monitoring systems can monitor single-dimensional electrical parameters such as current or voltage, these systems struggle to comprehensively and deeply capture the multidimensional state changes caused by the complex metallurgical behavior of titanium alloys and dynamic process fluctuations. Therefore, current technologies lack effective early prediction and warning capabilities for faults such as arc interruption caused by the characteristics of the welding materials themselves.

[0006] In summary, existing titanium alloy welding monitoring technologies suffer from shortcomings such as lag, simplification, and passivity. To ensure the stability of titanium alloy welding production and product yield, there is an urgent need in this field for an intelligent method and system capable of real-time and multi-dimensional sensing of welding status, and accurate prediction of faults such as arc cessation based on deep fusion of multi-source information. Summary of the Invention

[0007] The purpose of this invention is to provide a method, device, system, computer equipment, computer-readable storage product, and computer program product for monitoring titanium alloy welding faults, in order to solve the problems of lag, singularity, and passivity in existing titanium alloy welding monitoring technologies.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a method for monitoring welding faults in titanium alloys is provided, including: The system receives real-time visual data from a vision sensor, spectral data from a spectral sensor, power data from a power sensor, and temperature data from an infrared temperature sensor. The vision sensor is positioned directly in front of the titanium alloy weld pool to capture real-time visual data reflecting the weld pool image, weld pool orifice behavior, and / or plume morphology. The spectral sensor is also positioned directly in front of the titanium alloy weld pool to capture real-time spectral data reflecting the weld spectral composition. The power sensor is connected to the titanium alloy welding power supply to capture real-time power data reflecting the welding output current and / or welding arc voltage. The infrared temperature sensor is also positioned directly in front of the titanium alloy weld pool to capture real-time temperature data reflecting the temperature field distribution in the weld pool area and the heat-affected zone of the weld pool. Based on a unified time scale, the visual acquisition data, the spectral acquisition data, the power acquisition data, and the temperature acquisition data are processed in real time for time synchronization to obtain visual data, spectral data, power data, and temperature data with synchronized acquisition time. The visual data, spectral data, power data, and temperature data are processed in real time to extract multidimensional features to obtain the current multidimensional feature vector time series data. The multidimensional feature vector time series data contains multiple multidimensional feature vectors arranged in time sequence to reflect the instantaneous health state of titanium alloy welding. The multidimensional feature vector time series data is imported in real time into the titanium alloy welding fault identification model pre-trained based on the time series classification model to obtain the current titanium alloy welding fault identification result. Based on the titanium alloy welding fault identification results, the current titanium alloy welding decision is output in real time.

[0009] Based on the above-mentioned invention, an intelligent solution is provided that perceives the welding status in real time and identifies welding fault types to output response decisions in real time. This solution first receives visual data from a visual sensor, spectral data from a spectral sensor, power data from a power sensor, and temperature data from an infrared temperature sensor in real time. Then, it performs time synchronization and multi-dimensional feature extraction processing on the collected multi-source data to obtain a sequence of multi-dimensional feature vectors reflecting the instantaneous health status of the titanium alloy welding. Finally, this sequence of data is imported into a titanium alloy welding fault identification model pre-trained based on a time-series classification model to obtain the current fault identification result. Based on the identification result, the current titanium alloy welding decision is output. By fusing four-dimensional data (visual, spectral, power, and temperature) and performing real-time analysis based on a time-series model, comprehensive perception of the titanium alloy welding process and early, accurate prediction of faults such as arc termination can be achieved. This transforms passive detection into proactive intervention, effectively ensuring welding stability and yield, and facilitating practical application and promotion.

[0010] In one possible design, after obtaining the time-synchronized visual data, spectral data, power data, and temperature data, the method further includes: The images in the visual data are subjected to noise reduction, contrast enhancement, and / or region of interest extraction to obtain new visual data. And / or, the current signal and / or voltage signal in the power data are filtered to obtain new power data.

[0011] In one possible design, the multidimensional feature vector includes morphological features extracted from the visual data, spectral features extracted from the spectral data, electrical features extracted from the power data, and thermal imaging features extracted from the temperature data. The morphological features include the molten pool area, molten pool aspect ratio, molten pool trailing angle, molten pool orifice diameter, molten pool orifice fluctuation rate, plume area, and / or plume tremor frequency. The spectral features include the relative intensity ratios of characteristic spectral lines for each pair of elements in hydrogen, oxygen, and nitrogen. The electrical features include the mean current, current standard deviation, mean voltage, and / or voltage standard deviation. The thermal imaging features include the highest temperature at the molten pool tip and / or the temperature gradient of the molten pool heat-affected zone.

[0012] In one possible design, the titanium alloy welding fault identification model is pre-trained as follows: Historical multidimensional feature vector time series data are collected and used as model input, and titanium alloy welding fault labels based on expert experience / post-weld inspection results are obtained as model output. Then, the model input and model output are used as sample data. The titanium alloy welding fault labels are divided into "normal", "fault precursor" and "fault in progress" labels. Using all the sample data, the temporal classification model constructed based on a long short-term memory network or a temporal convolutional network is trained to obtain the titanium alloy welding fault identification model.

[0013] In one possible design, when the titanium alloy welding fault identification result includes the probability of being labeled "normal," "precursor to fault," and "in progress," the current titanium alloy welding decision is output in real time based on the titanium alloy welding fault identification result, including: If the probability of the “normal” label in the titanium alloy welding fault identification result is greater than the preset first probability threshold, the current first titanium alloy welding decision, which indicates that no adjustment action is required, is output in real time. And / or, if the probability of the “fault precursor” label in the titanium alloy welding fault identification result is greater than the preset second probability threshold, then the current second titanium alloy welding decision is output in real time and used to indicate the triggering of the audible and visual alarm and / or to send parameter fine-tuning suggestions to the welding controller. And / or, if the probability of the “faulty” label in the titanium alloy welding fault identification result is greater than the preset third probability threshold, then the current third titanium alloy welding decision, which is used to indicate sending an emergency stop command or a current reduction command to the welding machine, is output in real time.

[0014] Secondly, a titanium alloy welding fault monitoring device is provided, which includes a multi-source data receiving unit, a time synchronization processing unit, a multi-dimensional feature extraction unit, a welding fault identification unit, and a welding decision output unit that are sequentially connected in communication. The multi-source data receiving unit is used to receive in real time visual acquisition data from a vision sensor, spectral acquisition data from a spectral sensor, power acquisition data from a power sensor, and temperature acquisition data from an infrared temperature sensor. The vision sensor is positioned directly in front of the titanium alloy weld pool to capture the visual acquisition data in real time, reflecting the image of the weld pool, the behavior of the weld pool apertures, and / or the plume morphology. The spectral sensor is also positioned directly in front of the titanium alloy weld pool to capture the spectral acquisition data in real time, reflecting the spectral composition of the weld. The power sensor is connected to the titanium alloy welding power supply to capture the power acquisition data in real time, reflecting the magnitude of the welding output current and / or the level of the welding arc voltage. The infrared temperature sensor is also positioned directly in front of the titanium alloy weld pool to capture the temperature acquisition data in real time, reflecting the temperature field distribution in the weld pool area and the heat-affected zone of the weld pool. The time synchronization processing unit is used to perform real-time time synchronization processing on the visual acquisition data, the spectral acquisition data, the power acquisition data and the temperature acquisition data based on a unified time scale, so as to obtain visual data, spectral data, power data and temperature data with synchronized acquisition time. The multidimensional feature extraction unit is used to perform multidimensional feature extraction processing on the visual data, the spectral data, the power data and the temperature data in real time to obtain the current multidimensional feature vector time series data. The multidimensional feature vector time series data contains multiple multidimensional feature vectors arranged in time sequence to reflect the instantaneous health state of titanium alloy welding. The welding fault identification unit is used to import the multi-dimensional feature vector time series data into the titanium alloy welding fault identification model pre-trained based on the time series classification model in real time, so as to obtain the current titanium alloy welding fault identification result. The welding decision output unit is used to output the current titanium alloy welding decision in real time based on the titanium alloy welding fault identification result.

[0015] Thirdly, the present invention provides a titanium alloy welding fault monitoring system, including a visual sensor, a spectral sensor, a power sensor, an infrared temperature sensor, and a monitoring device, wherein the visual sensor, the spectral sensor, the power sensor, and the infrared temperature sensor are respectively communicatively connected to the monitoring device; The visual sensor is configured to be positioned directly in front of the titanium alloy welding molten pool to capture visual acquisition data in real time, reflecting the image of the molten pool, the behavior of the molten pool orifice, and / or the plume morphology, and to transmit the visual acquisition data to the monitoring device in real time. The spectral sensor is configured to be positioned directly opposite the titanium alloy weld pool to capture spectral data reflecting the spectral composition of the weld in real time, and to transmit the spectral data to the monitoring device in real time. The power sensor is used to connect to the titanium alloy welding power source to capture power acquisition data in real time, which reflects the magnitude of the welding output current and / or the level of the welding arc voltage, and to transmit the power acquisition data to the monitoring device in real time. The infrared temperature sensor is used to be positioned directly in front of the titanium alloy welding molten pool to capture temperature data in real time, reflecting the temperature field distribution of the molten pool area and the heat-affected zone of the molten pool, and to transmit the temperature data to the monitoring device in real time. The monitoring equipment is used to perform the titanium alloy welding fault monitoring method as described in the first aspect or any possible design in the first aspect.

[0016] Fourthly, the present invention provides a computer device comprising a storage module, a processing module, and a transceiver module connected in sequence for communication, wherein the storage module is used to store a computer program, the transceiver module is used to send and receive messages, and the processing module is used to read the computer program and execute the titanium alloy welding fault monitoring method as described in the first aspect or any possible design in the first aspect.

[0017] Fifthly, the present invention provides a computer-readable storage product storing instructions that, when executed on a computer, perform the titanium alloy welding fault monitoring method as described in the first aspect or any possible design in the first aspect.

[0018] In a sixth aspect, the present invention provides a computer program product, including a computer program or instructions, which, when executed by a computer, implement the titanium alloy welding fault monitoring method as described in the first aspect or any possible design in the first aspect.

[0019] The beneficial effects of the above scheme are: (1) This invention creatively provides an intelligent solution that can perceive the welding status in real time and identify the welding fault type to output response decisions in real time. That is, it first receives visual acquisition data from a visual sensor, spectral acquisition data from a spectral sensor, power acquisition data from a power sensor, and temperature acquisition data from an infrared temperature sensor in real time. Then, it performs time synchronization and multi-dimensional feature extraction processing on the collected multi-source data to obtain the sequence data of multi-dimensional feature vectors that reflect the instantaneous health status of titanium alloy welding. Finally, it imports the sequence data into a titanium alloy welding fault identification model pre-trained based on a time-series classification model to obtain the current fault identification result. Based on the identification result, it outputs the current titanium alloy welding decision. In this way, by integrating four-dimensional data of vision, spectrum, power and temperature, and performing real-time analysis based on a time-series model, it can realize the comprehensive perception of the titanium alloy welding process and the early accurate prediction of faults such as arc stoppage, thereby changing passive detection to active intervention and effectively ensuring welding stability and yield. (2) It can achieve forward-looking prediction, that is, through multi-dimensional feature fusion and time series model, it can identify subtle signs before the occurrence of faults, and realize the transformation from "passive detection" to "active prediction". (3) It has high accuracy, that is, it combines the sensitive indicators unique to titanium alloy welding (such as spectral contamination signal), making the identification model more targeted and the accuracy rate is much higher than that of the method based on a single electrical parameter. (4) Strong adaptability, that is, the time series classification model can learn the failure mode under different working conditions from historical data and has a certain degree of adaptability; (5) It can ensure quality and efficiency, that is, effectively prevent arc stoppage and major defects, reduce scrap rate, improve equipment utilization and production efficiency, and is especially suitable for automated welding of precious materials such as titanium alloys, which is convenient for practical application and promotion. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating the titanium alloy welding fault monitoring method provided in an embodiment of this application.

[0022] Figure 2 This is a schematic diagram of the structure of the titanium alloy welding fault monitoring device provided in the embodiments of this application.

[0023] Figure 3This is a schematic diagram of the structure of the titanium alloy welding fault monitoring system provided in the embodiments of this application.

[0024] Figure 4 A schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be briefly introduced below in conjunction with the accompanying drawings and descriptions of the embodiments or the prior art. Obviously, the following description of the structure of the accompanying drawings is only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these embodiments without creative effort. It should be noted that the description of these embodiments is for the purpose of helping to understand the present invention, but does not constitute a limitation of the present invention.

[0026] It should be understood that although the terms "first" and "second", etc., may be used herein to describe various objects, these objects should not be limited by these terms. These terms are only used to distinguish one object from another. For example, the first object may be referred to as the second object, and similarly, the second object may be referred to as the first object, without departing from the scope of the exemplary embodiments of the invention.

[0027] It should be understood that the term "and / or" that may appear in this document is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can mean: A exists alone, B exists alone, or A and B exist simultaneously. Another example is A, B and / or C, which can mean that any one of A, B, and C or any combination thereof exists. The term " / and" that may appear in this document describes another relationship between related objects, indicating that two relationships can exist. For example, A / and B can mean: A exists alone or A and B exist simultaneously. In addition, the character " / " that may appear in this document generally indicates that the related objects before and after it are in an "or" relationship.

[0028] Example like Figure 1 As shown, the titanium alloy welding fault monitoring method provided in the first aspect of this embodiment can be executed, but is not limited to, by a computer device with certain computing resources, such as... Figure 3 As shown, the monitoring equipment in the titanium alloy welding fault monitoring system performs the monitoring, wherein the titanium alloy welding fault monitoring system also includes, but is not limited to, visual sensors, spectral sensors, power sensors, and infrared temperature sensors that are respectively communicatively connected to the monitoring equipment. Figure 1 As shown, the titanium alloy welding fault monitoring method includes, but is not limited to, the following steps S1 to S5.

[0029] S1. Real-time reception of visual acquisition data from a vision sensor, spectral acquisition data from a spectral sensor, power acquisition data from a power sensor, and temperature acquisition data from an infrared temperature sensor, wherein the vision sensor is positioned facing the titanium alloy weld pool to capture the visual acquisition data in real time, reflecting the image of the weld pool, the behavior of the weld pool keyhole (if it is deep penetration welding), and / or the plume morphology, etc.; the spectral sensor is also positioned facing the titanium alloy weld pool to capture the spectral acquisition data in real time, reflecting the spectral composition of the weld, etc.; the power sensor is connected to the titanium alloy welding power supply to capture the power acquisition data in real time, reflecting the magnitude of the welding output current and / or the level of the welding arc voltage, etc.; and the infrared temperature sensor is also positioned facing the titanium alloy weld pool to capture the temperature acquisition data in real time, reflecting the temperature field distribution of the weld pool area and the heat-affected zone of the weld pool, etc.

[0030] In step S1, the visual acquisition data specifically refers to a sequence of images arranged chronologically. The molten pool, pinhole, and plume are all key phenomena in the welding process; therefore, the terms "molten pool image," "molten pool pinhole behavior," and "plume morphology" are all existing technical terms. The visual sensor is preferably an existing visual sensor with high-speed operation. The spectral acquisition data specifically refers to a sequence of spectra arranged chronologically. The welding spectral components are, for example, the spectral components of the welding arc plasma, and are used as an indirect monitoring result of the purity of the protective atmosphere, so as to determine whether air intrusion has occurred through specific spectral line intensities. The spectral sensor is preferably an existing spectral sensor with high-speed operation. The power acquisition data specifically refers to current sampling signals and / or voltage sampling signals. The power sensor is preferably an existing current sensor and / or voltage sensor with high-frequency sampling characteristics. The temperature acquisition data specifically refers to a sequence of temperature field distribution maps arranged chronologically. The infrared temperature sensor is preferably an existing infrared temperature sensor with high-speed operation.

[0031] S2. Based on a unified time scale, the visual acquisition data, the spectral acquisition data, the power acquisition data, and the temperature acquisition data are processed in real time for time synchronization to obtain visual data, spectral data, power data, and temperature data that are acquired in a time-synchronized manner.

[0032] In step S2, the unified time standard can be an external time standard such as Beijing time, or any time axis among the acquisition time axes of the visual acquisition data, the spectral acquisition data, the power acquisition data, and the temperature acquisition data. Considering that the time synchronization processing involves conventional interpolation processing, and video image data is not convenient for interpolation processing, this embodiment preferably uses the acquisition time axis of the visual acquisition data as the unified time standard. Then, based on this acquisition time axis, conventional data alignment and interpolation processing are performed on the spectral acquisition data, the power acquisition data, and the temperature acquisition data, respectively, to obtain processed spectral data, processed power data, and processed temperature data whose acquisition time is synchronized with the visual acquisition data. Finally, the visual acquisition data, the processed spectral data, the processed power data, and the processed temperature data are used as the visual data, spectral data, power data, and temperature data synchronized with the acquisition time. Furthermore, considering the potential presence of noise in the visual data, spectral data, power data, and / or temperature data, it is necessary to perform corresponding data preprocessing to improve data quality. Specifically, after obtaining the visual data, spectral data, power data, and temperature data synchronized with the acquisition time, the method may include, but is not limited to: performing noise reduction, contrast enhancement, and / or region of interest extraction on the images in the visual data to obtain new visual data; and / or filtering the current and / or voltage signals in the power data to obtain new power data.

[0033] S3. Perform multidimensional feature extraction processing on the visual data, the spectral data, the power data, and the temperature data in real time to obtain the current multidimensional feature vector time series data. The multidimensional feature vector time series data includes, but is not limited to, multiple multidimensional feature vectors arranged in chronological order to reflect the instantaneous health state of titanium alloy welding.

[0034] In step S3, each of the multidimensional feature vectors can be extracted based on the visual data, spectral data, power data, and temperature data collected concurrently (e.g., at the same acquisition timestamp). To ensure that the multidimensional feature vectors comprehensively reflect the instantaneous healthy state of the titanium alloy welding, preferably, the multidimensional feature vectors include, but are not limited to, morphological features extracted from the visual data, spectral features extracted from the spectral data, electrical features extracted from the power data, and thermal imaging features extracted from the temperature data. The morphological features include, but are not limited to, molten pool area, molten pool aspect ratio, molten pool trailing angle, molten pool orifice diameter, molten pool orifice fluctuation rate, plume area, and / or plume tremor frequency, to reflect molten pool stability, heat input conditions, and atmosphere protection effectiveness. The plume morphology changes significantly when the titanium alloy is poorly protected; the spectral characteristics include, but are not limited to, the relative intensity ratios of characteristic spectral lines of each pair of elements in hydrogen, oxygen, and nitrogen, to directly monitor air intrusion and contamination (a key indicator for preventing titanium alloy embrittlement); the electrical characteristics include, but are not limited to, mean current, standard deviation of current, mean voltage, and / or standard deviation of voltage, to reflect arc stability and the smoothness of energy input; the thermal imaging characteristics include, but are not limited to, the highest temperature at the front end of the molten pool and / or the temperature gradient of the heat-affected zone of the molten pool, to reflect the heat accumulation and prevent overheating or insufficient penetration. Furthermore, the terms such as molten pool area, molten pool aspect ratio, molten pool trailing angle, molten pool orifice diameter, molten pool orifice fluctuation rate, plume area, plume tremor frequency, relative intensity ratio of characteristic spectral lines of each pair of elements, mean current, standard deviation of current, mean voltage, standard deviation of voltage, highest temperature at the front end of the molten pool, and temperature gradient of the heat-affected zone of the molten pool are all existing technical terms and can be obtained through conventional statistical methods using existing data collection techniques.

[0035] S4. Import the multidimensional feature vector time series data into the titanium alloy welding fault identification model pre-trained based on the time series classification model in real time to obtain the current titanium alloy welding fault identification result.

[0036] In step S4, the titanium alloy welding fault identification model is an artificial intelligence model with the ability to identify titanium alloy welding fault types (such as "normal", "fault precursor", or "fault in progress") based on the multi-dimensional feature vector time series data. In order for the titanium alloy welding fault identification model to learn the dynamic evolution law before the fault occurs by utilizing the historical feature sequence over a period of time to have the aforementioned identification ability, preferably, the titanium alloy welding fault identification model can be pre-trained according to, but not limited to, the following steps S401 to S402.

[0037] S401. Collect the historical multidimensional feature vector time series data and use it as model input, and obtain titanium alloy welding fault labels based on expert experience / post-weld inspection results as the multidimensional feature vector time series data and use them as model output. Then, the model input and the model output are used as sample data. The titanium alloy welding fault labels are divided into "normal", "fault precursor" and "fault in progress" labels.

[0038] In step S401, the aforementioned historical multidimensional feature vector time-series data can be collected based on the visual acquisition data, spectral acquisition data, power acquisition data, and temperature acquisition data obtained through numerous historical welding experiments, combined with the aforementioned steps S1 to S3, and will not be elaborated further here. Furthermore, the term "fault precursor" specifically includes, but is not limited to, precursors to faults such as arc cessation or burn-through, and the term "in the midst of a fault" specifically includes, but is not limited to, being in the midst of a fault such as arc cessation or burn-through.

[0039] S402. Using all the sample data, train the temporal classification model constructed based on a long short-term memory network or a temporal convolutional network to obtain the titanium alloy welding fault identification model.

[0040] In step S402, the Long Short-Term Memory (LSTM) network is a type of recurrent neural network specifically designed to address the long-term dependency problem inherent in general RNNs (Recurrent Neural Networks). The Temporal Convolutional Network (TCN) is a convolutional neural network model specifically designed for processing time-series data. It captures temporal dependencies in time series data through convolution operations and is commonly used for tasks such as prediction and classification. Therefore, the titanium alloy welding fault identification model can be pre-trained based on all the sample data through a conventional calibration and verification modeling process (which specifically includes model calibration and verification processes, i.e., first comparing the model simulation results with the measured data, and then adjusting the model parameters based on the comparison results to make the simulation results match the actual results). This allows the titanium alloy welding fault identification model to learn to map specific feature change patterns (such as: increased pinhole fluctuations + increased oxygen spectral line intensity + abnormal voltage fluctuations) to "normal," "precursor to fault," or "fault in progress" labels.

[0041] S5. Based on the titanium alloy welding fault identification results, output the current titanium alloy welding decision in real time.

[0042] In step S5, the purpose of the titanium alloy welding decision output is to ensure that the titanium alloy welding process is always in a normal state. Specifically, when the titanium alloy welding fault identification result includes the classification probabilities of "normal," "fault precursor," and "fault in progress" labels (this classification probability can be conventionally obtained based on the identification confidence of the corresponding labels), the current titanium alloy welding decision is output in real time based on the titanium alloy welding fault identification result. This includes, but is not limited to: if the classification probability of the "normal" label in the titanium alloy welding fault identification result is greater than a preset first probability threshold (e.g., 61.8%), then the current first titanium alloy welding decision, indicating that no adjustment action is needed, is output in real time; if the classification probability of the "normal" label in the titanium alloy welding fault identification result is greater than a preset first probability threshold (e.g., 61.8%), then the current first titanium alloy welding decision is output in real time to indicate that no adjustment action is needed ... If the probability of the “precursor of failure” label in the titanium alloy welding fault identification result is greater than a preset second probability threshold (e.g., 61%), then a second titanium alloy welding decision is output in real time to indicate the triggering of an audible and visual alarm (which is used to alert the operator) and / or to send parameter fine-tuning suggestions to the welding controller (at which time a local device communication connection to the welding controller is required); if the probability of the “fault in progress” label in the titanium alloy welding fault identification result is greater than a preset third probability threshold (e.g., 60%), then a third titanium alloy welding decision is output in real time to indicate the sending of an emergency stop command or a current reduction command to the welding machine (at which time a local device communication connection to the welding machine is required), in order to avoid workpiece damage and safety accidents.

[0043] Therefore, based on the titanium alloy welding fault monitoring method described in steps S1 to S5 above, an intelligent solution is provided that can perceive the welding status in real time and identify the welding fault type to output response decisions in real time. Specifically, it first receives visual data from a visual sensor, spectral data from a spectral sensor, power data from a power sensor, and temperature data from an infrared temperature sensor in real time. Then, it performs time synchronization and multi-dimensional feature extraction processing on the collected multi-source data to obtain a sequence of multi-dimensional feature vectors reflecting the instantaneous health status of the titanium alloy welding. Finally, it imports this sequence of data into a titanium alloy welding fault identification model pre-trained based on a time-series classification model to obtain the current fault identification result. Based on the identification result, it outputs the current titanium alloy welding decision. By fusing four-dimensional data (visual, spectral, power, and temperature) and performing real-time analysis based on a time-series model, it can achieve comprehensive perception of the titanium alloy welding process and early accurate prediction of faults such as arc termination, thus transforming passive detection into proactive intervention, effectively ensuring welding stability and yield, and facilitating practical application and promotion.

[0044] like Figure 2As shown, the second aspect of this embodiment provides a virtual device for implementing the titanium alloy welding fault monitoring method described in the first aspect, including a multi-source data receiving unit, a time synchronization processing unit, a multi-dimensional feature extraction unit, a welding fault identification unit, and a welding decision output unit that are sequentially connected in communication. The multi-source data receiving unit is used to receive in real time visual acquisition data from a vision sensor, spectral acquisition data from a spectral sensor, power acquisition data from a power sensor, and temperature acquisition data from an infrared temperature sensor. The vision sensor is positioned directly in front of the titanium alloy weld pool to capture the visual acquisition data in real time, reflecting the image of the weld pool, the behavior of the weld pool apertures, and / or the plume morphology. The spectral sensor is also positioned directly in front of the titanium alloy weld pool to capture the spectral acquisition data in real time, reflecting the spectral composition of the weld. The power sensor is connected to the titanium alloy welding power supply to capture the power acquisition data in real time, reflecting the magnitude of the welding output current and / or the level of the welding arc voltage. The infrared temperature sensor is also positioned directly in front of the titanium alloy weld pool to capture the temperature acquisition data in real time, reflecting the temperature field distribution in the weld pool area and the heat-affected zone of the weld pool. The time synchronization processing unit is used to perform real-time time synchronization processing on the visual acquisition data, the spectral acquisition data, the power acquisition data and the temperature acquisition data based on a unified time scale, so as to obtain visual data, spectral data, power data and temperature data with synchronized acquisition time. The multidimensional feature extraction unit is used to perform multidimensional feature extraction processing on the visual data, the spectral data, the power data and the temperature data in real time to obtain the current multidimensional feature vector time series data. The multidimensional feature vector time series data contains multiple multidimensional feature vectors arranged in time sequence to reflect the instantaneous health state of titanium alloy welding. The welding fault identification unit is used to import the multi-dimensional feature vector time series data into the titanium alloy welding fault identification model pre-trained based on the time series classification model in real time, so as to obtain the current titanium alloy welding fault identification result. The welding decision output unit is used to output the current titanium alloy welding decision in real time based on the titanium alloy welding fault identification result.

[0045] The working process, working details and technical effects of the aforementioned device provided in the second aspect of this embodiment can be found in the titanium alloy welding fault monitoring method described in the first aspect, and will not be repeated here.

[0046] like Figure 3As shown, the third aspect of this embodiment provides a physical system for monitoring titanium alloy welding faults using the method described in the first aspect, including a visual sensor, a spectral sensor, a power sensor, an infrared temperature sensor, and a monitoring device, wherein the visual sensor, the spectral sensor, the power sensor, and the infrared temperature sensor are respectively communicatively connected to the monitoring device; The visual sensor is configured to be positioned directly in front of the titanium alloy welding molten pool to capture visual acquisition data in real time, reflecting the image of the molten pool, the behavior of the molten pool orifice, and / or the plume morphology, and to transmit the visual acquisition data to the monitoring device in real time. The spectral sensor is configured to be positioned directly opposite the titanium alloy weld pool to capture spectral data reflecting the spectral composition of the weld in real time, and to transmit the spectral data to the monitoring device in real time. The power sensor is used to connect to the titanium alloy welding power source to capture power acquisition data in real time, which reflects the magnitude of the welding output current and / or the level of the welding arc voltage, and to transmit the power acquisition data to the monitoring device in real time. The infrared temperature sensor is used to be positioned directly in front of the titanium alloy welding molten pool to capture temperature data in real time, reflecting the temperature field distribution of the molten pool area and the heat-affected zone of the molten pool, and to transmit the temperature data to the monitoring device in real time. The monitoring equipment is used to perform the titanium alloy welding fault monitoring method as described in the first aspect.

[0047] The working process, working details and technical effects of the aforementioned system provided in the third aspect of this embodiment can be found in the titanium alloy welding fault monitoring method described in the first aspect, and will not be repeated here.

[0048] like Figure 4As shown, the fourth aspect of this embodiment provides a computer device for executing the titanium alloy welding fault monitoring method as described in the first aspect. The device includes a storage module, a processing module, and a transceiver module connected in sequence. The storage module stores a computer program, the transceiver module sends and receives messages, and the processing module reads the computer program and executes the titanium alloy welding fault monitoring method as described in the first aspect. Specifically, the storage module may include, but is not limited to, random-access memory (RAM), read-only memory (ROM), flash memory, first-in-first-out (FIFO) memory, and / or first-in-last-out (FILO) memory, etc.; the processing module may, but is not limited to, use a microprocessor of the STM32F105 series. Furthermore, the computer device may also include, but is not limited to, a power supply module, a display screen, and other necessary components.

[0049] The working process, working details and technical effects of the aforementioned computer equipment provided in the fourth aspect of this embodiment can be found in the titanium alloy welding fault monitoring method described in the first aspect, and will not be repeated here.

[0050] This fifth aspect of the embodiment provides a computer-readable storage product that stores instructions comprising the titanium alloy welding fault monitoring method as described in the first aspect. Specifically, the computer-readable storage product stores instructions that, when executed on a computer, perform the titanium alloy welding fault monitoring method as described in the first aspect. The computer-readable storage product refers to a data storage medium, which may include, but is not limited to, computer-readable storage media such as floppy disks, optical disks, hard disks, flash memory, USB flash drives, and / or Memory Sticks. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0051] The working process, working details and technical effects of the aforementioned computer-readable storage product provided in the fifth aspect of this embodiment can be found in the titanium alloy welding fault monitoring method described in the first aspect, and will not be repeated here.

[0052] The sixth aspect of this embodiment provides a computer program product, including a computer program or instructions, which, when executed by a computer, implements the titanium alloy welding fault monitoring method as described in the first aspect. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.

[0053] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for monitoring welding faults in titanium alloys, characterized in that, include: The system receives real-time visual data from a vision sensor, spectral data from a spectral sensor, power data from a power sensor, and temperature data from an infrared temperature sensor. The vision sensor is positioned directly in front of the titanium alloy weld pool to capture real-time visual data reflecting the image of the weld pool, the behavior of the weld pool apertures, and / or the plume morphology. The spectral sensor is also positioned directly in front of the titanium alloy weld pool to capture real-time spectral data reflecting the spectral composition of the weld. The power sensor is connected to a titanium alloy welding power source to capture real-time power data reflecting the magnitude of the welding output current and / or the level of the welding arc voltage. The infrared temperature sensor is also positioned directly in front of the titanium alloy weld pool to capture real-time temperature data reflecting the temperature field distribution in the weld pool area and the heat-affected zone of the weld pool. Based on a unified time scale, the visual acquisition data, the spectral acquisition data, the power acquisition data, and the temperature acquisition data are processed in real time for time synchronization to obtain visual data, spectral data, power data, and temperature data with synchronized acquisition time. The visual data, spectral data, power data, and temperature data are processed in real time to extract multidimensional features to obtain the current multidimensional feature vector time series data. The multidimensional feature vector time series data contains multiple multidimensional feature vectors arranged in time sequence to reflect the instantaneous health state of titanium alloy welding. The multidimensional feature vector time series data is imported in real time into the titanium alloy welding fault identification model pre-trained based on the time series classification model to obtain the current titanium alloy welding fault identification result. Based on the titanium alloy welding fault identification results, the current titanium alloy welding decision is output in real time.

2. The titanium alloy welding fault monitoring method according to claim 1, characterized in that, After obtaining the time-synchronized visual data, spectral data, power data, and temperature data, the method further includes: The images in the visual data are subjected to noise reduction, contrast enhancement, and / or region of interest extraction to obtain new visual data. And / or, the current signal and / or voltage signal in the power supply data are filtered to obtain new power supply data.

3. The titanium alloy welding fault monitoring method according to claim 1, characterized in that, The multidimensional feature vector includes morphological features extracted from the visual data, spectral features extracted from the spectral data, electrical features extracted from the power data, and thermal imaging features extracted from the temperature data. The morphological features include the molten pool area, molten pool aspect ratio, molten pool trailing angle, molten pool orifice diameter, molten pool orifice fluctuation rate, plume area, and / or plume tremor frequency. The spectral features include the relative intensity ratio of characteristic spectral lines for each pair of elements in hydrogen, oxygen, and nitrogen. The electrical features include the mean current, current standard deviation, mean voltage, and / or voltage standard deviation. The thermal imaging features include the highest temperature at the molten pool tip and / or the temperature gradient of the molten pool heat-affected zone.

4. The titanium alloy welding fault monitoring method according to claim 1, characterized in that, The titanium alloy welding fault identification model was pre-trained in the following manner: Historical multidimensional feature vector time series data are collected and used as model input, and titanium alloy welding fault labels based on expert experience / post-weld inspection results are obtained as model output. Then, the model input and model output are used as sample data. The titanium alloy welding fault labels are divided into "normal", "fault precursor" and "fault in progress" labels. Using all the sample data, the temporal classification model constructed based on a long short-term memory network or a temporal convolutional network is trained to obtain the titanium alloy welding fault identification model.

5. The titanium alloy welding fault monitoring method according to claim 1, characterized in that, When the titanium alloy welding fault identification result includes the probability of being labeled "normal", "precursor to fault", and "fault in progress", the current titanium alloy welding decision is output in real time based on the titanium alloy welding fault identification result, including: If the probability of the "normal" label in the titanium alloy welding fault identification result is greater than the preset first probability threshold, the current first titanium alloy welding decision, which indicates that no adjustment action is required, is output in real time. And / or, if the probability of the "fault precursor" label in the titanium alloy welding fault identification result is greater than the preset second probability threshold, then the current second titanium alloy welding decision is output in real time and used to indicate the triggering of the audible and visual alarm and / or to send parameter fine-tuning suggestions to the welding controller. And / or, if the probability of the "fault in progress" label in the titanium alloy welding fault identification result is greater than the preset third probability threshold, then the current third titanium alloy welding decision, which is used to indicate sending an emergency stop command or a current reduction command to the welding machine, is output in real time.

6. A titanium alloy welding fault monitoring device, characterized in that, It includes a multi-source data receiving unit, a time synchronization processing unit, a multi-dimensional feature extraction unit, a welding fault identification unit, and a welding decision output unit, which are connected in sequence. The multi-source data receiving unit is used to receive in real time visual acquisition data from a vision sensor, spectral acquisition data from a spectral sensor, power acquisition data from a power sensor, and temperature acquisition data from an infrared temperature sensor. The vision sensor is positioned directly in front of the titanium alloy weld pool to capture the visual acquisition data in real time, reflecting the image of the weld pool, the behavior of the weld pool apertures, and / or the plume morphology. The spectral sensor is also positioned directly in front of the titanium alloy weld pool to capture the spectral acquisition data in real time, reflecting the spectral composition of the weld. The power sensor is connected to the titanium alloy welding power supply to capture the power acquisition data in real time, reflecting the magnitude of the welding output current and / or the level of the welding arc voltage. The infrared temperature sensor is also positioned directly in front of the titanium alloy weld pool to capture the temperature acquisition data in real time, reflecting the temperature field distribution in the weld pool area and the heat-affected zone of the weld pool. The time synchronization processing unit is used to perform real-time time synchronization processing on the visual acquisition data, the spectral acquisition data, the power acquisition data and the temperature acquisition data based on a unified time scale, so as to obtain visual data, spectral data, power data and temperature data with synchronized acquisition time. The multidimensional feature extraction unit is used to perform multidimensional feature extraction processing on the visual data, the spectral data, the power data and the temperature data in real time to obtain the current multidimensional feature vector time series data. The multidimensional feature vector time series data contains multiple multidimensional feature vectors arranged in time sequence to reflect the instantaneous health state of titanium alloy welding. The welding fault identification unit is used to import the multi-dimensional feature vector time series data into the titanium alloy welding fault identification model pre-trained based on the time series classification model in real time, so as to obtain the current titanium alloy welding fault identification result. The welding decision output unit is used to output the current titanium alloy welding decision in real time based on the titanium alloy welding fault identification result.

7. A titanium alloy welding fault monitoring system, characterized in that, It includes a visual sensor, a spectral sensor, a power sensor, an infrared temperature sensor, and a monitoring device, wherein the visual sensor, the spectral sensor, the power sensor, and the infrared temperature sensor are respectively communicatively connected to the monitoring device; The visual sensor is configured to be positioned directly in front of the titanium alloy welding molten pool to capture visual acquisition data in real time, reflecting the image of the molten pool, the behavior of the molten pool orifice, and / or the plume morphology, and to transmit the visual acquisition data to the monitoring device in real time. The spectral sensor is configured to be positioned directly opposite the titanium alloy weld pool to capture spectral data reflecting the spectral composition of the weld in real time, and to transmit the spectral data to the monitoring device in real time. The power sensor is used to connect to the titanium alloy welding power source to capture power acquisition data in real time, which reflects the magnitude of the welding output current and / or the level of the welding arc voltage, and to transmit the power acquisition data to the monitoring device in real time. The infrared temperature sensor is used to be positioned directly in front of the titanium alloy welding molten pool to capture temperature data in real time, reflecting the temperature field distribution of the molten pool area and the heat-affected zone of the molten pool, and to transmit the temperature data to the monitoring device in real time. The monitoring equipment is used to perform the titanium alloy welding fault monitoring method as described in any one of claims 1 to 5.

8. A computer device, characterized in that, The device includes a storage module, a processing module, and a transceiver module that are sequentially connected in communication. The storage module is used to store a computer program, the transceiver module is used to send and receive messages, and the processing module is used to read the computer program and execute the titanium alloy welding fault monitoring method as described in any one of claims 1 to 5.

9. A computer-readable storage product, characterized in that... The computer-readable storage product stores instructions that, when executed on a computer, perform the titanium alloy welding fault monitoring method as described in any one of claims 1 to 5.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or the instructions are executed by the computer, they implement the titanium alloy welding fault monitoring method as described in any one of claims 1 to 5.

Images