Stability closed-loop control method and system for double-wire indirect electric arc welding with electric signal feedback

Abstract

The application discloses a double-wire indirect electric arc welding stability closed-loop control method and system of electric signal feedback, relates to the technical field of metal material welding, acquires process data and control variables of a bypass coupled double-wire indirect electric arc welding ideal state, establishes a process database based on a mapping relationship of the process data and the control variables, collects electric signals in a welding process in real time, extracts features of the electric signals, and obtains electric characteristic quantities; based on a current process parameter combination, corresponding ideal electric characteristic quantities are obtained in the process database, based on the ideal electric characteristic quantities and the electric characteristic quantities, a deviation is obtained; the deviation is input to a neural network controller for adjustment control, an output wire feeding speed adjustment amount is output, and then an adjustment signal is obtained; based on the adjustment signal, the wire feeding speed is adjusted, stable combustion of a composite electric arc is realized, a process window is widened, and welding quality and process reliability are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of metal material welding, in particular to a double-wire indirect arc welding stability closed-loop control method and system based on electric signal feedback. BACKGROUND

[0002] The statements in this section merely provide background information related to the present disclosure and do not necessarily constitute the prior art.

[0003] Welding is an indispensable key connection technology in modern manufacturing industry. In order to pursue higher production efficiency and better welding quality, various efficient welding methods have emerged. Among them, the bypass coupled double-wire indirect arc welding (BC-TWIAW) as a new type of efficient and low heat input welding process, shows great application potential. The core principle of this process lies in its unique double-arc coupling structure and circuit design. Through a specially designed double-output welding power supply or two synchronously coordinated power supplies, a main indirect arc is ignited between the positive and negative welding wires, and a bypass direct arc is ignited between the positive welding wire and the workpiece. The total welding current is cleverly divided into the main indirect arc current and the bypass direct arc current. This design brings significant technical advantages: high deposition efficiency compared with single-wire arc welding, the energy of the main indirect arc is concentrated in the melting welding wire, thereby achieving a very high welding wire melting speed and deposition efficiency; low heat input to the base metal, as only the bypass direct arc current actually flows into the workpiece, by accurately controlling the size, the overall heat input to the workpiece can be significantly reduced under the premise of ensuring good fusion, effectively controlling the welding deformation and reducing the width of the heat-affected zone. Therefore, the BC-TWIAW process is particularly suitable for additive manufacturing with high deposition efficiency, heat-sensitive materials, and situations requiring high-efficiency surfacing.

[0004] Although the BC-TWIAW process has the above outstanding advantages, its process stability control is a major challenge in actual application, mainly reflected in the following aspects:

[0005] Compared with traditional arc welding, its process window is relatively narrow. The ideal state of this process is that the main indirect arc and the bypass direct arc form a stable and synchronous "composite arc". This state requires high matching accuracy of various process parameters, especially the matching of the wire feeding speeds of the two welding wires.

[0006] The current BC-TWIAW process lacks self-adjusting ability, because the current commonly used welding system, its wire feeding mechanism is usually open-loop control, that is, the wire is fed according to the preset speed (generally the speed of the positive welding wire is fixed, and the negative welding wire speed is adjusted). When arc length fluctuation, droplet transfer instability and other disturbances occur during welding, the system cannot automatically adjust the wire feeding speed to compensate for these changes.

[0007] If the welding process is unstable, for example, if the double-wire wire feeding speed and melting speed are mismatched, or affected by other disturbances, the synchronous combustion state of the composite arc will be destroyed. At this time, two undesirable phenomena are likely to occur: one is frequent short circuit, mainly caused by short circuit between the positive electrode wire and the workpiece due to excessive wire feeding speed or short arc length, which will cause severe current fluctuations and a large amount of spatter; the other is that the originally stable combustion of the composite arc becomes an alternating combustion of two arcs, the main circuit indirect arc and the bypass direct arc cannot exist synchronously, showing an alternating combustion state of this disappearing and that appearing, resulting in chaotic droplet transfer and unstable energy input, ultimately leading to poor weld formation and difficult to guarantee internal quality.

[0008] In the prior art, the operator mainly relies on experience to set parameters offline, which is difficult to cope with dynamic changes in the welding process. Although some closed-loop control schemes based on visual sensing are feasible, they have problems such as expensive equipment, complex algorithms, susceptibility to arc light and smoke interference, limited response speed, etc., and are not completely suitable for the welding environment of BC-TWIAW with high speed and strong arc light. SUMMARY

[0009] To overcome the shortcomings of the prior art, the present application provides a dual-wire indirect arc welding stability closed-loop control method and system based on electrical signal feedback, aiming to realize stable combustion of the composite arc, widen the process window, and improve welding quality and process reliability.

[0010] To achieve the above-mentioned purpose, one or more embodiments of the present application provide the following technical solutions:

[0011] In a first aspect, the present application provides a dual-wire indirect arc welding stability closed-loop control method based on electrical signal feedback, comprising:

[0012] Obtaining process data and control variables of the ideal state of bypass coupled dual-wire indirect arc welding, and establishing a process database based on the mapping relationship between the process data and the control variables;

[0013] Real-time acquisition of electrical signals during the welding process, and feature extraction of the electrical signals to obtain electrical characteristic quantities;

[0014] Obtaining corresponding ideal electrical characteristic quantities in the process database based on the current process parameter combination, and obtaining deviations based on the ideal electrical characteristic quantities and the electrical characteristic quantities;

[0015] Inputting the deviations into a neural network controller for adjustment and control, outputting a wire feeding speed adjustment amount, and then obtaining an adjustment signal, and adjusting the wire feeding speed based on the adjustment signal.

[0016] Further technical solutions, under a variety of process parameter combinations, welding experiments are performed to determine the ideal state of welding, specifically:

[0017] According to each process parameter combination, a welding experiment is designed, and an electric signal in a welding process is collected;

[0018] Through a multi-dimensional quality evaluation system, the weld quality is evaluated, and a comprehensive quality score of the weld is obtained;

[0019] The electric signal is processed, and process characteristics are extracted, and then a feature database is constructed;

[0020] According to the comprehensive quality score of the weld, all weld samples are sorted from high to low, and a predetermined number of weld samples are selected as an ideal sample set;

[0021] Based on the ideal sample set, a regression model is trained, and an ideal feature range is analyzed by using a model explanation tool, and then an ideal state of welding is obtained.

[0022] Further technical solutions, the process data includes process parameter combination and ideal electric characteristic quantity, and the control variable is the wire feeding speed.

[0023] Further technical solutions, the electric signal includes voltage signal and current signal of the main circuit arc and the bypass arc, the electric characteristic quantity is calculated based on the voltage signal and the current signal, and the electric characteristic quantity includes bypass arc short-circuit frequency, current waveform standard deviation and dynamic resistance change rate.

[0024] Further technical solutions, the deviation of each electric characteristic quantity from the ideal electric characteristic quantity is calculated respectively, and the bypass arc short-circuit deviation, the current standard deviation and the resistance change rate deviation are obtained.

[0025] Further technical solutions, the neural network controller adopts a multi-input single-output model, receives the deviation and the current wire feeding speed, and outputs a wire feeding speed adjustment amount.

[0026] Further technical solutions, based on the adjustment signal, a new wire feeding speed is obtained, and the new wire feeding speed is the sum of the current wire feeding speed and the adjustment signal.

[0027] In a second aspect, the present application provides a double-wire indirect arc welding stability closed-loop control system with electric signal feedback, comprising:

[0028] The database construction module is configured to: acquire process data and control variables of an ideal state of bypass coupling double-wire indirect arc welding, and establish a process database based on a mapping relationship between the process data and the control variables;

[0029] The feature extraction module is configured to: collect an electric signal in a welding process in real time, and extract features from the electric signal to obtain electric characteristic quantities;

[0030] a deviation calculation module configured to: obtain a corresponding ideal electrical characteristic quantity in the process database based on the current process parameter combination, and obtain a deviation based on the ideal electrical characteristic quantity and the electrical characteristic quantity;

[0031] a control adjustment module configured to: input the deviation into a neural network controller for adjustment control, output a wire feeding speed adjustment amount, and then obtain an adjustment signal, and adjust the wire feeding speed based on the adjustment signal.

[0032] In a third aspect, the present application provides a computer readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the steps in the dual-wire indirect arc welding stability closed-loop control method of electrical signal feedback.

[0033] In a fourth aspect, the present application provides a computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor executing the program to implement the steps in the dual-wire indirect arc welding stability closed-loop control method of electrical signal feedback.

[0034] The above one or more technical solutions have the following beneficial effects:

[0035] The present application can monitor the stability of the welding process in real time and automatically adjust the key process parameters (mainly the wire feeding speed), thereby realizing stable combustion of the composite arc, widening the process window, and improving the welding quality and process reliability.

[0036] The present application significantly improves the process stability through closed-loop control, and the system can automatically suppress arc disturbance, so that the main arc and the bypass arc can maintain a stable ideal state of synchronous stable combustion for a long time.

[0037] The present application has self-adaptive adjustment capability, reduces the harsh requirements for the accuracy of initial parameter setting, and enables the process to operate stably in a wider range of current, voltage and wire feeding speed, thereby widening the process window.

[0038] The stable welding process of the present application directly leads to uniform and beautiful weld formation, reduces welding defects such as spatter, incomplete fusion, and porosity, and ensures the high consistency of batch product welding quality.

[0039] The automation and intelligentization of the system of the present application greatly reduces the adjustment burden of welders, and even ordinary skill level operators can complete high-quality welding tasks, thereby reducing the dependence on operator skills.

[0040] The feedback path of the electric signal is short and fast, and is not disturbed by smoke, arc light and other field environment, so that the real-time and reliability of the control are ensured. BRIEF DESCRIPTION OF DRAWINGS

[0041] The accompanying drawings, which form a part of the specification, are included to provide a further understanding of the application and are incorporated herein for explanation by referring to the exemplary embodiments thereof.

[0042] Figure 1 is a flow chart of the double-wire indirect electric arc welding stability closed-loop control method of the electric signal feedback of the embodiment of the application;

[0043] Figure 2 is a schematic diagram of the double-wire indirect electric arc welding stability closed-loop control system of the electric signal feedback of the embodiment of the application, wherein, 101-welding torch, 102-negative electrode welding wire, 103-positive electrode welding wire, 104-welding power source, 105-workpiece, 106-double-channel wire feeder, 200-central processing and control unit, 201-signal sensing and acquisition unit, 202-feature extraction module, 203-deviation calculation module, 204-control adjustment module;

[0044] Figure 3 is a real-time current-voltage change diagram before and after the application of the closed-loop feedback system of the embodiment of the application, wherein, (a) is a real-time current-voltage change diagram before the application of the closed-loop feedback system, and (b) is a real-time current-voltage change diagram after the application of the closed-loop feedback system;

[0045] Figure 4 is a weld forming comparison diagram before and after the optimization of the embodiment of the application, wherein, (a) is a weld forming diagram before the optimization, and (b) is a weld forming diagram after the optimization. DETAILED DESCRIPTION

[0046] It should be noted that the following detailed description is exemplary in nature and is intended to provide further description of the application. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0047] It is to be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups thereof.

[0048] The embodiments in the application and the features in the embodiments can be combined with each other without conflict.

[0049] Terminology explanation:

[0050] Bypass-Coupled Twin-Wire Indirect Arc Welding (BC-TWIAW): A new type of welding process, whose circuit structure contains a main indirect arc burning between the positive and negative twin wires, and a bypass direct arc burning between the positive welding wire and the workpiece.

[0051] Main Indirect Arc: The arc formed between the ends of the positive and negative welding wires, whose current (I main ) is mainly used for efficient melting of the two welding wires.

[0052] Bypass Direct Arc: The arc formed between the positive welding wire and the workpiece, whose current (I bypass ) flows through the workpiece to provide the necessary heat input for the molten pool and base material.

[0053] Composite Arc: The ideal state of synchronous and stable combustion of the main indirect arc and the bypass direct arc, at which the welding process stability is best and the weld formation quality is optimal.

[0054] Electrical Characteristic Quantities: Quantitative indicators extracted from the current and / or voltage waveforms of the welding process, which can represent the stability of droplet transfer, arc shape, and process stability. For example, short circuit frequency, current waveform standard deviation, dynamic resistance change rate, etc.

[0055] Process Database: A set of pre-established mapping relationships that store the optimal combination of electrical characteristic quantities corresponding to different process parameters (such as welding current, voltage, materials, etc.) under ideal welding conditions.

[0056] Example 1

[0057] As shown in Figure 1 , the present embodiment discloses a dual-wire indirect arc welding stability closed-loop control method based on electrical signal feedback, which comprises the following steps:

[0058] S1: Obtain the process data and control variables of the ideal welding state, and establish a process database based on the mapping relationship between the process data and the control variables;

[0059] In this embodiment, welding tests are performed under a variety of typical process parameter combinations (such as material type, plate thickness, shielding gas, preset total current, voltage, etc.). The "ideal state" of the most stable welding process and the best weld formation is determined by means of high-speed photography, weld appearance analysis, etc.

[0060] The specific steps for determining the ideal state are as follows:

[0061] (1) Design experiments and collect data.

[0062] A relatively wide range of process parameters is determined according to the welding method, material, and thickness, and the key parameters include welding current, arc voltage, welding speed, wire feed speed, gas flow, and welding gun posture.

[0063] Orthogonal experiments, response surface methods, or full-factor experimental design are used to systematically change the process parameters, covering the entire state space from poor to good. Each experiment, i.e., each weld, is given a unique number.

[0064] When welding each sample, process data and metadata records are collected simultaneously. The process data is the time series waveform of the arc voltage and welding current during the welding process, and the data file is strictly matched with the number when recording. The metadata records are all the process parameters (such as current, voltage setting value, speed, etc.) and environmental conditions of each experiment.

[0065] (2) Quantitatively evaluate the weld quality. The optimal weld formation is determined by establishing a multi-dimensional quality evaluation system.

[0066] Macroscopic appearance evaluation: The geometric dimensions of each weld, i.e., weld reinforcement, weld width, penetration, and width, are measured using a 3D laser profile scanner or a microscope. The formation coefficient, i.e., the ratio of penetration to width and the ratio of reinforcement to width, is calculated based on the measured geometric dimensions. The optimal range of the formation coefficient is preset, and the calculated formation coefficient within the optimal range is considered as good macroscopic appearance, with a value of 1, otherwise, with a value of 0. The width , height , and shape coefficient are quantified.

[0067] Surface quality and defect evaluation: The weight of the collected plates around the workpiece before and after welding is weighed to calculate the spatter rate. The uniformity of the weld surface and the continuity of the corrugation are evaluated, and the image processing algorithm can be used to calculate the uniformity index of the surface texture. The length, depth, and number of defects such as undercut, incomplete penetration, and surface porosity are checked and recorded. The spatter rate , surface roughness , and undercut depth are quantified.

[0068] Internal quality evaluation (optional): X-ray non-destructive testing of the weld to detect internal porosity, slag inclusion, cracks, etc. Quantify the porosity area , crack length . This evaluation is a non-mandatory step.

[0069] Establish a comprehensive quality index. Normalize all the above quantitative indicators so that indicators of different units can be compared with each other. For defect indicators, the smaller the value, the better, for forming indicators, there is an optimal range.

[0070] According to the actual welding requirements, assign weights to each standardized indicator. For example, for load-bearing structural parts, the weight of penetration and internal defect indicators is very high, and for appearance parts, the weight of forming coefficient is higher.

[0071] The comprehensive quality score is represented as:

[0072]

[0073] wherein , , , are the weights of the normalized width, height, spatter rate and porosity area, , , , are the normalized width, height, spatter rate and porosity area scores. The formula of the comprehensive quality score only selects four indicators, and in actual application, quantitative indicators and corresponding weights can be selected according to actual welding requirements, without specific limitation.

[0074] (3) Extract process features. Process data collected, i.e. arc voltage and welding current are processed: first, remove the unstable part of the arc and the arc stage from each weld's electrical signal data, and only keep the steady-state welding section for analysis; calculate the feature quantities of the steady-state welding section of arc voltage and welding current data, including bypass short-circuit frequency, main circuit current standard deviation, bypass dynamic resistance change rate; based on the feature dimension and the feature quantity, construct a feature database.

[0075] (4) Determine the ideal state of the feature space. Select ideal sample set: according to the comprehensive quality score, sort all weld samples from high to low, select a certain number of samples with the highest comprehensive quality score, and define them as the ideal sample set.

[0076] Based on the ideal sample set, the regression model (such as random forest, gradient boosting tree, etc.) is trained to predict the comprehensive quality score through the characteristic quantity. The SHAP model is used to analyze the best value range of the characteristic quantity, that is, the ideal characteristic range. The weld in the ideal characteristic range is in the ideal state.

[0077] (5) Verification and closed-loop control. According to the determined ideal characteristic range, adjust the welding process parameters and conduct new welding experiments, the goal is to make the process characteristics of the new experiment fall within the ideal characteristic range; then evaluate the quality of the new weld and verify whether the comprehensive quality score is high, if so, the defined ideal state is reliable.

[0078] Synchronize high-speed acquisition and record the waveform data of the main arc current , bypass arc current and corresponding voltage , under the ideal state. From these data, the corresponding ideal electrical characteristic quantity (such as ideal short circuit frequency , ideal current standard deviation , etc.) and the optimal wire feeding speed at this time are calculated.

[0079] The process data includes process parameter combination and ideal electrical characteristic quantity, and the control variable is the optimal wire feeding speed. The mapping relationship between “process parameter combination” and “ideal electrical characteristic quantity” and “optimal wire feeding speed” is stored in the database, and the process database is established.

[0080] S2: Real-time acquisition of electrical signals during welding and feature extraction of electrical signals to obtain electrical characteristic quantities;

[0081] In this embodiment, during welding, the high-frequency sampling module is used to real-time and synchronously acquire the electrical signals (including current and voltage signals) of the main arc and bypass arc, and obtain real-time waveform data, i.e. main arc current waveform data , bypass arc current waveform data , main arc voltage waveform data , and bypass arc voltage waveform data . The sampling frequency is preferably above 10 kHz to ensure that the details of the dynamic process such as droplet transfer are captured.

[0082] The acquired real-time waveform data is sent to the signal processing unit, and the electrical characteristic quantities are calculated in real time within a preset time window (for example, 100 ms). The electrical characteristic quantities are preferably one or a combination of the following:

[0083] (1) Bypass arc short circuit frequency : The number of times the bypass arc voltage is below a preset threshold (such as 5V) per unit time is counted, which directly reflects the contact between the welding wire and the workpiece.

[0084] The short-circuit frequency directly quantifies the number of times the positive electrode welding wire and the workpiece make physical contact and cause the arc to extinguish per unit time. Too high a frequency means that the wire feeding speed is too fast relative to the melting speed, the welding wire frequently inserts into the molten pool, resulting in more spatter and an unstable arc; too low a frequency may lead to insufficient penetration or arc drift and arc breakage.

[0085] The bypass arc short-circuit frequency is calculated based on the bypass arc voltage waveform data, specifically as follows:

[0086] 1) Set a short circuit judgment threshold (voltage threshold). When the bypass arc voltage is lower than this threshold, the arc is judged to be in a short circuit state.

[0087] 2) Initialize the short-circuit count counter and status flags.

[0088] 3) Iterate through multiple bypass voltage data points within the time window. If the bypass arc voltage is lower than the short circuit judgment threshold, identify the start time of the short circuit state, increment the short circuit count counter, and update the status flag to the short circuit state.

[0089] 4) Continue to identify the recovery of the arc state. If the bypass arc voltage is greater than or equal to the short circuit judgment threshold, it means that the arc has resumed combustion and is in the arc state. Change the status flag to the arc state.

[0090] 5) After traversing all data points, the total number of short circuits within the time window is obtained, and then the bypass arc short circuit frequency is obtained.

[0091] (2) Standard deviation of current waveform ( ): Calculate within the time window or The standard deviation of the current characterizes the degree of fluctuation in the current; the smaller the fluctuation, the more stable the electric arc.

[0092] In welding electrical signal analysis, the standard deviation of the current waveform reflects the degree of fluctuation of the current signal around its average value. A small standard deviation indicates that the current value is relatively concentrated and the fluctuation is small, while a large standard deviation indicates that the current value fluctuates violently, corresponding to undesirable conditions such as unstable arc, irregular droplet transfer, or frequent short circuits.

[0093] The standard deviation of the current waveform is calculated based on the main arc current waveform data or the bypass arc current waveform data. Taking the standard deviation of the main arc current waveform data as an example, the specific steps are as follows:

[0094] 1) Calculate the arithmetic mean of all current data points within the time window.

[0095] 2) Calculate the square of the difference between each current data point and the arithmetic mean, and add them all together to get the sum of squares of deviations.

[0096] 3) Based on the sum of squares of deviations and the total number of current data points, the variance is obtained, and then the standard deviation is obtained.

[0097] (3) Dynamic resistance change rate (R) ): The dynamic resistance of the arc is calculated by , where is the voltage of the arc, and is the current of the arc, and the change rate is calculated. The dynamic resistance of a stable arc changes relatively slowly. The dynamic resistance change rate characterizes the speed of the physical state change of the arc. A small change rate indicates that the resistance value of the arc is relatively stable, and a large change rate indicates that the resistance of the arc changes rapidly.

[0098] The dynamic resistance change rate is calculated based on paired voltage and current data, such as bypass arc voltage waveform data and bypass arc current waveform data. The specific steps are as follows:

[0099] 1) Traverse each pair of synchronous sampling points in the time window, and calculate the instantaneous dynamic resistance. Considering the case where the current value is 0 or close to 0 (such as the zero-crossing point at the moment of short circuit or arc breaking), in order to avoid division by zero error, set a current lower limit. When the current data point is less than the current lower limit, the instantaneous dynamic resistance is equal to the set replacement value.

[0100] 2) According to the obtained resistance sequence, the change rate between adjacent points is calculated, and the change rate sequence is obtained; then the average value of the absolute value of the change rate sequence is calculated, and the dynamic resistance change rate is obtained, which represents a single numerical value of the stability of the entire time window.

[0101] S3: Based on the current process parameter combination, the corresponding ideal electrical characteristic quantity is obtained in the process database, and based on the ideal electrical characteristic quantity and the electrical characteristic quantity, the deviation is obtained.

[0102] In this embodiment, according to the current process parameter combination, the corresponding ideal electrical characteristic quantity is retrieved in the process database, and the real-time calculated electrical characteristic quantity is compared with the ideal electrical characteristic quantity to calculate the deviation.

[0103] The deviation of each electrical characteristic quantity from its ideal electrical characteristic quantity is calculated respectively to obtain the bypass arc short circuit deviation

[0104] , the current standard deviation deviation , and the resistance change rate deviation . As

[0105] , where is the bypass arc short circuit frequency, and is the ideal short circuit frequency.

[0106] ​S4: input the deviation to a controller for adjustment control, output an adjustment signal, adjust the wire feeding speed based on the adjustment signal.

[0107] In this embodiment, the deviation is input to a controller. The controller outputs an adjustment signal according to the size and positive and negative of the deviation . For example, if the deviation is positive, indicating that the actual short-circuit frequency is too high (the wire feeding is too fast), the controller outputs a negative adjustment signal; otherwise.

[0108] The controller uses a neural network controller, which determines the corresponding accurate adjustment signal in the current specific state by learning the complex and nonlinear mapping relationship.

[0109] The neural network controller includes two stages, an offline training stage and an online application stage.

[0110] (I) Offline training stage

[0111] Collect training data, i.e. a large number of pairs of welding process states and optimal adjustment actions (ideal wire feeding speed adjustment amount), record a large number of data pairs through a large number of experiments, and construct a training set. The state includes bypass arc short-circuit deviation , current standard deviation , resistance rate of change deviation , and current wire feeding speed.

[0112] Construct a neural network, including an input layer, a hidden layer, and an output layer, the input layer is provided with multiple branches in parallel, respectively receiving input states.

[0113] Train the constructed neural network based on the training set, i.e. the neural network receives the training data to obtain a predicted value, calculates the loss based on the deviation between the predicted value and the ideal value, and uses the back propagation algorithm and the optimizer to fine-tune the weights and biases of all hidden layers in the network based on the loss to obtain a trained neural network.

[0114] (II) Online application stage

[0115] Deploy the trained neural network to an actual welding controller, input each deviation and the current wire feeding speed to the trained neural network for processing, and output the wire feeding speed adjustment amount. The wire feeding speed adjustment amount output by the neural network is an accurate, nonlinearly calculated, and multi-factor considered adjustment value.

[0116] The controller obtains an adjustment signal from the wire feeding speed adjustment amount , applies the adjustment signal to the corresponding wire feeder drive system to adjust the wire feeding speed. The new wire feeding speed is represented as:

[0117]

[0118] wherein, is the new wire feed speed, is the current wire feed speed.

[0119] Through the above technical solution, the welding process is a multivariate coupling process, and the current wire feed speed, various deviations and other input parameters will simultaneously affect the welding process, and the neural network controller is adopted to realize the multi-input and output control.

[0120] By adjusting the wire feed speed of the positive and / or negative electrode, the arc length and the droplet transfer behavior are changed, so that the electrical characteristic quantity returns to the vicinity of the ideal value (ideal characteristic range), and the dynamic stability of the welding process is realized. Here, there is also a key point that the two sets of double-wire feeding mechanisms are changed from the traditional wire feeding mechanism to each being equipped with a precision wire feeder driven by a brushless servo motor. The motor has extremely high response speed and control accuracy, and can realize hundreds of times of wire feeding direction switching per minute, meeting the requirement of real-time adjustment of the wire feed speed.

[0121] S2 to S4 are repeated to form a closed loop control until the welding is completed.

[0122] In summary, the application discards the visual signal which is easy to be disturbed, and innovatively uses the electrical signal (main and bypass current / voltage) generated in the welding process itself as the feedback source of the closed loop control. The electrical signal directly reflects the physical state of the arc, has high signal-to-noise ratio and fast response speed.

[0123] The application proposes and defines the "electrical characteristic quantity" capable of accurately characterizing the stability of the BC-TWIAW process, such as the short-circuit frequency, the current waveform standard deviation, the dynamic resistance change rate and the like. These characteristic quantities have strong correlation with the droplet transfer mode and the arc synchronism.

[0124] The application converts the complex real-time analysis problem into a fast table lookup and comparison problem by pre-establishing the "ideal electrical characteristic quantity - optimal process parameter" mapping database, thereby greatly improving the response speed and decision efficiency of the control system.

[0125] The application constructs a complete closed loop control loop, can calculate the electrical characteristic quantity in real time, compare it with the ideal value in the database to generate a deviation, and adjust the wire feed speed in real time and dynamically through the controller (such as PID) according to the deviation, so that the welding process always tends to and remains in the optimal stable state.

[0126] Example one: closed loop control based on bypass arc short circuit frequency

[0127] Welding of 2mm thick low carbon steel sheet, using BC-TWIAW process.

[0128] Process database establishment: Experiments show that the ideal short circuit frequency of bypass direct arc under this condition is 5 Hz ± 1 Hz. When the frequency is too high (> 6 Hz), it indicates that the positive electrode wire feeding is too fast, and spatter is prone to occur; when the frequency is too low (< 4 Hz), it indicates that the arc length is too long, the arc is unstable, and the arc is prone to break. The ideal frequency range is stored in the database.

[0129] The control process is as follows:

[0130] (1) The system collects the bypass arc voltage in real time at a frequency of 20 kHz .

[0131] (2) The short circuit frequency (the number of times the voltage is lower than 5V) in this time window is calculated every 100ms . Suppose that at a certain moment, the calculated = 10 Hz.

[0132] (3) The controller calculates the deviation .

[0133] (4) The PID controller outputs a negative adjustment signal according to the positive deviation, instructing the positive wire feeder to reduce the speed by 5%.

[0134] After the wire feeding speed is reduced, the arc length increases, and the short circuit frequency decreases. In the next control cycle, the system will continue to monitor and fine-tune until stabilizes at about 5 Hz.

[0135] Example 2: Closed-loop control based on the standard deviation of the main arc current

[0136] High deposition efficiency and stable process are required for high-speed surfacing of wear-resistant layers.

[0137] Establishment of process database: Experiments show that when the indirect arc in the main circuit is most stable, the standard deviation of its current waveform reaches a minimum value. The database records the corresponding minimum under different surfacing currents.

[0138] The control process is as follows:

[0139] (1) The system collects the main circuit current in real time .

[0140] (2) The current standard deviation is calculated every 50ms .

[0141] (3) The target of the controller is to minimize ​​It does so by slightly perturbing the relative values of the positive and negative wire feed speeds (e.g., slightly increasing the positive wire speed while slightly decreasing the negative wire speed for one cycle) and observing the trend of

[0142] (4) An optimization algorithm such as hill climbing or gradient descent is used to continuously adjust the relative wire feed speeds of the two wires so that (the real-time monitored and calculated standard deviation of the main arc current) always approaches (the ideal standard deviation of the arc current that would result in the most stable arc burning under the current welding process parameters) in the database, thereby achieving the most stable burning of the main arc.

[0143] Example Three: Fuzzy Logic Control Based on Multi-Feature Fusion

[0144] Welding important aluminum alloy structures requires extremely high process stability and weld quality.

[0145] Process Database Establishment: A more complex database is established that contains not only single feature quantities but also multi-feature quantities (such as bypass short circuit frequency , main arc current standard deviation , bypass dynamic resistance change rate ) and fuzzy relationships with stability states.

[0146] Control Process:

[0147] (1) The system calculates the above three electrical feature quantities in real time.

[0148] (2) A fuzzy logic controller serves as the control decision unit, with its inputs being the deviations of the real-time values of the three feature quantities from the ideal values.

[0149] (3) The fuzzy rule base is pre-set with expert knowledge, such as: 1) Rule 1: IF ( deviation is "positive large") AND ( deviation is "positive medium") THEN (positive wire feed speed "significantly reduced"); 2) Rule 2: IF ( deviation is "zero") AND ( deviation is "positive large") THEN (negative wire feed speed "slightly increased").

[0150] (4) The controller, through fuzzy reasoning, comprehensively judges the current instability type and degree and outputs more accurate and smoother adjustment instructions to the positive and negative wire feeders. This approach can handle more complex working condition changes and achieve more robust control effects.

[0151] Example Four: Closed-Loop Control of Additive Manufacturing Process ​

[0152] The BC-TWIAW process is used for additive manufacturing.

[0153] Establishing the process database: tests show that when the composite arc is stable, the ideal short-circuit frequency of the bypass direct arc is 5 Hz ± 1 Hz under the working condition. The indirect arc current The ideal short-circuit frequency of the bypass direct arc is 5 Hz ± 1 Hz. This ideal frequency range is stored in the database.

[0154] Control process:

[0155] (1) The system collects the bypass arc voltage in real time at a frequency of 20 kHz.

[0156] (2) The number of short circuits (the number of times the voltage is lower than 5 V) in the time window is calculated every 100 ms, and the frequency is converted.

[0157] (3) miscellaneous If at a certain moment, it is calculated that .

[0158] (4) The controller calculates the deviation Error = 10 - 5 = +5 Hz.

[0159] (5) The PID controller outputs a negative adjustment signal according to the positive deviation, instructing the positive wire feeder to reduce the speed by 5%.

[0160] (6) After the wire feeding speed is reduced, the arc length increases, and the short-circuit frequency decreases. In the next control cycle, the system will continue to monitor and fine-tune until stabilize at about 5 Hz.

[0161] Actual welding results: the welding process parameters are initially selected as follows: indirect arc current , indirect arc current , indirect arc and direct arc voltage , positive wire feeding speed , negative wire feeding speed , and protective gas flow . During the welding process, the composite arc is disturbed and becomes turbulent due to the unevenness of the test piece, and the negative wire feeding speed is automatically adjusted to about 0.17 m / s, and the welding spatter is significantly suppressed, and the welding process returns to normal. The actual measured current and voltage waveforms are shown in Figure 3 , (a) is the real-time current and voltage variation diagram before the closed-loop feedback system is applied, and (b) is the real-time current and voltage variation diagram after the closed-loop feedback system is applied; the weld formation is as follows​Figure 4 As shown, (a) is the weld forming diagram before optimization, and (b) is the weld forming diagram after optimization.

[0162] Embodiment Two

[0163] The embodiment discloses a double-wire indirect electric arc welding stability closed-loop control system with electric signal feedback, comprising:

[0164] A database construction module configured to: acquire process data and control variables of bypass coupling double-wire indirect electric arc welding ideal state, and establish a process database based on a mapping relationship between the process data and the control variables;

[0165] A feature extraction module configured to: collect electric signals in real time during welding, and extract features of the electric signals to obtain electric characteristic quantities;

[0166] A deviation calculation module configured to: obtain corresponding ideal electric characteristic quantities in the process database based on a current process parameter combination, and obtain a deviation based on the ideal electric characteristic quantities and the electric characteristic quantities;

[0167] A control adjustment module configured to: input the deviation into a neural network controller for adjustment control, output a wire feeding speed adjustment amount, and then obtain an adjustment signal, and adjust the wire feeding speed based on the adjustment signal.

[0168] As shown in Figure 2 , the system comprises a welding execution subsystem, a signal sensing and collecting unit 201, a central processing and control unit 200, and an execution adjustment unit.

[0169] The welding execution subsystem comprises a welding torch 101, a negative electrode welding wire 102, a positive electrode welding wire 103, a welding power supply 104, a workpiece 105, and a double-channel wire feeder 106. The welding power supply 104 is a double-output power supply, which supplies power to the main circuit arc and the bypass arc respectively.

[0170] The signal sensing and collecting unit 201 comprises electric signal sensors S1 and S2 (such as Hall current sensors, voltage dividing resistors, etc.) arranged in the main circuit and the bypass circuit, which are used to detect signals such as , in real time. The collecting unit is responsible for converting the analog signals into digital signals at a high speed.

[0171] The central processing and control unit 200 is the core of the system, usually implemented by DSP, FPGA or high-performance single-chip microcomputer, which integrates a database construction module (stores the process database), a feature extraction module 202 (receives the collected digital signals and calculates the electrical characteristic quantities in real time), a deviation calculation module 203 (compares the real-time characteristic quantities with the ideal values in the database and calculates the deviation), and a control and adjustment module 204 (built-in PID control algorithm, generates adjustment instructions for the wire feeding speed according to the deviation).

[0172] The execution adjustment unit is the motor driver of the double-channel wire feeder 106 (including the positive wire feeder and the negative wire feeder). It receives the adjustment instructions from the controller and accurately adjusts the rotating speed of one or two wire feeding motors, thereby changing the wire feeding speed.

[0173] Embodiment Three

[0174] The purpose of this embodiment is to provide a computing device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method of embodiment one.

[0175] Embodiment Four

[0176] The purpose of this embodiment is to provide a computer-readable storage medium, a computer-readable storage medium having a computer program stored thereon, wherein the program is executed by a processor to perform the steps of the method of embodiment one.

[0177] The steps and methods involved in the above embodiments three and four correspond to embodiment one, and the specific embodiments can be referred to the relevant description part of embodiment one. The term "computer-readable storage medium" should be understood to include a single medium or multiple media of one or more instruction sets; it should also be understood to include any medium that can store, encode or carry instruction sets for execution by a processor and make the processor execute any method in the present application.

[0178] Those skilled in the art should understand that the above-mentioned modules or steps of the present application can be implemented by a general-purpose computer device, alternatively, they can be implemented by program codes executable by a computing device, so that they can be stored in a storage device for execution by a computing device, or they can be made into individual integrated circuit modules, or a plurality of modules or steps among them can be made into a single integrated circuit module. The present application is not limited to any specific combination of hardware and software.

[0179] The above merely describes the preferred embodiments of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0180] The above merely describes the preferred embodiments of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method of stability closed loop control of a twin wire indirect electric arc welding with electrical signal feedback, characterized in that, The method comprises the following steps: acquiring process data and control variables of an ideal state of bypass coupling double-wire indirect electric arc welding, establishing a process database based on a mapping relationship of the process data and the control variables, wherein the process data comprises a process parameter combination and an ideal electric characteristic quantity, and the control variable is a wire feeding speed; real-time acquisition of an electric signal in a welding process, feature extraction of the electric signal, and obtaining of an electric characteristic quantity, wherein the electric signal comprises a voltage signal and a current signal of a main circuit electric arc and a bypass electric arc, the electric characteristic quantity is calculated based on the voltage signal and the current signal, and the electric characteristic quantity comprises a bypass electric arc short-circuit frequency, a current waveform standard deviation, and a dynamic resistance change rate; based on a corresponding ideal electric characteristic quantity of a current process parameter combination in the process database, obtaining a deviation based on the ideal electric characteristic quantity and the electric characteristic quantity, and respectively calculating a deviation of each electric characteristic quantity from its ideal electric characteristic quantity to obtain a bypass electric arc short-circuit deviation, a current standard deviation deviation, and a resistance change rate deviation; inputting the deviation into a neural network controller for adjustment control, outputting a wire feeding speed adjustment amount, and then obtaining an adjustment signal, adjusting the wire feeding speed based on the adjustment signal, wherein the neural network controller adopts a multi-input single-output model, receives the deviation and a current wire feeding speed, and outputs the wire feeding speed adjustment amount.

2. The dual wire indirect electrical arc welding stability closed loop control method of electrical signal feedback as claimed in claim 1, wherein, Under a plurality of process parameter combinations, welding experiments are performed to determine an ideal state of welding, specifically as follows: designing welding experiments according to each process parameter combination, and acquiring an electric signal in a welding process; evaluating a weld quality through a multi-dimensional quality evaluation system to obtain a comprehensive quality score of the weld; processing the electric signal and extracting process features to construct a feature database; sorting all weld samples from high to low according to the comprehensive quality score of the weld, and selecting a preset number of weld samples as an ideal sample set; training a regression model based on the ideal sample set, and analyzing an ideal feature range by using a model interpretation tool to obtain an ideal state of welding.

3. The dual wire indirect electrical arc welding stability closed loop control method of claim 1, wherein, A new wire feeding speed is obtained based on the adjustment signal, and the new wire feeding speed is the sum of the current wire feeding speed and the adjustment signal.

4. A dual wire indirect electric arc welding stability closed loop control system with electrical signal feedback, characterized in that, The method comprises the following steps: a database construction module configured to acquire process data and control variables of an ideal state of bypass coupling double-wire indirect electric arc welding, and establish a process database based on a mapping relationship of the process data and the control variables, wherein the process data comprises a process parameter combination and an ideal electric characteristic quantity, and the control variable is a wire feeding speed; a feature extraction module configured to real-time acquisition of an electric signal in a welding process, feature extraction of the electric signal, and obtaining of an electric characteristic quantity, wherein the electric signal comprises a voltage signal and a current signal of a main circuit electric arc and a bypass electric arc, the electric characteristic quantity is calculated based on the voltage signal and the current signal, and the electric characteristic quantity comprises a bypass electric arc short-circuit frequency, a current waveform standard deviation, and a dynamic resistance change rate; a deviation calculation module configured to obtain a corresponding ideal electric characteristic quantity based on a current process parameter combination in the process database, obtain a deviation based on the ideal electric characteristic quantity and the electric characteristic quantity, and respectively calculate a deviation of each electric characteristic quantity from its ideal electric characteristic quantity to obtain a bypass electric arc short-circuit deviation, a current standard deviation deviation, and a resistance change rate deviation; The control adjustment module is configured to input the deviation into a neural network controller for adjustment control, output a wire feeding speed adjustment amount, and then obtain an adjustment signal, and adjust the wire feeding speed based on the adjustment signal, wherein the neural network controller adopts a multi-input single-output model, receives the deviation and a current wire feeding speed, and outputs the wire feeding speed adjustment amount.

5. A computer-readable storage medium having stored thereon a computer program, characterized in that, The program, when executed by a processor, implements the steps in the method for stabilizing closed-loop control of a double-wire indirect electric arc welding with electrical signal feedback according to any one of claims 1-3.

6. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The program, when executed by a processor, implements the steps in the method for stabilizing closed-loop control of a double-wire indirect electric arc welding with electrical signal feedback according to any one of claims 1-3.

Images