Search

CN-121988832-A - LabView-based welding penetration control method and system

CN121988832ACN 121988832 ACN121988832 ACN 121988832ACN-121988832-A

Abstract

The invention discloses a LabView-based welding penetration control method and a LabView-based welding penetration control system, and relates to the technical field of welding process control. The method comprises the steps of collecting welding voltage signals in a welding process, carrying out data format conversion, noise reduction treatment and frequency domain analysis on the welding voltage signals by using a LabView processor, extracting molten pool oscillation frequency, comparing the molten pool oscillation frequency with critical penetration frequency in an expert database, judging the penetration state of the current welding process, and controlling a welding system and/or a motion control system to execute corresponding control actions according to a judging result so as to adjust welding heat input and realize penetration state closed-loop control. The noise reduction treatment adopts a self-adaptive multi-scale related weighted generalized threshold wavelet noise reduction method, and the oscillation frequency of the molten pool is obtained through peak group duty ratio analysis. The invention can realize the on-line identification and real-time regulation of the welding penetration state and has the advantages of strong anti-interference capability, stable feature extraction and high control precision.

Inventors

  • LIAO BAOYI
  • CHEN FENGDAN
  • CHEN HAN
  • CHEN LEZHU

Assignees

  • 汕尾职业技术学院

Dates

Publication Date
20260508
Application Date
20260407

Claims (10)

  1. 1. The LabView-based welding penetration control method is characterized by comprising the following steps of: s1, acquiring welding voltage signals in a welding process through a data acquisition system, and transmitting the welding voltage signals to a LabView processor; S2, the LabView processor performs data format conversion and noise reduction processing on the welding voltage signals, and converts the noise-reduced welding voltage signals into frequency domain signals; carrying out peak statistics and duty ratio analysis on the frequency domain signals, extracting the frequency with the highest duty ratio as a molten pool oscillation frequency, and transmitting the molten pool oscillation frequency to a penetration controller; s3, comparing the oscillation frequency of the molten pool with critical penetration frequency in an expert database by the penetration controller, judging the penetration state of the current welding process according to a comparison result, and generating a control signal according to the penetration state; S4, the welding system and/or the motion control system execute corresponding control actions according to the control signals so as to change welding heat input and adjust welding penetration states; s5, circularly executing the steps S1 to S4 in the welding process until the current penetration state reaches the preset target penetration state.
  2. 2. The welding penetration control method according to claim 1, wherein the noise reduction processing in step S2 adopts an adaptive multi-scale correlation weighted generalized threshold wavelet noise reduction method, comprising the steps of: S21, carrying out mean value removal and normalization on an original welding voltage signal x (n) to obtain a preprocessed signal x 0 (n)=(x(n)-μ)/σ x , wherein mu is the mean value of the original welding voltage signal, and sigma x is the standard deviation of the original welding voltage signal; S22, performing time domain impulse noise coarse detection on the preprocessing signal, and calculating local abnormal indexes of each sampling point: Wherein I (n) is a local abnormality index of an nth sampling point, m (n) is a local median in an nth point adjacent domain, MAD (n) is a median absolute deviation in the nth point adjacent domain, and epsilon is an extremely small positive number; when I (n) > T p , marking the corresponding sampling point as a pulse interference candidate point, wherein T p is a pulse noise discrimination threshold; s23, selecting a wavelet base as sym8, enabling the decomposition layer number J to be 5, and performing discrete wavelet decomposition on the preprocessed signal: wherein, DWT is discrete wavelet transformation, A J is the J-th layer approximation coefficient, D j is the J-th layer detail coefficient; s24, estimating standard deviation of noise of each layer and calculating basic threshold value of each layer, wherein noise of each layer is estimated as The corresponding basic threshold is Wherein σ j is the j-th layer noise standard deviation estimation value, λ j is the j-th layer base threshold value, and N j is the j-th layer coefficient number; S25, constructing an energy self-adaptive adjusting factor, and firstly calculating the average energy of detail coefficients of each layer Reconstructing an adaptive correction factor And obtain the corrected threshold value Wherein E j is the average energy of the detail coefficient of the j-th layer, k is the position index of the wavelet coefficient, Σ k E k is the sum of all decomposition layer capacities, alpha j is the j-th layer threshold self-adaptive correction factor, beta is the energy self-adaptive adjustment parameter, and lambda' j is the j-th layer threshold after energy self-adaptive correction; s26, constructing scale-related weights, for j=1..j-1, interpolating and aligning the detail coefficients of the j+1th layer to the J th layer, and recording as Re-calculating scale correlation index And construct multi-scale correlation weights Thereby obtaining a local effective threshold Wherein ρ j (k) is a scale correlation index near the j-th layer and the j+1th layer at position k, γ is a scale correlation adjustment parameter, ω j (k) is a multi-scale correlation weight at the j-th layer at position k, and the top layer takes ω J (k) =1; S27, if the time domain position corresponding to a certain wavelet coefficient is positioned in the impulse interference candidate point area, introducing impulse noise enhancement penalty to obtain Wherein delta is the impulse noise penalty enhancement factor and delta >0, A threshold value after punishment for combining impulse noise; S28, performing threshold contraction on each detail coefficient D j (k) by using a sampling generalized continuous threshold function: Wherein sgn (-) is a sign function, p epsilon [1,2] is a shape parameter of a generalized threshold function, The j-th layer detail coefficient after the threshold value shrinkage treatment; S29, performing inverse wavelet reconstruction according to the J-th layer approximation coefficient and the detail coefficients of each layer after threshold shrinkage And obtaining a welding voltage signal after noise reduction, and performing inverse normalization processing.
  3. 3. The welding penetration control method according to claim 1, wherein the means for converting the noise-reduced welding voltage signal to the frequency domain signal in step S2 includes any one of a fast fourier transform, a Welch power spectrum estimation, or a short-time fourier transform, wherein: when the fast Fourier transform is adopted, the welding voltage signal after noise reduction is windowed and then the spectrum analysis is carried out so as to reduce the spectrum leakage and improve the frequency resolution; When Welch power spectrum estimation is adopted, dividing a welding voltage signal after noise reduction into a plurality of mutually overlapped time periods, respectively windowing each time period, calculating a power spectrum and taking an average value to reduce the influence of random noise on frequency domain peak detection and improve the stability of molten pool oscillation frequency extraction; When short-time Fourier transform is adopted, carrying out local spectrum analysis on the welding voltage signal after noise reduction in a plurality of time windows to obtain a time-frequency distribution result of the variation of the molten pool oscillation frequency along with time in the welding process, and determining the molten pool oscillation frequency based on a dominant frequency statistical result in the time windows; The window function adopted in the windowing processing comprises one or more of a hanning window, a hamming window and a blackman window, and a frequency analysis range is preset when frequency domain transformation is carried out so as to screen out interference frequency components beyond the effective frequency range of molten pool oscillation.
  4. 4. The welding penetration control method according to claim 1, wherein the peak statistics and the duty analysis of the frequency domain signal in step S2 includes the steps of: SA1, carrying out frequency domain transformation on the welding voltage signal after noise reduction to obtain a frequency spectrum, detecting local maxima in the frequency spectrum based on local background normalized peak saliency to obtain a plurality of candidate characteristic peaks, extracting peak frequency, peak amplitude and peak width parameters corresponding to each candidate characteristic peak, wherein the peak saliency of the kth frequency point is that Wherein P (f k ) is the power spectrum value of the kth frequency point, B (f k ) is the mean or median of the local background spectrum near the kth frequency point, and the characteristic peak is determined if the following conditions are satisfied, namely a) P (f k ) is the local maximum value, B) S (k) is more than or equal to T s , wherein T s is the peak significance threshold value, and c) the peak width is within a preset range, thereby obtaining a candidate peak set: Wherein f i is the frequency corresponding to the ith candidate peak, A i is the power spectrum peak value of the ith candidate peak, W i is the peak width of the ith candidate peak, and M is the total number of candidate peaks; SA2, aiming at each candidate characteristic peak, generating a fundamental frequency candidate set in a harmonic inversion mode, and carrying out cluster combination on the fundamental frequency candidate set to obtain a plurality of candidate fundamental frequencies; SA3, respectively calculating the peak group energy duty ratio, the peak number duty ratio and the harmonic consistency duty ratio of each harmonic correlation peak group, and defining the total energy of the peak group as the q-th harmonic correlation peak group Wherein alpha h is a harmonic weighting coefficient for properly attenuating higher order harmonic, the energy duty ratio of the q-th harmonic associated peak group is ; Defining the number of effective harmonics detected in the q-th harmonic correlation peak group as Wherein I h epsilon {0,1}, when I h =1, the h harmonic is detected, when I h =0, the h harmonic is not detected, and the peak ratio is ; Defining the q-th harmonic correlation peak group as a harmonic consistency score Wherein f q,h is the h harmonic frequency of the q candidate fundamental frequency, A q,h is the matched h harmonic peak value, 0,F q is the candidate fundamental frequency if not matched, Δh is the frequency offset tolerance, and the harmonic consistency ratio is SA4, obtaining comprehensive duty ratio of each harmonic associated peak group based on preset fusion weight And determining the fundamental frequency corresponding to the harmonic correlation peak group with the largest comprehensive duty ratio as the molten pool oscillation frequency, wherein lambda i is more than or equal to 0 and lambda 1 +λ 2 +λ 3 =1.
  5. 5. The welding penetration control method according to claim 1, wherein the critical penetration frequency in the expert database is obtained by: S31, respectively setting a plurality of groups of welding process parameters aiming at different welding materials and/or different plate thicknesses, and performing a welding test under each group of welding process parameters, wherein the welding process parameters comprise one or more of welding current, welding voltage, welding speed, wire feeding speed, welding gun posture and shielding gas parameters; s32, collecting welding voltage signals in each welding test process, and extracting molten pool oscillation frequency in a corresponding test process based on the welding voltage signals; S33, detecting a weld forming result after each welding test, and judging a penetration state of a corresponding welding test by combining a weld back forming state, penetration, melting width, residual height, back collapse amount and/or a section metallographic result; S34, establishing a correlation sample set of welding materials, plate thicknesses, welding technological parameters, penetration states and molten pool oscillation frequencies corresponding to each welding test; S35, screening samples in a penetration state in the associated sample set, and selecting the maximum molten pool oscillation frequency in the penetration state as the critical penetration frequency of the corresponding welding material under the conditions of the same welding material and the same or corresponding process parameter range; s36, storing the critical penetration frequency, the corresponding welding material types, plate thickness information and welding process parameter ranges into an expert database together to form a critical penetration frequency mapping relation; S37, in the subsequent welding process, calling the matched critical penetration frequency from an expert database according to the current welding materials, plate thicknesses and welding process parameters, and comparing the molten pool oscillation frequency with the critical penetration frequency to determine the penetration state of the current welding process.
  6. 6. The welding penetration control method according to claim 5, wherein the critical penetration frequency in the expert database is established according to the welding material type, the plate thickness specification and the welding process parameter interval, and specifically comprises the steps of a) classifying welding materials according to the base material type, the material grade, the material thickness and the joint form, b) grouping the welding process parameters according to one or more of a welding current interval, a welding voltage interval, a welding speed interval, a wire feeding speed interval, an arc length interval, a shielding gas type and a flow interval, c) establishing a corresponding critical penetration frequency data item for each material type, plate thickness specification and the welding process parameter interval respectively, recording the number of test samples, the frequency fluctuation range, the penetration state mark and the confidence level corresponding to the critical penetration frequency, d) calling the critical penetration frequency corresponding to the data item when the welding material type, the plate thickness specification and the welding process parameter of a workpiece to be welded are matched with one data item in the expert database, and obtaining a welding process parameter close, close proximity, proximity or interpolation weighted proximity, interpolation, weighting or interpolation, and subsequent iteration and welding process frequency correction are carried out on the prior to the expert database by adopting a new method, and the current oscillation and the welding process frequency is corrected, and the welding penetration result is obtained.
  7. 7. The welding penetration control method according to claim 1, wherein the penetration state in step S3 includes at least an un-penetration state, a critical penetration state, a full penetration state, and a penetration state, wherein: when the oscillation frequency of the molten pool is larger than the corresponding critical penetration frequency upper limit threshold value in the expert database, judging that the molten pool is in an unfermented state; when the oscillation frequency of the molten pool is located in a preset tolerance interval near the critical penetration frequency, judging the state as a critical penetration state; When the oscillation frequency of the molten pool is smaller than the critical penetration frequency and is in a preset complete penetration interval, judging that the molten pool is in a complete penetration state; When the oscillation frequency of the molten pool is lower than the lower limit of the complete penetration interval or the welding heat input change trend is combined to judge that the oscillation frequency exceeds a preset penetration allowance, judging that the molten pool is in an overpenetration state; The penetration controller is used for carrying out smooth correction on a single judgment result by combining the change trend, the change slope and the frequency fluctuation amplitude of the oscillation frequency of the molten pool in a plurality of continuous sampling periods when judging the penetration state.
  8. 8. The welding penetration control method of claim 1, wherein the controlling act in step S4 comprises controlling the welding system and/or the motion control system to perform a controlling act of increasing the welding heat input when it is determined that the current welding process is in an un-penetrated state, and controlling the welding system and/or the motion control system to perform a controlling act of decreasing the welding heat input when it is determined that the current welding process is in an over-penetrated state, wherein the controlling act comprises one or more of increasing or decreasing the welding current, increasing or decreasing the welding speed, and increasing or decreasing the arc length.
  9. 9. The welding penetration control method according to claim 1, wherein the preset target penetration state in step S5 is a critical penetration state or a complete penetration state, and: when the preset target penetration state is a critical penetration state, the penetration controller controls the welding system and/or the motion control system to adjust the real-time molten pool oscillation frequency to be within a preset frequency interval corresponding to the critical penetration frequency, so that the penetration risk is reduced while the welding forming quality is ensured; When the preset target penetration state is the complete penetration state, the penetration controller controls the welding system and/or the motion control system to adjust the real-time molten pool oscillation frequency to be in a target frequency interval lower than the critical penetration frequency so as to ensure that the welding seam meets the complete penetration requirement.
  10. 10. A welding penetration control system that implements the welding penetration control method of any one of claims 1 to 9, characterized by comprising: A welding system for performing a welding operation, the welding system including a welding power source, a welding gun, a shielding gas supply, and a cooling system; the data acquisition system is used for acquiring welding voltage signals in the welding process; the LabView processor is used for receiving the welding voltage signals acquired by the data acquisition system, and performing format conversion, noise reduction treatment, frequency domain transformation and molten pool oscillation frequency extraction on the welding voltage signals; The penetration controller is used for receiving the oscillation frequency of the molten pool output by the LabView processor, judging the penetration state of the current welding process by combining the critical penetration frequency in the expert database, and further generating a control signal; and the motion control system is used for adjusting the motion state and/or welding parameters of the welding system according to the control signal so as to change the welding heat input and adjust the welding penetration state.

Description

LabView-based welding penetration control method and system Technical Field The invention relates to the technical field of welding process control and intelligent manufacturing, in particular to welding penetration control, and particularly relates to a LabView-based welding penetration control method and system. Background In the arc welding process, the penetration state of the welding seam directly influences the forming quality, the connecting strength and the use reliability of the welding joint. If the heat input is insufficient in the welding process, the lack of penetration or incomplete penetration is easily caused, and the defect of the weld joint root is caused, and if the heat input is excessive, the lack of penetration, burning-through, back collapse or weld joint forming deterioration is easily caused. Therefore, the welding penetration state is accurately identified in real time, and the welding process parameters are adjusted accordingly, so that the method has important significance for improving the welding quality. Among the existing welding penetration control methods, one type of method relies on post-welding detection, such as detection by cross-section observation, back surface formation detection or metallographic analysis, to determine penetration state. Although the method can obtain more accurate detection results, the method generally cannot meet the requirements of on-line monitoring and real-time regulation in the welding process. Another type of method attempts to indirectly represent the penetration state by using electric signals, acoustic signals, visual images or temperature field information in the welding process, but the related signals have stronger noise due to strong electromagnetic interference, random pulse interference and process fluctuation in the welding process, so that the feature extraction difficulty is higher, and the recognition stability is insufficient. In particular, the welding voltage signal contains rich dynamic information of a molten pool, and has certain correlation with the oscillation behavior of the molten pool. However, the welding voltage signal generally has the characteristics of strong non-stationarity, multiple local abrupt changes, complex noise and the like, and if the frequency domain analysis is directly performed, a plurality of frequency peaks often appear, so that the dominant frequency representing the penetration state is difficult to accurately identify. In addition, under different welding materials, plate thicknesses and welding process parameters, corresponding penetration critical conditions are obviously different, and unified and reliable penetration judgment references are lacked. Meanwhile, the existing partial welding control scheme can detect welding process parameters, but most of the existing partial welding control scheme lacks a real-time closed-loop control mechanism based on a penetration state discrimination result, and can not dynamically adjust welding heat input according to the penetration state, so that stable target penetration state is difficult to maintain under complex working conditions. Therefore, it is needed to propose a new welding penetration control method and system, so as to collect welding voltage signals in real time in the welding process, perform effective noise reduction and stable feature extraction on the signals, and compare the extracted molten pool oscillation frequency with a pre-established critical penetration frequency expert database, thereby judging the current penetration state on line, dynamically adjusting welding heat input according to the judging result, and realizing closed-loop control of the welding penetration state. Disclosure of Invention The invention aims to provide a LabView-based welding penetration control method and a LabView-based welding penetration control system, which are used for solving the problems that in the prior art, the welding penetration state is difficult to identify on line, the welding voltage signal noise interference is serious, the extraction stability of the oscillation frequency of a molten pool is insufficient, the closed-loop regulation of welding heat input is lacked, and the like. The invention also aims to provide a critical penetration frequency acquisition and expert library construction method suitable for different welding materials, different plate thicknesses and different welding process parameters so as to realize intelligent judgment and self-adaptive control of the welding penetration state. In order to achieve the above purpose, the invention adopts the following technical scheme: the invention provides a LabView-based welding penetration control method, which comprises the following steps: Acquiring welding voltage signals in the welding process through a data acquisition system, and transmitting the welding voltage signals to a LabView processor; the LabView processor performs data format conversion and noise reduction