CN109324033B - Method and device for detecting state of aluminum alloy fusion welding process - Google Patents
Method and device for detecting state of aluminum alloy fusion welding process Download PDFInfo
- Publication number
- CN109324033B CN109324033B CN201810798526.7A CN201810798526A CN109324033B CN 109324033 B CN109324033 B CN 109324033B CN 201810798526 A CN201810798526 A CN 201810798526A CN 109324033 B CN109324033 B CN 109324033B
- Authority
- CN
- China
- Prior art keywords
- welding process
- metal
- aluminum alloy
- spectrum
- fusion welding
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/66—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light electrically excited, e.g. electroluminescence
- G01N21/67—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light electrically excited, e.g. electroluminescence using electric arcs or discharges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/66—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light electrically excited, e.g. electroluminescence
- G01N21/69—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light electrically excited, e.g. electroluminescence specially adapted for fluids, e.g. molten metal
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/66—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light electrically excited, e.g. electroluminescence
- G01N21/69—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light electrically excited, e.g. electroluminescence specially adapted for fluids, e.g. molten metal
- G01N2021/695—Molten metals
Abstract
The invention discloses a method for detecting the state of an aluminum alloy fusion welding process, which comprises the following steps of S1, obtaining a metal spectrum signal in the aluminum alloy fusion welding process, and separating the metal spectrum signal by adopting an envelope method to obtain a metal background spectrum and a metal wire spectrum; step S2, performing principal component analysis on the plurality of metal line spectrums to obtain a feature vector coefficient of a first principal component and a feature vector curve of the first principal component of each metal line spectrum; step S3, confirming the chemical element corresponding to each peak in the eigenvector curve obtained in step S2; and step S4, according to the characteristic vector coefficients obtained in the step S2, performing correlation qualitative analysis and sensitivity evaluation on the chemical elements determined in the step S3 according to the wavelength pixels obtained corresponding to the characteristic vector coefficients. The method is based on a spectral information deep mining and feature extraction method of metal spectrum principal component analysis, and can realize state detection of the aluminum alloy fusion welding process.
Description
Technical Field
The invention belongs to the technical field of aluminum alloy fusion welding; relates to the technical field of metal spectrum principal component analysis; in particular to a method for detecting the state of an aluminum alloy fusion welding process; still relate to an aluminum alloy fusion welding process state detection device.
Background
Aluminum alloy fusion welding is one of the main welding forming manufacturing methods in aerospace key components, and the welding quality is ensured to be very important. Under the background of the requirements of popularization and application of the robot and intelligent manufacturing, the realization of state monitoring and detection in the welding process has important significance for improving the stability of welding quality and promoting intelligent welding manufacturing. The efficient and stable sensing technology and the comprehensive process information are the key points for the robot to have enough 'intelligence' and realize intelligent manufacturing. The existing welding process detection technology is mainly based on a molten pool visual image, electric arc sound, magneto-optical imaging, X-ray, electric arc spectrum, multi-information fusion and the like. Compared with other information, the spectrum sensing has the advantages of abundant information quantity, sensitive real-time response, correlation with metallurgical defects in welding and the like. The metal spectral radiation mainly comes from welding tungsten electrodes, welding wires and base metals, and plasma radiation information containing a large number of chemical elements such as Al, Mg, Fe, Mn and the like is closely related to the welding state and quality. Abundant spectral information is fully excavated and utilized, and the method has important significance for improving the utilization rate of the spectral information, promoting scientific cognition on the physical process of the welding process, promoting the on-line detection technology of the welding quality and further promoting the intelligent manufacturing of the welding process of the robot.
The existing spectral feature extraction method mainly comprises a ratio method, namely, a metal spectral line is selected, and the ratio of the line spectrum to a background spectrum is used as a laser welding quality monitoring parameter; the plasma temperature method is that a plurality of spectral lines of the same kind of elements are utilized, and the stability of the welding process can be reflected according to boltzmann calculation. The publication number is CN 103878479B, and the patent name is that, the method is an on-line detection method for laser welding T-shaped overlap joint gap based on spectral analysis, the quantitative relation between the gap and the peak frequency is obtained by detecting the peak frequency of a spectral line of 500 nm-600 nm and utilizing origin calculation, thereby realizing the real-time detection of the gap and the timely adjustment of welding parameters and avoiding the generation of the gap and defects. The publication number is CN102615423B, the patent name is an online diagnosis method for the laser powder filling welding defects of galvanized steel based on characteristic spectrum, the relative intensity of a CuI324.8nm spectral line is used for changing in real time, whether defects occur in the welding process is monitored, and welding parameters are adjusted in real time.
The existing method for mining the spectrum knowledge is single, and manual selection of characteristic spectrum segments is mostly carried out by adopting manual experience; the utilization rate of spectral information is not high, and only one or more spectral lines of the same kind of elements are adopted; the adopted calculation method has higher requirements on equipment precision, has low universality and robustness on different welding materials and processes, and is difficult to popularize and apply in actual production. The difficulty of research is mainly due to the fact that the spectral signals contain a large amount of element radiation information participating in the dynamic process of the arc and are highly correlated. A disturbance tends to cause similar responses to multiple elemental spectral lines, and in addition to the complexity and dynamics of the welding process, the correlation between defects and spectral elemental spectral lines is relatively fuzzy and not unique.
Disclosure of Invention
The invention provides a method and a device for detecting the state of an aluminum alloy fusion welding process; the method is based on a spectral information deep mining and feature extraction method of metal spectrum principal component analysis, and can realize state detection of the aluminum alloy fusion welding process.
The technical scheme of the invention is as follows: a method for detecting the state of an aluminum alloy fusion welding process comprises the step S1 of obtaining a metal spectrum signal in the aluminum alloy fusion welding process, and separating the metal spectrum signal by an envelope method to obtain a metal background spectrum and a metal line spectrum; step S2, performing principal component analysis on the plurality of metal line spectrums to obtain a feature vector coefficient of a first principal component and a feature vector curve of the first principal component of each metal line spectrum; step S3, confirming the metal chemical element corresponding to each peak value in the characteristic vector curve obtained in the step S2; step S4, obtaining a first principal component characteristic vector coefficient value according to the step S2, and performing correlation qualitative analysis and sensitivity evaluation on the determined chemical elements; step S5, calculating the integral area of each peak value of the characteristic vector curve in the step S2 by adopting a trapezoidal approximation rule, wherein the integral area is a characteristic parameter for representing the state of the welding process; and step S6, constructing SPC control threshold values in the welding process according to the characteristic parameters obtained in the step S5, and using the SPC control threshold values in the welding process for state detection of the welding process.
Furthermore, the invention is characterized in that:
in step S1, acquiring a metal spectrum signal by using high-level external triggering; and the spectral band of the metal spectral signal is 200-600 nm.
In step S1, a fiber CCD spectrometer is used to obtain a metal spectrum signal; and the spectrometer probe of the fiber CCD spectrometer is finely adjusted in a horizontal-vertical orthogonal mode to be located in a position sensitive area.
The specific process of performing horizontal-vertical orthogonal fine adjustment on the spectrometer probe in the step S1 is as follows: the spectrometer probe and the center of the fusion welding arc are on the same horizontal line; the spectrometer probe performs micro displacement adjustment from top to bottom in the vertical direction of fusion welding; the spectrometer probe is rotated from left to right and the spectrometer probe is arranged in a position sensitive zone.
The specific process of quantitatively evaluating the sensitivity of the fusion welding process state of the chemical elements in step S4 is: and taking the characteristic vector coefficient value of the first principal component as the weight coefficient of the determined chemical element, carrying out corresponding positive correlation qualitative analysis or negative correlation qualitative analysis according to the positive or negative of the weight coefficient, and evaluating the sensitivity of the welding process state of the chemical element according to the absolute value of the weight coefficient.
The specific process of calculating the integrated area of each peak in step S5 is as follows: and (4) taking the peak point as the center to contain the whole spectrum peak, and performing integral operation on the spectrum peak value by adopting a trapezoidal method to obtain the integral area of the spectrum peak curve.
The formula of integral operation of the spectrum peak value by the trapezoidal method is as follows:
the invention also provides a device for detecting the state of the aluminum alloy fusion welding process, which is characterized by comprising a workbench for fixing a workpiece, wherein a welding gun is arranged on the workbench, and the distance between the tungsten electrode of the welding gun and the workpiece is 50-150 mm; the probe of the spectrometer is over against the tungsten electrode and the electric arc generated by the workpiece; the spectrometer probe is connected with the spectrometer through a data line; the spectrometer is connected with the computer through a data line; the computer can implement the aluminum alloy fusion welding process state detection method. Wherein the welding gun is fixed on the tail end shaft of the robot, and the computer controls the robot to drive the welding gun to move.
Compared with the prior art, the invention has the beneficial effects that: the method provides a spectral information deep mining method based on data principal component analysis, can release dependence on human experience, and extracts metal spectral information highly related to a welding state through deep analysis of data internal rules; the characteristic calculation method based on the spectral line area greatly reduces the requirement on the resolution of the spectrometer, can overcome the difficulty of inaccurate calculation caused by equipment precision drift, and simultaneously, compared with the traditional parameters such as spectral intensity, electronic temperature and the like, the provided principal component coefficient area characteristic parameter reflects the relative dynamic change in data and has higher robustness; repeated tests prove that the method is high in anti-interference capability, rapid and simple in calculation, high in expansibility and capable of achieving accurate detection of the state of the welding process.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a schematic representation of the front and back sides of a weld and X-ray inspection according to example 1 of the present invention;
FIG. 3 is a graph showing the spectral intensity variation during the fine tuning process of the spectrometer probe in embodiment 1 of the present invention;
FIG. 4 is a graph of the background metal spectrum and the line metal spectrum after separation in example 1 of the present invention;
FIG. 5 is a graph of the first and second principal component coefficients of PCA corresponding to the metal spectral bands of example 1 of the present invention;
FIG. 6 is a third and fourth principal component coefficient plot of PCA corresponding to the metal spectral band in example 1 of the present invention;
FIG. 7 is a graph of principal components of PCA and spectral intensity of metal in example 1 of the present invention;
FIG. 8 is a metal spectrum in a state of wire feeding and non-wire feeding in example 1 of the present invention;
FIG. 9 is a graph of characteristic parameters of different metal spectral lines in example 1 of the present invention;
FIG. 10 is a graph showing a physical map of a weld and FeI intensities of metal spectral lines and extracted characteristic parameters in example 2 of the present invention;
FIG. 11 is a pictorial view of a defective weld joint in example 3 of the present invention;
FIG. 12 is a detection curve of the welding process according to example 3 of the present invention;
FIG. 13 is a pictorial view of a defective weld joint in example 4 of the present invention;
FIG. 14 is a detection curve of the welding process of example 4 of the present invention;
FIG. 15 is a pictorial view of a defective weld joint in example 5 of the present invention;
FIG. 16 is a detection curve of the welding process of example 5 of the present invention;
FIG. 17 is a schematic structural diagram of the detecting device of the present invention.
In the figure: a is an in-situ welding wire-feeding-free area; b is a walking wire feeding area; c is a walking non-feeding area; d is a welding leakage area; e is an asphalt interference area; 1 is a welding gun; 2 is an electric arc; 3 is a workpiece; 4 is a spectrometer probe; 5 is an optical fiber; 6 is a spectrometer; and 7, a computer.
Detailed Description
The technical solution of the present invention is further explained with reference to the accompanying drawings and specific embodiments.
The invention provides a method for detecting the state of an aluminum alloy welding process, which has the specific process shown in figure 1, realizes the detection of the state of the aluminum alloy welding process based on the deep excavation and the characteristic extraction of an arc welding metal spectrum, and can be used for detecting the welding state in real time. Because different welding materials, welding wires and metal elements contained in the tungsten electrode are different, before the method is used for actual processing and manufacturing, welding experiments containing different states need to be completed, necessary data sources are provided for accurately extracting the metal spectral line characteristics, and the state detection of the welding process can be realized based on the method provided by the invention.
The invention also provides a device for detecting the state of the aluminum alloy fusion welding process, which comprises a workbench and a welding gun 4, wherein the workbench is used for fixing the workpiece 3, and the welding gun 4 is arranged on a tail end shaft of a robot; the welding device further comprises a spectrometer probe 4, wherein the spectrometer probe 4 is over against a gap between the tungsten electrode and the workpiece 3, and the spectrometer probe 4 acquires an electric arc 2 generated between the tungsten electrode and the workpiece 3 in the welding process; the spectrometer probe 4 is connected with the spectrometer 6 through a data line, preferably an optical fiber; the spectrometer 6 is connected to the computer 7 via a data line which is connected to a USB port via the computer 7. The computer 7 can implement the method for detecting the state of the aluminum alloy fusion welding process shown in the polar diagram 1, and meanwhile, the computer 7 also operates the robot to drive the welding gun 1 to move and control the wire feeding operation of the tungsten electrode. The preferred spectrometer 6 is AvaSpec-1350F-USB 2; the acquisition range of the spectrometer is 350nm-1100nm, the CCD pixels are 1350, the integration time is set to be 1.5ns, and the sampling rate is 70 Hz.
The specific embodiment of the invention comprises the following steps:
example 1
Step S1, fixing the workpiece 3 on a workbench by using a clamp, installing the welding gun 1 on the tail end shaft of the robot, and setting the distance between the tungsten electrode and the workpiece to be 100mm so as to ensure the electric arc 2 with required arc length; the spectrometer probe 4 is rigidly fixed on the bracket, the optical fiber 5 is connected with the probe and the spectrometer 6, and the spectrum information of the electric arc 2 collected in real time is transmitted to the computer 7 through the USB. The spectrometer 6 is a fiber optic spectrometer (AvaSpec-1350F-USB2), the acquisition range is 350nm-1100nm, the CCD pixels 1350, the integration time is set to be 1.5ms, and the sampling rate is about 70 Hz.
And step S2, adopting butt-jointed Y-shaped groove wire filling pulse tungsten electrode argon arc welding, wherein the peak current is 240A, the base value current is 50A, the pulse frequency is 1Hz, the wire feeding speed is 12mm/S, the flow of protective gas argon is 15L/min, the diameter of a tungsten electrode is 3.2mm, and the diameter of a welding wire is 1.6 mm. Before the welding is started, the welding speed of the welding gun 1 is controlled to be 16cm/min through a robot, the welding path is a straight line, and the walking length is 240 mm. After the arc starts, the robot controls the welding gun 1 to perform spot welding in situ, and the wire is not fed for preheating for 8 pulse periods; meanwhile, an arc starting signal is sent to the computer 7 after successful arc starting, the computer sends a high-level trigger signal to the spectrometer 6 after receiving the arc starting signal, and the spectrometer 6 controls the spectrometer probe 4 to start collecting arc spectrum signals.
Step S3, the spectrometer is extremely sensitive to the position of the probe, the probe is adjusted through vertical-horizontal continuous movement before the experiment, and the spectrum is collected in real time to analyze and select the more reasonable position of the probe; the specific spectrometer probe 4 moves up and down and then left and right, and is adjusted to a position sensitive area. First, the relative position distance D between the spectrometer probe 4 and the arc 2 is fixed to 100mm, and then the spectrometer probe 4 is slowly fine-tuned from top to bottom. As shown in fig. 3, the intensity curve of the line spectrum of the magnesium element is shown during the movement of the spectrometer probe 4, and it can be seen from the figure that when the spectrometer probe 4 moves up and down, the spectral intensity reaches the maximum in the area right near the arc 2, and the arc is too strong and may exceed the maximum range, and when slightly above or below the maximum range, the MgI intensity drops sharply, which indicates that the arc is shielded by the tungsten electrode seriously. Finally, the up-down position is determined in the position sensitive area. The spectrometer probe 4 is then moved from left to right as shown in figure 3, collecting at a fixed location in the sensitive area where the intensity is greater. Finally, the integration time was integrated and a 10% dimmer sheet 8 was added to bring the spectral intensity around the 2/3 range.
In the step, the position of the spectrometer probe 4 is adjusted to obtain a metal spectral line signal in the aluminum alloy fusion welding process, and the metal spectral line signal is used for performing the following metal spectral line characteristic extraction and process state detection method of Principal Component Analysis (PCA), and the specific process is as follows:
and S4, selecting a spectrum section with more active metal spectral line change between 200 nm and 600nm as a metal spectral signal to be analyzed according to the dynamic characteristics of the signals collected in real time in the step S3. And then separating a metal background spectrum and a metal line spectrum of the metal spectrum signal by using a method of taking down the envelope. FIG. 4 shows the separated metal line spectrum (382.2 nm-429.64 nm) and the metal background spectrum; as shown in step S101 in fig. 1.
In step S5, a total of N metal line spectrum samples are obtained in step S4, and Principal Component Analysis (PCA) processing is performed on the N metal line spectrum samples. The welding experiment of the embodiment comprises two states of no wire feeding and wire feeding, 60 samples are selected as a group, and PCA treatment is carried out on the samples with 81-dimensional characteristics; as shown in step S102 in fig. 1.
And step S6, performing qualitative correlation analysis and quantitative sensitivity evaluation by using the PCA principal component coefficient. As shown in fig. 5 and 6, the PCA coefficients of the four principal components, spectral line pixels with large absolute values of the PCA coefficients, are 64 pixels, 82-85 pixels, and 105 pixels, respectively; and determining the positive and negative correlation relations of different pixels in the same component according to the positive and negative values of the coefficient. If in the third component, 105 pixels are in negative correlation with 64 and 82, and the absolute value of the coefficient is the largest, the fourth component has a similar rule, which indicates that the spectral lines corresponding to the three pixels have a large negative correlation; as shown in step S103 in fig. 1.
Step S7, comparing and analyzing the PCA main component and the 105 pixel spectral line intensity curve to obtain that the No. 3 main component and the No. 1 main component have opposite change rules in the first half section of welding as shown in figure 7, and the negative correlation relationship between the No. 3 main component and the No. 1 main component is indirectly verified corresponding to the non-wire-feeding and wire-feeding polishing states in the welding process; in fig. 7, the negative value of the second-half intensity of the 3 rd main component is mainly contributed by 105 pixels, and it is assumed that the wire-feeding state is highly correlated with 105 pixels; and each coefficient in the 1 st principal component is a positive value, and the characteristic curve is the result of linear combination of each pixel coefficient.
And step S8, confirming the metal spectral line elements with larger coefficients through the NIST database. Due to the fact that the pixels of the CCD spectrometer 6 are limited, the content of each element in a welding plate and a welding wire, the radiation ionization difficulty, the theoretical intensity, the literature report and the like are comprehensively considered, and the final result is shown in the following table 1; this process is shown as step S104 in fig. 1.
TABLE 1
In step S9, the spectral signals in the wire-feeding and wire-non-feeding states are shown in FIG. 8. And analyzing whether the change of the wire feeding has a radiation influence rule on the metal spectral line. After wire feeding, the FeI radiation intensity of iron atoms is obviously increased, and the marked MgI and AlI do not obviously change, which shows that the change of wire feeding only affects the FeI spectral line, and the FeI spectral line and the marked MgI and AlI have a unique strong correlation relationship.
And step S10, calculating the integral area of the corresponding spectrum peak based on the PCA coefficient curve, and extracting the characteristic parameters of the metal spectral line. The peak point is taken as the center to contain the whole spectrum peak, and SOI-MgI, SOI-AlI and SOI-MgI are selected according to the following formula:
The result of this example is shown in fig. 9, in which the left side of a is in a thread-off state, the right side of B is in a thread-on state, and different characteristics of metal pu have different variation laws. After the wire feeding is started, the FeI spectral line coefficient area is increased sharply, which reflects that the dynamic change degree of FeI ionizing radiation is increased, the corresponding intensity curve has obvious peak value, and the characteristics of AlI and MgI are slightly reduced. It shows that the wire feeding has an instantaneous cooling effect on the electric arc, and the melting ionization of the welding wire consumes a part of the heat of the electric arc. The reasons may be: the Fe I has the maximum theoretical intensity in three types of metal spectral lines, the PCA coefficient is also the maximum, and the ionizing radiation degree of the FeI is increased after the Fe element in the welding wire enters an arc; secondly, the wire feeding has an instant cooling effect on the electric arc, a part of electric arc heat is consumed by melting and ionizing the welding wire, and the ionization voltage of Mg is maximum, the relative intensity is low, so the radiation intensity is minimum, and the negative correlation relationship among the three discovered by PCA is verified.
The characteristic values of all samples without wire feeding in example 1 are counted, normalized, and the upper threshold value is calculated by using SPC method. And designing welding experiments in different states, and performing repeatability verification. As shown in step S106 of fig. 5.
Example 2
The same welding process parameters as those of example 1 were adopted, and a part of metal was milled off on the front surface of the rear half of the workpiece, to simulate the manufacturing of different penetration states. The welding starts to weld and preheat in situ for 8 seconds without wire feeding, the welding continues until 3 pulses are generated after walking, then the continuous wire feeding is started for about 20 seconds, and then the wire feeding is stopped until the welding is finished; the workpiece design and weld sample results are shown in the physical representation of the weld in FIG. 10.
The metal spectrum FeI characteristic parameter curve shown in FIG. 10 is obtained by calculation through the method of the invention, and analysis is performed corresponding to different welding states. Under the coupling of the non-wire feeding state C and the welding leakage state D, the characteristic curve is still relatively stable, no response is made to the welding leakage defect, and the amplitude change is relatively large during wire feeding, so that the Fe I (407.84nm) characteristic value has unique correlation to the change of the wire feeding state and is not interfered by other factors.
Example 3
The same welding process parameters as those of example 1 were adopted, the surface of the weldment was polished to remove the oxide film, the polishing debris was removed by wiping with absolute ethanol, the workpiece was fixed on a welding table, the weld joints were filled with asphalt at three points E as shown in fig. 11, and the air hole defect was simulated. And (4) for the first 11 pulses, the wire is not fed, and the wire is fed in the first half section and not fed in the second half section of the weldment. According to the method of the invention, the characteristic parameters are calculated and compared with the SPC threshold line after normalization, as shown in FIG. 12, the characteristic values of the early-stage non-wire feeding state A and the latter-stage non-wire feeding state C are both near the threshold line, while the characteristic value of the wire feeding state B far exceeds the threshold, and the asphalt interference state E has no obvious influence on the monitoring characteristics.
Example 4
The same welding process parameters as in example 1 were adopted, and the weld joints are shown in fig. 13, with the first half of the welding being in a wire feeding traveling state B and the second half being in a non-wire feeding traveling state C. FeI characteristic parameters were calculated according to the method of the invention and normalized for comparison to SPC threshold lines as shown in fig. 14. The characteristic value of the early wire feeding state B is above a threshold value line, and the characteristic after wire feeding is stopped is lower than the threshold value.
Example 5
The welding process parameters same as those of the example 1 are adopted, the surface of the weldment is ground to remove the oxide film, the grinding scraps are removed through wiping of absolute ethyl alcohol, the weldment is fixed on a welding workbench, the welding line of the weldment is shown in figure 15, asphalt is filled at a plurality of positions (E point in figure 15) on the surface, and the wire feeding is not carried out in the whole welding process. The FeI signature was calculated and normalized and compared to the SPC threshold line as shown in fig. 16 to remove pre-weld instability, and the signature was substantially near the threshold and increased in magnitude compared to example 4 but still less than the signature in the wire feed.
It can be seen from the above experiment and embodiment results that by using the feature extraction method and the welding process on-line detection device of the present invention, the monitoring feature parameters with higher robustness can be obtained, the metal spectral line with the greatest correlation with the process state can be quickly found in real time by the principal component analysis method, the extracted feature parameters are simple in calculation method, good in sensitivity, low in precision requirement on a spectrometer, and strong in anti-interference capability, and accurate detection of the welding state can be realized by using the SPC control threshold, so that a foundation is laid for popularization to the process detection of the fusion welding process and the welding defects of other alloys.
Claims (9)
1. The method for detecting the state of the aluminum alloy fusion welding process is characterized by comprising the following steps of:
step S1, obtaining a metal spectrum signal in the aluminum alloy fusion welding process, and separating the metal spectrum signal by adopting an envelope method to obtain a metal background spectrum and a metal line spectrum;
step S2, performing principal component analysis on the plurality of metal line spectrums to obtain a feature vector coefficient of a first principal component and a feature vector curve of the first principal component of each metal line spectrum;
step S3, confirming the metal chemical element corresponding to each peak value in the characteristic vector curve obtained in the step S2;
step S4, according to the characteristic vector coefficients obtained in the step S2, performing correlation qualitative analysis and sensitivity evaluation on the chemical elements determined in the step S3 according to the wavelength pixels correspondingly obtained by the characteristic vector coefficients;
step S5, calculating the integral area of each peak value of the characteristic vector curve obtained in the step S2 by adopting a trapezoidal method, wherein the integral area is a characteristic parameter for representing the state of the welding process;
and step S6, constructing SPC control threshold values in the welding process according to the characteristic parameters obtained in the step S5, and using the SPC control threshold values in the welding process for state detection of the welding process.
2. A method for detecting the state of an aluminum alloy fusion welding process as recited in claim 1, wherein in said step S1, a high level external trigger is used to obtain a metal spectrum signal; and the spectral band of the metal spectral signal is 200-600 nm.
3. A method for detecting the state of an aluminum alloy fusion welding process according to claim 1, wherein in step S1, a fiber CCD spectrometer is used to obtain a metal spectrum signal; and the spectrometer probe of the fiber CCD spectrometer is finely adjusted in a horizontal-vertical orthogonal mode to be located in a position sensitive area.
4. The method for detecting the state of the aluminum alloy fusion welding process of claim 3, wherein the specific process of performing the horizontal-vertical orthogonal fine adjustment on the spectrometer probe in the step S1 is as follows: the spectrometer probe and the center of the fusion welding arc are on the same horizontal line; the spectrometer probe performs micro displacement adjustment from top to bottom in the vertical direction of fusion welding; the spectrometer probe rotates from left to right and is arranged in a position sensitive zone.
5. The method of inspecting the aluminum alloy fusion welding process state as claimed in claim 1, wherein the chemical elements determined in the step S3 are subjected to the qualitative analysis of correlation and the evaluation of sensitivity in the step S4 by the following specific procedures: and taking the characteristic vector coefficient value of the first principal component as a weight coefficient of the determined metal chemical element, carrying out corresponding positive correlation qualitative analysis or negative correlation qualitative analysis according to the positive or negative of the weight coefficient, and evaluating the sensitivity of the welding process state of the chemical element according to the absolute value of the weight coefficient.
6. A method for detecting the state of an aluminum alloy fusion welding process according to claim 1, wherein the specific process of calculating the integrated area of each peak value in step S5 is: and (4) taking the peak point as the center to contain the whole spectrum peak, and performing integral operation on the spectrum peak value by adopting a trapezoidal method to obtain the integral area of the spectrum peak curve.
7. A method for detecting the condition of an aluminum alloy fusion welding process as recited in claim 6, wherein said trapezoidal method is adapted to perform the integration of the peak values of the spectraThe formula is as follows:
8. a state detection device for an aluminum alloy fusion welding process is characterized by comprising a workbench for fixing a workpiece (3), wherein a welding gun (1) is arranged on the workbench, and a spectrometer probe (4) is over against a tungsten electrode and an electric arc (2) generated by the workpiece (3); the spectrometer probe (4) is connected with the spectrometer (6) through a data line; the spectrometer (6) is connected with the computer (7) through a data line; a computer (7) capable of implementing the aluminum alloy fusion welding process status detection method as claimed in claim 1.
9. A state detecting device for an aluminum alloy fusion welding process according to claim 8, characterized in that the welding torch (1) is fixed on a terminal shaft of a robot, and the computer (7) operates the robot to move the welding torch (1).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810798526.7A CN109324033B (en) | 2018-07-19 | 2018-07-19 | Method and device for detecting state of aluminum alloy fusion welding process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810798526.7A CN109324033B (en) | 2018-07-19 | 2018-07-19 | Method and device for detecting state of aluminum alloy fusion welding process |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109324033A CN109324033A (en) | 2019-02-12 |
| CN109324033B true CN109324033B (en) | 2020-06-19 |
Family
ID=65264061
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810798526.7A Active CN109324033B (en) | 2018-07-19 | 2018-07-19 | Method and device for detecting state of aluminum alloy fusion welding process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109324033B (en) |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101275913A (en) * | 2008-01-10 | 2008-10-01 | 中北大学 | MIG solder quality monitoring method based on electric arc multi-spectral coverage light information sensing |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9981341B2 (en) * | 2014-08-25 | 2018-05-29 | Jyoti Mazumder | Smart additive manufacturing system (SAMS) |
| CN106312310B (en) * | 2015-07-01 | 2018-05-08 | 湖南大学 | Laser welding optimization method and its device based on Laser induced plasma spectroscopy |
-
2018
- 2018-07-19 CN CN201810798526.7A patent/CN109324033B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101275913A (en) * | 2008-01-10 | 2008-10-01 | 中北大学 | MIG solder quality monitoring method based on electric arc multi-spectral coverage light information sensing |
Non-Patent Citations (4)
| Title |
|---|
| Al-Mg合金钨极气体保护焊接电弧的光谱特征提取方法;余焕伟等;《上海交通大学学报》;20131130;第47卷(第11期);第1655-1660页 * |
| EMD-based pulsed TIG welding process porosity defect detection and defect diagnosis using GA-SVM;Yiming Huang等;《Journal of Materials Processing Technology》;20160712;第92-102页 * |
| Online defect detection of Al alloy in arc welding based on feature extraction of arc spectroscopy signal;Zhifen Zhang等;《Int J Adv Manuf Technol》;20150314;第2067-2077页 * |
| 电弧光谱深度挖掘下的铝合金焊接过程状态检测;张志芬等;《焊接学报》;20190131;第40卷(第1期);第019-025页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109324033A (en) | 2019-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP1340576A1 (en) | Control method of arc welding and arc welder | |
| US10094782B2 (en) | Method and apparatus for fast quantitative analysis of a material by laser induced breakdown spectroscopy (LIBS) | |
| Zhang et al. | Spectral diagnosis of wire arc additive manufacturing of Al alloys | |
| CN109530955A (en) | The welding technological properties evaluating apparatus and method of gas shield welding wire | |
| Compernolle et al. | Direct determination of Pd, Pt and Rh in fire assay lead buttons by laser ablation-ICP-OES: automotive exhaust catalysts as an example | |
| CN111015006B (en) | Online monitoring method for laser-arc hybrid welding quality based on spectral information | |
| US8063340B2 (en) | Method for controlling the consistency of an arc welding process by monitoring welding voltage to determine weld droplet detachment | |
| Taparli et al. | In situ laser-induced breakdown spectroscopy measurements of chemical compositions in stainless steels during tungsten inert gas welding | |
| CN106353284A (en) | On-line diagnosis method for defects in laser material increasing manufacturing process on basis of spectrum diagnosis | |
| CN206567687U (en) | Detect the penetration control device of frequency of oscillation in a kind of pulse laser exciting TIG molten baths | |
| Li et al. | Robust sensing of arc length | |
| CN102749309A (en) | Laser-induced breakdown spectrograph for rapid analysis of metallurgical components and analysis method therefor | |
| CN109030461A (en) | A kind of laser induced breakdown spectroscopy electrochemistry combination heavy metal detection method | |
| CN106695103A (en) | Real-time monitoring detecting device for electron beam deep penetration welding and penetration forming | |
| ZHANG et al. | Visual sensing of weld pool in variable polarity TIG welding of aluminium alloy | |
| Ancona et al. | A sensing torch for on-line monitoring of the gas tungsten arc welding process of steel pipes | |
| CN109324033B (en) | Method and device for detecting state of aluminum alloy fusion welding process | |
| Barborak et al. | " Through-arc" process monitoring: techniques for control of automated gas metal arc welding | |
| Lv et al. | Automatic measuring and processing system of audio sensing for real-time arc height control of pulsed GTAW | |
| Yang et al. | The effects of process variables on the bead height of submerged-arc weld deposits | |
| d'Abzac et al. | Study of near infra red femtosecond laser induced particles using transmission electron microscopy and low pressure impaction: Implications for laser ablation–inductively coupled plasma-mass spectrometry analysis of natural monazite | |
| CN106841177B (en) | Online defect diagnosis method in laser cladding process | |
| Guo et al. | Quality monitoring in Wire-Arc additive manufacturing based on spectrum | |
| CN107214409A (en) | A kind of electron beam deep penetration welding penetrates the real-time monitor and detection method of shaping | |
| JP3943380B2 (en) | Arc welding control method and arc welding apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |