Weld evaluation device and weld evaluation method

Abstract

To highly accurately measure the depth of penetration according to welding.SOLUTION: A weld evaluation device comprises: an AE sensor 21 for detecting AE waves generated from a weld zone 2 during laser welding; a frequency analysis part 32 for analyzing the frequency of waveform data on an AE signal outputted from the AE sensor and processed; a frequency band extraction part 33 for extracting a frequency component of a frequency band contributing to evaluation of a depth of penetration at a weld zone, from an analysis result obtained in the frequency analysis part; a reverse frequency analysis part 34 for obtaining waveform data by applying reverse frequency analysis to a frequency component in a frequency band extracted by the frequency band extraction part; and a penetration depth calculation part 35 for calculating a depth of penetration at a weld zone, on the basis of the intensity of waveform data obtained by the reverse frequency analysis part.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] An embodiment of the present invention relates to a welding evaluation device and a welding evaluation method that measure and evaluate a penetration depth of a welded portion based on an elastic wave generated from the welded portion during welding. [Background technology]

[0002] In laser welding, a technology is known in which an AE sensor detects AE waves generated by acoustic emission (AE) at the welding point during welding, and the AE signal obtained by the detection of the AE waves is used for quality evaluation such as abnormality detection at the welding point and prediction of penetration depth. The AE waves generated during welding are a type of elastic wave. However, the detected AE waves may include AE ​​waves that do not contribute to quality evaluation, or environmental sounds (e.g. noise) generated from other than the welding point may be detected at the same time, which may affect the quality evaluation of the welding point. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2010-266404 A [Patent Document 2] JP 2007-111745 A [Patent Document 3] Patent No. 3292008 [Patent Document 4] JP 2006-181585 A Summary of the Invention [Problem to be solved by the invention]

[0004] In Patent Document 1, in addition to the AE waveforms generated by welding, there are cases where large AEs that do not contribute to welding are included, such as when the laser beam initially hits the workpiece, when spatter occurs, and when foreign matter adheres, and a threshold is set to prevent these from being judged as welding abnormalities, making it possible to determine abnormalities in the welded part. Also, in Patent Document 2, since the AE signal level differs depending on the change in the laser irradiation angle to the cylinder and the cylinder rotation angle, a threshold is set according to each angle to identify the penetration depth, weld lead length, and weld skips, and the presence or absence of a weld defect is determined by comparing the judgment value of the AE wave when a good product is welded.

[0005] The techniques described in the above Patent Documents 1 and 2 set a threshold value based on the AE signal in a normal state to determine quality defects. In addition, the measurement of the penetration depth is also judged to be normal or not by referring to the threshold value, and accurate monitoring of the penetration depth is not possible.

[0006] In Patent Document 3, the plasma light inside the keyhole is measured by a photodiode, and the quality of the molten state is judged from the intensity of a specific frequency signal. In Patent Document 4, the reflected light of the laser beam is detected by a photodiode during laser-arc hybrid welding to capture the change in frequency band that occurs when there is porosity or non-welding. The techniques described in Patent Documents 3 and 4 check the frequency band in more detail from the acquired information to extract the necessary information. However, in the techniques described in Patent Documents 3 and 4, the light emission during laser welding is measured, and evaluation is performed using the characteristic frequency band that occurs due to the molten state or poor welding.

[0007] The embodiments of the present invention have been made in consideration of the above circumstances, and an object of the present invention is to provide a welding evaluation device and a welding evaluation method that can measure the penetration depth of welding with high accuracy. [Means for solving the problem]

[0008] A welding evaluation device in an embodiment of the present invention is characterized in that it is configured to include a sensor that detects elastic waves generated from a welded portion during welding; a frequency analysis unit that performs frequency analysis on waveform data of a detection signal output from the sensor and processed; a frequency band extraction unit that extracts frequency components of a frequency band that contribute to the evaluation of a penetration depth at the welded portion from the analysis result obtained by the frequency analysis unit; an inverse frequency analysis unit that performs inverse frequency analysis on the frequency components of the frequency band extracted by the frequency band extraction unit to obtain waveform data; and a penetration depth calculation unit that calculates the penetration depth of the welded portion based on the intensity of the waveform data obtained by the inverse frequency analysis unit.

[0009] A welding evaluation method in an embodiment of the present invention is characterized by comprising the steps of: a step of a sensor detecting elastic waves generated from a welded area during welding; a step of a frequency analysis unit performing frequency analysis on waveform data of the detection signal output from the sensor and processed; a step of a frequency band extraction unit extracting frequency components of a frequency band that contribute to the evaluation of the penetration depth at the welded area from the analysis result obtained by the frequency analysis unit; a step of an inverse frequency analysis unit performing inverse frequency analysis on the frequency components of the frequency band extracted by the frequency band extraction unit to obtain waveform data; and a step of a penetration depth calculation unit calculating the penetration depth of the welded area based on the intensity of the waveform data obtained by the inverse frequency analysis unit. Effect of the Invention

[0010] According to an embodiment of the present invention, the penetration depth due to welding can be measured with high accuracy. [Brief description of the drawings]

[0011] [Figure 1] FIG. 1 is a block diagram showing a configuration of a laser welding evaluation device according to an embodiment. [Diagram 2] FIG. 2 is a block diagram showing the configuration of the CPU in FIG. 1. [Diagram 3]3A to 3C are diagrams for explaining the derivation of an envelope by the first envelope derivation unit in Figure 2, where (A) is a graph showing an AE signal output from the AE sensor in Figure 1 and processed, and (B) is a graph showing a converted AE signal obtained by converting the AE signal in Figure 3A, and an envelope derived from this signal. [Figure 4] 4(A) is a cross-sectional photograph showing the welded portion, and FIG. 4(B) is a graph showing the penetration depth at the welded portion of FIG. 4(A). [Diagram 5] Graph showing the relationship between the penetration depth in FIG. 4(B) and the envelope curve in FIG. 3(B). [Figure 6] 3 is a flowchart illustrating steps of a laser welding evaluation method using the laser welding evaluation device of FIGS. 1 and 2. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. Fig. 1 is a block diagram showing the configuration of a laser welding evaluation device according to an embodiment. Fig. 2 is a block diagram showing the configuration of a CPU in Fig. 1. A laser welding evaluation device 10 as a welding evaluation device shown in Figs. 1 and 2 calculates, measures, and evaluates a penetration depth of a welded portion 2 welded to a workpiece 1 by a laser welding device 11, and is configured with an AE sensor 21, a signal processing unit 22, and a CPU (Central Processing Unit) 23. Here, the laser welding evaluation device 10 may be provided integrally with the laser welding device 11, or may be installed outside the laser welding device 11.

[0013] 1 includes a laser control device 12, a laser oscillator 13, an optical fiber 14, and a welding head 15. The laser control device 12 controls the laser oscillator 13. For example, the laser control device 12 controls the power (output) of the laser light L output by the laser oscillator 13, the timing at which the laser oscillator 13 starts, interrupts, or ends laser irradiation, and the like. The laser light L output from the laser oscillator 13 is transmitted from the optical fiber 14 to the welding head 15, and is focused to a desired beam diameter by a lens 16 in the welding head 15, and is irradiated onto the workpiece 1, which is an object to be welded.

[0014] The AE sensor 21 is composed of, for example, a piezoelectric element, detects elastic waves (i.e., AE waves) due to AE (Acoustic Emission) generated at the welded portion 2 of the workpiece 1 during laser welding, and outputs an AE signal as a detection signal. In laser welding by the laser welding device 11, the deeper the penetration depth is, the larger the AE generated and the larger the AE waves generated, and therefore the larger the AE signal output from the AE sensor 21 becomes. When the contact-type AE sensor (for example, a piezoelectric element) 21 is fixed to the workpiece 1, a contact medium is applied to the contact surface to ensure close contact with no gaps, making it possible to detect AE waves with higher sensitivity and output an AE signal.

[0015] In addition, the AE signal acquired by the AE sensor 21 differs depending on the sensitivity characteristics of the AE sensor 21. As the AE sensor 21 in this embodiment, an AE sensor capable of outputting an AE signal including frequency components in a frequency band that contributes to evaluation of the penetration depth at the welded portion 2 of the workpiece 1 is used.

[0016] The signal processing unit 22 processes the AE signal output from the AE sensor 21, and includes a preamplifier 24, a filter 25, a main amplifier 26, and an A / D converter 27. That is, when the AE signal output from the AE sensor 21 is weak, the preamplifier 24 amplifies the AE signal based on a predetermined amplification condition. The filter 25 attenuates signals that are not necessary for measurement, such as noise, as necessary. The main amplifier 26 amplifies the AE signal transmitted from the filter 25 under a predetermined amplification condition, and the A / D converter 27 converts the AE signal from the main amplifier 26 from an analog signal to a digital signal.

[0017] The CPU 23 is a calculation unit that actually calculates the penetration depth, and as shown in FIG. 2, has a signal acquisition unit 31, a frequency analysis unit 32, a frequency band extraction unit 33, an inverse frequency analysis unit 34, and a penetration depth calculation unit 35, and further has a first envelope derivation unit 36 ​​and a second envelope derivation unit 37.

[0018] The signal acquiring unit 31 acquires, in a time series, the AE signal output from the AE sensor 21 and processed by the signal processing unit 22. The time series waveform data A of this AE signal is shown in Fig. 3(A). Furthermore, the frequency analysis unit 32 performs frequency analysis on the waveform data A of the AE signal from the signal acquiring unit 31, resolves it into frequency components, and obtains intensity information of each frequency component.

[0019] The frequency band extraction unit 33 extracts frequency components of a frequency band correlated with the penetration depth at the welded portion 2 of the laser welding (i.e., a frequency band that contributes to the evaluation of the penetration depth) from the analysis result obtained by the frequency analysis unit 32. For example, the frequency band extraction unit 33 extracts frequency components of a frequency band of 400 kHz or more, thereby making it possible to evaluate the penetration depth using an AE signal that is less affected by environmental sounds (such as vibrations occurring outside the welded portion 2 and cross-jet gas sounds flowing during welding).

[0020] The inverse frequency analysis unit 34 obtains waveform data by performing inverse frequency analysis on the frequency components in the frequency band extracted by the frequency band extraction unit 33. Therefore, the waveform data obtained by the inverse frequency analysis unit 34 is the waveform data of the AE signal obtained by performing inverse frequency analysis on the frequency components in the frequency band that contribute to the evaluation of the penetration depth, among the waveform data A of the AE signal output from the AE sensor 21 and processed by the signal processing unit 22.

[0021] The penetration depth calculation unit 35 calculates the penetration depth of the welding point 2 based on the intensity of the waveform data of the AE signal obtained by the inverse frequency analysis unit 34 and the laser welding conditions from the laser control device 12. Here, FIG. 3(A) shows the waveform data A of the AE signal output from the AE sensor 21 during laser welding by the laser welding device 11 and processed by the signal processing unit 22, and FIG. 4(A) and (B) show the fluctuation of the penetration depth D of the welding point 2 at this time. In the cross-sectional photograph of FIG. 4(A), the reference numeral 3 indicates the base material of the workpiece 1, and the reference numeral 4 indicates the boundary between the welding point 2 and the base material 3, that is, the penetration depth D. In FIG. 4(A), the laser welding is started from the point α, and the welding progresses to the right in the figure. As shown in these FIGS. 4(A) and (B), it can be seen that the penetration depth D fluctuates from the start of welding under the specified laser welding conditions.

[0022] The penetration depth D varies depending on the laser welding conditions, etc., so that the higher the laser output (power), the deeper the penetration depth D. The generated AE wave also varies proportionally depending on the laser welding conditions, etc., so that the intensity of the AE signal output from the AE sensor 21 varies. Furthermore, the lower the laser output and the shallower the penetration depth D, the smaller the fluctuations in the penetration depth D and the intensity of the AE signal become. Therefore, the penetration depth calculation unit 35 calculates the penetration depth D of the welding point 2 from the laser welding conditions and the intensity of the waveform data of the AE signal, based on the previously stored relationship between the penetration depth D of the welding point 2, the laser welding conditions from the laser control device 12, and the intensity of the waveform data of the AE signal from the inverse frequency analysis unit 34.

[0023] The first envelope deriving unit 36 ​​shown in Fig. 2 obtains an envelope C of the waveform data A of the AE signal output from the AE sensor 21 and processed by the signal processing unit 22 from the waveform data A. In other words, the waveform data B of the converted AE signal shown in Fig. 3(B) is obtained by converting the waveform data A of the AE signal oscillating to the + side and - side shown in Fig. 3(A) into waveform data oscillating to the + side. The first envelope deriving unit 36 ​​derives the envelope C of the waveform data A of the AE signal, for example, by averaging the waveform data B of the converted AE signal shown in Fig. 3(B) with data before and after the time axis and processing it smoothly. It can be seen that this envelope C is wavy and fluctuates.

[0024] When comparing the temporal fluctuation of the envelope C in the waveform data A of the AE signal described above with the time series data of the penetration depth D of the welding point 2, it is found that the fluctuation periods are in a matching relationship, as shown in Fig. 5. Therefore, the penetration depth calculation unit 35 shown in Fig. 2 is able to calculate the time series data of the penetration depth D of the welding point 2 from the temporal fluctuation of the envelope C, based on the pre-stored relationship between the temporal fluctuation of the envelope C in the waveform data A of the AE signal described above and the time series data of the penetration depth D of the welding point 2.

[0025] 2 obtains an envelope of the waveform data obtained by the inverse frequency analysis unit 34 from the waveform data. That is, the second envelope derivation unit 37 obtains an envelope of the waveform data of the AE signal obtained by inverse frequency analysis by the inverse frequency analysis unit 34 for frequency components in a frequency band that contributes to the evaluation of the penetration depth, from the waveform data A of the AE signal output from the AE sensor 21 and processed by the signal processing unit 22.

[0026] The fluctuation frequencies of the envelope derived by the second envelope derivation unit 37 and the time series data of the penetration depth D of the welding point 2 also match. Therefore, even in this case, the penetration depth calculation unit 35 can calculate the time series data of the penetration depth D of the welding point 2 from the temporal fluctuation of the envelope, based on the pre-stored relationship between the temporal fluctuation of the envelope derived by the second envelope derivation unit 37 and the time series data of the penetration depth D of the welding point 2.

[0027] When the envelope derived by the second envelope derivation unit 37 is used in this manner, it becomes possible to calculate the penetration depth D of the welded area 2 with high accuracy compared to when the envelope C derived by the first envelope derivation unit 36 ​​is used.

[0028] The penetration depth calculation unit 35 may calculate the penetration depth D of the welding point 2 using artificial intelligence (AI) that performs machine learning. That is, the penetration depth calculation unit 35 first learns the laser welding conditions from the laser control device 12, the time series data of the AE signal, and the amount of fluctuation in the penetration depth D of the welding point 2 as teacher data. Next, the penetration depth calculation unit 35 inputs the laser welding conditions and the time series data of the AE signal as explanatory variables to calculate the time series change in the penetration depth D of the welding point 2.

[0029] Here, the greater the number of pieces of training data, the higher the calculation accuracy of the penetration depth D. The training data and the time series data of the AE signal as the explanatory variable are waveform data of the AE signal output from the AE sensor 21 and processed by the signal processing unit 22, or waveform data obtained by performing inverse frequency analysis on the frequency components in the frequency band that contributes to the evaluation of the penetration depth by the inverse frequency analysis unit 34. When the time series data of the AE signal is the latter, the calculation accuracy of the penetration depth D of the welded portion 2 is higher than in the former case.

[0030] Next, each step of the laser weld evaluation method using laser weld evaluation device 10 configured as described above will be described mainly with reference to FIG. First, the AE sensor 21 detects an AE wave generated from a welding portion 2 of the workpiece 1 when the laser welding device 11 laser-welds the workpiece 1 (S1). Next, the signal acquiring unit 31 acquires in chronological order the AE signal output from the AE sensor 21 and processed by the signal processing unit 22 (S2).

[0031] Then, the frequency analysis unit 32 performs frequency analysis on the time-series waveform data of the AE signal acquired by the signal acquisition unit 31 (S3). Next, the frequency band extraction unit 33 extracts frequency components of a frequency band that contributes to the evaluation of the penetration depth at the welded portion 2 from the analysis result obtained by the frequency analysis unit 32 (S4).

[0032] Next, the inverse frequency analysis unit 34 performs inverse frequency analysis on the frequency components of the frequency band extracted by the frequency band extraction unit 33 to obtain waveform data (S5). After that, the penetration depth calculation unit 35 calculates the penetration depth D of the welded portion 2 based on the intensity of the waveform data obtained by the inverse frequency analysis unit 34 and the laser welding conditions from the laser control device 12 (S6).

[0033] Further, the first envelope deriving unit 36 ​​derives an envelope C from the time-series waveform data of the AE signal acquired in step S2 (S7). Then, the penetration depth calculating unit 35 calculates the time-series change in the penetration depth D of the welded portion 2 based on the time variation of the envelope C acquired by the first envelope deriving unit 36 ​​(S6).

[0034] Furthermore, the second envelope deriving unit 37 derives an envelope from the waveform data obtained by the inverse frequency analysis by the inverse frequency analysis unit 34 in step S5 (S8). Then, the penetration depth calculating unit 35 calculates the time-series change in the penetration depth D of the welded portion 2 based on the temporal variation of the envelope obtained by the second envelope deriving unit 37 (S6).

[0035] As configured as above, this embodiment provides the following advantages. The AE sensor 21 detects AE waves generated from the welding point 2 of the workpiece 1 during laser welding by the laser welding device 11. Of the waveform data of the AE signal output from this AE sensor 21 and processed by the signal processing unit 22, the frequency analysis unit 32, the frequency band extraction unit 33, and the inverse frequency analysis unit 34 determine waveform data of frequency components in a frequency band that contributes to the evaluation of the penetration depth at the welding point 2, and the penetration depth calculation unit 35 calculates the penetration depth D based on this waveform data. As a result, the penetration depth D by laser welding can be measured with high accuracy.

[0036] Although the embodiment of the present invention has been described above, this embodiment is presented as an example and is not intended to limit the scope of the invention. This embodiment can be implemented in various other forms, and various omissions, substitutions, changes, and combinations can be made without departing from the gist of the invention, and these substitutions, changes, and combinations are included in the scope and gist of the invention, as well as in the scope of the invention and its equivalents described in the claims. [Explanation of symbols]

[0037] 2...welding point, 10...laser welding evaluation device (welding evaluation device), 11...laser welding device, 21...AE sensor (sensor), 31...signal acquisition unit, 32...frequency analysis unit, 33...frequency band extraction unit, 34...inverse frequency analysis unit, 35...penetration depth calculation unit, 36...first envelope derivation unit, 37...second envelope derivation unit, A...waveform data, C...envelope, D...penetration depth, L...laser light

Claims

1. a sensor for detecting elastic waves generated from a welding point during welding; a frequency analysis unit that performs frequency analysis on waveform data of the detection signal output from the sensor and processed; a frequency band extraction unit that extracts, from the analysis result obtained by the frequency analysis unit, a frequency component of a frequency band that contributes to an evaluation of a penetration depth at the welded portion; an inverse frequency analysis unit that performs inverse frequency analysis on the frequency components of the frequency band extracted by the frequency band extraction unit to obtain waveform data; a penetration depth calculation unit that calculates the penetration depth of the welded portion based on the intensity of the waveform data obtained by the inverse frequency analysis unit.

2. a first envelope deriving unit for determining an envelope of waveform data of the detection signal output from the sensor and processed; 2. The welding evaluation device according to claim 1, wherein the penetration depth calculation unit is configured to calculate a time-series change in the penetration depth of the welded portion based on the temporal variation of the envelope obtained by the first envelope derivation unit.

3. A second envelope deriving unit is further provided for determining an envelope of the waveform data obtained by the inverse frequency analysis unit, 2. The welding evaluation device according to claim 1, wherein the penetration depth calculation unit is configured to calculate a time-series change in the penetration depth of the welded portion based on the temporal variation of the envelope obtained by the second envelope derivation unit.

4. 2. The welding evaluation device according to claim 1, wherein the penetration depth calculation unit further has a function of learning welding conditions, time series data of elastic waves, and a fluctuation amount of the penetration depth as teacher data, inputting the welding conditions and the time series data of the elastic waves as explanatory variables, and calculating the time series change of the penetration depth.

5. 5. The welding evaluation device according to claim 4, wherein the time-series data of the elastic waves is waveform data of a detection signal output from a sensor and processed, or waveform data obtained by performing inverse frequency analysis on frequency components in a frequency band that contributes to evaluation of a penetration depth.

6. A step of detecting elastic waves generated from a welding point during welding by a sensor; a frequency analysis unit performing frequency analysis on waveform data of the detection signal output from the sensor and processed; A step in which a frequency band extraction unit extracts frequency components of a frequency band that contributes to an evaluation of a penetration depth at the welded portion from the analysis result obtained by the frequency analysis unit; An inverse frequency analysis unit performs inverse frequency analysis on the frequency components of the frequency band extracted by the frequency band extraction unit to obtain waveform data; a penetration depth calculation unit calculating the penetration depth of the welded portion based on the intensity of the waveform data obtained by the inverse frequency analysis unit.

Images