JPS58143264A - Detection of weld defect by acoustic release - Google Patents

Abstract

PURPOSE:To detect high-temp. cracks and the positions thereof non-destructively in real time, by moving master and slave ultrasonic sensors having prescribed relative positions with a welding torch integrally with said torch and detecting and deciding weld defects. CONSTITUTION:A welding torch 3 and a master sensor M are fixed in relative positions in such a way as to correspond to the timing when the temp. range vulnerable to generation of defects is attained in the cooling stage of weld zones according to welding speeds. Slave sensors S1, S2 are provided respectively before and behind the sensor M, and the sensors M, S1, S2 move integrally with the torch 3. The ultrasonic detection outputs by the sensors S1, M, S2 are applied to a comparator 11 for time difference, and the arrival sequence of signals is compared. The detected output of the sensor M is taken in as a defect detection signal only when the output detected by the sensor M is faster than the detection outputs of the sensors S1, S2. Said signal is evaluated in an evaulation circuit 11 for the degree of harmness and the position is recorded with a recorder 7 and is marked with a marker 16, whereby the high-temp. cracking by welding and the position thereof is detected non-destructively in real time.

Description

[Detailed Description of the Invention] The present invention relates to a method for detecting welding defects by acoustic emission,
In particular, the present invention relates to a method for detecting internal defects and their locations by detecting acoustic emissions generated from cooling weld metal behind an arc during welding.

It is a well-known fact that when welding with deep penetration using high current welding such as submerged arc welding or MIG welding, hot cracking is likely to occur as a welding defect.1 This hot cracking occurs within the solidification temperature range of the weld metal. Or there are various types of cracks that occur at temperatures just below that temperature, such as vertical cracks, horizontal cracks, root cracks, or minute cracks that can be seen under a microscope, all of which are extremely dangerous and cause a decline in mechanical properties. These hot cracks may occur not only in the weld metal but also in coarse grain areas in the heat affected zone adjacent to the weld metal, and internal defects in these areas are difficult to detect by visual inspection.

An object of the present invention is to provide a method for non-destructively and real-time detection of welding defects, particularly hot cracks, regarding their extent and location behind the arc during welding. The aim is to improve quality.

In other words, this invention recognizes the occurrence of a welding defect by detecting the accompanying acoustic emitted sound, and the extent and location of the defect can be determined without being disturbed by arc noise, flux cracking noise, or before the transformation of the bath weld metal. This enables accurate detection in real time.

In one embodiment of the detection method of the present invention, a welding torch is moved at a predetermined welding speed, and the welding torch is located at a position directly behind the arc that corresponds to a time when defects are likely to occur during the cooling process of the welded part. a first slave sensor is located in front of the master sensor and rearward of the welding torch, and a second slave sensor is located rearward of the master sensor so that the master sensor and While moving the first and second slave sensors integrally with the welding torch without changing their mutual positional relationship, the acoustic emission sound during cooling of the weld metal propagating through the base metal and the weld metal is detected by these three slave sensors. The detection signal from the master sensor is compared with the signal arrival order of the detection signals from the first and second slave sensors, and the signal is detected only when the detection signal from the master sensor is in the first arrival order. In this way, only the acoustic emission sound generated near the master sensor is extracted, and its energy,
From the ringdown count or the effective voltage value,
The size and position of a defect within the weld bead corresponding to the range from the midpoint between the second slave sensor and the master sensor to the midpoint between the second slave sensor and the master sensor are determined.

Further, in another aspect of the detection method of the present invention, the first master sensor is arranged at a first position behind the point directly below the arc of the welding torch moving at a predetermined welding speed, The second position is located further back than the first position.
A master sensor is arranged, a first slave sensor is arranged immediately before the first master sensor, and a first slave sensor is arranged immediately before the first master sensor, and a first slave sensor is arranged immediately before the first master sensor.
A second slave sensor is placed immediately after the master sensor to connect the first and second master sensors and the first slave sensor during welding.
While moving the and second slave sensor integrally with the welding torch without changing their mutual positional relationship, these four sensors detect the acoustic emission sound during cooling of the weld metal that propagates through the base metal and the weld metal. The signal arrival orders of the detection signals from the first and second master sensors and the detection signals from the first and second slave sensors are compared, and the signal arrival order of the detection signals from the first and second master sensors is compared. Ranked first
By capturing the signal only earlier than that of the second slave sensor, it is possible to simulate only the acoustic emission sound generated within the weld bead corresponding to the range between the first and second master sensors. It is used to detect the size and location of occurrence.

To illustrate an embodiment of this invention, in FIG.
) is the base metal, (2) is the flux layer, (3) is the welding torch, (4) is the flux hopper nozzle, <5) is the hammer welding metal, and this example shows the case where it is applied to submerged arc welding. . 1 from the arc point directly below the welding torch (3)
・- Distance t behind in the welding direction (arrow) of H.

A master sensor for detecting acoustic emitted sound is placed at a position of 1. A similar first slave sensor (S,) across
A similar second slave sensor (S,) is also arranged behind the master sensor (B) at a distance of 4. These sensors (goods) (S +) (
SZ) are arranged in a line along the joint line on the surface side of the base metal, and as welding progresses, they move together with the torch (3) along the groove αη without changing their mutual positional relationship. Fresh urine. The position of the distance to, that is, the position where the master sensor (Incorporated) is placed, is a position on the ridge of the welding torch where the sensor is not affected by the thermal adverse effects of the welding arc, and at a predetermined welding speed. This position corresponds to the time when defects are likely to occur during the cooling process of the welded part. (
6) is a signal processing device that processes detection signals from these sensors (S + ) (SZ ), and its output data is recorded and displayed by a recording device (7). That is, in the signal processing device (6), the slave sensor (St
) Quantities related to noise removal based on the detection results of (SX) and the magnitude of acoustic emission sound based on the detection results of the master sensor (for example, ring down count, effective voltage value, etc.)
Energy, etc., and temporal determinations are made.

For the location of the acoustic emission, a certain length consisting of a)
It is expressed as. In other words, as shown in Figure 2, this means that the master sensor (goods) and slave sensor (S,) (Sz
)! 1, these detection results are recorded in the recording device (
7), the data is recorded for each welding material. Furthermore, if the defect is determined to be harmful, it is convenient to mark the position where the harmful defect occurs using a marker aQ, which will be convenient for cleaning in the subsequent process. It goes without saying that this marker αQ moves integrally with the sensors (goods) (81), (82), etc.

In FIG. 1, the signal processing device (6) has the following general configuration. That is, the sensor (goods) (S +) (S
The signals captured by t) are passed through a preamplifier (8), a filter (9), and a main amplifier (g), respectively, and these amplify only the necessary frequency components in the signal so that they can be easily processed electrically. The time difference comparison circuit (
Enter 11). Here, the detection signal of the master sensor M is compared with the detection signals of both slave sensors (s r Only this signal is taken out as a defect detection signal, the rest is removed as noise, and only this defect signal is sent to the signal detection circuit (6).The signal detection circuit α detects the amplitude, duration, etc. of the defect signal. , the detection result is sent to the harmful degree evaluation circuit α→ for judgment, and then the output is sent to the recording device (7).At the same time, for harmful defects, the marker operating circuit α is instructed to operate the marker αQ.

Figure 3 shows another embodiment of this invention.
The same reference numerals as in FIG. 1 indicate equivalent parts.

This example also shows the case of submerged arc welding, from the arc point directly below the torch to the welding direction (arrow) of the torch.
A first master sensor (M,) is arranged at a first position a distance t0 seven distances rearward with respect to , and a second master sensor (M, ) at a second position further rearward by a distance t0.
are arranged immediately before the first sister sensor (M,) with a distance of 1. A first slave sensor (St) is arranged at a distance of 4, and a second slave sensor (S,) is arranged immediately after the second master sensor (M,) at a distance of 4. These sensors (M+) (M*) (sl) (s
*) are all arranged in a line along the joint line on the base metal surface side in order to detect the acoustic emission sound based on the cracks that occur during cooling in the weld metal (5), and as the welding progresses. 1. The distance t from the torch (3) is such that it moves together with the torch (3) without changing the relative positional relationship. That is, the first position where the master sensor (M,) is located should be at a position behind the welding torch where the sensor is not affected by the heat of the welding arc. The second position where (M2) is placed is
If a flux removal device linked to the welding torch (3) is used, the location will not be affected by the flux removal device, and if no flux removal device is used, the location will take into account the attenuation of the acoustic emission sound. Note that the distances t1 and 4 are sensor ot+X
The lower limit is determined by the response of Mg)(s t)(sz) and the resolution of the signal processing system described below, and the value is selected as small as possible within this lower limit range, but is usually about twice the sensor diameter. The signal processing device [(6) K performs noise removal based on the detection signal of the slave sensor (SIXSg) and detects the magnitude and location of the acoustic emission sound based on the detection signal of the master sensor (M,) (Mz), This detection result is recorded as data and 1 for each welding material by a recording device (7). Furthermore, in the case of a harmful defect, it is also possible to mark the position where the harmful defect occurs using the marker (2), as in the previous embodiment.

In the signal processing device (6), first the master sensor (
Among the detection sounds of Ml) (Mz), the first master sensor (M
The acoustic component generated near the torch in front of the position of 1) is removed by comparing the relative arrival order with the detection output of the slave sensor (S,), and is similarly removed from the master sensor (Mj).
Of the sound detected by the slave sensor (S
,) is removed by comparison with the detection output of ), and this removes arc noise, which accounts for most of the acoustic components of the former, and cracking of flux or slack crackling, which accounts for most of the acoustic components of the latter, as noise. do. Furthermore, the signal processing device (6) uses the detection signal of the master sensor (Ml) from which noise has been removed as described above to determine the degree of harmfulness and location of the defect from the acoustic emission sound generated only within a certain distance. is detected and outputted to the recording device (7).In this case, the degree of harmfulness of the defect is determined by the amplitude and duration of the master sensor output, and the position of occurrence is determined by the direction in which the acoustic emission sound was detected first. As the distance X from the master sensor, x=(L-
Jt-v)/2.

In this equation, Δ( is the difference in detection time of the acoustic emission sound between the two manutar sensors, and ■ is the speed of sound in the weld metal or base material.

In FIG. 3, the signal processing device (6) has the following general configuration. That is, the sensor (Mθ(MzXs+)
The signals captured by (St) are transmitted to the preamplifier (8), respectively.
), filter (9), and main amplifier (2), only the necessary frequency components in the signal are amplified so that they can be easily electrically processed, and then enter the time difference comparator circuit a. Here, it is determined whether the signal is a defective signal or noise based on the order in which the signals arrive at the master sensor and the slave sensor, and only those recognized as defective signals are sent to the signal detection circuit (6) and the position locating circuit 0. The position locating circuit υ calculates the location of the defect from the signal arrival time difference Δt to the master cell (Ml) (Ml) measured by the time difference comparison circuit 09, and sends the result to the harmfulness evaluation circuit 4. The signal detection circuit (6) determines the amplitude, duration, etc. of the defective signal, and sends these results to the harmfulness evaluation circuit α-hiro. In the hazard evaluation circuit α4, the position evaluation circuit (2
) and the signal detection circuit (6),
The degree of harmfulness of the detected defect is determined and outputted to the recording device (7), and for defective parts that are judged to require repair, the marker activation circuit αQ is activated to place a marker at the position calculated by the position evaluation circuit member. Mark with αQ.

As described above, in this invention, it is possible to continuously monitor acoustic emissions only in a certain part of the weld metal that is being cooled behind the torch during welding, and in this part, welding changes from normal to normal. Therefore, the acoustic emission sound detected by continuous monitoring according to the present invention can be considered to be caused by a welding defect such as hot cracking in the weld metal or heat affected zone. It is easy to mark welded joints as defects based on the recorded data of their location and size, and it is also possible to significantly improve the quality of welded joints by performing early repair work on the marked locations. .

[Brief explanation of drawings]

Fig. 1 is a block diagram of a system configuration according to an embodiment of the present invention, Fig. 2 is a schematic diagram for explaining the area where the sound generation position is limited by the master sensor, and Fig. 6 is another embodiment of the present invention. It is a block diagram showing an example of. (1): Base metal, (2): Flux, (3): Welding torch, (4): Flux hopper, (5): Weld metal, (
6): Signal processing device, (7): Recording device, (8): Device amplifier, (9): Filter, (to): Main amplifier, αth:
Time difference ratio-circuit, a: signal detection circuit, 0: position locating circuit, a4: hazard evaluation circuit, 05 = marker activation circuit,
Q@: marker, α′i): welding groove, (MI) (MI
) (M): Master sensor, (81) (82) Nislepe sensor. Agent Patent Attorney Masatoshi Sato

Claims (2)

[Claims] (1) A master sensor is placed at a position e from the point directly below the arc of a welding torch that moves at a predetermined welding speed, and a first slave sensor is placed in front of this master sensor and at a position behind the welding torch. and a second slave sensor is arranged at a position beyond the master sensor, and the master sensor and the first and second slave sensors are integrated with the welding torch and their mutual positional relationship does not change during welding. These sensors detect the acoustic emission sound during cooling of the weld metal that propagates through the base metal and the weld metal while moving the weld metal, and compare the detection signal from the master sensor with the detection signals from the first and second slave sensors. Compare with the signal arrival order,
The signal is input only when the detection signal from the master sensor is in the first arrival order, and from the midpoint position between the first slave sensor and the master sensor to the second 1. A method for detecting a welding defect by acoustic emission, characterized by determining the presence or absence of a defect within a weld bead corresponding to a midpoint position between a slave and a master sensor. (2) A second master sensor is placed at a first position beyond the point directly below the arc of the welding torch that moves at a predetermined welding speed, and a second master sensor is placed further away from the first position. A second master sensor is placed at the position, a first slave sensor is placed immediately before the first master sensor, a second slave sensor is placed immediately after the second master sensor, and the welding is performed. The first and second master sensors and the first and second slave sensors are moved integrally with the welding torch and without changing their mutual positional relationship while propagating through the base metal and the weld metal. Acoustic emitted sound during cooling of the weld metal is detected by these sensors, and the signal arrival order of the detection signals by the first and second master sensors and the detection signals by the first and second slave sensors is compared, and the first and by capturing the signal only when the order of arrival of the signal to the second master sensor is earlier than that of the first and second slave sensors, welding corresponding to the range between the first and second master sensors. A method for detecting welding defects by acoustic emission, characterized in that only the acoustic emission sound generated within the bead is multiplied and the position of its occurrence is determined.