JPH0215822B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for detecting weld defects by acoustic emission. More specifically, the present invention includes:
Behind the arc during welding, the weld metal after solidification is
Relating to a method for detecting welding defects by detecting acoustic signals emitted from the defects by an AE sensor (acoustic emission sensor) in conjunction with a welding torch when the temperature range is between 700℃ and 500℃. It is something.

It is well known that when submerged arc welding is performed with high current and deep penetration, such as MIG welding, hot cracking is likely to occur as a welding defect. These hot cracks occur at or just below the solidification temperature range of the weld metal, and include vertical cracks, horizontal cracks, root cracks, and microscopic cracks that can be seen under a microscope. This is extremely dangerous as it can lead to 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. Furthermore, pressure vessels etc. to be welded,
As the plate thickness becomes extremely thick, it is difficult to inspect each weld layer by layer using current non-destructive inspection methods, and inspection must be performed after all layers have been welded, causing defects to be detected. Even if they were repaired, the number of man-hours required for repair and quality would be reduced.

Therefore, there has been a demand for a method that can detect welding defects in real time, but until now there has been no reliable method.

Here, the present invention attempts to provide a method that is not affected by various noises and can detect weld defects as early as possible after solidification of weld metal.

The method according to the present invention monitors the temperature range of 700°C to 500°C during the cooling process of the weld metal behind the welding torch, and reliably determines the extent and location of defects without being affected by arc noise, cracking noise of flux, etc. The feature is that the detection is done quickly and in real time.

Embodiments of the present invention will be described below based on the drawings. FIG. 1 is a configuration explanatory diagram showing an example of an apparatus for implementing the first method of the present invention.

In this figure, 1 is a base material, 2 is a flux layer, 3 is a welding torch, 4 is a flux hopper nozzle, and 5 is a weld metal. Here, an example of application to submerged arc welding is shown.

Distance behind the arc point directly below welding torch 3 in the welding direction of the torch (arrow direction)
A master sensor M for detecting acoustic emission sound is installed at the position l ₀ , and a similar first slave sensor M is installed immediately in front of this master sensor M and at a distance l ₁ from the torch. A sensor S ₁ is installed, and behind the master sensor M and at a spacing distance l ₂ from the torch, a similar second slave sensor S ₂ is installed.

Here, l ₀ , l ₁ , and l ₂ are the speed V (cm/sec) of the welding torch 3, and the time t ₁ (sec) required for the weld metal to cool down to 700°C after passing through the welding torch behind the welding torch. ), similarly, if the time required for the weld metal to cool down to 500°C is t ₂ (sec), it is defined by the relational expressions (1) and (2).

(l ₀ +l ₁ )/2=V×t ₁ …(1) (l ₀ +l ₂ )/2=V×t ₂ …(2) Therefore, in advance, V, t ₁ , t ₂ By measuring , l ₀ , l ₁ , and l ₂ (cm) can be determined. Note that, generally, l ₀ =l ₁ +l ₂ /2.

These sensors M, S ₁ and S ₂ are arranged in a line along the joint line on the surface side of the base metal 1, and as welding progresses, they are integrated with the welding torch 3 without changing their mutual positional relationship. It moves along the groove 17.

Reference numeral 6 denotes a signal processing device that processes detection signals from each of these sensors M, S ₁ and S ₂ , and the output data thereof is recorded and displayed by a recording device 7 . That is, within the signal processing device 6, noise removal is performed based on the detection results of the slave sensors S ₁ and S ₂ , and quantities related to the magnitude of the acoustic emission sound based on the detection results of the master sensor M, such as a ring-down count, Measurement of voltage effective value, energy, etc., and temporal determination are performed. The location where the acoustic emission occurs is represented by a certain length a. That is, in FIG. 1, this is represented as an area closer to master sensor M from the midpoint of each of master sensor M and slave sensors S ₁ and S ₂ .
Equation (3) is obtained.

a=l ₀ +l ₂ /2-l ₀ +l ₁ /2=l ₂ -l ₁ /2...
(3) The length a expressed by equation (3) is the length of the welding torch 3 during the time (t ₂ - t ₁ ) required for the weld metal to cool from 700°C to 500°C behind the welding torch 3. It corresponds to the traveling distance {V×(t ₂ −t ₁ )}. In other words, when the master sensor M advances from the right side of FIG. 1 together with the welding torch 3 and reaches the right side of the area indicated by the symbol a in the figure, the weld metal is at 700°C.
Proceeding further, the temperature is set so that the left side of the area is reached after a time (t ₂ −t ₁ ), and the weld metal at that time becomes 500°C.

These detection results are recorded by the recording device 7 as data for each welding material. Further, if the defect is determined to be harmful, the position where the harmful defect occurs is marked with the marker 16. With this, it is possible to instruct the position to be cleaned in a subsequent process. Note that the marker 16 is configured to move together with the sensors M, S ₁ , S ₂ and the like.

In the signal processing device 6, each sensor M, S ₁ ,
The signals captured by S ₂ pass through a preamplifier 8, a filter 9, and a main amplifier 10, and these amplify only the necessary frequency components in the signal so that they can be easily processed electrically, and the time difference comparator circuit 11. Here, the detection signal from the master sensor M is compared with the detection signals of both slave sensors S ₁ and S ₂ in terms of signal arrival order, and based on the result, the detection signal from the master sensor M is ranked first in the order of arrival. Only when this is the case, this is taken out as a defect occurrence signal. In other cases, it is removed as noise. Each signal from each sensor M, S ₁ , S ₂ inputted to the time difference comparison circuit 11 is, for example, A to A in FIG.
There are six types F, and among these, as shown in E and F, when the signal from the master sensor M arrives earlier than the signals from both slave sensors S ₁ and S ₂ , it is output as a defect detection signal. The defect detection signal output from the time difference comparison circuit 11 is sent to the signal detection circuit 12, where the defect detection signal (AE Find the ringdown count, effective voltage value, or energy value for the wave). The ring-down count number is determined by counting the number of times the AE wave crosses a predetermined threshold with a positive (or negative) slope. Also, the effective voltage value is AE
It can be obtained by finding the DC output proportional to the effective voltage of the wave.

Furthermore, the energy value is a concept expressed as ∫{V(t)} ² dt when the AE wave is expressed as V(t), and can be obtained by the following method. As shown in Figure 5, AE
The wave (FIG. 5a) is subjected to double detection to obtain a double detection signal (FIG. 5b). Then, the double detection signal is subjected to envelope detection, the envelope detection signal (Fig. 5c) is discriminated using a predetermined threshold value, the discriminated envelope detection signal is subjected to voltage-frequency conversion, and then the frequency The converted signals (FIG. 5d) are integrated, and the integrated signal (FIG. 5e) is output as an energy value indicating the relative magnitude of energy. The ringdown count, effective voltage value, or energy value obtained as described above is sent to the harmfulness evaluation circuit 14. The toxicity evaluation circuit 14 judges the sent detection results,
The degree of toxicity is evaluated and the result is output to the recording device 7. At the same time, if it is evaluated as a harmful defect, the marker 1 is sent to the marker via the marker activation circuit 15.
Command the operation of 6.

Through this series of operations, the cooling process of the weld metal near the master sensor M is 700℃~
By extracting only the acoustic emissions within a temperature range of 500°C, it is possible to accurately detect the extent and location of welding defects.

Here, set the weld metal monitoring range to weld metal temperature 700
The reason why it is included as a part of the temperature range from .degree. C. to 500.degree. C. will be explained next with reference to FIG.

Figure 2 shows the ring-down count of the first AE signal from the master sensor during welding of ASTM-A387 steel, displayed on the CCT curve of the weld metal along with the time-temperature curve. this
In the CCT curve, Ac ₃ is the temperature at which ferrite changes to austenite when heated while maintaining equilibrium when the carbon content is 0, F is the region where ferrite precipitates, A is the austenite region, and Zw is the intermediate structure. The region M is the martensite region. In this graph, the horizontal axis represents the elapsed time in seconds, with the time when the welding torch passes the AE sensor position as 0, and the time from 1 second to 10,000 seconds after the welding torch passes is expressed on a log scale. It is displayed. The vertical axis is the temperature of the weld metal (Temp) and the cumulative value of the ring-down count of the AE signal that first arrived at the master sensor. The curve Temp is the cooling curve of the weld metal.
It shows that the temperature changes from 1500℃ to around 170℃.
In addition, curves #2163 and #2161 are curves showing the cumulative value of AE ringdown counts, #2163 is data on defective welds due to hot cracking, and #2161 is data on defective welds due to hot cracking.
is the data for a sound weld.

In the data for #2161 and #2163, the difference is clearly visible between about 18 seconds and 25 seconds after passing the welding torch, which corresponds to a temperature range of about 700℃ to 500℃. ℃, and in this temperature range, the cumulative ring-down count of defective welds increases rapidly. On the other hand, no significant change was observed in the signal from the sound weld, and a comparison of both data revealed that AE signals associated with cracking occur frequently in this temperature range.

According to the temperature measurement results, under these welding conditions, the period around 20 seconds after the welding torch passes the master sensor position corresponds to the start of bainite transformation for this weld metal, but at this time the sound weld is
There is no sign of a rapid increase in AE, and it can be said that there are very few AE signals related to bainite transformation.

In this way, the experimental results shown in Figure 2 show that the signals from weld defects are noticeable in the temperature range of approximately 700°C to 500°C during the weld metal cooling process, and that the weld metal reaches this temperature range behind the welding torch. If the measurement monitoring area by the master sensor M is set in a certain manner, defects such as hot cracking can be reliably detected.

Figure 2 shows the properties of ASTM-387 steel, but in general, steel exhibits properties as an elastic solid;
The temperature that has the ability to generate "sound" due to large local plastic deformation, etc. is approximately 700 degrees Celsius or lower, and
At temperatures below approximately 500°C, the sound of slag peeling may become noise. Therefore, in order to detect defects in steel using AE signals, it is sufficient to measure in a temperature range of approximately 700°C to 500°C.

FIG. 3 is a configuration explanatory diagram showing an example of an apparatus for realizing the second method of the present invention. In this figure, blocks with the same reference numerals as in FIG. 1 have the same functions, and this example also illustrates a case where the block is applied to submerged arc welding.

In this device, the first master sensor M ₁ is placed at a first position at a distance lo ₁ behind the arc point directly below the torch in the welding direction (arrow direction) of the torch, and further behind the torch. A second master sensor M ₂ is placed at a second position at a distance lo ₂ (rear position separated by Δl from M ₁ ). In addition, a first slave sensor S 1 is arranged immediately before the first master sensor M ₁ at a distance of △l ₁ and at a _distance of l ₁ from the welding torch.
and a second master sensor
Immediately after M ₂ , a second slave sensor S ₂ is also arranged at a distance of l ₂ from the welding torch, separated by a spacing distance Δl ₂ . Here, the midpoint between the first master sensor M ₁ and the first slave sensor S ₁ corresponds to the position where the weld metal cooling process reaches 700°C,
Also, the midpoint between the second master sensor M ₂ and the second slave sensor S ₂ is the weld metal cooling process 500
They are each selected to correspond to the position where the temperature is ℃.

The torch speed is V (cm/sec), the time required for the weld metal to cool down to 700℃ after passing through the welding torch behind the welding torch is t ₁ (sec), and similarly the weld metal is
If the time required to cool down to 500℃ is t ₂ (sec), then l ₀₁ (cm), l ₀₂ (cm), l ₁ (cm), l ₂ (cm), △
l ₁
(cm) and Δl ₂ (cm) can be expressed by the relationships shown in equations (4) to (7).

△l ₁ +l ₁ =l ₀₁ ...(4) △l ₂ +l ₀₂ =l ₂ ...(5) l ₁ +l ₀₁ /2=V×t ₁ ...(6) l ₁ +l ₀₂ /2=V ×t ₂ ...(7) The first and second master sensors M ₁ and M ₂ and the first and second slave sensors S ₁ and S ₂ all detect defects that occur during cooling in the weld metal 5. They are arranged in a line along the joint line on the surface side of the base material 1, and as welding progresses, they move together with the torch 3 without changing their relative positional relationship. ing. Note that the distances △l ₁ and △l ₂ between the master sensor and slave sensor are
The lower limit is determined by the response of each sensor and the resolution of the signal processing system, 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 inputs the signals from each sensor, and performs noise removal based on the detection signals from the slave sensors S ₁ and S ₂ and the master sensor.
The magnitude and generation position of the acoustic emission sound are detected using the detection signals M ₁ and M ₂ , and the detection results are output to the recording device 7 . The recording device 7 records the detection results as data for each welding member. Further, when a harmful defect is detected, the position where the harmful defect occurs is marked with a marker 16 via the marker activation circuit 15.

In the signal processing device 6, first, among the sounds detected by the master sensors M ₁ and M ₂ , acoustic components generated near the torch ahead of the position of the first master sensor M ₁ are compared with the detection output of the slave sensor S ₁ . Similarly, among the sounds detected by master sensors M ₁ and M ₂ , the second
The acoustic components generated behind the position of the master sensor _M2 are removed by comparison with the detection output of the slave sensor _S2 . i.e. master sensor
Only when the detection signals of M ₁ and M ₂ are in the states E and F in Figure 4 (when the detection signals of M ₁ and M ₂ are earlier than the detection signals of S ₁ and S ₂ ), those signals are It is output to the signal detection circuit 12. This eliminates arc noise, which accounts for most of the acoustic components of the former, and cracking of flux, which accounts for most of the acoustic components of the latter, as well as crackling noises, etc., while the weld metal cools down to 700℃ behind the welding torch. 500
The monitoring area is only within the range of ℃. Furthermore, as described above, the signal processing device 6 uses the detection signals of the master sensors M ₁ and M ₂ from which noise has been removed to detect the acoustic emission sound generated only within the distance (l ₀₂ −l ₀₁ ). The degree of harmfulness and position of the defect are detected and outputted to the recording device 7. The defect occurrence position is determined by equation (8) as the distance X from the master sensor that detected the acoustic emission sound first.

X = (l ₀₂ - l ₀₁ ) - △t・V _AE /2 ... (8) However, △t: Detection time difference of acoustic emission sound at master sensors M ₁ and M ₂ V _AE : Weld metal or base metal In FIG. 3, the signal processing device 6 is configured as follows. That is, each sensor M ₁ ,
The signals captured by M ₂ , S ₁ , and S ₂ pass through a preamplifier 8, a filter 9, and a main amplifier 10, respectively.
Only the necessary frequency components in the signal are amplified to facilitate electrical processing and applied to the time difference comparison circuit 11. Here, it is determined whether the signal is a defective signal or noise based on the order in which the signal arrives at the master sensor and slave sensor. Then, only the detected defective signals are sent to the signal detection circuit 12 and the position locating circuit 13. In the position location circuit 13, the master sensor measured by the time difference comparison circuit 11
The location of the defect is calculated from the signal arrival time difference Δt to M ₁ and M ₂ , and the result is sent to the hazard evaluation circuit 14 . Similar to the embodiment shown in FIG. 1, the signal detection circuit 12 calculates the ring-down count number, The effective voltage value or energy value is determined and the result is sent to the hazard evaluation circuit 14. The harmfulness evaluation circuit 14 determines the harmfulness of the detected defect based on the information sent from the position locating circuit 13 and the signal detection circuit 12, outputs it to the recording device 7, and determines whether correction is necessary. Regarding the determined defective part, the marker operation circuit 15 is operated to mark the position calculated by the position locating circuit 13 with the marker 16.

In this method, the reason why the weld metal monitoring range is set to the weld metal temperature range of 700° C. to 500° C. is the same as that of FIG. 1 explained using FIG. 2.

As explained above, according to the method of the present invention,
Welding defects can be reliably detected in almost real time without being affected by various noises. Therefore, welding defects can be corrected at the optimal timing, and negative effects on process, quality, and cost can be minimized.

[Brief explanation of drawings]

1 and 3 are configuration explanatory diagrams showing an example of an apparatus for realizing the method of the present invention, and FIG.
The figure is a diagram showing an example of the experimental results that are one of the basis of the present invention. FIG. 4 is a time chart showing the relative relationship between the detection signals of the master sensor and the slave sensor, and FIG. 5 is a time chart showing the waveforms of each signal in the processing process for determining the energy value. 1...Base metal, 2...Flux, 3...Welding torch, 5...Weld metal, 6...Signal processing device,
M, M ₁ , M ₂ ... Master sensor, S ₁ , S ₂ ...
slave sensor.

Claims (1)

[Claims] 1. A first slave sensor that moves integrally with the welding torch, without changing its relative position, toward the rear of the point directly below the arc of the welding torch that moves at a predetermined welding speed. , a master sensor, and a second slave sensor are arranged in sequence, and the arrangement position of the first and second slave sensors is such that the midpoint between the first slave sensor and the master sensor corresponds to around 700°C during the weld metal cooling process. At the same time, the midpoint between the second slave sensor and the master sensor is selected to correspond to around 500°C during the weld metal cooling process, and each of the sensors is configured to detect the weld metal that propagates through the base metal and weld metal during welding. The acoustic emission sound from the welding defect that occurs during cooling of the metal is captured, and the signal arrival order of the detection signal by the master sensor and the detection signals by the first and second slave sensors is compared, and the detection signal by the master sensor is detected. A welding defect detection method, comprising: extracting the signal only when the signal is in the first arrival order, and determining the extent and position of the welding defect from the energy, ringdown count, or effective voltage value of this signal. 2. A first slave sensor and a first master sensor that move integrally with the welding torch without changing their relative position to the rear side of the point directly below the arc of the welding torch that moves at a predetermined welding speed. , a second master sensor, and a second slave sensor are sequentially arranged, the first master sensor is arranged at a position corresponding to the weld metal cooling process at around 700°C, and the second master sensor is arranged at a position corresponding to the weld metal cooling process. Each of the sensors is selected to correspond to a cooling process of around 500°C, and each sensor captures the acoustic emission sound from a welding defect that occurs during cooling of the weld metal that propagates through the base metal and weld metal during welding, 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 detection signals from the first and second master sensors are in the first arrival order. A welding defect detection method characterized by extracting the detection signal only at certain times and processing this signal to determine the extent and position of the welding defect.