JPH0140308B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for detecting welds in a strip, the object of which is to automatically detect the welds contained in the strip while continuously moving the strip.

When manufacturing strips (cold rolled steel strips), the strips are usually welded to make them continuous, and then formed into coils of a predetermined size and subjected to various manufacturing steps to produce the final product. The strips produced in this way contain welds, and in the final process it is necessary to remove the welds, which are the largest defect.

Conventionally, a method for removing the welded portion was to visually detect the welded portion using a moving strip. However, visual detection of welds has its own limitations and has the disadvantage of slowing down the moving speed of the strip, which reduces productivity.

The present invention advantageously overcomes the above-mentioned drawbacks and makes it possible to detect welds even when the strip is moved at high speed. providing two receiving coils corresponding to the positions of the transmitting coils on the other side of the strip, detecting a voltage difference between the two receiving coils, and filtering the differential voltage signal with a bandpass filter; The strip is characterized in that a filtered signal is multiplied by a gain constant, the signal is divided by a sensitivity correction signal, and the detection signal obtained by the division is compared with a welding part discrimination criterion to detect a welded part. This is a welding part detection method.

In other words, the present invention divides the voltage difference between the two receiving coils by the sensitivity correction signal, so that the thickness of the strip can be changed from medium gauge (approximately 0.7 mm) to extremely thin gauge (approximately 0.1 mm) without sensitivity correction. A feature of the present invention is that it is possible to detect a welded part in the following state. Next, the present invention will be explained in detail with reference to FIGS. 1 to 3.

In Fig. 1, the transmitting coils 2 and 3 and the receiving coils 4 and 5 have the same number of turns and the same shape, and the transmitting coils 2 and 3 are connected in series and connected to an AC power source 1 of e _a [V] and f [Hz]. do. Such a transmitting coil 2,
3. The receiving coils 4 and 5 are provided with transmitting coils 2 and 3 on one side with the strip 6 in between, receiving coils 4 and 5 on the other side, and receiving coil 4 corresponding to transmitting coil 2. A receiving coil 5 is provided corresponding to the transmitting coil 3, and the two transmitting and receiving coils (2 and 4, 3 and 5) are arranged so as to be electromagnetically coupled. From this, the voltages e _b and e _c induced in the receiving coils 4 and 5 are governed by the thickness values of the transmitting and receiving coils 2 and 4 and the strip 6 sandwiched between the transmitting and receiving coils 3 and 5, and the respective plate thickness values If they are the same, e _b = e _c .

Figure 2a shows the results of actual measurement of the thickness change of a pickled welded part with a plate thickness of 0.1 mm and Figure 3 a with a plate thickness of 0.7 mm. As is clear from the figure, the welded portions in the pickling process are usually hard and therefore remain approximately 0.1 mm thicker than other locations even after cold rolling.

Next, to describe the detection of the welded part, in FIG.
5 is arranged in the longitudinal direction of the strip 6 as shown in the figure, and the spacing between the strips is adjusted to the distance where the difference in plate thickness at the pickling weld is the largest (approximately 60 mm), and the connection between the receiving coils 4 and 5 is made at a voltage _e Connect in the direction where and e _c cancel each other out. The difference between the two received voltages ( _eb - e _c ) is input to the bandpass filter 7 . As a result, the voltage difference (e _b −e _c ) is such that only the weld signal is extracted to a large extent. The differential voltage (e _b −e _c ) is filtered with a bandpass filter 7 to remove noise, input as a signal e _f to the amplifier circuit 8, and multiplied by an appropriate gain constant K ₁ .
It is input to the division circuit 12 as K ₁ e _f .

Figures 2b and 3b are plate thicknesses of 0.1 mm as mentioned earlier.
and change in plate thickness of pickled welded part with plate thickness of 0.7 mm (Fig. 2a,
This figure shows the value of K ₁ e _f for section X in FIG. 3a. As is clear from the figure, the magnitude of K ₁ e _f changes depending on the plate thickness, and becomes extremely small as the plate thickness increases. This is because in the electromagnetic coupling between the transmitting coils 2 and 3 and the receiving coils 4 and 5, as the plate thickness increases, the leakage magnetic flux increases and the interlinking magnetic flux of the receiving coils 4 and 5 decreases.

Actual measurement results show that the values of the received voltages e _b and e _c decrease approximately in inverse proportion to the square of the plate thickness.
If you try to distinguish welds by comparing the value of K ₁ e _f with the weld discrimination criterion R described later, you need to change the gain constant of K ₁ or change the weld discrimination criterion R depending on the plate thickness. There must be. The present invention eliminates the need to change the gain constant K ₁ , the weld zone discrimination criterion R, etc., and is characterized by dividing the above-mentioned K ₁ e _f by the sensitivity correction signal.

This sensitivity correction signal will be explained with reference to FIG. Any induced voltage in the two receiving coils 4 and 5
Take out e _c or e _b (indicated as e _c in Figure 1), input it to the amplifier circuit 9, and multiply it by the gain constant K ₂ to obtain K ₂ e _c
is input to the rectifier circuit 10 as follows. The rectifier circuit 10 converts the f (Hz) alternating current signal K ₂ e _c into direct current and inputs it to the averaging circuit 11 as Ec'. Averaging circuit 11
Then, smoothing is performed using a time constant T that is sufficiently longer than the fluctuation time of the signal K ₁ e _f due to the welding part, which is determined by the moving speed v of the strip 6 and the spacing between the two sets of transmitter/receiver coils 2, 4 and 3, 5 arranged in the longitudinal direction. It is averaged and input to the divider 12 as a signal Ec. This Ec is a sensitivity correction signal, and the sensitivity correction signal may be Eb, and it is not particularly important as long as it is a signal that prevents a difference in plate thickness by dividing the above-mentioned K ₁ e _f .

Next, after inputting the sensitivity correction signal E _c to the divider 12, the divider 12 performs division by K ₁ e _f /E _c , and the result is input to the rectifier circuit 13, where it is converted to direct current and used as the detection signal. E _f and input to the comparator circuit 14. The comparison circuit 14 compares the signal E _f with the welding part discrimination standard R,
A signal E _f that is larger than the welded part discrimination criterion R is determined to be a welded part, and a welded part detection signal H is generated, and the notification buzzer 15 is sounded at the same time.

The reason for dividing by the sensitivity correction signal E _c is that the value of E _c is the length of the strip 6 determined by the moving speed v of the strip 6 and the averaging time T.
Since it corresponds to the average plate thickness of vT, the above
Like K ₁ e _f , it is a value that is inversely proportional to the plate thickness.
E _f , which is obtained by rectifying the division result of K ₁ e _f /E _c , becomes a value that is not affected by the plate thickness, and the weld zone can be determined by fixing the weld zone discrimination criterion R.

Next, in Figures 2c and 3c, the signal E _f at the welded part where the strip thickness is 0.1 mm and 0.7 mm
The actual measurement data is shown. As shown in these figures, the signal E _f is detected as constant even if the strip thickness is different, making it possible to detect welds without considering differences in plate thickness and without making any adjustments. This method is extremely advantageous in terms of improved accuracy and productivity compared to the method of visually detecting welded parts.

The weld detection signal H is used to control the line after detection, and the output method is a contact or pulse signal. For example, it is advantageous to automatically stop the line or to reject the weld in cooperation with a tracking device and a shear device. be.

[Brief explanation of drawings]

FIG. 1 is an explanatory diagram showing the overall structure of the method of the present invention. Figure 2 is a waveform diagram showing an example of detecting welds in a strip with a plate thickness of 0.1 mm, and Figure 3 is a waveform diagram showing an example of detecting a welded part in a strip with a plate thickness of 0.7 mm.
These are waveform diagrams showing examples of detecting welds in strips, where a in each figure is the actual measured thickness of the strip including the weld, b is the weld signal for the strip in a, and c is the weld part signal for the strip in a.
shows the welding part signal after plate thickness correction of the detection signal of b. In the drawing, 6 is the strip, 2 and 3 are the transmitting coils,
4 and 5 are receiving coils, E _c is a sensitivity correction signal, E _f is a detection signal, and R is a welding part discrimination standard.

Claims (1)

[Claims] 1. Two transmitting coils are provided on one side of the strip, two receiving coils are provided on the other side of the strip corresponding to the positions of the transmitting coils, and a voltage difference between the two receiving coils is detected. The differential voltage signal is filtered with a band pass filter, the filtered signal is multiplied by a gain constant, the signal is divided by a sensitivity correction signal, and the detection signal obtained by the division is compared with the welding part judgment standard. A method for detecting a welded part of a strip, characterized in that the welded part is detected by using a welded part.