JPS5841441B2 - Trolley wire sliding surface width measurement method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention is capable of eliminating noise before and after the electrical signal in time, even when the waveform of the electrical signal obtained from the reflected light from the trolley wire sliding surface is considerably distorted. The present invention relates to a method that can accurately measure the sliding surface width of a trolley wire even if there is a contact wire.

In order to measure the amount of wear on the trolley wires that make up overhead contact lines, electrical inspection vehicles have traditionally been equipped with optical trolley wire wear measurement devices.

This optical type trolley wire wear measuring device measures the time width from the waveform of an electric signal corresponding to the light reflected from the sliding surface of the trolley wire, and calculates the amount of wear by converting the time width.

To explain this with reference to FIG. 1, the laser beam 2 output from the laser light source 2 passes through the half mirror 3 and the scanning mechanism 4 in a direction perpendicular to the trolley wire 1, and applies a force perpendicular to the sliding surface 1' of the trolley wire 1. It is irradiated in the direction.

In this case, the laser beam 2 is transmitted to the trolley wire 1 by the scanning mechanism 4.
is scanned at an appropriate frequency and speed to cover the entire deflection range.

The light reflected by the sliding surface 1' travels in the opposite direction along the same path as described above, and a part of the light is reflected rightward at the half mirror 3 as shown in the figure and is input to the photoelectric exchanger 5.

In the photoelectric converter 5, the electric signal waveform (hereinafter referred to as a signal wave or signal) is changed for the time width in which the reflected light from the sliding surface 1' exists.

For example, Figure 2 shows a signal wave when the sliding surface 1' is smooth and has good reflection, and the rising and falling parts have slope curves determined by the laser spot diameter and scanning speed.
The entire waveform is approximately trapezoidal, and the time width T at a value of % of the wave height value V corresponds to the sliding surface width W.

In the time width measuring circuit 6 of FIG. 1, the time width T is measured from the signal wave, and then in the time-voltage conversion circuit 7, a voltage proportional to the time width T is changed, and the amount of wear is determined by conversion in the subsequent stages. This is how it is.

However, the peak value 2 of the signal wave constantly fluctuates from scan to scan due to the reflection state of the sliding surface 1', and it is impossible to take a fixed value in advance as the detection level. The detection level for the waveform of the current scan was set using the percentage value of the peak value obtained from the previous scan, but this is the best method as long as there is a fluctuation in the peak value as described above. I can't say that.

Therefore, as an improvement measure, a circuit system has been devised in which a delay circuit is used and a value of % of the peak value of the signal wave itself obtained for each scan is used.

To explain this with reference to FIG. 3, first, a signal wave opening is created which is delayed from signal wave A by a time At by a delay circuit.

The time At is approximately the time width required for the rise of the signal wave A.

In addition, create a signal wave C whose amplitude is the H value of the original signal wave A, use an analog peak hold circuit to set it to a constant voltage of /2, and measure the time width T of the delayed signal wave opening using this value. It is something.

The feature of this method is that the points of the rising and falling portions are detected using the percentage of the wave height value at the rising portion, so it is accurate as long as the waveform unevenness at the trapezoidal top of the signal wave A is small to some extent.

However, in reality, the sliding surface 1' is not only not smooth, but also has reflections due to rust, etc., and in general, the signal waveform often has a distorted shape with severe unevenness changes, and depending on the size of the unevenness, the above Even with this delay comparison method, large errors are unavoidable.

FIG. 4 shows an example of a distorted signal waveform, and shows that the wave height of the original signal wave A' is smaller than that of the signal wave B near the falling portion of the trailing edge, resulting in a measurement error T''.

Also, contrary to Figure 4, errors may occur even when the peak value is small near the rising edge of the leading edge.To accurately measure the time width for such distorted signal waveforms, the above-mentioned delay is required. It is necessary to further improve the circuit system.

An object of the present invention is to measure the wear amount of the trolley wire or the width of the trolley wire sliding surface more accurately than before even if there is a considerably large uneven shape near the trapezoidal top of the signal wave.

Furthermore, the present invention effectively eliminates the influence of noise on measurement, which interferes with the measurement of the amount of wear of the trolley wire or the width of the sliding surface if noise is present before and after the signal wave. It is possible to make accurate measurements even in the presence of noise.

For this purpose, the present invention requires that, if the spot diameter and scanning speed of the light irradiated onto the trolley wire sliding surface are constant, the signal wave rises from the ground level to the wave height near the leading edge. The first signal (1 -1/n) A second signal with an amplitude of 1/n and a time delay of (1-1/n) Tu is created, and an electrical signal and a third signal with an amplitude of (1-1/n) Tu are delayed. If the comparison output between the signal No. 3 and the comparison output between the first signal and the second signal is ANDed, the pulse width of the AND output signal is the amount of time corresponding to the trolley wire sliding surface width. T u than
This was done by focusing on the fact that the time decreases by /n.

This principle is the first feature of the present invention.

However, since the above-mentioned AND signal includes noise, it is necessary to remove the noise.

Since noise should exist before and after the true AND output signal in time, if the above-mentioned AND output signal is gate-controlled by some gate signal, then the noise will become an AND output signal. It can be made so that it is not included in the output signal.

According to the present invention, the gate signal for this purpose is a signal wave and a second
The second feature is that the influence of noise is removed by gate-controlling the above-mentioned AND output signal using these gate signals.

The present invention will be described below, but before that, the principle of the present invention will be explained in detail.

First, in order to understand the method of measuring time width from a signal waveform according to the present invention, the characteristics at the rising and falling parts of the signal waveform will be explained.

In FIG. 5, a laser spot 2' (not necessarily a laser) having a circular cross section with a diameter d and a uniform light intensity passes through the front edge 1'' of the trolley wire sliding surface 1' at an arrow A.
If scanning is performed in the direction of It is clear that

The interval Tu between the rising point tp and the saturation point tp' of this curve
is the so-called rise time.

When the spot 2' escapes from the sliding surface 1' via the trailing edge 1'', it becomes a falling curve, contrary to the above case, and a falling time Td exists.

These times Tu. Td is uniquely determined by the diameter d of the spot 2' when the scanning speed is constant, so in the above case Tu and Td are equal.

Note that the curve when the light intensity inside the spot 2' is not uniform is determined by its distribution state, but in that case as well, Tu
It remains true that y T d are also equal.

Here, when looking at the diameter d of the spot 2', the smaller d is, the steeper the slope curve becomes, and even if the detection level deviates somewhat from the peak value H, the error that occurs is small and can be ignored, but in reality, For various reasons, a spot diameter d of about 111111 is used, and the measurement accuracy is required to be Q, I II, so it is required to use a value as accurate as possible for the detection level. It is something that will be done.

FIG. 6 explains the basic idea of the detection level for measuring the time width T from the signal waveform in this invention, and the rising part X and the falling part (
3] Assume that both X- are straight lines, and that the peak values V, , V- near the rising and falling parts are generally different; however, the rising time T'u and the falling time T
d are equal and known according to the above discussion.

Here, the delayed signal wave H delayed by Tu/2 using the delay circuit is converted into a compressed signal wave with a peak value of S/2 for each of the original signal wave 2 and the delayed signal wave H using the amplitude compression circuit. , G is created, and G is used as a detection level for the compressed signal wave.

First, in the rising part, the comparison circuit detects the intersection of the rising straight line of the delayed wave E and the compressed signal wave, and the point at time t is detected.
Get b.

The point is also the midpoint of the straight line R.

Next, in the falling part, the falling straight line CD of the original signal wave 2
The intersection point C between and the delayed wave is detected in the same way at time t.

get it. Point C is also the midpoint of straight line CD.

T' can be changed by calculating the time width (to t6), but if Tu/2 is added to T' using a calculation circuit based on a simple proportional relationship, it is possible to calculate the correct time T for sliding surface 1. It is possible.

Although the above description assumes that the rising and falling portions are straight lines, it also holds true for any curved line.

Although Fig. 6 shows the case where the peak value V' of the falling part is smaller than the peak value ■ of the rising part, in general, the sizes of ■ and V' can be set arbitrarily; for example, It is clear that the above argument holds true even when the peak value (2) is smaller than the peak value vp' of the falling portion.

However, as stated at the beginning, if the reflecting tables on the sliding surface 1' are very large, the top of the trapezoid may become concave.

Furthermore, when the depth of the concave shape is large, erroneous measurements may occur in the detection method according to the present invention described above.

To explain this with reference to FIGS. 7a and 7b, in FIG. 1a, the original signal wave 2 has depressions E and F at approximately the center, and the depth thereof is less than % of the wave height value. , in this case, the time width T' is changed by the above-described detection method in the same manner as described above, and no particular error occurs.

However, in Fig. 7, the depth of the recess E/p/ exceeds % of the peak value, so the delayed compressed wave T and the original signal wave 2 intersect at point e, and the delayed wave H and the compressed wave cross at point f. As a result, one trolley wire is imaged as if it were two trolley wires, and each time is The width becomes T□, T2, which causes unreasonableness.

In this way, in order to deal with erroneous measurements such as one trolley wire being incorrectly determined to be two, a logic circuit makes corrections using the judgment condition that the width of the sliding surface 1' is less than a certain value. Although there has been a separate invention in which it is possible to vary the sliding surface width, here we will discuss a method to avoid such erroneous measurements as much as possible, regardless of this invention.

FIG. 8a is an explanatory diagram of a method in which the basic concept of setting the detection level as a percentage of the peak value is expanded to an arbitrary value of 1/n.

The figure is drawn with n=3, but the delay time of the delayed signal E is (1-1/n) Tu, and for the compressed signal wave, G is 1/n of the peak value vp of the original signal wave 2. is taking.

With these, by the same method as before, the intersection b' and the intersection g
are obtained respectively.

In this case, the position of point b/ is different in height, but on the horizontal axis as the time axis, it is always at the same position regardless of the value of n.

That is, tb=tb'. (On the other hand, the position of point g on the horizontal axis changes depending on H, and the measurement time width (
t g t b) (=T')GkoTd/n (=
By adding Tu/n), the correct time width T can be obtained.

Such a detection method is possible because n is appropriately determined in advance and Tu=Td is known.

Figure 8 shows an example in which the detection method (n = 3 > explained in Figure 8a) is applied to the original signal wave 2, which has a concave portion with a depth of approximately 60% of the peak value. Defects like water,
That is, it is shown that correct measurement is performed without a situation in which the time width is divided.

However, although FIG. 8 shows a case where the wave height values of the rising and falling parts are equal, it is of course possible to perform accurate measurement as described above even when these are different.

However, as the value of n increases, the amplitude of the compressed signal wave becomes smaller, resulting in a deterioration in accuracy. Therefore, when selecting the value of n, consider the depth of the recess in the signal waveform in advance and select an appropriate value (
This should be determined.

The above explanation is a countermeasure according to the present invention for the case where a signal waveform is distorted, but there is also the problem of erroneous measurement due to noise existing before and after the signal wave.

In other words, if there is noise during a time period during which no signal wave exists during the measurement scan time, this noise will also exist in the delayed signal wave and the compressed signal wave, and by comparing each signal wave with the comparator circuit, these noises will be detected. This is because each point of time corresponding to the intersection of is detected.

In the present invention, with this point in mind, a method for removing harmful detection signals due to noise has been devised and made practical, and this will be described later.

The present invention will be explained below with reference to FIGS. 9 and 10.

Of these, FIG. 9 shows a block diagram of one embodiment of the trolley wire sliding surface width measuring method according to the present invention, and FIG.
It is a figure which shows an example of the voltage waveform in the principal part of a figure.

In FIG. 9, the signal wave obtained from the reflected light from the trolley wire sliding surface is adjusted to an appropriate level by the level adjustment circuit 9 through the input terminal 8, and becomes the original signal wave 2 shown in FIG. (1-1,/n) in the delay circuit 10
It is delayed by Tu and becomes a delayed signal wave H.

Here, Tu is the rise time of a waveform known in advance, and n takes an appropriate value in consideration of the rate of change in the peak value of the original signal wave 2, as described above.

On the other hand, the amplitude of the original signal wave 2 is reduced by 1/1 by the amplitude compression circuit 11.
n compressed signal waves, these delayed signal waves H and compressed signal waves are compared in the comparison circuit 12, and their intersection point b ('=b') is at time t1. It has already been explained that the output waveform is detected as (=t6') and the output waveform is obtained.

However, as shown in the figure, the output waveforms are accompanied by irregular voltage changes ``chi'' and ``chi'' temporally before the time tb and after the constant voltage continues to some extent.

The reason why such voltage changes ``chi'' and ``chi'' occur is that there is noise 2' in the original signal wave 2 where there should be no signal, and this noise 2' is transferred to the delayed signal wave H and the compressed signal wave. This is because noises H' and H' are included similarly.

Therefore, in the comparator circuit 12, by detecting the intersection of noise ho' and noise to', voltage changes appear as chi' and chi' in the above-mentioned output waveform, and falling is also a harmful and unnecessary cause of measurement error. Of course it is a thing.

Now, a part of the delayed signal wave H is branched to the amplitude compression circuit 1.
3, a delayed compressed signal wave whose amplitude is compressed to ], /n is obtained, and this and the original signal wave 2 are compared in the comparison circuit 14, and the intersection of these two signal waves is detected and the output wave is obtained.

As in the case of the output wave, this output wave is also accompanied by harmful and unnecessary voltage changes by detecting the intersection of the noise 2' and the noise 2'.

In the present invention, the above-mentioned irregular and unnecessary voltage change portions are removed by the following method.

First, the peak value of the noise contained in the original signal wave 2 is determined in advance through experiments or the like.

Of course, the peak value of the noise is assumed to be considerably lower than the peak value of the signal wave.

■ A constant DC voltage with a value slightly higher than this noise level.

is determined and supplied from the noise removal comparison voltage generation circuit 15 to comparison circuits 16 and 17, and compared with the original signal wave 2 and the delayed signal wave H, respectively, to obtain output waves N and L.

Here, when looking at the rising points tk and t1 of the output waves N and R, as long as the value of ■o is a value that is somewhat smaller than the peak value of the original signal wave two, the time points tk and ti are both the same in terms of time. It is in front of tb.

Similarly, the falling time t 1 t i J of the output waves N and R
Both are behind time t2.

Therefore, if we use the AND circuit 18 to perform the logical product of the four output waves Q, S, N, and L, we can obtain the output wave O that continues from time tb to time t2, and the output waves Q and Z exist. The voltage changing portions ``chi'', ``, ``, ``, and ``, which were previously used, are removed.

As mentioned above, the pulse width of the output wave O obtained by the following method is shorter than the time width T corresponding to the sliding surface W by T u /n, so this correction is necessary, but the correction method Therefore, it is simple and practical to perform the voltage value correction described below without relying on time processing.

That is, in FIG. 9, the output wave O is the time-voltage conversion circuit (
(hereinafter referred to as the TV conversion circuit) 7, the voltage is integrated as shown in the integral wave wa and is proportional to time (tb tg).
The voltage vs is further held in the peak hold circuit 19 at the next stage during the scanning period.

On the other hand, the correction voltage generation circuit 22 generates a correction voltage Vc which is proportional to a pre-known time width (Tu/n), and is added to the above voltage (2) by the addition circuit 224 to create the sum of both voltages. However, this is the output signal 2 for the desired sliding surface width W, and is taken out from the output terminal 23.

Such a series of signal processing is performed for each scan, and a timing pulse synchronized with the scan is applied from the input terminal 20, and the TV conversion circuit 7 and sample hold circuit 19 are set and reset.

As described above, the trolley wire sliding surface width measurement method according to the present invention can measure the sliding surface width more accurately than the distorted signal waveform for a trolley wire sliding surface that is not smooth and has many opposites. The method of eliminating design noise caused by noise has been taken into consideration, the circuit configuration is extremely simple, all circuit elements are easily available, and design and manufacture are easy, making it superior in economic efficiency. , and is highly reliable, making a great contribution in this area.

In the present invention, the irradiation light is not limited to a laser beam.

Although the present invention has mainly been described with respect to the processing of signal waveforms obtained by scanning a laser beam, the trolley wire wear measurement device uses ordinary electric light irradiation (reflected light from an image pickup tube is received, and this image is captured inside the tube). A method based on the scanning principle is also used, and the measurement method according to the present invention can be directly applied to such a method.

[Brief explanation of drawings]

Figure 1 is a schematic block diagram of an example of a conventional trolley wire sliding surface width measuring method using a laser beam, and Figure 2 shows the reflected light from the trolley wire sliding surface and the electrical signal obtained from the reflected light. Figures 3 and 4, which show the relationship between waveforms, are explanatory diagrams regarding the detection level setting method in the conventional method,
FIGS. 5 to 8 a and b are diagrams for explaining the present invention in detail, and FIGS. 9 and 10 are schematic circuit configurations and voltages at important parts in one embodiment of the present invention. It is a figure which shows an example of a waveform. 8... Electric signal or signal wave input terminal, 10
... Delay circuit, 11, 13 ... Amplitude compression circuit, 12.14, 16, 17 ... Comparison circuit, 15 ... Comparison voltage for noise removal Generation circuit, 1
8...ANDIE circuit, 19...Peak hold circuit, 21...Correction voltage generation circuit, 2
2...Addition circuit.

Claims (1)

[Claims] 1 Trolley wire sliding surface (The reflected light from the trolley wire sliding surface obtained by scanning the trolley wire sliding surface width direction with a spot light irradiated orthogonally to the trolley wire sliding surface is converted into an electrical signal, Measuring the trolley wire sliding surface width from the signal waveform Two methods of measuring the trolley wire sliding surface width: If the spot diameter and scanning speed of the irradiation light are constant, the electrical signal will rise from the ground level to the wave height. The time T required to rise to
Under the condition that u and the time Td required for falling from the peak value to the ground level are the same, the amplitude of each electric signal obtained from the light reflected from the trolley wire sliding surface is 1.
/n (n is generally 2 or a positive real number greater than 2, the same applies hereafter), a second signal delayed by a time of (1-1/n) Tu, and an amplitude of i/n. Yes (1-1
7n) Create a third time-delayed signal of Tu, a comparison output between the electrical signal and the third signal, a comparison output between the first signal and the second signal, the electrical An AND output signal obtained by ANDing all of the comparison output between the signal and a level slightly higher than the noise level and the comparison output between the second signal and a level slightly higher than the noise level. A trolley wire sliding surface width measuring method characterized in that the time obtained by adding the time T u /n to the pulse width corresponds to the trolley wire sliding surface width.