JPS6145763B2 - - Google Patents

Description

[Detailed Description of the Invention] The present invention relates to a track height deviation calculation device that directly outputs the true track curve shape by processing measurement data of track height deviation obtained by a track inspection vehicle. be.

Conventionally, various amounts of track deviation (hereinafter simply referred to as track deviation) have been measured by track inspection vehicles on railway tracks, but three-point measurement method and two-point measurement method are used to measure the height deviation of the track. It is being This will be explained below.

FIG. 1 shows a conventionally used height deviation detection mechanism. According to this mechanism, wheels 2 on a rail 1 are directly connected to an axle box 3 that is supported vertically movably with respect to a car body 7. , a fixing portion 4-1 of the hinge 4 is fixed onto the axle box 3. On the other hand, a support hardware 6 is fixed to the vehicle body 7, and a mechanical-electrical angle converter (hereinafter simply referred to as a converter) 8 is attached to the support hardware 6 as shown in the figure. The tip of the movable part 4-2 of the hinge 4 is the connecting rod 5.
are connected. Due to this mechanism, when the rail 1 is out of height, that is, is displaced upward or downward, the wheel 2 in contact with it is displaced upward or downward with respect to the car body 7, and this displacement movement is caused by the hinge 4 and the connecting rod. 5 to rotate the rotating shaft 8-1. The converter 8 outputs an electric signal having an amplitude depending on the rotation angle of the rotating shaft 8-1, and thereby the displacement of the rail 1 in the vertical direction with respect to the vehicle body 7 can be detected.

In the track inspection vehicle, actual height deviations are detected by using three or two of the above-mentioned height deviation detection mechanisms for the left and right rails, respectively. Figure 2a shows three displacement mechanisms for the front wheels 2-
1. It is a principle explanatory diagram of the three-point measurement method provided on the central wheel 2-2 and the central wheel 2-3. The vertical displacement of the rail 1 is y ₁ from the front wheel 2-1, and y ₂ and y ₃ from the center wheel 2-2 and rear wheel 2-3, respectively.
are simultaneously detected, and a separate arithmetic circuit is used to calculate the formula y=(y ₁
+y ₃ )÷2−y ₂ is calculated in real time,
This y is considered to be a height deviation. This value of y is the second
As shown in Figure b, the front wheel 2-1 and the rear wheel 2-3
This is the straight line connecting the lines, that is, the vertical displacement of the rail 1 at the position of the center wheel 2-2 with respect to the chord 9, and this is the principle of the three-point measurement method.

Usually, in a track inspection vehicle, the front wheel 2-1 and the rear wheel 2
-3 spacing is 10 m, i.e. a chord length l of 10 m is used, and the central wheel 2-2 is placed exactly in the center of the chord 9. Here, since the height deviation of the orbit usually has a wave shape (for example, a sinusoidal wave shape), if the length is called a "wavelength", the above-mentioned 3
Measurement data y using the point measurement method can measure height deviations as long as the wavelength is less than the chord length l, that is, 10 m, but as the wavelength becomes longer than the chord length, the resolution decreases and the detection accuracy deteriorates. The detection performance is determined by the chord length.

In Shinkansen trains, high-speed operation may cause the train to become more jittery, and preventing this is an extremely important issue. According to research on the wavelength of track deviations and train vibrations, it has been found that height deviations with wavelengths of several tens of meters, exceeding the chord length l, have a significant effect on vibrations. The two-point measurement method described next was developed and put into practical use.

Figure 3 shows the principle of this two-point measurement method, in which the vertical displacements y ₁ and y ₃ of the front wheel 2-1 and rear wheel 2-3 in the three-point measurement method described above are used as they are. It's becoming like that. Furthermore, a gyroscope 10 is installed in the car body 7 to constantly detect the angle α formed between the level line GL or GL' and the car body 7, regardless of whether or not the car body 7 is shaking such as pitching.
An arithmetic circuit calculates the formula y=lcosαtan {α+tan ⁻¹ (y ₃ −y ₁ )/l} in real time, and this y is defined as the height deviation for long wavelengths. As shown in FIG. 3, the value of y in this equation means the height difference of the rail 1 at the respective positions of the front wheel 2-1 and the rear wheel 2-3 with respect to the level line GL'. Although this two-point method has made it possible to measure long-wavelength track height deviations in Shinkansen trains and has made progress, it still has shortcomings in detection performance when examined in detail. In other words, in the two-point measurement method, the relative height difference of the rail 1 is measured between two points separated by the chord length l, but this does not directly represent the curved shape of the rail 1; It merely represents the differential value or something close to the differential value of the curve. To give a specific example of such a measured value y, for example, as shown in FIG. 1 exceeds the height deviation part a, the fourth value is used as the value of the height deviation y.
A waveform shown by curve A' shown in Figure c is obtained, and then when the rear wheel 2-3 crosses the height deviation part A, a curve shown by A'' with the polarity reversed is obtained.In other words, one height deviation part I is obtained. In addition to obtaining a curve I′ with a shape close to the true shape as the measured value y for
It appears at a position separated by a time corresponding to
Obviously it does not show the actual curve condition. Furthermore, FIG. 4d shows a case where the pitch deviation is continuous at a wavelength equal to the chord length l, and in this case, no pitch deviation is detected at all. FIG. 4e shows a case where the wavelength is approximately twice the chord length l, and the pitch deviation y can only be measured in such a wavelength range. As shown in these examples, the two-point measurement method has inherent drawbacks, such as ghosting or being completely undetectable in the height deviation y values for wavelengths below the chord length l.
The two-point measurement method is used exclusively for measurements in long wavelength ranges, and the three-point measurement method is used for measurements only in short wavelength ranges, as mentioned above, and both are used in conjunction to evaluate orbit height deviations. is being carried out.

However, the two-point measurement method described above measures the height difference relative to the level line, so
This does not immediately indicate the shape of the trajectory curve, and its evaluation requires a technique based mainly on experience, which inevitably causes inconvenience and inconvenience. Therefore, it is desired to develop a method or means to eliminate this and obtain an absolute shape, that is, a true trajectory curve, instead of a relative shape.

The present invention eliminates the drawbacks of the measurement data obtained by the conventional device described above, and performs arithmetic processing on the measurement data on the track inspection vehicle in real time.
The object of the present invention is to provide a track height deviation calculation device that can directly calculate and output the true curve shape of track height deviation.

The principles and features of the present invention will be explained below with reference to FIGS. 5 to 8.

The first point of the present invention is the measurement data (hereinafter referred to as two-point measurement data) by the two-point measurement method described above.
The purpose is to calculate the true height deviation shape by performing the necessary calculations on the . As mentioned above, two-point measurement data y
In principle, is the differential value of the trajectory curve (hereinafter referred to as the original curve), so in the following it will be expressed as △y instead of y, but since the chord length l is finite, it exists in the original curve. The short wavelength portion of the curve is based on the principle of difference rather than differentiation, and therefore the original curve cannot be reconstructed by simple integration. However, it should be noted that curves with long wavelengths where the chord length l can be approximately regarded as a differential can be approximately restored by a simple integral method.

As shown in Figure 5, the traveling direction of the inspection vehicle is taken as the x-axis, the original curve of rail 1 is expressed as y = (x), the measured value at point b is △yb, points a, b
Letting the values of (x) in (a) and (b), respectively, the following equation holds true.

△yb=(b)-(a)=(b)-(b-l) ∴(b)=△yb+(b-l) ...(1) Similarly, the measured value of point c is △yc, point c Letting the value of (x) in (c) be the following equation.

(c)=△yc+(c-l)...(2) These equations (1) and (2) express the relationship between the measured values △yb and △yc at arbitrary points b and c and the original curve (x). From these, the following equation generally holds true for any point x.

(x)=△yx+(x-l)...(3) In other words, to find out the original curve y=(x), measure △y continuously at each point x, and add the chord length to this value. This is possible by adding the value (x-l) of the original curve at the point (x-l) that is l behind the point (x-l) that has already been determined, and this is the basic principle and feature of the present invention.

Now, when executing the addition in equation (3), it is assumed that the value (x-l) of the original curve has already been obtained, but
Once the measurement is started by giving a known initial value at the measurement start point, the obtained value of (x) is memorized appropriately and added to the upper constant value △y delayed by the chord length l. Therefore, the method of setting initial values and performing delays is important for this execution. Setting the initial value will be explained later, but the delay method will be described as follows.
Generally, since the running speed of the inspection vehicle is not constant, the delay time with respect to the chord length l cannot take a constant value. Therefore, in the present invention, the measured value △y is sampled using a distance proportional pulse (hereinafter simply referred to as distance pulse) that is generated every fixed travel distance to create data at fixed intervals, and this is sent to the distance pulse through a shift register. After delaying by the chord length l by shifting in synchronization with
I am trying to use a method that adds up to . In this case, it is advantageous to perform these arithmetic operations digitally in terms of accuracy and the like.

As described above, by processing the two-point measurement data Δy, the original curve y=(x) can be obtained regardless of its shape or wavelength length. At this time, consideration of various shapes of wavelengths will be omitted, but since the details of the wavelength are determined by the sampling interval, that is, the resolution, this point will be briefly described.

First, the sampling interval needs to be less than a fraction of the wavelength of the curve that requires measurement. Furthermore, it is absolutely necessary that the sampling interval is exactly an integer fraction of the chord length l. The reason for this is so that the distance-delayed backward data can be added to the new measured value △y at the correct position. If the sampling interval is not an integer fraction of the chord length l, This is because addition is performed and there is a risk that an error will occur. Conventionally, distance pulses of 1 m have been used in track inspection vehicles, but if the chord length l is 10 m, it can be used as is since it satisfies the integer ratio condition.

Next, we will discuss problems in setting initial values.

FIG. 6 shows a case where an ideal track section 1-1, where the rail 1 is laid horizontally and the height deviation is zero, is selected as the measurement starting point of the track inspection vehicle.
Assume that the train starts from this trajectory section 1-1 in the direction of arrow A, and if the initial value is set so that all delay shift registers are cleared before measurement starts, the trajectory curve ahead can be calculated. It is something. However, the measurement accuracy of the height deviation required by the track inspection vehicle is required to be highly accurate (for example, 0.1 mm), and compared to this, the actual track always includes a slight difference in height. It is necessary to determine the initial value according to the point by some method, and it is necessary to set this in the shift register. Furthermore, once an unexpected error is mixed into the above-mentioned calculated value for some reason, the error remains as it is and does not disappear thereafter due to the principle of this method. That is, when an error occurs at a certain point, it appears not only in the output data at that point but also cyclically thereafter with a period of chord length l, and it is necessary to eliminate this error.

The second point of the present invention relates to correction means for setting initial values at the start of measurement and eliminating errors that occur during measurement. As these means, in the present invention, the above-mentioned two-point measurement data Δy and three-point measurement data y' (herein, three-point measurement is indicated by ') are constantly monitored while these measurement data are being measured. , a position where both are continuously below a certain small limit value over a section longer than at least the chord length l is detected.
Such a position with respect to Δy means that the rail 1 is horizontal or close to horizontal, and that the height deviation of the short wavelengths present in the rail 1 is small with respect to y'. On the condition that the detected positions by these two methods match, the shift register is rewritten as soon as possible after that point. As a first approximation of this rewriting, it is conceivable to clear all the contents of each shift register to zero, but in order to further improve accuracy, the following sine wave interpolation method is performed.

FIG. 7 is for explaining this sine wave interpolation method. A certain small limit value y'lim is given to the three-point measurement data y', and y' Assume that a waveform B such that <y′lim is detected by the method described above. At this time, regardless of the actual shape of waveform B, assume a sine waveform C extending between points p and p' whose maximum value is the peak value y'p at peak point p'' of waveform B, and x of waveform C. The value for each sampling point on the axis is determined and used as data for rewriting.In this case, rather than using the sine waveform C as described above, it is better to use the value of the waveform B itself as the new data. Although it is the best method, the reason why all the values at each point of waveform B can be obtained in the three-point measurement method is because the center wheel 2-2 of the inspection vehicle moves from the point p to the point p' shown in the diagram. At the time when y′p at point p″ is obtained, the shape of the waveform B between points p″ and p′ is still unknown, and correction of the measurement data is performed by changing the measured value y′ for point p″. Since it is necessary to execute the measurement immediately at the point when the measured value Δy for point p' is obtained in the two-point measurement method, the sine waveform C is substituted as described above. Note that the track height deviation is often well approximated by a sine wave due to the rigidity of the rail, and by setting the limit value y'lim as small as possible, the sine wave approximation is reasonable and practical. It is something.

The above-described correction method can be applied not only to correction in the middle of measurement, but also to the setting of initial values at the departure or measurement start point. 500m), or you can do either of these as you like. Analog circuits are suitable for executing the above-mentioned correction in terms of processing speed, and an analog sample and hold circuit, an analog comparator, and an analog function generator are used, and the correction data is rewritten digitally.

The third aspect of the present invention relates to the step of chart recording the calculated output of the true trajectory curve obtained as described above. As mentioned above, the true trajectory curve,
In other words, since the original curve is an integrated curve in a broad sense, it exhibits a large height difference as a whole when measured over a long distance. However, since the height difference required is for wavelengths within 100 m at most, height differences in wavelengths longer than that can be excluded from chart recording, and may be large depending on the size of the recording paper, the structure of the recording pen, etc. It is inconvenient that scaling or switching is required for height differences. Therefore, in the present invention, a method is adopted in which wavelength components longer than a certain length are removed. This results in
At the same time, elevation differences that are accumulated over time due to changes in sea level are also removed. The method to achieve this is to use the value of the original curve y=(x) obtained by the method described above,
This is achieved by finding a moving average value M using a weighted average method called a moving average method for a certain length range, and subtracting the moving average value M from the original curve y=(x). To explain this using Figure 8,
The moving average value M and the waveform of the difference between them are shown for the original curve D, and the waveform shows the waveform after unnecessary long wavelength deviation components (including linear gradient components) are removed and the necessary wavelength is removed. Only the ingredients are shown.

The circuit that obtains the moving average described above is a type of digital filter, but its principle, method, or circuit configuration is partially disclosed in a patent application (filing date: April 22, 1952, patent application filed by the same inventor as this invention). Gansho 52
−45779, name: compound orbital deviation calculation processing unit)
The detailed description is omitted as it is already known.

Now, the present invention will be explained with reference to FIGS. 9 to 12.

FIG. 9 is a block diagram showing the overall configuration of one example of the orbit height deviation calculation device according to the present invention. According to this, the two-point measurement system 1 installed on the track inspection vehicle
The height deviation measurement data Δy from 1 is given to an analog-to-digital converter (hereinafter referred to as an A/D converter) 14. On the other hand, the A/D converter 14 samples the measurement data △y at fixed distance intervals by the sampling pulse SP from the sampling pulse generator 13, which is controlled by a distance pulse that is separately provided in the track inspection vehicle. Digital amount〔△
y] (hereinafter, [ ] is added to indicate a digital quantity), the original curve [y] is obtained in the next stage original curve calculation circuit 15, and further, the digital filter 21
The longer wavelength component is cut off and output as [ys] to the terminal 21'. Further, ys converted into an analog quantity by a digital-to-analog converter (hereinafter referred to as a D/A converter) 22 is outputted to a terminal 22' and supplied to a chart recorder.

As a correction circuit for initial value setting or intermediate, 3-point measurement system 1 is used together with 2-point measurement data △y.
The measurement data y' from 2 is input to the corrected position detection circuit 16, and when a predetermined condition is satisfied, a GO signal is output. In addition, the three-point measurement data y' is input to the correction data creation circuit 17, and the sine wave function generator 1
The analog quantity correction data already mentioned in 8 is created,
The data is digitized by an analog-to-digital converter (A/D converter) 19 and updated and stored in a correction data register 20 every sampling pulse SP.
Therefore, as mentioned above, the orbital conditions are satisfied and GO
When the signal is obtained, the supply of correction data to the original curve calculating circuit 15 is started by this GO signal, and the correction data is rewritten one point at a time for each sampling pulse SP.

FIG. 10 shows the original curve calculation circuit 1 shown in FIG.
5 is a block diagram of an example of No. 5; FIG. The two-point measurement data [Δy] is input to one terminal of a digital binary adder (hereinafter referred to as ADD) 30. 31 is 10 stages with the required number of digits (for example, distance pulse is 1m, L
is 10m) shift registers 31-1 to 31
This is a group of shift registers consisting of -10 shift registers, and is used to perform the distance delay of measurement data as described above. That is, the shift registers 31-1, 31-2, .
ADD30 is connected to the other input terminal. With this configuration, the output data [y] of the ADD30 is transferred every sampling pulse SP, delayed by 10 m in 10 stages, and added to the new input [△y], as shown in equation (3).
The calculations shown in are executed. In this case, since the output data [y] of the ADD 30 has already been calculated, it will be simply referred to as calculated data hereinafter, but it goes without saying that this represents the original curve y=(x).

The point where the front wheel 2-1 of the inspection vehicle is present at any given time is x ₀ , and the sampling points behind that point are expressed as x _-1 and x _-2 , and each data corresponding to each point is set to 0. , -1, . . . , the input/output data of the ADD 30 and the data stored in the shift registers 31-1 to 31-10 at the relevant time maintain the relationship shown in FIG. In other words, new measurement data [△y] ₀ and
Calculated data for 10m backward [y] _-10 is ADD3
The new operation data [y] ₀ is output and stored in the shift register 31-1, and the contents of the shift register 31-10 become [y] _-9 , which is used for the next addition. It is something that you will have to wait for.

Here, correction of measurement data during measurement will be explained. FIG. 11 is a somewhat detailed block diagram of the correction position detection circuit 16 and correction data creation circuit 17 in FIG. 9, and FIG. 12 is a chart showing the data processed by the circuit in FIG. 11 and their positional relationship. be.

In Fig. 11, 161 and 163 are peak value determiners consisting of an analog sample and hold circuit and an analog comparator, which sample and hold the 2-point measurement data △y and the 3-point measurement data y', respectively, every 1 m using the sampling pulse SP. They are compared with fixed limit values ±Δylim and ±ylim, respectively, and a pulse signal is output when they are below the limit values. Further, 162 and 164 are so-called "10" pulse counters that output an ON signal when 10 consecutive pulses are input, and the peak value determiners 161 and 163 10m minutes), 1
The 0 pulse counters 162 and 164 output ON signals. In Figure 12, △y exceeds the limit value indicated by the dotted line at point x _-10 , as shown by waveform A, but is within the limit value at 10 points from x _-9 to x ₀ , so At point x ₀ , the output a of the counter 162 is
Turn on. Regarding y′, at this point, the measurement has been completed up to point x _-5 , and the output b of the counter 164 is ON, assuming that the previous 10 points are continuously within the limit value shown by the dotted line. ing.

166 is a peak position detector consisting of an analog sample hold circuit and an analog comparator.
The absolute value of the three-point measurement data y' sampled and held each time is compared with the previous value, and an ON signal is output only when the current value is smaller than the previous value. In waveform B of FIG. 12, the value |y' _-4 | for the point x _-4 is smaller than the value |y' _-5 | for the point x _-5 , so the output c turns ON at this point. The ON signals of these three outputs a, b, and c are logically multiplied by the AND circuit 165, and when all of them are ON at the same time,
GO signal is output. It should be noted here that the measurement point of the 3-point measurement data y' is always 5 m behind the measurement point of the 2-point measurement data △y, so the point at which each of the above ON signals is output is the point at which the front wheel If the line is redirected to point 2-1, the position shown in FIG. 12 will be obtained, and the GO signal will be output at point _x1 .

On the other hand, the three-point measurement data y' is generated by the sine wave function generator 1.
8, and 11 y′｜sinθ｜ ₁ on the output side
(θ=0°, 18°, 36°, 54°, 72°, 90°, 108
゜, 126゜, 144゜, 162゜, 180゜) values are output in parallel, and these values are output in parallel to the A/D converter group 191.
each digitized and sampled pulse
The SP updates and stores the data in the correction data register 20 every 1 m. Looking at this in Figure 12, at the time the GO signal mentioned above is output, the value of the sine wave whose peak value is the value [yd] = [y′ _-4 |sinθ|] for point x _-4 is the point Calculated for x _-9 , x _-8 , ... x ₀ , X ₁ , and the correction values [yd] _-9 , [yd] _-8 , ...
[yd] Stored as ₀ in the correction data register 20. In this case, [y] _-1 for point x ₁ is discarded. As described above, when the GO signal is input to the correction data register control 202 at point x ₁ , the control 202 gives a data transfer command to the correction data register 20, and the above correction value [[ yd〕 _-9 ,
... are sequentially transferred to the operating contact 32-2 of the switch 32 via the output terminal 203, the input terminal 33 in FIG. 10, and the ADD 34. The switch 32 is operated by the GO signal at this point, and the correction data is added to the ADD 30 in place of the existing data in the shift register group 31, and new data [△y]
In this way, the initial value is set or the correction data is rewritten midway. Here, if we consider the past calculation data [y] _-9 , ... in the shift register group 31, since these are originally the integral of △y in a broad sense, the difference in elevation is accumulated and This includes altitude differences. On the other hand, the above-mentioned correction data is only obtained under the condition that the rail 1 is horizontal or close to horizontal and the height deviation is extremely small, so the component corresponding to the above-mentioned height difference is defective. Therefore, in the above-mentioned rewriting, discontinuous steps occur in the value of the calculation data [y]. In order to avoid such inconveniences, in the present invention, a "1 data" register 35 is provided to constantly update and store the contents of the shift register 31-10, that is, the above-mentioned altitude difference data.
Therefore, when the GO signal is input, the data is held during rewriting and is sent to the ADD3 via the switch 36.
4, it is added to the correction data [yd]. This avoids the occurrence of a step. However, when such a thing is not necessary, the above-mentioned addition can be stopped by opening the switch 36.

Furthermore, a circuit for automatically executing rewriting using the correction data described above at appropriate distance intervals, for example, every 500 m, will be described. In Fig. 10, 40 is a correction interval signal circuit, to which a sampling pulse SP and a 500m distance marker signal separately used in the inspection vehicle from input terminal 41 are added, and it is turned ON for a certain distance every 500m. Make sure you get the signal. This ON signal and the GO signal
The flip-flop 43 is set by AND synthesis, and the ON signal of the set output Q is used to switch the switch 3.
2. However, if the GO signal is not obtained within the above-mentioned fixed distance every 500 m, the correction is not performed, but since the correction does not necessarily have to be performed every 500 m, this does not pose any problem. Regarding the termination of the above-mentioned rewriting operation, when the AND circuit 45 and the 10-pulse counter 46 count 10 sampling pulses SP after the start of the switching operation, the flip-flop 4
3 is reset, the switch 32 is restored and the rewriting is completed.

Next is how to set the initial value at the measurement start point, etc., but normally there is a preparation process in which the track inspection vehicle is pulled out from the side track to a predetermined position on the main track, and the above correction method is applied during this pulling out. The correction method can be carried out in the same way, and there is essentially no difference from the correction method used during the measurement run.
However, instead of the 500 m distance marker signal, it is necessary to provide an appropriate signal to the correction section signal circuit 40 by manual operation or the like.

Note that the sampling pulses SP used in each of the above-mentioned circuits are all created in the sampling pulse generator 13, and phase adjustment is required for each circuit due to the sequential operation of the entire circuit. This is done by generator 13. In addition, although the sampling pulse SP was explained as being at 1m intervals,
Of course, it is also possible to use a shorter distance from the viewpoint of accuracy and the like.

As explained above, the track height deviation calculation device according to the present invention eliminates unnatural and unnecessary ghosts that were mixed in the height deviation measurement data using the two-point measurement method conventionally performed with track inspection vehicles. This provides a so-called absolute shape curve that accurately shows changes in the height of the trajectory, making it easy to judge and evaluate data on deviations in the height of the trajectory without requiring any experience. In addition, this device uses a rational circuit system for timely and automatic correction of unexpected errors that occur during measurement, and for setting initial values necessary before starting measurement, improving accuracy. is measured. Furthermore, when recording calculation results, unnecessary long wavelengths may be distorted in height (including altitude differences).
By using a digital filter to remove the differences in height and recording only the height deviations in the wavelength range that are truly necessary for track maintenance, significant improvements have been made in many respects, including improved recording accuracy on chart paper. This makes it possible to measure orbital height deviations with high precision, making it an extremely important contribution to this field.

[Brief explanation of the drawing]

Figure 1 is a diagram showing the rail height deviation detection mechanism, Figure 2 a and b are diagrams for explaining the principle of the three-point measurement method, and Figure 3 is a diagram for explaining the principle of the two-point measurement method. Figures 4a to 4e are diagrams for explaining the relationship between the trajectory curve shape and measurement data in the two-point measurement method, and Figure 5 is a diagram showing the calculation principle of the orbit height deviation calculation device according to the present invention. FIG. 6 is a state diagram of the track inspection vehicle for explaining the initial value setting in the track height deviation calculation device according to the present invention, and FIG. 7 is a diagram for explaining the track height deviation calculation according to the present invention. FIG. 8 is an explanatory diagram of the sine wave interpolation method for eliminating errors occurring in the device. FIG. 8 is an explanatory diagram of the moving average method used in the orbit height deviation calculation device according to the present invention. FIG. Figures 10 and 11 are block diagrams showing the overall configuration of an example of the orbit height deviation calculation device, and Figure 12 is a block diagram showing the configuration of an example of the main circuit in Figure 9. 12 is a diagram showing each data processed in the circuit shown in FIG. 11 and their positional relationship. 11...2-point measurement system, 12...3-point measurement system, 13...
Sampling pulse generator, 14, 19...A/D
Converter, 15... Original curve calculation circuit, 16... Correction position detection circuit, 17... Correction data creation circuit, 21... Digital filter, 22... D/A converter.

Claims (1)

[Scope of Claims] 1. An arithmetic device that measures height deviations in the vertical direction of a track based on three-point measurement data and two-point measurement data from a track height deviation detection mechanism provided on a track inspection vehicle, A sampling pulse generation circuit generates sampling pulses based on pulses generated every fixed distance traveled by the track inspection vehicle, and supplies the pulses to other components as necessary. an A/D conversion circuit that performs D conversion; an original curve calculation circuit that delays its own output by a certain travel distance of the track inspection vehicle via a delay means and adds it to the output data from the A/D conversion circuit; a digital filter circuit that removes components of a certain wavelength or more from calculation output data from the circuit; a D/A conversion circuit that converts the output of the circuit from D/A to analog;
A means for recording a chart using the output from the circuit as an input, and detecting that the two-point measurement data and three-point measurement data are at a certain limit value over a certain distance of the track inspection vehicle, and A trajectory height deviation calculation device comprising a data correction circuit that rewrites and corrects calculation data of the original curve calculation circuit using correction data created based on the correction data. 2. The data correction circuit includes an analog sample hold circuit that detects that two-point measurement data and three-point measurement data are below a certain limit value over a certain distance, a correction position detection circuit that includes an analog comparator, and a 10-pulse counter; a correction data creation circuit comprising a sine wave function generator, an A/D converter, and a data register to create correction data;
It is composed of a switch that switches between the calculation data of the original curve calculation circuit and the correction data, and the calculation data of the original curve calculation circuit is changed to the correction data via the switch according to the detection signal from the correction position detection circuit. An orbital height deviation calculation device according to claim 1, which is rewritten as follows. 3. The switching device is composed of an AND circuit, a flip-flop, and a 10-pulse counter, and outputs the logical product of the detection signal from the correction position detection circuit and the fixed distance marker signal or manual operation signal to control the track inspection vehicle's constant running at any time or at any time. 3. The trajectory height deviation calculation device according to claim 2, wherein the switching is performed for a fixed distance for each distance.