JPH05656B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Industrial Application] This invention relates to a wheel tread abnormality detection device that automatically detects abnormalities such as flats and peeling that occur on wheel treads, which are a source of noise and vibration in railways, from the ground. It is used to detect

[Conventional technology]

Until now, inspection and detection of abnormalities in wheel treads has been carried out on most railway lines, except for a few, using human tactics, such as reports from train crew members, inspection staff accompanying them, and interviews near operating lines. There were many things.

Wheel tread abnormality detection devices that eliminate this human-force tactic include (1) devices that use strain gauges, as described in Japanese Patent Publication No. 58-2373, and (2) devices that use vibration accelerometers. etc. have been proposed.

The former device uses strain gauges attached to several locations on the belly of the rail to record changes in wheel load and bending moment when a train passes.The principle is that if there is an abnormality in the wheel tread, the impact load is superimposed. This is what was used.

In addition, the latter device attaches a vibration accelerometer to the rail and attempts to detect abnormalities in the wheel tread based on the vibration level output from the vibration accelerometer.

[Problem to be solved by the invention]

Conventional devices using strain gauges can detect trains with almost constant wheel loads and passing speeds, such as the Shinkansen, with good relative accuracy; When used by being attached to a rail where different vehicles pass, the accuracy of detecting abnormalities in wheel treads becomes poor. Furthermore, since the equipment is large-scale, there are major problems in terms of cost and operation for a simple equipment, making it unsuitable for local commercial lines.

In addition, with conventional equipment using vibration accelerometers,
Although it is possible to manufacture a simple one, it does not take into account that the impact caused by the flat of the wheel tread is a high frequency wave. For this reason, if a vibration accelerometer is attached to a sleeper that absorbs a lot of high-frequency components, if the data is passed through a low-pass filter of 500Hz or less, or if the data sampling period is low, the detected waveform will become dull and the vibration peak cannot be detected accurately. impossible to catch. Furthermore, unless the abnormal spot on the tread happens to be directly above the vibration accelerometer, the vibration will be attenuated and the detection accuracy will further deteriorate.

Moreover, these devices can detect abnormalities on a vehicle bogie basis, but cannot identify abnormalities on a wheel-by-wheel basis.

This invention was made to solve the above-mentioned conventional inconveniences, and it is possible to capture vibration acceleration from many types of trains with different weights and types and vehicles with different speeds with high accuracy, and The object of the present invention is to provide a wheel tread abnormality detection device that can identify abnormalities in wheel treads, can be easily reduced in size and weight, and can be manufactured at low cost.

[Means to solve the problem]

In order to achieve the above object, in the device of the present invention, two or more vibration accelerometers are fixed at a set interval to the bottom upper surface of a pair of rails,
A train approach detector is fixed to one of the rails on the train approach side of the vibration accelerometers, and a wheel detector is disposed between each vibration accelerometer. And an amplifier that amplifies the output voltage from each vibration accelerometer is connected to a computer via an interface section, and an A/D conversion circuit that A/D converts the output from the amplifier input via the interface section. teeth,
Equipped with a peak hold circuit that responds to a frequency of up to 20KHz, the computer includes a speed calculation circuit that calculates speed from the outputs of the train entry detector and wheel detector, and an output from the A/D conversion circuit according to the train speed. A flat determination circuit that determines whether or not a wheel is flat; a processing range determination circuit that determines a data processing range for the wheel based on the train speed and the output of a wheel detector; a wheel counter circuit that calculates the number of wheels passing; The computer is connected to a recorder that records the processed data and the passing time of the train, and a printer that prints the data.
A wheel tread abnormality detection device is configured.

[Effect]

Since two or more vibration accelerometers are fixed at set intervals on each of a pair of rails,
The large impact that a flat wheel tread exerts on the rail is detected by a vibration accelerometer and input to an amplifier. The output from the amplifier is input through the interface section to an A/D conversion circuit equipped with a peak hold circuit that responds to frequencies up to 20KHz.
Even if the extremely sharp flat peak waveforms captured by a vibration accelerometer are flat, even if the sampling period is set to around 1KHz and the number of data is reduced, the sharp peak waveforms can be accurately captured and analyzed. .

The computer also includes a speed calculation circuit that calculates the speed based on the outputs of the train entry detector and the wheel detector, and a flat determination circuit that determines whether the output from the A/D conversion circuit is flat or not depending on the train speed. , a processing range determining circuit that determines the data processing range for the wheel based on the train speed and the output of the wheel detector, a wheel counter circuit that calculates the number of wheels passing, and a timer that records the train passing time. Therefore, flats on the wheel treads of many types of trains and vehicles with different speeds can be automatically detected on a wheel-by-wheel basis.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings shown in FIGS. 1 to 3.

The wheel tread abnormality detection device includes a vibration detection section A that detects vibrations generated by the flat, a data processing section B that performs arithmetic processing on the detected data,
It mainly consists of a data recording section C that records processed data (see FIG. 1).

The vibration detection unit A includes a plurality of vibration accelerometers 2 attached to each rail 1, one train entry detector 3 and one wheel detector 4 attached to one rail 1, and an amplifier 5.

In FIG. 2, a plurality of vibration accelerometers 2, 2, for example, two each, are horizontally fixed to a pair of up or down rails 1, 1 at set intervals, respectively. To attach the vibration accelerometer 2 to the rail 1, since the bottom top surface of each rail 1 is inclined with respect to the horizontal direction, the top surface is horizontal and the bottom surface is made of hard Bakelite with the same slope as the bottom top surface of the rail 1. Insert a wedge-shaped member made of
This is done by fixing these together with an adhesive.

A piezoelectric sensor is used as the vibration accelerometer 2, and the two vibration accelerometers 2, 2 are installed at an interval of about 1.3 m, which corresponds to about 1/2 of the circumferential length of the wheel 6. This is either vibration accelerometer 2,
The aim is to be able to capture vibrations above a certain level, and at worst to capture vibrations due to flat parts within 60 to 70 cm from the vibration accelerometer 2.

A wheel detector 4 is attached to the center of two vibration accelerometers 2, 2 fixed to one rail 1, and a wheel detector 4 is attached to the bottom of the rail 1 about 2 m away from the wheel detector 4 toward the train approach side. An entry detector 3 is attached. Therefore, train approach detector 3
is located closer to the train approach side than any of the vibration accelerometers 2, and the distance between the vibration accelerometer 2 on the train approach side and the train approach detector 3 is equal to the distance between the vibration accelerometers 2, 2, and Approximately 1/2 of the circumference length.

For example, one set of magnetic sensors is used as each of the train entry detector 3 and the wheel detector 4, and these magnetic sensors are fixed to the bottom of the rail 1 so as to detect the wheels passing on the rail 1. 6 to detect the axle 7 of the wheel 6. When the train entrance detector 3 detects the entrance of a train, it functions to activate the data processing section B, and the wheel detector 4 notifies the passing of a wheel 6 and detects the wheel 6 where a flat has occurred.
is used to identify the axle 7 to which it is attached.

The vibration acceleration captured by each vibration accelerometer 2 is amplified by an amplifier 5 and input to a data processing section B. On the other hand, the outputs from the train entry detector 3 and wheel detector 4 are directly input to the data processing section B.

The data processing section B includes an interface section 8, an A/D conversion circuit 9 with a peak hold circuit, and a personal computer (hereinafter referred to as a computer).
It mainly consists of 10.

The interface section 8 is a place where the vibration detection section A and the data processing section B, which are two different systems, are connected to each other and interacted with each other. The vibration acceleration from the train is amplified by the amplifier 5 and input to the interface section 8.
The detected signal is directly sent to the interface section 8.
input and waveform shaping.

The A/D conversion circuit with peak hold circuit 9 uses an A/D conversion circuit that responds to a frequency of up to 20KHz and holds peaks, and the vibration accelerometer 2
The analog signal from the computer is converted into a digital signal and input to the computer 10. It has been experimentally confirmed that by using an A/D conversion circuit having such characteristics (performance), it is possible to accurately convert a sharp vibration waveform caused by a flat. Therefore, even when sampling is performed at around 1KHz and the number of data is reduced, the A/D converter circuit with peak hold circuit 9 can respond up to a frequency of 20KHz.
This enables accurate determination of flatness.

The computer 10 has a speed calculation circuit 11 that calculates the train speed based on the outputs of the train entry detector 3 and the wheel detector 4, which are input via the interface section 8, and the outputs from the A/D conversion circuit 9. Based on the fact that the vibration acceleration increases as the train speed increases even on the same flat, for example, the threshold level S _L for flat determination is determined by taking the ratio of the passing train speed V to the reference train speed V ₀ , and S _L = a flat determination circuit 12 that determines a flat by giving weight to train speed using a formula such as a (V/V ₀ ) ² + b (V/V ₀ ) + c (where a, b, and c are constants); ,
A processing range determination circuit 13 that determines a data processing range for the detected wheel 6 based on the train speed calculated by the speed calculation circuit 11 and the output of the wheel detector 4, and a wheel that calculates the number of passing wheels 6. It mainly consists of a counter circuit 14 and a timer 15 that records the train passing time.

Therefore, based on the data processed by the computer 10, it is not only possible to uniquely identify the train in question from the train diagram, but also to ensure that the flatness of the wheel treads of vehicles running at different speeds is constant. This can be automatically detected on a wheel-by-wheel basis by replacing it with the ratio of the speed to the reference speed.

The data recording section C mainly includes a recorder 16 and a printer 17 connected to the computer 10. The recorder 16 records the data processed by the computer 10 and the passing time of the train whose vibration acceleration was detected, on a floppy disk, for example, and the printer 17 records the date, time, and passing time of the vibration detection. Prints information such as the number of train axles and speed, the location of the vehicle where the flat was detected, the axle location, left and right, and the magnitude of vibration.

Next, the flat determination process by the device will be explained with reference to FIG. 3 for wheels passing on the left rail in the train approach direction.

When a train passes through the device installed on the rails, the train entry detector 3 first detects the entry of the train, and then every time the axle 7 passes over the train entry detector 3, a detection signal is emitted. generated (third
Diagram a). On the other hand, the wheel detector 4 outputs the same detection signal as that detected by the train entrance detector 3 with a delay of the time required for the train to pass between the train entrance detector 3 and the wheel detector 4. Detected (3rd
Figure b).

On the other hand, the vibration accelerometers 2 and 2 attached to the rails 1 and 1 at intervals are always connected to the rails 1 and 1.
1. Vibrations from the train running above are detected. FIG. 3c shows a chart after the vibration detected by the vibration accelerometer 2 between the train approach detector 3 and the wheel detector 4 is amplified, and FIG. This is a chart after amplification of the vibration acceleration detected by the vibration accelerometer 2 provided on the rear side.

Figure 3 e and f are Figure 3 c and d, respectively.
This is a chart obtained by A/D conversion using the A/D conversion circuit 9. In these charts, the vibration acceleration before conversion, which was shown as a solid line, changes to a graph consisting of a set of points, and the starting point of the graph is when the train approach is first detected by the train approach detector 3. The process starts from the time when data processing section B is activated by a signal. In this conversion, the A/D conversion circuit 9
Since it has the ability to hold peaks in response to a frequency of up to 20 KHz, the sharp peaks shown in FIGS. 3c and 3d can be converted correctly even if the A/D conversion circuit 9 does not perform sampling at high speed.

The flatness of the wheels 6 attached to each axle 7 of a passing train is determined based on the passing time of each axle 7 detected by the wheel detector 4 as shown in Fig. 3e.
In this case, the vibration acceleration in front of the wheel 6 by the time required for it to pass half a revolution is adopted as the data for the half revolution of the first, second, ... n-th axes, and in Fig. 3 f,
The vibration acceleration of the first axis, second axis,
. . . is adopted as data for the remaining half revolution of the wheel 6 on the n-th axis.

The half-circle data of each wheel 6 thus adopted is sent to the flat determination circuit 1 in the computer 10.
It is compared with the threshold level (vibration acceleration) for the speed of the passing train calculated in step 2, and if there is a peak exceeding the threshold level in the vibration acceleration corresponding to one revolution of each wheel 6, the It is determined that wheel 6 has a flat. In Figures 3e and 3f, when the threshold level of the passing train is indicated by a chain double-dashed line, the left side of the second axis (L
It is determined that there is a flat in the wheel 6 on the side (side).

In determining the above-mentioned flat, the output from the A/D conversion circuit 9 is weighted according to the train speed to determine the level at which the flat is determined. Since the fact that the size increases is taken into account, the determination of flatness is performed correctly and efficiently. Further, the data processing range for each wheel 6 is determined by the processing range determining circuit 13 based on the train speed input to the processing range determining circuit 13 and the output from the wheel detector 4. Furthermore, the number of axles 7 that have passed the wheel detector 4 is counted by a wheel counter circuit 14, and the time of train passage is recorded by a timer 15.

The data processed by the data processing unit B is recorded by a recorder 16 connected to the computer 10, and printed out by a printer 17 if necessary.

Therefore, when comparing the train passing times recorded in the printed data with the train schedule for that day, even if many trains are tested,
It is convenient to be able to identify those trains later.

Additionally, if a train type sorting device is added so that train type data can be recorded at the same time, it will be easier to check against the train schedule.

In the above embodiment, the vibration accelerometers 2 are installed at intervals of 1/2 of the circumferential length of the wheel having the maximum diameter.
is 1/n (n
Three or more may be provided at an interval of ≧3). In this case, the range of valid data detected by each vibration accelerometer 2 is 1/1/2 of the circumference of the wheel 6.
3, 1/4, . . . 1/n, which is shorter than in the case of 1/2, so it is possible to more accurately determine flatness using data that is less affected by attenuation.

Further, the amplifier 5 provided separately from the vibration accelerometer 2 in the above embodiment may be omitted if the amplifier is incorporated in each vibration accelerometer 2.

〔Effect of the invention〕

According to the wheel tread abnormality detection device of this invention,
This produces the following effects.

(1) The A/D conversion circuit is equipped with a peak hold circuit that responds up to 20KHz, so even if the sampling frequency is as slow as 1KHz, sharp and flat waveforms can be faithfully detected and no peaks will be overlooked. Therefore, it is possible to reduce the number of data to be processed, and real-time processing can be advantageously performed.

(2) Since flats can be identified on a wheel-by-wheel basis, the wheel where the flat has occurred can be quickly checked against the actual vehicle.

(3) The flat determination circuit incorporates the relationship between train speed and flat vibration acceleration level, and flats can be determined based on the speed ratio to the standard speed, so flats can be detected for train speeds from low speed range to high speed range. This can be done with one device. Therefore, when this device is used on a commercial line where trains are running at various speeds, extremely excellent performance can be achieved with a small number of devices.

In addition, by performing statistical processing on data from the same vehicle repeatedly obtained using the device of the present application, it is possible to remove the influence of errors caused by differences in the contact between the rail and the wheel, and to obtain extremely accurate information about flats occurring in the wheels. data can be extracted.

(4) Since the process from wheel flat detection to maintenance planning can be performed consistently by computer, the efficiency of abnormality detection work can be improved and the number of workers can be significantly reduced.

Furthermore, since an A/D conversion circuit, a flat determination device, etc. can be incorporated into the computer, the device can be made small, lightweight, and inexpensive. For this reason,
Since the device can be easily moved to the detection location, costs can be significantly reduced compared to conventional large-scale devices that must be permanently installed at the detection location.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the overall configuration of an apparatus according to an embodiment of the present invention, and FIG. 2 is a perspective view showing an example of a state in which a vibration accelerometer, a train entry detector, and a wheel detector are attached to a rail. Figure 3 is a diagram schematically showing the flat determination process, where a is a chart of detection by a train entry detector, b is a chart of detection by a wheel detector, and c and d are attached to the rail on the same side. Charts e and f showing the respective vibration accelerations after amplification detected by the two vibration accelerometers are charts showing the vibration accelerations after A/D conversion of the charts c and d, respectively. 1...Rail, 2...Vibration accelerometer, 3...Train approach detector, 4...Wheel detector, 5...Amplifier, 6...Wheel, 7...Axle, 8...Interface section, 9...A/D with peak hold circuit Conversion circuit, 10... Computer, 11... Speed calculation circuit, 12... Flat determination circuit, 13... Processing range determination circuit, 14... Wheel counter circuit, 15... Timer, 16... Recorder, 1
7...Printer.

Claims (1)

[Scope of Claims] 1. Two or more vibration accelerometers are fixed to the upper surfaces of the bottoms of a pair of rails at predetermined intervals, and one of the vibration accelerometers is attached to one rail at a position closer to the train approach side than any of the vibration accelerometers. The detectors are fixed, and a wheel detector is arranged between each vibration accelerometer, and an amplifier that amplifies the output voltage from the train entry detector, wheel detector, and each vibration accelerometer is installed in the interface section. The A/D conversion circuit is connected to the computer via the
Equipped with a peak hold circuit that responds to a frequency of up to 20 KHz, the computer includes a speed calculation circuit that calculates the speed based on the outputs of the train entry detector and the wheel detector, and an output from the A/D conversion circuit according to the train speed. A flat determination circuit that determines whether or not a wheel is flat; a processing range determination circuit that determines a data processing range for the wheel based on the train speed and the output of a wheel detector; a wheel counter circuit that calculates the number of wheels passing; a timer that records the passing time of the train; the computer is connected to a recorder that records the processed data and the passing time of the train; and a printer that prints the data. Anomaly detection device.