WO2004009420A1

WO2004009420A1 - Method for onboard diagnosis of wheelset guiding elements

Info

Publication number: WO2004009420A1
Application number: PCT/EP2003/002937
Authority: WO
Inventors: Mario SÄGLITZ; Yong Guo; Rudibert King
Original assignee: Db Reise & Touristik Ag
Priority date: 2002-07-23
Filing date: 2003-03-20
Publication date: 2004-01-29
Also published as: DE10233527A1; DE50301945D1; EP1523436B1; ATE312742T1; ES2252663T3; DE10233527B4; AU2003216862A1; EP1523436A1

Abstract

The invention relates to a method for monitoring and diagnosing the condition of wheelset guiding devices on railway vehicles and a device for carrying out said method. It is generally very difficult to determine whether wheelset guide bushes made of elastomeric materials are sufficiently sturdy for operation or what the (remaining) service life thereof is because there are hardly any reliable calculation methods to do so. Furthermore, monitoring by controlling the dimensions thereof is almost impossible in the mounted state thereof because said wheelset guide bushes are difficult to access due to the assembly layout thereof within the bogie. According to the inventive method, the condition of the wheelset guide bush particularly the play thereof is monitored by evaluating an acceleration signal that is measured in real time in the area of at least one wheelset bearing, whereupon a displacement signal is estimated from said acceleration signal by means of double integration after separating the mean value. The range of said displacement signal is then subjected to a monitoring process.

Description

_VE RF _A FEES _AND VORRICHTU _N G FOR DIAGNOSIS OF ONBOARD RADSATZFUHRUNGSELEMENTEN

The invention relates to a method for condition diagnosis and monitoring of wheel set guiding devices on rail vehicles and a device for carrying out the method.

5 It is generally known to monitor the driving behavior of rail vehicles during their intended use by recording and evaluating operating parameters representing the driving behavior. It has proven to be particularly advantageous here to detect vibrations acting by means of accelerometers and to feed the signals generated by the accelerometers to a diagnostic device integrated on the rail vehicle.

DE 196 11 775 proposes the comparison of location-specific acceleration spectra with a "vibration profile" of a route to be traveled by the rail vehicle, which is defined as normal. However, individual components of the vehicle are not monitored here, but only the summary spectrum of all components is examined together DE 195 02 670 is also limited to the evaluation of an acceleration signal relating to the entire undercarriage with the aim of recognizing a possible derailment of the undercarriage DE 101 45 433 discloses a method in which the acceleration behavior 0 of at least two, one car body A rocker damper console assigned to a rail vehicle is monitored by forming a correlation coefficient from the two acceleration values, comparing it with a setpoint value and issuing an alarm message when it is exceeded. The method uses the dynamics of the chassis in relation to the vehicle body and correlates the resulting accelerations on the right and left wagon damper bracket on the car body side in a suitable manner. However, due to the different dynamics of the chassis and wheelset guide bushing, this method (including the location of the sensors) is not suitable for monitoring the wheel set guide bushings. 0 DE 199 53 677 teaches the use of a method for detecting a derailment of a rail vehicle, wherein an acceleration of a component of the rail vehicle that is directly or indirectly in contact with the lane is determined vertically to a direction of travel, the acceleration measured by means of an acceleration sensor, twice over time integrated and finally compared with an upper 5 and / or lower limit. It is a method in which a measured acceleration signal and one in the context of the Calculation to be performed eliminating a linear tendency a path difference is estimated. However, the diagnostic accuracy of the path signal that can be achieved in this way is insufficient in terms of its diagnostic quality for coupling to a vehicle control system. The guidance of the wheel sets in relation to the bogie frame of the rail vehicle represents a highly stressed system component that could not yet be recorded in the required quality with these previously known diagnostic systems. In the current series for high-speed trains of the ICE family, the precise wheelset guidance is carried out either via two guide pins pressed into the bogie frame and wheel set guide bushes mounted thereon or via wheel set guide bushes in the bearing of the linkage of the swing arm to the bogie frame. Regardless of the respective design solution, elastomer materials are used for the wheelset guide bushings, the material composition and properties of which are adapted to the intended use. In addition to achieving smooth vehicle running, the focus is also on enabling a limited radial adjustment of the wheel sets during cornering. Thus, the wheelset guide bushes are subject to a variety of wear-causing influences in daily use. In the vertical direction, they are exposed to the weight forces of the vehicle and dynamic forces from irregularities in the road (e.g. corrugations on the rail surface), in the horizontal longitudinal direction braking and acceleration forces act, in the transverse direction centrifugal forces act when cornering. A slight twist within the bushing can also be superimposed on these loads. In general, it is very difficult to prove sufficient operational strength or to specify a (remaining) service life for elastomer parts, since there are hardly any reliable calculation methods for this. Likewise, monitoring the wheelset guide bushes in the installed state by checking their dimensions is almost impossible, since they are hardly accessible due to the installation situation in the bogie.

The invention is therefore based on the object of developing a method for condition diagnosis and monitoring of wheel set guidance devices on rail vehicles and an apparatus for carrying out the method. A robust, railroad-compatible design is intended to ensure high diagnostic quality and reliable alarm rates. This object is achieved in connection with the preamble of patent claim 1 in that the state of the wheel set guide bushing of the rail vehicle - in particular the play - is monitored by evaluating an acceleration signal currently measured in the area of at least one wheel set bearing by using this acceleration signal by means of two integrations an estimation of a path signal is carried out after the mean value has been released. This estimated path signal w (t) represents a state variable that uniquely describes at least one coordinate of the current spatial position of the wheel set in relation to the other components of the bogie surrounding it. With an increased play of the wheel set guide bushing or also with impermissible changes in stiffness, a significant change in position with respect to this spatial coordinate can be expected in the operating state compared to the good state. However, deriving such a path signal estimate from the output signals of acceleration sensors by means of double integration initially seems impossible since neither the starting position nor the starting speed of the measuring point are known. The performance according to the invention consists in demonstrating a method which nevertheless enables a suitable estimation and monitoring of the spatial position of the wheel set guide bushings.

According to a particular embodiment of the subject matter of the invention, the range of the spacing of these wheel sets w _s , _{A from} one another is determined from the path signal estimates w (t) of at least two wheel sets associated with a bogie of a rail vehicle (variant A). This method is particularly advantageous due to the high significance of the signal determined and the associated high degree of error detection. In this context it should be noted that in the determination of W _S, A the result to the design clearance w _{A stat} i _SCH is adjusted between the two wheel set bearings.

An advantageous modification of the inventive concept provides that the range of the distance between this wheel set and a fixed point on the associated bogie frame W _{S, B of} the rail vehicle is analyzed from the path signal estimate w (t) of a wheel set (variant B). In contrast to the aforementioned embodiment, it is not only possible to draw conclusions about the side of a bogie which is affected by an irregularity in the wheel set guidance, but also to enable the specifically damaged pair of wheel set guide bushings of a wheel set to be identified. In this case too, when determining w _B Result to clean the constructive distance w _stat j _Scrι between the wheelset bearing and the fixed point on the bogie frame.

Further preferred refinements of the method according to the invention result from the other features mentioned in subclaims 4 to 10.

According to the invention, the object is further achieved by a device with the features mentioned in claim 11. As a result of the fact that means for detecting the acceleration behavior of at least one wheelset guide bushing on at least one wheelset bearing housing of the rail vehicle are available, by means of which a currently measured acceleration signal can be detected and the range w _{s of} a path signal w (t) estimated from this acceleration signal by double integration after the mean value has been released. can be subjected to monitoring, an apparatus for condition diagnosis and monitoring of wheelset guiding devices on rail vehicles can be provided in an advantageous manner.

In particular, it can be provided in a preferred embodiment that an evaluation device for monitoring the estimated path signal is an integral part of a higher-level device implemented on the rail vehicle for monitoring the driving properties of the rail vehicle. This takes a further step in the direction of a powerful, vehicle-integrated on-board diagnostic system.

The idea of the invention is explained in more detail below in an exemplary embodiment with reference to the associated FIG. 1.

Figure 1 shows schematically (top view) a bogie frame (1) and two wheel sets (2, 3), the bogie frame extending over a primary suspension system (not shown here for the sake of clarity) and two wheel set bearings (4, 5; 6, 7) ) is supported on the two wheel sets. There are two wheelset guide bushings in the immediate vicinity of a wheelset bearing.

In the first variant of the inventive idea, the range W _{S, A of} the distance w _A between the two wheel sets is monitored. For the estimation of w _A , two acceleration signals are required per bogie side, which originate from sensors which are mounted in the area of the wheelset guide bushings (4, 5; 6, 7) on at least two wheelset bearing housings and at least one parallel to it have investigating spatial coordinate extending detection direction. The following therefore applies: w _Λ (t) = w ₄ (t) - w ₆ (t) or w _A (t) = w ₅ (t) - w ₇ (t)

For the sake of simplicity, it is assumed below that the spatial coordinate to be examined corresponds to the direction of travel of the rail vehicle. Since there is a fixed assignment between the wheelset guide bushings and the acceleration sensor, the same reference numerals are used in FIG. 1 for the sake of clarity. In the modified solution variant according to claim 3, the span of the distance w _B between a wheel set (2) and a fixed point (8) on the associated bogie frame (1) is monitored.

The rail vehicle (not shown) comprises a data acquisition and evaluation device which is connected to the acceleration sensors. The evaluation unit of the rail vehicle controls warning displays for the driver and / or error messages to the depot and / or actuators to act on the vehicle control. The structure and mode of operation of such data acquisition and evaluation devices, like those of acceleration sensors, are generally known and are not further explained in this context.

The acceleration of the wheel sets (2; 3) in the direction of travel are recorded by means of the acceleration sensors. From the signals determined in this way, the spatial position of the wheel sets in the wheel set guide bushings (4, 5; 6, 7) should now be inferred. While the rail vehicle is traveling in the direction of travel, the wheel sets are each at different distances w _{A from} one another.

Since the spatial coordinate of a wheel set to be examined generally has upper and lower limit values, it can be assumed that the associated displacement measurement variables oscillate around a rest position within an interval. With additional consideration of the linear character of an integrated constant and the quadratic curve shape of an integrated straight line, it follows that when a series of measurement data with a non-zero mean value are available, both the first and the second integration curve tend towards infinity. This can be avoided by removing the mean value from the signals before integration. The following applies to the estimation of the speed and distance signals from the measured acceleration signal a: (1) the dynamic path signal has a mean value close to zero over a sufficiently long period of time

(2) the data section considered in each case contains a sufficient number of periods of the path signal components broken down according to the Fourier series with the smallest frequency that cannot be neglected

Assumption (1) is generally fulfilled on the one hand by the design and often applies to machine vibrations, especially when the machines are in a stable operating state. Assumption (2) applies if the path signal of interest oscillates at a high frequency or has components of a higher frequency. Each path signal of a typical machine vibration can be represented practically as the sum of theoretically infinitely many sine and cosine functions (Fourier series):

w (= ^ ⁺ ∑ ( ^α * - ^cos ( ^2π - * - - ') + ß * -sin (2π - k -f - t))

where: f = fundamental frequency α, β = Fourier coefficient

Since the path signal w (t) to be estimated swings around a rest position and has an average ot value of zero, the term - is negligible.

Furthermore, due to the data windowing, residual errors still arise, which, however, are compensated for in the overall view across all frequencies and, statistically, strive for an average of zero.

With - sin (t) = cos (t) and - cos (t) = -sin (t) it also applies that the by (double) δt δt

Differentiation of the path signal w (t) represented as a Fourier series

Speed and acceleration signals each have an average value of approximately

Have zero. The speed signal v (t) results from the acceleration for a selected time window as follows: f = f ₀ + / Δr ^v ι - ^v ( ^f o + ^{Z '} Δ = \ a _t dt + v ₀

where v = speed, v ₀ = speed at time t ₀ , a = acceleration, t = time,

Δt = duration between two scans, i = index of the measuring points n = number of measuring points of the data section

Since the mean of v must assume the value zero, the following applies:

This means that v ₀ can be partially compensated for by the integration and the subsequent exemption from the mean value. This makes it possible to determine the speed signal with a calculable error barrier.

According to the invention, the path signal is determined in an analogous manner from the speed signal.

In a departure from this, when determining the path size w _B (variant B), it should be noted that in this case, ie considering the relative path between the axlebox bearing and the bogie frame, the acceleration signal a _{d of} the bogie frame to be used as a reference can be dispensed with. It could be shown that the difference size a _r - a _d depends essentially on a _r , since the amplitude ratio a _r / a _{d takes on} large values (eg 50 and higher). where: a _d = acceleration signal of the bogie a _r = acceleration signal of the wheelset

The accelerations on the bogie frame are significantly lower and are also dominated by low-frequency components. This specific oscillation phenomenon can now be exploited in such a way that the signal a _r is freed from the linear tendency and the offset and the resulting signal can be used as a good approximation for the difference _quantity a _r - a _ά - provided the duration is sufficiently short , The signal processing sequence is as follows: the sensors in the area of the axlebox bearing housings ([4 and 5] or [6 and 7]) supply the required acceleration signals as an input variable. All data processing steps are preferably applied to a data window length of 2 seconds. In variant A, the two acceleration signals on one side of the bogie are subtracted from each other after the mean value has been released from each bogie and the difference value is then subjected to further data processing steps: after filtering, first integration, renewed mean value exemption and second integration, the desired range w _{s, A} des is then displayed - Estimated path signal w _A determined and buffered. After calculating and temporarily storing a sufficient number of span values, the mean value is formed. This is then used in the data acquisition and evaluation device of the rail vehicle as an online monitoring variable and is continuously compared against a fixed limit value determined during the parameterization. In the course of carrying out the tests, it has proven expedient to use a moving average over the 30 most recent span values. The oldest span value is replaced by the most recent span value determined each time a data window is run. As a result, the monitoring variable is smoothed over time and the false alarm rate can be minimized by setting the limit value more significantly. The diagnostic depth that can be achieved with variant A is limited to the bogie side (left or right in the direction of travel); damage cannot be assigned to a wheel. In contrast to this, subtraction of two acceleration signals is omitted in variant B due to the processing or use of only one acceleration signal in each case. The depth of diagnosis extends to the detection of damaged guide bushes in the area of a wheel.

When evaluating the mean value, it should also be noted that the monitoring variable shows fluctuations even in the faultless GOOD condition, which is due to the spring travel of the wheel set guide bushing. In the DEFECTIVE state, the enlarged wheel set guidance game overlaps this system-related basic game. It has been shown experimentally that this discrepancy or overlay can only be demonstrated in a sufficiently significant manner above a minimum speed of the rail vehicle. For example, with a spring travel of 0.3 to 0.5 mm, a minimum speed of 120 km / h was required. An expedient development of variant B according to the invention provides that, by means of an automatic feature selection, for example by means of SVM (Support Vector Machine), a selection of additional features depending on the driving speed is made. By using additional features, the lower significance of the monitored feature w _B - caused by the use of approximate solutions for difference formation and double integration - can be compensated for. In numerous test series, however, it has been shown that the result of the automated feature selection depends to a large extent on the vehicle speed range under consideration and on the averaging duration. It has therefore proven to be advisable to work in graded speed ranges with specific additional features in each case. The error detectability rate is increased with a tendency towards longer averaging times and narrower speed ranges.

LIST OF REFERENCE NUMBERS

w (t) estimated path signal

Wj path signal at the location of the wheelset guide bush i

Ws range of the travel signal w (t) w _A distance between the wheelsets of a bogie w _B distance between the wheelset and fixed point on the bogie frame

1 bogie frame

2, 3 wheelset

4, 5, 6, 7 wheelset guide bush or acceleration sensor 8 fixed point on bogie frame

Claims

claims

1. method for condition diagnosis and monitoring of wheel set guidance devices on rail vehicles,

characterized,

that the condition of the wheel set guide bushing (4, 5, 6, 7) of the rail vehicle - in particular the play - by evaluating one in the area of at least one

The currently measured acceleration signal of the wheelset bearing is monitored by making an estimate of a path signal w (t) from this acceleration signal by means of two integrations after the mean value has been freed in each case, which as a state variable is at least one coordinate of the current spatial position of the wheelset (2, 3) with respect to the surrounding other components of the

Rail vehicle clearly describes.

2. A method for the diagnosis and monitoring of wheel set guiding devices according to claim 1, characterized in that from the travel signal estimates w (t) at least two wheel sets (2, 3) belonging to a bogie of the rail vehicle determine the span w _s , _{A of} the distance between these wheel sets becomes.

3. A method for diagnosis and monitoring of wheel set guiding devices according to claim 1, characterized in that from the travel signal estimate w (t) of a wheel set (2) the span WS, _{B of} the distance between this wheel set and a fixed point (8) on the associated bogie frame ( 1) of the rail vehicle is determined.

4. A method for diagnosing and monitoring wheel set guiding devices according to claim 3, characterized in that linear tendencies and the offset from starting and braking processes of the rail vehicle are eliminated from the acceleration signal.

5. A method for diagnosing and monitoring wheel set guiding devices according to at least one of claims 3 or 4, characterized in that by means of an automatic feature selection, a travel speed-dependent selection of additional features is carried out.

6. Method for diagnosis and monitoring of wheel set guiding devices according to at least one of claims 1 to 5, characterized in that the evaluation of the double integrated acceleration signal is carried out within a preferred integration interval Δ t = 2 seconds.

7. The method for diagnosing and monitoring wheel set guiding devices according to at least one of claims 1 to 6, characterized in that the integration intervals Δ t ,, Δ t _{i + 1} begin sequentially contiguously.

8. A method for the diagnosis and monitoring of wheel set guiding devices according to at least one of claims 1 to 7, characterized in that spans w _s, j (t) of the path signals Wj determined by several integrations of the acceleration signal over the sequentially successive integration intervals Δ tj (t) mean values are formed which are compared continuously or at predefinable time intervals with a limit value determined in the course of the parameterization and which trigger alarm procedures as a function of the exceeding of these predefined limit values.

9. Method for diagnosis and monitoring of wheel set guiding devices according to at least one of claims 1 to 8, characterized in that the diagnostic method is only activated from a minimum speed of the railway vehicle.

10. Method for diagnosing and monitoring wheel set guiding devices according to at least one of claims 1 to 9, characterized in that the method is used as an integral part of a higher-level method implemented on the rail vehicle for diagnosis and monitoring of the driving properties of the rail vehicle.

11.Device for status diagnosis and monitoring of wheelset guidance devices on rail vehicles, characterized by means for detecting the acceleration behavior of at least one wheelset guidance bush (4, 5, 6, 7), by means of which an acceleration signal currently measured in the area of at least one wheelset bearing of the rail vehicle can be detected and the span a path variable obtained from this acceleration signal by double integration after the respective release of the mean value can be monitored for exceeding the limit value.

12. Device for condition diagnosis and monitoring of wheelset guidance devices according to claim 11, characterized in that an evaluation device for monitoring the path signal is an integral part of a higher-level device implemented on the rail vehicle for monitoring the driving properties of the rail vehicle.