Method and apparatus for determining changes in the longitudinal dynamic behavior of a railway vehicle

Abstract

A method for determining changes in the longitudinal-dynamic behavior is disclosed, in particular of an undercarriage, of a railway vehicle for identifying a current driving condition of the railway vehicle, wherein variables, which cannot be measured and which characterize the longitudinal-dynamic behavior, are reconstructed and evaluated via a system model of the railway vehicle by means of a cybernetic observer from a known or metrologically determined input signal and at least one measuring signal of the observed railway vehicle as an observed real reference system. The at least one measuring signal of the observed railway vehicle and a corresponding reconstructed measuring signal of the system model are compared and the deviation determined by comparison is recursively tracked with a regulator so that the determined deviation is minimized.

Claims

1. A method for determining changes in longitudinal-dynamic behavior, in particular of an undercarriage, of a railway vehicle for identifying a current driving condition of the railway vehicle, comprising: wherein variables, which cannot be measured and which characterize the longitudinal-dynamic behavior, are reconstructed and evaluated via a system model of the railway vehicle by means of a cybernetic observer from a known or metrologically determined input signal and at least one measuring signal of the observed railway vehicle as an observed real reference system, wherein the at least one measuring signal of the observed railway vehicle and a corresponding reconstructed measuring signal of the system model are compared and a deviation determined by comparison is recursively tracked with a regulator so that the determined deviation is minimized.

2. The method according to claim 1, wherein the input signal is additionally supplied to the system model.

3. The method according to claim 1, wherein the input signal and/or the at least one measuring signal are detected at one or more of the following components: a railcar body; an undercarriage of the railway vehicle; at least one bogie of the railway vehicle; at least one wheelset of the railway vehicle.

4. The method according to claim 3, wherein the at least one measuring signal is detected simultaneously on opposite sides of the undercarriage or the bogie or the wheelset.

5. The method according to claim 1, wherein a brake pressure of a brake actuator or a brake current for generating a braking force decelerating the railway vehicle is processed as the input signal.

6. The method according to claim 1, wherein a driving force or a motor current for generating a force which accelerates the railway vehicle is processed as the input signal.

7. The method according to claim 1, wherein a rotational speed or a change in rotational speed of at least one wheelset is processed as the at least one measuring signal.

8. The method according to claim 1, wherein a strain of a component which transmits a longitudinal force, in particular a pull/push-rod or a pivot pin or a lemniscate lever or a wheelset guide, is processed as the at least one measuring signal.

9. The method according to claim 1, wherein a spring deflection in one or more spring stages is processed as the at least one measuring signal.

10. The method according to claim 1, wherein the evaluation of the variables characterizing the longitudinal-dynamic behavior comprises a comparison of the variables of successive undercarriages or railcar bodies or wheelsets.

11. A computer program product which can be loaded directly into the internal memory of a digital control unit and comprises software code segments which, when the product runs on the control unit, carry out the method according to claim 1.

12. An apparatus for determining changes in the longitudinal-dynamic behavior, in particular of an undercarriage, of a railway vehicle for identifying a current driving condition of the railway vehicle, comprising: a control unit and at least one sensor unit for providing a respective measuring signal, wherein the control unit is configured to reconstruct and evaluate variables, which cannot be measured and which characterize the longitudinal-dynamic behavior, via a system model of the railway vehicle by means of a cybernetic observer from a known or metrologically determined input signal and the at least one measuring signal of the observed railway vehicle as an observed real reference system, wherein the control unit is further configured to compare the at least one measuring signal of the observed railway vehicle and a corresponding reconstructed measuring signal of the system model and to recursively track, with a regulator, the deviation determined by comparison so that the determined deviation is minimized.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention is explained in more detail below with respect to an embodiment. In the drawings are:

(2) FIG. 1 a schematic depiction of a block diagram of a cybernetic observer, as it is used in the method according to the invention;

(3) FIG. 2 a graphic depiction showing a comparison of an actual friction coefficient and a friction coefficient determined by the method according to the invention as a function of time; and

(4) FIG. 3 a graph showing a time-dependent, translational longitudinal speed of the railway vehicle as a function of the friction coefficient curve shown in FIG. 2.

DETAILED DESCRIPTION

(5) The below described method for determining changes in the longitudinal-dynamic behavior is used for a railway vehicle which is not shown in detail in the figures. Such a railway vehicle has one or more units which are movably connected to each other. A coupling device is provided for connecting the vehicle units. Depending on the design of the railway vehicle, each vehicle unit may have two undercarriages, each provided with at least one wheelset. Alternatively, a railway vehicle with two vehicle units may also have three undercarriages, each provided with at least one wheelset. An undercarriage usually comprises two wheelsets. The wheelsets of the undercarriages are each provided with wheels which are running on a rail.

(6) A plurality of sensor units (in short: sensors) can be provided on the railway vehicle. The railway vehicle may, for example, have one or more sensors for determining the vehicle speed and/or an acceleration or deceleration of the railway vehicle in the longitudinal direction of the vehicle. The acceleration may be a positive acceleration due to a force accelerating the railway vehicle, or a negative acceleration due to a braking force decelerating the railway vehicle. The (positive or negative) acceleration can be a total acceleration of the railway vehicle. If an acceleration sensor is attached to a respective vehicle unit, the acceleration can also be the respective (positive or negative) acceleration of the respective vehicle unit. Acceleration may, for example, be a deceleration occurring at an undercarriage or at the railcar body of a railway vehicle. The acceleration can be determined based on speed data. The deceleration can be inferred from a temporal course and/or a change in the vehicle speed. The deceleration can be determined by considering the course of the speed in time periods that are shorter than the duration of a (positive or negative) acceleration. It may thus be provided that at least one acceleration sensor is associated to each vehicle unit and/or undercarriage. Such sensors are often designated for monitoring the driving conditions, so that already existing sensors can be used to determine the (positive or negative) acceleration.

(7) For determining the vehicle speed, for example, a radar system, an optical sensor device and/or a communication device for receiving satellite data, to which a control device of the railway vehicle is or can be connected, can be provided.

(8) In addition, sensors may be provided to determine the rotational speeds and changes of rotational speed of at least one wheelset. The determination of the wheel speed can, for example, be used to determine a braking effect and is already installed in many railway vehicles. It is also conceivable that the vehicle speed is determined based on wheel speed data. From wheel speed data associated to individual wheels or wheelsets, it is possible to determine a speed associated to the respective wheel or wheel axis, for example a circumferential speed or wheel speed. In addition to wheel speed data, the radius of the wheel can also be taken into account.

(9) From a change of wheel speed of at least one wheel-set, for example, a (positive or negative) acceleration at an associated wheel-set or an associated wheel axis can be derived.

(10) Such a railway vehicle may be equipped with sensor units to detect pitching movements of individual components about the transverse axis of the vehicle. Such sensor units are preferably associated to a respective vehicle transverse axis. For example, acceleration sensors can be used to detect an acceleration about the vehicle's transverse axis.

(11) In addition, the railway vehicle may be equipped with at least one sensor unit which detects a spring deflection in respective spring stages of a vehicle unit of the railway vehicle. Such sensors can be realized optically, by means of cable pull measurement or by inductive plungers.

(12) By means of longitudinal force sensors, e.g. strain gauges, strains of longitudinal force transmitting components can be determined. Such a sensor unit can be associated to a respective push/pull rod, a respective pivot pin or a lemniscate lever or a respective wheelset guide.

(13) Furthermore, brake pressure sensors or brake current sensors and/or brake efficiency sensors such as brake force or brake torque sensors associated to the undercarriages or to the friction brake devices of an adhesion-dependent friction braking apparatus arranged on the undercarriages may be provided. In general, a brake pressure sensor or brake current sensor can be considered to be associated with a friction brake device if it is able to detect a brake pressure or brake current individually actuating the friction brake device. A braking force sensor or a braking torque sensor may be considered to be associated with a friction brake device or a set of wheels to be braked by the friction brake device if it is able to detect a braking force applied by the friction brake device or a corresponding braking torque.

(14) By means of the cybernetic observer described below, it is possible to emulate the frequently different undercarriage superstructures in a train set of a railway vehicle with a uniform model-based algorithm. An unambiguous estimation of the undercarriage dynamics of the railway vehicle can be ensured by a combination of a plurality of measuring signals.

(15) FIG. 1 shows a block diagram of the basic structure of a cybernetic observer 1, which helps to carry out the method for determining changes in the longitudinal-dynamic behavior of the railway vehicle. In a way known to a skilled person, cybernetic observer 1 comprises a system model 20 of the railway vehicle and a unit 26 for weighting the comparison result of system model 20 and an observed real reference system 10. The dynamics of observed real reference system 10, i.e. the observed railway vehicle, is influenced by an input signal u which is fed to observed real reference system 10 at a first input 11. The input signal u is a measurable signal. In the case of braking of the railway vehicle with friction brakes, the input signal u may correspond to a brake pressure of the brake system/device. If braking is carried out by means of an electric brake, the input signal may be a braking current for generating a braking force which decelerates the railway vehicle. If, on the other hand, a change in the longitudinal-dynamic behavior due to acceleration is to be detected, the input signal u can be a driving force or a motor current for generating the force which accelerates the railway vehicle.

(16) The dynamics of observed real reference system 10 is described by states x. Here, x can be a vector with a plurality of different states. Since observed real reference system 10 is provided with at least one sensor unit as described above, at least one measuring signal y is provided at an output 13. Here, y can be a vector whose number of vector entries corresponds to the number of (real) measuring signals. The acquired measuring signals can originate from sensor units of the same and/or a different type.

(17) The observed real reference system 10, i.e. the railway vehicle, can also be excited by non-measurable disturbances z. These non-measurable disturbances z are fed to reference system 10 at a second input 12. The disturbance variable z is defined as all those influences which influence the friction coefficient between the wheel and the rail and/or between the brake pad and the brake disc and/or the brake block and the wheel.

(18) This also includes those influences which influence the friction radius, i.e. the point of application of a brake pad to the brake disc. Furthermore, a total weight changing due to a changing loading condition of reference system 10, i.e. the railway vehicle, may occur as a disturbance z.

(19) The system model 20 represents a model of the dynamic behavior of reference system 10, i.e. the railway vehicle. The system model 20 can, for example, be formed by software. The system model 20, like reference system 10, is controlled by the input signal u. The input signal u is supplied to system model 20 at a first input 21. The system model 20 determines values for at least one reconstructed measuring signal , which is provided, e.g. also as a vector, at a first output 22. A respectively reconstructed vector entry of the measuring signal is assigned to a metrologically determined vector entry of the measuring signal y of reference system 10.

(20) Since the system model 20 generally cannot emulate the entire dynamics of reference system 10 and reference system 10 is also influenced by the non-measurable disturbance variables z, the dynamic behavior of system model 20 deviates a priori from the real behavior of reference system 10. For this reason, a comparison of the at least one reconstructed measuring signal (i.e. its vector entries) with the metrologically determined at least one measuring signal y (i.e. the assigned vector entries), which is provided at output 13 of reference system 10, is conducted. These two measuring signals are supplied to a comparator 25, which generates a difference. The deviation (y) is supplied to a unit 26 for weighting the comparison result. The L-weighted feedback of the deviation (y) is provided to system model 20 at a second input 24. The weighting by unit 26 is conducted such that, after a certain time period, the behavior of the reconstructed measuring signal calculated by system model 20 corresponds to the actually measured at least one measuring signal y, i.e. the deviation becomes zero after a certain time period. This process is performed automatically and recursively.

(21) At a second output 23 of system model 20, the desired dynamic values {circumflex over (x)} can then be read, which represent the longitudinal-dynamic behavior of the railway vehicle. These are, for example, non-measurable variables, such as speeds and friction coefficients between the wheel and the rail and between the brake pad and the brake disc, braking forces and braking torques and the like. Furthermore, disturbance variables 2 can be read at a third output 27 of system model 20.

(22) FIGS. 2 and 3 show the results of the procedure for a braking process on the basis of simulation results. FIG. 2 shows the friction coefficient curve (t) between the brake pad and the brake disc as a function of time t. FIG. 3 shows the change in the longitudinal speed v(t) as a function of the same time period. A period from t=30 s to t=80 s is shown. It is assumed that within the period from t=35 s to t=80 s a railway vehicle wheelset is braked with a constant braking pressure. The respective curves z.sub., x.sub.v shown with a solid line in FIGS. 2 and 3 show the time curve of the track model, while the broken lines show the calculated values , of system model 20. In this example it is assumed that, in addition to the wheelset speed , the translational longitudinal speed v of a wheelset is measured, i.e. they are provided as measuring signals y at output 13 of reference system 10.

(23) FIG. 2 shows the temporal course of the friction coefficient between the brake pad and the brake disc, as it results from fluctuating influences during braking with constant brake pressure. The depicted variation of the actual friction coefficient (solid line) is represented by means of the feedback of the above described variation (y) and the design of unit 26 selected (by numerical simulation, testing or calculation) to weight the comparison result from the system model. The longitudinal speed v as one of the states x of reference system 10 and as one of the variables characterizing the longitudinal-dynamic behavior {circumflex over (x)} of system model 20 is illustrated in FIG. 3. The feedback of the deviation (y) of the measured signal y provided at output 13 and of the reconstructed measured signal at first output 22 of the system model causes the motion calculated by the system model 20 to correlate with the actual system behavior.

(24) The effect of the reduced friction coefficient between t=40 s and t=70 s results in the longitudinal speed v decreasing less rapidly in the specified time range, which leads to a longer braking distance and can therefore represent a safety risk.

(25) In the example shown, the application of this method allows to determine the necessary braking pressure on the basis of the calculated friction coefficient between the brake pad and the brake disc, which is necessary to maintain a prescribed braking distance. This determination is made in a control unit whose design and procedure are not the subject of the present invention.

(26) Furthermore, the knowledge of the friction coefficient allows conclusions about the wear condition of the brake pad, which enables condition-oriented maintenance.

(27) The information thus obtained from leading wheelsets or undercarriages or railcar bodies may be provided in an appropriate manner as a prediction for subsequent wheelsets or undercarriages or railcar bodies. This evaluation is also carried out in a control unit and is not the subject of this considerations. By comparing the results of successive wheelsets or undercarriages or railcar bodies, it is then apparent whether these are track-related influences or vehicle-side effects. Changes that are detected at several sensor units with a time delay indicate track-side influences. In contrast, fluctuations that only occur at individual sensor units indicate vehicle-side influences.

LIST OF REFERENCE SIGNS

(28) 1 cybernetic observer 10 observed real reference system 11 first input for input signal u 12 second input for interference signal z 13 output for measuring signal y 20 System model 21 first input for input signal u 22 first output for reconstructed measuring signal 23 second output for observed variable(s) 24 second input 25 comparator 26 unit for weighting the comparative result 27 third output for reconstructed disturbance(s) {circumflex over (z)} u input signal y measuring signal x state variable reconstructed measuring signal reconstructed state variable z disturbance reconstructed disturbance