METHOD FOR CONTROLLING AND RECOVERING THE ADHESION OF THE WHEELS OF A CONTROLLED AXLE OF A RAILWAY VEHICLE

Abstract

The method comprises the steps of generating speed signals indicative of the angular speed of the wheels of said axle; generating an error signal indicative of the error or difference between a set point speed for the wheels, determined by means of a reference model, and the speed indicated by said speed signals; and generating a driving signal for torque-controlling apparatuses applied to the wheels of said axle, by adaptive filtering of an input signal which is a function of said set point speed, modifying parameters of the adaptive filtering as a function of said error signal, such as to make such speed error or difference tend to zero.

Claims

1. A method for controlling and possibly recovering the adhesion of the wheels of a controlled axle of a railway vehicle, comprising the steps of generating speed signals indicative of the angular speed of the wheels of said axle; generating an error signal indicative of the error or difference between a reference speed for the wheels, determined by means of a reference model in response to an assigned set point speed and the speed indicated by said speed signals; and generating a driving signal for torque controlling means applied to the wheels of said axle, by adaptive filtering of an input signal which is a function of said set point speed modifying parameters of the adaptive filtering as a function of said error signal, such as to make said speed error or difference tend to zero.

2. A method according to claim 1, wherein said input signal is a signal representative of said set point speed.

3. A method according to claim 1, wherein said input signal is a signal representative of said speed error or difference.

4. A method according to claim 1, wherein said driving signal is generated by means of an adaptive filter having a FIR-type structure, with a parallel integrator.

5. A method according to claim 3, wherein the signal provided by said integrator is subjected to adaptive calibration.

6. A method according to claim 1, wherein the driving signal is obtained by means of an adaptive filter having an IIR-type structure and preferably a PID-type configuration.

7. A method according to claim 1, wherein parameters or coefficients of the adaptive filtering are initialized with pre-calculated values stored in a non-volatile memory.

8. A method according to claim 1, wherein the parameters or coefficients of the adaptive filtering are limited in a pre-defined band of variation and are stored in a non-volatile memory.

9. A method according to claim 1, wherein by means of a leakage function of the adaptive filtering the parameters or coefficients of said filtering are reduced in a continuous manner in time.

Description

[0026] Further features and advantages of the invention will become apparent from the detailed description that follows, implemented with reference to the accompanying drawings, wherein:

[0027] FIG. 1 is a block diagram of an anti-slip control system of the wheels of a railway vehicle;

[0028] FIG. 2 is a block diagram of a closed loop control system of the axle's speed;

[0029] FIG. 3 is a diagram of a possible embodiment of an apparatus for controlling the torque applied to an axle;

[0030] FIGS. 4, 5 and 6 are block diagrams of different ways of building systems for the implementation of a control method according to the present invention;

[0031] FIG. 7 is a graph showing qualitatively the trend of the adhesion coefficient of a wheel, shown on the y-axis, as a function of the sliding , shown on the x-axis; and

[0032] FIG. 8 is a diagram illustrating forces applied to an axle's wheel.

[0033] The method according to the present invention applies adaptive-type techniques to the tuning and the dynamic correction of the control parameters of the slipping of the wheels of an axle, such techniques being performed continuously over time, in real time and not based on vectors or parametric tables mapped previously.

[0034] The invention, for this object, uses a control technique based on adaptive filtering, as described for example in B. Widrow and S. D. Stearns, Adaptive Signal Processing, New Jersey, Prentice-Hall, Inc., 1985.

[0035] Various known types of adaptive filters are known, suitable for use in a method according to the invention. By way of non-limiting example, the present invention provides the use of adaptive filters known as LMS (Least Mean Square) filters. For an accurate description of the general properties, features, convergence criteria and the variants of implementation of LMS filters, please refer to the available literature or the previously cited reference text.

[0036] The adaptive filters used may consist of both FIR-type (Finite Impulse Response) structures, and IIR-type (Infinite Impulse Response) structures.

[0037] According to the current symbolism for describing an adaptive filter, X(t) and Y(t) shall designate the input and output of such a filter.

[0038] In the description that follows and in the accompanying drawings, the time variable t will be denoted by the letter T to indicate that such time is understood in a discrete sense, namely that the method/system operates for finite samples with a period T.

[0039] FIG. 4 shows a first control system schematically illustrated for implementing the method according to the present invention.

[0040] The system according to FIG. 4 comprises a block RM that represents a reference model that receives as input the set point speed V.sub.SETPOINT for the wheels W of the controlled axle A and which provides at its output the speed V.sub.R(t) of the axle A in response to the set point speed and its variations.

[0041] The block RM has a transfer function G, which ideally is G=1. However, a more meaningful system, i.e. one adhering more to reality, for example (but not exclusively) a second order transfer function, approximating the expected model of the complex formed by the control module CM of the torque control apparatus TC and by the wheels W of FIG. 2.

[0042] In FIG. 4, AF indicates an adaptive filter, for example an LMS-type filter.

[0043] At the input X(T) of such filter AF a signal is supplied which is a function of the speed V.sub.SETPOINT, for example, a signal proportional to the vehicle speed, typically between 65% and 95% of the vehicle's speed.

[0044] The output Y(T) of the adaptive filter AF is a signal for driving the torque control apparatus TC, which is in turn coupled to the axle A and its wheels W.

[0045] The output V.sub.R(T) of the block RM is applied to an input of an adder ADD, to another input from which arrives a signal V.sub.M(T) indicative of the angular speed of the axle A actually measured by means of an SS detector and an associated acquisition and processing module APM.

[0046] The adder ADD provides as its output an error signal E(T) indicative of the error or difference between the speed V.sub.R(T) and the measured speed V.sub.M(T), i.e. the difference between the expected speed at the output of the RM block and the speed V.sub.M(T).

[0047] The error signal E(T) is fed to the adaptive filter AF, where it is used to implement a continuous correction of the parameters of this filter, as long as this error E(T) tends to zero.

[0048] The stabilization of the coefficients or parameters of the adaptive filter AF can happen quickly if the input signal X(T) has a harmonic content equivalent to the bandwidth of the process to be controlled.

[0049] In the case of the system according to FIG. 4, the speed V.sub.SETPOINT is a signal with a practically null harmonic content and thus relatively unsatisfactory as regards the need for a rapid retuning of the parameters or coefficients of the filter AF. Such a solution may still be effectively used by initializing the coefficients of the filter AF to standard values, pre-calculated on the basis of an initial situation of compromise and subsequently, in the course of a slip phenomenon, the variations of the error E(T) at the input of the filter AF are sufficient to implement in real-time a proper correction of the coefficients or parameters of the filter.

[0050] An alternative solution is illustrated in FIG. 5: the error E(T) is not only used as the correction factor of the adaptive algorithm implemented by the filter AF, but as input X(T) to the filter AF itself.

[0051] In fact, the error E(T) has an appropriate harmonic content for self-calibration of the filter AF and at the same time contains the information necessary for the generation of the corrections of the braking force acting on the controlled axle A.

[0052] The solution according to FIG. 5 allows a very fast dynamic calibration of the coefficients or parameters of the adaptive filter AF, even if initially these coefficients or parameters were completely zeroed.

[0053] As is known from the literature, the LMS-type adaptive filters can be realized both by using FIR structures and IIR structures.

[0054] The FIR structures are inherently stable, having no memory. However, this feature prevents the implementation of control functions having an integrative component, unless one uses the existing natural integrators downstream in the system, for example the natural integration represented by the braking cylinder.

[0055] FIG. 6 illustrates an embodiment which allows one to overcome the lack of an integrative component in a LMS-type adaptive filter AF produced by an FIR structure. An integrator I is provided substantially in parallel to this adaptive filter AF, between the input of this filter and an adder ADD1, which receives the output of the filter AF and the output of the integrator I; the output of the adder ADD1 is connected to the torque control apparatus TC.

[0056] A self-tuning apparatus STA may optionally be associated with the integrator I and it includes a dedicated LMS-type cell C, connected between the output of the integrator I and the adder ADD1 and driven as a function of the error signal E(T).

[0057] In general, in the implementation of a control method according to the present invention, in order to avoid problems of deviations in the adaptive filter coefficients during the execution of the method, it is possible to limit the variation of the adaptive filter coefficients to a range of safety values stored in nonvolatile memory.

[0058] In order to maintain control always responsive to new variations of external parameters of the system, the leakage function characteristic of adaptive filters is appropriately used, to perform a continuous de-tuning of the coefficients or parameters of the filter when the error E(T) is close to zero, or in any case within the limits of variation of the coefficients or parameters such as to permit the recovery of a correct tuning of these coefficients or parameters as soon as they re-present significant E(T) values.

[0059] Also included in the scope of the present invention are implementations wherein one uses an adaptive filter made with an IIR-type structure, which may take a PID-type (Proportional-Integrative-Derivative) configuration.

[0060] Naturally, without altering the principle of the invention, the embodiments and the details of construction may vary widely with respect to those described and illustrated purely by way of non-limiting example, without thereby departing from the scope of the invention as defined in the appended claims.