Method for estimating a coefficient of friction, brake control method and brake control device for a rail vehicle

Description

[0001] The invention relates to a method for estimating a coefficient of friction, a method for brake control for a rail vehicle, a brake control device for a rail vehicle, and a computer program product for carrying out the respective methods.

[0002] A well-known characteristic of friction brakes is that the coefficient of friction between the pad and disc is not constant as desired, but is influenced by various system parameters and conditions such as the clamping force, the brake disc temperature, and the friction speed. During braking, this has a direct impact on the instantaneous deceleration of the train and thus on the accuracy of the braking distance.

[0003] For example, typical braking behavior differs depending on the initial speed. Even with sufficient traction, the achieved deceleration values ​​and the resulting braking distances are subject to considerable tolerances. These tolerances primarily result from the fluctuating coefficient of friction of the brake pad-disc friction pairing. This coefficient of friction is mainly dependent on the changing brake disc temperature and, consequently, on the driving history within the given profile. The brake disc temperature is determined by the energy input, which is significantly influenced by the initial braking speed. Additionally, conditioning effects occur that can affect subsequent braking maneuvers. External factors, such as ambient temperature and humidity, also influence the coefficient of friction.

[0004] Furthermore, other tolerances of the braking system, such as the efficiency of the brake actuators and the imprecisely accounted-for diameters of worn tires, directly affect the resulting braking force at the wheelsets. These numerous variables result in a significant variation in braking distances.

[0005] The influence of the described system tolerances can be reduced with a train deceleration control system that reacts immediately to any deviation of the current deceleration from the desired value. This can significantly reduce the variation in braking distances. Potential braking distance exceedances due to tolerances in the braking equipment and variations in wheel diameter can be avoided. The required braking distances are thus largely independent of the friction characteristics of the brake lining.

[0006] A deceleration control system can compensate for effects averaged over the entire train. This train-wide (average) compensation can lead to different levels of traction on individual wheelsets due to potentially varying braking forces, posing a risk of exceeding the adhesion limits of some wheelsets, particularly when ED and EP brakes are used simultaneously on different bogies.

[0007] Therefore, a device or method for determining or estimating the instantaneous coefficient of friction per wheelset or bogie would be advantageous. Using an actual coefficient of friction makes it possible to compensate for the inaccuracies that arise when calculating the required brake pressure and the resulting clamping force using a constant reference coefficient of friction. This would minimize the respective deviations from the desired frictional values. By reducing the aforementioned averaging effects, such a method can also support deceleration control.

[0008] In currently known braking systems, a so-called µ-estimator is already used. For example, DE 10 2014 116 803 A1 discloses a braking device with a control unit in which at least one braking parameter profile is stored, based on a coefficient of friction profile that arises during deceleration from an initial speed to a target speed at constant normal force and a combination of test bench parameters. Determining the respective characteristic curves using test bench parameters is correspondingly complex.

[0009] In DE 10 2016 115 275 A1, the dependence of friction between the brake disc and the brake pad on the brake disc temperature, the relative velocity vf between the brake pad and the brake disc (and thus the speed of the brake disc in m / s), and the normal force acting between the brake pad and the brake disc is described by the equation µ a (vf , T s , F n ) = a 0 + a 1 vf + a 2 vf 2< + b 1 T s + b 2 T s 2< + c 1 F n + c 2 F n 2<, where vf is the relative velocity between the brake pad and the brake disc, T s is the temperature of a brake disc friction surface, and F n is the normal force acting between the brake pad and the brake disc, which can be calculated from a brake cylinder pressure. The nonlinear components of the equation increase the effort required to determine the experimentally determined constants that are characteristic of the respective brake pad-brake disc pairings.The increased effort is particularly disadvantageous when the coefficients need to be determined on-site, for example, directly on the train during commissioning. In such or similar situations, a simple and fast method is required. Determining the coefficients of nonlinear equations is hardly suitable in this respect and is also prone to errors.

[0010] Similarly, DE 10 2018 129 132 B3 discloses the calculation of the coefficient of friction COF using the empirical formula COF i = a + b * Fb (i-1) + c * Fb 2 < (i-1) + d * v (i-1) + e * v 2 < (i-1) + f * T (i-1) + j * T 2 < (i-1) , where a, b, c, d, e, f, j are suitably chosen parameters. Here, too, the disadvantage arises of the complex determination of the parameters in each case.

[0011] WO 2020 / 021665 A1 concerns a brake control system comprising a required braking force calculator for calculating a required braking force, which is a braking force generated by a mechanical braking device to achieve a deceleration specified by a braking command; an output velocity obtainer for obtaining an output velocity; a target pressure force calculator for calculating a target pressure force, which is a force required to press a brake shoe against a wheel to achieve the required braking force; and a target pressure calculator for calculating a target pressure, which specifies a pressure of a fluid within the brake cylinder necessary to achieve the target pressure force and for performing a feedback control to adjust the target pressure based on a feedback signal obtained from a pressure sensor.

[0012] In view of the foregoing, it is therefore an object of the present invention to provide a method for estimating a coefficient of friction, a method for brake control for a rail vehicle, a brake control device for a rail vehicle and a corresponding computer program product that enables a simplified estimation of the coefficient of friction or utilizes this simplified estimation.

[0013] The problem is solved by a method for estimating the coefficient of friction, a method for brake control for a rail vehicle, a corresponding brake control system, and a corresponding computer program according to the independent claims. Advantageous embodiments of the invention are contained in the dependent claims.

[0014] According to the invention, a method for estimating a coefficient of friction (µ) for a brake pad-brake disc pairing is provided, wherein the coefficient of friction is estimated as a function of a reference coefficient of friction via at least one linear function of the brake disc temperature and / or the friction speed and / or the clamping force.

[0015] The estimation is based on the assumption that the dependencies of the coefficient of friction on the system values ​​brake disc temperature, friction speed, and clamping force are independent of each other. The coefficient of friction can therefore be described independently in the form of at least one linear function. The reference coefficient of friction serves as a benchmark. Furthermore, the reference coefficient of friction is specifically a coefficient of friction that is established at comparatively low temperatures and moderate speeds and is therefore nearly constant, at least within a certain range. The estimation is thus performed using at least one generic property or characteristic that represents the respective functional dependence of the coefficients of friction on the selected system value, for example, in the form of a characteristic curve.

[0016] The estimation using at least one linear function includes both an estimation where the coefficient of friction is estimated in the form of a straight line, and an estimation where the coefficient of friction is estimated using several straight lines with at least partially different slopes, as will be described below, for example, in relation to the estimation using piecewise linear functions.

[0017] Estimating the coefficient of friction using at least one linear function allows us to compensate, to a certain extent, for the effects that would otherwise arise from using a constant coefficient of friction. At the same time, mapping the coefficient of friction onto linear functions and considering the system values ​​independently significantly simplifies both the determination of dependencies and the application of the estimation.

[0018] It should also be noted that the clamping force corresponds to the force applied or to be applied to the brake disc. For example, the clamping force is the force with which the brake pad is pressed, or is intended to be pressed, against the brake disc, or corresponds to this. The clamping force can also be represented by, or derived from, the brake pressure or clamping force. The frictional velocity corresponds to the relative velocity between the brake pad and the brake disc.

[0019] In In one embodiment of the process, the coefficient of friction is estimated according to the following equation: μ ≈ f 0 * f 2 v where f₀ represents the reference coefficient of friction µref and f₂(v) is the linear function as a function of the frictional speed (v). The frictional speed (v) can be derived from the wheel rotational speed or from the train speed determined by other means.

[0020] In brake control systems, for example, wheel speeds are recorded and a reference speed is calculated using the wheel radii. The friction radius (pad-disc) provides a direct relationship to the friction speed.

[0021] The constant f 0 can be equal to the reference coefficient of friction or a constant proportional to the reference coefficient of friction.

[0022] The above equation assumes that the dependence on clamping force and brake disc temperature can be neglected, meaning the model consists of only a single generic characteristic. This minimizes the need for additional parameters and eliminates the need for additional temperature measurements or estimations, such as those using a temperature model. This is of practical interest because it allows for compensation of the typical increase in the coefficient of friction during braking, thus enabling more precise braking. This is particularly advantageous for trains equipped with an ATO (Automatic Train Operation) system.

[0023] In other operating conditions and / or braking scenarios, the influence of the clamping force or the brake disc temperature may predominate, so that the coefficient of friction can also be adjusted as needed. μ ≈ f 0 * f 3 F b or μ ≈ f 0 * f 1 T can be estimated, where f 1 (T) is the linear function as a function of the brake disc temperature (T) and f 3 (F b ) is the linear function as a function of the clamping force (F b ).

[0024] The selection of the linear function used to estimate the coefficient of friction can depend not only on the relevant factors but also on which influencing factors are available at all or with sufficient accuracy. In other words, the equation µ ≈ f 0 * f 2 (v) can also be used preferentially, for example, when the frictional speed is sufficiently known, but information about the clamping force or brake disc temperature is not available or not available with sufficient accuracy for estimating the coefficient of friction.

[0025] According to a further development of the method, the coefficient of friction is estimated according to the following equation: μ ≈ f 0 * f 1 T * f 2 v where f 0 represents the reference friction coefficient µ ref and f 1 (T) is the linear function as a function of the brake disc temperature (T) and f 2 (v) is the linear function as a function of the friction speed (v).

[0026] In this version, it is assumed that the dependence on the clamping force can be neglected, and thus the model consists of only two generic characteristics, which allow for a still simple, but more accurate compensation of the friction coefficient's behavior. However, this requires a temperature measurement or a corresponding temperature estimate, for example, using a temperature model.

[0027] The above equation (I) can be supplemented with the linear function f₁(T) if necessary. Accordingly, equation (I) would be equal to equation (II) with f₁(T) = 1. Alternatively, when taking the brake disc temperature into account, a different linear function f₂(v) than in equation (I) can be used for equation (II).

[0028] As described above for equation (I), the coefficient of friction can also be determined with respect to given boundary conditions, such as operating conditions, respective braking processes and / or the availability of the respective values ​​of the influencing variables via μ ≈ f 0 * f 1 T * f 3 F b or μ ≈ f 0 * f 2 v * f 3 F b to be estimated.

[0029] According to a further development, the coefficient of friction is estimated according to the following equation: μ ≈ f 0 * f 1 T * f 2 v * f 3 F b where f 0 represents the reference friction coefficient µ ref and f 1 (T) is the linear function as a function of the brake disc temperature (T), f 2 (v) is the linear function as a function of the friction speed (v) and f 3 (F b ) is the linear function as a function of the clamping force (F b ).

[0030] Equation (II) can therefore also be mapped to equation (III) with f₃(Fb) = 1, and equation (I) to equation (III) with f₁(T) = 1 and f₃(Fb) = 1. However, other functions than those used in equations (I) and / or (II) can also be substituted for f₁(T) and f₃(Fb). In general, the respective linear functions can be used individually or in any combination within a specific model to estimate the coefficient of friction. The specific selection and / or combination can be determined by the required accuracy and the expected values ​​of the system parameters. Thus, the estimation of the coefficient of friction is achieved not only by mapping it to at least one linear function with respect to the respective influencing factors, but also by factoring the coefficient of friction into these influencing-factor-dependent linear functions.

[0031] In one embodiment of the procedure, the functions f 1 (T), f 2 (v) and / or f 3 (F b ) are at least piecewise linear.

[0032] Accordingly, the respective coefficients of friction can be described by different linear functions depending on the respective system value in different ranges of definition. For example, the coefficient of friction represented by the function f₁(T) may change more rapidly in a range between 200 °C and 300 °C than in a range greater than or equal to 300 °C. Alternatively or additionally, it is possible that certain dependencies only become relevant to one of the above equations once a specific system value is reached. Thus, the constant f₀ can, for example, be assumed to be the coefficient of friction to be estimated over a certain range of the defined values, while only after reaching a specific brake disc temperature, friction speed, and / or clamping force do the respective linear functions enter equation (III), or were previously only considered as constants.In other words, at least one linear function for estimating the coefficient of friction can be represented by several linear functions, or the estimation can be carried out using several of these at least one linear function.

[0033] Based on this approach of friction coefficient factorization starting from a reference friction coefficient, the friction coefficient can be estimated with sufficient approximation in a simple manner. It is therefore not necessary to represent the friction coefficient using a universally valid empirical formula and the parameters that must be determined for this purpose; instead, the equation can be decomposed into linear sub-functions and their applicability for predetermined domains of definition depending on the respective system value.

[0034] The piecewise linear functions are chosen such that they approximate the real-world dependencies with sufficient accuracy. Since each subfunction depends on only one influencing factor, this can be done independently of the other influencing factors and therefore with less effort.

[0035] As an example, it should be noted that in a brake element with a brake pad-brake disc pairing, f1(T) typically decreases with increasing brake disc temperature T. In ranges where f1(T) changes only relatively slightly over a predetermined temperature range, i.e., in the sense of estimating a coefficient of friction with tolerable effects, f1(T) can be approximated as a constant. From this, an estimate of the coefficient of friction can be derived with increasing brake disc temperature, in which it is constant in a first sub-function, then decreases linearly with further temperature increases according to a second sub-function, and finally assumes a constant value again according to a third sub-function.The course, the domain and the number of sub-functions can be described, among other things, depending on the material pairing of the brake pad-brake disc pairing and / or the required accuracy.

[0036] Similarly, f₂(v) typically decreases with increasing frictional speed. In other words, f₂(v) increases during braking until the vehicle comes to a standstill, and thus as frictional speeds decrease. Here too, f₂(v) can be formulated as three partial functions, where only the middle partial function decreases linearly with increasing frictional speed over a predetermined domain, while the partial functions preceding and following the middle partial function are each defined by constants.

[0037] The course of the coefficient of friction as a function of the clamping force f 3 (F b ) can be influenced, among other things, by the material pairing of the brake pad-brake disc pairing and other physical factors, such as size dimensions.

[0038] In particular, the functions f 1 (T), f 2 (v) and / or f 3 (F b ) are normalized to the reference friction coefficient.

[0039] As described above and illustrated, for example, by equation (III), the coefficient of friction is estimated taking into account the reference coefficient of friction. Normalization thus adjusts the reference coefficient of friction to become the actual coefficient of friction. In other words, the respective coefficient of friction is obtained by multiplying the reference coefficient of friction, or a value representing it, by at least one of the linear functions f₁(T), f₂(v), and / or f₃(Fb). The normalized functions f₁(T), f₂(v), and / or f₃(Fb) can be described, for example, as normalized functions of the piecewise linear, absolute-valued function F₁(T), F₂(v), and / or F₃(Fb), as follows: f 1 T : = 1 μ ref ∗ F 1 T , f 2 v : = 1 μ ref ∗ F 2 v and / or f 3 F b : = 1 μ ref ∗ F 3 F b .

[0040] The term "absolute-valued function" refers to functions that are not normalized. In other words, the functions Fi represent the actual coefficient of friction as a function of T, v, and Fb, while the functions fi, according to the equations above, represent scaled or normalized values ​​of the coefficient of friction depending on the absolute-valued functions Fi.

[0041] In one embodiment, the ranges of values ​​of the functions f 1 (T), f 2 (v) and / or f 3 (F b ) in the domain of the functions are limited by limiting parameters to a lower and / or upper bound c T , cv and / or c Fb.

[0042] The linear function over an entire domain or a predetermined subdomain can be formulated such that the values ​​achievable by the respective functions, which represent the friction coefficient dependent solely on this system value, do not exceed or fall below a certain upper and / or lower limit. Alternatively, the linear function can be formulated such that it may exceed or fall below an upper limit within the corresponding domain, but the characteristic curve or the estimated friction coefficient, in the event of such an exceedance or fall below, assumes the value of the respective predetermined lower or upper limit cT, cv, and / or cFb.

[0043] When designing linear functions, both safety-related aspects, such as ensuring compliance with the maximum permissible braking distance, and system-related limitations, such as preventing overload of a brake actuator due to excessive braking force, must be considered. Therefore, a limit can be defined for the approximate value of the friction coefficient represented by each of these functions. Depending on the gradient of the friction coefficient, this limit can be applied upwards, downwards, or, if necessary, in both directions.

[0044] The method therefore distinguishes between parameters used for modeling physical behavior, i.e., model parameters, and parameters used to control a function based on these model parameters, such as a braking function, referred to here as limiting parameters. The lower and upper bounds cT, cv, and / or cFb can be used as model parameters, for example, to reduce the computational effort when the estimated coefficient of friction changes significantly. The lower and upper bounds cT, cv, and / or cFb can also serve as limiting parameters to specific braking or safety requirements. With regard to different requirements for various braking scenarios, such as emergency braking or rapid braking versus service braking, the lower and upper bounds cT, cv, and cFb can be used to define different braking or safety requirements.The upper limits cT, cv, and / or cFb can also be adjusted as limiting parameters depending on the respective braking specifications. The lower and upper limits cT, cv, and / or cFb can therefore also be referred to as model parameters cMT, cMv, and / or cMFb, and the limiting parameters with respect to a control function as cST, cSv, and / or cSFb. The distinction between model and limiting parameters is also applicable analogously to the further explanations regarding usable limits and the weighting factors described later.

[0045] According to a further development, the function f 1 (T) is chosen such that the upper limit c T,max is less than or equal to a value representing the reference friction coefficient and / or the lower limit c T,min represents a maximum clamping force to be applied.

[0046] The upper limit cT,max corresponds, for example, to the increase in the coefficient of friction observed in practice at low temperatures. Based on this, the upper limit cT,max ensures that the braking force required for clamping is always greater than a predefined minimum value, regardless of the temperature, even in the critical low-temperature range. This prevents a reduction in braking force and thus ensures compliance with the resulting maximum permissible braking distances. In other words, the upper limit cT,max can serve to maintain a minimum braking force or minimum clamping force with appropriate braking force control. For example, the function f1(T) is defined such that f1(T) is always less than or equal to the reference coefficient of friction µref, or, with appropriate normalization, less than or equal to 1.

[0047] The lower limit c T,min at high temperatures corresponds to a braking or clamping force which, if increased further, could damage the braking system responsible for applying and / or absorbing the force. The lower limit c T,min can therefore serve as overload protection.

[0048] The preceding statements regarding the upper and lower limits with respect to low and higher temperatures, respectively, refer to the actual phenomenological behavior of the coefficient of friction in these temperature ranges. In other words, the limits are defined here with regard to the behavior of the coefficient of friction in these temperature ranges: the upper limit for higher coefficients of friction at low temperatures and the lower limit for lower coefficients of friction at higher temperatures.

[0049] Alternatively or additionally, the function f 2 (v) is chosen such that the lower bound cv,min is greater than or equal to a value representing the reference coefficient of friction and / or the upper bound cv,max represents a minimum clamping force (F b ) to be applied.

[0050] The lower limit cv,min is relevant in practice, for example, at higher speeds to prevent overloading the brake actuators. To prevent the resulting braking forces from becoming too low and thus increasing the braking distance, the characteristic curve can be limited by the upper limit cv,max. Such a functional description thus compensates, especially at low speeds, for the increasing coefficient of friction, particularly during the braking process.

[0051] Alternatively or additionally, the function f 3 (F b ) is chosen such that the upper limit c Fb,max and / or the lower limit c Fb,min is / are adapted to a brake pad-brake disc pairing.

[0052] The limitation arises from analogous considerations as already mentioned for the functions f 1 (T) and f 2 (v), whereby a material pairing dependency can also be taken into account for f 1 (T) and f 2 (v).

[0053] In one embodiment of the previously described method, different weighting factors are assigned to the linear functions f 1 (T), f 2 (v) and / or f 3 (F b), in particular at least piecewise depending on the respective domain of definition, in the equation for estimating the coefficient of friction (µ).

[0054] Estimating the coefficient of friction using the methods described above generally allows for the compensation of the effects of a state-dependent coefficient of friction to a certain extent. When using a constant coefficient of friction in the control system, this is only possible for a specific scenario. The characteristic curves underlying the respective compensation, which are formed by the linear functions, can be assigned weighting factors to regulate their compensation behavior. Weighting factors can range, for example, from 0% (no compensation or a constant coefficient of friction depending on the respective system value) to 100% (full effectiveness of the characteristic curve).One reason for applying weighting factors may be safety aspects, since the current state of the art for brake systems mostly refers to constant friction coefficient considerations and can therefore limit the effect of compensation measures from a safety perspective.

[0055] The weighting factors can be adjusted depending on the respective definition domain and / or predetermined value ranges. Alternatively or additionally, the weighting can also be variable, if required, depending on certain boundary conditions such as the train's state of motion (e.g., driving, braking, standstill), the states of the braking system (e.g., depending on braking control types, such as with and without deceleration control), braking types (e.g., emergency or service braking), the state of a braking device (e.g., new versus old brake pads), and / or time-limited.

[0056] When applying weighting factors, a distinction can be made between weighting factors for estimating the coefficient of friction as a model value or model parameter and weighting factors for control purposes or as limiting parameters. For example, with regard to estimating the coefficient of friction in the sense of a model representing physical conditions, the linear functions within a factorization can be weighted. The model value of the coefficient of friction can therefore be expressed, for example, as µ M ≈ f 0 * g M1 * f 1 (T) * g M2 * f 2 (v) * g M3 * f 3 (F b ), where g M1 to g M3 represent the respective weighting factors of the linear functions according to the model. For control purposes, these weighting factors can again assume different values ​​as needed.A corresponding friction coefficient µS estimated for such control purposes can thus be expressed, for example, as µS ≈ f0 * gS1 * f1(T) * gS2 * f2(v) * gS3 * f3(Fb), where gS1 to gS3 represent the respective weighting factors of the linear functions according to the control purpose. The weighting factor gM thus represents a model parameter, analogous to the explanations regarding the lower and upper bounds, while the weighting factor gS represents a limiting parameter.

[0057] Instead of assigning individual weighting factors to the respective linear functions, the estimated coefficient of friction µS can also be expressed directly for control purposes using a weighting factor assignable to the estimated coefficient of friction µM, i.e., as µS ≈ gSM * µM, where gSM corresponds to the weighting factor with respect to the model value. In one embodiment, it is not the model value of the coefficient of friction µM that is directly weighted, but rather a deviation of the coefficient of friction µM from the reference coefficient of friction µref. The estimated coefficient of friction µS for control purposes can thus be expressed, for example, as µS ≈ µref + gref * (µM - µref), with gref as the reference deviation weighting factor. The weighting factor g SM corresponds here to g SM = µ ref / µ M + gref * (1 - µ ref / µ M).

[0058] In a further aspect, the invention also relates to a method for brake control for a rail vehicle, wherein a central brake control device of the rail vehicle controls one or more local primary brake controls. Each local primary control (e.g., per car or bogie) controls at least one brake device and is adjusted depending on the coefficient of friction estimated by applying the method described above.

[0059] In the aforementioned method, the estimated coefficient of friction is used to adapt a primary brake control system. Accordingly, at least one brake device can be controlled more precisely, for example, based on the estimated coefficient of friction compared to a constant coefficient of friction, with regard to current friction conditions. Alternatively or additionally, the adaptation can also consist of adjusting the control behavior itself, and not just the value of the coefficient of friction used in the control system. With regard to adapting the control behavior, one variant can also provide that the estimated coefficient of friction is only considered in the primary control system in predetermined cases, such as during certain braking operations or states of the brake device, particularly in cases that are especially sensitive to changes in the coefficient of friction.This selective consideration can relieve the primary control system.

[0060] The brake control method can be used, in particular, to control one or more local primary brake control units for respective brake devices via the central brake control device, whereby the primary brake control is adapted locally according to the locally estimated coefficient of friction. If locally different coefficients of friction exist, a locally adapted primary brake control can increase the overall accuracy of the primary brake controls.

[0061] For example, the central brake control device controls the local primary brake control units distributed throughout a train according to a predefined control strategy, whereby the primary brake control is adjusted locally based on locally estimated coefficients of friction. The train can typically have several local primary brake control units, e.g., at least one per car or bogie. The central brake control unit distributes the total train braking force to the individual local primary brake control units according to specific criteria. The primary brake control used for this purpose and its local adjustment, as already mentioned, has a positive effect on local force application and thus also on the overall accuracy of the train's total braking force and consequently on the train's deceleration. Braking distances can therefore be adhered to more effectively.

[0062] In one embodiment, the primary brake control is implemented as a feedforward control of a deceleration control unit for at least one brake device.

[0063] For example, this allows for the combination of train-wide deceleration control with a local friction coefficient estimation, as described above. The friction coefficient estimations of the individual wheelsets or bogies can act as feedforward control for the deceleration control unit, such as a train-wide deceleration controller, via the primary brake control and its adjustment. This supports the controller's function, meaning it only needs to compensate for minimal differences in the resulting train deceleration. Similarly, in combination with the deceleration control, the accuracy requirements for the friction coefficient estimator are moderate, as the train-wide deceleration control can compensate for deviations or inaccuracies in the friction coefficient estimation overall.For this reason, determining the parameters for the characteristic curves is significantly simplified, as, among other things, fewer test drives directly with the vehicle and / or fewer test bench runs may be sufficient.

[0064] In a training course, a regulation via the delay control unit is implemented if a difference between a requested and actual delay exceeds a predetermined upper limit.

[0065] The braking control of at least one braking device is therefore only carried out via a control intervention by the deceleration control unit if a predetermined deceleration deviation exists. Otherwise, the braking control corresponds to the feedforward control by the primary control unit. The deceleration control unit is thus further relieved of its workload.

[0066] The predetermined upper limit can be adjustable. For example, a higher upper limit for deceleration deviation can be provided for service braking than in the case of emergency braking. Similarly, this can also be linked to the state of the braking system, whereby a braking performance classified as reduced based on the state may necessitate a lower upper limit.

[0067] In one variant of the brake control procedure, the primary brake control is adjusted at least when the difference between a requested and actual deceleration exceeds a predetermined upper limit.

[0068] This method variant can apply to cases where the primary brake control itself controls the braking device or is used as feedforward control for a deceleration control system. The adaptation in this context refers in particular to adjusting the control behavior by taking into account the estimated coefficient of friction and / or adjusting weighting factors.

[0069] The primary control is therefore only adjusted if there is a corresponding deviation between the requested and actual deceleration according to the predetermined upper limit, and can thus be limited to such cases. The upper limit can also be defined variably, for example, depending on external boundary conditions and / or optimization criteria. Alternatively or additionally, the respective limit values ​​and / or the adjustment of the primary control itself can also be determined depending on an expected influence of the system values ​​on the coefficient of friction and, consequently, on a deviation between the requested and actual deceleration. For example, an adjustment can be made when a specific brake disc temperature, friction speed, and / or clamping force is reached or exceeded.

[0070] In another variant, the estimated coefficient of friction is only taken into account by the primary control system in predetermined application cases.

[0071] For example, the normal operation of a rail vehicle can be relatively tolerant of deviations between the actual coefficient of friction and a coefficient of friction estimated to be constant. In emergency situations, such as an emergency stop, or with regard to further optimization requirements, such as noise protection limited to a section of track, a more precise estimate can become essential.

[0072] In another aspect, the invention relates to a brake control device for a rail vehicle for providing a clamping force on a brake disc, wherein the brake control device is configured such that the clamping force can be controlled as a function of the coefficient of friction by applying the method for estimating the coefficient of friction or the method for brake control for a rail vehicle.

[0073] The brake control can be a central brake control system as described in the brake control procedure, but it can also be implemented locally. Furthermore, the estimation of the coefficient of friction in a central brake control system need not be limited to adapting a primary control system, but can also be directly integrated into the direct control of brake components or the provision of the clamping force on a brake disc.

[0074] Furthermore, in another aspect, the invention relates to a computer program product with program code stored on a machine-readable medium, which, when executed on a data processing device, is configured to cause the data processing device to execute the method described above for estimating a coefficient of friction or the method described above for brake control for a rail vehicle.

[0075] The advantages arise analogously to the above explanations with regard to the respective procedures.

[0076] The invention will now be explained in more detail with reference to the accompanying figures. The figures show, in detail: Figure 1a model for estimating the coefficient of friction as a function of a brake disc temperature, a friction speed and a clamping force as a first exemplary embodiment of a method according to the invention for estimating a coefficient of friction; Figure 2 a model for estimating the coefficient of friction as a function of a brake disc temperature and a friction speed as a second exemplary embodiment of the inventive method for estimating a coefficient of friction; Figure 3 a model for estimating the coefficient of friction as a function of a frictional speed as a third exemplary embodiment of the inventive method for estimating a coefficient of friction; Figure 4 a block diagram as an exemplary first embodiment of a method according to the invention for brake control; Figure 5a block diagram as an exemplary second embodiment of a brake control method according to the invention.

[0077] Figure 1Figure 1 shows a model for estimating a coefficient of friction µ as a function of a brake disc temperature T, a frictional speed v, and a clamping force Fb as a first exemplary embodiment of a method according to the invention for estimating a coefficient of friction µ. First, the coefficients of friction µT, µv, and µFb are determined as a function of their respective system values: brake disc temperature T, frictional speed v, and clamping force Fb. The coefficient of friction µT corresponds to a coefficient of friction at different brake disc temperatures T, where the clamping force Fb, here exemplified as 25 kN, and the frictional speed v, here exemplified as 15 m / s, are held constant or assumed to be constant. Similarly, the coefficient of friction µv corresponds to a coefficient of friction at different frictional speeds v, where the brake disc temperature T and the frictional speed v are held constant or assumed to be constant.The mean brake disc temperature Tm, here exemplified as 100 °C, and the clamping force Fb, here exemplified as 25 kN, are held constant or assumed to be constant. Similarly, the coefficient of friction µFb corresponds to a coefficient of friction at different clamping forces Fb, where the brake disc temperature TT or the mean brake disc temperature Tm, here exemplified as 100 °C, and the friction velocity v, here exemplified as 15 m / s, are held constant or assumed to be constant. The coefficients of friction that depend on one of the system values ​​are thus independent of changes in the other system values ​​mentioned. The respective coefficients of friction µT, µv, and µFb can be determined empirically, for example, using test bench experiments, and / or by means of other model calculations.

[0078] The behavior of the determined friction coefficients µT, µv, and µFb is approximated in a second step using piecewise linear functions. For example, the friction coefficient µT as a function of the brake disc temperature T is described by the piecewise linear function f1(T). In a range of brake disc temperatures T < T0, i.e., in a domain T < T0, f1(T) is constant. In a range of brake disc temperatures T0 ≤ T1 ≤ T1, i.e., in a domain T0 ≤ T1 ≤ T1, the function value of f1(T), representing the friction coefficient µT, decreases linearly with increasing temperature. The function f1(T) in the domain T0 ≤ T1 ≤ T1 is chosen such that the predetermined lower bound cT,min is not undercut.

[0079] Similarly, the coefficient of friction µv is described as a function of the frictional speed v by the piecewise linear function f2(v). In a region of frictional speed v0 ≤ v ≤ v1, i.e., in a domain v0 ≤ v ≤ v1, the function value of f2(v), representing the coefficient of friction µv, decreases linearly with frictional speed. The function f2(v) in the domain v0 ≤ v ≤ v1 is chosen such that the predetermined upper limit cv,max is not exceeded. In a region of frictional speed v < v0, i.e., in a domain v < v0, f2(v) is constant.

[0080] The coefficient of friction µFb is described as a function of the clamping force Fb by the piecewise linear function f3(Fb). In a region of clamping force Fb < F0, i.e., in a domain Fb < F0, f3(Fb) is constant. In a region of clamping force F0 ≤ Fb ≤ F1, i.e., in a domain F0 ≤ Fb ≤ F1, the function value of f3(Fb), representing the coefficient of friction µFb, increases linearly with increasing friction velocity. The function f3(Fb) in the domain F0 ≤ Fb ≤ F1 is chosen such that the predetermined upper limit cFb,max is not exceeded.

[0081] The estimated friction value µ is estimated from the piecewise linear functions f 1 (T), f 2 (v) and f 3 (F b ) according to the following equation: μ ≈ f 0 * f 1 T * f 2 v * f 3 F b , where f 0 represents the reference friction coefficient µ ref and f 1 (T) is the linear function as a function of the brake disc temperature (T), f 2 (v) is the linear function as a function of the friction speed (v) and f 3 (F b ) is the linear function as a function of the clamping force (F b ).

[0082] The reference coefficient of friction µref is the coefficient of friction that occurs at low temperatures and moderate speeds. In these ranges, the coefficient of friction is nearly constant, resulting in an approximate reference friction coefficient plateau. The value f0 can be a constant, either directly equal to the reference coefficient of friction µref or at least proportional to it.

[0083] Based on the above approach of friction coefficient factorization, the current friction coefficient µ or µ(T, v, F b ) can therefore be estimated with a good approximation as the product of the friction coefficient factors from the individual characteristic curves multiplied by the reference friction coefficient µ ref or the constant f 0 representing this reference friction coefficient µ ref.

[0084] To estimate the coefficient of friction µ in relation to the reference coefficient of friction µ ref, the piecewise linear functions f 1 (T), f 2 (v) and f 3 (F b ) are normalized to the reference coefficient of friction µ ref.

[0085] For example, the piecewise linear functions f 1 (T), f 2 (v) and f 3 (F b ) are obtained from the piecewise linear functions F 1 (T), F 2 (v) and F 3 (F b ) as follows: f 1 T : = 1 μ ref ∗ F 1 T mit der unteren Schranke C T , min : = 1 μ ref ∗ minF 1 T , f 2 v : = 1 μ ref ∗ F 2 v mit der oberen Schranke C v , max : = 1 μ ref ∗ maxF 2 v and f 3 F b : = 1 μ ref ∗ F 3 F b mit der oberen Schranke C Fb , max : = 1 μ ref ∗ maxF 3 F b .

[0086] It should be noted that when using the temperature characteristic curve, the current brake disc temperature must be known, at least as a good approximation. If sensor-based temperatures cannot be used as measured values, the implementation of a correctly parameterized temperature model is preferably used for estimation.

[0087] Figure 2Figure 1 shows a model for estimating the coefficient of friction µ as a function of a brake disc temperature T and a friction speed v as a second exemplary embodiment of the inventive method for estimating a coefficient of friction. The second embodiment differs from the first embodiment in that the dependence of the coefficient of friction on the clamping force Fb is neglected, and thus the model consists of only two generic characteristics or system value dependencies, which allows for an even simpler, but in many cases sufficiently accurate, compensation of the behavior of the brake pad-brake disc friction coefficient.

[0088] The estimated friction value µ is therefore estimated from the piecewise linear functions f 1 (T) and f 2 (v) according to the following equation: μ ≈ f 0 * f 1 T * f 2 v where f 0 represents the reference friction coefficient µ ref and f 1 (T) is the linear function as a function of the brake disc temperature (T), and f 2 (v) is the linear function as a function of the friction speed (v).

[0089] Figure 3 Figure 1 shows a model for estimating the coefficient of friction µ as a function of a frictional speed v as a third exemplary embodiment of the inventive method for estimating a coefficient of friction.

[0090] This further simplification is based on the fact that the dependence on the clamping force Fb and the temperature T can be neglected, meaning the model consists of only a single generic characteristic. This minimizes the need for additional parameters and eliminates the need for additional temperature measurements or estimations, for example, based on a temperature model. This can be of practical interest because it allows for compensation of the typical increase in the coefficient of friction during a braking maneuver, thus enabling more precise braking. As mentioned earlier, this can be particularly advantageous for trains with an ATO (Automatic Train Operation) system.

[0091] The estimated friction value µ is therefore estimated from the piecewise linear function f 2 (v) according to the following equation: μ ≈ f 0 * f 2 v where f 0 represents the reference friction coefficient µ ref and f 2 (v) is the linear function as a function of the friction speed (v).

[0092] Figure 4 Figure 1 shows a block diagram as an exemplary first embodiment of a brake control method according to the invention using a brake control device 1. The method uses an estimated coefficient of friction µ according to the preceding explanations.

[0093] The brake control device 1 comprises a central brake control device 10, a deceleration control unit 25, and a primary brake control unit 20 for a brake device 30. The deceleration control unit 25 can be part of the central brake control unit 10 and thus also be usable across multiple trains. Even if in Fig. 4While only one primary brake control unit 20 is shown as an example, the central brake control unit 10 or the deceleration control unit 25 can be connected to more than one primary brake control unit 20. These can be respective local primary brake control units 20, which in turn are connected to at least one locally assigned brake device 30. The primary brake control for the local primary brake control unit 20 shown here is determined by the central brake control device 10 or the deceleration control unit 25.

[0094] Furthermore, the brake control device includes an estimation unit 40, also referred to as FCA. The estimation unit, which acts directly on the primary brake control unit 20, serves as a feedforward control for the deceleration control unit 25. For this purpose, a friction coefficient µ, estimated via the estimation unit 40, can be transmitted to the primary brake control unit 20. The estimation unit 40 estimates the friction coefficient µ according to one of the aforementioned method variants based on currently determined and / or estimated system values ​​of the brake disc temperature T, the friction velocity v, and / or the clamping force Fb. The estimation unit 40 can, for example, be configured separately or as part of the primary control unit 20. The estimated friction coefficient is continuously available to the primary brake control unit 20 as an example, or is taken into account in the primary brake control as needed.Alternatively, the primary brake control unit 20 can be used, as described later in the procedure variant. Figure 5 As described further below, the estimated coefficient of friction is only taken into account in predefined situations, for example, depending on the specific braking process, such as service braking or emergency braking, and / or the brake system condition. In other cases, the primary brake control can also be based on a constant coefficient of friction.

[0095] The primary brake control provided by the primary brake control unit 20 thus forms a feedforward control signal for the higher-level deceleration control by the deceleration control unit 25. Because the estimation unit 40 controls the brake pressure and thus the braking force more precisely for each assigned primary brake control unit 20 to adapt the primary brake control, smaller deviations in train deceleration also result, which the higher-level deceleration control has to compensate for less. Therefore, the sum of the local estimation units 40, in conjunction with the respective local primary brake control units 20, serves as feedforward control for the higher-level, train-wide deceleration control.In other words, the use of local primary brake control units 20 and local estimation units 40 enables local adjustment of the force application, which overall more accurately predicts the train-wide deceleration and thus reduces the load on the deceleration control system. It should be noted that local estimation units can improve the accuracy of the primary brake control adjustment, but a train-wide estimation unit 40 can also be used. While this may be less accurate, it can simplify the required data acquisition and / or reduce the number of sensors required.

[0096] In the Figure 4In the illustrated method configuration, the braking device 30 is monitored directly or indirectly via a monitoring unit 50. The monitoring unit 50 determines, for example, the actual deceleration a IST of the braking device 30. If the difference between the actual deceleration a IST and the requested deceleration a SOLL exceeds a predetermined upper limit Δa max, the deceleration control is activated via the deceleration control unit 25, as indicated by the arrow marked "yes" in Figure 4 This is indicated. The delay mechanism therefore only works when needed.

[0097] However, it can also be provided that the deceleration control and feedforward control or primary brake control operate in parallel and permanently, irrespective of monitoring unit 50 and its associated monitoring results or other conditions. This is in Figure 4The dashed arrow indicates the path from the brake device 30 to the deceleration control unit 25. Data from the brake device 30 is transmitted to the deceleration control unit 25. Alternatively or additionally, data to be forwarded to the deceleration control unit 25 can also be provided by the monitoring unit 50. With sufficiently accurate feedforward control based on the estimated coefficient of friction, the deceleration control unit can be sufficiently relieved of the load to only have to compensate for small differences or intervene at all. The deceleration deviation thus directly affects the deceleration control unit 25 and its control of the primary control 20 and the brake device 30. However, the intervention of the deceleration control, triggered above by exceeding a limit value, can, for example, offer the advantage that the intervention condition can be easily adjusted.Provided the deceleration control is permanently running, changed boundary conditions can also be taken into account and compensated for alternatively or additionally via the deceleration control unit, for example via a changed control behavior depending on various factors, such as the type of braking process and / or the state of the braking device 30.

[0098] Figure 5 Figure 1 shows a block diagram as an exemplary second embodiment of a brake control method according to the invention, using a brake control device 1'. The method also uses an estimated coefficient of friction µ as described above.

[0099] The brake control device 1' comprises, as already mentioned in section 1 Figure 4As described, a central brake control device 10 specifies a primary brake control for a brake device 30 to a primary brake control unit 20. The primary brake control unit 20 can also be part of the central brake control unit 10 or be designed separately from it.

[0100] The method of the second embodiment differs from the first embodiment essentially in that the focus here is on adapting the primary brake control via the primary brake control unit 20. In a further development, the primary brake control unit 20 can also be combined with a deceleration control unit 25, as described for the first embodiment, or with a comparable deceleration control system. The deceleration control unit 25 can be combined with the central brake control device 10.

[0101] In the illustrated embodiment, the brake control device 1' also includes a monitoring unit 50, which monitors the actual deceleration a IST by the primary brake control. If the difference between the actual deceleration a IST and the requested deceleration a SOLL exceeds a predetermined upper limit Δa max, the primary brake control of the primary brake control unit 20 is adjusted. For this purpose, a coefficient of friction µ, estimated via an estimation unit 40, is transmitted to the primary brake control unit 20. The estimation unit 40 estimates the coefficient of friction µ according to one of the preceding method variants based on currently determined and / or estimated system values ​​of the brake disc temperature T, the friction velocity v, and / or the clamping force F b. The estimation unit 40 can, for example, be configured separately or as part of the primary control unit 20.Furthermore, the estimation unit 40 can estimate the coefficient of friction µ continuously or periodically, regardless of whether the upper limit Δa max is exceeded, and transmit the estimate only as needed. Alternatively, the estimation unit 40 can estimate the coefficient of friction µ only when the upper limit Δa max is exceeded, in order to transmit the estimated coefficient of friction µ to the primary control unit 20. The adaptation of the primary brake control unit 20 is described in the [reference to be added]. Figure 4The described method variant focuses on taking into account the friction coefficient estimated under the respective boundary conditions, depending on the delay deviation. In general, however, the term "adaptation of the primary control" also encompasses the use of a friction coefficient that changes according to the estimate and the associated adjustment of the manipulated variable. An adaptation can also entail a change in the control behavior itself, for example, an adjustment of the transfer function.

[0102] The invention is not limited to the described embodiments. Even if, for example, a linear dependence of the coefficient of friction on the clamping force is not taken into account in the second and third embodiments of the method for estimating a coefficient of friction, it can be used instead of the brake disc temperature and / or friction speed, provided that not all system values ​​are taken into account equally anyway. REFERENCE MARK LIST

[0103] 1,1'Brake control device 10 Central brake control device 20 Primary brake control unit 25 Deceleration control unit 30 Brake device 40 Estimation unit 50 Monitoring unit µ Estimated coefficient of friction µ max Upper limit (coefficient of friction) µ min Lower limit (coefficient of friction) µ ref Reference coefficient of friction C Fb,min / max Lower or upper limit for f 3 (F b ) CT,min / max Lower or upper limit for f 1 (T) C v,min / max Lower or upper limit for f 2 (v) F b Clamping force T Brake disc temperature v Friction speed a ACTUAL Actual deceleration a SET Requested deceleration Δa max Upper limit

Claims

1. Method for estimating a coefficient of friction (µ) for a brake pad / brake disc pairing, wherein the coefficient of friction (µ) is estimated depending on a reference coefficient of friction (µref) via at least one linear function of the brake disc temperature (T) and / or of the frictional speed (v) and / or of the clamping force (Fb).

2. Method according to Claim 1, wherein the coefficient of friction (µ) is estimated according to the following equation: μ ≈ f 0 * f 2 v , wherein f0 represents the reference coefficient of friction µref and f2(v) is the linear function depending on the frictional speed (v).

3. Method according to Claim 1, wherein the coefficient of friction (µ) is estimated according to the following equation: μ ≈ f 0 * f 1 T * f 2 v , wherein f0 represents the reference coefficient of friction µref and f1(T) is the linear function depending on the brake disc temperature (T) and f2(v) is the linear function depending on the frictional speed (v).

4. Method according to Claim 1, wherein the coefficient of friction (µ) is estimated according to the following equation: μ ≈ f 0 * f 1 T * f 2 v * f 3 F b , wherein f0 represents the reference coefficient of friction µref and f1(T) is the linear function depending on the brake disc temperature (T), f2(v) is the linear function depending on the frictional speed (v) and f3(Fb) is the linear function depending on the clamping force (Fb).

5. Method according to any one of the preceding claims, wherein the functions f1(T), f2(v) and / or f3(Fb) are at least piecewise linear functions.

6. Method according to any one of the preceding claims, wherein the functions f1(T), f2(v) and / or f3(Fb) are standardized to the reference coefficient of friction (µref).

7. Method according to any one of the preceding claims, wherein the value ranges of the functions f1(T), f2(v) and / or f3(Fb) in the definition range of the functions are limited by limiting parameters to a lower and / or upper limit cT, cv and / or cFb.

8. Method according to Claim 7, wherein the function f1(T) is selected in such a manner that the upper limit cT,max is smaller than or equal to a value representing the reference coefficient of friction (µref) and / or the lower limit cT,min represents a maximally applicable clamping force (Fb).

9. Method according to Claim 7 or 8, wherein the function f2(v) is selected in such a manner that the lower limit cv,min is greater than or equal to a value representing the reference coefficient of friction (µref) and / or the upper limit cv,max represents a minimally applicable clamping force (Fb).

10. Method according to any one of Claims 7 to 9, wherein the function f3(Fb) is selected in such a manner that the upper limit cFb,max and / or the lower limit cFb,min are / is adapted to a brake pad / brake disc pairing.

11. Method according to any one of the preceding claims, wherein the linear functions f1(T), f2(v) and / or f3(Fb), in particular at least piecewise depending on the respective definition range, are assigned different weighting factors in the equation for estimating the coefficient of friction (µ).

12. Method for brake control for a rail vehicle, wherein a central brake control device (10) of the rail vehicle predetermines a primary brake control for at least one brake device (30), and the primary brake control is adapted depending on the coefficient of friction (µ) estimated by use of the method according to any one of Claims 1 to 11.

13. Method according to Claim 12, wherein the primary brake control is performed as a pilot control of a deceleration regulation unit (25) for the at least one brake device (30).

14. Method according to Claim 13, wherein regulation by the deceleration regulation unit (25) is performed if a difference between a requested and actual deceleration (aREQ, aACT) exceeds a predetermined upper limit value (Δamax).

15. Brake control device (1) for a rail vehicle for supplying a clamping force (Fb) to a brake disc, wherein the brake control device (1) is configured in such a manner that the clamping force (Fb) can be controlled depending on the coefficient of friction (µ) using the method according to any one of Claims 1 to 11 or the method according to any one of Claims 12 to 14.

16. Computer program product having a program code which is stored on a machine-readable medium and is designed, if it is executed on a data processing device, to cause the data processing device to carry out the method according to any one of Claims 1 to 11 or the method according to any one of Claims 12 to 14.