METHOD AND DEVICE FOR DETERMINING WHEEL GEOMETRY AND RAIL VEHICLE

Description

[0001] The invention relates to a method for determining wheel geometry for vehicles, in particular for rail vehicles, wherein vertical accelerations of at least one first wheel are determined by means of at least one first sensor, wherein driving speeds are processed, and wherein amplitude spectra are formed on the basis of vertical acceleration signals which characterize a vibration behavior of the at least first wheel.

[0002] Vehicles often require precise knowledge of wheel geometry. For example, a wheel radius, wheel diameter, or wheel circumference is needed to determine wheel or wheelset rotational speeds, for certain monitoring and diagnostic functions concerning wheels or wheelsets, or for determining speed via odometry, etc.

[0003] For example, EP 1 197 415 A2 is known from the prior art, in which a method for detecting bearing damage is described. In this method, a defined number of outer ring rolling harmonics are determined from acceleration signals acquired in the area of ​​an axle bearing, and from these a characteristic value is derived that indicates a bearing condition.

[0004] Furthermore, EP 1 197 417 A1 discloses a method for detecting wheel damage in railway vehicles. In this method, a defined number of wheel out-of-roundness harmonics are determined from acceleration signals detected in the area of ​​an axle bearing, and a characteristic value is derived from these harmonics, which indicates the condition of the wheel.

[0005] Furthermore, WO 2018 / 059937 A1 shows a wheel arrangement for a rail vehicle with a speed sensor.

[0006] Furthermore, WO 95 / 30886 A1 discloses a device for detecting defects in wheelsets of railway vehicles and in track for railway vehicles. The device comprises a processor in which measurement results relating to rotation rates on the one hand and vertical accelerations or vibrations on the other hand are processed to detect a defect in a wheelset or in a track.

[0007] The invention is based on the objective of providing a vibration-based method for determining wheel geometry.

[0008] According to the invention, this problem is solved by a method according to claim 1, in which an amplitude sum spectrum is generated from the amplitude spectra, wherein at least one wheel harmonic is determined from at least one frequency of the amplitude sum spectrum, which is assigned to an amplitude maximum of the amplitude sum spectrum, and wherein a wheel geometry is determined from a standard driving speed, which defines a characteristic vehicle operating behavior, and from the at least one wheel harmonic. This measure results in a precise and, at the same time, material-efficient method for determining wheel geometry, which can, for example, also be implemented in vehicles that do not have speed sensors. In such implementations, the costly retrofitting of speed sensors and associated wiring can be avoided.However, it is also possible that the inventive method is used redundantly with a wheel geometry determination based on measurements from a speed sensor, thus ensuring the reliability of the determination results.

[0009] Furthermore, many vehicles have acceleration sensors in the area of ​​their wheels that measure vertical accelerations. For example, for wheelset bearing diagnostics in the bogies of rail vehicles, acceleration sensors for detecting vertical accelerations are frequently arranged on the wheelset bearing housings of the bogies. Such equipment can be used by the method according to the invention.

[0010] In the method according to the invention, the wheel harmonic can be identified, for example, as the frequency of the sum of amplitudes in the spectrum that corresponds to an amplitude maximum within that spectrum. Alternatively, the wheel harmonic can be selected from a plurality of frequencies that correspond to amplitude maxima of different search intervals within the sum of amplitudes. The standard driving speed can be, for example, an average speed over a defined period or a defined constant, etc. Driving speeds can be measured, determined, etc. (for example, by deriving position data from a tracking device over time).

[0011] Further advantageous embodiments of the method according to the invention are set out in the dependent claims.

[0012] For example, it is advantageous if the wheel geometry uses a wheel radius defined by a radius calculation rule. r = v n 2 π ⋅ f h or a wheel diameter by means of a diameter formation rule R = v n π ⋅ f h The wheel radius or diameter is determined. Knowing the wheel radius or diameter is advantageous for various applications. For example, the determined wheel radius or diameter can be used in an odometric method for determining vehicle speed using speed sensors or in a wheel speed determination process.

[0013] In equivalent to the wheel radius or wheel diameter, a wheel circumference can also be determined, for example.

[0014] A preferred solution is achieved when at least one wheel harmonic is determined, excluding interfering frequencies caused by processes other than the rotation of at least the first wheel.

[0015] This allows for more accurate determination of the wheel geometry, as interfering frequencies are filtered out. This measure enables the creation of a quality metric, which increases, for example, with the number of frequencies used in the wheel geometry determination (i.e., frequencies that are not filtered out). The higher the quality metric, the greater the accuracy or plausibility of a determined wheel geometry value.

[0016] Parameterization and thus an increase in the efficiency of wheel geometry determination are made possible if at least one wheel harmonic is derived from a search area consisting of the standard driving speed, a defined maximum permissible wheel radius, a defined minimum permissible wheel radius, and an order factor with k = 1 ... K from the first intervals. k ⋅ v n 2 π ⋅ r max k ⋅ v n 2 π ⋅ r min or from the standard driving speed, a defined maximum permissible wheel diameter, a defined minimum permissible wheel diameter, and an order factor with k = 1 ... K from second intervals k ⋅ v n π ⋅ R max k ⋅ v n π ⋅ R min The selected frequencies are determined.

[0017] Robust results for determining wheel geometry, in contrast to statistical outliers, are achieved when at least one wheel harmonic is determined as the median from the frequencies selected from the search area, which are assigned to amplitude maxima of the sum amplitude spectrum within the first intervals or the second intervals.

[0018] Determining wheel speed without the use of speed sensors is made possible if a wheel speed can be calculated from a value of the vehicle speed and the wheel radius using a first speed calculation rule. n = v 2 π ⋅ r or from a value of the driving speeds and the wheel diameter by means of a second speed calculation rule n = v π ⋅ R is determined.

[0019] However, for determining rotational speed, it can also be helpful if at least one core wheel harmonic is derived from frequencies within a core search area, which is based on initial search values. k ⋅ v 2 π ⋅ r , which are formed from a value of the driving speeds, the wheel radius and an order factor with k = 1 ... K or by second search values k ⋅ v π ⋅ R , which is formed from a value of the vehicle speeds, the wheel diameter and an order factor with k = 1 ... K, is determined, whereby a wheel speed is determined from the at least one core wheel harmonic and from at least one value for the order factor, for which a frequency selected from the core search area as the at least one core wheel harmonic is assigned to an amplitude maximum of the amplitude sum spectrum, by means of a third speed formation rule n = f h 1 k or as a statistical value from quotients formed by dividing a plurality of values ​​for the core-wheel harmonic by the values ​​for the order factor assigned to each of the core-wheel harmonic values.

[0020] This measure ensures sufficiently accurate and plausible results for wheel speed even with unreliable or inaccurate information regarding vehicle speeds. The vehicle speed value is only used to determine the core wheel harmonics, where the core wheel harmonic or a plurality of core wheel harmonics is searched for within the core search area, which is, for example, smaller than the search area for the wheel harmonic or a plurality of wheel harmonics. The wheel speed is then determined from the core wheel harmonic or the plurality of core wheel harmonics, for which the vehicle speed value is not directly required.

[0021] High accuracy of determination results with regard to wheel speed is ensured if the wheel speed is determined when at least one wheel harmonic or at least one core wheel harmonic is determined from a number of frequencies equal to or greater than a frequency number threshold.

[0022] This prevents, for example, the wheel harmonic or the core wheel harmonic from being determined from a frequency spectrum which has, for example, a high number of interfering frequencies to be filtered out.

[0023] Risks regarding unreliable wheel geometry determination results are reduced if the amplitude spectra and / or the sum amplitude spectrum are then generated and / or updated when the vehicle speeds are equal to or greater than a vehicle speed threshold and when vehicle accelerations are equal to or less than a vehicle acceleration limit.

[0024] Vehicle accelerations can be mathematically positive or mathematically negative (i.e., decelerations). This can be accounted for by specifying the sign of the vehicle acceleration limit or by calculating the absolute value of the vehicle accelerations. Similarly, driving or braking forces can be compared using a driving or braking force limit.

[0025] Efficient utilization of available computing capacity is made possible if the amplitude spectra are formed in a first time interval and the sum amplitude spectrum is updated in a second time interval, where the second time interval is larger than the first time interval.

[0026] In this context, it is also helpful to form partial sum amplitude spectra from the amplitude spectra, which are temporarily stored in a third time interval that is larger than the first time interval and smaller than the second time interval, whereby the sum amplitude spectrum is formed from the partial sum amplitude spectra.

[0027] A promising application of the method according to the invention is made possible in a device comprising at least one first sensor, at least one computing unit and at least one device for locating, determining or recording driving speed, if the at least one computing unit for determining the geometry of at least one first wheel is connected to the at least first sensor for detecting vertical accelerations and to the at least one device for locating, determining or recording driving speed by means of signal transmission.

[0028] In particular, it is advantageous if a rail vehicle is equipped with at least one device according to the invention.

[0029] For rail vehicles, there is often a need for simple but precise wheel geometry determination (for example in connection with diagnostic and / or monitoring devices, which can be used to detect wheel damage, etc.).

[0030] The invention will now be explained in more detail using exemplary embodiments.

[0031] They show, for example: Fig. 1: A flowchart of an exemplary embodiment of a method according to the invention, with which wheel geometry parameters and wheel speeds are determined, and Fig. 2: A side view of a section of an exemplary embodiment of a rail vehicle according to the invention with an exemplary embodiment of a device according to the invention for carrying out a method according to the invention for determining wheel geometry.

[0032] Fig. 1 shows a flowchart for an exemplary embodiment of a method according to the invention, with which wheel geometry parameters and wheel speeds are determined.

[0033] The procedure according to Fig. 1 is intended for a rail vehicle, such as the one found in Fig. 2 is illustrated. By means of a first sensor 1, as exemplified in Fig. 2 The vertical accelerations of a first wheel 2 of a wheelset 3, as exemplified in [reference], are shown. Fig. 2 is shown, determined. From a similarly in Fig. 2 The satellite-based positioning device 4 of the rail vehicle, shown as an example, receives position information of the rail vehicle, which is transformed by temporal derivation into travel speeds v to be processed in connection with the method according to the invention.

[0034] Based on vertical acceleration signals from the first sensor 1, which characterize the vibration behavior of the first wheel 2, amplitude spectra are generated in a first time interval of 1 s. These spectra are then normalized to a standard driving speed vn, which is calculated as an average, i.e., a statistical value, from the driving speeds v recorded within a day (amplitude spectrum generation 5). According to the invention, it is also possible to use a different value for the standard driving speed vn (e.g., a constant that characterizes a typical or particularly frequent operating speed). In any case, the standard driving speed vn describes a characteristic operating behavior of a vehicle.

[0035] The amplitude spectra are accumulated, thereby forming partial sum amplitude spectra (accumulation 6) and presented in a Fig. 2 The data from the railway vehicle's database 7 shown are temporarily stored (intermediate storage 8). The partial sum amplitude spectra are then aggregated into a sum amplitude spectrum in a second time interval of 1 day (aggregation 9). This intermediate storage takes place in a third time interval of 1 hour. The second time interval is therefore longer than the first, and the third time interval is longer than the first but shorter than the second.

[0036] The amplitude spectrum formation 5, the accumulation 6, the intermediate storage 8, the aggregation 9 and all further steps of the method according to the invention are carried out under the condition that the driving speeds v are equal to or greater than a defined driving speed threshold and when driving accelerations are equal to or less than a defined driving acceleration limit.

[0037] The vehicle accelerations can be mathematically positive or mathematically negative (i.e., decelerations). This is taken into account by specifying the sign of the vehicle acceleration limit or by calculating the absolute value of the vehicle accelerations. In the exemplary embodiment of the method according to the invention, the vehicle speed threshold is... Fig. 1 A value of 5 km / h is set, and a value of 0.01 m / s² is set as the driving acceleration limit. However, according to the invention, it is also possible to choose other values.

[0038] From frequencies of the amplitude sum spectrum, which are assigned to amplitude maxima of the amplitude sum spectrum, a wheel harmonic fh is determined (harmonic determination 10).

[0039] Harmonics are spectral lines whose frequencies have an integer ratio to each other. This occurs particularly in Fourier transforms of periodic, non-sinusoidal signals. A fundamental harmonic is, for example, the harmonic with the lowest frequency that corresponds to the reciprocal of one period of such a signal.

[0040] In harmonic determination 10, the wheel harmonic fh is determined from selected frequencies. The frequencies are selected from a search area defined by the standard driving speed vn, a defined maximum permissible wheel radius r max, a defined minimum permissible wheel radius r min, and an order factor k with k = 1 ... K (order maximum K with K = 12, whereby other values ​​for the order maximum K can also be used according to the invention) from first intervals k ⋅ v n 2 π ⋅ r max k ⋅ v n 2 π ⋅ r min is formed. The maximum permissible wheel radius r max and the minimum permissible wheel radius r min are operating limit dimensions.

[0041] According to the invention, it is also conceivable that the search area consists of second intervals. k ⋅ v n π ⋅ R max k ⋅ v n π ⋅ R min is formed in which, instead of the maximum permissible wheel radius r max and the minimum permissible wheel radius r min, a defined maximum permissible wheel diameter R max and a defined minimum permissible wheel diameter R min are inserted.

[0042] The selected frequencies are assigned to amplitude maxima of the sum-amplitude spectrum within the first intervals (or the second intervals if the search area is formed from these). That is, those frequencies are selected from frequency-amplitude profiles where the profiles exhibit amplitude maxima.

[0043] The selected frequencies are normalized to the order factor k assigned to the respective first or second interval from which a frequency is selected. For example, a frequency from the first interval with an order factor of k = 2 is divided by k = 2 for normalization.

[0044] Then, from the selected and normalized frequencies, interfering frequencies outside a defined frequency band, caused by processes other than the rotation of the first wheel 2 (e.g., recurring, dynamic disturbances from contact between the first wheel 2 and a rail), are filtered out. This allows the wheel harmonic fh to be determined while excluding these interfering frequencies. The greater the number of selected and unfiltered frequencies, the higher the value assigned to a quality measure for wheel geometry determination. The higher the quality measure, the more accurate and reliable the wheel geometry determination.

[0045] The radian harmonic fh is then determined as the median of the normalized and filtered frequencies selected from the search area. This completes the harmonic determination process.

[0046] Subsequently, a wheel geometry is determined from the standard driving speed vn and the wheel harmonic fh (geometry determination 11). The harmonic determination 10 and the geometry determination 11 are performed once daily.

[0047] The wheel geometry is defined by a wheel radius r using a radius calculation rule. r = v n 2 π ⋅ f h determined. According to the invention, however, it is also conceivable that, for example, a wheel diameter R could be determined as the wheel geometry by means of a diameter formation rule. R = v n π ⋅ f h The vertical acceleration of a second wheel of the wheelset 3 is determined. According to the invention, it is further conceivable that a second sensor is used to determine the vertical acceleration of a second wheel of the wheelset 3, and thus a wheel geometry determination is also carried out for the second wheel according to the scheme described above. The wheel radius r or the wheel diameter R can be determined, for example, as the mean radius or as the mean diameter with respect to the first wheel 2 and the second wheel.

[0048] Based on the determined wheel radius r or the determined wheel diameter R, a wheel rotational speed n is then determined (rotational speed determination 12), provided that the wheel harmonic fh is determined from a number of frequencies equal to or greater than a frequency threshold. Therefore, in order for the wheel rotational speed n to be determined, the aforementioned quality measure must be equal to or greater than the frequency threshold. The frequency threshold for the exemplary embodiment of a method according to the invention is as follows: Fig. 1 defined with 10. However, according to the invention, it is also possible to set other values ​​as the frequency threshold.

[0049] In the speed determination 12, the wheel speed n is calculated from a current value of the vehicle speeds v and the wheel radius r using a first speed calculation rule. n = v 2 π ⋅ r determined. However, according to the invention, it is also possible to determine the wheel speed n from a value of the vehicle speed v and the wheel diameter R by means of a second speed calculation rule. n = v π ⋅ R to determine.

[0050] Furthermore, especially in the case of unreliable information about the vehicle speeds v, it is conceivable that as an alternative method for determining rotational speed 12 a core wheel harmonic f h1 from frequencies within a core search area, which is centered around initial search values k ⋅ v 2 π ⋅ r , which are formed from a value of the driving speeds v, the wheel radius r and an order factor k with k = 1 ... K or to second search values k ⋅ v π ⋅ R , which are formed from a value of the driving speeds v, the wheel diameter R and an order factor k with k = 1 ... K, is determined, whereby the wheel speed n is determined from the core wheel harmonic f h1 and from that value for the order factor k for which a frequency selected from the core search area as the core wheel harmonic f h1 is assigned to an amplitude maximum of the amplitude sum spectrum, by means of a third speed formation rule n = f h 1 k The search area is determined. Centering around the first or second search values ​​is achieved by forming intervals. For example, third intervals around the first search values ​​are formed by replacing the wheel radius r with a lower limit between the values ​​for the wheel radius r and the maximum permissible wheel radius rmax in the first search values. Similarly, for an upper limit of the third intervals, the wheel radius r is replaced with an upper limit between the values ​​for the wheel radius r and the minimum permissible wheel radius rmin. The core search area is therefore narrower than the search area for the wheel harmonic fh mentioned above.

[0051] For the second search values, the same procedure can be used when forming intervals based on the wheel diameter R, the maximum permissible wheel diameter R max and the minimum permissible wheel diameter R min, whereby fourth intervals are formed around the second search values ​​by replacing the wheel diameter R with a lower limit value that lies between the values ​​for the wheel diameter R and the maximum permissible wheel diameter R max, and replacing the wheel diameter R with an upper limit value that lies between the values ​​for the wheel diameter R and the minimum permissible wheel diameter R min, for a lower limit value of the fourth intervals.

[0052] If a plurality of values ​​for the core wheel harmonic f h1 are selected from the core search area (e.g., for each value of the order factor k, a value for the core wheel harmonic f h1 is selected at an order maximum K of K > 1), the wheel speed n is determined as a statistical value (e.g., as the median) from quotients formed by dividing the plurality of values ​​for the core wheel harmonic f h1 by the values ​​for the order factor k assigned to each of the values ​​for the core wheel harmonic f h1.

[0053] Even in the alternative speed determination 12 using the core-wheel harmonics f h1, a frequency number threshold can be assumed.

[0054] If the wheel radius r or the wheel diameter R is determined and can be used for determining the rotational speed 12, then the amplitude spectrum generation 5, the accumulation 6, and the intermediate storage 8 can be performed in parallel with the rotational speed determination 12 (for example, by performing the amplitude spectrum generation 5, the accumulation 6, and the intermediate storage 8 using a first processor or a first processor core, and the rotational speed determination 12 using a second processor or a second processor core). The rotational speed determination 12 is performed every second.

[0055] In Fig. 2 is a side view of a section from an exemplary embodiment of a rail vehicle according to the invention with an exemplary embodiment of a device according to the invention for carrying out a method according to the invention for determining wheel geometry, as is exemplified in connection with Fig. 1 shown, depicted.

[0056] The rail vehicle has a bogie 13 and a car body 14, the car body 14 being supported on the bogie 13 by an air secondary spring arrangement 15. A first sensor 1 is connected to a cover of a first axle bearing housing 16 of a first axle bearing. The first sensor 1 is designed as a piezoelectric accelerometer.

[0057] A in Fig. 2 A second sensor, not visible and designed as a piezoelectric accelerometer, is equipped with a Fig. 2 The second wheelset bearing housing of the running gear 13, which is also not visible, is connected. A first wheelset 3, comprising a first wheel 2 and a [unclear], is connected via the first wheelset bearing and the first wheelset bearing housing 16, and the second wheelset bearing and the second wheelset bearing housing. Fig. 2 A second, non-visible wheel is sprung and coupled to a chassis frame 17.

[0058] The first sensor 1 detects vertical accelerations of the first wheel 2, and the second sensor detects vertical accelerations of the second wheel.

[0059] In the car body 14 a computing device 18 with an implemented database 7 is arranged, by means of which signals from the first sensor 1 and the second sensor are processed.

[0060] A satellite-based tracking device 4 with an antenna is connected to the roof of the car body 14. The tracking device 4 is designed as a Global Positioning System (GPS). The tracking device 4 is used to determine the travel speeds v of the rail vehicle. However, according to the invention, it is also conceivable to arrange a device for detecting or measuring the travel speeds v in or on the car body 14 (for example, a device for Doppler radar speed measurement, etc.) to determine the travel speeds v.

[0061] Based on the signals regarding vertical accelerations and vehicle speeds v, those related to Fig. 1 described steps of the exemplary embodiment of a method according to the invention, i.e. an amplitude spectrum formation 5, an accumulation 6, an intermediate storage 8, an aggregation 9, a harmonic determination 10, a geometry determination 11 and a speed determination 12 in the computing device 18.

[0062] The first sensor 1, the second sensor, the computing unit 18 and the locating unit 4 are parts of an exemplary embodiment of a device according to the invention for carrying out the method according to the invention for determining wheel geometry, as exemplified in connection with Fig. 1 is described.

[0063] The computing device 18 is used to determine a wheel radius r and a wheel rotational speed n, which are also used in connection with Fig. 1 are named, connected to the first sensor 1 and the second sensor as well as to the locating device 4 in a signal-transmitting manner.

[0064] A first signal line 19 is provided between the first sensor 1 and the computing unit 18, and a second signal line 20 between the tracking device 4 and a vehicle bus 22, which runs in the car body 14. The vehicle bus 22 is connected to the computing unit 18 via a third signal line 21 for data transmission from the tracking device 4 to the computing unit 18. The second sensor is connected via a Fig. 2 The fourth signal line, not shown, is connected to the computing unit 18.

[0065] According to the invention, it is also conceivable to provide radio connections between the computing device 18 on the one hand and the first sensor 1, the second sensor and the locating device 4 on the other. List of designations

[0066] 1 First sensor 2 First wheel 3 Wheelset 4 Locating device 5 Amplitude spectrum generation 6 Accumulation 7 Database 8 Intermediate storage 9 Aggregation 10 Harmonic determination 11 Geometry determination 12 Speed ​​determination 13 Chassis 14 Car body 15 Air secondary spring assembly 16 First wheelset bearing housing 17 Chassis frame 18 Computing device 19 First signal line 20 Second signal line 21 Third signal line 22 Vehicle bus vV ...

Claims

1. Method for wheel geometry determination for vehicles, in particular for rail vehicles, wherein vertical accelerations of at least one first wheel (2) are determined by means of at least one first sensor (1), wherein travel speeds (v) are processed, and wherein amplitude spectra are formed on the basis of vertical acceleration signals which characterise an oscillating behaviour of the at least one first wheel (2), characterised in that an amplitude sum spectrum is formed from the amplitude spectra, wherein at least one wheel harmonic (fh) is determined from at least one frequency of the amplitude sum spectrum which is assigned to an amplitude maximum of the amplitude sum spectrum, and wherein a wheel geometry is determined from a standard travel speed (vn) which defines a characteristic vehicle operating behaviour, and from the at least one wheel harmonic (fh).

2. Method according to claim 1, characterised in that a wheel radius (r) is determined as the wheel geometry by means of a radius formation rule r = v n 2 π ⋅ f h or a wheel diameter (R) is determined by means of a diameter formation rule R = v n π ⋅ f h .

3. Method according to claim 1 or 2, characterised in that the amplitude spectra are standardised to the standard travel speed (vn).

4. Method according to one of claims 1 to 3, characterised in that the at least one wheel harmonic (fh) is determined excluding interference frequencies which are caused by processes other than a rotation of the at least one first wheel (2).

5. Method according to one of claims 1 to 4, characterised in that the at least one wheel harmonic (fh) determines selected frequencies from a search area consisting of the standard travel speed (vn), a defined maximum permissible wheel radius (rmax), a defined minimum permissible wheel radius (rmin) and an order factor (k) with k = 1 ... K from first intervals k ⋅ v n 2 π ⋅ r max k ⋅ v n 2 π ⋅ r min or of the standard travel speed (vn), a defined maximum permissible wheel diameter (Rmax), a defined minimum permissible wheel diameter (Rmin) and an order factor (k) with k = 1 ... K from second intervals k ⋅ v n π ⋅ R max k ⋅ v n π ⋅ R min .

6. Method according to claim 5, characterised in that the selected frequencies are standardised to the order factor (k).

7. Method according to claim 5 or 6, characterised in that the at least one wheel harmonic (fh) is determined as the median of the frequencies selected from the search area, which are assigned to amplitude maxima of the amplitude sum spectrum within the first intervals or the second intervals.

8. Method according to one of the claims 2 to 7, characterised in that a wheel speed (n) is determined from a value of the travel speeds (v) and the wheel radius (r) by means of a first speed formation rule n = v 2 π ⋅ r or from a value of the travel speeds (v) and the wheel diameter (R) by means of a second speed formation rule n = v π ⋅ R .

9. Method according to one of claims 2 to 7, characterised in that at least one core wheel harmonic (fh1) is determined from frequencies within a core search area which is centred around first search values k ⋅ v 2 π ⋅ r , which are formed from a value of the travel speeds (v), the wheel radius (r) and an order factor (k) with k = 1 ... K or around second search values k ⋅ v π ⋅ R , which are formed from a value of the travel speeds (v), the wheel diameter (R) and an order factor (k) with k = 1 ... K, wherein a wheel speed (n) is determined from the at least one core wheel harmonic (fh1) and from at least one value for the order factor (k), for which a frequency selected from the core search area as the at least one core wheel harmonic (fh1) is assigned to an amplitude maximum of the amplitude sum spectrum, by means of a third speed formation rule n = f h 1 k or as a statistical value from quotients which are formed by dividing a plurality of values for the core wheel harmonic (fh1) by the values for the order factor (k) respectively assigned to the values for the core wheel harmonic (fh1).

10. Method according to claim 8 or 9, characterised in that the wheel speed (n) is determined when the at least one wheel harmonic (fh) or the at least one core wheel harmonic (fh1) is determined from a number of frequencies which is equal to or greater than a frequency number threshold value.

11. Method according to one of claims 1 to 10, characterised in that the amplitude spectra and / or the amplitude sum spectrum are formed and / or updated when the travel speeds (v) are equal to or greater than a travel speed threshold and when travel accelerations are equal to or less than a travel acceleration threshold.

12. Method according to one of claims 1 to 11, characterised in that the amplitude spectra are formed in a first time interval and the sum amplitude spectrum is updated in a second time interval, wherein the second time interval is greater than the first time interval.

13. Method according to claim 12, characterised in that the partial sum the amplitude spectra are formed from the amplitude spectra, which are temporarily stored in a third time interval which is greater than the first time interval and less than the second time interval, wherein the sum amplitude spectrum is formed from the partial sum amplitude spectra.

14. Apparatus for carrying out a method according to one of claims 1 to 13, the apparatus comprising at least one first sensor (1), at least one computing facility (18) and at least one facility for positioning, for determining the travel speed or for detecting the travel speed, characterised in that the at least one computing facility (18) for geometry determination (11) of at least one first wheel (2) is connected in a signal-transmitting manner to the at least first sensor (1) for detecting vertical accelerations and to the at least one facility for positioning, for determining the travel speed or for detecting the travel speed.

15. Rail vehicle with at least one apparatus according to claim 14.