Method for damping vibrations, and torque assembly

Abstract

The invention relates to a method for damping (10) vibrations (74) of a torque assembly (14) of a vehicle (24), having the steps of providing (12) the torque assembly (14) that has at least one torque component (18), which transmits a drive torque (16) in order to move the vehicle (24) and rotates at a rotational speed, detecting (40) at least one vibration parameter (42) of a present or immediately upcoming undesired vibration (74) which relates to the torque assembly (14), and carrying out (60) a damping measure (62) in order to damp the vibration (74) on the basis of the vibration parameter (42), wherein the at least one vibration parameter (42) is calculated by a trained neural network (44) using analysis information (46) assigned to the vibration (74) as input data (48). The invention additionally relates to a torque assembly (14).

Description

[0001] Vibration damping method and torque assembly

[0002] Description introduction

[0003] The invention relates to a method for vibration damping according to the preamble of claim 1. Furthermore, the invention relates to a torque assembly.

[0004] DE 102004 032 150 A1 describes a method for reducing clutch judder in a vehicle's drivetrain. The judder is reduced by generating a superimposed torque and adding it to the clutch's transmission torque. The superimposed torque is calculated based on a filtered vehicle acceleration curve.

[0005] The object of the present invention is to increase the speed and accuracy of vibration damping.

[0006] At least one of these objects is achieved by a method for vibration damping having the features of claim 1. This allows the vibration to be damped more quickly and precisely. The vehicle can be operated more comfortably and safely.

[0007] The vehicle can be a motor vehicle, a goods vehicle or a two-wheeled vehicle.

[0008] The torque assembly can be arranged in a vehicle's drivetrain. The torque assembly can be an electric axle (E-axle). The torque assembly can include an electric motor, an inverter, and / or a transmission. The transmission can be single- or multi-speed.

[0009] The drive torque can be provided by a drive element, in particular an electric motor. The torque assembly can comprise the drive element. The electric motor can have a stator and a rotor rotatable relative to the stator. The torque component can be a rotor shaft connected to the rotor or a component connected to the rotor shaft in a torque-transmitting manner.

[0010] The vibration may be undesirable as it may affect the safety, comfort and / or reliability of the torque assembly and / or the vehicle.

[0011] The vibration can be a torsional, rotary, and / or linear vibration. The vibration can originate from the torque assembly via the vehicle wheel. The vibration can originate from a road surface irregularity, such as a pothole or bump. The vibration can originate from the torque assembly's mount or suspension. The vibration can originate from the rotational movement used to transmit the drive torque. The torsional vibration can be triggered by a change in the drive torque.

[0012] The evaluation information can comprise time series data. The evaluation information can comprise a time profile of a rotational speed, for example a rotor speed of the rotor or a wheel speed of at least one vehicle wheel of the vehicle, an acceleration, an angular acceleration, and / or a torque, in particular the drive torque. A Fourier transformation and / or wavelet transformation for calculating the vibration parameter based on the evaluation information can be omitted. The evaluation information can indicate the upcoming vibration predictively or causally. The evaluation information can comprise information on road surface irregularities, for example the geopositions of the road surface irregularities and / or a geolocal vehicle position of the vehicle. The evaluation information can be measured. The evaluation information can be information derived from measured values.The evaluation information can be unprocessed or processed sensor data.

[0013] The vibration parameter can include or be a slope, frequency, amplitude, and / or phase in a time signal. The vibration parameter can include or be a vibration indicator that indicates the presence of the disturbing vibration. The vibration parameter can be detected predictively and even before the vibration occurs.

[0014] The neural network can be trained using training data. The training data can include a plurality of assignments between evaluation information and respective vibration parameters. The training data can include labeled data. The training data can contain time series data and marked regions exhibiting disturbing vibrations, optionally including associated vibration parameters. The training data can include practical data, in particular recorded data from at least one ferry operation of the vehicle or comparable vehicles, and / or simulated data.

[0015] The neural network can be tested and / or validated. Validation can be performed using validation data. Testing can be performed using test data. The test data and / or validation data can include a plurality of assignments between evaluation information and respective vibration parameters.

[0016] The neural network can be a convolutional neural network (CNN). A CNN typically consists of multiple convolutional layers that automatically and adaptively extract features from the input data. Each convolutional layer applies a set of filters that slide over the input data and capture local features. Following convolutional layers, the CNN can have downstream pooling layers that reduce the spatial size of the representation and help mitigate overfitting, as well as one or more fully connected layers responsible for classification or regression based on the extracted features.

[0017] The neural network can be a recurrent neural network (RNN). An RNN typically consists of at least one input layer and at least one output layer, as well as recurrent layers that process the sequential data. The RNN can comprise a Long Short Term Memory (LSTM) and / or Gated Recurrent Unit (GRU) architecture. Furthermore, one or more fully connected layers can be present, which are responsible for classification or regression based on the extracted features.

[0018] The neural network can comprise an encoder-decoder architecture, particularly an autoencoder or a transformer incorporating attention mechanisms. The encoder is responsible for processing the input data and converting it into a compact, internal representational state. This state summarizes the essential information of the input data. The encoder can contain recurrent neurons (as in an RNN, LSTM, or GRU) or convolutional neurons (as in a CNN). The decoder takes the internal state generated by the encoder and uses it to generate the output data, in this case, at least one oscillation parameter. The decoder can also have recurrent or convolutional layers.

[0019] The neural network may be a deep neural network. The neural network may have a plurality of intermediate layers with a plurality of neurons between at least one input layer and one output layer.

[0020] The damping measure can be implemented immediately upon availability of the calculated vibration parameter. This allows for a rapid response to the unwanted vibration. In a preferred embodiment of the invention, it is advantageous if the vibration parameter is calculated by the neural network based on evaluation information covering a maximum of one oscillation period. This allows the vibration parameter to be calculated more quickly, and vibration damping can be implemented more quickly.

[0021] In a preferred embodiment of the invention, the evaluation information comprises at most a fraction of the oscillation period. The evaluation information can comprise at most the first half, in particular the first third, of the oscillation period.

[0022] In a specific embodiment of the invention, it is advantageous if the oscillation period is the initial first oscillation period of the oscillation. This allows for a rapid response to the onset of oscillation.

[0023] In a preferred embodiment of the invention, the damping measure comprises applying a control torque to the torque component. The control torque can be applied by the drive element, preferably the electric motor, which also provides the drive torque.

[0024] In a specific embodiment of the invention, it is advantageous if the control torque is superimposed on the drive torque. The control torque can be introduced in addition to the drive torque. The control torque can be introduced as an addition to the drive torque.

[0025] In a specific embodiment of the invention, it is advantageous if the damping measure is implemented as a function of at least one damping variable calculated by the neural network or another trained neural network using the at least one vibration parameter as input data. The description given with regard to the features of the neural network can also apply to the additional neural network. If the damping variable is calculated using the neural network, which includes the evaluation information as input data, the damping variable can be calculated internally as a function of the vibration parameter. The damping measure can be implemented as an indirect function of the vibration parameter and as a direct function of the damping variable.

[0026] In a preferred embodiment of the invention, it is advantageous if the damping variable includes preset values ​​for the control torque. The preset values ​​can be dynamic. This allows for adaptive vibration damping. In an advantageous embodiment of the invention, the calculation of the damping variable includes at least one operating parameter of the vehicle and / or one driving parameter of the vehicle. The operating parameter and / or the driving parameter can be assigned to the input data for the neural network or the further neural network for calculating the damping variable.

[0027] The operating parameter can be a current drive torque.

[0028] The driving parameter can be the vehicle's speed and / or position. The vehicle's position can be calculated using navigation data.

[0029] Furthermore, within the scope of the invention, a torque assembly having the features of claim 10 is proposed to solve at least one of the above-mentioned problems.

[0030] Further advantages and advantageous embodiments of the invention emerge from the description of the figures and the illustrations.

[0031] Character description

[0032] The invention is described in detail below with reference to the figures. They show in detail:

[0033] Figure 1: A method for vibration damping in a specific embodiment of the invention.

[0034] Figure 2: Evaluation information as input data when carrying out a method for vibration damping in a special embodiment of the invention.

[0035] Figure 3: A side view of a torque assembly in a specific embodiment of the invention.

[0036] Figure 4: A spatial view of a torque assembly in another specific embodiment of the invention.

[0037] Figure 1 shows a method for vibration damping in a specific embodiment of the invention. The method for vibration damping 10 of a torque assembly of a vehicle comprises providing 12 the torque assembly 14, which has a torque component 18 that transmits a drive torque 16 for propulsion of the vehicle and rotates at a speed. The torque assembly 14 is, for example, an electric axle 20 of a drive train 22 of the vehicle 24. The torque assembly 14 comprises an electric motor 26, an inverter 28 that electrically operates the electric motor, and a transmission 30. The transmission 30 is connected to a vehicle axle 32 in a torque-transmitting manner. The vehicle axle 32 has at least two vehicle wheels 34. The electric motor 26 comprises a rotor 36, and the torque component 18 can be a rotor shaft 38 of the rotor 36, which is connected to the transmission 30 in a torque-transmitting manner.

[0038] At least one vibration parameter 42 of an existing or imminent undesired vibration affecting the torque assembly 14 is detected 40 by a trained neural network 44 using evaluation information 46 associated with the vibration as input data 48. The vibration can be a torsional vibration of at least the torque component 18 and can act on the torque component 18, for example, due to road surface irregularities via the vehicle wheels 34, the vehicle axle 32, and the transmission 30. The torsional vibration can lead to undesired noises and vibrations of the torque assembly 14.

[0039] The vibration parameter 42 can be a vibration indicator 50 that indicates the presence of the vibration, or more specifically, a parameter, in particular the frequency of the vibration. The evaluation information 46 can be time series data 52, for example, a rotor speed 54 of the rotor 36. This allows an existing vibration to be detected and evaluated. The vibration parameter 42 can alternatively be calculated with the neural network 44 depending on a vehicle position 56 with previously geolocated, stored road surface irregularities 58 as evaluation information 46. This allows the vibration parameter 42 of an imminent, i.e., not yet initiated, vibration to be calculated.

[0040] A subsequent execution 60 of a damping measure 62 for damping the vibration depends on the vibration parameter 42. The damping measure 62 preferably comprises initiating a control torque 64 superimposed on the drive torque 16 via the electric motor 26. The damping measure 62 is executed depending on at least one damping variable 68 calculated by the neural network 44 or another trained neural network 66 using the vibration parameter 42 as input data, and thus indirectly depending on the vibration parameter 42. The damping variable 68 can include default values ​​for the control torque 64. The calculation of the damping variable 68 can further include an operating parameter 70 of the vehicle 24 and / or a driving parameter 72 of the vehicle 24 as input data. Figure 2 shows evaluation information as input data when carrying out a method for vibration damping in a specific embodiment of the invention.The evaluation information 46 is, for example, time series data of a value A of the rotor speed 54 of the rotor as a function of time t. The rotor speed 54 may begin to oscillate at time t0 upon impact on the vehicle wheel, for example, due to a bump in the road surface. Preferably, the oscillation parameter is calculated at the initial occurrence of the oscillation 74, in particular at time t1, and the damping measure 62 is initiated. This allows the oscillation 74 to be damped, and a rotor speed 76 with damped oscillation can be set, compared to the lack of oscillation damping, which causes the rotor speed 54 to oscillate.

[0041] The vibration parameter can be calculated as evaluation information 46 depending on rotor speed information comprising a maximum of a first vibration period 78 of the vibration 74, preferably a first third of the vibration period 78.

[0042] Figure 3 shows a torque assembly in a specific embodiment of the invention. The torque assembly 14 is, for example, an electric axle 20 that provides drive torque to a vehicle axle 32 of the vehicle 24. The vehicle axle 32 is mounted on a vehicle frame 82 via a chassis suspension 80. The torque assembly 14 has a center of gravity 84 that is offset from the vehicle axle 32. A force 86 acting on the vehicle axle 32 due to uneven road surfaces can cause a vibration 74 of the received torque assembly 14 due to the offset center of gravity 84 of the torque assembly 14.

[0043] Figure 4 shows a perspective view of a torque assembly in another specific embodiment of the invention. The torque assembly 14 is designed as an electric axle 20 and is mounted on a vehicle axle 32 of the vehicle. The torque assembly 14 includes an electric motor 26, an inverter 28, and a transmission 30. List of Reference Symbols

[0044] 10 methods for vibration damping

[0045] 12 Provision

[0046] 14 Torque assembly

[0047] 16 Drive torque

[0048] 18 Torque component

[0049] 20 electric axles

[0050] 22 Drivetrain

[0051] 24 vehicles

[0052] 26 electric motor

[0053] 28 inverters

[0054] 30 gearboxes

[0055] 32 vehicle axles

[0056] 34 vehicle wheel

[0057] 36 Rotor

[0058] 38 Rotor shaft

[0059] 40 Capture

[0060] 42 vibration parameters

[0061] 44 neural network

[0062] 46 Evaluation information

[0063] 48 input data

[0064] 50 vibration indicator

[0065] 52 time series data

[0066] 54 rotor speed

[0067] 56 Vehicle position

[0068] 58 Road unevenness

[0069] 60 Execute

[0070] 62 Damping measure control torque neural network

[0071] Damping size

[0072] Operating parameters

[0073] Driving parameters

[0074] vibration

[0075] Rotor speed

[0076] Oscillation period

[0077] chassis suspension

[0078] vehicle frame

[0079] Focus

[0080] force application

Claims

Patent claims 1. A method for vibration damping (10) of vibrations (74) of a torque assembly (14) of a vehicle (24), comprising the steps of providing (12) the torque assembly (14) which has at least one torque component (18) which transmits a drive torque (16) for moving the vehicle (24) and rotates at a speed, Detecting (40) at least one vibration parameter (42) of an existing or imminent undesired vibration (74) affecting the torque assembly (14), carrying out (60) a damping measure (62) for damping the vibration (74) depending on the vibration parameter (42), characterized in that the at least one vibration parameter (42) is calculated by a trained neural network (44) with evaluation information (46) associated with the vibration (74) as input data (48).

2. Method for vibration damping (10) according to claim 1, characterized in that the vibration parameter (42) is calculated by the neural network (44) as a function of evaluation information (46) comprising a maximum of one vibration period (78) of the vibration (74).

3. Method for vibration damping (10) according to claim 2, characterized in that the evaluation information (46) comprises at most a fraction of the vibration period (78).

4. A method for vibration damping (10) according to claim 2 or 3, characterized in that the oscillation period (78) is the initial first oscillation period (78) of the oscillation (74).

5. Method for vibration damping (10) according to one of the preceding claims, characterized in that the damping measure (62) comprises introducing a control torque (64) onto the torque component (18).

6. Method for vibration damping (10) according to claim 5, characterized in that the control torque (64) is superimposed on the drive torque (16).

7. Method for vibration damping (10) according to one of the preceding claims, characterized in that the damping measure (62) is carried out as a function of at least one damping variable (68) calculated by the neural network (44) or a further trained neural network (66) with the at least one vibration parameter (42) as input data (48).

8. Method for vibration damping (10) according to claim 5 or 6 and claim 7, characterized in that the damping variable (68) comprises default values for the control torque (64).

9. Method for vibration damping (10) according to claim 7 or 8, characterized in that the calculation of the damping variable (68) includes at least one operating parameter (70) of the vehicle (24) and / or a driving parameter (72) of the vehicle (24).

10. A torque assembly (14) for a vehicle (24), comprising at least one torque component (18) transmitting a drive torque (16) for moving the vehicle (24) and rotating at a speed, which may exhibit undesirable vibrations (74), and which is configured to be operated by a vibration damping method (10) according to any one of the preceding claims.

Images