DE102023211096B3 - Method and system for determining wear of a component of a machine for maintenance of the component - Google Patents

Abstract

A method for determining wear of a component (4) of a machine (2) for maintenance of the component (4) is described. The method comprises acquiring (S1) data relating to a vibration of the component (4) and determining (S2) audio data as a function of the acquired vibration data. The method further comprises outputting (S3) the audio data to a user at a location independent of the component (4) and reading in (S5) a user input relating to wear of the component (4) based on the user hearing (S4) the output audio data. A system is further described which is configured to carry out steps of the method.

Description

Technical area

The present invention relates to a method for determining wear of a component of a machine for maintenance of the component. Furthermore, the present invention relates to a system configured to carry out steps of such a method.

State of the art

During use, machines experience wear on mechanically moving parts. Therefore, such machines require maintenance. For example, wear on train wheels occurs, for example, as bumps on the wheel treads. During maintenance, the wheels are turned to eliminate these bumps. During turning, material is removed from the wheel, which is why this type of maintenance can only be performed at a limited frequency. Therefore, it is necessary to determine whether wheel turning is actually necessary as part of maintenance to ensure continued use of the wheel. It is known from the prior art that technicians ride along with the train during a test run and listen for certain noises from the wheels. Based on the noises, technicians can then determine the wheel wear. This can prevent unnecessary wheel turning, which can extend the service life of a wheel. However, maintenance can be performed if the technician has determined it is necessary based on the noise. Furthermore, the US7640139B2 Describes how a system and device are configured to diagnose such wear. A diagnostic processor is used to determine the wear of a machine. This is done based on vibrations and noises of the machine, which are processed by the processor.

Description of the invention

In a first aspect, the present invention relates to a method for determining wear of a component of a machine for servicing the component. Steps of the method can be carried out using a computing device. The machine can be, for example, a train or a wind turbine. The component can be, for example, a gearbox of a wind turbine or a wheel of a train. Wear of the component can, for example, be a flat spot on the wheel. During servicing of the component, an inspection of the component and, alternatively or additionally, a repair of the component can be carried out. For example, if the component is a wheel of a train, the wheel can be turned as part of the maintenance.

The method comprises acquiring data relating to a vibration of the component. Data relating to a vibration of the component can, for example, include data relating to structure-borne noise of the component. The acquiring data relating to a vibration of the component can, for example, be carried out using a suitable sensor. The acquiring of the vibration can be carried out during operation of the machine, for example while the train is traveling. The sensor can be configured to directly acquire the structure-borne noise of the component. Alternatively or additionally, the sensor can be configured to acquire the structure-borne noise of the component indirectly, for example by directly acquiring a further vibration induced by the structure-borne noise of the component.

The method further comprises determining audio data depending on the acquired vibration data. The audio data can be derived from the component vibration data. For example, the component vibration can generate airborne sound. For example, a wheel vibration can generate a noise in the train. When determining audio data depending on the acquired vibration data, the audio data can be determined such that it emulates and replicates the airborne sound generated by the component vibration. For example, the audio data can be determined such that it replicates the noise in the train.

The method further comprises outputting the audio data to a user at a location independent of the component. The user can be a technician. The output is performed using a loudspeaker. The loudspeaker is located at a different location than the component and can be spatially independent of the component. For example, the loudspeaker is stationary, while the machine, and thus the component, is mobile. The loudspeaker can be located, for example, at an office workstation of the technician. The machine, in the form of a train, can, however, be mobile. By outputting the audio data, a user can hear the audio data.

The method further comprises reading in a user input regarding the wear of the component based on the user listening to the output audio data. The user input can be independent of artificial intelligence. The input can, for example, only depend on the subjective hearing of the given audio data by the user. For example, the user can hear in the output audio data that a certain noise or tone occurs in the audio data. For example, a certain frequency with a certain amplitude, i.e. a certain sound pressure, can be contained in the audio data. The user's input regarding wear can then be carried out depending on the frequency and the specific amplitude relative to the frequency. For example, a certain degree of wear on a wheel has a certain frequency, and the user can recognize this when listening and thus enter it.

The method thus represents a condition monitoring system. The condition of a component, for example, a train wheel, can be monitored remotely by the user via audio data at a location independent of the component. Various types of damage to a wheel can be detected by the user via the audio data, for example, via different frequencies in the audio data. This means that the method for determining component wear can be carried out regardless of the location of the user and the component. It is not necessary for the user to be at the location of the component's vibration, for example, in the train, and to hear the component's vibration on site. A so-called listening test on a train test track can thus be dispensed with. Rather, the component's vibration can be generated during normal operation, and data on such vibrations can be recorded during normal operation. This recorded data can then be recorded as audio data regardless of location and output to the user. Alternatively or additionally, the user can also listen to the output audio data and enter an input at a later time after the vibration data has been recorded. During the input, the user can determine the level of wear. For example, if the user indicates that a certain level of wear is present on the component, maintenance can be performed on the component, for example, the wheel can be turned. Only if the user also indicates that a certain level of wear is present on the component can such a turning be performed. In other cases, if the user does not determine that a certain level of wear is present on the wheel based on listening to the audio data, turning can be omitted, thus saving on the wheel's material.

The method is further characterized in that data relating to a condition of the component is acquired using a sensor. The sensor can be a hardware sensor of the machine. The sensor for acquiring data relating to a condition of the component can be a different sensor or the same sensor as the sensor for acquiring data relating to a vibration of the component. The sensor for acquiring data relating to a condition can, for example, be an acceleration sensor configured to acquire data relating to an acceleration state of the component. Furthermore, information relating to a condition of the component is determined depending on the acquired data relating to the condition. For example, the information relating to the condition can be determined by a server. For this purpose, the acquired data relating to a condition of the component can be transmitted and sent via Bluetooth to a gateway of the machine and then from the gateway to the server. Determining the information relating to the condition of the component can be a damage prediction. This damage prediction can be carried out using an algorithm and, alternatively or additionally, using a trained neural network, for example, with a predetermined confidence level. Furthermore, the determined information relating to the condition is output to the user. For example, output can be performed via a display using a display device. This allows information about the condition, for example a specific damage status of the component, to be output or displayed to the user. Furthermore, the input is read in based on the output information about the condition. This allows a user to determine the wear of the component based on the output audio data and also based on the output specific information about the condition and to specify this in the form of the input during the read in. For example, data about the condition of the component can be recorded first, then the information about the condition of the component can be determined and thus a damage prediction can be carried out. This damage prediction can represent a preliminary damage prediction regarding wear of the component and be output to the user. In addition, vibration data can be recorded and audio data can be determined based on this. This is output to the user in addition to the specific information about the condition, for example the damage prediction. The user then decides whether the component is wearing or not based on the damage prediction and the audio data. This can be taken into account when reading in the user's input. When reading the user input, the determination of information about the state of the component can be improved.

The method thus enables a simple evaluation of the component's damage prediction. The damage prediction can be determined automatically using the algorithm or the neural network. This automatically determined damage prediction can then be evaluated manually and by the user by listening to the audio data and making the input. For example, the damage prediction may include that a wheel needs to be repaired, for example due to a flat spot. The user then listens to the audio data and decides that the damage prediction and thus the damage is not that serious. The user then decides, in the form of the input regarding the component's wear, that the wheel is not yet so worn that it needs to be repaired, for example by turning it down.

Alternatively, the damage prediction is confirmed during evaluation by the user based on the audio data and, for example, the wheel is repaired by turning it. The method thus provides an efficient and simple evaluation of the damage prediction. The component, for example a wheel, is only repaired when and exactly when it is necessary, i.e. when the user confirms based on the audio data that the damage prediction is correct and the component needs to be repaired. Evaluating the damage prediction of the component, for example by riding along and testing it on a train, is therefore no longer necessary. In addition to the damage prediction, the user can evaluate the wear of the component based on the automatically generated audio data and enter an input for reading.

According to a further embodiment, the method can be characterized in that a repeated determination of the information on the condition of the component can be carried out depending on at least one read-in input regarding the wear of the component. The repeated determination can, for example, be carried out after the user's input regarding the wear of the component has been read in, based on the user listening to the output audio data. Several steps of the repeated determination can be carried out after further steps of acquiring data on the vibration, determining the audio data, outputting the audio data to the user, and reading in an input from the user. As part of the repeated determination of the information regarding the condition of the component, a damage prediction of the component can be repeatedly determined. The repeated determination of the information regarding the condition of the component can, like the first determination of the information regarding a condition of the component, be carried out, for example, by a neural network. The neural network can be trained depending on at least one read-in input regarding the wear of the component. The neural network can therefore be trained based on the input read in regarding the wear of the component, or alternatively based on multiple inputs read in regarding the wear of the component. Training can be carried out iteratively. The neural network can therefore be trained differently for steps of determining the information on the condition of the component that are carried out at different times, i.e. for an initial determination and for a repeated determination. This allows the determination to be carried out differently by the neural network. For example, in a first, initial step of determining the information on the condition of the component in the form of a damage prediction, a damage condition on the wheel was determined, and based on the audio data as input from the user, no wear on the component was read in. The neural network can then be trained depending on the input. Thus, when the information on the condition is repeatedly determined based on identical data on the condition, no damage condition can be determined by the newly trained neural network.

This can improve damage prediction and the determination of information on the condition of the component. This can make damage prediction more precise, and user evaluation can then increasingly become a verification of the damage prediction. By iteratively determining information on the condition of the component, determining and outputting audio data, and reading in an input, a training data set for training the neural network can be updated and improved. The training data can, for example, be used in supervised learning to train the neural network and, for example, improve it over the various iteration steps. The aim is to achieve a state where the damage prediction for a particularly well-trained neural network corresponds almost or completely 100% to a user input on wear based on audio data. Alternatively or additionally, the determination of information on the condition of the component and the damage prediction can be adapted to a specific customer and, alternatively or additionally, to a specific component of the machine. The feedback from reading in the user input to the output audio data can thus be used to update the initially general and unspecific fish trained neural network can be trained more specifically, for example customer-specifically.

According to a further embodiment, the method can be characterized in that, when acquiring data relating to a vibration of the component, acceleration data of the component can be acquired using an acceleration sensor. This acceleration sensor can, but does not have to, be the same sensor as used for acquiring data relating to the condition of the component as described above. Therefore, the acceleration sensor used for acquiring acceleration data as described here does not have to be the same as that used for determining information relating to the condition of the component and thus for damage prediction. The acceleration sensor can, for example, be arranged on a bogie and alternatively or additionally on the wheel. Alternatively, the acceleration sensor can be arranged approximately half a meter or a meter away from the wheel or the bogie. Furthermore, the determination of audio data can be carried out depending on the acquired acceleration data. In the step of determining audio data, the acceleration data can be converted into audio data. Certain accelerations that have been measured in the form of acceleration data of the component can, for example, be converted into certain frequencies and amplitudes in the audio data. These accelerations of the component can induce noise, and the recorded audio data can emulate these noises. Before determining audio data based on the acquired acceleration data, this acceleration data can be post-processed, for example, digitally post-processed. During this post-processing, damage-related information can be identified from the recorded vibration data. This damage-related information can be converted into audio data, and alternatively or additionally, the audio data can be determined based on this damage-related information.

This allows data on the vibration of the component to be recorded via the acceleration sensor and then converted into audio data. This audio data can then be output to the user. This means that a recorded acceleration of the component can also be used to evaluate the damage prediction. The method can therefore avoid a classic audio test on a train, where vibrations of the component, triggered for example by a journey, can induce airborne sound and noise. Even during the audio test, a user can at least indirectly hear the acceleration of the component in the form of airborne sound induced by the acceleration of the component and thus determine wear on the component or evaluate a damage prediction. The method offers the option of doing this remotely.

According to a further embodiment, the method can be characterized in that, when data relating to a vibration of the component are acquired, acceleration data of the component can be acquired using an acceleration sensor. Refinements of the embodiment can be found in the corresponding embodiment described above. Furthermore, the method can be characterized in that information relating to a state of the component can be determined as a function of acquired data relating to the state of the component. Refinements of the embodiment can be found in the corresponding embodiment described above. Furthermore, the method can be characterized in that the acquisition of acceleration data is carried out as soon as the information relating to the state of the component has been determined. Thus, the acquisition of acceleration data can be carried out after the information relating to the state of the component has been determined for the first time, for example continuously and then triggered periodically.

This method thus continuously obtains acceleration data from the initial determination of information about the component's condition, allowing audio data to be determined for the user based on this information. After a damage prediction, audio data can be determined for any point in time to evaluate the damage prediction.

According to a further embodiment, the method can be characterized in that, when acquiring data relating to a vibration of the component, sound data of the component can be acquired using a microphone. The microphone can be arranged, for example, in the train, or alternatively outside the train. The microphone can be arranged at a location on the machine, such as the train, from which sound from the component can be clearly heard, for example due to damage to the component, and recorded with the microphone. The microphone can be static or mobile. Furthermore, the determination of audio data can be carried out depending on the acquired sound data. Before determining the audio data depending on the acquired sound data, the sound data can be post-processed, for example digitally post-processed, in order to The audio data can be determined based on the identified damage-related information. For example, the audio data can emulate the airborne sound that prevails at the location of the microphone due to the vibration of the component. Acquiring sound data can be performed independently of acquiring acceleration data. Determining audio data based on sound data can be performed independently of determining audio data based on acceleration data.

By capturing component vibration data using a microphone, the process of a conventional test drive can be emulated as accurately as possible. During a potential test drive, the user perceives noises by listening to the sound data and thus evaluates a damage prediction for the component or makes decisions about the extent to which the component is worn. By using the microphone to capture the sound data as part of the process, this process can be realistically emulated by the process. If the microphone is omitted, the audio data can be determined based on the acceleration data, for example, which can be more cost-effective due to lower development effort. For the same period, for example, both sound data and acceleration data can be captured, and different audio data can be determined depending on them. The user thus receives different audio data in order to determine the wear of a component as accurately as possible. In an alternative embodiment, the audio data is determined for a first period based on the captured sound data, and in a second period, which follows the first period, the determination is carried out based on the captured acceleration data. For example, it is not always necessary to record sound data.

According to a further embodiment, the method can be characterized in that, when data relating to a vibration of the component is acquired, sound data of the component can be acquired using a microphone. Embodiments of the embodiment can be found in the corresponding embodiment described above. Furthermore, the method can be characterized in that information relating to a condition of the component can be determined as a function of acquired data relating to the condition of the component. Embodiments of the embodiment can be found in the corresponding embodiment described above. Furthermore, the acquisition of sound data can be carried out if a damage condition has been determined as the condition of the component. For example, the acquisition of sound data can only be carried out if a damage condition has been determined as the condition of the component. For example, if a flat spot on the wheel has been determined as the damage condition within the framework of the damage prediction, the acquisition of sound data can be carried out. The microphone can then be switched on and used to acquire the sound data of the component.

This reduces the microphone's power consumption, providing a particularly efficient process.

According to a further embodiment, the method can be characterized in that the recorded data on the vibration of the component can be sent to a server. For example, the recorded data on the vibration of the component can be sent to the server either from the microphone and alternatively or additionally from the acceleration sensor via a gateway, such as a train gateway. Communication to the gateway can be via Bluetooth and from the gateway to the server via mobile radio. The determination of audio data can be carried out by the server. As part of the determination of the audio data, post-processing, for example with filters, of acceleration data and alternatively or additionally of sound data can also be carried out by the server. The server can then send the audio data to a loudspeaker for output, for example via mobile radio.

Thus, the computing power of a server, for example designed as a cloud, can be used to carry out computationally intensive steps of the process.

Computationally intensive steps can include, for example, determining the audio data and, in the process, post-processing the audio data. This can increase the efficiency of the method by eliminating the need for additional devices configured to perform other steps of the method to perform computationally intensive steps of the method.

According to a further embodiment, the method can be characterized in that, when outputting the audio data, the audio data can be uploaded to a dashboard. The dashboard can be configured for interaction with the user, such as the technician. Via the dashboard, the user can perform various Select audio data, for example, for listening. You can also upload it to the dashboard to output the audio data visually, for example, by graphically representing, outputting, and displaying the frequency and sound amplitude of the audio data.

This allows the user to listen to the output audio data and thus prepare the input by allowing them to perceive the audio data not only audibly but also visually. Furthermore, the user can select from a variety of different audio data, for example, several different audio data sets from different driving scenarios for a specific component. This facilitates evaluation and thus the determination of component wear.

A second aspect of the present invention relates to a system configured to carry out steps of the method according to an embodiment of the first aspect of the present invention. The system comprises a vibration detection device, such as a microphone or an acceleration sensor, which can be configured to collect data relating to a vibration of the component. The system further comprises a server, for example, configured as a cloud, a loudspeaker, and an input device. The system can comprise a sensor, such as an acceleration sensor, for collecting data relating to a condition of the component. The server can be configured to determine information relating to a condition of the component based on the collected data. The server can therefore perform a damage prediction for a component. The system can further comprise a gateway for communication between the sensor, a vibration detection device, and the server. The system can further comprise a dashboard for uploading the audio data.

Short description of the characters

• 1 shows schematically steps of a method for determining wear of a component of a machine for maintenance of the component.
• 2 shows schematically means of a system for carrying out steps of the schematically in 1 procedure shown.

Detailed description of embodiments

1 schematically shows steps of a method for determining wear of a component 4 of a machine 2 for maintenance of the component 4. 2 shows schematically a system which is set up to carry out steps of the schematically shown 1 to carry out the method shown. The system has component 4 of the machine 2 designed as a train. In this embodiment, component 4 is a wheel of the train. A sensor 6, which is designed as an acceleration sensor, is arranged close to component 4. Furthermore, a further acceleration sensor 7 and a microphone 8 are arranged on component 4. The acceleration sensors 6, 7 and the microphone 8 are communicatively connected to a gateway 10 via wired or wireless communication, in the present example via Bluetooth. This gateway 10 is communicatively connected via mobile radio to a server 12, which is not part of the machine 2. The server 12 is communicatively connected via mobile radio to a dashboard 14, a loudspeaker 16 and an input device 18, which are each communicatively connected to one another via a wired connection.

One of the vibration detection devices 7, 8 collects S1 data relating to a vibration of component 4. Vibration detection devices 7, 8 comprise acceleration sensor 7 and microphone 8. The collected data relating to the vibration of component 4 is then sent S9 to server 12. This is done via gateway 10, with the collected data relating to the vibration of component 4 being sent via Bluetooth to gateway 10 and then further via mobile radio to server 12. Server 12 then determines S2 audio data based on the collected vibration data. The determined audio data is then sent from server 12 to loudspeaker 16. Loudspeaker 16 outputs S3 the audio data to a user at a location independent of component 4. The loudspeaker 16 is thus spatially independent of the machine 2. When the audio data is output S3, the audio data is uploaded S3.1 to the dashboard 14. The user selects audio data on the dashboard 14 to listen to it. They also visually view the audio data, for example, in the form of specific frequencies and amplitudes. The user then listens S4 to the output audio data. The user then reads S5 the user's input regarding the wear of component 4 via the input device 18. The user indicates the extent to which the wheel, as component 4, is worn and needs to be repaired.

Before the acquisition S1, an acquisition S6 of data is carried out on a state of the component component 4 with the sensor 6. In the embodiment shown here, the sensor 6 and the acceleration sensor 7 are designed separately. In an alternative embodiment not shown in detail, the sensor 6 and the acceleration sensor 7 are designed as one and the same sensor. The sensor 6 in the embodiment shown is also designed as an acceleration sensor for recording acceleration data of the component 4. Furthermore, the recorded data on the state of the component 4 is sent from the sensor 6 via the gateway 10 to the server 12. The server 12 determines S7 information on a state of the component 4 depending on the recorded data on the state. Using a trained neural network, the server 12 thus determines a damage state of the component 4. This determines a damage prediction for the component 4. The determined information on the state is then output S8 to the user, here via the dashboard 14. The input is read S5 based on the output information on the state. Before the audio data is determined S2, a damage prediction is first performed using the trained neural network of server 12. This damage prediction indicates, for example, that component 4 is damaged. This information is output to the user. The audio data is then also output to the user. Based on these two output data, the user now evaluates the damage prediction using the audio data. Thus, a user can automatically evaluate certain damage predictions by server 12 and either correct or confirm them.

After at least a first reading S5 of the user's input, a repeated determination S7.0 of the information on the condition of component 4 is carried out. This repeated determination S7.0 is carried out as a function of at least one read input on the wear of component 4. Thus, the damage prediction is first evaluated at least once by the user. Based on this evaluated damage prediction, a training data set is then generated and the neural network is at least partially retrained with this generated training data. This newly trained neural network is then used to repeatedly determine S7.0 the information on the condition of component 4, i.e., for a new damage prediction of component 4. Thus, the determination of the information on the condition of component 4 is iteratively improved for an ever-improving damage prediction of component 4.

When data relating to a vibration of component 4 is acquired S1, acceleration data of component 4 is acquired S1.1 using acceleration sensor 7. This acquired acceleration data is sent via Bluetooth to gateway 10 and then on to server 12 via mobile radio. The determination S2 of audio data by server 12 is carried out depending on the acquired acceleration data. In this process, damage-specific accelerations are identified, and the audio data is determined based on this. The acquisition S1.1 of acceleration data is carried out as soon as the determination S7 of the information relating to the condition of component 4 has been carried out. Thus, as soon as an initial damage prediction has been determined, the acquisition S1.1 of the acceleration data of component 4 is carried out.

In the embodiment shown here, the machine 2 has the microphone 8 in addition to the acceleration sensor 7. When data relating to a vibration of the component 4 is acquired S1, a capture S1.2 of sound data from the component 4 is carried out using the microphone 8. In one embodiment, this is carried out in parallel and independently of the capture S1.1 of acceleration data from the component 4 using the acceleration sensor 7. In an embodiment not shown here, the steps of capturing S1.1 and S1.2 of acceleration sound data are carried out independently of one another. This acquired sound data is sent via Bluetooth to the gateway 10 and then via mobile radio to the server 12. In the server 12, the determination S2 of audio data is carried out depending on the acquired sound data. In this process, damage-specific information is filtered from the sound data, and the audio data is determined based on this damage-specific information. The capture S1.2 of sound data is carried out when a damage state has been determined as the state of the component 4. Thus, as part of the damage prediction, a damage state of component 4 is determined, and only if a damage state has been determined for component 4 as part of the damage prediction is the acquisition S1.2 of sound data carried out with microphone 8. This reduces the power consumption of microphone 8. Only if the damage prediction indicates that an evaluation of a potential damage state is necessary is microphone 8 switched on to acquire S1.2 of sound data.

Reference symbol

2

machine

4

component

6

sensor

7

Accelerometer

8

microphone

10

Gateway

12

server

14

Dashboard

16

Speaker

18

Input device

S1

Collecting data on a component vibration

S1.1

Collecting acceleration data of the component

S1.2

Collecting sound data from the component

S2

Determining audio data

S3

Outputting the audio data

S3.1

Uploading the audio data

S4

Listening to the output audio data

S5

Reading a user input

S6

Collecting data about a component's state

S7

Determining information about a state of the component

S7.0

Repeatedly determining the information about the state of the component

S8

Output the specific information about the state

S9

Sending the recorded component vibration data to the server

Claims (10)

Method for determining wear of a component (4) of a machine (2) for maintenance of the component (4), comprising the steps of: detecting (S1) data relating to a vibration of the component (4); determining (S2) audio data as a function of the detected vibration data; outputting (S3) the audio data to a user at a location independent of the component (4) by means of a loudspeaker (16) arranged at a different location than the component (4); and reading in (S5) an input from the user regarding the wear of the component (4) based on the user listening (S4) to the output audio data, characterized in that a detection (S6) of data relating to a state of the component (4) is carried out using a sensor (6), that a determination (S7) of information relating to a state of the component (4) is carried out as a function of the detected data relating to the state, that an output (S8) of the determined information relating to the state is carried out to the user, and that the reading in (S5) of the input is carried out based on the output information relating to the state. Method according to the preceding claim, characterized in that a repeated determination (S7.0) of the information on the state of the component (4) is carried out as a function of at least one read-in input on the wear of the component (4). Method according to one of the preceding claims, characterized in that when detecting (S1) data relating to a vibration of the component (4), a detection (S1.1) of acceleration data of the component (4) is carried out using an acceleration sensor (7), and that the determination (S2) of audio data is carried out as a function of the detected acceleration data. Method according to one of the preceding claims, characterized in that when detecting (S1) data relating to a vibration of the component (4), a detection (S1.1) of acceleration data of the component (4) is carried out using an acceleration sensor (7), that a determination (S7) of information relating to a state of the component (4) is carried out as a function of detected data relating to the state of the component (4), and that the detection (S1.1) of acceleration data is carried out as soon as the determination (S7) of the information relating to the state of the component has been carried out. Method according to one of the preceding claims, characterized in that when detecting (S1) data relating to a vibration of the component (4), a detection (S1.2) of sound data of the component (4) is carried out using a microphone (8), and that the determination (S2) of audio data is carried out as a function of the detected sound data. Method according to one of the preceding claims, characterized in that when data relating to a vibration of the component (4) are recorded (S1.2), sound data of the component (4) is recorded using a microphone (8), that information relating to a state of the component (4) is determined (S7) as a function of recorded data relating to the state of the component (4), and that the recording (S1.2) of sound data is carried out when a damage state has been determined as the state of the component (4). Method according to one of the preceding claims, characterized in that a transmission (S9) of the acquired data on the vibration of the component (4) is carried out to a server (12), and that the determination (S2) of audio data is carried out by the server (12). Method according to one of the preceding claims, characterized in that when outputting (S3) the audio data, an upload (S3.1) of the audio data to a dashboard (14) is carried out. Method according to one of the preceding claims, characterized in that when reading in (S5) the user's input, the determination of the information (S7; S7.0) on a state of the component is improved. System comprising a vibration detection device (7; 8), a server (12), a loudspeaker (16) and an input device (18), the system being arranged to carry out steps of the method according to one of the preceding claims.

Images