ACOUSTIC ANALYSIS OF A MACHINE'S CONDITION

Description

[0001] The present invention relates to a method for generating at least one reference signal for the acoustic analysis of a machine's state, a method for the acoustic analysis of a machine's state, wherein an acoustic operating signal is recorded during operation of the machine, a corresponding analysis device, a computer program, and a computer-readable storage medium.

[0002] For example, in industrial manufacturing, data analysis methods are becoming increasingly widespread, which aim, among other things, at predictive planning of maintenance work on machines in order to minimize downtime or quality losses in the manufacturing process ("predictive maintenance").

[0003] In established methods, process-acquired state data from a machine is analyzed using machine learning techniques to derive meaningful trends. This may require large amounts of data to make reliable predictions, and consequently, correspondingly high computing power.

[0004] From BE 1 018 513 A3, a working method for the early detection of damage to a machine by means of an acoustic signature or several signatures recorded by means of several microphones is already known.

[0005] Further disadvantages of known methods include their sometimes large instrumentation costs, a limited scope of application, or a lack of automation.

[0006] Against this background, it is an object of the present invention to provide an improved concept for the acoustic analysis of a machine's state, which can reduce the amount of data to be processed and the computational effort required for processing.

[0007] According to the invention, this problem is solved by the subject matter of the independent claims. Advantageous further developments and embodiments are the subject matter of the dependent claims.

[0008] The improved concept is based on the idea of ​​providing individual acoustic signals for individual components of the machine and generating a reference signal based on these.

[0009] According to a first independent aspect of the improved concept, a method for generating at least one reference signal for the acoustic analysis of a machine's state is presented. For each of at least two machine components, at least one individual acoustic signal is provided. A computing unit determines an expected overall acoustic signal based on these individual signals. The computing unit then generates the reference signal based on this overall signal.

[0010] The acoustic analysis of a machine can be understood, in particular, as an analysis of the machine's operating noise. Specifically, acoustic analysis includes the analysis of airborne sound waves generated by the machine's operation.

[0011] The computing unit can contain one or more subunits or processor units, which may, for example, be spatially distributed. In particular, different process steps can be carried out by different subunits of the computing unit.

[0012] One component of the machine is a noise-emitting component. This component may, for example, include a motor, a gearbox, and / or other electrical components or mechanically moving parts.

[0013] Providing the individual acoustic signals involves, in particular, generating the respective individual acoustic signals. This generation can be achieved by capturing the individual operating noises of the corresponding components using a transducer. Alternatively, supplementarily, or additionally, generating the individual signals can involve a simulation, i.e., a synthetic generation, of the individual acoustic signals.

[0014] In this and subsequent text, a sound transducer can be understood as a sound sensor or a sound receiver, i.e., a sensor or receiving unit designed to convert airborne sound into electrical signals. In particular, a sound transducer can be implemented as a microphone.

[0015] The sound transducer or the computing unit is part of an analysis device for the acoustic analysis of a machine's condition.

[0016] For each of the at least two components of the machine, at least one associated individual acoustic signal is generated depending on the respective individual operating noises of the corresponding component, in particular by means of a sound transducer.

[0017] The individual acoustic signals are generated based on the individual operating noises of the respective components in a reference or new state of the machine.

[0018] In particular, an associated acoustic signal is provided or generated for each component of the machine and for each individual operating state of the component, wherein each individual operating state of a component is selected from a predetermined set of individual operating states of the component.

[0019] The individual acoustic signals for a given component of at least two components are, in particular, independent of other components of the at least two components. This means that other components of the at least two do not contribute to the respective individual operating noise of the given component. In other words, the individual operating noises of the individual components are recorded separately, especially while the other components of the at least two are out of operation or acoustically isolated from the given component.

[0020] Here and in the following, recording or capturing a noise, in particular an operating or individual operating noise, can be understood as capturing the respective operating noise, in particular by means of at least one sound transducer, and the subsequent storage of the signal thereby generated on a computer-readable storage medium, for example the processing unit.

[0021] The individual signals can be provided, in particular, as time-dependent signals or as signals in a frequency domain. Accordingly, the overall signal and / or the reference signal can each be generated as a time-dependent signal or as a signal in the frequency domain.

[0022] By generating or providing one or more individual, associated acoustic signals for each of the at least two components, and using these signals to generate the overall and reference signals, the individual components of the machine are considered independently of one another in order to generate the reference signal. This allows for the targeted consideration of additional influences that contribute to the overall signal or the reference signal, such as environmental or interference influences, for example, based on empirical data, simulation data, or theoretical or partially theoretical models.

[0023] Because the individual acoustic signals are known, which can be understood, for example, as nominal signals or nominal noises of the individual components in the reference or new state of the machine, the subsequent analysis of the machine's condition with reference to the reference signal only requires recording the overall operating noise of the entire machine. In particular, it is not necessary to record the individual operating noises of the individual components as part of the analysis of the machine's condition, since the influences of the individual components on the overall noise or total operating noise of the machine can be separated or isolated. Accordingly, a recording of the operating noise from a relatively distant component can be made for the analysis of the machine's condition; a recording close to the components is not required.

[0024] This offers an advantage over other acoustic analysis methods: the improved concept requires less instrumentation, particularly regarding the number and positioning of one or more transducers. Specifically, according to the improved concept, a single transducer may suffice to record the machine's operating noise for condition analysis.

[0025] According to at least one embodiment, a first individual signal is recorded for each of the at least two components while the respective component is in its corresponding first individual operating state. The total signal is determined by the processing unit depending on the first individual signals. A first operating state of the machine is defined by a combination, particularly a predefined combination, of the first individual operating states of the at least two components, and in particular of all of the at least two components.

[0026] In other words, a first individual state is defined for each of the at least two components. The combination of all first individual states defines or constitutes the first operating state of the machine. In particular, every possible operating state of the machine can be defined as a predefined combination of corresponding individual operating states of the at least two components of the machine.

[0027] An individual operating state of one of at least two components can, for example, include one or more predefined operating parameters for the component and / or one or more environmental characteristics for the component. Depending on the type of machine and component, the operating parameters can include, for example, rotational speed, gear ratio, feed rate, load, or the like. The environmental characteristics of the component can include, for example, temperature or other environmental conditions in the component's environment.

[0028] The overall signal is determined, for example, depending on the first individual signals, by using those of the first individual signals to generate the overall signal that correspond to the respective individual operating states of the components which together constitute the first operating state of the machine.

[0029] By adjusting or varying the individual operating states or the overall operating state of the machine, a model for generating the overall signal can be improved. This improves the accuracy of the reference signal and thus the analysis of the machine's condition.

[0030] According to at least one embodiment, a second operating state of the machine is defined by a combination of the respective second individual operating states of the at least two components. A second individual signal is recorded for each of the at least two components while the respective component is in the corresponding second individual operating state. The overall signal is determined based on the second individual signals.

[0031] In particular, the second operating state of the machine differs from the first operating state. Accordingly, at least one of the second individual operating states of the at least two components differs from the corresponding first individual operating state of the respective component. It is possible, but not mandatory, that all first individual operating states differ from the corresponding second individual operating states.

[0032] According to at least one embodiment, the reference signal is equal to the total signal.

[0033] According to at least one embodiment, the reference signal is generated by determining several total signals for different operating states of the machine, as described above, and by generating the reference signal depending on the several total signals, for example by interpolation.

[0034] This eliminates the need to determine a separate overall acoustic signal for each relevant operating state of the machine that may require acoustic analysis by combining separately recorded individual signals. Instead, it suffices to select representative operating states of the machine as reference points and generate the overall signal for them as described. The reference signal for a representative operating state then corresponds to the respective overall signal, while the reference signal for an operating state not among the representative states is generated based on the overall signals of two or more representative operating states, for example, by interpolation or combination.

[0035] According to at least one embodiment, the overall signal is generated based on a model that depends on the individual signals. In the simplest case, the model can consist of adding the individual signals or adding them with weights, where the model includes the corresponding weighting factors. In more complex embodiments, the model can take into account additional influences besides the individual signals, such as background noise or other disturbances, the geometric design of the machine, the machine's environmental conditions, and / or material-specific noise signatures of the machine, in order to generate the overall signal.

[0036] According to at least one embodiment, the overall acoustic signal is determined by the computing unit depending on a pose of one of the at least two components or depending on two or more poses of two or more of the at least two components.

[0037] The pose of a component, as used here and in the following, is understood to be a combination of the position and orientation of the respective component in three-dimensional space. The position and orientation of the component can be defined relative to one or more of the at least two other components, relative to another part of the machine, or relative to an object in the machine's environment, such as a building floor, a building wall, or another machine.

[0038] The position of the components can influence how the individual signals are weighted to form the expected overall signal. Furthermore, the influence of external factors can also depend on the position. Therefore, considering the position improves the accuracy of the reference signal and ultimately the analysis of the machine's condition.

[0039] According to at least one embodiment, the pose of one of the at least two components, or the poses of all at least two components, are stored on a storage medium, for example, the processing unit. The processing unit can access the stored pose or poses to generate the overall signal.

[0040] According to at least one embodiment, the overall acoustic signal is determined by the computing unit depending on a nominal environmental condition of the machine.

[0041] The nominal environmental condition is not a measured value, but rather one or more model parameters used to generate or determine the overall signal from the individual signals. Depending on the environmental condition, different individual signals may be reflected differently in the overall signal, and / or the external influences on the overall signal may depend on the environmental condition.

[0042] The nominal environmental condition may, for example, be the temperature or humidity in the vicinity of the machine, background or disturbing noise, the size or nature of the room in which the machine is located, and / or the type and positioning or orientation of other objects or machines in the room.

[0043] By taking the nominal environmental conditions into account in the model for determining the overall signal, the accuracy in generating the overall signal can be improved.

[0044] According to at least one embodiment, the current environmental condition of the machine is determined, and in particular measured, by means of a sensor device, in particular the analysis device. The reference signal is determined by the processing unit depending on the current environmental condition, in particular depending on a difference value between the current environmental condition and the nominal environmental condition.

[0045] By taking into account the actual environmental conditions or the deviation of the actual environmental conditions from the nominal environmental conditions, the determination of the overall signal can be further improved.

[0046] According to a further independent aspect of the improved concept, a method for the acoustic analysis of a machine's state is described, wherein an acoustic operating signal, in particular dependent on the machine's current operating noise, is recorded during machine operation, especially by means of at least one transducer. A reference signal is provided by means of a method for generating at least one reference signal for the acoustic analysis of a machine's state according to the improved concept. The machine's state is analyzed based on a deviation of the operating signal from the reference signal.

[0047] The reference signal is provided in particular by generating it using a method for generating a reference signal according to the improved concept.

[0048] As explained above, a single transducer or a few transducers may be sufficient to record the operating signal, and yet a component-specific analysis of the machine's condition can be performed.

[0049] Compared to analysis methods based on structure-borne sound measurements or thermographic measurements, the instrumentation effort is reduced with a method based on the improved concept.

[0050] Furthermore, the improved concept allows the condition analysis to be carried out automatically, as no manual investigations are necessary, such as investigations of wear residues in machine oil, as is the case in ferrographic methods.

[0051] Another advantage of the improved concept is the universal applicability of the methods to almost any machine with noise-emitting components.

[0052] By analyzing the data using the improved concept, particularly with a reference signal generated according to this concept, the amount of data required for analysis is relatively small. Therefore, complex data analysis methods, such as those used in deep learning or machine learning, can be avoided. This also reduces the computational effort required for the analysis. Consequently, it is not necessary to process the data to be analyzed on central backend or cloud servers. Instead, the analysis can be performed on-premises.

[0053] According to at least one embodiment of the method for acoustically analyzing the state of a machine according to the improved concept, the current operating state of the machine is determined by means of the computing unit. The reference signal is provided by the computing unit depending on the current operating state.

[0054] To determine the current operating state, current operating parameters can be read from a machine control unit, particularly from a control program for the machine, using the computing unit. For this purpose, the computing unit can be coupled to the machine control unit.

[0055] The machine's current operating state, for example, is deterministically determined by the machine's control program, particularly at any given time. This allows the expected nominal acoustic behavior of the machine to be predicted within a very narrow timeframe by providing the appropriate reference signal.

[0056] To provide the reference signal depending on the current operating state, a composite signal generated for a representative operating state can be selected as the reference signal, or the reference signal can be generated based on composite signals from two or more representative operating states, as described above. The respective composite signals can be generated after the current operating state has been determined using the model, or they can be generated and stored in advance.

[0057] According to at least one embodiment, a differential signal is generated by the processing unit based on the deviation of the operating signal from the reference signal. The state is then analyzed by the processing unit based on this differential signal.

[0058] This can reduce the influence of noise as well as the complexity of the evaluation.

[0059] According to at least one embodiment, a frequency signal is generated by the processing unit based on the difference signal. The state is then analyzed by the processing unit based on this frequency signal.

[0060] To generate the frequency signal, the difference signal is subjected to a Fourier transform.

[0061] In such embodiments, the difference signal, and in particular the reference signal and the operating signal, are present as time-dependent signals.

[0062] In the frequency domain, noise components can be filtered out or removed, for example. This allows for a clearer identification of anomalous signal signatures or signal signatures resulting from system wear and tear.

[0063] According to at least one embodiment, the computing unit uses an analysis of the state to check whether an anomaly exists in the operation of the machine.

[0064] An anomaly is, in particular, a deviation of the current operating state of the machine or the current operating states of individual components from expected states. The anomaly can, for example, be caused by a fault or defect.

[0065] According to at least one embodiment, the computing unit uses analysis to identify which of the at least two components exhibits the anomaly.

[0066] This is possible in particular due to the potential separation of the individual noise signatures of the components.

[0067] According to at least one embodiment, the degree of wear of the machine is determined by means of the computing unit based on an analysis of the condition.

[0068] According to at least one embodiment, the computing unit determines the individual wear level of one or more of the at least two components based on an analysis of their condition.

[0069] In particular, characteristic signatures in the difference signal or the frequency signal can be used to distinguish between a possible anomaly and signs of wear.

[0070] Further embodiments of the method for generating a reference signal according to the improved concept result directly from the various configurations of the method for analyzing a state of a machine according to the improved concept and vice versa.

[0071] According to a further independent aspect of the improved concept, an analysis device for the acoustic analysis of a machine's condition is specified, wherein the analysis device comprises at least one sound transducer and a processing unit. The at least one sound transducer is configured to generate individual acoustic signals depending on the respective operating noises of at least two machine components. The processing unit is configured to determine an expected overall acoustic signal based on the individual signals. The processing unit is configured to generate a reference signal dependent on the overall signal.

[0072] For example, the at least one sound transducer can consist of exactly one sound transducer. In such embodiments, the sound transducer can, for example, be positioned at different locations to record the respective individual operating noises and / or the sound transducer can be used to record the respective operating noises of the different components one after the other.

[0073] Alternatively, the at least one transducer can contain two or more transducers, with one or more of the transducers being used to record each of the individual operating noises.

[0074] The individual acoustic signals are stored, in particular on a storage medium, especially by means of the processing unit.

[0075] In other words, at least two individual acoustic signals are generated, each of which belongs to a corresponding component of the at least two components.

[0076] According to at least one embodiment, the analysis device includes a sensor device configured to determine the current environmental conditions of the machine. The processing unit is configured to determine the overall signal as a function of these current environmental conditions.

[0077] According to at least one embodiment, the at least one sound transducer is configured to generate an acoustic operating signal depending on the operating noise of the machine. The processing unit is configured to analyze the state based on a deviation of the operating signal from the reference signal.

[0078] The operating noise of the machine is in particular an overall operating noise of the machine, i.e. an operating noise that is generated by several of the at least two components together.

[0079] Further embodiments of the analysis device result directly from the described configurations of the various methods according to the improved concept and vice versa.

[0080] According to another independent aspect of the improved concept, a computer program is specified by instructions. When the computer program is executed by a computer system, in particular by an analysis device according to the improved concept, for example, a computing unit of the analysis device, the instructions cause the computer system, in particular the analysis device, for example, the computing unit, to perform a method for generating at least one reference signal for the acoustic analysis of a state of a machine according to the improved concept.

[0081] According to another independent aspect of the improved concept, a further computer program with additional instructions is specified. When the further computer program is executed by an analysis device according to the improved concept, in particular by a computing unit of the analysis device, the additional instructions cause the analysis device, in particular the computing unit, to perform a method for the acoustic analysis of a machine's state according to the improved concept.

[0082] According to another independent aspect of the improved concept, a computer-readable storage medium is specified on which a computer program and / or another computer program according to the improved concept is stored.

[0083] The invention is explained in more detail below with reference to specific embodiments and associated schematic drawings. In the figures, identical or functionally equivalent elements may be designated with the same reference numerals. The description of identical or functionally equivalent elements is not necessarily repeated with respect to different figures.

[0084] The figures show: FIG 1 a schematic representation of an exemplary embodiment of an analysis device according to the improved concept; FIG 2 a flowchart of an exemplary embodiment of a method according to the improved concept, in particular a method for analyzing a state of a machine according to the improved concept; and FIG 3 signals in the context of a method according to FIG 2 .

[0085] In FIG 1 Figure 8 shows a machine which can be designed, for example, as an industrial robot or as a machine for subtractive or additive manufacturing of components.

[0086] Machine 8, for example, has several components, as shown in the schematic example of the FIG 1 Three components 14, 15, 16. The components 14, 15, 16 can, for example, include one or more corresponding motors or other noise-emitting components. A noise-emitting component is defined as such, for example, if it generates noise during the operation of machine 8, in particular during a specific operating state of machine 8.

[0087] The machine 8 also has a control unit 18 which is coupled to at least one of the components 14, 15, 16 in order to control them, among other things.

[0088] In FIG 1 Furthermore, a schematic representation of an exemplary embodiment of an analysis device 9 for the acoustic analysis of a state of the machine 8 is shown.

[0089] The analysis device 9 has a computing unit 13, which is designed, for example, as a computer system with one or more processor units. The computing unit 13 is, for example, coupled to the control unit 18. Furthermore, the analysis device 9 has at least one transducer, in the non-limiting example of the FIG 1 For example, three microphones 10, 11, 12. A first microphone 10 can be arranged in the vicinity of a first component 14, a second microphone 11 can, for example, be arranged in the vicinity of a second component 15, and a third microphone 12 can, for example, be arranged in the vicinity of a third component 16.

[0090] The described arrangement of microphones 10, 11, 12 can, for example, be used to carry out a method for generating at least one reference signal according to the improved concept as shown. In particular, the described arrangement of microphones 10, 11, 12 can serve to record individual acoustic signals from components 14, 15, 16 in order to generate the reference signal.

[0091] After the individual acoustic signals have been recorded, one, several, or all microphones 10, 11, 12 can be arranged differently, particularly to record the operating noise of the machine 8, especially the overall operating noise of the machine 8, in further steps of an acoustic analysis method according to the improved concept. For this purpose, all microphones 10, 11, 12 can be used in various embodiments. In alternative embodiments, not all microphones 10, 11, 12 are used; for example, only one of the microphones 10, 11, 12 is used to record the operating noise of the machine 8.

[0092] Optionally, the analysis device 9 includes a sensor device 17 which is connected or coupled to the computing unit 13, in particular to determine a current environmental condition, for example an ambient temperature, of the machine 8.

[0093] Optionally, the computing unit 13 has a computer-readable storage medium 19, in particular according to the improved concept, or the computing unit 13 is coupled with such a computer-readable storage medium 19.

[0094] The computing unit 13 can, for example, execute a computer program according to the improved concept, which is stored on the computer-readable storage medium 19, in order to cause the analysis device 9 to carry out a procedure according to the improved concept.

[0095] The functioning of the analysis device 9 of the FIG 1 The following will be explained with reference to an exemplary embodiment of a method for the acoustic analysis of a machine's state according to the improved concept, in particular with reference to the FIG 2, FIG 3 The described procedural steps 1-7 do not necessarily have to be carried out in the order suggested by their numbering.

[0096] In FIG 2 A flowchart of an exemplary embodiment of a method for the acoustic analysis of a state of machine 8 according to the improved concept is shown.

[0097] In step 1 of the procedure, for example, the individual acoustic signals of the individual components 14, 15, 16 are recorded using the corresponding microphones 10, 11, 12. The recording of each individual signal of a specific component 14, 15, 16 is performed in isolation from the other components. For example, when the individual signal for the first component 14 is recorded, any operating noise from other components 15, 16 is shielded by the first microphone 10. If possible, some of the other components 15, 16 can also be deactivated while the individual signal of the first component 14 is being recorded. The individual signals are generated, for example, for different individual operating states of the components 14, 15, 16.

[0098] As an alternative to using dedicated microphones 10, 11, 12 for each of the components 14, 15, 16, the individual signals of all components 14, 15, 16 can also be recorded with fewer microphones, for example with only one microphone, for example one after the other.

[0099] In optional step 2 of the procedure, for example, the environmental conditions of the machine 8 are recorded using the sensor device 17. Alternatively or additionally, in optional step 2, information about an arrangement and / or orientation, i.e., the position and / or orientation of the components 14, 15, 16, for example with respect to each other, can be loaded by the processing unit 13, for example from the storage medium 19 or from the control unit 18.

[0100] In step 3 of the procedure, the current operating state of machine 8 is determined. For this purpose, the computing unit 13 can, for example, access the control unit 18, for example, a control program that is executed by the control unit 18, in order to determine current operating parameters or state parameters of machine 8.

[0101] In step 4, the information from steps 1 and 3, and optionally from step 2, is processed to provide a reference signal.

[0102] In particular, the processing unit 13 determines, based on the current operating state of the machine 8, which of the individual signals are to be combined with each other and, if necessary, with further information, such as the information from step 2 or other disturbances or environmental influences, in order to generate the reference signal. Alternatively, several reference signals can be generated in advance, and the processing unit 13 selects one of them depending on the current operating state.

[0103] In step 5 of the procedure, an operating noise is detected during the current operating state of the machine 8 by at least one of the microphones 10, 11, 12 or by means of another microphone (not shown) and recorded as an operating signal.

[0104] In step 6 of the procedure, for example, the computing unit 13 generates a differential signal 22 based on a deviation of the operating signal 21 from the reference signal.

[0105] Also in step 6, the difference signal 22 can be converted into a frequency signal, for example by Fourier transformation of the difference signal 22.

[0106] In step 7, the computing unit 13 can, for example, analyze the difference signal 22 or the frequency signal to analyze the state of the machine 8.

[0107] For example, the computing unit 13 can use the difference or frequency signal to determine whether an anomaly exists in the operation of the machine 8 and, if applicable, which component 14, 15, 16 is responsible for the anomaly.

[0108] Alternatively or additionally, the computing unit 13 can infer a wear level of the machine 8 and / or a respective wear level of one or more of the components 14, 15, 16 based on the differential signal 22 or the frequency signal.

[0109] In FIG 3 The signals described are shown schematically. FIG 3The figure shows the respective signal amplitudes of the operating signal 21, an inverted reference signal 20, and the difference signal 22 as a function of time. The difference signal is the sum of signals 21 and 20.

[0110] The improved concept provides a method for analyzing the condition of a machine, or a method for generating a corresponding reference signal, which does not require large amounts of data and therefore not high computing power, yet still allows for predictive planning of maintenance work in order to minimize downtime or quality losses in a manufacturing process using the machine.

[0111] According to the improved concept, an acoustic method for the early detection of wear or anomalies in mechanical machine components is specified. The deviation between the machine's operating noise, recorded during a process using one or more microphones, and an expected nominal noise level serves as a measure of the wear condition or the presence of an anomaly. The nominal noise level is predicted as a reference signal. Modeling or generating this reference signal can be considered, for example, as providing a digital twin of the machine being analyzed.Predicting the noise development of the machine using the digital twin utilizes a deterministic concept of the nominal noise for different operating states of the machine, so that the correlation with different wear levels already provides more reliable statements with a smaller data basis than conventional approaches. Reference symbol list

[0112] 1 Process step 2 Process step 3 Process step 4 Process step 5 Process step 6 Process step 7 Process step 8 Machine 9 Analysis device 10 Microphone 11 Microphone 12 Microphone 13 Computing unit 14 Component 15 Component 16 Component 17 Sensor device 18 Control unit 19 Storage medium 20 Inverse reference signal 21 Operating signal 22 Differential signal

Claims

1. Method for generating at least one reference signal (20) for acoustic analysis of a state of a machine (8), characterised in that at least one acoustic individual signal is provided for each of at least two components (14, 15, 16) of the machine (8), on the basis of the individual signals, an expected acoustic overall signal is determined by means of a computing unit (13); and on the basis of a deviation of an acoustic operating signal (21) from the reference signal (20), the reference signal (20) for analysing the state of the machine (8) is generated by means of the computing unit (13) depending on the overall signal.

2. Method according to claim 1, characterised in that a respective first individual signal is recorded for each of the at least two components (14, 15, 16), while the respective component (14, 15, 16) is in a corresponding first individual operating state, wherein a combination of the first individual operating states of the at least two components (14, 15, 16) defines an operating state of the machine (8); and the overall signal is determined by means of the computing unit (13) depending on the first individual signals.

3. Method according to one of claims 1 or 2, characterised in that the acoustic overall signal is determined by means of the computing unit (13) depending on a pose of one of the at least two components (14, 15, 16), wherein the pose is a combination of position and orientation of the respective component in three-dimensional space.

4. Method according to one of claims 1 to 3, characterised in that the acoustic overall signal is determined by means of the computing unit (13) depending on a nominal ambient condition of the machine (8).

5. Method according to one of claims 1 to 4, characterised in that a current ambient condition of the machine (8) is determined by means of a sensor apparatus (17); and the reference signal (20) is determined by means of the computing unit (13) depending on the current ambient condition.

6. Method for acoustic analysis of a state of a machine (8), wherein the acoustic operating signal (21) is recorded during operation of the machine (8); characterised in that a reference signal (20) generated based on a method according to one of claims 1 to 5 is provided; and the state is analysed based on the deviation of the operating signal (21) from the reference signal (20).

7. Method according to claim 6, characterised in that a current operating state of the machine (8) is determined by means of the computing unit (13); and the reference signal (20) is provided depending on the current operating state.

8. Method according to one of claims 6 or 7, characterised in that a difference signal (22) is generated by means of the computing unit (13) on the basis of the deviation of the operating signal (21) from the reference signal (20); and the state is analysed by means of the computing unit (13) on the basis of the difference signal (22).

9. Method according to claim 8, characterised in that on the basis of the difference signal (22), a frequency signal is generated by means of the computing unit (13); and the state is analysed by means of the computing unit (13) on the basis of the frequency signal.

10. Method according to one of claims 6 to 9, characterised in that based on the state analysis, it is checked by means of the computing unit (13) whether there is an anomaly during operation of the machine (8); and / or based on the state analysis, a degree of wear of the machine (8) is determined by means of the computing unit (13).

11. Analysis apparatus for acoustic analysis of a state of a machine (8), wherein the analysis apparatus (9) has at least one sound transducer (10, 11, 12) and a computing unit (13); characterised in that the at least one sound transducer (10, 11, 12) is configured to generate acoustic individual signals in each case depending on individual operating noises from at least two components (14, 15, 16) of the machine (8) in each case; on the basis of the individual signals, the computing unit (13) is configured to determine an expected acoustic overall signal; and the computing unit (13) is configured to generate a reference signal (20) depending on the overall signal.

12. Analysis apparatus according to claim 11, characterised in that the analysis apparatus (9) has a sensor apparatus (17) configured to determine a current ambient condition of the machine (8); and the computing unit (13) is configured to determine the overall signal depending on the current ambient condition.

13. Analysis apparatus according to one of claims 11 or 12, characterised in that the at least one sound transducer (10, 11, 12) is configured to generate an acoustic operating signal (21) depending on an operating noise of the machine (8); and the computing unit (13) is configured to analyse the state based on a deviation of the operating signal (21) from the reference signal (20).