Signal processing method, fault detection method, device, equipment and storage medium

Abstract

The present disclosure relates to a signal processing method, a fault detection method, a device, equipment and a storage medium, and relates to the technical field of signal processing. The method processes multiple signal components in the same frequency band of multiple sound signals based on a joint signal component representing the overall situation of the multiple signal components, so that the signal component of each monitoring position can be filtered by referring to the signal components of other monitoring positions, avoiding the problem of poor processing effect caused by separately filtering the signal component of a certain monitoring position when the signal component is abnormal, and improving the filtering effect of the signal component. On this basis, the signal components filtered by the multiple sound signals are fused to obtain a to-be-detected sound signal, which not only improves the signal-to-noise ratio, but also makes the to-be-detected sound signal include more comprehensive sound characteristics because it fuses the sound signals of multiple monitoring positions, thereby improving the accuracy of fault detection.

Description

Technical Field

[0001] This disclosure relates to the field of signal processing technology, and in particular to a signal processing method, a fault detection method, an apparatus, a device, and a storage medium. Background Technology

[0002] During operation, the equipment to be monitored (such as mechanical parts) may malfunction. Therefore, sensors need to be installed at a monitoring location around the equipment to collect its sound signals, which are then used for fault detection. In scenarios with multiple devices to be monitored, the collected sound signals may contain environmental noise, so filtering is necessary to eliminate this noise.

[0003] However, since the device under monitoring is in a dynamic state during operation, while the monitoring position and direction of the sensor are fixed, the sound signal collected by the sensor may be mostly noise signal, that is, invalid signal, which will reduce the accuracy of fault detection based on the sound signal. Summary of the Invention

[0004] This disclosure provides a signal processing method, a fault detection method, an apparatus, a device, and a storage medium. The technical solution of this disclosure is as follows.

[0005] According to one aspect of the embodiments of the present disclosure, a signal processing method is provided, the method comprising:

[0006] Multiple sound signals from the device to be monitored are acquired, and the multiple sound signals are obtained at multiple monitoring locations of the device to be monitored.

[0007] Obtain the first component sequence of each of the plurality of sound signals, wherein the first component sequence of each sound signal includes a plurality of first signal components of the sound signal arranged according to frequency bands;

[0008] For each component group, a joint signal component of each component group is obtained. Each component group includes multiple first signal components located at the same position in each first component sequence. The joint signal component is used to represent the overall situation of multiple first signal components within the component group.

[0009] Based on the joint signal components of each component group, the multiple first signal components in each component group are filtered to obtain the second signal components corresponding to each of the multiple first signal components in each component group.

[0010] Multiple second signal components of the multiple sound signals are fused to obtain the sound signal to be tested of the device to be monitored, and the sound signal to be tested is used to detect faults in the device to be monitored.

[0011] In some embodiments, obtaining the joint signal component of each component group includes:

[0012] For each component group, multiple first signal components within the component group are concatenated to obtain the first component matrix of the component group.

[0013] The step of filtering multiple first signal components within each component group based on the joint signal components of each component group to obtain second signal components corresponding to each of the multiple first signal components within each component group includes:

[0014] For each component group, singular value decomposition filtering is performed on the first component matrix of the component group to obtain a second component matrix. The singular value decomposition filtering is used to remove the same noise in multiple first signal components within the component group.

[0015] Based on the multiple signal components separated from the second component matrix, the second signal components corresponding to each of the multiple first signal components in the component group are obtained.

[0016] In some embodiments, performing singular value decomposition filtering on the first component matrix of the component group to obtain the second component matrix includes:

[0017] Singular value decomposition is performed on the first component matrix to obtain a first diagonal matrix. The diagonal of the first diagonal matrix includes multiple singular values, which are arranged in order of magnitude on the diagonal. The multiple singular values ​​are used to indicate the energy weight of different sound components in the multiple first signal components.

[0018] Set the singular values ​​in the last preset position among the plurality of singular values ​​to zero to obtain the second diagonal matrix;

[0019] The component matrix is ​​restored based on the second diagonal matrix to obtain the second component matrix.

[0020] In some embodiments, the plurality of signal components separated from the second component matrix include signal components corresponding to each of the plurality of first signal components within the component group; obtaining the second signal components corresponding to each of the plurality of first signal components within the component group based on the plurality of signal components separated from the second component matrix includes:

[0021] Determine the mean value among the plurality of signal components to obtain the reference signal component of the component group;

[0022] For each first signal component in the component group, a first filtering process is performed on the signal component corresponding to the first signal component based on the reference signal component of the component group to obtain a third filtered component. A second filtering process is performed on the signal component corresponding to the first signal component based on the reference signal component of the component group to obtain a fourth filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise.

[0023] The third and fourth filtered components are fused to obtain the second signal component corresponding to the first signal component.

[0024] In some embodiments, obtaining the joint signal component of each component group includes:

[0025] For each component group, the mean value among multiple first signal components within the component group is determined to obtain the reference signal component of the component group.

[0026] The step of filtering multiple first signal components within each component group based on the joint signal components of each component group to obtain second signal components corresponding to each of the multiple first signal components within each component group includes:

[0027] For each first signal component within the component group, a first filtering process is performed on the first signal component based on the reference signal component of the component group to obtain a first filtered component, and a second filtering process is performed on the first signal component based on the reference signal component of the component group to obtain a second filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise.

[0028] The first filtered component and the second filtered component are fused to obtain the second signal component corresponding to the first signal component.

[0029] In some embodiments, each filtered component includes signal values ​​at multiple times, and the fusion coefficient between the first filtered component and the second filtered component includes sub-fusion coefficients for the signal values ​​at each of the multiple times. The method further includes:

[0030] For each time step, based on the target data of the previous time step, a sub-fusion coefficient of the signal value at that time step is determined. The target data includes at least one of the following: the sub-fusion coefficient of the signal value at the previous time step, the error of the first filter component and the second filter component at the previous time step, and the signal value of the first filter component at the previous time step. The error at the previous time step represents the error between the signal value of the filter component at the previous time step and the reference signal value at the previous time step. The filter component includes the first filter component and the second filter component.

[0031] In some embodiments, performing a first filtering process on the first signal component based on the reference signal component of the component group to obtain a first filtered component includes:

[0032] Based on the reference signal components of the component group, the first filter coefficients corresponding to the first filtering process are determined, and the first filtering process is performed on the first signal component based on the first filter coefficients to obtain the first filtered component.

[0033] The second filtering process performed on the first signal component based on the reference signal component of the component group to obtain the second filtered component includes:

[0034] Based on the reference signal components of the component group, the second filter coefficients corresponding to the second filtering process are determined, and the first signal component is subjected to the second filtering process based on the second filter coefficients to obtain the second filtered component.

[0035] In some embodiments, fusing the multiple second signal components of the plurality of sound signals to obtain the sound signal to be tested from the device under test includes:

[0036] For each component group, multiple second signal components corresponding to the component group are fused to obtain the fused component corresponding to the component group.

[0037] The fusion components corresponding to each component group are fused to obtain the sound signal to be tested from the device under test.

[0038] In some embodiments, obtaining the first component sequence of each of the plurality of sound signals includes:

[0039] The plurality of sound signals are decomposed to obtain the second component sequence of each of the plurality of sound signals;

[0040] For each sound signal, remove the first signal component that is not in the first preset position in the second component sequence of the sound signal to obtain the first component sequence of the sound signal. The frequency band corresponding to the first signal component that is in the first preset position is higher than the frequency band corresponding to the second signal component.

[0041] According to another aspect of the embodiments of this disclosure, a fault detection method is provided, the method comprising:

[0042] Multiple sound signals from the device to be monitored are acquired, and the multiple sound signals are obtained at multiple monitoring locations of the device to be monitored.

[0043] The plurality of sound signals are processed by the signal processing method described in any of the above embodiments to obtain the sound signal to be tested of the device to be monitored;

[0044] Based on the sound signal to be tested, the device to be monitored is subjected to fault detection, and the fault detection result of the device to be monitored is obtained.

[0045] According to another aspect of the present disclosure, a signal processing apparatus is provided, the apparatus comprising:

[0046] The first acquisition module is used to acquire multiple sound signals from the device to be monitored, wherein the multiple sound signals are obtained at multiple monitoring locations of the device to be monitored.

[0047] The second acquisition module is used to acquire the first component sequence of each of the plurality of sound signals, wherein the first component sequence of each sound signal includes a plurality of first signal components of the sound signal arranged according to frequency bands;

[0048] The third acquisition module is used to acquire a joint signal component for each component group. Each component group includes multiple first signal components located at the same position in each first component sequence. The joint signal component is used to represent the overall situation of the multiple first signal components in the component group.

[0049] The processing module is used to filter multiple first signal components in each component group based on the joint signal components of each component group, so as to obtain the second signal components corresponding to each of the multiple first signal components in each component group.

[0050] The fusion module is used to fuse multiple second signal components of the multiple sound signals to obtain the sound signal to be tested of the device under test, and the sound signal to be tested is used to detect faults in the device under test.

[0051] In some embodiments, the third acquisition module is configured to:

[0052] For each component group, multiple first signal components within the component group are concatenated to obtain the first component matrix of the component group.

[0053] The processing module is used for:

[0054] For each component group, singular value decomposition filtering is performed on the first component matrix of the component group to obtain a second component matrix. The singular value decomposition filtering is used to remove the same noise in multiple first signal components within the component group.

[0055] Based on the multiple signal components separated from the second component matrix, the second signal components corresponding to each of the multiple first signal components in the component group are obtained.

[0056] In some embodiments, the processing module is configured to:

[0057] Singular value decomposition is performed on the first component matrix to obtain a first diagonal matrix. The diagonal of the first diagonal matrix includes multiple singular values, which are arranged in order of magnitude on the diagonal. The multiple singular values ​​are used to indicate the energy weight of different sound components in the multiple first signal components.

[0058] Set the singular values ​​in the last preset position among the plurality of singular values ​​to zero to obtain the second diagonal matrix;

[0059] The component matrix is ​​restored based on the second diagonal matrix to obtain the second component matrix.

[0060] In some embodiments, the plurality of signal components separated from the second component matrix include the signal components corresponding to each of the plurality of first signal components within the component group; the processing module is configured to:

[0061] Determine the mean value among the plurality of signal components to obtain the reference signal component of the component group;

[0062] For each first signal component in the component group, a first filtering process is performed on the signal component corresponding to the first signal component based on the reference signal component of the component group to obtain a third filtered component. A second filtering process is performed on the signal component corresponding to the first signal component based on the reference signal component of the component group to obtain a fourth filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise.

[0063] The third and fourth filtered components are fused to obtain the second signal component corresponding to the first signal component.

[0064] In some embodiments, the third acquisition module is configured to:

[0065] For each component group, the mean value among multiple first signal components within the component group is determined to obtain the reference signal component of the component group.

[0066] The processing module is used for:

[0067] For each first signal component within the component group, a first filtering process is performed on the first signal component based on the reference signal component of the component group to obtain a first filtered component, and a second filtering process is performed on the first signal component based on the reference signal component of the component group to obtain a second filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise.

[0068] The first filtered component and the second filtered component are fused to obtain the second signal component corresponding to the first signal component.

[0069] In some embodiments, each filtered component includes signal values ​​at multiple times, and the fusion coefficient between the first filtered component and the second filtered component includes sub-fusion coefficients for the signal values ​​at each of the multiple times. The apparatus further includes a determining module for:

[0070] For each time step, based on the target data of the previous time step, a sub-fusion coefficient of the signal value at that time step is determined. The target data includes at least one of the following: the sub-fusion coefficient of the signal value at the previous time step, the error of the first filter component and the second filter component at the previous time step, and the signal value of the first filter component at the previous time step. The error at the previous time step represents the error between the signal value of the filter component at the previous time step and the reference signal value at the previous time step. The filter component includes the first filter component and the second filter component.

[0071] In some embodiments, the processing module is configured to:

[0072] Based on the reference signal components of the component group, the first filter coefficients corresponding to the first filtering process are determined, and the first filtering process is performed on the first signal component based on the first filter coefficients to obtain the first filtered component.

[0073] Based on the reference signal components of the component group, the second filter coefficients corresponding to the second filtering process are determined, and the first signal component is subjected to the second filtering process based on the second filter coefficients to obtain the second filtered component.

[0074] In some embodiments, the fusion module is configured to:

[0075] For each component group, multiple second signal components corresponding to the component group are fused to obtain the fused component corresponding to the component group.

[0076] The fusion components corresponding to each component group are fused to obtain the sound signal to be tested from the device under test.

[0077] In some embodiments, the second acquisition module is configured to:

[0078] The plurality of sound signals are decomposed to obtain the second component sequence of each of the plurality of sound signals;

[0079] For each sound signal, remove the first signal component that is not in the first preset position in the second component sequence of the sound signal to obtain the first component sequence of the sound signal. The frequency band corresponding to the first signal component that is in the first preset position is higher than the frequency band corresponding to the second signal component.

[0080] According to another aspect of the present disclosure, a fault detection apparatus is provided, the apparatus comprising:

[0081] The acquisition module is used to acquire multiple sound signals from the device to be monitored, wherein the multiple sound signals are obtained at multiple monitoring locations of the device to be monitored;

[0082] The processing module is used to process the plurality of sound signals using the signal processing method described in any of the above embodiments to obtain the sound signal to be tested of the device to be monitored;

[0083] The detection module is used to perform fault detection on the device under monitoring based on the sound signal to be tested, and to obtain the fault detection result of the device under monitoring.

[0084] According to another aspect of the embodiments of the present disclosure, a computer device is provided, the computer device comprising:

[0085] processor;

[0086] Memory used to store the processor's executable instructions;

[0087] The processor is configured to execute the instructions to implement the signal processing method or fault detection method described above.

[0088] According to another aspect of the present disclosure, a computer-readable storage medium is provided, which, when the instructions in the computer-readable storage medium are executed by a processor of a computer device, enables the computer device to perform the above-described signal processing method or fault detection method.

[0089] According to another aspect of the present disclosure, a computer program product is provided, the computer program product including a computer program that, when executed by a processor, implements the above-described signal processing method or fault detection method.

[0090] This application provides a signal processing method. The method sets up multiple monitoring locations for the device under test to acquire its sound signals, and acquires signal components of each sound signal in multiple frequency bands. Since signal components in different frequency bands represent different sound components, filtering based on signal components can improve the targeting and accuracy of the processing, avoiding the problem of poor processing effect caused by processing the entire sound signal. Furthermore, when filtering signal components, for multiple signal components in the same frequency band of multiple sound signals, processing is performed based on a joint signal component representing the overall situation of these multiple signal components. This allows the signal components at each monitoring location to be filtered with reference to the signal components at other monitoring locations, avoiding the problem of poor processing effect caused by filtering a single signal component when it is abnormal, thus improving the filtering effect of the signal components. Based on this, the filtered signal components of multiple sound signals are fused to obtain the sound signal to be tested. This not only improves the signal-to-noise ratio, but also, because the sound signal to be tested incorporates sound signals from multiple monitoring locations, makes the sound features included in the sound signal to be tested more comprehensive, thereby improving the accuracy of fault detection.

[0091] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0092] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0093] Figure 1 This is a schematic diagram illustrating an implementation environment according to an exemplary embodiment.

[0094] Figure 2 This is a flowchart illustrating a signal processing method according to an exemplary embodiment.

[0095] Figure 3 This is a flowchart illustrating a signal processing method according to an exemplary embodiment.

[0096] Figure 4 This is a flowchart illustrating a signal processing method according to an exemplary embodiment.

[0097] Figure 5This is a flowchart illustrating a signal processing method according to an exemplary embodiment.

[0098] Figure 6 This is a flowchart illustrating a signal processing method according to an exemplary embodiment.

[0099] Figure 7 This is a flowchart illustrating a fault detection method according to an exemplary embodiment.

[0100] Figure 8 This is a block diagram illustrating a signal processing apparatus according to an exemplary embodiment.

[0101] Figure 9 This is a block diagram illustrating a fault detection device according to an exemplary embodiment.

[0102] Figure 10 This is a schematic diagram of the structure of a terminal according to an exemplary embodiment.

[0103] Figure 11 This is a schematic diagram of the structure of a server according to an exemplary embodiment. Detailed Implementation

[0104] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0105] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0106] It should be noted that all information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.), and signals involved in this disclosure are authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. For example, the sound signals involved in this disclosure were obtained with full authorization.

[0107] The signal processing method provided in this application can be executed by a computer device, which can be at least one of a terminal and a server. The following is a schematic diagram illustrating the implementation environment of the signal processing method provided in this application. See also... Figure 1 , Figure 1 This is a schematic diagram of an implementation environment provided in an embodiment of this application. The implementation environment includes: terminal 101 and server 102.

[0108] In some embodiments, terminal 101 may run a client application with a target application, which provides a function to monitor in real time whether the device under monitoring has malfunctioned. This application does not limit the implementation form of the target application; for example, it may be an application that requires downloading and installation, a mini-program that does not require installation, a web application, etc.

[0109] In this embodiment, terminal 101 can be installed at multiple monitoring locations corresponding to the device to be monitored, and is used to acquire the sound signals of the device to be monitored. Server 102 is used to provide background services for the target application. After terminal 101 acquires the sound signals of the device to be monitored at multiple monitoring locations, it transmits the sound signals to server 102. Server 102 processes the multiple sound signals to obtain the sound signal to be tested of the device to be monitored, and then performs fault detection based on the sound signal to be tested, and sends the detected faults to terminal 101 for output.

[0110] In other embodiments, the terminal 101 itself may also acquire multiple sound signals of the device to be monitored and process the multiple sound signals to obtain the sound signal to be tested of the device to be monitored, and then perform fault detection based on the sound signal to be tested, and output the detected fault.

[0111] Terminal 101 can be a computer device such as an acoustic sensor with signal processing and output display functions, a mobile phone, a tablet computer, a multimedia playback device, a PC (Personal Computer), a wearable device, a VR (Virtual Reality) device, an AR (Augmented Reality) device, or a MR (Mixed Reality) device. Server 102 can be a standalone physical server, a server cluster consisting of multiple physical servers, or a distributed file system. It can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Server 102 and terminal 101 are directly or indirectly connected via wired or wireless communication.

[0112] Figure 2 This is a flowchart illustrating a signal processing method according to an exemplary embodiment, such as... Figure 2 As shown, the method is performed by a computer device and includes at least one of the following steps.

[0113] 201. Acquire multiple sound signals from the device to be monitored, which are obtained at multiple monitoring locations of the device.

[0114] In the embodiments of this application, the device to be monitored can be a device under various scenarios, and the device to be monitored can be a mechanical component, such as a gearbox, bearing, or motor of a battery swapping station, or a traction fan of a locomotive, without specific limitations. Furthermore, the device to be monitored in the embodiments of this application can be not only a device that is in a non-moving state during operation, but also a device that is in a dynamically moving state during operation.

[0115] Optionally, sensors are installed at multiple monitoring locations of the device under monitoring. These sensors are used to collect the sound signals from the device. The multiple monitoring locations are distributed around the device, and their specific positions can be set as needed. For example, if there are three monitoring locations, they can be arranged in a triangular shape around the device; if there are four monitoring locations, they can be arranged in a rectangular shape around the device. The distance between each monitoring location and the device under monitoring does not exceed a preset distance to facilitate the collection of the device's sound signals at the monitoring location.

[0116] In this embodiment, the sensor at each monitoring location can continuously collect sound signals. These multiple sound signals are sound signals collected in real time at multiple monitoring locations within the same time period, so as to perform real-time fault detection on the device under monitoring through sound signals. Each sound signal includes signal values ​​at multiple moments, which are also the amplitude of the signal.

[0117] 202. Obtain the first component sequence of each of the multiple sound signals, wherein the first component sequence of each sound signal includes multiple first signal components of the sound signal arranged according to frequency band.

[0118] Since sound signals are collected at each monitoring location, the collected sound signals include not only the sound generated by the device under monitoring, but also the sound in the environment where the device under monitoring is located. That is, the sound signals include multiple sound components, and multiple sound components can correspond to multiple frequency bands. Therefore, each sound signal can correspond to a first component sequence, which includes the signal components of the sound signal in multiple frequency bands.

[0119] In this first component sequence, the multiple first signal components are arranged according to frequency bands, either in descending order of frequency band or in ascending order of frequency band. In this embodiment, the arrangement of the multiple first signal components in the first component sequence in descending order of frequency band is used as an example for illustration.

[0120] 203. For each component group, obtain the joint signal component of each component group. Each component group includes multiple first signal components located at the same position in each first component sequence. The joint signal component is used to represent the overall situation of multiple first signal components in the component group.

[0121] In this embodiment, the number of first signal components in each first component sequence is the same. First signal components located at the same position in each first component sequence correspond to the same frequency band. The number of component groups is the same as the number of first signal components in each first component sequence.

[0122] Optionally, the joint signal component is determined based on multiple first signal components within the component group. The joint signal component can be the sum, mean, or weighted sum of multiple first signal components, etc., and is not specifically limited here.

[0123] 204. Based on the joint signal components of each component group, filter the multiple first signal components in each component group to obtain the second signal components corresponding to each of the multiple first signal components in each component group.

[0124] The second signal component is also the first signal component after filtering.

[0125] 205. The second signal components of multiple sound signals are fused to obtain the sound signal to be tested of the device under test. The sound signal to be tested is used to detect faults in the device under test.

[0126] Each sound signal has multiple second signal components, which correspond one-to-one with multiple first signal components in the first component sequence of the sound signal. The second signal components are the filtered first signal components. The multiple second signal components of multiple sound signals are fused, that is, the multiple first signal components of each sound signal are filtered and then fused to obtain the sound signal to be tested.

[0127] This application provides a signal processing method. The method sets multiple monitoring locations for the device under test to acquire its sound signals, and acquires signal components of each sound signal in multiple frequency bands. Since different signal components represent different sound components, filtering based on signal components improves the targeting and accuracy of the processing, avoiding the poor processing effect caused by processing the entire sound signal. Furthermore, when filtering signal components, for multiple signal components in the same frequency band of multiple sound signals, processing is performed based on a joint signal component representing the overall situation of these multiple signal components. This allows the signal components at each monitoring location to be filtered with reference to the signal components at other monitoring locations, avoiding the poor processing effect caused by filtering a single signal component when it is abnormal, thus improving the filtering effect of the signal components. Based on this, the filtered signal components of multiple sound signals are fused to obtain the sound signal to be tested. This not only improves the signal-to-noise ratio, but also, because the sound signal to be tested incorporates sound signals from multiple monitoring locations, makes the sound features included in the sound signal to be tested more comprehensive, thereby improving the accuracy of fault detection.

[0128] The above Figure 2 The diagram shown is merely the basic process of this disclosure. The following section, based on a specific implementation method, further elaborates on the solution provided in this disclosure. See also... Figure 3 , Figure 3 This is a flowchart illustrating a signal processing method according to an exemplary embodiment, the method being performed by a computer device, the method comprising at least one of the following steps.

[0129] 301. Acquire multiple sound signals from the device to be monitored, which are obtained at multiple monitoring locations of the device.

[0130] Step 301 is similar to step 201, and will not be repeated here.

[0131] 302. Decompose multiple sound signals separately to obtain the second component sequence of each sound signal. The second component sequence of each sound signal includes multiple first signal components of the sound signal arranged according to frequency band.

[0132] In the embodiments of this application, sound signals can be decomposed in various ways, including but not limited to wavelet decomposition, VMD (Variational Mode Decomposition), LMD (Local Mean Decomposition), EMD (Empirical Mode Decomposition), and ICEEMDAN (Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise) decomposition.

[0133] In this embodiment, the decomposition of a sound signal using the ICEEMDAN method is used as an example. The signal components obtained by ICEEMDAN decomposition can also be called IMFs (Intrinsic Mode Functions). In this implementation, multiple Gaussian white noises are first obtained, each with a mean of 0. The number of Gaussian white noises is I, where I is an integer greater than 1. For each Gaussian white noise, a noise figure is determined based on that Gaussian white noise. This noise figure refers to the coefficient assigned to the Gaussian white noise by the sound signal, and it is used to determine the signal components.

[0134] In the process of decomposing a sound signal using the ICEEMDAN decomposition method, the sound signal is decomposed step by step to obtain multiple first signal components. The first signal component that appears earlier in the decomposition order, that is, the lower the order of the first signal component, corresponds to a higher frequency band.

[0135] In the ICEEMDAN decomposition method, the noise figures used to determine the first-order signal component and the other-order signal components are different, and the methods for determining the noise figures are also different. Taking the decomposition of the sound signal into K signal components as an example, where K is an integer greater than 1, the process of determining the noise figures can be referred to in the following formulas (1) and (2). Formula (1) is used to determine the noise figures used to determine the first-order signal component, and formula (2) is used to determine the noise figures used to determine the other-order signal components.

[0136]

[0137] ε(k,i) =α0std(R) k ), k≠0 (2)

[0138] Among them, EMD k (·) represents the k-th noise signal component obtained by EMD decomposition of Gaussian white noise; EMD1(·) represents the 1-th noise signal component obtained by EMD decomposition of Gaussian white noise; W i Let R represent the i-th Gaussian white noise (i = 1, 2, 3, 4, 5... I), where I represents the number of Gaussian white noises; k Let std(R) represent the noise residual signal obtained by decomposing Gaussian white noise k times. k ) represents R k The standard deviation; α0 represents a preset fixed value; ε (k,i) This represents the noise figure used to determine the first signal component of order k.

[0139] Among them, the sound signal is decomposed based on the noise figure to obtain the signal components. The process of obtaining the first-order signal component and other-order signal components based on the noise figure is also different, as shown in the following formulas (3)-(7). Formulas (3)-(4) are used to determine the first-order signal component based on the noise figure, and formulas (5)-(7) are used to determine other-order signal components based on the noise figure.

[0140]

[0141]

[0142] Where R0 represents the initial residual signal and X represents the sound signal, then R0 = X; This represents the noise-added audio signal corresponding to X; EMD1(W i ) represents the first noise signal component obtained by EMD decomposition of the i-th Gaussian white noise; ε (0,i) This represents the noise figure of the i-th Gaussian white noise when k=0; since there are I Gaussian white noises, the I noise-added tone signals corresponding to the I Gaussian white noises are obtained through formula (3); This represents the first signal component obtained by performing EMD decomposition on the i-th noisy audio signal. Mean(·) represents the mean operator. That is, it represents the mean of the first signal component of I noise-added audio signals. IMF1 represents the first-order first signal component of the audio signal, that is, the first signal component decomposed from the audio signal.

[0143] R k =R k-1 -IMF k (5)

[0144] Y i =R k +EMD k (W i )*ε (k,i) (6)

[0145] IMF k+1 =Mean(EMD1(Yi i (7)

[0146] Among them, IMF k R represents the k-th first signal component of the sound signal; k-1 R represents the residual signal corresponding to the (k-1)th first signal component; k EMD represents the residual signal corresponding to the k-th first signal component. k (W i Y represents the k-th noise signal component obtained by EMD decomposition of the i-th Gaussian white noise; i EMD1(Yi) represents the noise-added tone signal corresponding to the i-th Gaussian white noise. i ) indicates that for Yi i The first signal component obtained by EMD decomposition, Mean(EMD1(Y) i That is, it means I Y i The mean of the first signal component; IMF k+1 Let K denote the (k+1)th first signal component of the sound signal, that is, the (k+1)th first signal component decomposed from the sound signal (k = 1, 2, 3... K), where K represents the number of first signal components decomposed from the sound signal.

[0147] In this embodiment, the sound signal is decomposed stepwise according to the above formulas (3)-(7) to obtain K signal components (IMF, Intrinsic Mode Function) and a residual term R. K Then the sound signal can be represented by the following formula (8).

[0148]

[0149] Where X represents the sound signal; IMF k This represents the k-th first signal component; K represents the number of first signal components. The number of first signal components can be set as needed, such as determining the number based on the complexity of the sound signal. K can be any number from 6 to 10. The higher the complexity of the sound signal, the more first signal components are required.

[0150] In this embodiment, three monitoring locations, P1, P2, and P3, are used as examples. The sound signals at the three monitoring locations can be represented as Xp1(t), Xp2(t), and Xp3(t), respectively. The second component sequence of the sound signals at the three monitoring locations can be represented as...

[0151] In this embodiment, since the monitored device is in a dynamic moving state during operation, the collected sound signals have multi-band coupling, transient impact, and non-stationary characteristics. Furthermore, the sensitive frequency bands of the sound signals collected at different monitoring locations differ. The sound signal is decomposed using the ICEEMDAN method, which is suitable for non-stationary signals and multi-component coupling scenarios. This method can adaptively decompose multiple signal components without requiring a preset basis function, effectively separating signal components of different frequency bands. These different frequency bands contain different sound characteristics. For example, high-frequency signal components include gear meshing noise and bearing local defect impact (sensitive to early faults) noise, while mid-to-low-frequency signal components reflect the fundamental frequency of shaft rotation and structural resonance characteristics. Independent decomposition of sound signals from multiple monitoring locations avoids spatial aliasing and preserves the sensitive frequency bands in each monitoring location. For instance, if a monitoring location is closer to the monitored device, the high-frequency signal components in the sound signal acquired at that location are more significant.

[0152] 303. For each sound signal, remove the first signal component that is not in the first preset position in the second component sequence of the sound signal to obtain the first component sequence of the sound signal. The frequency band corresponding to the first signal component that is in the first preset position is higher than the frequency band corresponding to the first signal component that is in the second preset position.

[0153] The frequency band corresponding to the first signal component in the sequence is higher than that corresponding to the second signal component in the sequence. The high-frequency signal components contain the main noise and the noise corresponding to the transient impact of the equipment, while the low-frequency signal components contain the stable sound characteristics of the equipment. In other words, the high-frequency signal components are mainly used for fault detection of the monitored equipment. Therefore, only the high-frequency signal components are retained to avoid the interference caused by the low-frequency signal components to fault detection, improve the accuracy of fault detection, and reduce the amount of data to be processed in the subsequent process, thereby improving the efficiency of signal processing.

[0154] It should be noted that steps 302-303 above are only one optional implementation of obtaining the first component sequence of each of the multiple sound signals. This process can also be implemented in other ways, such as directly using the second component sequence as the first component sequence.

[0155] 304. For each component group, multiple first signal components within the component group are spliced ​​together to obtain the first component matrix of the component group. Each component group includes multiple first signal components located at the same position in each first component sequence.

[0156] Taking the first signal component in each first component sequence as an example, the component group containing these multiple first signal components is called the first component group. The first component matrix of the first component group can then be represented as follows: p1, p2, and p3 each represent a monitoring location. These represent the first signal components of the sound signals at these three monitoring locations, respectively, and T represents transpose.

[0157] 305. For each component group, perform singular value decomposition filtering on the first component matrix of the component group to obtain the second component matrix. Singular value decomposition filtering is used to remove the same noise in multiple first signal components within the component group.

[0158] The process of performing Singular Value Decomposition (SVD) filtering on the first component matrix of the component group to obtain the second component matrix includes the following steps: performing Singular Value Decomposition on the first component matrix to obtain a first diagonal matrix, the diagonal of the first diagonal matrix containing multiple singular values ​​arranged in order of magnitude on the diagonal, the multiple singular values ​​being used to indicate the energy weights of different sound components in the multiple first signal components; setting the singular values ​​in the last preset position of the multiple singular values ​​to zero to obtain the second diagonal matrix; and restoring the component matrix based on the second diagonal matrix to obtain the second component matrix.

[0159] In this embodiment, taking an example of 3 sound signals, each containing m time-time signal values, the process of performing singular value decomposition on the first component matrix of the k-th component group can be represented as follows: C k Let C represent the first component matrix. k It is an m x 3 matrix; U k U represents the left singular vector matrix. k For an m x m matrix, U k This represents the fluctuation pattern of a sound signal, that is, the weighted distribution of common sound components in the time domain; ∑ k Let ∑ represent the first diagonal matrix. k A matrix with m rows and 3 columns; V k V represents a right singular vector matrix. k V is a 3x3 matrix. k This represents the weight distribution of the first signal component across multiple monitoring locations.

[0160] The first diagonal matrix contains multiple singular values ​​arranged in ascending order, with larger singular values ​​appearing earlier in the order. The later singular values ​​represent the energy weights of the same noise within multiple first signal components, such as environmental interference noise. Setting the later singular values ​​in the first diagonal matrix to zero yields the second diagonal matrix, achieving dimensionality reduction. This ensures that the energy proportion of the sound components corresponding to the retained singular values ​​is greater than a preset energy proportion threshold, which, based on practical experience, can be set to 70%.

[0161] It should be noted that in this singular value decomposition (SVD) filtering technique, singular values ​​are elements on the diagonal of the first diagonal matrix. Therefore, the number of singular values ​​is the smaller of the number of rows and columns of the first diagonal matrix. Taking a sound signal with 3 signals, each including signal values ​​at m time points, as an example, the first diagonal matrix is ​​an m-3 matrix. Since the singular values ​​are elements on the diagonal of the first diagonal matrix, and m is generally greater than 3, the number of singular values ​​is also 3. These 3 singular values ​​are the elements in the first row and first column, the second row and second column, and the third row and third column of the first diagonal matrix. The first two singular values ​​are retained, and the last singular value is set to zero. This way, the main sound components corresponding to the retained first two singular values ​​account for more than 85%, effectively removing common noise among multiple first signal components.

[0162] The second diagonal matrix can be represented as ∑' k Then the second component matrix obtained based on the second diagonal matrix can be expressed as:

[0163] Therefore, setting the singular value ranked last among multiple singular values ​​to zero removes the common noise in multiple first signal components. This common noise is generally environmental interference, which affects fault detection. Removing it in this way reduces the impact of environmental interference on fault detection. This method utilizes the spatial distribution characteristics of multiple monitoring locations to suppress common noise in multiple signal components, enhancing the local sound characteristics of each signal component relative to its monitoring location, thus achieving noise suppression across monitoring locations. Furthermore, since environmental interference noise is common across multiple signal components, joint processing of multiple first signal components effectively removes the common noise, and removing noise from multiple first signal components at once improves processing efficiency. After removing the common noise from each first signal component, the local sound characteristics of the first signal component based on its monitoring location are enhanced, facilitating fault detection based on these local sound characteristics.

[0164] 306. Based on the multiple signal components split from the second component matrix, the second signal components corresponding to each of the multiple first signal components in the component group are obtained.

[0165] In this process, each row of elements is extracted from the second component matrix, and the vector formed by each row of elements is a decomposed signal component. As described above... Taking this as an example, the multiple signal components that are separated are represented as follows:

[0166] Among them, multiple second signal components are obtained based on the multiple signal components that have been split off. These multiple signal components can be directly used as multiple second signal components, or the multiple signal components that have been split off can be further processed to obtain multiple second signal components. The specific process will not be described in detail here.

[0167] In this embodiment, sound signals from the monitored device are collected at multiple monitoring locations. Environmental interference signals (such as conducted noise from the motor base) then appear as common noise across these multiple sound signals, while the actual fault signal of the monitored device exhibits local spatial correlation. In this embodiment, a matrix is ​​constructed from the first signal components of the same frequency band in the multiple sound signals. After SVD filtering, the common noise in the multiple first signal components is suppressed, and the local correlation features indicating fault characteristics in each first signal component are enhanced. Thus, SVD filtering is used to collaboratively reduce noise in multiple first signal components, achieving effective noise reduction. Furthermore, even if a first signal component at a certain monitoring location fails due to electromagnetic interference, it can still be reconstructed using the principal components of other first signal components, improving the robustness of signal processing.

[0168] 307. For each component group, fuse the multiple second signal components corresponding to the component group to obtain the fused component corresponding to the component group.

[0169] Each component group includes multiple first signal components, and the corresponding multiple second signal components are also the multiple second signal components corresponding to these multiple first signal components. Taking the number of sound signals as P as an example, each component group includes P first signal components, and each component group also corresponds to P second signal components. If the first component sequence of each sound signal includes K first signal components, then K fused components are obtained, where P and K are both integers greater than 1.

[0170] The fusion of multiple second signal components can be achieved through summation or weighted summation. In this embodiment, the sum of multiple second signal components is used as an example for explanation. Since each second signal component includes signal values ​​at multiple times, the fused component also includes signal values ​​at multiple times. The signal value at each time moment in the fused component is the sum of the signal values ​​of the multiple second signal components at that time moment.

[0171] 308. The corresponding fusion components of each component group are fused to obtain the sound signal to be tested of the device under test. The sound signal to be tested is used to detect faults in the device under test.

[0172] The fusion of multiple components can be performed by summation or weighted summation. In this embodiment, the sum of the multiple fusion components is used as an example to illustrate the process. Since the fusion components include signal values ​​at multiple times, the sound signal under test also includes signal values ​​at multiple times. The signal value at each time in the sound signal under test is the sum of the signal values ​​of the multiple fusion components at that time. The signal value at each time in the sound signal under test can be expressed as: y sum (t) represents the signal value of the sound signal at time t in the measured sound signal, y k (t) represents the signal value at time t in the k-th fused component.

[0173] In this embodiment of the application, steps 307-308 described above are used to fuse multiple second signal components of multiple sound signals to obtain the sound signal to be tested from the device under test. This process can also be implemented in other ways. For example, for each sound signal, multiple second signal components of the sound signal are fused to obtain a processed sound signal, and then the multiple processed sound signals are fused to obtain the sound signal to be tested.

[0174] In this embodiment, the second signal components corresponding to each component group are fused, realizing the fusion of signal components after filtering in the same frequency band. Then, the fused components of multiple frequency bands are fused to obtain the sound signal to be tested. This achieves the fusion of the signal components after filtering each sound signal to obtain the sound signal to be tested, improving the signal-to-noise ratio of the sound signal to be tested and making the sound signal to be tested include more comprehensive sound characteristics of the device to be monitored. Therefore, based on the sound signal to be tested, fault detection can be performed, which can improve the accuracy of fault detection.

[0175] In this embodiment, a matrix is ​​constructed from the first signal components of the same frequency band among multiple sound signals. After SVD filtering, common noise in the multiple first signal components is suppressed, and the local correlation features used to indicate fault characteristics in each first signal component are enhanced. Thus, SVD filtering is used to collaboratively reduce noise in multiple first signal components, achieving effective noise reduction. The first signal components after processing multiple sound signals are fused to obtain the sound signal to be tested. This not only improves the signal-to-noise ratio but also, because the sound signal to be tested incorporates sound signals from multiple monitoring locations, makes the sound features included in the sound signal to be tested more comprehensive, thereby improving the accuracy of fault detection.

[0176] The above Figure 3 The embodiment is illustrated by taking the processing of multiple first signal components based on a splicing matrix between multiple first signal components as an example. The following is a description of the process. Figure 4 An embodiment is described using the process of processing multiple first signal components based on the average value among multiple first signal components. See [link to example]. Figure 4 , Figure 4 This is a flowchart illustrating a signal processing method according to an exemplary embodiment, the method being performed by a computer device, the method comprising at least one of the following steps.

[0177] 401. Acquire multiple sound signals from the device to be monitored, which are obtained at multiple monitoring locations of the device to be monitored.

[0178] 402. Decompose multiple sound signals separately to obtain the second component sequence of each sound signal. The second component sequence of each sound signal includes multiple first signal components of the sound signal arranged according to frequency band.

[0179] 403. For each sound signal, remove the first signal component that is not in the first preset position in the second component sequence of the sound signal to obtain the first component sequence of the sound signal. The frequency band corresponding to the first signal component that is in the first preset position is higher than the frequency band corresponding to the first signal component that is in the second preset position.

[0180] Steps 401-403 are similar to steps 301-303, and will not be repeated here.

[0181] 404. For each component group, determine the mean value among multiple first signal components within the component group to obtain the reference signal component of the component group. Each component group includes multiple first signal components located at the same position in each first component sequence.

[0182] Each signal component includes signal values ​​at multiple times, and the reference signal component includes reference signal values ​​at multiple times. Taking a component group consisting of three first signal components as an example, the reference signal component of the kth component group can be obtained by the following formula (9).

[0183]

[0184] Where t represents time, d k (t) represents the reference signal value at time t in the reference signal component; p1, p2, p3 represent three monitoring positions. It represents the sum of the signal values ​​of the three first signal components in the k-th component group at time t.

[0185] 405. For each first signal component in the component group, the first signal component is subjected to a first filtering process based on the reference signal component of the component group to obtain a first filtered component. The first signal component is subjected to a second filtering process based on the reference signal component of the component group to obtain a second filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise.

[0186] Wherein, the reference signal component is used to obtain the filter coefficients. The process of performing a first filtering process on the first signal component based on the reference signal component of the component group to obtain the first filtered component includes the following steps: determining the first filter coefficients corresponding to the first filtering process based on the reference signal component of the component group; and performing a first filtering process on the first signal component based on the first filter coefficients to obtain the first filtered component. Similarly, the process of performing a second filtering process on the first signal component based on the reference signal component of the component group to obtain the second filtered component includes the following steps: determining the second filter coefficients corresponding to the second filtering process based on the reference signal component of the component group; and performing a second filtering process on the first signal component based on the second filter coefficients to obtain the second filtered component.

[0187] The first and second filtering processes can be set as needed. In this embodiment, the first filtering process is LMS (Least Mean Square Filtering) and the second filtering process is RLS (Recursive Least Squares Filtering) as an example.

[0188] Each first signal component includes signal values ​​at multiple times. Performing a first filtering process on the first signal component includes performing a first filtering process on the signal value at each time. This process includes the following steps: For each time, the signal value at that time is subjected to a first filtering process based on the first filter coefficients at that time to obtain the first filtered signal value at that time. The first filtered component also includes the first filtered signal values ​​at multiple times.

[0189] The first filter coefficient is determined based on the reference signal component. Optionally, the first filter coefficient at each time step is determined based on the reference signal value at the previous time step. Since there is no previous time step for the first time step, the first filter coefficient at the first time step can be set, such as 0.

[0190] For each signal value at any given time, the signal value is subjected to a first filtering process based on the first filter coefficient at that time to obtain the signal value after the first filtering process. This process is described in the following formula (10).

[0191] y LMS (t)=w LMS (t)*x(t) (10)

[0192] in, w represents the signal value at time t in the k-th first signal component at monitoring location p. LMs (t) represents the first filter coefficient at time t; y LMS (t) represents the signal value after the first filtering process at time t.

[0193] For each time moment, the error at that time moment is determined based on the reference signal value and the signal value after the first filtering process. Based on the error, signal value and first filter coefficient at that time moment, the first filter coefficient at the next time moment is determined. The process of determining the first filter coefficient is illustrated by any first signal component in the kth component group. See formulas (11)-(12) below.

[0194] e LMS (t)=d k (t)-y LMS (t) (11)

[0195] w LMS (t+1)=w LMS (t)+μe LMS (t)x(t) (12)

[0196] Where, d k (t) represents the reference signal value at time t; e LMS(t) represents the error at time t; μ represents the step size used to control the convergence rate; w LMS (t) represents the first filter coefficient at time t, w LMS (t+1) represents the first filter coefficient at time t+1, which is also the adaptive spatial weight vector.

[0197] Each first signal component includes signal values ​​at multiple times. The second filtering process for the first signal component includes the second filtering process for the signal value at each time. The process includes the following steps: For each time, the signal value at that time is subjected to the second filtering process based on the second filter coefficients at that time to obtain the second filtered signal value at that time. The second filtered component also includes the second filtered signal values ​​at multiple times.

[0198] The second filter coefficients are determined based on the reference signal components. Optionally, the second filter coefficients at each time step are determined based on the reference signal value of the previous time step. Since there is no previous time step for the first time step, the second filter coefficients for the first time step can be set, such as being 0.

[0199] For each signal value at any given time, a second filtering process is performed on the signal value based on the second filter coefficients at that time to obtain the signal value after the second filtering process. This process is described in the following formula (13).

[0200] y RLS (t)=w RLS (t)*x(t) (13)

[0201] in, w represents the signal value at time t in the k-th first signal component at monitoring location p. RLS (t) represents the second filter coefficient at time t; y RLS (t) represents the signal value after the second filtering process at time t.

[0202] For each time step, the error at that time step is determined based on the reference signal value and the second filtered signal value at that time step; the gain vector at that time step is determined based on the input vector at that time step and the inverse correlation matrix of the previous time step. The input vector at each time step is a vector obtained by arranging the signal value at that time step and the signal values ​​before that time step in reverse chronological order. Based on the gain vector, error, and second filter coefficients at that time step, the second filter coefficients at the next time step are determined. Taking any first signal component in the k-th component group as an example, the process of determining the second filter coefficients is described in the following formulas (14)-(16).

[0203] e RLS (t)=d k(t)-y RLS (t) (14)

[0204]

[0205] w RLS (t)=w RLs (t-1)+r RLs (t)k(t) (16)

[0206] Where, d k (t) represents the reference signal value at time t, e RLS (t) represents the error at time t, k(t) represents the gain vector at time t, X(t) represents the input vector at time t, P(t-1) represents the inverse correlation matrix at time t-1, λ represents the forgetting factor, and w RLS (t-1) represents the second filter coefficient at time t-1, w RLS (t) represents the second filter coefficient at time t.

[0207] The inverse correlation matrix at each time step is determined based on the input vector, gain vector, and inverse correlation matrix of the previous time step. That is, the inverse correlation matrix at multiple time steps is determined iteratively. The inverse correlation matrix at the first time step can be set as needed. The process of determining the inverse correlation matrix at each time step is shown in the following formula (17).

[0208] P(t=λ -1 [P(t-1)-k(t)X T [(t)P(t-1)] (17)

[0209] Where P(t) represents the inverse correlation matrix at time t; P(t-1) represents the inverse correlation matrix at time t-1; X T (t) represents the transpose of the input vector at time t.

[0210] In this embodiment, the filter coefficients in the two filtering processes are determined based on the reference signal component. Since the reference signal component is the average of the signal components in the same frequency band at multiple monitoring locations, the problem of inaccurate reference signal components caused by invalid single sound signals or local interference is suppressed. Therefore, the filter coefficients determined based on the reference signal component have high accuracy and can effectively filter the first signal component, thereby improving the accuracy of the filtering results.

[0211] 406. The first filtered component and the second filtered component are fused to obtain the second signal component corresponding to the first signal component.

[0212] The first filter component and the second filter component are fused based on the fusion coefficient. Each filter component includes signal values ​​at multiple times. Therefore, the fusion coefficient between the first filter component and the second filter component includes sub-fusion coefficients of the signal values ​​at multiple times. The fusion process of the signal values ​​at each time time is shown in the following formula (18).

[0213] y(t)=a(t)y LMS (t)+[1-a(t)]y RLS (t) (18)

[0214] Where a(t) represents the sub-fusion coefficient of the signal value at time t; y LMS (t) represents the signal value of the first filtered component at time t; y RLS (t) represents the signal value of the second filtered component at time t.

[0215] Optionally, the sub-fusion coefficients of the signal value at each time step are dynamically determined based on the data from the previous time step. This process includes the following steps: For each time step, based on the target data from the previous time step, the sub-fusion coefficients of the signal value at that time step are determined. The target data includes at least one of the following: the sub-fusion coefficients of the signal value at the previous time step, the errors of the first and second filtered components at the previous time step, and the signal value of the first filtered component at the previous time step. The error at the previous time step represents the error between the signal value of the filtered component at the previous time step and the reference signal value at the previous time step. The filtered component includes the first and second filtered components. The reference signal value is also the reference signal value in the reference signal component. The process for determining the sub-fusion coefficients at each time step is shown in the following formula (19).

[0216]

[0217] θ(t)=θ(t-1)+η[e RLS (t-1)-e LMS (t-1)]y LMS (t-1)a(t-1)[1-a(t-1)] (19)

[0218] Where a(t) represents the sub-fusion coefficient at time t; θ(t) represents the combined weight at time t; η represents the update step size; e LMS (t-1) represents the error of the first filtered component at time t-1; e RLs (t-1) represents the error of the second filtered component at time t-1; y LMS (t-1) represents the signal value of the first filtered component at time t-1; a(t-1) represents the sub-fusion coefficient at time t-1. This method restricts the fusion coefficient to between 0 and 1, realizing dynamic control of the fusion coefficient by the Sigmoid function.

[0219] In this embodiment, the sub-fusion coefficient at each time step is dynamically adjusted based on the sub-fusion coefficient at the previous time step and error data, making the sub-fusion coefficient at each time step more targeted and more accurate.

[0220] In this embodiment, steps 404-406 described above achieve the following process: for each component group, obtain the joint signal component of each component group; based on the joint signal component of each component group, filter multiple first signal components within each component group to obtain the second signal components corresponding to each of the multiple first signal components within each component group. In this embodiment, the average value of the first signal components at multiple monitoring locations is determined as the reference signal component. This utilizes spatial redundancy to reduce the impact of sound signal anomalies at a single monitoring location. Compared to a scheme where each first signal component determines its reference signal component based solely on itself, this scheme suppresses the problem of inaccurate reference signal components caused by invalid single sound signals or local interference. By filtering multiple first signal components based on the reference signal component determined by this scheme, the weight of the sound signal at each monitoring location is dynamically allocated. When a sound signal is invalid, its weight is automatically reduced, ensuring robustness and improving the effectiveness of the filtering process. Furthermore, during the operation of the monitored equipment, the sound signal at each monitoring location may experience instantaneous signal-to-noise ratio fluctuations due to transient faults in the monitored equipment or external noise. In this embodiment, the sound signal is processed jointly using two filtering methods. Since the first filtering process, once in a steady state, only requires multiplication and addition operations to output the filtered value, the computational load is low. It also automatically subtracts the slow drift caused by temperature and aging of sensor components to retain the true dynamic monitoring signal, thus suppressing slow time-varying drift of the sensor. Because the drift changes extremely slowly, this ensures a small steady-state error and further reduces the computational load. The second filtering process can quickly track sudden noise, i.e., filter out sudden noise, resulting in fast convergence. This scheme combines two filtering methods, accelerating initial convergence and ensuring later stability through the second filtering process. Furthermore, the fusion coefficient is automatically adjusted using the Sigmoid function, achieving a balance between convergence speed and steady-state error, thereby improving the filtering effect.

[0221] 407. For each component group, fuse the multiple second signal components corresponding to the component group to obtain the fused component corresponding to the component group.

[0222] 408. The corresponding fusion components of each component group are fused to obtain the sound signal to be tested of the device under test. The sound signal to be tested is used to detect faults in the device under test.

[0223] Steps 407-408 are similar to steps 307-308, and will not be repeated here.

[0224] In this embodiment, the average of the first signal components from multiple monitoring locations is determined as the reference signal component. This utilizes spatial redundancy to reduce the impact of sound signal anomalies from a single monitoring location. Compared to schemes where each first signal component determines its reference signal component solely based on itself, this approach suppresses the problem of inaccurate reference signal components caused by invalid single sound signals or local interference. Therefore, filtering multiple first signal components based on the reference signal component determined by this approach improves the effectiveness of the filtering process. Furthermore, by combining two filtering methods and automatically adjusting the fusion coefficient using the Sigmoid function, a balance is achieved between convergence speed and steady-state error. Fusing the filtered first signal components from multiple sound signals to obtain the test sound signal not only improves the signal-to-noise ratio but also, because the test sound signal incorporates sound signals from multiple monitoring locations, includes more comprehensive sound features, thereby improving the accuracy of fault detection.

[0225] In the above Figure 3 and Figure 4 In the embodiments described, signal processing was performed in two different ways. The following describes the process of combining these two implementation methods for signal processing. See also... Figure 5 , Figure 5 This is a flowchart illustrating a signal processing method according to an exemplary embodiment, the method being performed by a computer device, the method comprising at least one of the following steps.

[0226] 501. Acquire multiple sound signals from the device to be monitored, which are obtained at multiple monitoring locations of the device.

[0227] 502. Decompose multiple sound signals to obtain the second component sequence of each sound signal. The second component sequence of each sound signal includes multiple first signal components of the sound signal arranged according to frequency band.

[0228] 503. For each sound signal, remove the first signal component that is not in the first preset position in the second component sequence of the sound signal to obtain the first component sequence of the sound signal. The frequency band corresponding to the first signal component that is in the first preset position is higher than the frequency band corresponding to the first signal component that is in the second preset position.

[0229] 504. For each component group, multiple first signal components within the component group are spliced ​​together to obtain the first component matrix of the component group. Each component group includes multiple first signal components located at the same position in each first component sequence.

[0230] 505. Perform singular value decomposition filtering on the first component matrix of the component group to obtain the second component matrix. Singular value decomposition filtering is used to remove the same noise in multiple first signal components within the component group.

[0231] Steps 501-505 are similar to steps 301-305, and will not be repeated here.

[0232] 506. Extract multiple signal components corresponding to the component group from the second component matrix, determine the mean value among the multiple signal components, and obtain the reference signal component of the component group. The multiple signal components extracted from the second component matrix include the signal components corresponding to the multiple first signal components in the component group.

[0233] In this embodiment, the signal components separated from the second component matrix correspond one-to-one with multiple first signal components within the component group. Subsequent signal processing is performed on the separated signal components to obtain the sound signal to be tested.

[0234] 507. For each first signal component in the component group, the signal component corresponding to the first signal component is subjected to a first filtering process based on the reference signal component of the component group to obtain a third filtered component. The signal component corresponding to the first signal component is subjected to a second filtering process based on the reference signal component of the component group to obtain a fourth filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise.

[0235] 508. The third and fourth filtered components are fused to obtain the second signal component corresponding to the first signal component.

[0236] The process of processing the signal components in steps 506-508 to obtain the second signal component is the same as the process of processing the first signal component in steps 404-406 to obtain the second signal component, and will not be described again here.

[0237] 509. For each component group, fuse the multiple second signal components corresponding to the component group to obtain the fused component corresponding to the component group.

[0238] 510. The corresponding fusion components of each component group are fused to obtain the sound signal to be tested of the device under test. The sound signal to be tested is used to detect faults in the device under test.

[0239] Steps 509-510 are similar to steps 307-308, and will not be repeated here.

[0240] In this embodiment, singular value decomposition (SVD) filtering achieves coordinated noise reduction of signal components in the same frequency band from multiple monitoring locations. Dynamic weighted fusion of signal components obtained from two filtering methods balances convergence speed and steady-state error. Furthermore, fusing signal components processed from multiple audio signals enables focused fault feature analysis across the entire frequency band. This three-stage processing significantly improves the detection rate of complex faults in monitored equipment in motion, making it particularly suitable for industrial monitoring scenarios with multiple monitoring locations but limited installation space.

[0241] For example, see Figure 6 , Figure 6 This is a flowchart illustrating a signal processing method according to an exemplary embodiment. Taking three monitoring locations as an example, the sound signal at each monitoring location is first decomposed to obtain multiple Information Function Factors (IMFs) for each sound signal. Then, the IMFs of the same frequency band among the multiple sound signals are subjected to collaborative noise reduction through SVD filtering to obtain new IMFs of the same frequency band. A reference signal component is obtained based on the new IMFs of the same frequency band. Based on the reference signal component, the IMFs of the same frequency band at each monitoring location are processed using LMS and RLS filtering methods. The processed IMFs are then fused to obtain a fused component corresponding to the frequency band. Finally, the fused components of the multiple frequency bands are summed to obtain the sound signal to be measured.

[0242] In related technologies, when the sound signal from a single monitoring location is filtered using SANC (Self-Adaptive Noise Cancelling), the suppression of non-stationary noise in the sound signal is limited, and useful sound components are easily mistakenly eliminated when the ambient noise and the target signal spectrum overlap. However, the solution provided in this application performs multi-band decomposition and collaborative noise reduction on sound signals from multiple monitoring locations. Through singular value decomposition filtering and two filtering methods, it effectively suppresses common noise and local interference noise, significantly improving noise reduction performance. Furthermore, the SANC method has high iterative computational complexity (O(n...). 2The current method suffers from poor real-time performance and may over-smooth transient sound components (such as bearing fault pulses) during sound signal processing. The solution provided in this application improves computational efficiency through parallel signal component processing and a hybrid convergence strategy using two filtering methods. Furthermore, it enhances impact characteristics through wavelet thresholding and envelope detection of signal components, resulting in strong feature preservation during sound signal processing. In related technologies, sound signals from a single monitoring location are susceptible to noise interference along the path, leading to high signal loss rates in scenarios where the monitored device is moving. A single monitoring location can also cause missed detection of sound features from moving monitored devices. The solution provided in this application, however, uses multiple sensors at multiple monitoring locations to collect sound signals and adaptively processes these locations, improving environmental adaptability. Moreover, by covering the movement trajectory of the monitored device from multiple monitoring locations, it collects comprehensive sound signals, further enhancing adaptability to various movement scenarios.

[0243] This application provides a signal processing method. The method sets multiple monitoring locations for the device under test to acquire its sound signals, and acquires signal components of each sound signal in multiple frequency bands. Since different signal components represent different sound components, filtering based on signal components improves the targeting and accuracy of the processing, avoiding the problem of poor processing results caused by processing the entire sound signal. Furthermore, when filtering signal components, for multiple signal components in the same frequency band of multiple sound signals, processing is performed based on a joint signal component representing the overall situation of these multiple signal components. This allows the signal components at each monitoring location to be filtered with reference to the signal components at other monitoring locations, avoiding the problem of poor processing results caused by filtering a single signal component when it is abnormal, thus improving the filtering effect of the signal components. Moreover, fusing the filtered signal components of multiple sound signals to obtain the sound signal to be tested not only improves the signal-to-noise ratio, but also, because the sound signal to be tested incorporates sound signals from multiple monitoring locations, makes the sound features included in the sound signal to be tested more comprehensive, thereby improving the accuracy of fault detection.

[0244] The sound signal obtained through any of the above embodiments is used for fault detection of the device to be monitored. The fault detection process is described in [reference needed]. Figure 7 , Figure 7 This is a flowchart illustrating a fault detection method according to an exemplary embodiment, the method comprising at least one of the following steps.

[0245] 701. Acquire multiple sound signals from the device to be monitored, which are obtained at multiple monitoring locations of the device.

[0246] Step 701 is similar to step 201, and will not be repeated here.

[0247] 702. Process multiple sound signals to obtain the sound signal to be tested from the device to be monitored.

[0248] In the embodiments of this application, it can be achieved through the above... Figures 2-5 Any embodiment of the present invention processes multiple sound signals to obtain the sound signal to be tested; the specific process will not be described in detail here.

[0249] 703. Based on the sound signal to be tested, perform fault detection on the device to be monitored and obtain the fault detection result of the device to be monitored.

[0250] In the embodiments of this application, the process of detecting faults in the device under monitoring based on the sound signal under test and obtaining the fault detection result of the device under monitoring includes at least one of the following implementation methods.

[0251] (1) Based on the characteristic frequencies of various faults and the sound signal to be tested, fault detection is performed on the equipment to be monitored. If the sound signal to be tested includes the characteristic frequency of any fault, it is determined that the equipment to be monitored has the fault.

[0252] In some embodiments, before fault detection is performed based on the sound signal to be tested, the sound signal to be tested is preprocessed. This process includes performing Hilbert envelope detection, FFT (Fast Fourier Transform) transformation and envelope spectrum analysis on the sound signal to be tested in sequence to obtain the envelope spectrum corresponding to the sound signal to be tested.

[0253] The process involves performing a Hilbert transform on the sound signal to be measured to obtain conjugate components. These conjugate components are then combined with the sound signal to be measured to synthesize an analytic signal, the magnitude of which is the envelope signal. The envelope signal is then smoothed, and finally, an FFT transform is performed on the smoothed envelope signal to obtain the envelope spectrum.

[0254] Accordingly, if the envelope spectrum includes the characteristic frequency of any fault, it is determined that the device under monitoring has the fault. That is, the envelope spectrum is analyzed based on the characteristic frequencies of multiple faults. If the characteristic frequency of any fault appears in the envelope spectrum, it is determined that the device under monitoring has the fault.

[0255] Taking the bearing as an example of the equipment to be monitored, the characteristic frequencies of various faults can include the frequency characteristics of outer race faults (BPFO, Ball Pass Frequency of Outer race), the frequency characteristics of inner race faults (BPFI, Ball Pass Frequency of Inner race), the frequency characteristics of gear meshing faults (GMF, Gear Mesh Frequency), the frequency characteristics of gear sideband faults, and the frequency characteristics of bearing fault impacts, etc.

[0256] (2) Input the sound signal to be tested into the fault detection model, and the fault detection model processes the sound signal to be tested to obtain the fault detection result of the device to be monitored.

[0257] The fault detection model can be an artificial intelligence model, used to obtain fault detection results for the monitored device based on its sound signals. Using an artificial intelligence model for fault detection improves both efficiency and accuracy.

[0258] Optionally, the fault detection model is trained based on multiple sets of training samples. Each set of training samples includes sample sound signals and fault labels corresponding to the sample sound signals. Thus, the trained fault detection model can obtain accurate fault detection results based on the sound signals. Furthermore, some training samples include sample sound signals corresponding to compound faults and compound fault labels. Therefore, the trained fault detection model can also detect compound faults in the monitored equipment.

[0259] In this embodiment of the application, the sound signals processed from multiple monitoring locations are integrated to realize the comprehensive information of multiple sound signals across the entire frequency band, thereby improving the signal-to-noise ratio of the sound signal under test. In cases where the impact characteristics of early faults (such as gear microcracks) are easily masked by strong rotating fundamental frequencies, they can also be detected through the sound signal under test. Furthermore, it can significantly improve the detection rate of complex faults in the monitored equipment that is in motion during operation.

[0260] In this application embodiment, based on the above Figure 2-5 In any embodiment, multiple sound signals from the monitoring device are processed to obtain a sound signal to be tested. Since the sound signal to be tested not only has a high signal-to-noise ratio, but also incorporates sound signals from multiple monitoring locations, the sound features included in the sound signal to be tested are more comprehensive. Therefore, fault detection based on the sound signal to be tested can improve the accuracy of fault detection.

[0261] Figure 8 This is a block diagram illustrating a signal processing apparatus according to an exemplary embodiment. (Refer to...) Figure 8 The device includes:

[0262] The first acquisition module 801 is used to acquire multiple sound signals from the device to be monitored, and the multiple sound signals are obtained at multiple monitoring positions of the device to be monitored.

[0263] The second acquisition module 802 is used to acquire the first component sequence of each of the multiple sound signals, wherein the first component sequence of each sound signal includes multiple first signal components of the sound signal arranged according to frequency bands;

[0264] The third acquisition module 803 is used to acquire the joint signal component of each component group for each component group. Each component group includes multiple first signal components located at the same position in each first component sequence. The joint signal component is used to represent the overall situation of multiple first signal components in the component group.

[0265] The processing module 804 is used to filter multiple first signal components in each component group based on the joint signal components of each component group, so as to obtain the second signal components corresponding to each of the multiple first signal components in each component group.

[0266] The fusion module 805 is used to fuse multiple second signal components of multiple sound signals to obtain the sound signal to be tested of the device under test, which is used for fault detection of the device under test.

[0267] In some embodiments, the third acquisition module 803 is configured to:

[0268] For each component group, multiple first signal components within the component group are concatenated to obtain the first component matrix of the component group.

[0269] Processing module 804 is used for:

[0270] For each component group, singular value decomposition filtering is performed on the first component matrix of the component group to obtain the second component matrix. Singular value decomposition filtering is used to remove the same noise in multiple first signal components within the component group.

[0271] Based on the multiple signal components separated from the second component matrix, the second signal components corresponding to each of the multiple first signal components in the component group are obtained.

[0272] In some embodiments, the plurality of signal components separated from the second component matrix include the signal components corresponding to each of the plurality of first signal components within the component group; the processing module 804 is configured to:

[0273] Determine the mean value among multiple signal components to obtain the reference signal component of the component group;

[0274] For each first signal component in the component group, the signal component corresponding to the first signal component is subjected to a first filtering process based on the reference signal component of the component group to obtain a third filtered component. The signal component corresponding to the first signal component is subjected to a second filtering process based on the reference signal component of the component group to obtain a fourth filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise.

[0275] The third and fourth filtered components are fused to obtain the second signal component corresponding to the first signal component.

[0276] In some embodiments, the processing module 804 is configured to:

[0277] Singular value decomposition is performed on the first component matrix to obtain the first diagonal matrix. The diagonal of the first diagonal matrix contains multiple singular values, which are arranged in order of magnitude on the diagonal. These multiple singular values ​​are used to indicate the energy weights of different sound components in the multiple first signal components.

[0278] Set the singular value in the last preset position among the multiple singular values ​​to zero to obtain the second diagonal matrix;

[0279] The second component matrix is ​​obtained by restoring the component matrix based on the second diagonal matrix.

[0280] In some embodiments, the third acquisition module 803 is configured to:

[0281] For each component group, the mean value among multiple first signal components within the component group is determined to obtain the reference signal component of the component group.

[0282] Processing module 804 is used for:

[0283] For each first signal component within the component group, the first signal component is subjected to a first filtering process based on the reference signal component of the component group to obtain a first filtered component. The first signal component is subjected to a second filtering process based on the reference signal component of the component group to obtain a second filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise.

[0284] The first filtered component and the second filtered component are fused to obtain the second signal component corresponding to the first signal component.

[0285] In some embodiments, each filtered component includes signal values ​​at multiple times, and the fusion coefficient between the first filtered component and the second filtered component includes sub-fusion coefficients for the signal values ​​at multiple times. The apparatus further includes a determining module for:

[0286] For each time step, based on the target data of the previous time step, the sub-fusion coefficients of the signal value at that time step are determined. The target data includes at least one of the following: the sub-fusion coefficients of the signal value at the previous time step, the errors of the first and second filtered components at the previous time step, and the signal value of the first filtered component at the previous time step. The error at the previous time step represents the error between the signal value of the filtered component at the previous time step and the reference signal value at the previous time step. The filtered component includes the first filtered component and the second filtered component.

[0287] In some embodiments, the processing module 804 is configured to:

[0288] Based on the reference signal components of the component group, the first filter coefficients corresponding to the first filtering process are determined, and the first filtering process is performed on the first signal component based on the first filter coefficients to obtain the first filtered component.

[0289] Based on the reference signal components of the component group, the second filter coefficients corresponding to the second filtering process are determined. Based on the second filter coefficients, the first signal component is subjected to the second filtering process to obtain the second filtered component.

[0290] In some embodiments, the fusion module 805 is configured to:

[0291] For each component group, multiple second signal components corresponding to the component group are fused to obtain the fused component corresponding to the component group.

[0292] The corresponding fusion components of each component group are fused to obtain the sound signal to be measured from the device under test.

[0293] In some embodiments, the second acquisition module 802 is configured to:

[0294] The multiple sound signals are decomposed separately to obtain the second component sequence of each sound signal;

[0295] For each sound signal, remove the first signal component that is not in the first preset position in the second component sequence of the sound signal to obtain the first component sequence of the sound signal. The frequency band corresponding to the first signal component in the first position is higher than the frequency band corresponding to the second signal component in the second position.

[0296] Figure 9 This is a block diagram illustrating a fault detection device according to an exemplary embodiment. (Refer to...) Figure 9 The device includes:

[0297] The acquisition module 901 is used to acquire multiple sound signals from the device under monitoring, and the multiple sound signals are obtained at multiple monitoring positions of the device under monitoring.

[0298] Processing module 902 is used to process multiple sound signals using the signal processing method of any of the above embodiments to obtain the sound signal to be tested of the device to be monitored;

[0299] The detection module 903 is used to perform fault detection on the device under monitoring based on the sound signal to be tested, and to obtain the fault detection result of the device under monitoring.

[0300] Regarding the apparatus in any of the above embodiments, the specific manner in which each unit performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0301] In some embodiments, the computer device is provided as a terminal. Figure 10 A schematic diagram of a terminal 1000 provided in an exemplary embodiment of this disclosure is shown. The terminal 1000 may be a smartphone, tablet computer, MP3 player (Moving Picture Experts Group Audio Layer III), MP4 player (Moving Picture Experts Group Audio Layer IV), laptop computer, or desktop computer. The terminal 1000 may also be referred to as a user device, portable terminal, laptop terminal, desktop terminal, or other names.

[0302] Typically, terminal 1000 includes a processor 1001 and a memory 1002.

[0303] Processor 1001 may include one or more processing cores, such as a 4-core processor, a 9-core processor, etc. Processor 1001 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). Processor 1001 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 1001 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, processor 1001 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0304] The memory 1002 may include one or more computer-readable storage media, which may be non-transitory. The memory 1002 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the non-transitory computer-readable storage media in the memory 1002 are used to store at least one program code, which is executed by the processor 1001 to implement the signal processing method or fault detection method provided in the method embodiments of this disclosure.

[0305] In some embodiments, the terminal 1000 may also optionally include a peripheral device interface 1003 and at least one peripheral device. The processor 1001, memory 1002, and peripheral device interface 1003 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 1003 via a bus, signal line, or circuit board. Specifically, the peripheral device includes at least one of the following: a radio frequency circuit 1004, a display screen 1005, a camera assembly 1006, an audio circuit 1007, and a power supply 1008.

[0306] Peripheral device interface 1003 can be used to connect at least one I / O (Input / Output) related peripheral device to processor 1001 and memory 1002. In some embodiments, processor 1001, memory 1002 and peripheral device interface 1003 are integrated on the same chip or circuit board; in some other embodiments, any one or two of processor 1001, memory 1002 and peripheral device interface 1003 can be implemented on separate chips or circuit boards, which is not limited in this embodiment.

[0307] The radio frequency (RF) circuit 1004 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The RF circuit 1004 communicates with communication networks and other communication devices via electromagnetic signals. The RF circuit 1004 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals back into electrical signals. Optionally, the RF circuit 1004 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, etc. The RF circuit 1004 can communicate with other terminals via at least one wireless communication protocol. This wireless communication protocol includes, but is not limited to: metropolitan area networks (MANs), various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area networks (WLANs), and / or WiFi (Wireless Fidelity) networks. In some embodiments, the RF circuit 1004 may also include circuitry related to NFC (Near Field Communication), which is not limited in this disclosure.

[0308] Display screen 1005 is used to display a UI (User Interface). The UI may include graphics, text, icons, videos, and any combination thereof. When display screen 1005 is a touch display screen, it also has the ability to collect touch signals on or above its surface. These touch signals can be input as control signals to processor 1001 for processing. In this case, display screen 1005 can also be used to provide virtual buttons and / or a virtual keyboard, also known as soft buttons and / or a soft keyboard. In some embodiments, there may be one display screen 1005, serving as the front panel of terminal 1000; in other embodiments, there may be at least two display screens, respectively disposed on different surfaces of terminal 1000 or in a folded design; in still other embodiments, display screen 1005 may be a flexible display screen, disposed on a curved or folded surface of terminal 1000. Furthermore, display screen 1005 may also be configured as a non-rectangular, irregular shape, i.e., a non-rectangular screen. The display screen 1005 can be made of materials such as LCD (Liquid Crystal Display) and OLED (Organic Light-Emitting Diode).

[0309] The camera assembly 1006 is used to acquire images or videos. Optionally, the camera assembly 1006 includes a front-facing camera and a rear-facing camera. Typically, the front-facing camera is located on the front panel of the terminal, and the rear-facing camera is located on the back of the terminal. In some embodiments, there are at least two rear-facing cameras, which are any one of a main camera, a depth-sensing camera, a wide-angle camera, and a telephoto camera, to achieve background blurring by fusion of the main camera and the depth-sensing camera, panoramic shooting by fusion of the main camera and the wide-angle camera, VR (Virtual Reality) shooting, or other fusion shooting functions. In some embodiments, the camera assembly 1006 may also include a flash. The flash can be a single-color temperature flash or a dual-color temperature flash. A dual-color temperature flash refers to a combination of a warm-light flash and a cool-light flash, which can be used for light compensation at different color temperatures.

[0310] The audio circuit 1007 may include a microphone and a speaker. The microphone is used to collect sound waves from the user and the environment, converting the sound waves into electrical signals that are input to the processor 1001 for processing, or input to the radio frequency circuit 1004 for voice communication. For stereo sound acquisition or noise reduction purposes, multiple microphones may be used, each positioned at a different location on the terminal 1000. The microphone may also be an array microphone or an omnidirectional microphone. The speaker is used to convert electrical signals from the processor 1001 or the radio frequency circuit 1004 into sound waves. The speaker may be a conventional diaphragm speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, it can convert electrical signals not only into audible sound waves but also into inaudible sound waves for purposes such as distance measurement. In some embodiments, the audio circuit 1007 may also include a headphone jack.

[0311] The power supply 1008 is used to power the various components in the terminal 1000. The power supply 1008 can be AC ​​power, DC power, a disposable battery, or a rechargeable battery. When the power supply 1008 includes a rechargeable battery, the rechargeable battery can support wired charging or wireless charging. The rechargeable battery can also be used to support fast charging technology.

[0312] Those skilled in the art will understand that Figure 10 The structure shown does not constitute a limitation on terminal 1000 and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0313] Figure 11 This is a schematic diagram of a server structure according to an embodiment of this application. The server 1100 can vary significantly due to different configurations or performance. It may include one or more Central Processing Units (CPUs) 1101 and one or more memories 1102. The memories 1102 are used to store executable program code, and the processors 1101 are configured to execute the executable program code to implement the signal processing method or fault detection method provided in the various method embodiments described above. Of course, the server 1100 may also have wired or wireless network interfaces, a keyboard, and input / output interfaces for input and output. The server 1100 may also include other components for implementing device functions, which will not be elaborated here.

[0314] In an exemplary embodiment, a computer-readable storage medium is also provided, which, when executed by a processor of a computer device, enables the computer device to perform the aforementioned signal processing method or fault detection method. Optionally, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, or optical data storage device, etc.

[0315] In an exemplary embodiment, a computer program product is also provided, the computer program product including a computer program that, when executed by a processor, implements the above-described signal processing method or fault detection method.

[0316] In some embodiments, the computer program product involved in this disclosure may be deployed and executed on a computer device, or on multiple computer devices located in one location, or on multiple computer devices distributed in multiple locations and interconnected through a communication network. Multiple computer devices distributed in multiple locations and interconnected through a communication network may constitute a blockchain system.

[0317] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims. All the above-described optional technical solutions can be combined in any way to form optional embodiments of this application, and will not be elaborated upon here.

[0318] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A signal processing method, characterized in that, The method includes: Multiple sound signals from the device to be monitored are acquired, and the multiple sound signals are obtained at multiple monitoring locations of the device to be monitored. Obtain the first component sequence of each of the plurality of sound signals, wherein the first component sequence of each sound signal includes a plurality of first signal components of the sound signal arranged according to frequency bands; For each component group, multiple first signal components within the component group are spliced ​​together to obtain the first component matrix of the component group. Each component group includes multiple first signal components located at the same position in each first component sequence. Singular value decomposition is performed on the first component matrix to obtain a first diagonal matrix. The diagonal of the first diagonal matrix includes multiple singular values, which are arranged in order of magnitude on the diagonal. The multiple singular values ​​are used to indicate the energy weight of different sound components in the multiple first signal components. Set the singular values ​​in the last preset position among the plurality of singular values ​​to zero to obtain the second diagonal matrix; The component matrix is ​​restored based on the second diagonal matrix to obtain the second component matrix. The multiple signal components separated from the second component matrix include the signal components corresponding to the multiple first signal components in the component group. Determine the mean value among the plurality of signal components to obtain the reference signal component of the component group; For each first signal component in the component group, a first filtering process is performed on the signal component corresponding to the first signal component based on the reference signal component of the component group to obtain a third filtered component. A second filtering process is performed on the signal component corresponding to the first signal component based on the reference signal component of the component group to obtain a fourth filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise. The third and fourth filtered components are fused to obtain the second signal component corresponding to the first signal component. Multiple second signal components of the multiple sound signals are fused to obtain the sound signal to be tested of the device to be monitored, and the sound signal to be tested is used to detect faults in the device to be monitored.

2. The method according to claim 1, characterized in that, The method further includes: For each component group, the mean value among multiple first signal components within the component group is determined to obtain the reference signal component of the component group. For each first signal component within the component group, a first filtering process is performed on the first signal component based on the reference signal component of the component group to obtain a first filtered component, and a second filtering process is performed on the first signal component based on the reference signal component of the component group to obtain a second filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise. The first filtered component and the second filtered component are fused to obtain the second signal component corresponding to the first signal component.

3. The method according to claim 2, characterized in that, Each filtered component includes signal values ​​at multiple times, and the fusion coefficient between the first filtered component and the second filtered component includes sub-fusion coefficients for the signal values ​​at each of the multiple times. The method further includes: For each time step, based on the target data of the previous time step, a sub-fusion coefficient of the signal value at that time step is determined. The target data includes at least one of the following: the sub-fusion coefficient of the signal value at the previous time step, the error of the first filter component and the second filter component at the previous time step, and the signal value of the first filter component at the previous time step. The error at the previous time step represents the error between the signal value of the filter component at the previous time step and the reference signal value at the previous time step. The filter component includes the first filter component and the second filter component.

4. The method according to claim 2, characterized in that, The first filtering process performed on the first signal component based on the reference signal component of the component group to obtain the first filtered component includes: Based on the reference signal components of the component group, the first filter coefficients corresponding to the first filtering process are determined, and the first filtering process is performed on the first signal component based on the first filter coefficients to obtain the first filtered component. The second filtering process performed on the first signal component based on the reference signal component of the component group to obtain the second filtered component includes: Based on the reference signal components of the component group, the second filter coefficients corresponding to the second filtering process are determined, and the first signal component is subjected to the second filtering process based on the second filter coefficients to obtain the second filtered component.

5. The method according to claim 1, characterized in that, The process of fusing multiple second signal components of the multiple sound signals to obtain the sound signal to be tested from the device under test includes: For each component group, multiple second signal components corresponding to the component group are fused to obtain the fused component corresponding to the component group. The fusion components corresponding to each component group are fused to obtain the sound signal to be tested from the device under test.

6. The method according to claim 1, characterized in that, The step of obtaining the first component sequence of each of the plurality of sound signals includes: The plurality of sound signals are decomposed to obtain the second component sequence of each of the plurality of sound signals; For each sound signal, remove the first signal component that is not in the first preset position in the second component sequence of the sound signal to obtain the first component sequence of the sound signal. The frequency band corresponding to the first signal component that is in the first preset position is higher than the frequency band corresponding to the second signal component.

7. A fault detection method, characterized in that, The method includes: Multiple sound signals from the device to be monitored are acquired, and the multiple sound signals are obtained at multiple monitoring locations of the device to be monitored. The plurality of sound signals are processed by the signal processing method according to any one of claims 1-6 to obtain the sound signal to be tested of the device to be monitored; Based on the sound signal to be tested, the device to be monitored is subjected to fault detection, and the fault detection result of the device to be monitored is obtained.

8. A signal processing apparatus, characterized in that, The device includes: The first acquisition module is used to acquire multiple sound signals from the device to be monitored, wherein the multiple sound signals are obtained at multiple monitoring locations of the device to be monitored. The second acquisition module is used to acquire the first component sequence of each of the plurality of sound signals, wherein the first component sequence of each sound signal includes a plurality of first signal components of the sound signal arranged according to frequency bands; The third acquisition module is used to splice multiple first signal components within each component group to obtain a first component matrix of the component group. Each component group includes multiple first signal components located at the same position in each first component sequence. The processing module is configured to perform singular value decomposition on the first component matrix to obtain a first diagonal matrix. The first diagonal matrix includes multiple singular values ​​arranged in ascending order on the diagonal, which indicate the energy weights of different sound components in the multiple first signal components. The module then sets the singular values ​​at the last preset position in the sequence to zero, resulting in a second diagonal matrix. Based on the second diagonal matrix, the component matrix is ​​reconstructed to obtain a second component matrix. The multiple signal components extracted from the second component matrix include the signal components corresponding to each of the multiple first signal components within the component group. The module then determines... The average value among the multiple signal components is used to obtain the reference signal component of the component group. For each first signal component within the component group, a first filtering process is performed on the signal component corresponding to the first signal component based on the reference signal component of the component group to obtain a third filtered component. A second filtering process is performed on the signal component corresponding to the first signal component based on the reference signal component of the component group to obtain a fourth filtered component. The first filtering process is used to remove stable noise, and the second filtering process is used to filter out burst noise. The third filtered component and the fourth filtered component are fused to obtain the second signal component corresponding to the first signal component. The fusion module is used to fuse multiple second signal components of the multiple sound signals to obtain the sound signal to be tested of the device under test, and the sound signal to be tested is used to detect faults in the device under test.

9. A fault detection device, characterized in that, The device includes: The acquisition module is used to acquire multiple sound signals from the device to be monitored, wherein the multiple sound signals are obtained at multiple monitoring locations of the device to be monitored; The processing module is used to process the plurality of sound signals using the signal processing method according to any one of claims 1-6 to obtain the sound signal to be tested of the device to be monitored; The detection module is used to perform fault detection on the device under monitoring based on the sound signal to be tested, and to obtain the fault detection result of the device under monitoring.

10. A computer device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the signal processing method as claimed in any one of claims 1 to 6 or the fault detection method as claimed in claim 7.

11. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of a computer device, the computer device is able to perform the signal processing method of any one of claims 1 to 6 or the fault detection method of claim 7.

12. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the signal processing method according to any one of claims 1 to 6 or the fault detection method according to claim 7.

Images