Mechanical Fault Diagnosis Methods and Systems
By acquiring and fusing sound and vibration data from multiple monitoring locations in GIS equipment, and calculating the signal-to-noise ratio and stability confidence, accurate fault diagnosis in complex environments is achieved, improving the accuracy of fault type identification and location.
Patent Information
- Application Number
- CN202611560968.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-10-10
- Publication Date
- 2026-11-10
AI Technical Summary
Existing methods for diagnosing mechanical faults in GIS equipment struggle to fully utilize sound and vibration information from different monitoring locations in complex interference environments, leading to inaccurate fault type matching and fault location.
By acquiring sound and vibration monitoring data from multiple monitoring locations of GIS equipment, coordinate feature sets are extracted, and signal-to-noise ratio confidence and stability confidence are calculated. Weighted fusion is then performed to determine the coordinate feature set of the target monitoring equipment. This data is then combined with fault type feature information for matching and fault location.
It improves the accuracy and reliability of fault type identification and fault location, and reduces the impact of noise interference and characteristic fluctuations on diagnostic results.
Smart Images

Figure CN123065745A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of fault diagnosis technology, and in particular to mechanical fault diagnosis methods and systems. Background Technology
[0002] Gas-insulated switchgear (GIS) typically comprises multiple internal components such as circuit breakers, disconnectors, and grounding switches. During operation, mechanical faults such as loosening, jamming, or abnormal contact in these internal components can generate abnormal sounds and vibrations. Therefore, sound or vibration signals can be collected at different locations within the GIS, and the mechanical faults can be identified by analyzing the characteristics of the collected signals.
[0003] Existing methods for diagnosing mechanical faults in GIS equipment typically rely primarily on vibration signals, or supplemented by sound signals. However, in actual operating environments, the distance between different monitoring locations and the monitored objects, signal propagation paths, and installation conditions vary. Furthermore, sound signals are easily affected by environmental noise, and vibration signals are susceptible to interference from adjacent equipment vibrations and structural transmission. Consequently, sound and vibration signals collected at the same monitoring location may exhibit different noise immunity and characteristic stability, and the data quality collected from different monitoring locations may also vary significantly.
[0004] Without differentiating the reliability of sound and vibration characteristics at each monitoring location, directly fusing these characteristics can lead to interference from monitoring data exhibiting reduced signal quality or abnormal fluctuations due to interference. This interference can negatively impact the fusion results, affecting fault type matching and fault location. Therefore, existing technologies struggle to fully utilize sound and vibration information from different monitoring locations in complex interference environments for accurate and reliable mechanical fault diagnosis of multiple monitored objects in GIS equipment. Summary of the Invention
[0005] In view of the shortcomings of the prior art described above, the purpose of this disclosure is to provide a mechanical fault diagnosis method and system to solve the problems in the related art.
[0006] This disclosure provides a mechanical fault diagnosis method applied to a GIS device, wherein the GIS device contains multiple monitoring objects. The method includes: in each acquisition cycle, acquiring a set of sound monitoring data and vibration monitoring data for each of the multiple monitoring devices on the GIS device, and extracting a coordinate feature set for each monitoring device. The coordinate feature set includes a sound feature set and a vibration feature set; wherein the sound feature set and the vibration feature set respectively include sound feature information and vibration feature information, and both include the coordinates of the monitoring location of the monitoring device and the monitoring object; acquiring a first signal-to-noise ratio confidence level and a first stability confidence level corresponding to the sound feature information of the same coordinate feature set of the monitoring device, and a second signal-to-noise ratio confidence level and a second stability confidence level corresponding to the vibration feature information; determining the sound feature confidence level corresponding to the monitoring location based on the first signal-to-noise ratio confidence level and the first stability confidence level, and determining the vibration feature confidence level corresponding to the monitoring location based on the second signal-to-noise ratio confidence level and the second stability confidence level. The confidence levels of the sound features and vibration features are positively correlated with their respective signal-to-noise ratio confidence and stability confidence, respectively. Based on the sound fusion weight and vibration fusion weight corresponding to each monitoring device coordinate feature set, the confidence levels of the sound features and vibration features corresponding to the same monitoring device coordinate feature set are weighted and fused to obtain a fusion confidence. The monitoring device coordinate feature sets whose fusion confidence reaches a preset confidence threshold are determined as target monitoring device coordinate feature sets, and associated feature information containing the sound feature information and vibration feature information associated with the target monitoring device coordinate feature sets is obtained. The associated feature information is matched with preset fault type feature information, and the fault type of the fault monitoring object is determined during matching. In the target monitoring device coordinate feature sets of each monitoring device of the fault monitoring object, a preset number of target monitoring device coordinate feature sets with the highest fusion confidence are determined, and the fault point coordinates on the fault monitoring object are located according to the monitoring position coordinates contained in the set of target monitoring device coordinate feature sets.
[0007] In the first aspect of the embodiment, before acquiring a set of sound monitoring data and vibration monitoring data from multiple monitoring devices on the GIS equipment in each acquisition cycle, and extracting the coordinate feature set of the monitoring devices respectively, the method further includes: determining multiple candidate installation locations of the sound signal acquisition unit and the vibration signal acquisition unit on the GIS equipment or its structural model; under multiple sets of environmental interference conditions, acquiring sound evaluation data and vibration evaluation data for multiple evaluation cycles at each of the candidate installation locations, and performing feature processing to obtain sound evaluation feature information and vibration evaluation feature information corresponding to each of the candidate installation locations; calculating the sound at the same candidate installation location within the multiple evaluation cycles. The fluctuation variance of the evaluation feature information and vibration evaluation feature information is assessed, and the fluctuation variance is normalized and then inverted to determine the sound evaluation stability confidence and vibration evaluation stability confidence corresponding to each candidate installation location. Based on the signal parameters in the sound evaluation data and vibration evaluation data corresponding to each candidate installation location, the signal quality evaluation result corresponding to each candidate installation location is determined. The candidate installation locations where the sound evaluation stability confidence, the vibration evaluation stability confidence, and the signal quality evaluation result all reach their respective preset screening thresholds are determined as available installation locations. Multiple locations whose monitoring ranges jointly cover each monitored object are selected from the available installation locations as the monitoring locations.
[0008] In a first aspect embodiment, determining a set of target monitoring device coordinate feature sets with the highest fusion confidence from the target monitoring device coordinate feature set of each monitoring device of the fault monitoring object, and locating the fault point coordinates on the fault monitoring object based on the monitoring location coordinates included in the set of target monitoring device coordinate feature sets, includes: determining a set of target monitoring device coordinate feature sets with the highest fusion confidence from the target monitoring device coordinate feature set associated with the same fault monitoring object and obtained by matching associated feature information, and using the monitoring location corresponding to each selected target monitoring device coordinate feature set as the location monitoring location; obtaining each location monitoring location pair The arrival time of the sound fault signal generated when the fault monitoring object malfunctions, as per the sound monitoring data; for any two positioning monitoring locations, the difference in the arrival time of the sound fault signal at the two positioning monitoring locations is calculated to obtain the corresponding sound signal arrival time difference; each sound signal arrival time difference is multiplied by the corresponding sound signal propagation speed to obtain multiple sets of distance differences, wherein each set of distance differences is the difference in distance from the fault point of the fault monitoring object to the corresponding two positioning monitoring locations; based on the coordinates of each positioning monitoring location and each set of distance differences, the spatial position coordinates that simultaneously satisfy the constraints of each set of distance differences are obtained, and these coordinates are used as the fault point coordinates of the fault monitoring object.
[0009] In the first aspect of the embodiment, the step of determining a set of target monitoring device coordinate feature sets with the highest fusion confidence from the target monitoring device coordinate feature set that is associated with the same fault monitoring object and whose associated feature information is matched to obtain the same fault type, and taking the monitoring position corresponding to each selected target monitoring device coordinate feature set as the location monitoring position, includes: matching the associated feature information corresponding to each target monitoring device coordinate feature set associated with the same fault monitoring object with the preset fault type feature information to obtain the fault type corresponding to each target monitoring device coordinate feature set, and dividing the target monitoring device coordinate feature sets with the same fault type into the same group;
[0010] Grouping a set of coordinate features of at least a preset number of the target monitoring devices is defined as a positioning group, and the monitoring positions corresponding to each set of coordinate features of the target monitoring devices within the same positioning group are used as a group of positioning monitoring positions, so as to jointly determine the coordinates of the corresponding fault point by utilizing the signal arrival time difference between each positioning monitoring position.
[0011] In the first aspect of the embodiment, a structural model pre-constructed based on the actual structural data of the GIS device is obtained, and the spatial range corresponding to each component of the fault monitoring object in the structural model is obtained; the coordinates of the fault point are compared with the spatial ranges corresponding to each component in the same coordinate system, and the component whose spatial range includes the coordinates of the fault point is determined as the fault part of the fault monitoring object.
[0012] In a first aspect embodiment, the step of acquiring the first signal-to-noise ratio (SNR) confidence level and the first stability confidence level corresponding to the sound feature information of the same monitoring device coordinate feature set, and the second SNR confidence level and the second stability confidence level corresponding to the vibration feature information; determining the sound feature confidence level corresponding to the monitoring location based on the first SNR confidence level and the first stability confidence level, and determining the vibration feature confidence level corresponding to the monitoring location based on the second SNR confidence level and the second stability confidence level, includes: calculating the SNR corresponding to the sound feature information and the vibration feature information of the same monitoring device coordinate feature set, and respectively... The signal-to-noise ratio (SNR) is normalized to obtain the first SNR confidence level and the second SNR confidence level. The amplitude-frequency characteristic fluctuation variances of the sound feature information and the vibration feature information in the same coordinate feature set of the monitoring device are calculated respectively, and the amplitude-frequency characteristic fluctuation variances are normalized and their correlation is inverted to obtain the first stability confidence level and the second stability confidence level. Corresponding weighting coefficients are applied to the first SNR confidence level and the first stability confidence level, and then summed to obtain the sound feature confidence level. Similarly, corresponding weighting coefficients are applied to the second SNR confidence level and the second stability confidence level, and then summed to obtain the vibration feature confidence level.
[0013] In an embodiment of the first aspect, the sound fusion weight is positively correlated with the ratio of the sound feature confidence to the vibration feature confidence, and the vibration fusion weight is negatively correlated with the ratio of the sound feature confidence to the vibration feature confidence. The calculation method of the sound fusion weight and the vibration fusion weight includes: determining the sound fusion weight and the vibration fusion weight as the proportions of the sound feature confidence and the vibration feature confidence to the sum of the sound feature confidence and the vibration feature confidence, respectively.
[0014] In a first aspect embodiment, the preset fault type feature information includes: sound type fault type feature information and vibration type fault type feature information. Matching the associated feature information with the preset fault type feature information and determining the fault type of the fault monitoring object during matching includes: comparing the confidence levels of sound features and vibration features corresponding to the target monitoring device coordinate feature set, and determining the matching priority of each of the sound feature information and the vibration feature information; wherein the matching priority of each feature information is positively correlated with its corresponding feature confidence level; and matching the feature information with the higher matching priority among the sound feature information and the vibration feature information. The dominant matching feature information is determined; the dominant matching feature information is matched with the corresponding fault type feature information to obtain a matching result; in response to the matching result satisfying the matching condition, the fault type corresponding to the matching result is determined as the fault type of the fault monitoring object; in response to the matching result not satisfying the matching condition, from the set of coordinate features of the target monitoring devices associated with the same monitoring object and not yet matched for fault type, the next set of coordinate features of the target monitoring devices is selected in descending order of fusion confidence, and the matching is performed again based on the sound feature information and vibration feature information corresponding to the next set of coordinate features of the target monitoring devices.
[0015] In an embodiment of the first aspect, the method further includes: simulating mechanical failures of each of the monitored objects individually in the structural model of the GIS equipment, respectively, and obtaining sound evaluation feature information and vibration evaluation feature information corresponding to each available installation location when each of the monitored objects individually experiences any of the preset failure types; fusing the sound evaluation feature information and vibration evaluation feature information obtained when the same available installation location experiences the same preset failure type when the same monitored object individually, to obtain evaluation association feature information; calculating the feature difference value between each monitored object and each other monitored object under the corresponding preset failure type, respectively, and determining the minimum value among the feature difference values as the object distinguishability of the monitored object at the available installation location; selecting at least one available installation location of the monitored object as a target installation location from each of the available installation locations according to the priority order of the object distinguishability from large to small, and taking the target installation location corresponding to each monitored object as the monitoring location.
[0016] The second aspect of this disclosure provides a mechanical fault diagnosis system applied to a GIS device including multiple monitoring objects. The system includes: multiple monitoring devices, each set at its corresponding monitoring location, and each monitoring location associated with a corresponding monitoring object; wherein each monitoring device includes a sound acquisition unit and a vibration acquisition unit, the sound acquisition unit being used to acquire sound monitoring data at the corresponding monitoring location, and the vibration acquisition unit being used to acquire vibration monitoring data at the corresponding monitoring location; and a processing terminal, which is communicatively connected to the multiple sound acquisition units and the multiple vibration acquisition units, and configured to execute the mechanical fault diagnosis method as described in any of the first aspects.
[0017] As described above, this disclosure provides a mechanical fault diagnosis method and system. The method includes: acquiring, in each acquisition cycle, a set of sound monitoring data and vibration monitoring data from multiple monitoring devices on the GIS equipment, and extracting a coordinate feature set of the monitoring devices, wherein the coordinate feature set of the monitoring devices includes a sound feature set and a vibration feature set; wherein the sound feature set and the vibration feature set respectively include sound feature information and vibration feature information, and both include the coordinates of the monitoring location where the monitoring device is located and the monitoring object; acquiring, respectively, a first signal-to-noise ratio confidence score and a first stability confidence score corresponding to the sound feature information of the same coordinate feature set of the monitoring devices, and a second signal-to-noise ratio confidence score and a second stability confidence score corresponding to the vibration feature information; determining the sound feature confidence score corresponding to the monitoring location based on the first signal-to-noise ratio confidence score and the first stability confidence score, and determining the vibration feature confidence score corresponding to the monitoring location based on the second signal-to-noise ratio confidence score and the second stability confidence score; wherein the sound feature confidence score... The confidence levels of the vibration features and the sound feature confidence level are positively correlated with their respective signal-to-noise ratio confidence level and stability confidence level. Based on the sound fusion weight and vibration fusion weight corresponding to each of the monitoring device coordinate feature sets, the sound feature confidence level and the vibration feature confidence level corresponding to the same monitoring device coordinate feature set are weighted and fused to obtain a fusion confidence level. The monitoring device coordinate feature set whose fusion confidence level reaches a preset confidence threshold is determined as the target monitoring device coordinate feature set, and the associated feature information containing the sound feature information and vibration feature information associated in the target monitoring device coordinate feature set is obtained. The associated feature information is matched with the preset fault type feature information, and the fault type of the fault monitoring object is determined during the matching. In the target monitoring device coordinate feature set of each monitoring device of the fault monitoring object, a preset number of target monitoring device coordinate feature sets with the highest fusion confidence level are determined, and the fault point coordinates on the fault monitoring object are located according to the monitoring position coordinates contained in the target monitoring device coordinate feature set. This disclosure constructs a monitoring device coordinate feature set by associating the sound and vibration feature information of the same monitoring device with the monitoring location coordinates and the monitoring object, thus giving the feature set a clear location and object source. By comprehensively determining the signal-to-noise ratio confidence and stability confidence of the sound and vibration feature information, and filtering the target monitoring device coordinate feature set based on the fused confidence, the impact of noise interference and monitoring data with large feature fluctuations on fault type matching can be reduced. After determining the fault type of the fault monitoring object, the target monitoring device coordinate feature set with the highest fused confidence is further selected for fault point location, thereby improving the accuracy and reliability of fault type identification and fault point location. Attached Figure Description
[0018] Figure 1A schematic diagram illustrating a scenario application of the mechanical fault diagnosis method in one embodiment of this disclosure is shown.
[0019] Figure 2 A flowchart illustrating a mechanical fault diagnosis method according to an embodiment of this disclosure is shown.
[0020] Figure 3 A flowchart illustrating the calculation method of sound feature confidence and vibration feature confidence in one embodiment of this disclosure is shown.
[0021] Figure 4 A flowchart illustrating a method for determining a fault type according to an embodiment of this disclosure is shown.
[0022] Figure 5 A flowchart illustrating a method for determining fault point coordinates in one embodiment of this disclosure is shown.
[0023] Figure 6 A flowchart illustrating a method for determining a monitoring location according to an embodiment of this disclosure is shown.
[0024] Figure 7 A flowchart illustrating a monitoring location selection method according to an embodiment of this disclosure is shown.
[0025] Figure 8 An example diagram illustrating the selection of monitoring locations based on object distinguishability in one embodiment of this disclosure is shown.
[0026] Figure 9 A schematic diagram of the structure of an electronic device according to an embodiment of the present disclosure is shown. Detailed Implementation
[0027] The following specific examples illustrate the implementation of this disclosure. Those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed herein. This disclosure can also be implemented or applied through other different specific embodiments, and various details in this disclosure can also be modified or changed according to different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this disclosure can be combined with each other.
[0028] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings, so that those skilled in the art to which this disclosure pertains can readily implement it. This disclosure may be embodied in many different forms and is not limited to the embodiments described herein.
[0029] In the description of this disclosure, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this disclosure. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or a group of embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this disclosure, as well as the features of those different embodiments or examples.
[0030] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this disclosure, "a set" means two or more, unless otherwise explicitly specified.
[0031] For the purpose of clarity, devices unrelated to the description are omitted, and the same or similar components throughout the specification are given the same reference numerals.
[0032] Throughout this specification, when it is said that a device is "connected" to another device, this includes not only "direct connection" but also "indirect connection" by placing other components in between. Furthermore, when it is said that a device "comprises" a certain constituent element, unless otherwise stated otherwise, this does not exclude other constituent elements, but rather implies that other constituent elements may be included.
[0033] While the terms first, second, etc., are used in some examples herein to describe various elements, these elements should not be limited by these terms. These terms are used only to distinguish one element from another. For example, descriptions such as first interface and second interface, etc. Furthermore, as used herein, the singular forms “a,” “an,” and “the” are intended to also include the plural forms unless the context indicates otherwise. It should be further understood that the terms “comprising,” “including,” indicate the presence of the stated feature, step, operation, element, module, item, kind, and / or group, but do not exclude the presence, occurrence, or addition of one or more other features, steps, operations, elements, modules, items, kinds, and / or groups. The terms “or” and “and / or” as used herein are interpreted as inclusive, or mean any one or any combination thereof. Thus, “A, B, or C” or “A, B, and / or C” means “any one of: A; B; C; A and B; A and C; B and C; A, B, and C.” Exceptions to this definition occur only when combinations of elements, functions, steps, or operations are inherently mutually exclusive in some way.
[0034] The technical terms used herein are for reference only to specific embodiments and are not intended to limit the scope of this disclosure. The singular form used herein includes the plural form unless the statement explicitly states otherwise. The word "comprising" as used in this specification means to specify a particular characteristic, region, integer, step, operation, element, and / or component, and does not exclude the presence or addition of other characteristics, regions, integers, steps, operations, elements, and / or components.
[0035] Although not explicitly defined, all terms, including technical and scientific terms used herein, shall have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. Terms defined in commonly used dictionaries shall be further interpreted as having a meaning consistent with the relevant technical literature and the message of the present disclosure, and shall not be over-interpreted as having an ideal or overly formulaic meaning unless otherwise defined.
[0036] Gas-insulated switchgear (GIS) typically comprises multiple internal components such as circuit breakers, disconnectors, and grounding switches. During operation, mechanical faults such as loosening, jamming, or abnormal contact in these internal components can generate abnormal sounds and vibrations. Therefore, sound or vibration signals can be collected at different locations within the GIS, and the mechanical faults can be identified by analyzing the characteristics of the collected signals.
[0037] Existing methods for diagnosing mechanical faults in GIS equipment typically rely primarily on vibration signals, or supplemented by sound signals. However, in actual operating environments, the distance between different monitoring locations and the monitored objects, signal propagation paths, and installation conditions vary. Furthermore, sound signals are easily affected by environmental noise, and vibration signals are susceptible to interference from adjacent equipment vibrations and structural transmission. Consequently, sound and vibration signals collected at the same monitoring location may exhibit different noise immunity and characteristic stability, and the data quality collected from different monitoring locations may also vary significantly.
[0038] Without differentiating the reliability of acoustic and vibration characteristics at each monitoring location, directly fusing these characteristics can lead to interference from monitoring data with low signal quality or large fluctuations. This can affect the determination of effective monitoring locations and the matching of fault types. Therefore, existing technologies struggle to fully utilize acoustic and vibration information from different monitoring locations in complex interference environments to accurately and reliably diagnose mechanical faults in multiple monitored objects within GIS equipment.
[0039] In view of this, some embodiments of this disclosure may provide a mechanical fault diagnosis method. As an example, the mechanical fault diagnosis method may be executed by a mechanical fault diagnosis system and applied to the following scenarios.
[0040] refer to Figure 1 The diagram illustrates a scenario application of the mechanical fault diagnosis method. (Example:) Figure 1 As shown, a mechanical fault diagnosis system is applied to a GIS device 100, which contains multiple monitoring objects 110. Exemplarily, the monitoring objects 110 may include at least one of a voltage transformer 111, a disconnecting switch 112, a circuit breaker 113, and a current transformer 114. In actual arrangements, different monitoring objects 110 may be respectively located in different chambers, adjacent chambers, or adjacent functional modules within the GIS device 100.
[0041] The mechanical fault diagnosis system includes multiple monitoring devices and a processing terminal 300. The multiple monitoring devices are respectively located at their corresponding monitoring positions, and each monitoring position is associated with a corresponding monitoring object 110 and is located near the corresponding monitoring object 110 inside the GIS equipment. Each monitoring device includes a sound acquisition unit 210 and a vibration acquisition unit 220. The sound acquisition unit 210 is used to acquire sound monitoring data at the corresponding monitoring position, and the vibration acquisition unit 220 is used to acquire vibration monitoring data at the corresponding monitoring position. The processing terminal 300 is communicatively connected to the sound acquisition unit 210 and the vibration acquisition unit 220 in each of the monitoring devices and is configured to execute the mechanical fault diagnosis method of this disclosure embodiment.
[0042] refer to Figure 2 , Figure 2 This embodiment illustrates a flowchart of the mechanical fault diagnosis method. The method is applied to a GIS device, which contains multiple monitoring objects. The method includes the following steps:
[0043] S110: In each acquisition cycle, acquire a set of sound monitoring data and vibration monitoring data from each of the multiple monitoring devices on the GIS equipment, and extract the coordinate feature set of the monitoring devices, which includes the sound feature set and the vibration feature set.
[0044] The sound feature set and vibration feature set include sound feature information and vibration feature information, respectively, and both include the coordinates of the monitoring location of the monitoring device and the monitoring object.
[0045] In some embodiments, the sound feature information includes the amplitude-frequency characteristics corresponding to the sound monitoring data, and the vibration feature information includes the amplitude-frequency characteristics corresponding to the vibration monitoring data. The amplitude-frequency characteristics of the sound monitoring data are obtained by frequency domain transformation of the sound monitoring data, and the amplitude-frequency characteristics of the vibration monitoring data are obtained by frequency domain transformation of the vibration monitoring data. For example, the frequency domain transformation can employ Fourier transform, short-time Fourier transform, or other transformation methods capable of obtaining frequency domain characteristics. In some embodiments, a GIS equipment monitoring model can be pre-constructed. The GIS equipment monitoring model is used to record the installation coordinates of each sound acquisition unit and vibration acquisition unit, the corresponding monitoring positions, and the monitoring objects associated with each monitoring position. The processing end can, based on the GIS equipment monitoring model, associate and store each set of sound feature information and vibration feature information with the corresponding monitoring position coordinates to construct a sound feature set and a vibration feature set. For example, if the electrical equipment inside the GIS enclosure at the first monitoring position is a disconnect switch, and the electrical equipment inside the GIS enclosure at the second monitoring position is a circuit breaker, then the first sound feature set is denoted as... , The coordinates of the first monitoring location The first set of sound features is the sound characteristic information of the first monitoring location, and the second set of sound features is denoted as... , The coordinates of the second monitoring location. This refers to the acoustic characteristics of the second monitoring location.
[0046] S120: Obtain the first signal-to-noise ratio confidence level and the first stability confidence level corresponding to the sound feature information of the same monitoring device coordinate feature set, and the second signal-to-noise ratio confidence level and the second stability confidence level corresponding to the vibration feature information; determine the sound feature confidence level corresponding to the monitoring location based on the first signal-to-noise ratio confidence level and the first stability confidence level, and determine the vibration feature confidence level corresponding to the monitoring location based on the second signal-to-noise ratio confidence level and the second stability confidence level.
[0047] The confidence level of the sound features is positively correlated with the confidence level of the first signal-to-noise ratio and the confidence level of the first stability, and the confidence level of the vibration features is positively correlated with the confidence level of the second signal-to-noise ratio and the confidence level of the second stability.
[0048] In some embodiments, reference Figure 3 , Figure 3 A flowchart illustrating the calculation methods for sound feature confidence and vibration feature confidence, S120 includes:
[0049] S121: Calculate the signal-to-noise ratio (SNR) corresponding to the sound feature information and the vibration feature information respectively, and normalize the SNR to obtain the first SNR confidence level. and the second signal-to-noise ratio confidence level .
[0050] In some embodiments, a frequency domain filtering method can be used to distinguish the effective signal components and noise signal components in the corresponding monitoring data based on the amplitude-frequency characteristics of the sound monitoring data and the vibration monitoring data, and to determine the signal-to-noise ratio (SNR) of the sound monitoring data and the vibration monitoring data respectively based on the power ratio of the effective signal components to the noise signal components. The SNR is then normalized to obtain the first SNR confidence level. and the second signal-to-noise ratio confidence level .
[0051] Specifically, firstly, based on the amplitude-frequency characteristics corresponding to the sound monitoring data, one or more effective sound frequency bands can be determined, and frequency bands outside the effective sound frequency bands can be defined as sound noise frequency bands. Similarly, based on the amplitude-frequency characteristics corresponding to the vibration monitoring data, one or more effective vibration frequency bands can be determined, and frequency bands outside the effective frequency bands can be defined as vibration noise frequency bands.
[0052] Secondly, for sound characteristic information, the signal power of the sound monitoring data within the effective sound frequency band is determined as the effective sound signal power, and the signal power of the sound monitoring data within the sound noise frequency band is determined as the sound noise power. The signal-to-noise ratio (SNR) of the sound monitoring data is then determined based on the ratio between the effective sound signal power and the sound noise power. Similarly, for vibration characteristic information, the signal power of the vibration monitoring data within the effective vibration frequency band is determined as the effective vibration signal power, and the signal power of the vibration monitoring data within the vibration noise frequency band is determined as the vibration noise power. The SNR of the vibration monitoring data is then determined based on the ratio between the effective vibration signal power and the vibration noise power.
[0053] Finally, the signal-to-noise ratios (SNRs) corresponding to the sound monitoring data and the vibration monitoring data are normalized respectively, and the normalization results are determined as the first SNR confidence level. and the second signal-to-noise ratio confidence level The signal-to-noise ratio (SNR) is positively correlated with the corresponding SNR confidence level.
[0054] S122: Calculate the amplitude-frequency characteristic fluctuation variance of the sound feature information and the vibration feature information respectively, and then normalize and invert the correlation of the amplitude-frequency characteristic fluctuation variance to obtain the first stability confidence. and the second stability confidence level The amplitude-frequency characteristic fluctuation variance is used to characterize the degree of fluctuation of the corresponding amplitude-frequency characteristic between different acquisition periods.
[0055] In some embodiments, sound feature information and vibration feature information corresponding to the same monitoring location in multiple consecutive acquisition cycles can be obtained, and sound amplitude-frequency features and vibration amplitude-frequency features corresponding to each acquisition cycle can be extracted respectively.
[0056] Specifically, for sound feature information, the amplitude changes at corresponding frequency points within multiple consecutive acquisition cycles can be compared for one or more effective sound frequency bands. The amplitude-frequency characteristic fluctuation variance corresponding to the sound feature information is determined based on the deviation between the amplitude value and the average amplitude value for each acquisition cycle. Similarly, for vibration feature information, the amplitude changes at corresponding frequency points within multiple consecutive acquisition cycles can be compared for one or more effective vibration frequency bands. The amplitude-frequency characteristic fluctuation variance corresponding to the vibration feature information is determined based on the deviation between the amplitude value and the average amplitude value for each acquisition cycle. Further, normalized inverse correlation processing is performed on the amplitude-frequency characteristic fluctuation variances corresponding to the sound feature information and the vibration feature information, respectively. The normalized inverse correlation results are then determined as the first stability confidence level. and the second stability confidence level The smaller the variance of the amplitude-frequency characteristic fluctuation, the more stable the corresponding amplitude-frequency characteristic is across different acquisition periods, and the higher the stability confidence obtained after normalization and correlation inversion. The larger the variance of the amplitude-frequency characteristic fluctuation, the more obvious the fluctuation of the corresponding amplitude-frequency characteristic, and the lower the stability confidence obtained after normalization and correlation inversion. The amplitude-frequency characteristic fluctuation variance is negatively correlated with the stability confidence.
[0057] In some examples, the normalized inverse correlation result D can be expressed as:
[0058]
[0059] in, This represents the first or second stability confidence level. This refers to the variance of the amplitude-frequency characteristic fluctuation corresponding to sound or vibration feature information. and These are the preset maximum variance and the preset minimum variance, respectively. The preset maximum variance... and preset minimum fluctuation variance Based on historical monitoring data or sample data, when D exceeds the normalization interval, D can be restricted to the normalization interval.
[0060] S123: Confidence level of the first signal-to-noise ratio and the first stability confidence Apply the corresponding weighting coefficients respectively , Then sum them up to obtain the confidence level of the sound features. and the confidence level of the second signal-to-noise ratio. and the second stability confidence level Apply the corresponding weighting coefficients , And sum them up to obtain the confidence level of the vibration feature. .
[0061] Among them, the confidence level of the sound features The calculation formula can be expressed as: The confidence level of the vibration characteristics The calculation formula can be expressed as:
[0062] (1)
[0063] Therefore, the signal-to-noise ratio (SNR) confidence level reflects the degree of noise interference in the current monitoring data, while the stability confidence level reflects the consistency of the corresponding amplitude-frequency characteristics across different acquisition cycles. By combining the SNR and stability confidence levels to determine the feature confidence level, we can avoid evaluating feature reliability solely based on the SNR or solely based on cross-cycle stability, thus improving the ability of feature confidence levels to characterize the actual quality of monitoring data.
[0064] S130: Based on the sound fusion weights and vibration fusion weights corresponding to the coordinate feature sets of each monitoring device, determine the confidence level of the sound features corresponding to the coordinate feature set of the same monitoring device. and the confidence level of the vibration characteristics Perform weighted fusion to obtain the fusion confidence level. To achieve the fusion confidence level The set of coordinate features of monitoring devices that reaches the preset confidence threshold is determined as the set of coordinate features of target monitoring devices, and associated feature information containing the sound feature information and vibration feature information associated with the set of coordinate features of target monitoring devices is obtained.
[0065] In some embodiments, the sound fusion weight is positively correlated with the ratio of the sound feature confidence level to the vibration feature confidence level, and the vibration fusion weight is negatively correlated with the ratio of the sound feature confidence level to the vibration feature confidence level. Specifically, the calculation method for the sound fusion weight and the vibration fusion weight includes: determining the sound fusion weight and the vibration fusion weight as the proportions of the sound feature confidence level and the vibration feature confidence level to the sum of the sound feature confidence level and the vibration feature confidence level, respectively.
[0066] For example, the sound fusion weights can be represented as The vibration fusion weight can be expressed as fusion confidence The calculation formula can be expressed as:
[0067] (2)
[0068] Therefore, when the confidence levels of sound features and vibration features change, the corresponding fusion weights can be adjusted accordingly, so that modes with higher feature confidence levels make a greater contribution to the fusion confidence level, thereby reducing the impact of low-confidence modes on the selection results of the target monitoring equipment coordinate feature set.
[0069] In some embodiments, the method for fusing associated feature information of the target monitoring device coordinate feature set includes concatenating the sound feature information and the vibration feature information according to a preset feature arrangement order. Specifically, the sound feature information and the vibration feature information are normalized respectively, and the normalized sound feature information and vibration feature information are combined into a joint feature vector according to a preset arrangement order to obtain the associated feature information. The preset arrangement order can be sound feature information first, vibration feature information second, or a fixed arrangement order can be adopted according to actual needs. Through feature concatenation, a joint sound-vibration feature for fault type matching can be formed while retaining the respective feature dimensions of the sound feature information and vibration feature information. For example, if the normalized sound feature information includes multiple sound feature components and the vibration feature information includes multiple vibration feature components, the feature components can be arranged in the order of sound feature components first, vibration feature components second, to form a joint feature vector that simultaneously represents sound and vibration features.
[0070] In other embodiments, the sound feature information and the vibration feature information can be feature-encoded according to a preset vibration-sound fusion rule, and the resulting encoding results can be fused to obtain joint feature information that simultaneously represents the sound feature information and the vibration feature information, which serves as the associated feature information corresponding to the coordinate feature set of the target monitoring device.
[0071] Therefore, sound feature information and vibration feature information can be associated in the same feature representation, and the complementary representation of mechanical faults by the two modes can be used to improve the ability to distinguish between different fault types.
[0072] S140: Match the associated feature information with the preset fault type feature information, and determine the fault type of the fault monitoring object during the matching process.
[0073] The associated feature information includes interrelated sound feature information and vibration feature information. When performing fault type matching, the feature confidence levels corresponding to the sound feature information and the vibration feature information can be compared, and the feature information with the higher feature confidence level is preferentially matched with the corresponding fault type feature information. The fault monitoring object is the monitoring object that has experienced a fault, determined based on the matching results.
[0074] In some embodiments, the fault type is a periodic mechanical fault type, and the corresponding fault signal has a periodic repetition characteristic. The periodic repetition characteristic refers to the corresponding fault responses in the sound fault signal and vibration fault signal repeating themselves at a certain period. During the repetition phase of the fault response, multiple consecutive acquisition cycles under the same or similar operating conditions are selected. Each acquisition cycle uses the same sampling parameters, acquisition duration, and frequency domain transformation processing method, and the acquisition period covers an integer number of complete cycles of the fault signal to compare the amplitude-frequency characteristics corresponding to each acquisition cycle. The stability confidence level is used to characterize the consistency of the above amplitude-frequency characteristics across different acquisition cycles. Therefore, for fault signals with periodic repetition characteristics, under the above acquisition conditions, the amplitude-frequency characteristics corresponding to different acquisition cycles can remain similar.
[0075] In some embodiments, the preset fault type database stores corresponding sound fault type feature information and vibration fault type feature information for each fault type. The sound fault type feature information and the vibration fault type feature information may include one or more combinations of fault feature amplitude, fault feature frequency, or other feature parameters used to characterize mechanical faults under the corresponding fault type.
[0076] For example, the monitoring object may include at least one of the following: GIS equipment casing, circuit breaker, disconnector, voltage transformer and current transformer. The preset fault type and corresponding sound fault type feature information and vibration fault type feature information corresponding to each monitoring object may include at least one of the following.
[0077] For GIS equipment enclosures, preset fault types may include enclosure loosening. Enclosure loosening can be caused by loose or detached flange bolts, leading to an increase in the overall vibration amplitude of the enclosure during GIS equipment operation and generating abnormal vibration components at the power frequency and its harmonic frequencies. Corresponding acoustic fault characteristics may include regular low-frequency abnormal sounds and increased low-frequency sound energy. Vibration fault characteristics may include higher vibration amplitude in the loose area compared to the normal area and an increase in power frequency harmonic components.
[0078] For circuit breakers, preset fault types can include component wear and poor contact. The acoustic fault characteristics corresponding to component wear can include persistent high-frequency abnormal noise and an increased proportion of high-frequency sound energy. The vibration fault characteristics corresponding to poor contact can include persistent micro-vibrations at the contact points and abnormal pulse components in the time-domain waveform of the vibration signal.
[0079] For disconnecting switches, the preset fault types can include contact failure. The audible fault characteristics corresponding to contact failure can include intermittent abnormal noise and abnormal pulse components in the sound signal.
[0080] For voltage transformers, preset fault types can include core loosening and winding fastening defects. The acoustic and vibration fault characteristics corresponding to core loosening can include persistent low-frequency abnormal noise and increased core vibration amplitude. The vibration fault characteristics corresponding to winding fastening defects can include an increased proportion of high-frequency vibration components and increased energy in the high-frequency band of the vibration signal.
[0081] For current transformers, preset fault types can include loose mounting bolts and wear of the core insulation structure. Vibration fault characteristics corresponding to loose mounting bolts may include an increase in the overall vibration amplitude of the current transformer, with the vibration amplitude near the loose location being higher than at other locations; acoustic fault characteristics corresponding to wear of the core insulation structure may include continuous abnormal sounds. The corresponding specific frequency range, amplitude threshold, and energy threshold can be predetermined based on the normal operation data and fault sample data of different models of GIS equipment.
[0082] In some embodiments, reference Figure 4 , Figure 4 A flowchart illustrating a method for determining fault types is provided. The fault type characteristic information includes: fault type characteristic information for sound types and fault type characteristic information for vibration types. The method includes the following steps:
[0083] S141: Compare the confidence levels of the sound features and vibration features corresponding to the coordinate feature set of the target monitoring device, and determine the matching priority of the sound feature information and the vibration feature information respectively.
[0084] The matching priority of each feature information is positively correlated with its corresponding feature confidence level.
[0085] S142: The feature information with higher matching priority among the sound feature information and the vibration feature information is determined as the dominant matching feature information.
[0086] Specifically, when the matching priority of sound feature information is high, sound feature information is determined as the dominant matching feature information. When the matching priority of vibration feature information is high, vibration feature information is determined as the dominant matching feature information.
[0087] S143: Match the dominant matching feature information with the corresponding fault type feature information to obtain the matching result.
[0088] Specifically, when the dominant matching feature information is sound feature information, it is matched with the fault type feature information of the sound type corresponding to each fault type; when the dominant matching feature information is vibration feature information, it is matched with the fault type feature information of the vibration type corresponding to each fault type.
[0089] In some embodiments, the cosine similarity between the feature vector corresponding to the dominant matching feature information and the feature vector corresponding to the fault type feature information of each corresponding type can be calculated, and each cosine similarity can be used as the matching degree of the corresponding fault type. The matching result may include the fault type with the highest matching degree and its corresponding matching degree. In addition to cosine similarity, Euclidean distance, correlation coefficient, or a trained feature classification model can also be used to obtain the matching result.
[0090] S144: In response to the matching result satisfying the matching condition, the fault type corresponding to the matching result is determined as the fault type of the fault monitoring object.
[0091] Specifically, the matching condition may include the highest matching degree in the matching results being greater than a preset matching threshold.
[0092] S145: In response to the matching result not meeting the matching condition, from the set of coordinate features of the target monitoring devices associated with the same monitoring object and not yet matched for fault type, select the next set of coordinate features of the target monitoring devices in descending order of fusion confidence, and re-match according to the sound feature information and vibration feature information corresponding to the next set of coordinate features of the target monitoring devices.
[0093] Specifically, for the next target monitoring device coordinate feature set, the confidence levels of the sound features and vibration features are re-compared to determine the corresponding dominant matching feature information, and steps S141 to S145 are re-executed. If there is no target monitoring device coordinate feature set that has not yet been matched for fault type, the fault type matching of the monitored object in the current acquisition cycle ends, the determined fault type is not output, and monitoring continues in the next acquisition cycle.
[0094] Therefore, by prioritizing the use of feature information with higher confidence for fault type matching, and when the current target monitoring device coordinate feature set fails to match, calling the next target monitoring device coordinate feature set in descending order of fusion confidence, we can prioritize the use of monitoring features with higher reliability to complete fault type judgment, reduce the interference of low-confidence features and low fusion confidence monitoring data on the matching results, and improve the efficiency and reliability of fault type judgment.
[0095] In some embodiments, after determining the fault type of the monitored object, vibration amplitude characteristics from the vibration characteristic information used to determine the fault type can be obtained, and the vibration amplitude characteristics can be compared with a preset severity threshold corresponding to the fault type. The severity of the mechanical fault can be determined based on the comparison result. For example, the severity of the mechanical fault can be classified into at least one of minor, moderate, and severe according to the degree to which the vibration amplitude characteristics exceed a normal vibration amplitude reference. Different fault types can correspond to different preset severity thresholds to avoid affecting the judgment result of fault severity due to differences between the normal vibration amplitude and the fault vibration amplitude corresponding to different fault types.
[0096] In some embodiments, after determining the fault type of the monitored object corresponding to each target monitoring device coordinate feature set, a corresponding fault type judgment report can be generated. The fault type judgment report includes at least one of the fault type, fusion confidence level, sound feature information, vibration feature information, monitoring location coordinates, the severity of the mechanical fault, and the type of monitored object in the target monitoring device coordinate feature set. Therefore, the fault type judgment report can record the fault type judgment result and the sound feature information, vibration feature information, and fusion confidence level on which it is based, making the fault type judgment process traceable.
[0097] S150: In the target monitoring device coordinate feature set of each monitoring device of the fault monitoring object, determine a set of target monitoring device coordinate feature sets with the highest fusion confidence, and locate the fault point coordinates on the fault monitoring object according to the monitoring position coordinates contained in the set of target monitoring device coordinate feature sets.
[0098] refer to Figure 5 , Figure 5 A flowchart illustrating the fault point coordinate determination method in this embodiment is shown. After determining the fault type of the monitored object, this embodiment further utilizes the time difference of arrival of the same fault signal received by at least a preset number of positioning monitoring locations set in different spatial locations to determine the spatial location of the fault point within the GIS equipment. Preferably, sound monitoring data collected by the sound acquisition unit is used for fault point location. The method includes the following steps.
[0099] S151: From the set of target monitoring device coordinate features that are associated with the same fault monitoring object and whose associated feature information is matched to obtain the same fault type, determine a set of target monitoring device coordinate feature sets with the highest fusion confidence, and use the monitoring positions corresponding to each selected set of target monitoring device coordinate feature sets as the location monitoring positions. The preset number is at least four.
[0100] In some embodiments, step S151 includes: matching the associated feature information corresponding to each target monitoring device coordinate feature set associated with the same fault monitoring object with the preset fault type feature information to obtain the fault type corresponding to each target monitoring device coordinate feature set; dividing the target monitoring device coordinate feature sets with the same corresponding fault type into the same group; determining the group containing at least a preset number of target monitoring device coordinate feature sets as a positioning group; selecting the preset number of target monitoring device coordinate feature sets with the highest fusion confidence from the positioning group according to the order of fusion confidence from high to low; and taking the monitoring positions corresponding to each selected target monitoring device coordinate feature set as a group of positioning monitoring positions, so as to jointly determine the corresponding fault point coordinates by using the signal arrival time difference between each positioning monitoring position. In this embodiment, the preset number is also at least four.
[0101] In some embodiments, the occurrence time and waveform correlation of fault signals collected from monitoring locations corresponding to the coordinate feature sets of each target monitoring device within the same group can be compared. When the fault signals collected from the monitoring locations corresponding to the coordinate feature sets of each target monitoring device are within the same preset time window and the waveform correlation reaches a preset correlation threshold, it is determined that these fault signals originate from the same fault event. This avoids erroneously using multiple fault events generated by the same monitoring object at different times for the same fault location.
[0102] In some embodiments, when the number of target monitoring device coordinate feature sets contained in the same positioning group is greater than the preset number, the preset number of target monitoring device coordinate feature sets with the highest fusion confidence are selected from the positioning group in descending order of fusion confidence, and the monitoring positions corresponding to each selected target monitoring device coordinate feature set are used as positioning monitoring positions.
[0103] S152: Obtain the arrival time of the sound fault signal generated when the fault monitoring object malfunctions in the sound monitoring data corresponding to each of the positioning and monitoring locations.
[0104] Specifically, when the associated feature information in the coordinate feature set of the target monitoring device corresponding to the location successfully matches the preset fault type feature information, the sound monitoring data that generates the associated feature information is determined as monitoring data containing the corresponding fault signal. Based on the fault type feature information corresponding to the successfully matched fault type, the corresponding sound fault signal is extracted from the monitoring data, and the arrival time of the sound fault signal at each location monitoring position is determined.
[0105] In some embodiments, the arrival time can be determined when the amplitude of the fault signal reaches a preset trigger threshold; alternatively, the relative arrival time between fault signals can be determined by cross-correlation processing of fault signals corresponding to different positioning monitoring locations.
[0106] S153: For any two of the positioning and monitoring locations, calculate the difference in arrival times of the sound fault signal at the two positioning and monitoring locations to obtain the corresponding sound signal arrival time difference.
[0107] Specifically, because the same fault signal travels different distances from the fault point to different location monitoring positions, the time at which the fault signal is received at different location monitoring positions differs. The sign of the signal arrival time difference can characterize the order in which the fault signal arrives at the two location monitoring positions, and its absolute value can characterize the distance difference between the two propagation paths.
[0108] S154: Multiply the arrival time difference of each sound signal by the corresponding sound signal propagation speed to obtain multiple sets of distance differences, wherein each set of distance differences is the difference in distance from the fault point of the fault monitoring object to the corresponding two positioning monitoring positions.
[0109] The effective sound propagation speed can be determined based on the propagation medium and propagation path of the sound fault signal. The propagation medium and propagation path can involve at least one of the insulating gas inside the GIS equipment, the casing structure of the GIS equipment, and the air outside the GIS equipment.
[0110] In some embodiments, the distance difference can be expressed as:
[0111] (3)
[0112] in, Coordinates from the fault point to the monitoring location and monitoring location coordinates The distance difference The speed of signal propagation can be the speed of propagation of a sound fault signal. Coordinates of the signal arrival at the monitoring location and monitoring location coordinates The time difference is maintained as a non-negative value by switching the subtraction relationship between the two moments.
[0113] In some embodiments, the effective propagation speed corresponding to different propagation paths can be determined in advance through structural model simulation or on-site calibration. For example, test excitation can be applied at a known location of the GIS equipment, and the propagation speed of the sound signal can be calibrated based on the time it takes for the test signal to reach each acquisition unit and the corresponding propagation path length. The propagation speed can also be corrected based on temperature, gas pressure, or structural connection status.
[0114] For any two location monitoring positions, multiplying the time difference of their corresponding signal arrival by the corresponding signal propagation speed will give the difference in distance from the fault point to these two location monitoring positions.
[0115] S155: Based on the coordinates of each of the positioning and monitoring locations and the distance differences of each group, obtain the spatial position coordinates that simultaneously satisfy the distance difference constraints of each group, and use them as the fault point coordinates of the fault monitoring object.
[0116] In some embodiments, the coordinates of the fault point can be obtained by substituting the coordinates of different monitoring points i ( , , ), coordinates of different monitoring points j ( , , ) and the corresponding The resulting system of distance difference equations can be expressed as follows:
[0117] (4)
[0118] For example, the three-dimensional coordinates of four monitoring locations corresponding to the same monitoring object are obtained. ( , , ), ( , , ), ( , , ), ( , , The distance difference between the fault point and the first and second monitoring positions is calculated using equation (3). The distance difference between the fault point and the first and third monitoring positions. The distance difference between the fault point and the first and fourth monitoring positions, respectively. Substituting the three-dimensional coordinates and distance differences into equation (4) to construct a system of equations, the fault point can be solved. , , The three-dimensional coordinates of ).
[0119] Therefore, by establishing distance difference constraints through the signal arrival time difference between different positioning monitoring locations, the spatial location of the fault point can be determined without knowing the absolute generation time of the fault signal in advance, thus refining the fault diagnosis results from fault type identification to the determination of the specific fault location of the monitored object.
[0120] In some embodiments, after step S155, the method further includes: obtaining a pre-constructed structural model based on the actual structural data of the GIS device, and obtaining the spatial range corresponding to each component of the fault monitoring object in the structural model; comparing the coordinates of the fault point with the spatial range corresponding to each component in the same coordinate system, and determining the component whose spatial range includes the coordinates of the fault point as the fault part of the fault monitoring object.
[0121] In some embodiments, after determining the coordinates of the fault point and the location of the fault in the monitored object, a fault analysis report can be generated. The fault analysis report includes at least one of the following: the coordinates of the fault point, the location of the fault, the fault type, each location monitoring position involved in fault location, associated feature information from the coordinate feature set of the target monitoring equipment corresponding to each location monitoring position, and a fault type judgment report. Thus, the fault analysis report can record and associate the fault type, fault severity, fault point coordinates, fault location, and corresponding diagnostic criteria, facilitating the inspection or maintenance of the monitored object.
[0122] refer to Figure 6 , Figure 6This embodiment illustrates a flowchart of the method for determining the monitoring location. Before executing step S110, this embodiment pre-screens the possible installation locations of the sound acquisition unit and the vibration acquisition unit to ensure that the subsequently determined monitoring location can stably acquire the sound and vibration signals generated by the corresponding monitoring object.
[0123] Figure 6 The method for determining the monitoring location includes the following steps:
[0124] S210: Determine multiple candidate installation locations for the sound signal acquisition unit and the vibration signal acquisition unit on the GIS equipment or its structural model.
[0125] In some embodiments, when the candidate installation location is set on an actual GIS device, sound evaluation data and vibration evaluation data for multiple evaluation cycles can be collected under different environmental interference conditions or device operating states.
[0126] In other embodiments, when the candidate installation location is set on the structural model of the GIS equipment, the corresponding sound evaluation data and vibration evaluation data can be obtained through simulation.
[0127] The structural model of the GIS equipment can be implemented as a basic GIS equipment model including the equipment shell, equipment air chamber, monitoring object, equipment connection structure, and support structure. The basic GIS equipment model can also include material parameters, wall thickness, connection relationships, and vibration and sound propagation parameters corresponding to each structural part. The monitoring object can include at least one of a circuit breaker, disconnector, voltage transformer, and current transformer. The structural model can be used to simulate the propagation process of sound and vibration signals generated when the monitoring object experiences a mechanical failure within the GIS equipment. Specifically, multiple candidate installation locations can be determined on the structural model based on the installable area of the GIS equipment shell, the spatial location of the monitoring object, the feasibility of sensor installation, and maintenance space. The candidate installation locations can be located inside the GIS equipment shell. For each candidate installation location, a sound signal acquisition unit and a vibration signal acquisition unit can be set up.
[0128] S220: Under multiple sets of environmental interference conditions, sound evaluation data and vibration evaluation data are collected for multiple evaluation cycles at each of the candidate installation locations, and feature processing is performed to obtain sound evaluation feature information and vibration evaluation feature information corresponding to each of the candidate installation locations.
[0129] The environmental interference conditions may include at least one of different background noise levels, temperature, humidity, electromagnetic interference, external mechanical vibration, equipment operating status, and installation contact status. For each candidate installation location, the sound and vibration signals generated when a preset mechanical fault occurs in each monitored object can be simulated in the structural model, and the background noise, external vibration, or propagation attenuation effects under the corresponding environmental interference conditions can be superimposed to obtain sound evaluation data and vibration evaluation data for multiple consecutive evaluation periods.
[0130] In some embodiments, the sound evaluation data and vibration evaluation data can be subjected to frequency domain transformation according to the methods for obtaining sound feature information and vibration feature information in the foregoing embodiments to obtain sound evaluation feature information and vibration evaluation feature information, respectively.
[0131] S230: Calculate the fluctuation variance of the sound evaluation feature information and vibration evaluation feature information of the same candidate installation location in the multiple evaluation periods, and normalize the fluctuation variance and invert the correlation to determine the sound evaluation stability confidence and vibration evaluation stability confidence corresponding to each candidate installation location.
[0132] In some embodiments, the calculation methods for the first and second stability confidence scores described in the foregoing embodiments are used. For the same candidate installation location, corresponding sound evaluation feature values and vibration evaluation feature values are obtained in multiple consecutive evaluation periods, and the fluctuation variance of each evaluation feature value relative to its corresponding average value is calculated. The fluctuation variances of the sound evaluation feature information and the vibration evaluation feature information can be normalized and then inverted to determine the stability confidence scores for sound and vibration evaluations. The smaller the fluctuation variance of the sound evaluation feature information or the vibration evaluation feature information, the more stable the corresponding features collected at the candidate installation location under different environmental interference conditions and different evaluation periods, and the higher the corresponding stability confidence score.
[0133] S240: Determine the signal quality evaluation result corresponding to each of the candidate installation locations based on the signal parameters in the sound evaluation data and vibration evaluation data corresponding to each of the candidate installation locations.
[0134] Specifically, at least one signal parameter can be extracted from the sound evaluation data and vibration evaluation data, including effective frequency band energy, noise frequency band energy, signal-to-noise ratio, signal attenuation degree, and characteristic amplitude, respectively, and the sound signal quality evaluation value and vibration signal quality evaluation value can be determined accordingly.
[0135] In some embodiments, the lower of the sound signal quality evaluation value and the vibration signal quality evaluation value is determined as the signal quality evaluation result corresponding to the candidate installation location. Therefore, only when both the sound signal and vibration signal have high signal quality will the corresponding candidate installation location have a high signal quality evaluation result, avoiding the selection of installation locations where only one modal signal has good quality while the other modal signal is difficult to acquire effectively.
[0136] In other embodiments, preset evaluation value weights can be applied to the sound signal quality evaluation value and the vibration signal quality evaluation value respectively, and then summed to obtain the signal quality evaluation result corresponding to the candidate installation position.
[0137] S250: The candidate installation locations where the sound evaluation stability confidence level, the vibration evaluation stability confidence level, and the signal quality evaluation results all reach their respective preset screening thresholds are determined as available installation locations, and multiple locations whose monitoring ranges jointly cover each of the monitored objects are selected from the available installation locations as the monitoring locations.
[0138] Specifically, the confidence level of sound stability assessment can be compared with a preset sound stability threshold, the confidence level of vibration stability assessment can be compared with a preset vibration stability threshold, and the signal quality assessment result can be compared with a preset signal quality threshold. Only when all three comparison results meet the corresponding threshold conditions is the candidate installation location determined as a usable installation location.
[0139] In some embodiments, for each available installation location, a monitoring object whose sound and vibration signals that can be collected at that location meet preset identifiable conditions is identified, and this monitoring object is determined as an object within the monitoring range of that available installation location. Multiple monitoring ranges from each available installation location are selected to cover all monitoring objects, and this location is chosen as the final monitoring location.
[0140] By using the above method, before actually installing the acquisition unit, the structural model is used to evaluate the stability and signal quality of vibration and acoustic acquisition at different candidate installation locations. This can avoid installation locations with strong background noise interference, severe signal attenuation, or large characteristic fluctuations, and ensure that the multiple monitoring locations selected at the end jointly cover each monitoring object, thereby improving the reliability of the monitoring data used for subsequent mechanical fault diagnosis.
[0141] In some embodiments, reference Figure 7 , Figure 7This embodiment illustrates a flowchart of the monitoring location selection method. In the aforementioned structural model, this embodiment simulates the sound and vibration signals generated when different monitored objects experience mechanical failures. Based on the evaluation characteristics collected from each available installation location and their ability to distinguish between different monitored objects, it selects the actual monitoring locations for installing the sound signal acquisition unit and the vibration signal acquisition unit. The monitoring location selection method can be executed after determining the available installation locations in step S250, or it can be executed independently before step S110. The method further includes:
[0142] S261: In the structural model of the GIS equipment, simulate the mechanical failure of each of the monitoring objects individually for each of the preset fault types, and obtain the sound evaluation feature information and vibration evaluation feature information of each available installation location under the condition that each of the monitoring objects individually experiences any of the preset fault types.
[0143] The preset fault types and their corresponding mechanical abnormalities, sound fault characteristics, and vibration fault characteristics can be determined by referring to the correspondence in the aforementioned preset fault type database. The available installation locations can be pre-set based on experience, or they can be... Figure 6 In the embodiments, the sound evaluation stability confidence level of the candidate installation location, the vibration evaluation stability confidence level, and the signal quality evaluation results are all obtained by filtering through a preset screening threshold.
[0144] Specifically, for any monitored object and any preset fault type, corresponding fault excitation parameters can be set in the structural model, and the process of sound and vibration signals generated by the fault propagating along the internal structure and shell of the GIS equipment to each available installation location can be simulated. Subsequently, sound evaluation data and vibration evaluation data corresponding to each available installation location are acquired and characterized to obtain sound evaluation feature information and vibration evaluation feature information. The sound evaluation feature information and vibration evaluation feature information can be amplitude-frequency feature information. By simulating the mechanical fault of each monitored object individually, a clear association can be established between the evaluation features collected at each available installation location and the corresponding fault source, avoiding feature aliasing caused by multiple monitored objects generating signals simultaneously.
[0145] In some embodiments, the fault excitation parameters used to simulate each of the preset fault types can be determined based on historical fault test data of the corresponding mechanical fault. Specifically, the monitored object can be placed in the corresponding preset mechanical fault state in the test environment, and the sound signals and vibration signals generated in this state can be collected. Based on the collection results, at least one fault excitation parameter can be determined from the following: fault location, fault excitation frequency, fault excitation amplitude, duration, and variation pattern.
[0146] In other embodiments, the corresponding fault excitation parameters can be determined by mechanical simulation or acoustic-vibration coupling simulation based on the structural parameters, material parameters, connection relationships and mechanical fault mechanisms of the monitored object, and the fault excitation parameters can be calibrated using historical fault test data.
[0147] S262: The sound evaluation feature information and vibration evaluation feature information obtained when the same available installation location of the same monitoring object individually occurs with any of the preset fault types are fused together to obtain evaluation association feature information.
[0148] In some embodiments, the method for obtaining evaluation-related feature information from the sound evaluation feature information and vibration evaluation feature information can refer to the implementation method of obtaining related feature information based on sound feature information and vibration feature information in the target monitoring device coordinate feature set in the foregoing embodiments.
[0149] S263: Calculate the feature difference value between each monitored object and each other monitored object under any of the preset fault types in the corresponding evaluation association feature information, and determine the minimum value among the feature difference values as the object distinguishability of the monitored object in the available installation location.
[0150] Specifically, the feature difference value can be exemplarily represented by Euclidean distance, that is, calculating the difference between corresponding feature vectors in two evaluation-related feature information, and summing the squares and taking the square root of each difference. Furthermore, when there are multiple other monitoring objects and multiple preset fault types, the feature difference values between the current monitoring object and each of the other monitoring objects under their respective preset fault types can be calculated separately, and the minimum value among the obtained feature difference values can be selected as the object distinguishability of the current monitoring object at that available installation location. The object distinguishability characterizes the ability to distinguish the current monitoring object from other monitoring objects and their corresponding fault types that are most likely to cause feature confusion at the corresponding available installation location. Therefore, the higher the object distinguishability, the better the identification ability of the current monitoring object at that available installation location.
[0151] S264: According to the priority order of the object distinguishability from large to small, select at least one available installation location of the monitored object from each available installation location as the target installation location, and take the target installation location corresponding to each monitored object as the monitoring location.
[0152] In some embodiments, for each monitored object, available installation locations can be sorted in descending order of object distinguishability, and the target installation location corresponding to the monitored object can be selected from the top-ranked available installation locations. For example, when an available installation location can cover the corresponding monitored object and its object distinguishability meets a preset distinguishability requirement, the available installation location can be determined as the target installation location. When the monitoring range of an available installation location cannot cover the entire monitored area of the corresponding monitored object, or when multiple monitoring locations need to be set up for redundant monitoring, other available installation locations can continue to be selected in descending order of object distinguishability. If the same available installation location can monitor multiple monitored objects, the available installation location can be associated with multiple monitored objects simultaneously, without needing to repeatedly set up acquisition units for different monitored objects.
[0153] In some embodiments, reference Figure 8 , Figure 8 This illustration shows an example diagram of selecting monitoring locations based on object discrimination in one embodiment of the present disclosure, which can be used as... Figure 7 The example shown illustrates the application of the monitoring location filtering method.
[0154] like Figure 8 As shown, the monitoring object in the structural model It may include the first monitoring object Second monitoring object and the third monitoring object and for each monitoring object Set the available installation locations separately. The second monitoring target Available installation locations Including available installation locations To a usable installation location Solid arrows indicate the direction of the monitoring object in the structural model. Apply monitoring to the respective targets Corresponding preset fault types The fault excitation. The dashed arrows indicate the monitoring objects. The sound and vibration signals generated under the corresponding fault excitation are transmitted to each available installation location. Propagation. After the sound and vibration signals propagate to each available installation location, they are collected by virtual sound acquisition units and virtual vibration acquisition units set in the structural model at the corresponding available installation locations, resulting in corresponding sound evaluation data and vibration evaluation data.
[0155] With the second monitoring object and available installation locations For example, the fault types in the first preset fault type group are: , , Simulate the first monitoring object respectively Preset fault type occurs alone Second monitoring object Preset fault type occurs alone and the third monitoring object Preset fault type occurs alone Obtain available installation locations Corresponding assessment related feature information , , .Will respectively with and Comparison yields feature difference values. and .
[0156] Using the same method, the first monitoring object Second monitoring object Third monitoring object Each fault type in the second preset fault type group occurs individually. , , , obtain the feature difference value and The minimum value among the four characteristic differences mentioned above will be selected as the usable installation location. For the second monitoring target object distinguishability Each preset fault type group can include a preset fault type corresponding to each monitored object.
[0157] Furthermore, using the same method, available installation locations can be determined. To a usable installation location For the second monitoring target object distinguishability Therefore, the greater the object distinguishability of an available installation location, the greater the difference in the evaluation correlation characteristics between that location and other easily confused monitoring objects and preset fault types. This means that the available installation location can be prioritized as the target installation location for the monitoring object. When multiple target installation locations need to be selected, the available installation locations can continue to be selected in descending order of object distinguishability to achieve monitoring redundancy requirements.
[0158] Therefore, this embodiment not only eliminates unsuitable locations for installing the acquisition unit based on the assessment of stability confidence and signal quality, but also determines the monitoring location based on the ability of each available installation location to distinguish faults in different monitored objects, thereby reducing the possibility of vibration and acoustic signals from faults in adjacent monitored objects propagating to the same location and causing confusion about the source of the fault.
[0159] like Figure 9 The diagram shown illustrates the structure of an electronic device according to an embodiment of the present disclosure.
[0160] The electronic device can execute computer program instructions to perform tasks such as Figure 2 or Figure 3 or Figure 4 or Figure 5 or Figure 6 or Figure 7 The method described herein. Exemplarily, the electronic device may be a distributed computing node system, server group / server, desktop computer, laptop computer, etc., for running, such as... Figure 2 The mechanical fault diagnosis method is used to complete the mechanical fault diagnosis of GIS equipment.
[0161] The electronic device 900 includes a bus 901, a processor 902, and a memory 903. The processor 902 and the memory 903 can communicate via the bus 901. The memory 903 can store program instructions. The processor 902 implements the method steps in the previous embodiments by running the program instructions in the memory 903, for example... Figure 2 or Figure 3 or Figure 4 or Figure 5 or Figure 6 or Figure 7 The method in the middle.
[0162] Bus 901 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of description, although only one thick line is used in the diagram, this does not indicate that there is only one bus or one type of bus.
[0163] In some embodiments, processor 902 may be implemented as a central processing unit (CPU), microcontroller unit (MCU), system-on-chip (SoC), or field-programmable array (FPGA). Memory 903 may include volatile memory for temporary data storage during program execution, such as random access memory (RAM).
[0164] The memory 903 may also include non-volatile memory for data storage, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state disk (SSD).
[0165] In some embodiments, the electronic device 900 may further include a communicator 904. The communicator 904 is used for communication with external devices. In specific examples, the communicator 904 may include one or more wired and / or wireless communication circuit modules. For example, the communicator 904 may include one or more of, such as a wired network card, a USB module, a serial interface module, etc. The wireless communication protocols followed by the wireless communication module include, for example, Nearfield Communication (NFC) technology, Infrared (IR) technology, Global System for Mobile Communications (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Time-Division Code Division Multiple Access (TD-SCDMA), Long Term Evolution (LTE), Bluetooth (BT), Global Navigation Satellite System (GNSS), etc.
[0166] This disclosure also provides a computer-readable storage medium, characterized in that it stores program instructions, which are executed and perform, for example... Figure 2 or Figure 3 or Figure 4 or Figure 5 or Figure 6 or Figure 7 Mechanical fault diagnosis method in the embodiments.
[0167] That is, the method steps in the above embodiments are implemented as software or computer code that can be stored in a recording medium (such as CD ROM, RAM, floppy disk, hard disk or magneto-optical disk), or implemented as computer code that is originally stored in a remote recording medium or a non-transitory machine-readable medium and will be stored in a local recording medium after being downloaded via a network, so that the method described herein can be processed by software stored on a recording medium of a general-purpose computer, a special processor or programmable or special hardware (such as ASIC or FPGA).
[0168] In summary, this disclosure provides a mechanical fault diagnosis method and system. The method includes, in each acquisition cycle, acquiring a set of sound monitoring data and vibration monitoring data from multiple monitoring devices on the GIS equipment, and extracting a coordinate feature set of each monitoring device. The coordinate feature set of the monitoring device includes a sound feature set and a vibration feature set; wherein the sound feature set and vibration feature set respectively include sound feature information and vibration feature information, and both include the coordinates of the monitoring location of the monitoring device and the monitoring object; acquiring, respectively, a first signal-to-noise ratio confidence score and a first stability confidence score corresponding to the sound feature information of the same monitoring device coordinate feature set, and a second signal-to-noise ratio confidence score and a second stability confidence score corresponding to the vibration feature information; determining the sound feature confidence score corresponding to the monitoring location based on the first signal-to-noise ratio confidence score and the first stability confidence score, and determining the vibration feature confidence score corresponding to the monitoring location based on the second signal-to-noise ratio confidence score and the second stability confidence score; wherein, the sound feature confidence score... The vibration feature confidence scores are positively correlated with their respective signal-to-noise ratio confidence scores and stability confidence scores. Based on the sound fusion weights and vibration fusion weights corresponding to each monitoring device coordinate feature set, the sound feature confidence scores and vibration feature confidence scores corresponding to the same monitoring device coordinate feature set are weighted and fused to obtain a fusion confidence score. The monitoring device coordinate feature sets whose fusion confidence scores reach a preset confidence threshold are determined as target monitoring device coordinate feature sets, and associated feature information containing the sound feature information and vibration feature information associated with the target monitoring device coordinate feature sets is obtained. The associated feature information is matched with preset fault type feature information, and the fault type of the fault monitoring object is determined during matching. In the target monitoring device coordinate feature sets of each monitoring device of the fault monitoring object, a preset number of target monitoring device coordinate feature sets with the highest fusion confidence scores are determined, and the fault point coordinates on the fault monitoring object are located according to the monitoring position coordinates contained in the target monitoring device coordinate feature sets. This disclosure can improve the reliability of fault type identification and fault point location.
[0169] The above embodiments are merely illustrative of the principles and effects of this disclosure and are not intended to limit this disclosure. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this disclosure. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this disclosure should still be covered by the claims of this disclosure.
Claims
1. A method for diagnosing mechanical faults, characterized in that, Applied to GIS equipment, wherein the GIS equipment contains multiple monitoring objects, the method includes: In each acquisition cycle, a set of sound monitoring data and vibration monitoring data from multiple monitoring devices on the GIS equipment are acquired respectively, and the coordinate feature sets of the monitoring devices are extracted respectively. The coordinate feature sets of the monitoring devices include sound feature sets and vibration feature sets. The sound feature sets and vibration feature sets respectively include sound feature information and vibration feature information, and both include the coordinates of the monitoring location of the monitoring device and the monitoring object. The system acquires the first signal-to-noise ratio (SNR) confidence level and the first stability confidence level corresponding to the sound feature information of the same monitoring device coordinate feature set, and the second SNR confidence level and the second stability confidence level corresponding to the vibration feature information. Based on the first SNR confidence level and the first stability confidence level, the system determines the sound feature confidence level corresponding to the monitoring location, and based on the second SNR confidence level and the second stability confidence level, the system determines the vibration feature confidence level corresponding to the monitoring location. The sound feature confidence level and the vibration feature confidence level are positively correlated with their respective corresponding SNR confidence level and stability confidence level. Based on the sound fusion weight and vibration fusion weight corresponding to each of the monitoring device coordinate feature sets, the sound feature confidence and vibration feature confidence corresponding to the same monitoring device coordinate feature set are weighted and fused to obtain a fusion confidence. The monitoring device coordinate feature set whose fusion confidence reaches a preset confidence threshold is determined as the target monitoring device coordinate feature set, and associated feature information containing the sound feature information and vibration feature information associated in the target monitoring device coordinate feature set is obtained. The associated feature information is matched with the preset fault type feature information, and the fault type of the fault monitoring object is determined during the matching process. In the target monitoring device coordinate feature set of each monitoring device of the fault monitoring object, determine a set of target monitoring device coordinate feature sets with the highest fusion confidence, and locate the coordinates of the fault point on the fault monitoring object according to the coordinates of each monitoring position contained in the set of target monitoring device coordinate feature sets.
2. The mechanical fault diagnosis method according to claim 1, characterized in that, Before acquiring a set of sound monitoring data and vibration monitoring data from multiple monitoring devices on the GIS equipment in each acquisition cycle, and extracting the coordinate feature sets of the monitoring devices respectively, the method further includes: Multiple candidate installation locations for the sound signal acquisition unit and the vibration signal acquisition unit are determined on the GIS equipment or its structural model; Under multiple environmental interference conditions, sound evaluation data and vibration evaluation data for multiple evaluation cycles are collected at each of the candidate installation locations, and feature processing is performed to obtain sound evaluation feature information and vibration evaluation feature information corresponding to each candidate installation location. The fluctuation variances of the sound evaluation feature information and vibration evaluation feature information of the same candidate installation location in the multiple evaluation periods are calculated respectively. The fluctuation variances are normalized and the correlation is inverted to determine the sound evaluation stability confidence and vibration evaluation stability confidence corresponding to each candidate installation location. Based on the signal parameters in the sound evaluation data and vibration evaluation data corresponding to each of the candidate installation locations, the signal quality evaluation result corresponding to each of the candidate installation locations is determined; Candidate installation locations where the sound evaluation stability confidence level, the vibration evaluation stability confidence level, and the signal quality evaluation results all reach their respective preset screening thresholds are determined as available installation locations. Multiple locations whose monitoring ranges collectively cover each monitored object are selected from these available installation locations as the monitoring locations.
3. The mechanical fault diagnosis method according to claim 1, characterized in that, The step involves determining a set of target monitoring device coordinate feature sets with the highest fusion confidence level from the target monitoring device coordinate feature sets of each monitoring device of the fault monitoring object, and locating the fault point coordinates on the fault monitoring object based on the monitoring location coordinates included in this set of target monitoring device coordinate feature sets, including: From the set of target monitoring device coordinate features that are associated with the same fault monitoring object and whose associated feature information is matched to obtain the same fault type, determine a set of target monitoring device coordinate feature sets with the highest fusion confidence, and use the monitoring position corresponding to each selected set of target monitoring device coordinate feature sets as the location monitoring position; Obtain the arrival time of the sound fault signal generated when the fault monitoring object malfunctions in the sound monitoring data corresponding to each of the positioning and monitoring locations; For any two of the aforementioned positioning and monitoring locations, the difference in arrival times of the sound fault signal at the two aforementioned positioning and monitoring locations is calculated to obtain the corresponding sound signal arrival time difference; The arrival time difference of each sound signal is multiplied by the corresponding sound signal propagation speed to obtain multiple sets of distance differences, wherein each set of distance differences is the difference in distance from the fault point of the fault monitoring object to the corresponding two positioning monitoring locations; Based on the coordinates of each of the aforementioned positioning and monitoring locations and the distance differences of each group, the spatial location coordinates that simultaneously satisfy the distance difference constraints of each group are obtained, and these coordinates are used as the fault point coordinates of the fault monitoring object.
4. The mechanical fault diagnosis method according to claim 3, characterized in that, The process involves determining a set of target monitoring device coordinate feature sets with the highest fusion confidence from a set of target monitoring device coordinate feature sets that are associated with the same fault monitoring object and whose associated feature information is matched, and using the monitoring location corresponding to each selected target monitoring device coordinate feature set as the location monitoring location, including: The associated feature information corresponding to each target monitoring device coordinate feature set associated with the same fault monitoring object is matched with the preset fault type feature information to obtain the fault type corresponding to each target monitoring device coordinate feature set, and the target monitoring device coordinate feature sets with the same fault type are divided into the same group; Grouping a set of coordinate features of at least a preset number of the target monitoring devices is defined as a positioning group, and the monitoring positions corresponding to each set of coordinate features of the target monitoring devices within the same positioning group are used as a group of positioning monitoring positions, so as to jointly determine the coordinates of the corresponding fault point by utilizing the signal arrival time difference between each positioning monitoring position.
5. The mechanical fault diagnosis method according to claim 3, characterized in that, The method further includes: Obtain a pre-constructed structural model based on the actual structural data of the GIS equipment, and obtain the spatial range of each component of the fault monitoring object in the structural model; The coordinates of the fault point are compared with the spatial ranges corresponding to each component in the same coordinate system, and the component whose spatial range includes the coordinates of the fault point is determined as the fault location of the fault monitoring object.
6. The mechanical fault diagnosis method according to claim 1, characterized in that, The steps of acquiring the first signal-to-noise ratio (SNR) confidence level and the first stability confidence level corresponding to the sound feature information of the same monitoring device coordinate feature set, and the second SNR confidence level and the second stability confidence level corresponding to the vibration feature information; determining the sound feature confidence level corresponding to the monitoring location based on the first SNR confidence level and the first stability confidence level, and determining the vibration feature confidence level corresponding to the monitoring location based on the second SNR confidence level and the second stability confidence level, include: Calculate the signal-to-noise ratio (SNR) of the sound feature information and the vibration feature information in the same coordinate feature set of the monitoring device, and normalize the SNR to obtain the first SNR confidence level and the second SNR confidence level. The amplitude-frequency characteristic fluctuation variances of the sound feature information and the vibration feature information in the same coordinate feature set of the monitoring device are calculated respectively, and the amplitude-frequency characteristic fluctuation variances are normalized and then inverted to obtain the first stability confidence and the second stability confidence. The sound feature confidence is obtained by applying corresponding weighting coefficients to the first signal-to-noise ratio confidence and the first stability confidence, and then summing them. The vibration feature confidence is obtained by applying corresponding weighting coefficients to the second signal-to-noise ratio confidence and the second stability confidence.
7. The mechanical fault diagnosis method according to claim 1, characterized in that, The sound fusion weight is positively correlated with the ratio of the sound feature confidence level to the vibration feature confidence level, and the vibration fusion weight is negatively correlated with the ratio of the sound feature confidence level to the vibration feature confidence level. The calculation methods for the sound fusion weight and the vibration fusion weight include: The proportions of the confidence levels of the sound features and the vibration features to the sum of the confidence levels of the sound features and the vibration features are determined as the sound fusion weight and the vibration fusion weight, respectively.
8. The mechanical fault diagnosis method according to claim 1, characterized in that, The preset fault type feature information includes: sound type fault type feature information and vibration type fault type feature information. The step of matching the associated feature information with the preset fault type feature information and determining the fault type of the fault monitoring object during matching includes: The confidence levels of sound features and vibration features corresponding to the coordinate feature set of the target monitoring device are compared, and the matching priorities of the sound feature information and the vibration feature information are determined respectively; wherein, the matching priority of each feature information is positively correlated with its corresponding feature confidence level; The feature information with higher matching priority among the sound feature information and the vibration feature information is determined as the dominant matching feature information; The dominant matching feature information is matched with the corresponding fault type feature information to obtain the matching result; In response to the matching result satisfying the matching condition, the fault type corresponding to the matching result is determined as the fault type of the fault monitoring object; In response to the matching result not meeting the matching condition, the next target monitoring device coordinate feature set is selected from the target monitoring device coordinate feature set associated with the same monitoring object and not yet matched for fault type, in descending order of fusion confidence, and the matching is performed again based on the sound feature information and vibration feature information corresponding to the next target monitoring device coordinate feature set.
9. The mechanical fault diagnosis method according to claim 1 or 2, characterized in that, The method further includes: In the structural model of the GIS equipment, mechanical failures of each of the monitoring objects are simulated individually for each of the preset failure types, and sound evaluation feature information and vibration evaluation feature information of each available installation location are obtained when each of the monitoring objects individually experiences any of the preset failure types. The sound evaluation feature information and vibration evaluation feature information obtained when the same available installation location and the same monitoring object individually experience any of the preset fault types are fused together to obtain evaluation correlation feature information; Calculate the feature difference value between each monitoring object and each other monitoring object under any of the preset fault types, and determine the minimum value among the feature difference values as the object distinguishability of the monitoring object under the available installation location; According to the priority order of the object distinguishability from large to small, at least one available installation location of the monitored object is selected from each of the available installation locations as the target installation location, and the target installation location corresponding to each monitored object is used as the monitoring location.
10. A mechanical fault diagnosis system, characterized in that, The system is applied to GIS equipment that includes multiple monitoring objects, and includes: Multiple monitoring devices are respectively set at their respective monitoring locations, and each monitoring location is associated with a corresponding monitoring object; wherein, each monitoring device includes a sound acquisition unit and a vibration acquisition unit, the sound acquisition unit is used to collect sound monitoring data of the corresponding monitoring location, and the vibration acquisition unit is used to collect vibration monitoring data of the corresponding monitoring location; The processing unit is communicatively connected to the plurality of sound acquisition units and the plurality of vibration acquisition units, and is configured to execute the mechanical fault diagnosis method as described in any one of claims 1 to 9.
