Fault detection method, device, equipment and storage medium

Abstract

The present disclosure relates to a fault detection method, device, equipment and storage medium, and relates to the technical field of signal processing. The method obtains sound signals and rotation speed data of a fan device in a first time period; determines phases corresponding to multiple time points based on the rotation speed of the multiple time points; determines multiple sound amplitudes corresponding to multiple target phases based on the phases corresponding to the multiple time points and the sound amplitudes, wherein the sound amplitude corresponding to each target phase is the sound amplitude of the corresponding time point in a rotation period, and the rotation period corresponds to one complete rotation of an impeller; for each target phase, fuses the multiple sound amplitudes corresponding to the target phase to obtain a fused sound amplitude of the target phase; and determines fault information of the impeller based on the fused sound amplitudes of the multiple target phases. The method can accurately identify the acoustic characteristics corresponding to weak faults, and effectively improves the accuracy and reliability of the imbalance fault detection of the impeller.

Description

Technical Field

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

[0002] Fans are power devices equipped with rotating impellers, adaptable to various industrial applications, and widely used in thermal power, chemical, metallurgical, data center, and central air conditioning systems. Because fans operate in high humidity, water mist, corrosive, dusty, and continuous environments, impellers are prone to imbalance due to factors such as scale buildup, corrosion, wear, deformation, installation misalignment, or counterweight detachment. Impeller imbalance can lead to various problems during fan operation. Therefore, early fault detection for fans is crucial.

[0003] In related technologies, the common approach is to collect sound signals from the wind turbine during operation, then extract acoustic features such as spectral peak value, frequency band energy, sound pressure level, and spectral entropy, and use these features to determine if the wind turbine is malfunctioning. However, due to background noise interference from sources such as water spray noise, airflow noise, motor noise, noise from adjacent equipment, and environmental wind noise near the wind turbine, the acoustic features corresponding to impeller imbalance are easily drowned out by background noise, leading to inaccurate fault detection. Summary of the Invention

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

[0005] According to one aspect of the embodiments of this disclosure, a fault detection method is provided, the method comprising: Acquire sound signals and rotational speed data of the wind turbine equipment within a first time period, wherein the sound signals include sound amplitude values ​​at multiple times, and the rotational speed data includes the rotational speed of the impeller of the wind turbine equipment at the multiple times; Based on the rotational speed at the multiple moments, determine the phase corresponding to each of the multiple moments; Based on the phase and sound amplitude corresponding to each of the multiple moments, multiple sound amplitudes corresponding to each of the multiple target phases are determined. The sound amplitude corresponding to each target phase is the sound amplitude at the moment corresponding to the target phase within one rotation cycle. One rotation cycle corresponds to one complete rotation of the impeller. For each target phase, multiple sound amplitudes corresponding to the target phase are fused to obtain the fused sound amplitude of the target phase; Based on the fused sound amplitude of each of the multiple target phases, the fault information of the impeller is determined.

[0006] In some embodiments, determining the impeller fault information based on the fused sound amplitude of each of the plurality of target phases includes: Based on the fused sound amplitude of each of the multiple target phases, the acoustic characteristics of the sound signal in multiple dimensions are determined; Based on the acoustic characteristics of the sound signal in multiple dimensions, the fault information of the impeller is determined.

[0007] In some embodiments, determining the acoustic features of the sound signal in multiple dimensions based on the fused sound amplitude of each of the plurality of target phases includes: A discrete Fourier transform is performed on the fused sound amplitude of each of the multiple target phases to obtain an order spectrum, which includes the order amplitude corresponding to each of the multiple orders. Based on the order spectrum, the acoustic features of the sound signal in multiple dimensions are determined.

[0008] In some embodiments, determining the acoustic features of the sound signal in multiple dimensions based on the order spectrum includes at least one of the following: Based on the first-order amplitude corresponding to the order spectrum, the frequency-reversing acoustic energy of the sound signal is determined, and the frequency-reversing acoustic energy is used to characterize the periodic disturbance of the impeller's first-order frequency imbalance. Based on the order amplitude values ​​corresponding to each order other than the first order in the order spectrum, the harmonic acoustic energy of the sound signal is determined, and the harmonic acoustic energy is used to characterize the composite periodic disturbance of the impeller harmonic nonlinearity. Based on the first-order amplitude corresponding to the first order in the order spectrum, the frequency conversion acoustic energy of the sound signal is determined. Based on the order amplitude corresponding to each order other than the first order in the order spectrum, the harmonic acoustic energy of the sound signal is determined. Based on the frequency conversion acoustic energy and the harmonic acoustic energy, the proportion of the frequency conversion acoustic energy in the total order acoustic energy is determined.

[0009] In some embodiments, determining the acoustic features of the sound signal in multiple dimensions based on the order spectrum includes at least one of the following: Based on the order amplitude corresponding to the target order in the order spectrum, the blade passage order energy of the sound signal is determined. The target order corresponds to the number of blades of the impeller. The blade passage order energy is used to characterize the energy intensity of the aerodynamic noise generated by the blade cutting the airflow. Based on the order amplitude of the order adjacent to the target order in the order spectrum, the modulation sideband energy of the sound signal is determined. The modulation sideband energy is used to characterize the degree of disturbance of the aerodynamic noise by the impeller imbalance.

[0010] In some embodiments, determining the acoustic features of the sound signal in multiple dimensions based on the fused sound amplitude of each of the plurality of target phases includes: Determine the statistical values ​​of the fused sound amplitude of each of the plurality of target phases, wherein the statistical values ​​include at least one of standard deviation, mean, and root mean square value; Based on the statistical values, the sound distribution characteristics of the impeller are determined, and the sound distribution characteristics are used to characterize the degree of non-uniformity of sound pressure distribution within a single rotation cycle.

[0011] In some embodiments, determining the impeller fault information based on the acoustic characteristics of the sound signal in multiple dimensions includes: For each dimension of acoustic features, a feature deviation value for that dimension is determined based on the standard acoustic features of that dimension, which are obtained under normal operating conditions of the wind turbine equipment; By integrating the feature deviation values ​​of the multiple dimensions, an imbalance index is obtained, which is used to indicate the overall degree of deviation of the acoustic features of the multiple dimensions relative to the standard acoustic features of the multiple dimensions. Based on the imbalance index, the fault information of the impeller is determined.

[0012] In some embodiments, determining the impeller's fault information based on the imbalance index includes: The imbalance index is matched with the imbalance index thresholds of multiple fault levels, and the fault level corresponding to the matched imbalance index threshold is determined as the fault level of the impeller.

[0013] In some embodiments, before determining the feature deviation value of a dimension based on the standard acoustic features of that dimension for each dimension's acoustic features, the method further includes: Based on the current operating condition type of the wind turbine equipment, a standard acoustic feature set corresponding to the current operating condition type is determined from the standard acoustic feature sets corresponding to each of the multiple operating condition types. Each standard acoustic feature set includes standard acoustic features of multiple dimensions.

[0014] In some embodiments, determining the impeller's fault information based on the imbalance index includes: Obtain the imbalance index of the sound signals from multiple second time periods; Based on the imbalance index of the sound signals in the first time period and the plurality of second time periods, the information on the change of the imbalance index is determined; Based on the information about the change in the imbalance index, the fault information of the impeller is determined.

[0015] In some embodiments, fusing multiple sound amplitudes corresponding to the target phase to obtain a fused sound amplitude of the target phase includes: Based on the weights of the rotation periods in which the multiple sound amplitudes are located, the multiple sound amplitudes are weighted and averaged to obtain the fused sound amplitude of the target phase.

[0016] In some embodiments, determining the phase corresponding to each of the plurality of time points based on the rotational speed at the plurality of time points includes: Based on the rotational speed at the multiple times, the rotational frequency at the multiple times is determined; For each time moment, the frequency conversion integral of the time moment and all time moments before the time moment is used to obtain the cumulative phase. Based on the cumulative phase and the initial phase, the phase corresponding to the time moment is obtained.

[0017] In some embodiments, the method further includes: A single rotation cycle is divided according to a preset angular interval to obtain the multiple target phases.

[0018] According to another aspect of the present disclosure, a fault detection apparatus is provided, the apparatus comprising: The acquisition module is used to acquire the sound signal and rotational speed data of the wind turbine equipment in a first time period. The sound signal includes the sound amplitude at multiple times, and the rotational speed data includes the rotational speed of the impeller of the wind turbine equipment at the multiple times. The determining module is used to determine the phase corresponding to each of the plurality of times based on the rotational speed at the plurality of times; The determining module is further configured to determine multiple sound amplitudes corresponding to multiple target phases based on the phases and sound amplitudes corresponding to the multiple times, wherein the sound amplitude corresponding to each target phase is the sound amplitude of the target phase at the time corresponding to one rotation cycle, and one rotation cycle corresponds to one complete rotation of the impeller; The fusion module is used to fuse multiple sound amplitudes corresponding to each target phase to obtain the fused sound amplitude of the target phase. The determining module is also used to determine the impeller's fault information based on the fused sound amplitude of each of the multiple target phases.

[0019] In some embodiments, the determining module is configured to: Based on the fused sound amplitude of each of the multiple target phases, the acoustic characteristics of the sound signal in multiple dimensions are determined; Based on the acoustic characteristics of the sound signal in multiple dimensions, the fault information of the impeller is determined.

[0020] In some embodiments, the determining module is configured to: A discrete Fourier transform is performed on the fused sound amplitude of each of the multiple target phases to obtain an order spectrum, which includes the order amplitude corresponding to each of the multiple orders. Based on the order spectrum, the acoustic features of the sound signal in multiple dimensions are determined.

[0021] In some embodiments, the determining module is configured to perform at least one of the following: Based on the first-order amplitude corresponding to the order spectrum, the frequency-reversing acoustic energy of the sound signal is determined, and the frequency-reversing acoustic energy is used to characterize the periodic disturbance of the impeller's first-order frequency imbalance. Based on the order amplitude values ​​corresponding to each order other than the first order in the order spectrum, the harmonic acoustic energy of the sound signal is determined, and the harmonic acoustic energy is used to characterize the composite periodic disturbance of the impeller harmonic nonlinearity. Based on the first-order amplitude corresponding to the first order in the order spectrum, the frequency conversion acoustic energy of the sound signal is determined. Based on the order amplitude corresponding to each order other than the first order in the order spectrum, the harmonic acoustic energy of the sound signal is determined. Based on the frequency conversion acoustic energy and the harmonic acoustic energy, the proportion of the frequency conversion acoustic energy in the total order acoustic energy is determined.

[0022] In some embodiments, the determining module is configured to perform at least one of the following: Based on the order amplitude corresponding to the target order in the order spectrum, the blade passage order energy of the sound signal is determined. The target order corresponds to the number of blades of the impeller. The blade passage order energy is used to characterize the energy intensity of the aerodynamic noise generated by the blade cutting the airflow. Based on the order amplitude of the order adjacent to the target order in the order spectrum, the modulation sideband energy of the sound signal is determined. The modulation sideband energy is used to characterize the degree of disturbance of the aerodynamic noise by the impeller imbalance.

[0023] In some embodiments, the determining module is configured to: Determine the statistical values ​​of the fused sound amplitude of each of the plurality of target phases, wherein the statistical values ​​include at least one of standard deviation, mean, and root mean square value; Based on the statistical values, the sound distribution characteristics of the impeller are determined, and the sound distribution characteristics are used to characterize the degree of non-uniformity of sound pressure distribution within a single rotation cycle.

[0024] In some embodiments, the determining module is configured to: For each dimension of acoustic features, a feature deviation value for that dimension is determined based on the standard acoustic features of that dimension, which are obtained under normal operating conditions of the wind turbine equipment; By integrating the feature deviation values ​​of the multiple dimensions, an imbalance index is obtained, which is used to indicate the overall degree of deviation of the acoustic features of the multiple dimensions relative to the standard acoustic features of the multiple dimensions. Based on the imbalance index, the fault information of the impeller is determined.

[0025] In some embodiments, the determining module is configured to: The imbalance index is matched with the imbalance index thresholds of multiple fault levels, and the fault level corresponding to the matched imbalance index threshold is determined as the fault level of the impeller.

[0026] In some embodiments, the determining module is further configured to: Based on the current operating condition type of the wind turbine equipment, a standard acoustic feature set corresponding to the current operating condition type is determined from the standard acoustic feature sets corresponding to each of the multiple operating condition types. Each standard acoustic feature set includes standard acoustic features of multiple dimensions.

[0027] In some embodiments, the determining module is configured to: Obtain the imbalance index of the sound signals from multiple second time periods; Based on the imbalance index of the sound signals in the first time period and the plurality of second time periods, the information on the change of the imbalance index is determined; Based on the information about the change in the imbalance index, the fault information of the impeller is determined.

[0028] In some embodiments, the fusion module is configured to: Based on the weights of the rotation periods in which the multiple sound amplitudes are located, the multiple sound amplitudes are weighted and averaged to obtain the fused sound amplitude of the target phase.

[0029] In some embodiments, the determining module is configured to: Based on the rotational speed at the multiple times, the rotational frequency at the multiple times is determined; For each time moment, the frequency conversion integral of the time moment and all time moments before the time moment is used to obtain the cumulative phase. Based on the cumulative phase and the initial phase, the phase corresponding to the time moment is obtained.

[0030] In some embodiments, the apparatus further includes a partitioning module for: A single rotation cycle is divided according to a preset angular interval to obtain the multiple target phases.

[0031] According to another aspect of the embodiments of the present disclosure, a computer device is provided, the computer device comprising: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the above-described fault detection method.

[0032] According to another aspect of the present disclosure, a computer-readable storage medium is provided, which, when the instructions in the computer-readable storage medium are executed by a processor of a computer device, enables the computer device to perform the above-described fault detection method.

[0033] According to another aspect of the present disclosure, a computer program product is provided, the computer program product including a computer program that, when executed by a processor, implements the above-described fault detection method.

[0034] This application provides a fault detection method. This method simultaneously collects sound signals and impeller speed from a fan during the same time period. Based on the real-time speed, it calculates the phase corresponding to each moment, achieving a precise mapping between acoustic amplitude and phase. Then, it divides the impeller's complete rotation cycle into multiple target phases, collects and fuses all sound amplitudes corresponding to the same target phase within different rotation cycles. Since the acoustic disturbances caused by impeller imbalance faults are periodically repetitive, they will stably appear at the same phase position in each rotation cycle. Thus, after fusing the sound amplitudes of the same phase across multiple rotation cycles, the periodic acoustic characteristics caused by impeller imbalance are superimposed and enhanced, while irregular background noise such as random spray noise, environmental noise, and random airflow disturbances are significantly attenuated, greatly improving the signal-to-noise ratio of the acoustic characteristics corresponding to the fault. Based on this, impeller fault detection is performed using the fused sound amplitudes of each target phase after signal-to-noise ratio optimization, which can accurately identify the acoustic characteristics corresponding to weak faults, effectively improving the accuracy and reliability of impeller imbalance fault detection.

[0035] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0036] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure, and are not intended to unduly limit this disclosure.

[0037] Figure 1 This is a schematic diagram illustrating an implementation environment according to an exemplary embodiment.

[0038] Figure 2 This is a flowchart illustrating a fault detection method according to an exemplary embodiment.

[0039] Figure 3 This is a flowchart illustrating a fault detection method according to an exemplary embodiment.

[0040] Figure 4 This is a schematic diagram of a multi-layer architecture according to an exemplary embodiment.

[0041] Figure 5 This is a flowchart illustrating a fault detection method according to an exemplary embodiment.

[0042] Figure 6 This is a block diagram illustrating a fault detection device according to an exemplary embodiment.

[0043] Figure 7 This is a schematic diagram of the structure of a terminal according to an exemplary embodiment.

[0044] Figure 8 This is a schematic diagram of the structure of a server according to an exemplary embodiment. Detailed Implementation

[0045] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings.

[0046] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0047] It should be noted that all information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.), and signals involved in this disclosure are authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. For example, the sound signals and speed data involved in this disclosure were obtained with full authorization.

[0048] The fault detection method provided in this application can be executed by a computer device, which can be at least one of a terminal and a server. The following is a schematic diagram of the implementation environment of the fault detection method provided in this application. See also... Figure 1 , Figure 1This is a schematic diagram of an implementation environment provided in an embodiment of this application. The implementation environment includes: terminal 101 and server 102.

[0049] In some embodiments, terminal 101 may run a client application with a target application, which provides a function to detect the operation of the wind turbine equipment in real time based on sound signals and impeller speed. This application embodiment does not limit the implementation form of the target application; for example, it may be an application that requires downloading and installation, a mini-program that does not require installation, a web application, etc.

[0050] In this embodiment, terminal 101 can be installed at multiple monitoring locations of the wind turbine equipment under test to acquire the equipment's sound signals and impeller speed. Server 102 provides background services for the target application. After terminal 101 acquires the sound signals and impeller speed data of the wind turbine equipment, it transmits the sound signals and speed data to server 102. Server 102 processes the sound signals and speed data to obtain the sound amplitude values ​​of multiple target phases in multiple rotation cycles. After fusing, it obtains the impeller fault information based on the fused sound amplitude values ​​and sends the fault information to terminal 101 for output.

[0051] In other embodiments, the terminal 101 itself can also acquire the sound signal and speed data of the wind turbine equipment, process the sound signal and speed data to obtain the sound amplitude of multiple target phases in multiple rotation cycles, and obtain the impeller fault information based on the fused sound amplitude after fusion, and output the fault information.

[0052] Terminal 101 can be a computer device such as an acoustic acquisition unit with signal processing and output display functions, a speed acquisition unit, a mobile phone, a tablet computer, a multimedia playback device, a PC (Personal Computer), a wearable device, a VR (Virtual Reality) device, an AR (Augmented Reality) device, or a MR (Mixed Reality) device. Server 102 can be an independent physical server, a server cluster consisting of multiple physical servers, or a distributed file system. It can also be a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN (Content Delivery Network), and big data and artificial intelligence platforms. Server 102 and terminal 101 are directly or indirectly connected via wired or wireless communication.

[0053] Figure 2This is a flowchart illustrating a fault detection method according to an exemplary embodiment, such as... Figure 2 As shown, the method is performed by a computer device and includes at least one of the following steps.

[0054] 201. Acquire the sound signal and speed data of the fan equipment in the first time period. The sound signal includes the sound amplitude at multiple times, and the speed data includes the speed of the impeller of the fan equipment at multiple times.

[0055] A fan is a power device equipped with a rotating impeller, suitable for various industrial applications. For example, a fan can be a cooling tower fan, an induced draft fan, a forced draft fan, or a centrifugal fan. The impeller has several blades that rotate synchronously with the shaft. During operation, the blades periodically cut the airflow, generating aerodynamic noise. When the impeller has defects such as fouling, wear, or uneven weight distribution, it will produce periodic imbalance disturbances, thus affecting the sound amplitude corresponding to each phase.

[0056] In this embodiment, the sound signal can be acquired by an acoustic acquisition unit, and the sound signal is the airborne sound signal acquired during the operation of the fan equipment. The acoustic acquisition unit can be an industrial microphone, a directional microphone, an array microphone, multiple distributed acoustic sensors, or a sound pressure sensor, etc., and is not specifically limited here. The type of sound amplitude value can be set as needed; for example, the unit of measurement for sound amplitude can be decibel (dB) or palladium (Pa). When the unit of sound amplitude is palladium, the sound amplitude specifically corresponds to the acquired instantaneous sound pressure value.

[0057] It should be noted that the acoustic acquisition unit and the fan equipment are arranged in a non-contact manner. The installation location of the fan equipment must comprehensively consider the following four conditions: effective reception intensity of impeller aerodynamic noise, avoidance of direct impact from spray water, isolation of interference noise from adjacent fans, and ensuring convenient maintenance and repair of the equipment in the future. For example, the acoustic acquisition unit can be deployed on the air inlet side, air outlet side, outer area of ​​the fan, or around the equipment maintenance platform.

[0058] The impeller is rigidly connected to the shaft of the fan and rotates synchronously with the shaft, outputting a unique real-time impeller speed at any given moment. The impeller speed can be acquired by a speed acquisition unit.

[0059] 202. Based on the rotational speed at multiple times, determine the phase corresponding to each of the multiple times.

[0060] The phase corresponding to each moment, which is also the rotation angle of the impeller at that moment, is used to identify the circumferential position of the impeller within a 360-degree rotation cycle, so as to achieve precise matching between the sound signal and the rotation angle.

[0061] 203. Based on the phase and sound amplitude corresponding to each of the multiple moments, determine the multiple sound amplitudes corresponding to each of the multiple target phases. The sound amplitude corresponding to each target phase is the sound amplitude of the target phase at the moment corresponding to one rotation cycle. One rotation cycle corresponds to one complete rotation of the impeller.

[0062] The rotation angle corresponding to one complete rotation of the impeller is 360 degrees. Therefore, any target phase appears only once in a single rotation cycle, corresponding to a unique moment within that single rotation cycle. The sound amplitude corresponding to that target phase is the sound amplitude at that corresponding moment.

[0063] The division rules of the target phase and the total number of target phases can be flexibly configured according to the fault detection accuracy and the computing power requirements of the equipment. This application embodiment does not make specific limitations here.

[0064] 204. For each target phase, fuse the multiple sound amplitudes corresponding to the target phase to obtain the fused sound amplitude of the target phase.

[0065] For each target phase, the fusion method for multiple sound amplitudes can be set as needed, such as summation, averaging, weighted summation, weighted average, median fusion, or robust fusion, etc., without specific limitations here. Median fusion sorts multiple sound amplitudes corresponding to the same target phase according to their numerical values ​​and selects the sound amplitude at the middle position of the sequence as the fusion result; if the number of sound amplitudes is even, the average of the two middle sound amplitudes is taken as the fusion output. Robust fusion is a general term for a class of fusion algorithms with the ability to resist outliers. It first identifies and removes abnormal sound amplitudes that deviate from the overall distribution, and then performs summation, averaging, weighted summation, or weighted average on the remaining valid sound amplitudes.

[0066] 205. Based on the fused sound amplitude of multiple target phases, determine the impeller's fault information.

[0067] The impeller fault information can be set as needed. In this embodiment, taking impeller imbalance fault as an example, the impeller fault information can include the degree of imbalance, imbalance level, etc. The causes of impeller imbalance faults include, but are not limited to, impeller fouling, corrosion, wear, deformation, installation misalignment, or counterweight detachment.

[0068] In this embodiment of the application, the sound signal and rotation speed data can be collected in real time during the operation of the wind turbine equipment, thereby enabling real-time fault detection of the impeller of the wind turbine equipment.

[0069] This application provides a fault detection method. This method simultaneously collects sound signals and impeller speed from a fan during the same time period. Based on the real-time speed, it calculates the phase corresponding to each moment, achieving a precise mapping between acoustic amplitude and phase. Then, it divides the impeller's complete rotation cycle into multiple target phases, collects and fuses all sound amplitudes corresponding to the same target phase within different rotation cycles. Since the acoustic disturbances caused by impeller imbalance faults are periodically repetitive, they will stably appear at the same phase position in each rotation cycle. Thus, after fusing the sound amplitudes of the same phase across multiple rotation cycles, the periodic acoustic characteristics caused by impeller imbalance are superimposed and enhanced, while irregular background noise such as random spray noise, environmental noise, and random airflow disturbances are significantly attenuated, greatly improving the signal-to-noise ratio of the acoustic characteristics corresponding to the fault. Based on this, impeller fault detection is performed using the fused sound amplitudes of each target phase after signal-to-noise ratio optimization, which can accurately identify the acoustic characteristics corresponding to weak faults, effectively improving the accuracy and reliability of impeller imbalance fault detection.

[0070] The above Figure 2 The diagram shown is merely the basic process of this disclosure. The following section, based on a specific implementation method, further elaborates on the solution provided in this disclosure. See also... Figure 3 , Figure 3 This is a flowchart illustrating a fault detection method according to an exemplary embodiment, the method being performed by a computer device, the method comprising at least one of the following steps.

[0071] 301. Acquire the sound signal and speed data of the fan equipment in the first time period. The sound signal includes the sound amplitude at multiple times, and the speed data includes the speed of the impeller of the fan equipment at multiple times.

[0072] In some embodiments, if the sound signal is acquired by a microphone, the sound amplitude at each moment is obtained based on the microphone's output voltage and sensitivity coefficient at that moment. This process can be described by the following formula (1): (1) in, Represents the time in the sound signal The amplitude of sound, measured in Pa (Pascal). Indicates the microphone at time The output voltage is expressed in volts (V). This indicates the microphone's sensitivity coefficient, measured in V / Pa.

[0073] In some embodiments, the speed acquisition unit can directly acquire the impeller speed at each moment, or it can acquire the speed of the drive motor of the fan equipment and then obtain the impeller speed based on the speed of the drive motor. When obtaining the impeller speed through the speed of the drive motor, that is, for each moment, the ratio of the speed of the drive motor at that moment to the transmission ratio (reduction ratio) is determined as the impeller speed at that moment. This process can be referred to in the following formula (2): (2) in, Indicates the time of the impeller rotational speed, Indicates the time when the drive motor is rotational speed, Indicates the transmission ratio or reduction ratio.

[0074] In some embodiments, before performing the fault detection method, the microphone sensitivity, acoustic sampling rate, number of blades, fixed delay between the acoustic channel and the rotational speed channel, sensor installation position and installation direction are calibrated. This calibration process is a multi-source sensor joint calibration, which mainly includes the following two items. (1) Acoustic channel calibration: This process applies a known sound pressure signal at a fixed distance using a standard sound source (such as a calibrator or a known sound pressure level generator), measures the microphone output voltage to obtain the sensitivity coefficient, and then establishes the conversion relationship between the microphone output voltage and the sound pressure (sound amplitude) based on the sensitivity coefficient. (2) Multi-channel time calibration: The fixed delay between the acoustic channel and the rotational speed channel is determined by a synchronous trigger signal or a step excitation source for subsequent time alignment. In addition, the number of blades, sensor installation position and direction are structural parameters, which are obtained by geometric measurement and equipment parameter solidification input system during the installation stage. They are used for subsequent acoustic feature calculations, such as calculating the energy of the blade passing through the order.

[0075] In some embodiments, the current operating condition parameters of the fan equipment are also collected by the operating condition acquisition unit. These parameters include, but are not limited to, one or more of the following: motor current, power, inverter frequency, spray status, cooling water flow rate, ambient temperature and humidity, ambient wind speed, and tower group operating status. Based on these parameters, the current operating condition type of the fan equipment can be determined. Specifically, based on these collected parameters, at least one operating condition can be determined, such as speed range, spray status, load status, or environmental conditions. Then, the current operating condition type is determined based on this at least one operating condition. In this embodiment, the determination of the operating condition type based on these four operating conditions is used as an example for explanation.

[0076] Optionally, various combinations of speed ranges, spray conditions, load conditions, and environmental conditions can be performed in advance to obtain multiple operating condition types. Then, based on the current operating conditions, the matching operating condition type is determined, thus obtaining the current operating condition type.

[0077] In some embodiments, timing alignment of multi-source signals such as sound signals, speed data, and operating parameters can be achieved using a unified time reference at the data acquisition end or data input port. For example, when synchronously sampling multi-source signals through an industrial acquisition card, edge acquisition terminal, or DAQ (Data Acquisition) module, a unified system timestamp is appended to each sampled data stream, generating a standardized data stream of "sampled value + timestamp," relying on a globally unified clock source (e.g., NTP network time protocol, PTP precise time protocol, or device local high-precision crystal oscillator clock). If sound signals, speed data, and operating parameters are acquired separately by multiple independent acquisition devices, the original timestamps of each stream are calibrated and aligned using a unified clock synchronization protocol and cross-device time deviation correction mechanism. This ensures that all multi-source signals belong to the same global time coordinate axis, achieving one-to-one matching of various monitoring data at the same time. This provides a precise timing foundation for subsequent data processing steps such as time window slicing, speed integral phase calculation, and sound amplitude and phase mapping.

[0078] In some embodiments, there may be a fixed delay when the sound acquisition channel, speed acquisition channel, and operating condition acquisition channel acquire data, so the delay needs to be compensated. The sound acquisition channel is the channel for acquiring sound signals, the speed acquisition channel is the channel for acquiring speed data, and the operating condition acquisition channel is the channel for acquiring operating condition parameters. Taking the sound acquisition channel and speed acquisition channel as examples, if there is a fixed delay between the sound acquisition channel and the speed acquisition channel, the delay can be compensated using the following formula (3): (3) in, For the moment in the sound signal Compensation for delayed sound amplitude, Indicates time sound amplitude, This involves a pre-defined fixed delay between the two channels. This step ensures that any moment in the audio signal accurately corresponds to the real-time rotational speed and phase of the impeller at that moment, avoiding phase mismatch caused by various delays in the two acquisition channels and clock deviations from multiple acquisition devices.

[0079] It should be noted that if the time synchronization error, i.e. the fixed delay, in the first time period exceeds the preset delay range, the first time period will be marked as a low-quality time period or an invalid time period and will not be used for fault detection for the time being, or the confidence level of the detection results in the first time period will be reduced when the detection results are output in the subsequent period.

[0080] In some embodiments, the audio signal after delay compensation is also processed. Preprocessing is performed, including but not limited to at least one of the following: DC removal, bandpass filtering, abnormal sound amplitude removal, background noise suppression, amplitude normalization, and signal quality evaluation. The purpose of preprocessing is not to directly conclude that the data is unbalanced, but to remove obviously invalid data and reduce the impact of water spray noise, occasional impacts, environmental wind noise, and acquisition anomalies on subsequent steps.

[0081] The process of removing DC can be found in the following formula (4): (4) in, Indicates time The sound amplitude after removing DC, Indicates time The sound amplitude before DC current is applied. This represents the average sound amplitude after delay compensation at multiple points within the first time period.

[0082] In this embodiment, the effective sound amplitude after DC removal, bandpass filtering, abnormal sound amplitude removal, background noise suppression, and amplitude normalization is uniformly denoted as... The process can be expressed as the following formula (5): (5) in, This indicates that the sound amplitude after DC removal at time t has undergone bandpass filtering, abnormal sound amplitude removal, and background noise suppression. This indicates that the sound amplitude at time t has been normalized.

[0083] In some embodiments, bandpass filtering and background noise suppression are performed on the sound signal based on the current rotational frequency, harmonics, blade pass order, and modulation sideband range of the wind turbine equipment. Specifically, the bandpass interval is determined based on the current rotational frequency, the number of blades, and the blade pass order. Then, bandpass filtering and background noise suppression are performed based on the bandpass interval and harmonics. For example, taking a rotational frequency of AHz, 3 blades, a blade pass order of 3AHz, a 4th-order modulation sideband, and a rotational frequency harmonic band of 1 to 5 times as an example, the bandpass interval is [3A-4×A, 3A+4×A]Hz, with a lower limit of 0Hz, that is, the actual bandpass interval is [0, 7A]Hz. The rotational frequency harmonic band is taken as 1 to 5 times: A, 2A, 3A, 4A, 5AHz, with each frequency point ±BHz narrow band. In this way, the bandpass interval only retains the above discrete harmonics + continuous sideband interval within 0~7AHz, and all background noise above 7AHz is directly filtered out.

[0084] In some embodiments, when collecting operating condition parameters, the sound signal can also be normalized based on these parameters. Operating condition normalization refers to the process where, during the operation of the wind turbine, changes in operating conditions such as spray status, load level, ambient temperature and humidity, and wind speed significantly affect the sound amplitude. Therefore, by introducing these operating condition parameters, the sound signal is corrected for comparability, mapping data under different operating conditions to a unified reference standard, thereby eliminating amplitude shifts or energy incomparability caused by differences in operating conditions. For example, the acoustic signal can be proportionally normalized (e.g., scaling the sound pressure amplitude according to rotational speed or load) to obtain a sound signal with a consistent statistical scale under different operating conditions, ensuring stable comparison and calculation of subsequently extracted features.

[0085] In some embodiments, to address the data misalignment problem caused by the mismatch in sampling frequencies among three types of multi-source signals—audio signals, speed data, and operating parameters—multi-sampling rate granularity normalization can be performed. To address the frequency difference between the high sampling frequency of audio signals and the low sampling frequencies of speed data and operating parameters, the sampling frequencies of the acquired signals are unified: for high-frequency audio signals, downsampling is used to reduce their sampling frequency to the target frequency; for low-frequency speed signals and operating parameter signals, sliding window average aggregation is used to coarseen their time granularity to the target frequency, thereby ensuring that the high-frequency audio signals, low-frequency speed signals, and operating parameter signals have consistent data point time intervals.

[0086] 302. Based on the rotational speed at multiple times, determine the phase corresponding to each of the multiple times.

[0087] In some embodiments, the process of determining the phase corresponding to each of the multiple times based on the rotational speed at multiple times includes the following implementation: determining the rotational frequency at multiple times based on the rotational speed at multiple times; for each time moment, integrating the rotational frequency at that time moment and all times before that time moment to obtain the cumulative phase; and obtaining the phase corresponding to that time moment based on the cumulative phase and the initial phase.

[0088] In this embodiment, the rotational speed at each moment is also the instantaneous rotational speed, and the rotational frequency at each moment is also the instantaneous rotational frequency. This process can be described by the following formula (6): (6) in, Indicates time frequency conversion, Indicates time The rotational speed, expressed in r / min. Indicates time The single rotation cycle. Based on the single rotation cycle, the sound signal in the first time period can be divided into multiple rotation cycles.

[0089] In this embodiment, the phase corresponding to the current moment is obtained based on the accumulated phase and the initial phase, that is, the sum of the accumulated phase and the initial phase is determined as the phase corresponding to the current moment. Accordingly, the cumulative phase is obtained by integrating the frequency of the current moment and all moments before the current moment. The process of obtaining the phase corresponding to the current moment based on the accumulated phase and the initial phase can be found in the following formula (7): (7) in, Indicates time The corresponding phase, Indicates the initial phase. Represents time from time 0 to time 1 The frequency integral represents the accumulated phase. The initial phase can be obtained through the phase acquisition unit, which serves as the reference for calculating the phase at each moment. Optionally, the phase acquisition unit uses a phase reference signal per revolution or a multi-pulse coded signal. If a phase reference signal per revolution is used, it means that one zero-point pulse is generated per revolution of the impeller, used to determine the starting point and cycle boundary of each revolution; if a multi-pulse coded signal is used, it means that multiple pulses are generated per revolution of the impeller, used to more precisely determine the specific angular position of the impeller within one rotation cycle.

[0090] In this embodiment, this step can convert the analysis reference of the acoustic signal from a fixed time coordinate to a phase coordinate of the impeller rotation. In this way, even if the impeller speed fluctuates, the first-order frequency, second-order frequency, blade passing order energy, and adjustment sideband energy related to impeller imbalance can still be stably tracked under a unified phase coordinate or order coordinate, avoiding the drift of the fixed frequency band characteristics of the sound signal due to changes in speed.

[0091] In this embodiment, the phase is solved by instantaneous frequency integration, adapting to the variable speed operation of the impeller. The cumulative phase is calculated using a time-by-time frequency integration accumulation method, which can accurately reproduce the continuous angular changes of the impeller's rotation. It can completely capture subtle angular shifts under continuous speed increases and decreases and instantaneous fluctuations. Compared to the estimation method of directly dividing 360 degrees at a fixed sampling interval, the phase calculation accuracy is significantly improved, and phase calculation errors caused by sudden changes in speed are avoided. The cumulative phase obtained by integration is uniformly calibrated based on the initial phase. The phase at all sampling moments is mapped to the same angular coordinate system, ensuring that the same physical rotation angle corresponds to the same phase under different rotation cycles and different speed ranges. This provides a precise phase matching basis for subsequent collection of sound amplitudes with the same phase across multiple rotation cycles, avoiding the weakening of fault characteristics and the decrease in signal-to-noise ratio caused by phase misalignment from the source.

[0092] In this embodiment, the phase at each moment is obtained by means of the rotation speed at each moment. In other embodiments, the phase at each moment can also be obtained directly by the phase acquisition unit.

[0093] 303. Based on the phase and sound amplitude corresponding to each of the multiple moments, determine the multiple sound amplitudes corresponding to each of the multiple target phases. The sound amplitude corresponding to each target phase is the sound amplitude of the target phase at the moment corresponding to one rotation cycle. One rotation cycle corresponds to one complete rotation of the impeller.

[0094] In some embodiments, multiple target phases can be obtained by dividing a single rotation cycle according to a preset angular interval to obtain multiple target phases.

[0095] The preset angle interval can be set as needed and is not specifically limited here. For example, if it is necessary to capture the acoustic features corresponding to subtle imbalances, the preset angle interval can be reduced to increase the number of target phases; if it is necessary to ensure computational efficiency, the preset angle interval can be increased to reduce the number of target phases.

[0096] For example, a single rotation cycle is divided into Then the equal-angle sampling points, then the The angle point (target phase) can be found in the following formula (8): (8) in, Indicates the first One target phase.

[0097] Accordingly, within each rotation cycle, the sound amplitude corresponding to each target phase is expressed by the following formula (9): (9) in, Indicates the first Within the first rotation cycle, the first The sound amplitude corresponding to each target phase Indicates the first Within the first rotation cycle, the first The time corresponding to each target phase This indicates the sound amplitude at that moment.

[0098] In this embodiment, a standardized, equally spaced angular coordinate system can be constructed by dividing the rotation cycle using a uniform fixed angular interval. Rotation cycles with different speeds and durations can be mapped to the same set of angular scales, eliminating the difference in rotation cycle duration caused by speed fluctuations. This ensures that positions with the same physical rotation angle in all rotation cycles correspond to the same target phase, providing a unified and regular matching benchmark for collecting in-phase sound amplitudes across rotation cycles.

[0099] In this embodiment, the multiple sound amplitudes corresponding to the multiple target phases constitute a multi-period angular domain acoustic matrix. It can be expressed as the following formula (10): (10) in, Indicates the number of rotation cycles. The matrix represents the number of target phases. The row dimensions correspond to different rotation periods in the matrix. Each column dimension corresponds to the same target phase. In this embodiment, this step unifies the sound signals collected at different rotation speeds and different single-rotation times into a sound signal expression based on the impeller rotation angle, i.e., the phase, providing a data foundation for subsequent processing steps.

[0100] In some embodiments, the sound signals for each of the multiple rotation cycles can also undergo periodic correction processing. This periodic correction can be performed based on standard parameters pre-calibrated by the fan equipment, where the standard parameters at least include a standard rotation cycle. Optionally, the current actual rotation cycle is first obtained based on the real-time rotation speed or phase pulse interval, and the difference between the actual rotation cycle and the standard rotation cycle is compared. If the deviation exceeds a preset threshold, the actual rotation cycle is modified using correction methods such as interpolation resampling, time axis stretching / compression, or phase compensation.

[0101] In the interpolation resampling method, real-time phase pulse signals are synchronously acquired to divide the single-cycle sound signal and establish a mapping relationship between time and angle. A uniform grid with equal angles corresponding to the standard rotation cycle is generated, such as 0°, 1°, 2°…359°. For each target angle, the corresponding target sound amplitude is calculated by interpolation between the two original angles on the left and right. That is, based on the ratio of the target angle to the distance between the two angles, different weighting coefficients are assigned to the actual sound amplitude. The sum of the product of the actual sound amplitude and the weighting coefficients is used as the target sound amplitude for that target angle. The target sound amplitudes corresponding to multiple target angles together form the corrected sound signal. After interpolation resampling is performed sequentially for all rotation cycles, the signals are spliced ​​or superimposed to complete the rotation cycle correction of the multi-cycle sound signal.

[0102] In the time-axis stretching and compression method, real-time phase pulse signals are synchronously acquired to divide the single-turn sound signal. A time scaling factor is calculated based on the actual rotation period and the standard rotation period, where the time scaling factor is the ratio of the standard rotation period to the actual rotation period. Based on the time scaling factor, the actual single-turn time-domain time axis is scaled as a whole, forcibly transforming the actual rotation period duration to the standard rotation period. Then, using the actual sound signal data as the reference data, upsampling or downsampling is performed to obtain the sound amplitude corresponding to the time point on the scaled time axis, which is also the sound amplitude corresponding to the standard rotation time axis. In this way, the duration of the single-turn sound signal is unified to the standard rotation period duration. After performing time-axis stretching and compression sequentially on all rotation periods, the signals are spliced ​​or superimposed to complete the rotation period correction of the multi-cycle sound signal.

[0103] In the phase compensation method, real-time phase pulse signals are synchronously acquired to divide the single-cycle sound signal, which is then converted to the complex frequency domain by Fourier transform. The phase compensation amount is calculated based on the current actual rotation period and the standard rotation period, and the reverse phase compensation amount is uniformly superimposed on the actual phase to offset the phase deviation. The inverse Fourier transform is performed using the complex spectrum after phase compensation to obtain the phase-aligned single-cycle time-domain sound signal. After phase compensation processing is performed on all rotation periods in sequence, the signals are spliced ​​or superimposed to complete the rotation period correction of the multi-cycle sound signal.

[0104] 304. For each target phase, fuse the multiple sound amplitudes corresponding to the target phase to obtain the fused sound amplitude of the target phase.

[0105] In some embodiments, the process of fusing multiple sound amplitudes corresponding to the target phase to obtain the fused sound amplitude of the target phase includes the following implementation: weighted averaging of multiple sound amplitudes based on the weights of the rotation periods in which each of the multiple sound amplitudes is located to obtain the fused sound amplitude of the target phase.

[0106] The weight of each rotation cycle can be set as needed and is not specifically limited here. For example, the weight is determined based on the quality of the sound signal in each rotation cycle, which can be based on factors such as the signal-to-noise ratio, amplitude stability, proportion of abnormal amplitudes, packet loss, and jet impact interference. When the quality of the sound signal is consistent across rotation cycles, the weights of multiple rotation cycles are the same, such as all being 1.

[0107] In this embodiment, the process of obtaining the fused sound amplitude can be referred to the following formula (11): (11) in, Indicates the first The fused sound amplitude of each target phase, Indicates the first Weights for each rotation cycle.

[0108] It should be noted that impeller imbalance typically leads to uneven distribution of aerodynamic loads, sound pressure levels, and periodic disturbances within one impeller rotation cycle. These acoustic disturbances usually have a stable correspondence with the impeller rotation phase. Therefore, during multi-rotation cycle synchronous fusion, the imbalanced acoustic components synchronized with the impeller rotation phase can be coherently enhanced. However, spray water noise, random airflow noise, ambient wind noise, adjacent equipment noise, and occasional impact noise typically do not have a stable correspondence with the impeller rotation phase and will be attenuated during multi-rotation cycle fusion. Therefore, this step essentially involves phase-locked demodulation or periodic acoustic demodulation of the wind turbine's acoustic characteristics using the real-time rotational speed as a synchronization reference. This demodulation process highlights the periodic imbalanced acoustic components hidden within complex background noise. This treatment improves the ability to identify early imbalance states such as blade fouling, minor deformation, installation deviations, corrosion wear, or counterweight misalignment.

[0109] In this embodiment, the sound amplitude of multiple rotation cycles in the same phase is weighted according to the weight of each rotation cycle. This can give high weight to high-quality rotation cycles and low weight to low-quality rotation cycles, avoiding the defect of simple arithmetic average being susceptible to single random noise pollution, strengthening the periodic fault characteristics and suppressing noise interference.

[0110] 305. Based on the fused sound amplitude of multiple target phases, determine the acoustic characteristics of the sound signal in multiple dimensions.

[0111] In some embodiments, the process of determining the acoustic features of a sound signal in multiple dimensions based on the fused sound amplitudes of multiple target phases includes the following implementation: performing a discrete Fourier transform on the fused sound amplitudes of multiple target phases to obtain an order spectrum, the order spectrum including the order amplitudes corresponding to multiple orders; and determining the acoustic features of the sound signal in multiple dimensions based on the order spectrum.

[0112] The process of performing discrete Fourier transform on the fused sound amplitudes of multiple target phases to obtain the order spectrum can be found in the following formula (12): (12) in, Indicates the first The magnitude of the order; When =1, The order amplitude corresponding to the first frequency conversion component, =1; When =2, The order amplitude corresponding to the second-fold frequency conversion component, =2; When =Z, it corresponds to the order amplitude of the order component passed by the blade, where Z is the number of blades. =Z. The maximum value can be set as needed, and no specific limit is given here.

[0113] In this embodiment, the discrete Fourier transform is first performed on the fused sound amplitudes of each target phase obtained from multi-rotation cycle fusion to generate an order spectrum, and then acoustic features are obtained based on the order spectrum. Since the multi-rotation cycle fusion step has significantly reduced background noise, the signal-to-noise ratio of the sound amplitude after discrete Fourier transform is significantly improved, and the amplitude of the fault-related order components in the generated order spectrum is more prominent, making the order spectrum features more recognizable. The order spectrum is divided into orders based on the impeller rotation cycle, with the horizontal axis representing the rotation order rather than a fixed frequency. This avoids the problem of traditional spectral features drifting with rotational speed and difficulty in locating fault components under variable impeller speed conditions. In the order spectrum, regardless of whether the impeller speed fluctuates, imbalance, blade modulation-related fault components, etc., are fixedly located at the first order, the blade passage order, and its sideband order. Thus, based on the order spectrum, multiple acoustic features such as frequency conversion acoustic energy, harmonic acoustic energy, blade passage order energy, and modulation sideband energy can be stably output, ensuring the stability and accuracy of fault detection.

[0114] In some embodiments, the process of determining the acoustic characteristics of a sound signal in multiple dimensions based on the order spectrum includes at least one of the following implementations: determining the frequency conversion acoustic energy of the sound signal based on the order amplitude corresponding to the first order in the order spectrum, wherein the frequency conversion acoustic energy is used to characterize the periodic disturbance of the impeller's first-order frequency imbalance; determining the harmonic acoustic energy of the sound signal based on the order amplitude corresponding to each order other than the first order in the order spectrum, wherein the harmonic acoustic energy is used to characterize the composite periodic disturbance of the impeller's harmonic nonlinearity; determining the frequency conversion acoustic energy of the sound signal based on the order amplitude corresponding to the first order in the order spectrum, determining the harmonic acoustic energy of the sound signal based on the order amplitude corresponding to each order other than the first order in the order spectrum, and determining the proportion of the frequency conversion acoustic energy in the total order acoustic energy based on the frequency conversion acoustic energy and the harmonic acoustic energy.

[0115] The process of determining the frequency-reversing acoustic energy of a sound signal based on the first-order amplitude corresponding to the first-order spectrum includes the following implementation: the square of the absolute value of the first-order amplitude corresponding to the first-order signal is determined as the frequency-reversing acoustic energy. This process can be referred to in the following formula (13): (13) in, Indicates frequency-reversing acoustic energy. This represents the magnitude of the first-order value.

[0116] The process of determining the harmonic acoustic energy of a sound signal based on the order amplitudes corresponding to each order other than the first order in the order spectrum includes the following implementation: the sum of the squares of the absolute values ​​of the order amplitudes corresponding to each order is determined as the harmonic acoustic energy. This process can be found in the following formula (14): (14) in, This represents harmonic acoustic energy. It represents the highest order in the order spectrum.

[0117] In some embodiments, the process of determining the proportion of the frequency-reversing acoustic energy in the total order acoustic energy based on the frequency-reversing acoustic energy and the harmonic acoustic energy includes the following implementation: determining the sum of the frequency-reversing acoustic energy and the harmonic acoustic energy to obtain the total order acoustic energy, and determining the ratio of the frequency-reversing acoustic energy to the total order acoustic energy as the proportion. Alternatively, the total order acoustic energy can be obtained by directly determining the sum of the squares of the absolute values ​​of the order amplitudes corresponding to each order in the order spectrum, and then obtaining the proportion based on the total order acoustic energy. This process can be referred to in the following formula (15): (15) in, Indicates percentage, This represents the total order acoustic energy. It is a very small constant to prevent the denominator from being zero.

[0118] In this embodiment, the first-order component of the order spectrum is extracted separately to calculate the frequency-reversing acoustic energy, which precisely corresponds to the imbalance disturbance that occurs only once per revolution of the impeller. This allows for a quantitative characterization of the strength of the first-order frequency-reversing imbalance fault, enabling a separate measurement of the core fault source. All higher-order components (excluding the first order) are aggregated to calculate the harmonic acoustic energy, uniformly summarizing the acoustic contributions of various nonlinear composite disturbances such as blade aerodynamic distortion and component misalignment. This provides a complete characterization of the overall level of non-imbalance-related high-order periodic noise during impeller operation. By calculating the energy ratio between the frequency-reversing acoustic energy and the harmonic acoustic energy, the dominant cause of acoustic anomalies can be intuitively distinguished: a high ratio indicates that the acoustic distortion is mainly caused by impeller imbalance; a low ratio indicates that the noise anomaly originates more from nonlinear airflow disturbances of the blades, assembly deviations, and other factors. Through the above-mentioned hierarchical quantification and ratio analysis, the acoustic changes caused by imbalance faults and other periodic noises can be distinguished, improving the reliability of impeller imbalance fault detection.

[0119] In some embodiments, the process of determining the acoustic characteristics of a sound signal in multiple dimensions based on the order spectrum includes at least one of the following implementation methods: determining the blade passage order energy of the sound signal based on the order amplitude corresponding to the target order in the order spectrum, wherein the target order corresponds to the number of blades of the impeller, and the blade passage order energy is used to characterize the energy intensity of the aerodynamic noise generated by the blade cutting the airflow; and determining the modulation sideband energy of the sound signal based on the order amplitude of the order adjacent to the target order in the order spectrum, wherein the modulation sideband energy is used to characterize the degree of disturbance of the aerodynamic noise by the impeller imbalance.

[0120] The process of determining the blade-passing energy of the sound signal based on the order amplitude corresponding to the target order in the order spectrum includes the following implementation: the square of the absolute value of the order amplitude corresponding to the target order is determined as the blade-passing energy. This process can be found in the following formula (16): (16) in, This indicates that the blade passes through order energy. This represents the magnitude of the order corresponding to the target order.

[0121] In some embodiments, the process of determining the modulation sideband energy of the sound signal based on the order amplitudes of orders adjacent to the target order in the order spectrum includes the following implementation: the sum of the squares of the absolute values ​​of the order amplitudes corresponding to each adjacent order is determined as the modulation sideband energy, which can be referred to in the following formula (17): (17) in, The modulated sideband energy is represented by L, which represents the order of the adjacent (sideband) bands participating in the statistics. If L is 1, the adjacent orders are Z-1 and Z+1; if L is 2, the adjacent orders are Z-1, Z+1, Z-2 and Z+2.

[0122] In this embodiment, the blade passage order energy is calculated separately for each target order corresponding to the number of blades. This allows for precise quantification of the fundamental inherent aerodynamic noise intensity generated by the periodic cutting of airflow by the blades. This energy reflects the baseline acoustic level under normal impeller operation without fault interference, providing a reliable benchmark for fault disturbances. Furthermore, the modulation sideband energy is obtained based on the amplitude of adjacent orders, specifically characterizing the amplitude modulation disturbance strength of impeller imbalance on aerodynamic noise. Since the periodic weight imbalance impact caused by impeller imbalance periodically changes the airflow velocity, it modulates the amplitude of aerodynamic noise, generating sideband components only on both sides of the blade passage order. The sideband energy can be intuitively and separately quantified to reflect the modulation effect caused by imbalance. By quantifying the fundamental aerodynamic noise benchmark energy and the imbalance modulation disturbance energy separately, a comparative relationship is established between the two. This shields against fundamental aerodynamic noise interference, accurately captures the acoustic characteristics corresponding to weak imbalances, and improves fault detection sensitivity.

[0123] In the above embodiments, acoustic features obtained based on the order spectrum are used as an example for illustration. In other embodiments, acoustic features can also be obtained in other ways. The process includes the following implementation: determining the statistical values ​​of the fused sound amplitude of each of multiple target phases, the statistical values ​​including at least one of standard deviation, mean and root mean square value; and determining the sound distribution characteristics of the impeller based on the statistical values, the sound distribution characteristics being used to characterize the non-uniformity of the sound pressure distribution within a single rotation cycle.

[0124] Taking the statistical values, including the standard deviation and the mean, as an example, the sound distribution characteristics determined based on the statistical values ​​include the periodic sound pressure fluctuation coefficient. The periodic sound pressure fluctuation coefficient is determined based on the absolute values ​​of the standard deviation and the mean. This process can be seen in the following formula (18): (18) in, This represents the periodic sound pressure fluctuation coefficient. Indicates standard deviation, This represents the mean. It is a very small constant to prevent the denominator from being zero.

[0125] Taking the root mean square (RMS) value as an example, the sound distribution characteristics determined based on the statistical values ​​include angular domain acoustic nonuniformity. The angular domain acoustic nonuniformity is determined based on the RMS value, the maximum value and the minimum value among multiple fused sound amplitudes. This process can be seen in the following formula (19): (19) in, Indicates acoustic inhomogeneity in the angular domain. Represents the root mean square value. Indicates the maximum value. This represents the minimum value. It is a very small constant to prevent the denominator from being zero.

[0126] In this embodiment, statistical calculations are performed based on the high signal-to-noise ratio sound amplitude after in-phase fusion of multiple rotation cycles. The mean value reflects the average level of the overall acoustic amplitude within a single rotation cycle, the root mean square quantifies the total sound pressure energy, and the standard deviation intuitively reflects the degree of dispersion and fluctuation of the sound amplitude in each phase. These three types of statistical values ​​characterize the sound pressure distribution within the cycle from three basic dimensions: average amplitude, total energy, and amplitude fluctuation. Impeller imbalance faults disrupt the uniformity of circumferential weight, causing significant differences in airflow compression intensity in different phases of the same rotation cycle, which directly manifests as increased sound amplitude dispersion and a significant increase in standard deviation. By quantifying the uniformity of sound pressure distribution within the cycle using the above statistical values, the probability of missed or false alarms due to imbalance faults can be effectively reduced, thereby improving the reliability of fault detection.

[0127] In this application embodiment, the above-mentioned acoustic features are used as examples for illustration. In other embodiments, acoustic consistency features and feature trend slopes over multiple rotation periods can also be extracted to determine whether the periodic acoustic components have a stable phase relationship and whether the imbalance fault continues to develop. The acoustic consistency feature of each acoustic feature can be represented by the standard deviation, mean, etc., of the acoustic feature over multiple rotation periods, and the feature trend slope of each acoustic feature can be represented by the trend slope of the acoustic feature over multiple rotation periods.

[0128] In the embodiments of this application, the above-mentioned acoustic features are all extracted based on the sound signal after synchronization with the rotational speed, which can form a clear correspondence with the impeller rotation period, frequency order, blade passing order energy and modulation sideband energy, thereby improving the mechanism interpretability of the fault detection results.

[0129] 306. Based on the acoustic characteristics of sound signals in multiple dimensions, determine the fault information of the impeller.

[0130] In some embodiments, the process of determining impeller fault information based on the acoustic characteristics of sound signals in multiple dimensions includes the following implementation: for each dimension of acoustic characteristics, determining the feature deviation value of that dimension based on the standard acoustic characteristics of that dimension, wherein the standard acoustic characteristics are obtained under normal operating conditions of the wind turbine equipment; fusing the feature deviation values ​​of multiple dimensions to obtain an imbalance index, wherein the imbalance index is used to indicate the overall degree of deviation of the acoustic characteristics of multiple dimensions relative to the standard acoustic characteristics of multiple dimensions; and determining the impeller fault information based on the imbalance index.

[0131] The feature deviation value can be directly the difference between the standard acoustic feature and the acoustic feature. In some other embodiments, the standard acoustic feature is acquired under normal operating conditions of the fan equipment. Therefore, the standard acoustic feature can be acquired multiple times under normal operating conditions, resulting in multiple standard acoustic features. Correspondingly, based on the standard deviation and mean of the multiple standard acoustic features, and the acoustic feature itself, the feature deviation value is determined. Specifically, the absolute value of the difference between the acoustic feature and the mean is determined, and the ratio of this absolute value to the standard deviation is determined to obtain the feature deviation value. This process can be referred to in the following formula (20): (20) in, Indicates the first Feature deviation values ​​in each dimension Indicates the first Acoustic characteristics in several dimensions Indicates the first The mean of multiple standard acoustic features in each dimension, Indicates the first The standard deviation of multiple standard acoustic features in each dimension. It is a very small constant to prevent the denominator from being zero.

[0132] In this embodiment, standard acoustic features pre-collected under normal operating conditions are used as a benchmark. Feature deviation values ​​are calculated in different dimensions. The abnormal offset of each dimension of acoustic features relative to the normal state can be quantified separately, accurately locating the feature dimension that causes the distortion due to the fault, and making the source of the anomaly traceable. Furthermore, the independent feature deviation values ​​of multiple dimensions are integrated into a single quantitative imbalance index, which normalizes and integrates the scattered and multi-dimensional acoustic anomaly indicators. The imbalance index comprehensively reflects the overall degree of anomaly in all acoustic dimensions. Compared with the method of judging faults based on single-dimensional features, it can avoid the error of single-dimensional features and greatly improve the accuracy of impeller imbalance fault detection.

[0133] In some embodiments, the process of fusing feature deviations from multiple dimensions to obtain an imbalance index includes the following implementation methods: determining the average of feature deviations from multiple dimensions as the imbalance index; or determining the weighted sum of feature deviations from multiple dimensions as the imbalance index.

[0134] The weight of each dimension can be set as needed. Optionally, the weight of each dimension can be determined based on at least one of the acoustic characteristics of each dimension, such as sensitivity to impeller imbalance, stability, anti-interference ability, and false alarm rate. The process of obtaining the imbalance index can be found in the following formula (21): (twenty one) in, Indicates an imbalance index, This represents the total number of multiple dimensions. Indicates the first Weights of acoustic features in each dimension. Indicates the first Feature deviation values ​​in each dimension.

[0135] In some embodiments, the process of determining impeller fault information based on the imbalance index includes the following implementation: matching the imbalance index with imbalance index thresholds of multiple fault levels, and determining the fault level corresponding to the matched imbalance index threshold as the fault level of the impeller.

[0136] The number of fault levels can be set as needed, and the imbalance index threshold for each fault level can also be set as needed, with multiple imbalance index thresholds. , , Perform fault classification. < < For multiple fault levels, please refer to the following formula (22): (twenty two) Four fault levels were determined based on three imbalance index thresholds.

[0137] In this embodiment, the multi-level threshold matching design can accurately distinguish different fault degrees such as slight, moderate, and severe impeller imbalance, going beyond simple fault presence or absence judgment. It enables refined fault classification diagnosis, providing precise data support for differentiated operation and maintenance of wind turbine equipment. This facilitates targeted inspections, maintenance, and shutdown repairs based on different fault levels. Furthermore, this classification matching method has simple computational logic and strong stability. It can quickly output fault levels based on standardized threshold comparison, adapting to online real-time monitoring scenarios for wind turbine equipment and efficiently and stably completing long-term fault classification and early warning.

[0138] In some embodiments, different operating condition types correspond to different standard acoustic features. For each dimension of acoustic features, before determining the feature deviation value of the dimension based on the standard acoustic features of the dimension, the following steps are also included: based on the current operating condition type of the wind turbine equipment, determining the standard acoustic feature set corresponding to the current operating condition type from the standard acoustic feature sets corresponding to each of the multiple operating condition types, wherein each standard acoustic feature set includes standard acoustic features of multiple dimensions.

[0139] Among them, the various operating conditions correspond to the normal operating conditions of the wind turbine equipment, that is, the standard acoustic feature sets corresponding to each of the various operating conditions are obtained in advance under the normal operating conditions of the wind turbine equipment.

[0140] In this embodiment, the inherent noise baseline of the wind turbine equipment varies significantly under different operating conditions. If a single fixed standard acoustic feature is used for comparison, changes in operating conditions will cause a huge baseline shift, which can easily lead to false alarms. However, this embodiment pre-stores independent standard acoustic feature sets according to different operating conditions, which can achieve adaptive benchmark matching and eliminate baseline interference caused by changing operating conditions. Furthermore, by establishing standards for different operating condition types, it is possible to accurately distinguish between inherent noise fluctuations under normal operating conditions and acoustic distortions caused by impeller imbalance. This significantly reduces misjudgments caused by fluctuations in normal operating conditions such as airflow regulation, start-stop load changes, and changes in ambient temperature and humidity. This allows for comparison calculations using a benchmark that fits the current operating conditions, regardless of the operating state of the wind turbine equipment, thereby improving the accuracy of the detection results.

[0141] In some embodiments, to reduce false alarms caused by short-term noise, occasional impacts, sudden changes in spray conditions, or instantaneous speed fluctuations, one or more of the following criteria are further introduced: duration criterion, continuous window consistency criterion, and trend criterion. Accordingly, the process of determining impeller fault information based on the imbalance index further includes the following implementation: acquiring the imbalance index of the sound signals for multiple second time periods; determining imbalance index change information based on the imbalance index of the sound signals for the first time period and the multiple second time periods; and determining impeller fault information based on the imbalance index change information.

[0142] The duration of the first time period and each of the second time periods can be the same or different. Each time period corresponds to an analysis window, and the window length of each analysis window is the duration of the corresponding time period.

[0143] In this context, multiple second time periods can all be periods preceding the first time period, periods following the first time period, or periods encompassing both. The first time period and multiple second time periods are collectively referred to as multiple time periods, arranged chronologically. Taking multiple time periods of equal duration as an example, the overall sound signal of the wind turbine equipment includes sound signals from multiple time periods. The multiple time periods involved in the calculation can be consecutive, or there can be one or more time periods between adjacent periods; no specific limitation is made here. For example, the overall sound signal of the wind turbine equipment includes sound signals from time periods 1, 2, 3, 4, 5, 6, 7, 8, and 9, but the time periods involved in the calculation only include time periods 1, 3, 5, 7, and 9.

[0144] The information on the change in the imbalance index can be the trend slope of the imbalance index over multiple time periods. The trend slope of the imbalance index can be calculated according to the following formula (23): (twenty three) in, Indicates the first time period among multiple time periods The slope of the imbalance index trend at each time period Indicates the first Imbalance index for each period This indicates the number of time intervals between any two adjacent time intervals in the multiple time intervals involved in the calculation. Then it means the ( ) Imbalance index for each time period This represents the time interval between any two adjacent time periods in the calculation, that is, the time interval between the start points of two time periods.

[0145] Taking the slope of the imbalance index trend over multiple time periods as an example, the explanation also includes multiple fault levels based on multiple slope thresholds. These fault levels can be the same as those in formula (22). The slope threshold corresponding to each fault level can be set as needed. Accordingly, when determining fault information based on the slope of the imbalance index trend over multiple time periods, if the slope of the imbalance index trend exceeds a certain slope threshold in a preset number of time periods, then the fault level corresponding to that slope threshold is determined as the fault level of the impeller.

[0146] In other embodiments, taking the imbalance index change information as an example of imbalance index over multiple time periods, if the imbalance index of a preset number of time periods exceeds a certain imbalance index threshold, that is, under abnormal conditions for a relatively long period of time, then the fault level corresponding to the imbalance index threshold is determined as the fault level of the impeller. It should be noted that if only the imbalance index of a single time period is abnormal, that is, the imbalance index of that time period is greater than the minimum imbalance index threshold, while the other time periods are not abnormal, then the fault result of this detection can be directly output as a low-confidence fault prompt, and the fault level will not be upgraded for the time being.

[0147] In this embodiment, relying solely on the instantaneous imbalance index for a single time period is prone to misjudgment due to accidental interference such as instantaneous airflow disturbances, spray impacts, and short-term load fluctuations. However, by introducing a multi-time period dimension and forming a time-series trend through multiple sets of imbalance indices, sudden changes in the instantaneous index caused by single random interference can be filtered out, significantly reducing errors caused by instantaneous anomalies. Furthermore, by obtaining imbalance index change information based on multi-time period imbalance indices, the evolution law of impeller imbalance faults can be fully captured. This allows for accurate detection of established stable imbalance faults and early detection of slowly deteriorating, subtle imbalance trends, enabling fault warnings and facilitating advance maintenance planning by operation and maintenance personnel. Compared to a single-time period static judgment method, this approach offers richer detection dimensions and is suitable for industrial applications requiring long-term online continuous monitoring of wind turbine equipment.

[0148] In some embodiments, the output of a lightweight machine learning model can be incorporated into the calculation of the imbalance index to assist in the determination. This lightweight machine learning model takes a feature vector composed of extracted multi-dimensional acoustic features as input and outputs the impeller imbalance probability or imbalance risk score. The lightweight machine learning model can employ lightweight network structures such as Logistic Regression (LR), Support Vector Machine (SVM), Shallow Random Forest (RF), or Small-Layer Perceptron (MLP), capable of performing nonlinear fusion operations on multi-dimensional acoustic features. The output of this lightweight machine learning model is then fused with the previously calculated imbalance index. For example, the fusion method involves weighting and correcting the original imbalance index based on the model output value to generate the final imbalance index used for fault detection. By leveraging the nonlinear fitting capability of the machine learning model for multi-dimensional features, the robustness and scenario generalization ability of impeller fault detection results under complex industrial conditions can be effectively improved.

[0149] In this embodiment, steps 305-306 above are used to determine the impeller fault information based on the fused sound amplitude of multiple target phases. In this embodiment, after fusing the in-phase amplitudes of multiple rotation cycles to obtain the fused sound amplitudes corresponding to each target phase, stable and high signal-to-noise ratio acoustic features can be extracted from multiple dimensions for fault identification. On the one hand, periodic acoustic distribution features in the angular domain can be extracted, and the mean, standard deviation, root mean square, periodic sound pressure fluctuation coefficient, and angular domain acoustic non-uniformity can be calculated based on the fused amplitudes to quantify the degree of sound pressure distribution non-uniformity within a single rotation cycle. On the other hand, order domain features can be constructed to solve for the frequency conversion acoustic energy, harmonic acoustic energy, blade passage order energy, and its modulation sideband energy, quantifying the intensity of periodic aerodynamic noise modulation disturbance caused by imbalance. In this way, the multi-dimensional acoustic features corroborate each other and are jointly judged, further reducing the probability of misjudgment and missed judgment caused by random noise, spray water flow, and interference from adjacent machines, and significantly improving the accuracy of impeller imbalance fault identification.

[0150] In this embodiment of the application, fault information is used as an example to illustrate the fault level. After obtaining the fault level, a fault report can be output based on it. The fault report includes, but is not limited to, the number of the wind turbine equipment, detection time, current operating speed, acoustic signal quality evaluation, imbalance index, fault level, anomaly confidence level, acoustic characteristics of the main anomalies, imbalance trend curve and suggested measures.

[0151] Among them, signal quality evaluation corresponds to the data validity screening mechanism in the preprocessing stage and before diagnosis. It mainly includes indicators such as signal-to-noise ratio assessment, amplitude stability, abnormal amplitude ratio, packet loss, and spray impact interference. These indicators are used to determine whether the current period is reliable, that is, whether the quality meets the standard. If the quality does not meet the standard, the weight of the data in that period participating in the fusion calculation can be reduced, or all sampled data in that period can be directly removed to avoid low-quality data interfering with the accuracy of the final result.

[0152] In some embodiments, recommended measures are generated in advance for different fault levels. For example, see Table 1 below, based on the imbalance index. With the imbalance index threshold , , Based on the relationship, the impeller failure level can be divided into four levels, and there are corresponding recommended measures for each level.

[0153] Table 1

[0154] Specifically, when the fault level is minor imbalance, the system outputs early warning information, prompting maintenance personnel to strengthen inspections and pay special attention to issues such as blade fouling, minor deformation, installation deviations, or counterweight misalignment. When the fault level is moderate imbalance, the system recommends scheduling blade cleaning, installation inspection, counterweight verification, or planned maintenance. When the fault level is severe imbalance and persists for more than the set abnormal duration, the system can output high-level alarms, maintenance work orders, speed reduction suggestions, shutdown suggestions, or linkage protection signals to PLCs (Programmable Logic Controllers), DCSs (Distributed Control Systems), SCADA (Supervisory Control And Data Acquisition) systems, industrial IoT platforms, equipment health management platforms, or maintenance management systems.

[0155] In this embodiment, suggestions are provided based on the fault level, including routine monitoring, enhanced inspections, planned maintenance, speed reduction, shutdown checks, or linkage protection. It should be noted that for outputs involving speed reduction, shutdown, or linkage protection, on-site safety policies, control permissions, and manual confirmation requirements must be met first to avoid triggering high-risk control actions based solely on a single abnormal period.

[0156] In some embodiments, the operation and maintenance results, re-inspection results, and manual confirmation information can also be fed back to the healthy acoustic benchmark library for subsequent threshold calibration and model updates, forming a closed-loop system from field data collection to operation and maintenance feedback. The healthy acoustic benchmark library includes standard acoustic features under various operating conditions. The updated model refers to the calculation model involved in this method, including but not limited to the following: healthy acoustic benchmark model, imbalance index calculation model, and classification and threshold model. The healthy acoustic benchmark model is used to establish data such as the mean and variance of standard acoustic features under normal operating conditions of the wind turbine equipment. The imbalance index calculation model is used to obtain the imbalance index. The classification and threshold model is used to classify the fault level. Accordingly, subsequent threshold calibration and model updates can include the following implementation methods: adjusting the mean and variance of standard acoustic features to adapt to equipment aging or changes in operating conditions; optimizing the weights of each acoustic feature to improve sensitivity to actual imbalance; and correcting the classification imbalance index threshold to reduce false alarms or missed alarms.

[0157] In this embodiment, the airborne acoustic signal during the operation of the wind turbine is the main monitoring object, and the real-time speed and phase information of the impeller or drive motor are used as the synchronous reference. Through non-contact acoustic signal sampling, speed / phase synchronous registration, angle domain phase resampling, periodic acoustic demodulation, order acoustic feature extraction, imbalance index calculation and graded fault determination, the online identification, quantitative assessment and trend early warning of impeller imbalance faults of the wind turbine are realized.

[0158] The method provided in this application differs from traditional acoustic feature analysis. Instead of directly sampling a fixed time window or fixed frequency band to diagnose the original sound signal, this method establishes a correspondence between the sampling time of the acoustic amplitude and the impeller rotation phase based on the real-time rotational speed and phase reference of the impeller. This converts the time-domain sound signal into an angle-domain sound signal synchronized with the impeller rotation cycle. In this way, the periodic acoustic components caused by impeller imbalance, such as rotational frequency, harmonics, blade passage order and its modulation sideband, are enhanced, while asynchronous interferences such as spray water noise, random airflow noise, environmental wind noise and noise from adjacent equipment are weakened. This improves the stability and interpretability of impeller imbalance fault diagnosis under complex wind turbine equipment operating conditions.

[0159] The method provided in this application embodiment can be implemented using a five-layer layered architecture. For example, see... Figure 4 , Figure 4This is a schematic diagram of a multi-layer architecture according to an exemplary embodiment. From bottom to top, it includes a perception layer, a synchronization and preprocessing layer, a speed-synchronized acoustic demodulation and feature extraction layer, a diagnostic decision and status assessment layer, and an application and early warning layer. The five layers sequentially transmit data and processing results according to the logic of "field signal acquisition—multi-source data alignment—speed-synchronized acoustic generation—imbalance status judgment—alarm and maintenance closed loop." Lower layers provide reliable input to upper layers, and upper layers provide diagnostic interpretation and application feedback to the output of lower layers, thus forming a complete online detection link for impeller imbalance faults in wind turbine equipment.

[0160] The perception layer, located at the bottom of the system, is responsible for acquiring field signals and inputting raw data. It mainly consists of an acoustic acquisition unit, a speed / phase acquisition unit, and an optional auxiliary acquisition unit for operating conditions. The perception layer does not perform fault diagnosis; its role is to provide subsequent layers with time-continuous, clearly sourced, and calibrable basic data. The specific functions of this layer have been described in detail in the above embodiments and will not be repeated here.

[0161] The synchronization and preprocessing layer, located above the sensing layer, addresses inconsistencies in sampling frequency, transmission link, and timestamps between multi-source signals. This layer first aligns the audio signal, rotational speed data, and operating parameters under a unified time reference, compensates for fixed delays between different acquisition channels based on the calibration results, and preprocesses the audio signal. The specific functions of this layer have been described in detail in the above embodiments and will not be repeated here.

[0162] The rotational speed synchronized acoustic demodulation and feature extraction layer is the core computational layer of this scheme. It receives the valid acoustic signal output from the synchronization and preprocessing layer, as well as rotational speed / phase reference information. This layer converts the acoustic signal at a fixed time coordinate into an acoustic expression based on the impeller rotation angle, enabling acoustic data at different rotational speeds and different single-rotation times to be compared in the same angular coordinate. The specific functions of this layer have been described in detail in the above embodiments and will not be repeated here.

[0163] The diagnostic decision and condition assessment layer receives synchronous acoustic characteristics of the rotational speed, standard acoustic characteristics, and operating condition type to quantitatively determine the impeller imbalance state. This layer first selects a matching standard acoustic characteristic based on the current rotational speed range, spray status, load status, and environmental conditions, and then calculates the degree of deviation of the real-time characteristics from the standard acoustic characteristics. The specific functions of this layer have been described in detail in the above embodiments and will not be repeated here.

[0164] The application and early warning layer is located at the top of the system. It is used to present the detection results to on-site personnel and the upper-level system, and to interact with PLCs, DCS, SCADA, industrial IoT platforms, equipment health management platforms, or operation and maintenance management systems to output fault reports. The specific functions of this layer have been described in detail in the above embodiments and will not be repeated here.

[0165] The method provided in this application is based on the above-described five-layer layered architecture. See [link to application]. Figure 5 , Figure 5 This is a flowchart illustrating a method according to an exemplary embodiment. The overall process is executed in the following order: installation calibration and standard establishment—multi-source signal synchronous acquisition—time alignment and delay compensation—acoustic preprocessing and quality evaluation—instantaneous frequency conversion, single-cycle period and continuous phase calculation—rotational period framing and angle domain resampling—multi-cycle synchronous fusion and acoustic demodulation—order acoustic feature extraction—standard comparison, imbalance index calculation and fault classification—detection result output and trend warning.

[0166] The solutions provided in this application can achieve at least the following technical effects.

[0167] 1. Improved reliability of non-contact detection. This solution uses an acoustic acquisition unit to collect the sound signal of the fan operation, eliminating the need to install contact vibration sensors on the bearing housing, reducer housing, or impeller structure. This reduces problems such as sensor loosening, cable aging, and signal drift under high humidity, water mist, corrosion, and long-term vibration environments.

[0168] 2. Enhanced fault localization. This solution extracts acoustic features directly related to the impeller rotation cycle from the sound signal. Through acoustic features such as frequency conversion acoustic energy, harmonic acoustic energy, periodic sound pressure fluctuations, and angular domain acoustic inhomogeneity, it more accurately characterizes impeller mass distribution unevenness and aerodynamic disturbance anomalies.

[0169] 3. Improved robustness in complex noise environments. This solution performs periodic alignment, angular domain resampling, and synchronous demodulation of the acoustic signal based on real-time rotational speed / phase. This enhances the unbalanced acoustic components that are synchronized with the impeller rotation cycle, while suppressing asynchronous interference such as spray water noise, airflow noise, ambient wind noise, and noise from adjacent equipment.

[0170] 4. Improved adaptability to variable speed conditions. This solution maps acoustic signals from the time domain to the rotating phase domain or order domain, enabling features such as frequency-modulated acoustic energy, harmonic acoustic energy, blade passage order energy, and their modulation sideband energy to adaptively track the impeller's real-time speed, reducing feature drift and misjudgment risks caused by speed fluctuations.

[0171] 5. Enhanced early imbalance detection capability. This solution enhances the characteristics of periodic aerodynamic disturbances caused by impeller imbalance through synchronous acoustic demodulation of the impeller speed, enabling the identification of early imbalance conditions such as blade fouling, minor deformation, installation deviation, corrosion wear, or counterweight misalignment.

[0172] 6. Enhanced Quantification and Interpretability of Diagnostic Results. This solution constructs an imbalance index based on the impeller imbalance mechanism and outputs fault level, anomaly confidence level, and trend change information, which can support graded early warning, trend tracking, maintenance decision-making, and predictive maintenance.

[0173] This application provides a fault detection method. This method simultaneously collects sound signals and impeller speed from a fan during the same time period. Based on the real-time speed, it calculates the phase corresponding to each moment, achieving a precise mapping between acoustic amplitude and phase. Then, it divides the impeller's complete rotation cycle into multiple target phases, collects and fuses all sound amplitudes corresponding to the same target phase within different rotation cycles. Since the acoustic disturbances caused by impeller imbalance faults are periodically repetitive, they will stably appear at the same phase position in each rotation cycle. Thus, after fusing the sound amplitudes of the same phase across multiple rotation cycles, the periodic acoustic characteristics caused by impeller imbalance are superimposed and enhanced, while irregular background noise such as random spray noise, environmental noise, and random airflow disturbances are significantly attenuated, greatly improving the signal-to-noise ratio of the acoustic characteristics corresponding to the fault. Based on this, impeller fault detection is performed using the fused sound amplitudes of each target phase after signal-to-noise ratio optimization, which can accurately identify the acoustic characteristics corresponding to weak faults, effectively improving the accuracy and reliability of impeller imbalance fault detection.

[0174] Figure 6 This is a block diagram illustrating a fault detection device according to an exemplary embodiment. (Refer to...) Figure 6 The device includes: The acquisition module 601 is used to acquire the sound signal and speed data of the wind turbine equipment in the first time period. The sound signal includes the sound amplitude at multiple times, and the speed data includes the speed of the impeller of the wind turbine equipment at multiple times. The determining module 602 is used to determine the phase corresponding to each of the multiple rotational speeds at multiple times. The determining module 602 is also used to determine multiple sound amplitudes corresponding to multiple target phases based on the phases and sound amplitudes corresponding to multiple moments. The sound amplitude corresponding to each target phase is the sound amplitude of the target phase at the moment corresponding to one rotation cycle. One rotation cycle corresponds to one complete rotation of the impeller. The fusion module 603 is used to fuse multiple sound amplitudes corresponding to each target phase to obtain the fused sound amplitude of the target phase. The determination module 602 is also used to determine the impeller fault information based on the fused sound amplitude of multiple target phases.

[0175] In some embodiments, the determining module 602 is configured to: Based on the fused sound amplitude of multiple target phases, the acoustic characteristics of the sound signal in multiple dimensions are determined; Based on the acoustic characteristics of sound signals in multiple dimensions, the fault information of the impeller is determined.

[0176] In some embodiments, the determining module 602 is configured to: The discrete Fourier transform is performed on the fused sound amplitude of multiple target phases to obtain the order spectrum, which includes the order amplitude corresponding to each of the multiple orders. Based on the order spectrum, the acoustic characteristics of the sound signal in multiple dimensions are determined.

[0177] In some embodiments, the determining module 602 is configured to perform at least one of the following: Based on the first-order amplitude corresponding to the order spectrum, the frequency-reversing acoustic energy of the sound signal is determined. The frequency-reversing acoustic energy is used to characterize the periodic disturbance of the impeller's first-order frequency imbalance. Based on the order amplitudes corresponding to each order other than the first order in the order spectrum, the harmonic acoustic energy of the sound signal is determined. The harmonic acoustic energy is used to characterize the composite periodic disturbance of the impeller's harmonic nonlinearity. Based on the order amplitude corresponding to the first order in the order spectrum, the frequency conversion acoustic energy of the sound signal is determined. Based on the order amplitude corresponding to each order other than the first order in the order spectrum, the harmonic acoustic energy of the sound signal is determined. Based on the frequency conversion acoustic energy and the harmonic acoustic energy, the proportion of frequency conversion acoustic energy in the total order acoustic energy is determined.

[0178] In some embodiments, the determining module 602 is configured to perform at least one of the following: Based on the order amplitude corresponding to the target order in the order spectrum, the blade passage order energy of the sound signal is determined. The target order corresponds to the number of blades in the impeller. The blade passage order energy is used to characterize the energy intensity of the aerodynamic noise generated by the blade cutting the airflow. Based on the order amplitude of the order adjacent to the target order in the order spectrum, the modulation sideband energy of the sound signal is determined. The modulation sideband energy is used to characterize the degree of disturbance of aerodynamic noise by impeller imbalance.

[0179] In some embodiments, the determining module 602 is configured to: Determine the statistical values ​​of the fused sound amplitude for each of the multiple target phases, including at least one of the standard deviation, mean, and root mean square value; Based on statistical values, the sound distribution characteristics of the impeller are determined, and these characteristics are used to characterize the non-uniformity of sound pressure distribution within a single rotation cycle.

[0180] In some embodiments, the determining module 602 is configured to: For each dimension of acoustic features, the feature deviation value of the dimension is determined based on the standard acoustic features of the dimension. The standard acoustic features are obtained under normal operating conditions of the wind turbine equipment. By integrating feature deviation values ​​from multiple dimensions, an imbalance index is obtained. The imbalance index is used to indicate the overall degree of deviation of acoustic features from standard acoustic features in multiple dimensions. Based on the imbalance index, the impeller's fault information is determined.

[0181] In some embodiments, the determining module 602 is configured to: The imbalance index is matched with the imbalance index thresholds of multiple fault levels, and the fault level corresponding to the matched imbalance index threshold is determined as the fault level of the impeller.

[0182] In some embodiments, the determining module 602 is further configured to: Based on the current operating condition type of the wind turbine equipment, the standard acoustic feature set corresponding to the current operating condition type is determined from the standard acoustic feature sets corresponding to each of the various operating condition types. Each standard acoustic feature set includes standard acoustic features of multiple dimensions.

[0183] In some embodiments, the determining module 602 is configured to: Obtain the imbalance index of the sound signals from multiple second time periods; Based on the imbalance index of the sound signals in the first time period and multiple second time periods, the information on the change of the imbalance index is determined; Based on the information on the change in the imbalance index, the fault information of the impeller is determined.

[0184] In some embodiments, the fusion module 603 is configured to: Based on the weights of the rotation periods in which each of the multiple sound amplitudes resides, a weighted average of the multiple sound amplitudes is calculated to obtain the fused sound amplitude of the target phase.

[0185] In some embodiments, the determining module 602 is configured to: Based on the rotational speed at multiple times, determine the rotational frequency at multiple times; For each time step, the frequency conversion is integrated over time step and all time steps prior to time step to obtain the cumulative phase. Based on the cumulative phase and the initial phase, the phase corresponding to time step is obtained.

[0186] In some embodiments, the apparatus further includes a partitioning module for: A single rotation cycle is divided into multiple target phases by a preset angular interval.

[0187] This application provides a fault detection device that simultaneously collects sound signals and impeller speed of a fan equipment at the same time period. Based on the real-time speed, it calculates the phase corresponding to each moment, achieving a precise mapping between acoustic amplitude and phase. Then, it divides the impeller's complete rotation cycle into multiple target phases, collects and fuses all sound amplitudes corresponding to the same target phase within different rotation cycles. Since the acoustic disturbances caused by impeller imbalance faults are periodically repetitive, they will stably appear at the same phase position in each rotation cycle. Thus, after fusing the sound amplitudes of the same phase across multiple rotation cycles, the periodic acoustic characteristics caused by impeller imbalance are superimposed and enhanced, while irregular background noise such as random spray noise, environmental noise, and random airflow disturbances are significantly attenuated, greatly improving the signal-to-noise ratio of the acoustic characteristics corresponding to the fault. Based on this, impeller fault detection is performed using the fused sound amplitudes of each target phase after signal-to-noise ratio optimization, which can accurately identify the acoustic characteristics corresponding to weak faults, effectively improving the accuracy and reliability of impeller imbalance fault detection.

[0188] Regarding the apparatus in the above embodiments, the specific manner in which each unit performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0189] In some embodiments, the computer device is provided as a terminal. Figure 7 A schematic diagram of the structure of a terminal 700 provided in an exemplary embodiment of this disclosure is shown. The terminal 700 may be a smartphone, tablet computer, MP3 player (Moving Picture Experts Group Audio Layer III), MP4 player (Moving Picture Experts Group Audio Layer IV), laptop computer, or desktop computer. The terminal 700 may also be referred to as a user device, portable terminal, laptop terminal, desktop terminal, or other names.

[0190] Typically, terminal 700 includes a processor 701 and a memory 702.

[0191] Processor 701 may include one or more processing cores, such as a quad-core processor, an octa-core processor, etc. Processor 701 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). Processor 701 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, processor 701 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, processor 701 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0192] The memory 702 may include one or more computer-readable storage media, which may be non-transitory. The memory 702 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In some embodiments, the non-transitory computer-readable storage media in the memory 702 are used to store at least one program code, which is executed by the processor 701 to implement the fault detection method provided in the method embodiments of this disclosure.

[0193] In some embodiments, the terminal 700 may also optionally include a peripheral device interface 703 and at least one peripheral device. The processor 701, memory 702, and peripheral device interface 703 can be connected via a bus or signal line. Each peripheral device can be connected to the peripheral device interface 703 via a bus, signal line, or circuit board. Specifically, the peripheral device includes at least one of the following: a radio frequency circuit 704, a display screen 705, a camera assembly 706, an audio circuit 707, and a power supply 708.

[0194] Peripheral device interface 703 can be used to connect at least one I / O (Input / Output) related peripheral device to processor 701 and memory 702. In some embodiments, processor 701, memory 702 and peripheral device interface 703 are integrated on the same chip or circuit board; in some other embodiments, any one or two of processor 701, memory 702 and peripheral device interface 703 can be implemented on separate chips or circuit boards, which is not limited in this embodiment.

[0195] The radio frequency (RF) circuit 704 is used to receive and transmit RF (Radio Frequency) signals, also known as electromagnetic signals. The RF circuit 704 communicates with communication networks and other communication devices via electromagnetic signals. The RF circuit 704 converts electrical signals into electromagnetic signals for transmission, or converts received electromagnetic signals back into electrical signals. Optionally, the RF circuit 704 includes: an antenna system, an RF transceiver, one or more amplifiers, a tuner, an oscillator, a digital signal processor, a codec chipset, a user identity module card, etc. The RF circuit 704 can communicate with other terminals through at least one wireless communication protocol. This wireless communication protocol includes, but is not limited to: metropolitan area networks (MANs), various generations of mobile communication networks (2G, 3G, 4G, and 5G), wireless local area networks (WLANs), and / or WiFi (Wireless Fidelity) networks. In some embodiments, the RF circuit 704 may also include circuitry related to NFC (Near Field Communication), which is not limited in this disclosure.

[0196] Display screen 705 is used to display a UI (User Interface). This UI may include graphics, text, icons, videos, and any combination thereof. When display screen 705 is a touch display screen, it also has the ability to collect touch signals on or above its surface. These touch signals can be input as control signals to processor 701 for processing. In this case, display screen 705 can also be used to provide virtual buttons and / or a virtual keyboard, also known as soft buttons and / or a soft keyboard. In some embodiments, there may be one display screen 705, which serves as the front panel of terminal 700; in other embodiments, there may be at least two display screens 705, respectively disposed on different surfaces of terminal 700 or in a folded design; in still other embodiments, display screen 705 may be a flexible display screen, disposed on a curved or folded surface of terminal 700. Furthermore, display screen 705 may be configured as a non-rectangular, irregular shape, i.e., a non-rectangular screen. Display screen 705 may be made of materials such as LCD (Liquid Crystal Display) or OLED (Organic Light-Emitting Diode).

[0197] The camera assembly 706 is used to acquire images or videos. Optionally, the camera assembly 706 includes a front-facing camera and a rear-facing camera. Typically, the front-facing camera is located on the front panel of the terminal, and the rear-facing camera is located on the back of the terminal. In some embodiments, there are at least two rear-facing cameras, which are any one of a main camera, a depth-sensing camera, a wide-angle camera, and a telephoto camera, to achieve background blurring by fusion of the main camera and the depth-sensing camera, panoramic shooting by fusion of the main camera and the wide-angle camera, VR (Virtual Reality) shooting, or other fusion shooting functions. In some embodiments, the camera assembly 706 may also include a flash. The flash can be a single-color temperature flash or a dual-color temperature flash. A dual-color temperature flash refers to a combination of a warm-light flash and a cool-light flash, which can be used for light compensation at different color temperatures.

[0198] The audio circuit 707 may include a microphone and a speaker. The microphone is used to collect sound waves from the user and the environment, converting the sound waves into electrical signals that are input to the processor 701 for processing, or input to the radio frequency circuit 704 for voice communication. For stereo sound acquisition or noise reduction purposes, multiple microphones may be used, each located at a different part of the terminal 700. The microphone may also be an array microphone or an omnidirectional microphone. The speaker is used to convert the electrical signals from the processor 701 or the radio frequency circuit 704 into sound waves. The speaker may be a conventional diaphragm speaker or a piezoelectric ceramic speaker. When the speaker is a piezoelectric ceramic speaker, it can convert electrical signals not only into audible sound waves but also into inaudible sound waves for purposes such as distance measurement. In some embodiments, the audio circuit 707 may also include a headphone jack.

[0199] Power supply 708 is used to power the various components in terminal 700. Power supply 708 can be AC ​​power, DC power, a disposable battery, or a rechargeable battery. When power supply 708 includes a rechargeable battery, the rechargeable battery can support wired charging or wireless charging. The rechargeable battery can also be used to support fast charging technology.

[0200] Those skilled in the art will understand that Figure 7 The structure shown does not constitute a limitation on terminal 700, and may include more or fewer components than shown, or combine certain components, or use different component arrangements.

[0201] Figure 8 This is a schematic diagram of a server structure according to an embodiment of this application. The server 800 can vary considerably due to different configurations or performance. It may include one or more Central Processing Units (CPUs) 801 and one or more memories 802. The memories 802 are used to store executable program code, and the processors 801 are configured to execute the executable program code to implement the fault detection methods provided in the various method embodiments described above. Of course, the server may also have wired or wireless network interfaces, a keyboard, and input / output interfaces for input and output. The server may also include other components for implementing device functions, which will not be elaborated here.

[0202] In an exemplary embodiment, a computer-readable storage medium is also provided, which, when executed by a processor of a computer device, enables the computer device to perform the aforementioned fault detection method. Optionally, the computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, or optical data storage device, etc.

[0203] In an exemplary embodiment, a computer program product is also provided, the computer program product including a computer program that, when executed by a processor, implements the above-described fault detection method.

[0204] In some embodiments, the computer program product involved in this disclosure may be deployed and executed on a computer device, or on multiple computer devices located in one location, or on multiple computer devices distributed in multiple locations and interconnected through a communication network. Multiple computer devices distributed in multiple locations and interconnected through a communication network may constitute a blockchain system.

[0205] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims. All the above-described optional technical solutions can be combined in any way to form optional embodiments of this application, and will not be elaborated upon here.

[0206] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A fault detection method, characterized in that, The method includes: Acquire sound signals and rotational speed data of the wind turbine equipment within a first time period, wherein the sound signals include sound amplitude values ​​at multiple times, and the rotational speed data includes the rotational speed of the impeller of the wind turbine equipment at the multiple times; Based on the rotational speed at the multiple moments, determine the phase corresponding to each of the multiple moments; Based on the phase and sound amplitude corresponding to each of the multiple moments, multiple sound amplitudes corresponding to each of the multiple target phases are determined. The sound amplitude corresponding to each target phase is the sound amplitude at the moment corresponding to the target phase within one rotation cycle. One rotation cycle corresponds to one complete rotation of the impeller. For each target phase, multiple sound amplitudes corresponding to the target phase are fused to obtain the fused sound amplitude of the target phase; Based on the fused sound amplitude of each of the multiple target phases, the fault information of the impeller is determined.

2. The method according to claim 1, characterized in that, The determination of the impeller's fault information based on the fused sound amplitude of each of the multiple target phases includes: Based on the fused sound amplitude of each of the multiple target phases, the acoustic characteristics of the sound signal in multiple dimensions are determined; Based on the acoustic characteristics of the sound signal in multiple dimensions, the fault information of the impeller is determined.

3. The method according to claim 2, characterized in that, The determination of the acoustic features of the sound signal in multiple dimensions based on the fused sound amplitude of each of the multiple target phases includes: A discrete Fourier transform is performed on the fused sound amplitude of each of the multiple target phases to obtain an order spectrum, which includes the order amplitude corresponding to each of the multiple orders. Based on the order spectrum, the acoustic features of the sound signal in multiple dimensions are determined.

4. The method according to claim 3, characterized in that, The determination of the acoustic features of the sound signal in multiple dimensions based on the order spectrum includes at least one of the following: Based on the first-order amplitude corresponding to the order spectrum, the frequency-reversing acoustic energy of the sound signal is determined, and the frequency-reversing acoustic energy is used to characterize the periodic disturbance of the impeller's first-order frequency imbalance. Based on the order amplitude values ​​corresponding to each order other than the first order in the order spectrum, the harmonic acoustic energy of the sound signal is determined, and the harmonic acoustic energy is used to characterize the composite periodic disturbance of the impeller harmonic nonlinearity. Based on the first-order amplitude corresponding to the first order in the order spectrum, the frequency conversion acoustic energy of the sound signal is determined. Based on the order amplitude corresponding to each order other than the first order in the order spectrum, the harmonic acoustic energy of the sound signal is determined. Based on the frequency conversion acoustic energy and the harmonic acoustic energy, the proportion of the frequency conversion acoustic energy in the total order acoustic energy is determined.

5. The method according to claim 3, characterized in that, The determination of the acoustic features of the sound signal in multiple dimensions based on the order spectrum includes at least one of the following: Based on the order amplitude corresponding to the target order in the order spectrum, the blade passage order energy of the sound signal is determined. The target order corresponds to the number of blades of the impeller. The blade passage order energy is used to characterize the energy intensity of the aerodynamic noise generated by the blade cutting the airflow. Based on the order amplitude of the order adjacent to the target order in the order spectrum, the modulation sideband energy of the sound signal is determined. The modulation sideband energy is used to characterize the degree of disturbance of the aerodynamic noise by the impeller imbalance.

6. The method according to claim 2, characterized in that, The determination of the acoustic features of the sound signal in multiple dimensions based on the fused sound amplitude of each of the multiple target phases includes: Determine the statistical values ​​of the fused sound amplitude of each of the plurality of target phases, wherein the statistical values ​​include at least one of standard deviation, mean, and root mean square value; Based on the statistical values, the sound distribution characteristics of the impeller are determined, and the sound distribution characteristics are used to characterize the degree of non-uniformity of sound pressure distribution within a single rotation cycle.

7. The method according to claim 2, characterized in that, The process of determining the impeller's fault information based on the acoustic characteristics of the sound signal in multiple dimensions includes: For each dimension of acoustic features, a feature deviation value for that dimension is determined based on the standard acoustic features of that dimension, which are obtained under normal operating conditions of the wind turbine equipment; By integrating the feature deviation values ​​of the multiple dimensions, an imbalance index is obtained, which is used to indicate the overall degree of deviation of the acoustic features of the multiple dimensions relative to the standard acoustic features of the multiple dimensions. Based on the imbalance index, the fault information of the impeller is determined.

8. The method according to claim 7, characterized in that, The process of determining the impeller's fault information based on the imbalance index includes: The imbalance index is matched with the imbalance index thresholds of multiple fault levels, and the fault level corresponding to the matched imbalance index threshold is determined as the fault level of the impeller.

9. The method according to claim 7, characterized in that, Before determining the feature deviation value of each dimension based on the standard acoustic features of that dimension, the method further includes: Based on the current operating condition type of the wind turbine equipment, a standard acoustic feature set corresponding to the current operating condition type is determined from the standard acoustic feature sets corresponding to each of the multiple operating condition types. Each standard acoustic feature set includes standard acoustic features of multiple dimensions.

10. The method according to claim 7, characterized in that, The process of determining the impeller's fault information based on the imbalance index includes: Obtain the imbalance index of the sound signals from multiple second time periods; Based on the imbalance index of the sound signals in the first time period and the plurality of second time periods, the information on the change of the imbalance index is determined; Based on the information about the change in the imbalance index, the fault information of the impeller is determined.

11. The method according to claim 1, characterized in that, The process of fusing multiple sound amplitudes corresponding to the target phase to obtain the fused sound amplitude of the target phase includes: Based on the weights of the rotation periods in which the multiple sound amplitudes are located, the multiple sound amplitudes are weighted and averaged to obtain the fused sound amplitude of the target phase.

12. The method according to claim 1, characterized in that, Determining the phase corresponding to each of the multiple rotational speeds at the multiple times includes: Based on the rotational speed at the multiple times, the rotational frequency at the multiple times is determined; For each time moment, the frequency conversion integral of the time moment and all time moments before the time moment is used to obtain the cumulative phase. Based on the cumulative phase and the initial phase, the phase corresponding to the time moment is obtained.

13. The method according to claim 1, characterized in that, The method further includes: A single rotation cycle is divided according to a preset angular interval to obtain the multiple target phases.

14. A fault detection device, characterized in that, The device includes: The acquisition module is used to acquire the sound signal and rotational speed data of the wind turbine equipment in a first time period. The sound signal includes the sound amplitude at multiple times, and the rotational speed data includes the rotational speed of the impeller of the wind turbine equipment at the multiple times. The determining module is used to determine the phase corresponding to each of the plurality of times based on the rotational speed at the plurality of times; The determining module is further configured to determine multiple sound amplitudes corresponding to multiple target phases based on the phases and sound amplitudes corresponding to the multiple times, wherein the sound amplitude corresponding to each target phase is the sound amplitude of the target phase at the time corresponding to one rotation cycle, and one rotation cycle corresponds to one complete rotation of the impeller; The fusion module is used to fuse multiple sound amplitudes corresponding to each target phase to obtain the fused sound amplitude of the target phase. The determining module is also used to determine the impeller's fault information based on the fused sound amplitude of each of the multiple target phases.

15. A computer device, characterized in that, include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the fault detection method as described in any one of claims 1 to 13.

16. A computer-readable storage medium, characterized in that, When the instructions in the computer-readable storage medium are executed by the processor of a computer device, the computer device is able to perform the fault detection method according to any one of claims 1 to 13.

17. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the fault detection method according to any one of claims 1 to 13.

Images