Ship environmental vibration decoupling analysis PHM system

Abstract

This invention relates to equipment condition monitoring and PHM data processing technology for ships and offshore floating equipment. It discloses a ship environmental vibration decoupling analysis PHM system. Addressing the problem that when ship motion, wind and wave excitation, and external equipment vibrations are at the same or near-the same frequency as the fault vibrations of the equipment under test, the true fault characteristics are easily suppressed, and the intrinsic vibrations of the equipment are inaccurately extracted, leading to distorted PHM assessments, the system uses rigid body kinematic mapping, candidate frequency identification, same-frequency fault discrimination, fault protection and purification, and PHM health management modules in conjunction to generate external background vibration prediction signals and candidate frequencies. Based on same-frequency fault evidence information, it performs background suppression, hybrid weakening, or fault feature protection on the vibration data of the equipment under test to obtain the intrinsic vibration data of the equipment. The system outputs PHM health index, fault severity, and remaining life prediction results, which can be used for the health management of shipborne equipment.

Description

Technical Field

[0001] This invention relates to the field of ship equipment condition monitoring and PHM data processing technology, and in particular to a ship environmental vibration decoupling analysis PHM system. Background Technology

[0002] In existing technologies, marine equipment health management and PHM systems typically collect equipment vibration signals through vibration sensors, accelerometers, or data acquisition cards. After data preprocessing, filtering, time-domain feature extraction, frequency-domain feature extraction, power spectrum analysis, feature matrix construction, and normalization, neural networks, support vector machines, logistic regression, statistical pattern recognition, or predictive models are used to determine equipment status, classify faults, assess health, and predict performance. Alarms, spectrum information, health status, runtime, and prediction results are then output through a host computer, network layer, or application layer. Such systems can already achieve online equipment vibration monitoring, data storage, fault alarms, health status display, and a certain degree of predictive maintenance, making them suitable for health management scenarios involving marine equipment, rolling bearings, and general machinery.

[0003] The problem with the aforementioned existing technologies is that they typically use the vibration data of the equipment under test directly as the analysis object, or only perform conventional filtering, feature extraction, and classification prediction on the collected vibration signals. They do not fully consider the situation where the vibration data of the equipment measurement points is mixed in by the ship's motion, wind and wave excitation, and the vibration of external equipment transmitted through the ship's structure. In particular, when the external background vibration frequency is the same as or similar to the fault characteristic frequency of the equipment under test, conventional frequency domain processing is prone to mistakenly treating the real fault characteristics as background vibration attenuation, resulting in inaccurate extraction of the intrinsic vibration of the equipment, which in turn causes the PHM health index, fault severity, and remaining life prediction results to be distorted.

[0004] Publication No. CN114715354A discloses a health management device for marine equipment and a method for equipment status detection and fault diagnosis. This device collects vibration information from marine equipment through a signal acquisition unit, performs Fourier transform on the signal processing unit, and uploads the time-domain and frequency-domain data to a host computer. The host computer then performs equipment level judgment, pre-alarm, alarm, operational status judgment, and operational time statistics. Furthermore, it achieves equipment status detection through time-domain features, frequency-domain features, power spectrum features, and BP neural network training. While this solution and the present invention both address marine equipment health management and vibration fault diagnosis, the former primarily solves the problems of real-time vibration monitoring, alarm, and health status judgment of marine equipment. However, it does not address the issue that when ship motion, wind and wave excitation, and external equipment vibrations are mixed into the equipment measurement data, the true fault characteristics are easily falsely suppressed under conditions of the same or near-same frequency.

[0005] Publication No. CN114997220A discloses a device PHM system based on a CPS architecture. This system collects vibration signals from mechanical equipment through an intelligent sensing layer, performs data preprocessing, feature extraction, and feature selection, executes health assessment, fault diagnosis, and performance prediction through a data decision layer, and achieves data communication and operational status monitoring through a network layer and application layer, thus forming a data closed loop of status perception, real-time analysis, scientific decision-making, and execution feedback. While this solution and the present invention both involve device PHM systems, vibration data processing, health assessment, and performance prediction, the latter focuses on the hierarchical architecture of general mechanical equipment PHM and vibration characteristic analysis. It does not address the problem of inaccurate intrinsic vibration data and distorted PHM predictions caused by the coupling of external background vibration and equipment fault vibration in a marine environment. Summary of the Invention

[0006] The technical problem to be solved by this invention is that the existing technology has the core problem of coupling between the external background vibration of the ship and the fault vibration of the equipment under test, especially when they are at the same frequency or close to the same frequency, which easily leads to the misfiltering of the true fault characteristics. That is, the disconnect between conventional vibration decoupling and PHM health assessment leads to inaccurate extraction of the intrinsic vibration of the equipment and distortion of PHM health index and remaining life prediction. To this end, we propose a ship environmental vibration decoupling analysis PHM system.

[0007] To achieve the above objectives, this application adopts the following technical solution: a ship environmental vibration decoupling analysis (PHM) system, including a data acquisition module, a data preprocessing module, a rigid body kinematics mapping module, a candidate frequency identification module, a same-frequency fault discrimination module, a fault protection and purification module, a PHM health management module, and a closed-loop optimization module.

[0008] The data acquisition module is used to collect motion data of the hull reference point, vibration data of the device under test, and operating status data of the device under test. The motion data of the hull reference point is used to characterize the overall motion or local motion state of the hull. The vibration data of the device under test is collected from vibration measuring points set at the device under test or its connecting structure. The operating status data of the device under test includes at least one of rotational speed, current, temperature, load, and pressure.

[0009] The data preprocessing module is used to synchronize and perform quality screening on the motion data of the ship's reference point, the vibration data of the equipment under test, and the operating status data of the equipment under test. The synchronization and alignment is to unify the time reference of the data from different acquisition channels. The quality screening includes filtering, correcting, or marking data with packet loss, saturation, outliers, persistent zero values, or drift anomalies.

[0010] The rigid body kinematics mapping module is used to generate an external background vibration prediction signal at the measurement point of the device under test based on the motion data of the hull reference point and the positional relationship between the hull reference point and the measurement point of the device under test. The external background vibration prediction signal is an equivalent predicted vibration signal transmitted to the measurement point of the device under test by the hull motion, wind and wave excitation or external equipment vibration through the hull structure. This signal is not used as the fault vibration signal of the device under test, but is used to determine the candidate frequency of the external background vibration and assist in the identification of the same frequency source.

[0011] Preferably, the motion data of the hull reference point includes translational components in the X direction, translational components in the Y direction, translational components in the Z direction, and rotational components about the X direction. Rotational component about the Y direction and rotational component about the Z direction The rigid body kinematics mapping module generates a position-equivalent background vibration prediction signal at the measurement point of the device under test based on the position vector from the hull reference point to the measurement point of the device under test.

[0012] Specifically, under the condition of small-angle ship motion, the position-equivalent background vibration prediction signal satisfies:

[0013] ,

[0014] ,

[0015] ,

[0016] , , These represent the translational components of the hull reference point in three orthogonal directions. , These represent the rotational components of the hull reference point about the three orthogonal directions, respectively. , , These represent the positional equivalent background vibration prediction signals at the measuring points of the device under test in three orthogonal directions. , , These represent the distances of the measuring point of the device under test relative to the reference point of the ship in the three orthogonal directions. The above formula is used to characterize the equivalent translational background vibration formed at the measuring point of the device under test after the ship's rotation is acted upon by the spatial lever arm.

[0017] Furthermore, the rigid body kinematics mapping module also includes a multi-reference fusion unit; the multi-reference fusion unit is used to generate corresponding external background vibration prediction signals based on at least two hull reference points, and to fuse multiple external background vibration prediction signals according to the data quality or noise level of the data corresponding to each hull reference point to obtain a fused external background vibration prediction signal; the data quality includes at least one of signal integrity, synchronization reliability, amplitude stability, noise level, and abnormal data ratio.

[0018] The candidate frequency identification module is used to determine the candidate frequency of external background vibration based on the external background vibration prediction signal; the candidate frequency of external background vibration is the frequency in the external background vibration prediction signal that satisfies the spectral energy condition, the time-frequency stability condition, or the continuous time window tracking condition; the candidate frequency of external background vibration can be combined with the candidate bandwidth to form a candidate frequency set; the candidate frequency set is used for subsequent background suppression, hybrid weakening, or fault feature protection processing.

[0019] Preferably, the candidate frequency identification module includes a time-frequency analysis unit, a dominant frequency filtering unit, and a frequency tracking unit; the time-frequency analysis unit is used to perform fast Fourier transform, short-time fast Fourier transform, wavelet transform, or order spectrum analysis on the external background vibration prediction signal; the dominant frequency filtering unit is used to filter external background vibration candidate frequencies according to the spectral amplitude sorting results and amplitude thresholds; the frequency tracking unit is used to smoothly track the external background vibration candidate frequencies within a continuous time window and generate a candidate frequency set containing the external background vibration candidate frequencies and candidate bandwidths.

[0020] The same-frequency fault discrimination module is used to generate same-frequency fault evidence information based on the external background vibration prediction signal, the vibration data of the device under test, and the operating status data of the device under test when the candidate frequency of the external background vibration is the same as or similar to the fault characteristic frequency of the device under test. The same or similar means that the candidate frequency of the external background vibration and the fault characteristic frequency of the device under test are within a preset frequency neighborhood. The fault characteristic frequency of the device under test includes at least one of the following: rotational frequency, harmonic frequency, bearing fault characteristic frequency, gear meshing frequency, pump impeller frequency, or electromagnetic fault frequency. The same-frequency fault evidence information is data used to characterize whether the vibration component corresponding to the candidate frequency of the external background vibration contains the fault characteristics of the device under test.

[0021] Preferably, the same-frequency fault discrimination module includes a background interpretation quantity calculation unit, a fault interpretation quantity calculation unit, and a same-frequency competition discrimination unit; the background interpretation quantity calculation unit is used to generate a background interpretation quantity based on the coherence relationship between the vibration data of the device under test and the external background vibration prediction signal at the candidate frequency of the external background vibration; the coherence relationship includes at least one of correlation, cross-spectral coherence, phase consistency, or propagation relationship consistency at the candidate frequency of the external background vibration; the fault interpretation quantity calculation unit is used to generate a fault interpretation quantity based on the correlation between the candidate frequency of the external background vibration and the fault characteristic frequency of the device under test, the phase of the device rotation speed, and the operating status data of the device under test; the same-frequency competition discrimination unit is used to generate a same-frequency competition discrimination result based on the background interpretation quantity and the fault interpretation quantity; the same-frequency competition discrimination result includes at least one of background vibration component, device fault vibration component, or mixed vibration component.

[0022] Furthermore, the same-frequency competition discrimination unit can generate a same-frequency competition index based on the background interpretation quantity and the fault interpretation quantity. The same-frequency competition index is used to characterize whether the vibration component corresponding to the candidate frequency of the external background vibration is closer to the source of the external background vibration or closer to the source of the fault of the device under test. The same-frequency competition index satisfies:

[0023] ,

[0024] in, Indicates the first The same-frequency competition index corresponding to the candidate frequencies of external background vibration. This indicates the amount of background information provided. Indicates the fault interpretation quantity. To prevent positive numbers with a denominator of zero, the same-frequency fault discrimination module determines the vibration component corresponding to the candidate frequency of external background vibration as an external background vibration component, a mixed vibration component, or an equipment fault vibration component based on the same-frequency competition index.

[0025] Furthermore, the same-frequency fault discrimination module also includes a fractional-order chaotic phase transition detection unit; the fractional-order chaotic phase transition detection unit is used to construct a fractional-order nonlinear oscillator model corresponding to the fault characteristic frequency of the equipment under test when the same-frequency competition discrimination result characterizes the mixed vibration component or equipment fault vibration component corresponding to the candidate frequency of the external background vibration, and generate same-frequency fault evidence information according to the phase state change of the fractional-order nonlinear oscillator model; the phase state change includes the change between chaotic state, critical chaotic state, periodic state or large-scale periodic state; the same-frequency fault evidence information includes at least one of phase state transition result, phase state transition response time, phase space dispersion degree, phase locking degree or same-frequency fault evidence index.

[0026] Furthermore, the fractional-order chaotic phase transition detection unit can generate a synchronous fault evidence index based on the phase space dispersion, phase transition response time, and phase locking degree. The synchronous fault evidence index satisfies the following:

[0027] ,

[0028] in, Indicates the first The same-frequency fault evidence index corresponding to the candidate frequencies of external background vibration. This indicates the degree of phase space discretization in the critical chaotic state. This indicates the degree of phase space discretization after injecting normalized narrowband vibration components. Indicates the phase transition response time. Indicates the reference response time. This indicates the degree of phase lock between the candidate frequency of the external background vibration and the phase of the rotational speed of the device under test. This indicates the upper limit of the evidence index for synchronous failure. Represents the amplitude limiting function. This indicates a positive number to prevent the denominator from being zero; the same-frequency fault evidence index is used to participate in the fault feature protection of the fault protection and purification module and the calculation of the PHM health index of the PHM health management module.

[0029] The fault protection and purification module is used to perform background suppression, hybrid weakening, or fault feature protection on the vibration data of the equipment under test based on the candidate frequencies of external background vibration and evidence information of the same frequency fault, thereby obtaining the intrinsic vibration data of the equipment. The background suppression is to suppress the candidate frequencies of external background vibration that are determined to be external background vibration components. The hybrid weakening is to perform low-intensity suppression on the mixed vibration components that cannot be completely determined to be background vibration or equipment fault vibration. The fault feature protection is to not perform deep suppression on the candidate frequencies of external background vibration that have evidence information of the same frequency fault, and to retain the amplitude, envelope, phase, or order correlation characteristics at that frequency. The intrinsic vibration data of the equipment is the vibration data used to characterize the operating state and degradation state of the equipment under test after background suppression, hybrid weakening, or fault feature protection.

[0030] Furthermore, the fault protection and purification module includes a background suppression unit, a hybrid weakening unit, and a fault protection unit; the background suppression unit is used to perform background suppression processing on external background vibration candidate frequencies that are determined to be external background vibration components; the hybrid weakening unit is used to perform weakening processing on external background vibration candidate frequencies that are determined to be hybrid vibration components; the fault protection unit is used to perform fault feature retention processing on external background vibration candidate frequencies that have evidence of faults at the same frequency, and output the processed vibration data as intrinsic vibration data of the equipment.

[0031] The PHM health management module generates a PHM health index, fault severity, and remaining life prediction results based on the intrinsic vibration data of the equipment and evidence of faults at the same frequency. The PHM health index is an indicator used to quantify the degree of degradation of the equipment under test. The fault severity is the degree of equipment fault classified according to the PHM health index or its changing trend. The remaining life prediction results are life prediction data generated based on the time evolution trend of the PHM health index and the failure threshold.

[0032] Preferably, the PHM health management module includes a baseline matching unit and a health index calculation unit; the baseline matching unit is used to match a health baseline under the corresponding operating condition based on the operating status data of the equipment under test; the operating condition is the equipment operating condition determined by at least one of the following state variables: speed, equipment rotation speed, load, sea state, temperature, and pressure; the health baseline is the time-domain baseline, spectral baseline, impact characteristic baseline, or state data baseline formed under the corresponding operating condition when the equipment under test is in a healthy state; the health index calculation unit is used to generate a PHM health index based on the residual signal between the equipment's intrinsic vibration data and the health baseline, the energy of the fault-sensitive frequency band, the impact characteristics, and the evidence information of the same-frequency fault.

[0033] Furthermore, the PHM health management module can generate a PHM health index based on the energy deviation of the fault-sensitive frequency band, the impact characteristic deviation, and the total amount of evidence of faults at the same frequency. The PHM health index satisfies the following:

[0034] ,

[0035] in, express PHM health index at any time This indicates the deviation of the fault-sensitive frequency band energy of the equipment's intrinsic vibration data relative to the healthy baseline. This indicates the deviation of the equipment's intrinsic vibration data from the impact characteristics of a healthy baseline. It represents the total amount of evidence of synchronous faults within the current time window; the PHM health index is used to characterize the degree of degradation of the device under test relative to the health baseline, and is used to generate fault severity and remaining life prediction results.

[0036] Furthermore, the PHM health management module also includes a degradation trend analysis unit and a remaining life prediction unit; the degradation trend analysis unit is used to smooth the PHM health index within a continuous time window and extract the health degradation trend; the health degradation trend includes at least one of the health index change rate, accelerated degradation trend, or cross-condition degradation consistency; the remaining life prediction unit is used to generate the remaining life prediction result of the device under test based on the PHM health index, health degradation trend, and failure threshold.

[0037] Furthermore, the system includes a closed-loop optimization module; the closed-loop optimization module includes a decoupling quality evaluation unit, a fault fidelity evaluation unit, a same-frequency evidence evaluation unit, and a parameter update unit; the decoupling quality evaluation unit is used to evaluate the degree of external background vibration suppression; the fault fidelity evaluation unit is used to evaluate the degree of fault feature retention at the fault protection frequency; the same-frequency evidence evaluation unit is used to evaluate the strength of same-frequency fault evidence information; the parameter update unit is used to update the candidate frequency set, candidate bandwidth, or fault protection frequency set according to the degree of external background vibration suppression, the degree of fault feature retention, and the strength of same-frequency fault evidence information.

[0038] The technical effects and advantages of this invention are as follows:

[0039] In this invention, a predicted signal of external background vibration at the measurement point of the device under test is generated based on rigid body kinematic mapping, so that the identification of candidate frequencies of external background vibration does not depend on the mixed vibration data of the device under test; based on same-frequency fault discrimination and fractional-order chaotic phase transition detection, same-frequency fault evidence information is generated when the candidate frequencies of external background vibration are the same as or similar to the fault characteristic frequencies of the device under test; based on the fault protection and purification module, background suppression, mixed weakening or fault characteristic protection are performed on candidate frequencies of external background vibration from different sources, so that the intrinsic vibration data of the device retains the fault characteristics while suppressing external background vibration; based on the intrinsic vibration data of the device and the same-frequency fault evidence information, PHM health index, fault severity and remaining life prediction results are generated, making the PHM health management results more applicable to the complex vibration environment of ships. Attached Figure Description

[0040] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts:

[0041] Figure 1 A schematic diagram of the PHM system structure for decoupling analysis of ship environmental vibration;

[0042] Figure 2 Flowchart for same-frequency fault discrimination and fractional-order chaotic phase transition detection;

[0043] Figure 3 This is a flowchart of the fault protection and purification process.

[0044] Figure 4 A flowchart for PHM health management and closed-loop optimization;

[0045] Figure 5 This is a schematic diagram illustrating the effect of same-frequency fault detection.

[0046] Figure 6This is a schematic diagram illustrating the trend of the PHM health index.

[0047] Figure 7 This is a schematic diagram of the closed-loop optimization process. Detailed Implementation

[0048] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0049] like Figure 1 As shown, this embodiment provides a ship environmental vibration decoupling analysis (PHM) system, including a data acquisition module, a data preprocessing module, a rigid body kinematics mapping module, a candidate frequency identification module, a same-frequency fault discrimination module, a fault protection and purification module, a PHM health management module, and a closed-loop optimization module.

[0050] The data acquisition module is used to collect background reference measurement point data, raw vibration data from equipment measurement points, and equipment operating status data. Background reference measurement points are set on the ship's hull structure to collect background vibration information caused by ship motion, wind and wave excitation, and structural propagation. Equipment measurement points are set on the equipment under test to collect raw vibration data from that equipment. Equipment operating status data includes at least one of the following: rotational speed, current, temperature, load, and pressure.

[0051] Background reference measurement points collect six degrees of freedom data; the six degrees of freedom data include translational vibration waveforms in the X direction, translational vibration waveforms in the Y direction, translational vibration waveforms in the Z direction, and rotational waveforms about the X direction. , Rotational waveform around the Y direction and the rotating waveform around the Z direction The equipment's measuring points acquire raw vibration waveforms in the X, Y, and Z directions. Background reference measuring points and equipment measuring points can be wired or wireless sensors, achieving synchronous acquisition through unified time synchronization, timed acquisition, or timestamp correction. Since candidate frequency identification is primarily performed in the frequency domain, this implementation does not require absolute synchronization among all measuring points; millisecond-level synchronization is sufficient for frequency identification and band-stop filter frequency setting requirements.

[0052] The data preprocessing module is used to synchronize and align background reference measurement point data, raw vibration data from equipment measurement points, and equipment operating status data, as well as to filter out abnormal data and unify coordinates. Synchronization and alignment unify the time base of different acquisition channels. Abnormal data filtering identifies packet loss, saturation, outliers, persistent zero values, and drift anomalies. Coordinate unification transforms the coordinate systems of the background reference measurement points and equipment measurement points to the same ship coordinate system.

[0053] In this embodiment, the ship's coordinate system is used as the unified coordinate system. The X-direction of the ship's coordinate system is the bow direction, the Y-direction is the transverse direction, and the Z-direction is the vertical direction. If the background reference sensor coordinate system is inconsistent with the ship's coordinate system, the data preprocessing module first converts the sensor data to the ship's coordinate system using an installation attitude matrix. Let the installation attitude matrix from the background reference sensor coordinate system to the ship's coordinate system be... The background reference measuring point sensor directly collects the three-dimensional translational data as follows: The background reference measuring point sensor directly collects the three-dimensional rotation data as follows: Then, the three translational components and three rotational components after transformation to the ship's coordinate system are as follows:

[0054] ,

[0055] ,

[0056] in, , T represents vector transpose. The mounting attitude matrix can be determined by the sensor mounting angle or through the calibration process after installation.

[0057] The rigid body kinematics mapping module is used to generate a predicted external background vibration signal at the device measuring point based on the six degrees of freedom data of the background reference measuring point and the positional relationship between the background reference measuring point and the device measuring point; the positional relationship includes the distance of the device measuring point relative to the background reference measuring point in the X direction. Distance in the Y direction Distance in the Z direction

[0058] In this embodiment, the rigid body kinematics mapping module does not directly map the background reference measurement points. , , The translational vibration waveform is used as a background reference signal at the measuring point of the equipment, but it is combined with... , , The rotation waveform and the spatial positional relationship between the equipment measuring point and the background reference measuring point are used to generate a position-equivalent background vibration prediction signal at the equipment measuring point. This processing is used to reflect the equivalent translational background vibration formed at the equipment measuring point after the ship's rotational motion is acted upon by the spatial lever arm.

[0059] Let the position vector of the device measuring point relative to the background reference measuring point be:

[0060] ,

[0061] in, This indicates the distance of the device's measuring point relative to the background reference measuring point in the X direction. This indicates that the device measuring point is relative to the background reference measuring point in the Y direction. This indicates the distance of the device's measuring point relative to the background reference measuring point in the Z direction.

[0062] Under small-angle ship motion conditions, the rigid body kinematics mapping module can be generated using a first-order position equivalent mapping method. , , External background vibration prediction signals in three directions. In this embodiment, the position-equivalent background vibration prediction signal at the device measuring point can be expressed as:

[0063] ,

[0064] in, , representing the positional equivalent background vibration prediction signal at the device's measuring point; Let represent the antisymmetric matrix composed of rotational components, and its expression is:

[0065] ,

[0066] Expanding the above matrix, we get:

[0067] ,

[0068] ,

[0069] ,

[0070] in, , , The above formula represents the predicted equivalent background vibration signals at the equipment measuring point in three directions. It indicates that the rotational motion of the background reference measuring point, through the spatial distance between the background reference measuring point and the equipment measuring point, creates additional equivalent translational background vibration at the equipment measuring point.

[0071] When the background reference measuring points and the equipment measuring points are mainly located on the same deck surface or in the same compartment plane, or When the influence of the lever arm corresponding to the directional distance is negligible, a simplified formula can be used:

[0072]

[0073]

[0074]

[0075] This simplification ignores In and The coupling terms, and In and The coupling term is suitable for installation scenarios where the height difference between the background reference measuring point and the device measuring point is small. However, if the sensor installation orientation or the positive direction of the ship's coordinate system is different, , , The signs of the corresponding items are adjusted according to the actual coordinate conventions. In the above calculations, , , , , , , , , All data are time series data. , , Since it is a position scalar, the mapping is a vectorized computation performed at each sampling time.

[0076] For scenarios involving large swings, turns, acceleration / deceleration, or requiring acceleration domain processing, the rigid body kinematics mapping module can also employ a complete rigid body acceleration mapping method. Let the position vector from the background reference measuring point to the device measuring point be... The translational acceleration of the background reference measuring point is angular velocity is angular acceleration is Then the predicted external background vibration acceleration at the measuring point of the equipment is:

[0077]

[0078] in, This represents the predicted acceleration of external background vibration at the measuring point of the equipment. This indicates the translational acceleration of the background reference measuring point. Indicates the angular acceleration of the background reference measurement point. Indicates the angular velocity of the background reference measurement point. This represents the position vector from the background reference measuring point to the device measuring point, with the symbol... This represents the vector cross product operation. This acceleration mapping method, along with the aforementioned first-order position equivalent mapping, is used to generate the predicted external background vibration signal at the device's measuring point. The specific method chosen depends on the type of data acquired and the system deployment requirements.

[0079] like Figure 2 As shown, the candidate frequency identification module receives the external background vibration prediction signal and extracts candidate frequencies of the external background vibration. In this embodiment, the candidate frequency identification module prioritizes... , , The process involves frame segmentation, followed by a Fast Fourier Transform (FFT) for each frame. A Hanning window can be used for frame segmentation, with a frame length of 1024 sampling points and a frame overlap rate of 50%. The candidate frequency identification module obtains the values ​​for each frame. , , After obtaining the corresponding Fourier spectrum, the spectral amplitudes are sorted from largest to smallest, and the dominant frequency of external background vibration is selected based on the amplitude threshold.

[0080] The amplitude threshold can be set as a certain percentage of the maximum amplitude of a single frame's spectrum, or it can be determined based on the spectral noise floor, the percentile of the spectral amplitude, or the historical background vibration spectrum level. In one embodiment, the amplitude threshold is set to 5% of the maximum amplitude of a single frame's spectrum. The candidate frequency identification module extracts the dominant frequencies with amplitudes greater than the amplitude threshold, thus obtaining... , , to A candidate frequency set for external background vibration is formed; the candidate frequency set for external background vibration can also be used as a notch filter frequency list.

[0081] It should be noted that the candidate frequencies for external background vibration are derived from the predicted external background vibration signal, rather than directly from the original vibration data of the equipment measuring points. In other words, the predicted external background vibration signal is used for frequency determination, while the original vibration data of the equipment measuring points is used for frequency correction. This dual-signal-channel processing method avoids the equipment fault frequency being mistakenly selected as the background dominant frequency simply because it has a high amplitude in the original vibration data of the equipment measuring points.

[0082] like Figure 3 As shown, the same-frequency fault discrimination module is used to generate same-frequency fault evidence information when the candidate frequency of external background vibration is the same as or close to the fault characteristic frequency of the equipment under test. The system pre-establishes a fault characteristic frequency library for the equipment under test. For bearings, the fault characteristic frequency library includes the fault characteristic frequencies of the inner ring, outer ring, rolling elements, and cage; for gearboxes, the fault characteristic frequency library includes the rotational frequency, harmonics, meshing frequency, and sideband frequency; for pumps, fans, or motors, the fault characteristic frequency library includes the rotational frequency, pump impeller frequency, electromagnetic fault frequency, and their related harmonics.

[0083] The background interpretation unit determines whether the candidate frequency of the external background vibration is closer to the source of the external background vibration based on the coherence relationship between the original vibration data from the equipment measuring point and the predicted external background vibration signal at the candidate frequency of the external background vibration. The coherence relationship includes cross-spectral coherence, phase consistency, and propagation relationship consistency. If the original vibration data from the equipment measuring point at the candidate frequency of the external background vibration has high coherence with the predicted external background vibration signal, and the phase relationship matches the propagation relationship obtained from rigid body kinematic mapping, then the background interpretation quantity increases.

[0084] The fault interpretation calculation unit determines whether the candidate frequency of external background vibration is closer to the source of the fault in the equipment based on the correlation between the candidate frequency of external background vibration and the characteristic frequency of equipment fault, the phase of equipment rotation speed, and the operating status data of the equipment under test. If the candidate frequency of external background vibration is close to the characteristic frequency of equipment fault and maintains phase lock with the order of equipment rotation speed, while the current, temperature, load, or pressure deviates synchronously from the healthy baseline, the fault interpretation value increases.

[0085] The same-frequency competition discrimination unit compares the background interpretation quantity and the fault interpretation quantity. When the background interpretation quantity is dominant, the candidate frequency of external background vibration is determined to be the external background vibration component. When the fault interpretation quantity is dominant, the candidate frequency of external background vibration is determined to be the equipment fault vibration component. When the two are close, the candidate frequency of external background vibration is determined to be the mixed vibration component.

[0086] Furthermore, the same-frequency competition discrimination unit can calculate the same-frequency competition index based on the background interpretation quantity and the fault interpretation quantity. The calculation formula is as follows:

[0087] ,

[0088] in, Indicates the first The competition index corresponding to each candidate frequency of external background vibration represents the amount of background explanation. Indicates the fault interpretation quantity. This indicates a positive number that prevents the denominator from being zero. If If the value is greater than the first threshold, the corresponding vibration component is determined to be an external background vibration component; if... If the vibration component is less than the second threshold, the corresponding vibration component is determined to be a device fault vibration component; if If the vibration component is located between the first threshold and the second threshold, then the corresponding vibration component is determined to be a mixed vibration component.

[0089] Figure 5This is a schematic diagram of the same-frequency fault discrimination effect. The horizontal axis represents the background interpretation quantity, and the vertical axis represents the fault interpretation quantity. When the background interpretation quantity is high and the fault interpretation quantity is low, it corresponds to the background vibration component. When the background interpretation quantity and the fault interpretation quantity are close, it corresponds to the mixed vibration component. When the fault interpretation quantity is high and the background interpretation quantity is low, it corresponds to the equipment fault vibration component. Figure 5 This is used to illustrate how the same-frequency competition discrimination unit classifies the sources of external background vibration candidate frequencies based on background interpretation and fault interpretation.

[0090] For candidate frequencies of external background vibration identified as mixed vibration components or equipment fault vibration components, the fractional-order chaotic phase transition detection unit further detects weak fault characteristics at the same frequency. This unit extracts narrowband vibration components near the candidate frequencies of external background vibration from the original vibration data of the equipment measurement points. Narrowband vibration components can be obtained through bandpass filtering, time-frequency local reconstruction, or order tracking. After extraction, their amplitudes are normalized, using the root mean square value or envelope mean value of the narrowband vibration components within the current time window as the normalization scale.

[0091] The fractional-order chaotic phase transition detection unit is constructed using a fractional-order nonlinear oscillator model corresponding to the fault characteristic frequency:

[0092]

[0093] in, Indicates the first Oscillator state variables corresponding to candidate frequencies of external background vibration This represents the first derivative of the state variable with respect to time. The fractional derivative of the state variable. Describes a fractional differential operator. This represents the fractional order corresponding to the i-th candidate frequency of external background vibration. Indicates the damping coefficient. Indicates the amplitude of the driving force. Indicates the frequency of fault characteristics of the equipment under test. This indicates the fault order phase related to the equipment speed. Indicates the injection coefficient of the narrowband signal at the same frequency. This represents the normalized narrowband vibration component.

[0094] The fractional-order chaotic phase transition detection unit initializes the oscillator state under healthy baseline data or historical stable data, placing it in a critical chaotic state. During operation, normalized narrowband vibration components are input to the oscillator. If a weak periodic component related to the equipment fault characteristic frequency and the phase of the equipment rotation speed exists at the candidate frequency of the external background vibration, the oscillator state transitions from a critical chaotic state to a periodic state or a large-scale periodic state. This phase change is recorded as evidence of a synchronous fault; the synchronous fault evidence information includes at least one of the following: phase transition result, phase transition response time, phase space dispersion, phase locking degree, or synchronous fault evidence index.

[0095] Furthermore, the fractional-order chaotic phase transition detection unit can calculate the same-frequency fault evidence index based on the phase space dispersion, phase transition response time, and phase locking degree. The calculation formula is as follows:

[0096] ,

[0097] in, Indicates the first The same-frequency fault evidence index corresponding to the candidate frequencies of external background vibration. This indicates the degree of phase space discretization in the critical chaotic state. This indicates the degree of phase space discretization after injecting normalized narrowband vibration components. Indicates the phase transition response time. Indicates the reference response time. This indicates the degree of phase lock between the candidate frequency of the external background vibration and the phase of the rotational speed of the device under test. This indicates the upper limit of the evidence index for synchronous failure. Represents the amplitude limiting function. This indicates a positive number to prevent the denominator from being zero. If the phase space dispersion decreases, the phase transition response time shortens, and the phase locking degree increases after injecting normalized narrowband vibration components, the co-frequency fault evidence index increases.

[0098] like Figure 3 As shown, the fault protection and purification module performs background suppression, hybrid weakening, or fault feature protection on the original vibration data of the equipment measuring points based on the external background vibration candidate frequency set, the same-frequency competition discrimination result, and the same-frequency fault evidence information.

[0099] In this embodiment, the fault protection and purification module constructs a multi-band-stop filter bank based on the set of candidate frequencies of external background vibration. Each candidate frequency of external background vibration corresponds to a band-stop filter unit. The center frequency of the band-stop filter unit is determined by the candidate frequency of external background vibration, and the stopband width is determined by the candidate bandwidth. Multiple band-stop filter units form a multi-band-stop filter bank. This multi-band-stop filter bank acts on the original vibration waveform synchronously acquired by the equipment's measuring points, rather than acting on... , , External background vibration prediction signal.

[0100] The raw vibration data from the equipment's measuring points includes the raw vibration waveforms in the X, Y, and Z directions, which can be denoted as... , , After processing by multiple band-stop filter banks, the purified three-dimensional equipment vibration data is output, which can be denoted as... , , . , , As intrinsic vibration data of the equipment, it is used for subsequent fault diagnosis and PHM health management.

[0101] In one implementation, the candidate bandwidth w is 2Hz. When the candidate frequency of the external background vibration is 0.8Hz, the stopband can be set to 0Hz to 1.8Hz; when the candidate frequency of the external background vibration is 1.6Hz, the stopband can be set to 0.6Hz to 2.6Hz; when the candidate frequency of the external background vibration is 3.2Hz, the stopband can be set to 2.2Hz to 4.2Hz; and when the candidate frequency of the external background vibration is 12.5Hz, the stopband can be set to 11.5Hz to 13.5Hz. The above-mentioned candidate frequency, candidate bandwidth, and stopband are only one implementation method, and can be adjusted according to the sampling frequency, frequency resolution, sea state changes, and equipment fault characteristic frequencies in practical applications.

[0102] When the same-frequency fault discrimination module does not find that the external background vibration candidate frequency is the same as or similar to the equipment fault characteristic frequency, the fault protection and purification module can perform background suppression processing according to the set of external background vibration candidate frequencies. When the external background vibration candidate frequency is determined to be an external background vibration component, the corresponding band-stop filter unit performs background suppression. When the external background vibration candidate frequency is determined to be a mixed vibration component, the corresponding band-stop filter unit reduces the suppression intensity and performs mixed weakening. When there is evidence of a same-frequency fault at the external background vibration candidate frequency, the corresponding frequency does not perform deep suppression, and the fault protection unit retains the amplitude, envelope, phase, or order characteristics at that frequency.

[0103] like Figure 4 As shown, the PHM health management module generates the PHM health index, fault severity, and remaining life prediction results based on the intrinsic vibration data of the equipment, the operating status data of the equipment under test, and evidence of faults at the same frequency.

[0104] The PHM (Health Management) module includes a baseline matching unit, a health index calculation unit, a degradation trend analysis unit, and a remaining life prediction unit. The baseline matching unit matches a health baseline based on ship speed, equipment rotation speed, load, sea state, temperature, or pressure. The health baseline is established when the equipment is confirmed to be in good health, and its content includes a time-domain baseline, a spectral baseline, an impact characteristic baseline, and a condition data baseline.

[0105] The health index calculation unit first calculates the residual signal between the intrinsic vibration data of the equipment and the health baseline. Then, it extracts the energy deviation, impact characteristic deviation, and in-frequency fault evidence information of the residual signal within the fault-sensitive frequency band. The fault-sensitive frequency band is determined around the fault characteristic frequency of the equipment under test. Impact characteristics include peak value, kurtosis, impulse factor, margin factor, or envelope impact characteristics. The health index calculation unit normalizes and fuses the above degradation evidence to obtain the PHM health index.

[0106] Furthermore, the PHM health management module can calculate the PHM health index based on the energy deviation of the fault-sensitive frequency band, the impact characteristic deviation, and the total amount of evidence of faults at the same frequency. The calculation formula is as follows:

[0107] ,

[0108] in, express PHM health index at any time This indicates the deviation of the fault-sensitive frequency band energy of the equipment's intrinsic vibration data relative to the healthy baseline. This indicates the deviation of the equipment's intrinsic vibration data from the impact characteristics of a healthy baseline. This represents the total amount of evidence of faults occurring at the same frequency within the current time window. This total amount of evidence can be obtained by summing, root mean square, or maximizing the indices of all fault evidence at the same frequency within the fault protection frequency set. A higher PHM health index indicates a greater deviation of the device under test from the health baseline.

[0109] Figure 6 This diagram illustrates the PHM (Health and Well-being Management) index trend. The horizontal axis represents operating time, and the vertical axis represents the PHM health index. This diagram demonstrates how changes in the health index reflect equipment degradation trends after incorporating evidence of synchronous failures. The failure threshold indicates when the PHM health index reaches a preset threshold. The remaining life prediction unit can then generate a remaining life prediction result based on the relationship between the health index trend and the failure threshold. Figure 6 This is a trend illustration chart, and does not limit the actual health index value and failure time of the equipment.

[0110] The degradation trend analysis unit smooths the PHM health index within a continuous time window. The smoothing process can be achieved by moving average, median filtering, trend fitting, or outlier removal. The remaining lifetime prediction unit generates the remaining lifetime prediction result based on the smoothed PHM health index, health degradation trend, and failure threshold.

[0111] In this embodiment, the intrinsic vibration data of the equipment can also be input into a fixed-ground equipment fault diagnosis model or a land-based equipment health diagnosis model. The fault diagnosis model extracts the effective value, peak value, kurtosis, envelope spectrum characteristics, fault-sensitive frequency band energy, and order characteristics, and identifies bearing faults, gear faults, shaft misalignment, blade damage, or other equipment faults based on the above characteristics. Based on the output results of the fault diagnosis model, the PHM health management module combines the intrinsic vibration data of the equipment, evidence information of faults at the same frequency, and health baseline to generate the PHM health index, fault severity, and remaining life prediction results.

[0112] The closed-loop optimization module updates the candidate frequency set, candidate bandwidth, and fault protection frequency set based on the operation results. The decoupling quality evaluation unit evaluates the degree of external background vibration suppression based on the energy changes before and after purification at the background vibration frequency. The fault fidelity evaluation unit evaluates the degree of fault feature retention based on the retention of amplitude, envelope, phase, or order characteristics at the fault protection frequency before and after purification. The same-frequency evidence evaluation unit evaluates the strength of same-frequency fault evidence information based on phase transition results, phase transition response time, and phase locking degree. The parameter update unit adjusts the frequency range or suppression intensity of background suppression processing when external background vibration suppression is insufficient, reduces the suppression intensity of the corresponding external background vibration candidate frequency when fault feature retention is insufficient, or adds the external background vibration candidate frequency to the fault protection frequency set. When same-frequency fault evidence information is enhanced, the priority of the external background vibration candidate frequency in the PHM health index calculation is increased.

[0113] Furthermore, the parameter update unit can update the candidate bandwidth based on the degree of external background vibration suppression and the degree of fault feature retention, as shown in the following update formula:

[0114]

[0115] in, Indicates the first Candidate bandwidth of each external background vibration candidate frequency in the next time window This indicates the candidate bandwidth within the current time window. and These represent the lower limit and upper limit of the candidate bandwidth, respectively. Indicates the degree of suppression of external background vibration. Indicates the degree of target inhibition. Indicates the degree to which fault characteristics are preserved. Indicates the minimum fault characteristic retention requirement. and This represents the adjustment coefficient. This indicates that the value is 0 when it is less than 0. When the external background vibration suppression is insufficient and the fault feature retention meets the requirements, the candidate bandwidth increases; when the fault feature retention is insufficient, the candidate bandwidth is limited or decreased.

[0116] Figure 7 This is a schematic diagram of the closed-loop optimization process. Figure 7 The horizontal axis represents the time window, and the vertical axis represents the candidate bandwidth. Figure 7 This describes how the closed-loop optimization module updates the candidate bandwidth based on the degree of external background vibration suppression and the degree of fault feature retention. When the external background vibration suppression is insufficient but the degree of fault feature retention meets the requirements, the system can increase the candidate bandwidth or enhance background suppression; when the degree of fault feature retention is insufficient, the system reduces the suppression intensity of the corresponding external background vibration candidate frequency or adds the external background vibration candidate frequency to the fault protection frequency set. Figure 7 The graph illustrates the parameter update trend, without limiting the actual numerical changes in the candidate bandwidth.

[0117] In a full-process application scenario, the system first collects data through background reference measurement points. , , , , , Six degrees of freedom data are collected through device measurement points. , , The three-dimensional original vibration waveform is simultaneously acquired, along with the equipment's rotational speed, current, and temperature. The rigid body kinematics mapping module maps the six-degree-of-freedom data to [data based on the positional relationship between the background reference measuring point and the equipment measuring point]. , , The candidate frequency identification module... , , Perform FFT and extract based on amplitude threshold. to The fault protection purification module is based on to Construct a multi-band-stop filter bank and apply the filter bank to , , If the same-frequency fault discrimination module detects that a candidate frequency of external background vibration is the same as or similar to the characteristic frequency of equipment fault, it further generates same-frequency fault evidence information and selects background suppression, hybrid weakening, or fault feature protection based on this evidence information; after processing... , , Input the PHM health management module and output the PHM health index, fault severity, and remaining life prediction results.

[0118] This implementation method is consistent with the basic decoupling filtering process. Specifically, the six-degree-of-freedom background reference measurement point data is mapped to the location equivalent to obtain the predicted external background vibration signal at the equipment measurement point. The mapped signal is used for FFT main frequency extraction, and the main frequency forms a band-stop filter frequency list. The band-stop filter is applied to the original vibration waveform at the equipment measurement point. Building upon this, this implementation method further incorporates same-frequency fault evidence information, fault protection and purification, and PHM health management processes to address the issue of potential false suppression of fault characteristics when the candidate frequency of external background vibration is the same as or close to the equipment fault characteristic frequency.

[0119] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.

Claims

1. A ship environmental vibration decoupling analysis (PHM) system, comprising a data acquisition module, a data preprocessing module, and a PHM health management module, wherein the data acquisition module is used to acquire motion data of the ship's reference points, vibration data of the equipment under test, and operating status data of the equipment under test; the data preprocessing module is used to perform synchronization alignment and quality screening on the acquired data; and the PHM health management module is used to generate equipment health management results, characterized in that... Also includes: The rigid body kinematics mapping module is used to generate the predicted signal of the external background vibration at the measuring point of the device under test based on the motion data of the hull reference point and the positional relationship between the hull reference point and the measuring point of the device under test. The candidate frequency identification module is used to determine the candidate frequency of external background vibration based on the predicted external background vibration signal; the same-frequency fault discrimination module is used to generate same-frequency fault evidence information based on the predicted external background vibration signal, the vibration data of the device under test, and the operating status data of the device under test when the candidate frequency of external background vibration is the same as or similar to the fault characteristic frequency of the device under test; the fault protection and purification module is used to perform background suppression, hybrid weakening, or fault feature protection on the vibration data of the device under test based on the candidate frequency of external background vibration and the same-frequency fault evidence information to obtain the intrinsic vibration data of the device; the PHM health management module generates PHM health index, fault severity, and remaining life prediction results based on the intrinsic vibration data of the device and the same-frequency fault evidence information.

2. The PHM system for decoupling analysis of ship environmental vibration according to claim 1, characterized in that, The motion data of the hull reference point includes translational components in three orthogonal directions and rotational components around the three orthogonal directions. The rigid body kinematics mapping module is used to generate a position equivalent background vibration prediction signal at the measurement point of the device under test under small-angle ship motion conditions, based on the translational components, the rotational components, and the position vector from the hull reference point to the measurement point of the device under test. The location-equivalent background vibration prediction signal satisfies: ; ; ; in, , , These represent the translational components of the hull reference point in three orthogonal directions. , , , , and represent the rotational components of the hull reference point about the three orthogonal directions, respectively. , , These represent the positional equivalent background vibration prediction signals at the measuring points of the device under test in three orthogonal directions. , , These represent the distances of the measuring point of the device under test relative to the reference point of the ship's hull in the three orthogonal directions, respectively.

3. The PHM system for decoupling analysis of ship environmental vibration according to claim 2, characterized in that, The rigid body kinematics mapping module also includes a multi-reference fusion unit, which is used to generate corresponding external background vibration prediction signals based on at least two hull reference points, and fuse multiple external background vibration prediction signals according to the data quality or noise level of the data corresponding to each hull reference point to obtain the fused external background vibration prediction signal.

4. The PHM system for decoupling analysis of ship environmental vibration according to claim 1, characterized in that, The candidate frequency identification module includes a time-frequency analysis unit, a dominant frequency filtering unit, and a frequency tracking unit. The time-frequency analysis unit is used to perform fast Fourier transform, short-time fast Fourier transform, wavelet transform, or order spectrum analysis on the external background vibration prediction signal. The dominant frequency filtering unit is used to filter the external background vibration candidate frequencies according to the spectral amplitude sorting results and amplitude thresholds. The frequency tracking unit is used to smoothly track the external background vibration candidate frequencies within a continuous time window and generate a candidate frequency set containing the external background vibration candidate frequencies and candidate bandwidths.

5. The PHM system for decoupling analysis of ship environmental vibration according to claim 4, characterized in that, The same-frequency fault discrimination module includes a background interpretation quantity calculation unit, a fault interpretation quantity calculation unit, and a same-frequency competition discrimination unit. The background interpretation quantity calculation unit is used to generate a background interpretation quantity based on the coherence relationship between the vibration data of the device under test and the external background vibration prediction signal at the candidate frequency of the external background vibration. The fault interpretation quantity calculation unit is used to generate a fault interpretation quantity based on the correlation between the candidate frequency of the external background vibration and the fault characteristic frequency of the device under test, the phase of the device rotation speed, and the operating status data of the device under test. The same-frequency competition discrimination unit is used to generate a same-frequency competition discrimination result based on the background interpretation quantity and the fault interpretation quantity.

6. The PHM system for decoupling analysis of ship environmental vibration according to claim 5, characterized in that, The same-frequency fault discrimination module further includes a fractional-order chaotic phase transition detection unit. The fractional-order chaotic phase transition detection unit is used to construct a fractional-order nonlinear oscillator model corresponding to the fault characteristic frequency of the equipment under test when the same-frequency competition discrimination result characterizes the mixed vibration component or the equipment fault vibration component corresponding to the candidate frequency of the external background vibration, and generate same-frequency fault evidence information based on the phase change of the fractional-order nonlinear oscillator model.

7. The PHM system for decoupling analysis of ship environmental vibration according to claim 6, characterized in that, The fault protection and purification module includes a background suppression unit, a hybrid weakening unit, and a fault protection unit. The background suppression unit is used to perform background suppression processing on candidate frequencies of external background vibrations that are determined to be external background vibration components. The hybrid weakening unit is used to perform weakening processing on candidate frequencies of external background vibrations that are determined to be hybrid vibration components. The fault protection unit is used to perform fault feature retention processing on candidate frequencies of external background vibrations that have evidence of faults at the same frequency, and outputs the processed vibration data as intrinsic vibration data of the equipment.

8. The PHM system for decoupling analysis of ship environmental vibration according to claim 7, characterized in that, The PHM health management module includes a working condition baseline matching unit and a health index calculation unit. The working condition baseline matching unit is used to match the health baseline under the corresponding working condition based on the operating status data of the equipment to be tested. The health index calculation unit is used to generate the PHM health index based on the residual signal between the intrinsic vibration data of the equipment and the health baseline, the energy of the fault-sensitive frequency band, the impact characteristics, and the evidence information of the same frequency fault.

9. A ship environmental vibration decoupling analysis PHM system according to claim 8, characterized in that, The PHM health management module also includes a degradation trend analysis unit and a remaining life prediction unit. The degradation trend analysis unit is used to smooth the PHM health index within a continuous time window and extract the health degradation trend. The remaining life prediction unit is used to generate the remaining life prediction result of the device under test based on the PHM health index, health degradation trend and failure threshold.

10. A ship environmental vibration decoupling analysis (PHM) system according to claim 9, characterized in that, It also includes a closed-loop optimization module, which comprises a decoupling quality evaluation unit, a fault fidelity evaluation unit, a same-frequency evidence evaluation unit, and a parameter update unit. The decoupling quality evaluation unit is used to evaluate the degree of external background vibration suppression, the fault fidelity evaluation unit is used to evaluate the degree of fault feature retention at the fault protection frequency, the same-frequency evidence evaluation unit is used to evaluate the strength of same-frequency fault evidence information, and the parameter update unit is used to update the candidate frequency set, candidate bandwidth, or fault protection frequency set based on the degree of external background vibration suppression, the degree of fault feature retention, and the strength of same-frequency fault evidence information.

Images