Single-sensor-based engine fault detection method, device, equipment and medium

Abstract

The application discloses an engine fault detection method and device based on a single sensor, equipment and a medium, and relates to the technical field of fault detection. The method comprises the following steps: acquiring an eddy current original signal and a fiber pulse signal of a fiber eddy current composite sensor; the fiber eddy current composite sensor is installed at an aero-engine casing position; calculating a blade radial displacement signal according to the eddy current original signal and calculating a blade circumferential displacement signal according to the fiber pulse signal; filtering the blade circumferential displacement signal to obtain a target circumferential displacement signal; resampling the blade radial displacement signal to obtain a time sequence blade tip clearance signal; and determining engine blade fault information according to the target circumferential displacement signal and the time sequence blade tip clearance signal. In this way, the eddy current original signal and the fiber pulse signal are directly measured by the single sensor to detect faults, the transmission path is shortened, the sensitivity, accuracy and measuring point utilization rate of diagnosis are improved, and the complexity of sensor arrangement is reduced.

Description

Technical Field

[0001] This invention relates to the field of fault detection technology, and in particular to an engine fault detection method, device, equipment and medium based on a single sensor. Background Technology

[0002] Currently, condition monitoring of aero-engines mainly relies on multi-sensor fusion methods, which typically include acceleration sensors for measuring casing vibration, displacement sensors for measuring rotor shaft trajectory, blade tip sensors for measuring blade vibration, and pressure sensors for detecting surge. However, these methods require multiple sensors to work together, have a large number of measurement points, and are less sensitive to early, minor faults. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to overcome the shortcomings of the prior art and provide an engine fault detection method, device, equipment and medium based on a single sensor, which is used to directly acquire the original eddy current signal and fiber optic pulse signal by installing a single composite sensor at the location of the aero-engine casing, and to use this information to further diagnose and warn of various engine faults, so as to achieve efficient monitoring of aero-engines with single measurement point and multi-source sensing.

[0004] This invention provides the following technical solution:

[0005] In a first aspect, this application proposes an engine fault detection method based on a single sensor, comprising: acquiring the raw eddy current signal and fiber pulse signal of an optical fiber eddy current composite sensor; the optical fiber eddy current composite sensor being installed at the casing position of an aero-engine; calculating the radial displacement signal of the blade based on the raw eddy current signal, and calculating the circumferential displacement signal of the blade based on the fiber pulse signal; filtering the circumferential displacement signal of the blade to obtain a target circumferential displacement signal; resampling the radial displacement signal of the blade to obtain a time-series tip clearance signal; and determining engine blade fault information based on the target circumferential displacement signal and the time-series tip clearance signal.

[0006] In one embodiment, the step of calculating the blade radial displacement signal based on the original eddy current signal and the blade circumferential displacement signal based on the fiber optic pulse signal includes: filtering the original eddy current signal to obtain an eddy current filtered signal; acquiring the current rotational speed of the aero-engine and determining the dynamic sensitivity corresponding to the current rotational speed; converting the eddy current filtered signal into the blade radial displacement signal based on the dynamic sensitivity; performing time point conversion on the fiber optic pulse signal to obtain the number of measurement time points; determining the time point difference between the preset theoretical time point number and the number of measurement time points, and obtaining the blade circumferential displacement signal by multiplying the time point difference by the current rotational speed.

[0007] In one embodiment, resampling the blade radial displacement signal to obtain a time-series blade tip gap signal includes: resampling the blade radial displacement signal to obtain a blade tip gap value; correcting and compensating the blade tip gap value based on a preset blade static machining dimension to obtain a corrected blade tip gap value; and performing traveling wave analysis on the corrected blade tip gap value to obtain the time-series blade tip gap signal.

[0008] In one embodiment, determining engine blade fault information based on the target circumferential displacement signal includes: obtaining a single-blade signal based on the target circumferential displacement signal, and acquiring a blade phase based on the single-blade signal; calculating a coefficient of variation based on the blade phase to obtain a variation characteristic parameter; obtaining a phase characteristic evaluation result based on the variation characteristic parameter; obtaining an amplitude warning threshold based on the blade phase, and obtaining an amplitude characteristic evaluation result based on the amplitude warning threshold and the target circumferential displacement signal; and determining the engine blade fault information as surge information, resonance information, or foreign object impact fault information based on the phase characteristic evaluation result and the amplitude characteristic evaluation result.

[0009] In one embodiment, determining engine blade fault information based on the time-series blade tip clearance signal includes: fitting the time-series blade tip clearance signal to obtain a radial relative motion state; obtaining an unbalance amplitude and an unbalance phase based on the radial relative motion state using a cross-correlation algorithm; and determining the engine blade fault information as engine unbalance fault information based on the unbalance amplitude and the unbalance phase.

[0010] In one embodiment, determining engine blade fault information based on the time-series blade tip clearance signal includes: detecting the time-series blade tip clearance signal based on an anomaly detection algorithm to obtain an anomaly signal; performing time-frequency analysis on the anomaly signal to obtain a time-frequency energy spectrum; and determining the engine blade fault information as rubbing fault information if there are rubbing characteristics in the time-frequency energy spectrum.

[0011] In one embodiment, the method further includes: performing discrete wavelet decomposition on the circumferential displacement signal of the blade to obtain wavelet coefficients; calculating the modulus square of the wavelet coefficients and obtaining a time-frequency spectrum based on the modulus square; determining the blade vibration frequency based on the time-frequency spectrum; determining the frequency change rate based on the blade vibration frequency and a preset modal frequency; and determining the engine blade fault information as blade crack information if the frequency change rate is greater than a preset threshold.

[0012] Secondly, this application proposes an engine fault detection device based on a single sensor, comprising:

[0013] The acquisition module is used to acquire the raw eddy current signal and fiber pulse signal of the fiber optic eddy current composite sensor; the fiber optic eddy current composite sensor is installed in the casing of the aero-engine.

[0014] The calculation module is used to calculate the blade radial displacement signal based on the original eddy current signal and the blade circumferential displacement signal based on the fiber pulse signal.

[0015] The processing module is used to filter the circumferential displacement signal of the blade to obtain the target circumferential displacement signal; and to resample the radial displacement signal of the blade to obtain the time-series blade tip clearance signal.

[0016] The determination module is used to determine engine blade fault information based on the target circumferential displacement signal and the timing blade tip clearance signal.

[0017] Thirdly, this application proposes a computer device including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the engine fault detection method based on a single sensor as described in the first aspect.

[0018] Fourthly, this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the single-sensor-based engine fault detection method as described in the first aspect.

[0019] This application discloses a single-sensor-based engine fault detection method, apparatus, equipment, and medium. It acquires the raw eddy current signal and fiber pulse signal from a fiber optic eddy current composite sensor. The fiber optic eddy current composite sensor is installed in the aero-engine casing. The radial displacement signal of the blade is calculated based on the raw eddy current signal, and the circumferential displacement signal is calculated based on the fiber pulse signal. The circumferential displacement signal is filtered to obtain a target circumferential displacement signal. The radial displacement signal is resampled to obtain a time-series tip clearance signal. Engine blade fault information is determined based on the target circumferential displacement signal and the time-series tip clearance signal. This method directly measures the raw eddy current signal and fiber pulse signal reflecting the dynamic characteristics of the blade using a single sensor, shortening the transmission path, improving diagnostic sensitivity, accuracy, and measurement point utilization, and reducing sensor layout complexity. It possesses advantages such as strong real-time performance, high diagnostic accuracy, and wide applicability, and can be used for aero-engine condition monitoring, fault early warning, and health management, improving engine operation safety and reliability. Attached Figure Description

[0020] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope of protection of the present invention. In the various drawings, similar components are numbered similarly.

[0021] Figure 1 A flowchart of the engine fault detection method based on a single sensor proposed in this embodiment is shown.

[0022] Figure 2 A schematic diagram illustrating the principle of synchronous monitoring of blade vibration and tip clearance under single-sensor conditions is shown.

[0023] Figure 3 Another flowchart of the engine fault detection method based on a single sensor proposed in this embodiment is shown;

[0024] Figure 4 This diagram illustrates another flow chart of the engine fault detection method based on a single sensor proposed in this embodiment.

[0025] Figure 5 This embodiment presents a schematic diagram of a single-sensor resonance early warning case based on blade tip timing.

[0026] Figure 6 This embodiment presents a schematic diagram of a single-sensor surge early warning system based on blade tip timing.

[0027] Figure 7 This embodiment illustrates the principle of rotor vibration fitting and unbalance parameter identification based on blade tip clearance.

[0028] Figure 8 This illustration shows a practical diagnostic case of imbalance parameter identification proposed in this embodiment;

[0029] Figure 9 This embodiment illustrates a real-world case of time-domain early warning for rubbing faults based on blade tip clearance.

[0030] Figure 10 This illustration shows a practical case of frequency domain diagnosis of rubbing faults based on blade tip clearance proposed in this embodiment;

[0031] Figure 11 A schematic diagram of the engine fault detection device based on a single sensor proposed in this embodiment is shown.

[0032] Explanation of reference numerals in the attached diagram:

[0033] 1101 - Acquisition module; 1102 - Calculation module; 1103 - Processing module; 1104 - Determination module. Detailed Implementation

[0034] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0035] The components of the embodiments of the invention described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0036] In the following, the terms “comprising,” “having,” and their cognates, which may be used in various embodiments of the invention, are intended only to indicate a particular feature, number, step, operation, element, component, or combination thereof, and should not be construed as excluding, firstly, the presence of one or more other features, numbers, steps, operations, elements, components, or combinations thereof, or adding the possibility of one or more features, numbers, steps, operations, elements, components, or combinations thereof.

[0037] Furthermore, the terms "first," "second," and "third" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0038] Unless otherwise specified, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the invention pertain. Terms (such as those defined in commonly used dictionaries) shall be interpreted as having the same meaning as in their contextual meaning in the relevant technical field and shall not be interpreted as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of the invention.

[0039] Example 1

[0040] This disclosure provides a single-sensor-based engine fault detection method, which uses a single composite sensor installed at the engine casing to directly acquire the raw eddy current signal and fiber optic pulse signal. This information is then used to further diagnose and warn of various engine faults, thereby achieving efficient monitoring of aero-engines with single-point, multi-source sensing.

[0041] Please see Figure 1 The engine fault detection method based on a single sensor includes steps S101 to S104, and each step is described in detail below.

[0042] Step S101: Obtain the raw eddy current signal and fiber pulse signal from the fiber optic eddy current composite sensor; the fiber optic eddy current composite sensor is installed in the casing of the aero-engine.

[0043] In this embodiment, the fiber optic eddy current composite sensor is a fiber-optic-eddy current coupled blade vibration-tip clearance composite sensor. The fiber-optic-eddy current coupled blade vibration-tip clearance composite sensor is installed in the aero-engine casing, and the raw eddy current signal and fiber pulse signal from the sensor are synchronously acquired at high frequency by a host computer monitoring system. The principle of synchronous monitoring of blade vibration and tip clearance under single-sensor conditions is as follows: Figure 2 As shown.

[0044] It should be noted that the fiber-optic-eddy current coupled blade vibration-tip clearance composite sensor is mounted at a single point on the casing. The center of the sensor probe must be aligned with the blade tip, and the sensor probe must not interfere with the blade's rotation path. Given the high-frequency vibration and high linear velocity of the blade, both the raw eddy current signal and the fiber optic pulse signal require high-frequency sampling.

[0045] It should be noted that by using a single fiber-optic-eddy current composite sensor located in the engine casing to simultaneously measure blade vibration and tip clearance, the number of measurement points is reduced, the sensor layout is simplified, and the complexity and maintenance cost of the monitoring system are decreased compared to traditional multi-sensor systems.

[0046] Step S102: Calculate the blade radial displacement signal based on the original eddy current signal, and calculate the blade circumferential displacement signal based on the fiber pulse signal.

[0047] In this embodiment, the raw eddy current signal reflects the radial displacement of the engine blades, while the fiber optic pulse signal reflects the circumferential displacement. Therefore, the raw eddy current signal can be converted into a radial displacement signal, and the fiber optic pulse signal can be converted into a circumferential displacement signal. Preliminary processing of the fiber optic pulse signal and the raw eddy current signal using signal preprocessing methods improves data quality and ensures data reliability and analytical accuracy.

[0048] Please see Figure 3 In one specific embodiment, step S102 includes steps S1021 to S1025, and each step is described in detail below.

[0049] Step S1021: Filter the original eddy current signal to obtain the filtered eddy current signal.

[0050] In this embodiment, the original eddy current signal is filtered to remove noise introduced by high-frequency sampling, resulting in a filtered eddy current signal.

[0051] Step S1022: Obtain the current speed of the aero-engine and determine the dynamic sensitivity corresponding to the current speed.

[0052] In this embodiment, dynamic sensitivity represents the response characteristics of the sensor output as a function of rotational speed. It is typically a nonlinear function and requires experimental calibration. The dynamic sensitivity corresponding to the current rotational speed of the aero-engine is then queried.

[0053] Step S1023: Convert the eddy current filter signal into the blade radial displacement signal based on the dynamic sensitivity.

[0054] In this embodiment, the original eddy current signal is analyzed, and the rotor rotation period T and the arrival time difference t between adjacent blades are calculated based on the characteristic blades. w The specific calculation formula is: T = t OPR,p+1 -t OPR,p , t w =T / n b In the formula, n b This indicates the number of blades, OPR indicates a bonded phase blade, and p indicates the number of turns.

[0055] Furthermore, based on the rotor rotation period T and the arrival time difference t between adjacent blades... w Calculate the time range of arrival for a single blade and extract the peak value within this time range. Each peak value is the extracted blade tip gap voltage.

[0056] Furthermore, the tip gap voltage V is converted into the tip gap value D, i.e., the blade radial displacement signal, based on the dynamic sensitivity S. The specific expression is: D = V / S.

[0057] Step S1024: Convert the optical fiber pulse signal into a number of time points to obtain the number of measurement time points.

[0058] In this embodiment, the fiber optic pulse signal is converted into a number of measurement time points using a counter.

[0059] Step S1025: Determine the time difference between the preset theoretical time point number and the measured time point number, and obtain the blade circumferential displacement signal based on the product of the time point difference and the current rotational speed.

[0060] In this embodiment, the time difference ΔT between the preset theoretical time point number and the measured time point number is determined, and the circumferential displacement signal D of the blade is obtained by multiplying the time difference ΔT with the current rotational speed V. c The specific expression is: D c=VΔT.

[0061] Step S103: Filter the circumferential displacement signal of the blade to obtain the target circumferential displacement signal; resample the radial displacement signal of the blade to obtain the time-series blade tip gap signal.

[0062] In this embodiment, the blade circumferential displacement signal is filtered to obtain the target circumferential displacement signal. For example, the blade circumferential displacement signal is smoothed and denoised using a data smoothing and differential calculation method based on local polynomial least squares fitting, and offset information is removed to obtain the target circumferential displacement signal with high-frequency noise removed. Simultaneously, the blade radial displacement signal is resampled to obtain the time-series blade tip gap signal.

[0063] Please see Figure 4 In one specific embodiment, step S103 includes steps S1031 to S1033, and each step is described in detail below.

[0064] Step S1031: Resample the radial displacement signal of the blade to obtain the blade tip gap value.

[0065] In this embodiment, the radial displacement signal of the blade is resampled to obtain the blade tip gap value of each blade.

[0066] Step S1032: Based on the preset static processing dimensions of the blade, the blade tip gap value is corrected and compensated to obtain the corrected blade tip gap value.

[0067] In this embodiment, a preset static machining dimension D of the blade is used. s The blade tip clearance value D is corrected and compensated to obtain the corrected blade tip clearance value D. B The specific expression is: D B =D+D s .

[0068] Step S1033: Perform traveling wave analysis on the corrected blade tip gap value to obtain the time-series blade tip gap signal.

[0069] In this embodiment, traveling wave analysis was performed on the corrected blade tip gap value to obtain the temporal blade tip gap variation of different blades, so as to obtain the temporal blade tip gap signal.

[0070] It should be noted that the average tip clearance of each blade and the minimum tip clearance during the operating cycle can be calculated based on the time-series tip clearance changes. The average tip clearance can be used to describe the engine's operating status, and the minimum tip clearance can also be used for early warning of aero-engine malfunctions.

[0071] Step S104: Determine engine blade fault information based on the target circumferential displacement signal and the timing blade tip clearance signal.

[0072] In this embodiment, by comprehensively analyzing the target circumferential displacement signal reflecting the vibration or deformation of the blade along the circumferential direction, and the time-series blade tip clearance signal reflecting the change of the blade-casing clearance over time, engine blade faults can be effectively identified, improving the sensitivity and accuracy of fault detection, enhancing the health management level of aero engines, and solving problems such as the complex arrangement of multiple sensors, the difficulty of data synchronization and fusion, and the strong independence and lack of uniformity in fault diagnosis in existing aero engine condition monitoring methods.

[0073] In one specific embodiment, step S104 includes: obtaining a single-blade signal based on the target circumferential displacement signal, and acquiring the blade phase based on the single-blade signal; calculating the coefficient of variation based on the blade phase to obtain a variation characteristic parameter; obtaining a phase characteristic evaluation result based on the variation characteristic parameter; obtaining an amplitude warning threshold based on the blade phase, and obtaining an amplitude characteristic evaluation result based on the amplitude warning threshold and the target circumferential displacement signal; determining the engine blade fault information as surge information, resonance information, or foreign object impact fault information based on the phase characteristic evaluation result and the amplitude characteristic evaluation result.

[0074] In this embodiment, the individual blade signals of each blade are obtained based on the target circumferential displacement signal, and the blade phase of each blade is calculated based on the individual blade signals. The specific expression is: In the formula, A is the tip circumferential displacement coefficient. This is phase offset information.

[0075] Furthermore, based on the blade phase of each blade The phase standard deviation was calculated. and average phase difference And based on the phase standard deviation and average phase difference The coefficient of variation is calculated to obtain the characteristic parameter U of the variation. The specific expression is:

[0076] Furthermore, the phase feature evaluation result is obtained based on the variation feature parameters. Exemplarily, if the variation feature parameter is lower than a preset range, the corresponding phase feature return value is set to "-1"; if the variation feature parameter exceeds the preset range, the corresponding phase feature return value is set to "1". The preset range can be (-60, +60), and this embodiment does not limit it.

[0077] Furthermore, the amplitude warning threshold D is obtained based on the blade phase. maxThe specific expression is: In the formula, N is the number of rotor blades.

[0078] Furthermore, an amplitude feature evaluation result is obtained based on the amplitude warning threshold and the target circumferential displacement signal. For example, if the vibration amplitude of the blade, i.e., the target circumferential vibration displacement signal, exceeds the amplitude warning threshold, the amplitude feature return value is "large".

[0079] Furthermore, based on the evaluation results of the phase characteristics and amplitude characteristics of each blade, the engine blade fault information is determined to be surge, resonance, or foreign object impact fault information. For example, if the phase characteristic return value of all blades is "-1" and the amplitude characteristic return value is "large", then the engine blade fault is resonance; if the phase characteristic return value of all blades is "1" and the amplitude characteristic return value is "large", then the engine blade fault is surge; if the phase characteristic return value of some blades is "1" and the amplitude characteristic return value is "large", then the engine blade fault is foreign object impact fault.

[0080] like Figure 5 The compressor tip timing monitoring data shown is as follows: the resonance begins at the 200th revolution and recovers after the 530th revolution. The figure shows the displacement vibration waveform, amplitude characteristics, alarm threshold, and phase characteristics of the blade. It can be seen that when the resonance occurs, the amplitude characteristic curve exceeds the alarm threshold and the phase characteristic curve shows -1. After the resonance ends, the amplitude characteristic and phase characteristic curves recover.

[0081] like Figure 6 The figure shows the blade tip timing monitoring data during the compressor surge. The surge started at the 356th revolution and recovered after the 447th revolution. The figure shows the blade displacement vibration waveform, amplitude characteristics, alarm threshold, and phase characteristics. It can be seen that when the surge occurs, the amplitude characteristic curve exceeds the alarm threshold, and the phase characteristic curve shows 1. After the surge ends, the amplitude characteristic and phase characteristic curves recover.

[0082] In one specific embodiment, step S104 includes: fitting the timing blade tip clearance signal to obtain the radial relative motion state; obtaining the unbalance amplitude and unbalance phase based on the radial relative motion state using a cross-correlation algorithm; and determining the engine blade fault information as engine unbalance fault information based on the unbalance amplitude and the unbalance phase.

[0083] In this embodiment, the time-series blade tip clearance signal is fitted to restore the discrete time-series blade tip clearance to the continuous radial relative motion state between the blade and the casing within the rotation cycle of the aero-engine, thereby reflecting the overall vibration characteristics of the rotor. Fitting methods include polynomial fitting and sine wave fitting.

[0084] Furthermore, based on the cross-correlation algorithm, the amplitude and phase of the rotor's first harmonic component are extracted from the radial relative motion state to obtain the unbalance amplitude and unbalance phase, thereby identifying the rotor's unbalance vibration parameters.

[0085] The identified unbalanced phase is calculated with reference to the characteristic blade under the condition of no bond phase, that is, the identification of unbalanced amplitude and unbalanced phase only requires a single sensor.

[0086] Furthermore, based on the unbalance amplitude and unbalance phase, the current engine blade fault is determined to be an engine unbalance fault. The principle of rotor vibration fitting and unbalance parameter identification based on blade tip clearance is as follows: Figure 7 As shown, the left side depicts the engine main shaft and an impeller. During rotation, the engine shaft vibrates in real time, and the blade tip clearance signal reflects the radial displacement of the blades, thus reflecting the radial vibration of the rotor. A practical example of time-domain early warning for rubbing faults based on blade tip clearance is provided. Figure 8 As shown in the figure, the effect of diagnosing imbalance through the blade tip gap signal is demonstrated: the imbalance correlation component is the signal after filtering, fitting and decomposing the blade tip gap signal. The data corresponding to this method is the result of cross-correlation calculation (representing the imbalance amplitude and imbalance phase). The numerical simulation setting is the true value. It can be seen that the true value is basically consistent with the calculation result of the method in this paper, which is used to reflect the practical effect of the proposed calculation method.

[0087] In one specific embodiment, step S104 includes: detecting the time-series blade tip gap signal based on an anomaly detection algorithm to obtain an anomaly signal; performing time-frequency analysis on the anomaly signal to obtain a time-frequency energy spectrum; and determining that the engine blade fault information is a rubbing fault information if there are rubbing features in the time-frequency energy spectrum.

[0088] In this embodiment, anomaly detection is performed on the time-series blade tip clearance signal based on an anomaly detection algorithm to identify abnormal signals, which can then provide an early warning that the aero-engine is in an abnormal state. Further, time-frequency analysis is performed on the abnormal signal to obtain the time-frequency energy spectrum. If rubbing characteristics are present in the time-frequency energy spectrum, the engine blade fault is determined to be a rubbing fault. Rubbing characteristics include peak clipping at the time of rubbing and time-frequency non-stationarity. The specific expression for time-frequency analysis of the abnormal signal is as follows: In the formula, D c It is a decomposed abnormal signal, y i These are the individual decomposition components, and r is the remaining decomposition component. These are the analytical solutions of each component, representing the solution to the previous equation. N represents the number of components, j is the complex unit, and a i Let i be the amplitude of the i-th decomposed component. This represents the phase angle of the i-th decomposed component.

[0089] In one specific embodiment, the method further includes: performing discrete wavelet decomposition on the circumferential displacement signal of the blade to obtain wavelet coefficients; calculating the modulus square of the wavelet coefficients and obtaining a time-frequency spectrum based on the modulus square; determining the blade vibration frequency based on the time-frequency spectrum; determining the frequency change rate based on the blade vibration frequency and a preset modal frequency; and determining the engine blade fault information as blade crack information if the frequency change rate is greater than a preset threshold. A practical example of time-domain early warning for rubbing faults based on blade tip clearance is provided. Figure 9 As shown, when rubbing occurs, the blade tip clearance is used to calculate time-domain characteristic parameters and trigger a time-domain anomaly warning. 0 on the right indicates normal operation, and 1 indicates anomaly. The presence of an anomaly suggests the possible occurrence of rubbing. A practical case study of frequency-domain diagnosis of rubbing faults based on blade tip clearance is provided. Figure 10 As shown, by judging the abnormal frequency domain, we can see the peak clipping phenomenon, which is a characteristic of the rubbing fault.

[0090] In this embodiment, the circumferential displacement signal of the blade is denoted as the original signal x(t), where t represents time. Since actual signals are often subject to noise interference, preprocessing is required to improve analysis accuracy.

[0091] The preprocessing procedure is as follows: wavelet thresholding denoising is used, and Symlet wavelet basis functions are selected. Given the decomposition level J, perform discrete wavelet decomposition on x(t) to obtain wavelet coefficients as the decomposition scale w. j,k The translation parameter k. The soft thresholding denoising formula applied to the wavelet coefficients is: In the formula, λ is the noise filtering threshold, which is usually taken as... σ is the noise standard deviation, and N is the signal length, i.e., the number of blade rotations.

[0092] Use wavelet coefficients for noise reduction To reconstruct the signal, we obtain the preprocessed original signal x. denoised (t), and perform wavelet transform on it: In the formula, 'a' is the scale parameter, which is inversely proportional to the frequency. b is the translation parameter, representing the time position, and is generally taken as equal to the rotational speed frequency, i.e.

[0093] Furthermore, by calculating the squared modulus |W(a,b)| of the wavelet coefficients W(a,b), 2 A time-frequency spectrum is generated to reflect the energy distribution of the signal at different times and frequencies. In the time-frequency spectrum, the blade vibration frequency is represented by a ridge line where energy is concentrated. Therefore, the blade vibration frequency is obtained by extracting the ridge line. The specific expression is: In the formula, a max (t) represents the scale with the maximum energy at time t in the time-spectrum graph, f sThe frequency at which a single sensor is used at this moment is called the frequency conversion.

[0094] Cracks reduce blade stiffness, causing the vibration frequency to deviate from the normal modal frequency. Therefore, comparing the extracted blade vibration frequency with the modal frequency can diagnose crack faults. Specifically, the blade vibration frequency f under fault-free conditions is calculated. nolmal Calculate its average frequency Calculate the average frequency of blades suspected of having crack faults. If the rate of change of frequency If the deviation exceeds the empirical value by 2.5%, it can be determined that the blade has a blade crack.

[0095] It should be noted that the target circumferential displacement signal and the time-series blade tip clearance signal can be integrated to match different fault characteristics, ultimately achieving efficient comprehensive diagnosis and evaluation of various faults in aero-engines.

[0096] This embodiment proposes a single-sensor-based engine fault detection method. It acquires the raw eddy current signal and fiber pulse signal from a fiber optic eddy current composite sensor, which is installed in the engine casing. The method calculates the blade radial displacement signal based on the raw eddy current signal and the blade circumferential displacement signal based on the fiber pulse signal. The circumferential displacement signal is filtered to obtain the target circumferential displacement signal. The radial displacement signal is resampled to obtain the time-series tip clearance signal. Engine blade fault information is determined based on the target circumferential displacement signal and the time-series tip clearance signal. This method directly measures the raw eddy current signal and fiber pulse signal, reflecting the dynamic characteristics of the blade, using a single sensor. This shortens the transmission path, improves diagnostic sensitivity, accuracy, and measurement point utilization, and reduces sensor layout complexity. It offers advantages such as strong real-time performance, high diagnostic accuracy, and wide applicability, and can be used for aero-engine condition monitoring, fault early warning, and health management, improving engine operation safety and reliability.

[0097] Example 2

[0098] Furthermore, this disclosure provides an engine fault detection device 1100 based on a single sensor; please refer to [link to relevant documentation]. Figure 11 The device includes:

[0099] The acquisition module 1101 is used to acquire the raw eddy current signal and fiber pulse signal of the fiber optic eddy current composite sensor; the fiber optic eddy current composite sensor is installed in the casing of the aero-engine.

[0100] Calculation module 1102 is used to calculate the blade radial displacement signal based on the original eddy current signal and to calculate the blade circumferential displacement signal based on the fiber optic pulse signal.

[0101] Processing module 1103 is used to filter the circumferential displacement signal of the blade to obtain the target circumferential displacement signal; and to resample the radial displacement signal of the blade to obtain the time-series blade tip gap signal.

[0102] The determination module 1104 is used to determine engine blade fault information based on the target circumferential displacement signal and the timing blade tip clearance signal.

[0103] Optionally, the calculation module 1102 is further configured to filter the original eddy current signal to obtain an eddy current filtered signal; acquire the current rotational speed of the aero-engine and determine the dynamic sensitivity corresponding to the current rotational speed; convert the eddy current filtered signal into the radial displacement signal of the blade according to the dynamic sensitivity; perform time point conversion on the fiber optic pulse signal to obtain the number of measurement time points; determine the time point difference between the preset theoretical time point number and the number of measurement time points, and obtain the circumferential displacement signal of the blade according to the product of the time point difference and the current rotational speed.

[0104] Optionally, the processing module 1103 is further configured to resample the radial displacement signal of the blade to obtain the blade tip gap value; correct and compensate the blade tip gap value based on the preset static machining dimensions of the blade to obtain the corrected blade tip gap value; and perform traveling wave analysis on the corrected blade tip gap value to obtain the time-series blade tip gap signal.

[0105] Optionally, the determining module 1104 is further configured to obtain a single-blade signal based on the target circumferential displacement signal, and obtain a blade phase based on the single-blade signal; calculate a coefficient of variation based on the blade phase to obtain a variation characteristic parameter; obtain a phase characteristic evaluation result based on the variation characteristic parameter; obtain an amplitude warning threshold based on the blade phase, and obtain an amplitude characteristic evaluation result based on the amplitude warning threshold and the target circumferential displacement signal; and determine whether the engine blade fault information is surge information, resonance information, or foreign object impact fault information based on the phase characteristic evaluation result and the amplitude characteristic evaluation result.

[0106] Optionally, the determining module 1104 is further configured to fit the timing blade tip clearance signal to obtain the radial relative motion state; based on the cross-correlation algorithm, obtain the unbalance amplitude and unbalance phase according to the radial relative motion state; and determine the engine blade fault information as engine unbalance fault information based on the unbalance amplitude and the unbalance phase.

[0107] Optionally, the determining module 1104 is further configured to detect the timing blade tip gap signal based on an anomaly judgment algorithm to obtain an anomaly signal; perform time-frequency analysis on the anomaly signal to obtain a time-frequency energy spectrum; and if there are rubbing characteristics in the time-frequency energy spectrum, determine that the engine blade fault information is rubbing fault information.

[0108] Optionally, the determining module 1104 is further configured to perform discrete wavelet decomposition on the circumferential displacement signal of the blade to obtain wavelet coefficients; calculate the modulus square of the wavelet coefficients, and obtain a time-frequency spectrum based on the modulus square; determine the blade vibration frequency based on the time-frequency spectrum; determine the frequency change rate based on the blade vibration frequency and a preset modal frequency; if the frequency change rate is greater than a preset threshold, then determine the engine blade fault information as blade crack information.

[0109] The apparatus provided in this disclosure can perform the steps of the engine fault detection method based on a single sensor provided in Embodiment 1. To avoid repetition, the steps will not be repeated.

[0110] This embodiment proposes a single-sensor-based engine fault detection device that acquires the raw eddy current signal and fiber pulse signal from a fiber optic eddy current composite sensor. The fiber optic eddy current composite sensor is installed in the aero-engine casing. The device calculates the blade radial displacement signal based on the raw eddy current signal and the blade circumferential displacement signal based on the fiber pulse signal. The circumferential displacement signal is filtered to obtain the target circumferential displacement signal. The radial displacement signal is resampled to obtain the time-series tip clearance signal. Engine blade fault information is determined based on the target circumferential displacement signal and the time-series tip clearance signal. This method directly measures the raw eddy current signal and fiber pulse signal, reflecting the dynamic characteristics of the blade, using a single sensor. This shortens the transmission path, improves diagnostic sensitivity, accuracy, and measurement point utilization, and reduces the complexity of sensor placement. It possesses advantages such as strong real-time performance, high diagnostic accuracy, and wide applicability, and can be used for aero-engine condition monitoring, fault early warning, and health management, improving the safety and reliability of engine operation.

[0111] Example 3

[0112] Furthermore, this disclosure provides a computer device including a memory and a processor. The memory stores a computer program, which, when executed by the processor, implements the engine fault detection method based on a single sensor as described in Embodiment 1.

[0113] The device provided in this embodiment can perform the steps of the engine fault detection method based on a single sensor provided in Embodiment 1. To avoid repetition, the steps will not be repeated.

[0114] Example 4

[0115] This disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the engine fault detection method based on a single sensor as described in Embodiment 1.

[0116] In this embodiment, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0117] The computer-readable storage medium provided in this embodiment can implement the engine fault detection method based on a single sensor provided in Embodiment 1. To avoid repetition, it will not be described again here.

[0118] In all examples shown and described herein, any specific values ​​should be interpreted as merely exemplary and not as limitations; therefore, other examples of exemplary embodiments may have different values.

[0119] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0120] The above-described embodiments are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. A method for engine fault detection based on a single sensor, characterized in that, include: Acquire the raw eddy current signal and fiber pulse signal of the fiber optic eddy current composite sensor; The fiber optic eddy current composite sensor is installed in the casing of the aero-engine. The radial displacement signal of the blade is calculated based on the original eddy current signal, and the circumferential displacement signal of the blade is calculated based on the fiber pulse signal. The circumferential displacement signal of the blade is filtered to obtain the target circumferential displacement signal; The radial displacement signal of the blade is resampled to obtain the time-series blade tip clearance signal; Engine blade fault information is determined based on the target circumferential displacement signal and the timing blade tip clearance signal; The determination of engine blade fault information based on the target circumferential displacement signal includes: The single-blade signal is obtained based on the target circumferential displacement signal, and the blade phase is obtained based on the single-blade signal; The coefficient of variation is calculated based on the blade phase to obtain the variation characteristic parameters; The phase feature evaluation result is obtained based on the aforementioned variation feature parameters; An amplitude warning threshold is obtained based on the blade phase, and an amplitude feature evaluation result is obtained based on the amplitude warning threshold and the target circumferential displacement signal. Based on the phase characteristic evaluation results and the amplitude characteristic evaluation results, the engine blade fault information is determined to be surge information, resonance information, or foreign object impact fault information. The method further includes: Discrete wavelet decomposition is performed on the circumferential displacement signal of the blade to obtain wavelet coefficients; Calculate the squared modulus of the wavelet coefficients, and obtain the time-frequency spectrum based on the squared modulus; The blade vibration frequency is determined based on the aforementioned time-frequency spectrum. The frequency change rate is determined based on the blade vibration frequency and the preset modal frequency; If the rate of change of frequency is greater than a preset threshold, the engine blade fault information is determined to be blade crack information.

2. The engine fault detection method based on a single sensor according to claim 1, characterized in that, The step of calculating the blade radial displacement signal based on the original eddy current signal and calculating the blade circumferential displacement signal based on the fiber optic pulse signal includes: The original eddy current signal is filtered to obtain the filtered eddy current signal. Obtain the current rotational speed of the aircraft engine and determine the dynamic sensitivity corresponding to the current rotational speed; Based on the dynamic sensitivity, the eddy current filter signal is converted into the blade radial displacement signal; The fiber optic pulse signal is converted into a number of measurement time points. The difference between the preset theoretical time point number and the measured time point number is determined, and the circumferential displacement signal of the blade is obtained by multiplying the time point difference with the current rotational speed.

3. The engine fault detection method based on a single sensor according to claim 1, characterized in that, The step of resampling the radial displacement signal of the blade to obtain the time-series blade tip clearance signal includes: The radial displacement signal of the blade is resampled to obtain the blade tip clearance value; The blade tip clearance value is corrected and compensated based on the preset static processing dimensions of the blade to obtain the corrected blade tip clearance value. Traveling wave analysis was performed on the corrected blade tip gap value to obtain the time-series blade tip gap signal.

4. The engine fault detection method based on a single sensor according to claim 1, characterized in that, Determining engine blade fault information based on the timing blade tip clearance signal includes: The radial relative motion state is obtained by fitting the time-series blade tip gap signal; Based on the cross-correlation algorithm, the unbalanced amplitude and unbalanced phase are obtained according to the radial relative motion state; The engine blade fault information is determined to be engine imbalance fault information based on the imbalance amplitude and the imbalance phase.

5. The engine fault detection method based on a single sensor according to claim 1, characterized in that, Determining engine blade fault information based on the timing blade tip clearance signal includes: An anomaly detection algorithm is used to detect the timing blade tip gap signal to obtain anomaly signals; Time-frequency analysis was performed on the abnormal signal to obtain the time-frequency energy spectrum; If the time-frequency energy spectrum contains rubbing characteristics, then the engine blade fault information is determined to be rubbing fault information.

6. An engine fault detection device based on a single sensor, characterized in that, include: The acquisition module is used to acquire the raw eddy current signal and fiber pulse signal of the fiber optic eddy current composite sensor. The fiber optic eddy current composite sensor is installed in the casing of the aero-engine. The calculation module is used to calculate the blade radial displacement signal based on the original eddy current signal and the blade circumferential displacement signal based on the fiber pulse signal. The processing module is used to filter the circumferential displacement signal of the blade to obtain the target circumferential displacement signal; The radial displacement signal of the blade is resampled to obtain the time-series blade tip clearance signal; The determination module is used to determine engine blade fault information based on the target circumferential displacement signal and the timing blade tip clearance signal; The determining module is further configured to: obtain a single-blade signal based on the target circumferential displacement signal; acquire a blade phase based on the single-blade signal; calculate a coefficient of variation based on the blade phase to obtain a variation characteristic parameter; obtain a phase characteristic evaluation result based on the variation characteristic parameter; obtain an amplitude warning threshold based on the blade phase; obtain an amplitude characteristic evaluation result based on the amplitude warning threshold and the target circumferential displacement signal; and determine whether the engine blade fault information is surge information, resonance information, or foreign object impact fault information based on the phase characteristic evaluation result and the amplitude characteristic evaluation result. The determining module is further configured to perform discrete wavelet decomposition on the circumferential displacement signal of the blade to obtain wavelet coefficients; calculate the modulus square of the wavelet coefficients and obtain a time-frequency spectrum based on the modulus square; determine the blade vibration frequency based on the time-frequency spectrum; determine the frequency change rate based on the blade vibration frequency and a preset modal frequency; if the frequency change rate is greater than a preset threshold, then the engine blade fault information is determined to be blade crack information.

7. A computer device, characterized in that, It includes a memory and a processor, the memory storing a computer program that, when executed by the processor, implements the engine fault detection method based on a single sensor as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by a processor, implements the engine fault detection method based on a single sensor as described in any one of claims 1 to 5.

Images