Vibration data analysis method and system based on narrowband spectrum filtering

Abstract

The embodiment of the invention discloses a vibration data analysis method and system based on narrowband spectrum filtering. The method comprises the steps that original vibration data are imported, engine working condition recognition is conducted, the original vibration data are intercepted and windowed according to the engine working condition, and data to be analyzed are obtained; carrying out filtering and spectrum analysis on the data to be analyzed by using a band-pass filter and Cooley-Tukey FFT (Fast Fourier Transform) transform; performing vibration effective component screening on a spectral analysis result; calculating characteristic parameters based on the effective vibration components; and state monitoring and fault diagnosis are carried out on the whole engine and parts according to the characteristic parameters. According to the method, the Cooley-Tukey FFT is introduced for vibration data analysis, the defects of DFT / FFT can be overcome, support is provided for healthy management development of an engine, noise influences caused by factors such as electromagnetic interference and vibration signal interference of a non-monitoring object vibration source can be effectively overcome, and the calculation error of the total vibration amount is reduced.

Description

Technical Field

[0001] This invention relates to the field of engine condition monitoring technology, specifically to a vibration data analysis method and system based on narrowband spectral filtering. Background Technology

[0002] The main structure and principle of existing vibration data analysis methods are as follows:

[0003] The engine health monitoring unit can acquire, process, and record sensor signals such as engine vibration and speed. This unit can also identify and diagnose typical engine faults and evaluate engine performance. Furthermore, it can download data via Ethernet, allowing for further in-depth analysis and fault diagnosis using vibration and speed data in a ground-based analysis system. Based on the engine data collected and recorded by the monitoring unit, characteristic parameters are calculated through time-domain and frequency-domain analysis of vibration and speed data. These characteristic parameters are then used to monitor engine condition and identify and diagnose typical faults.

[0004] As a vibration system, the engine is a multi-degree-of-freedom vibration system with diverse vibration sources. These primarily include the turbine rotor and blades, while the components in the gearbox and transmission structure of the accessory gearbox are also significant vibration sources. The combined effect of these vibration sources reflects the overall vibration status of the engine. During engine operation, sensors measure engine vibration data and calculate its characteristic quantities. This serves two purposes: firstly, it allows for online monitoring of typical mechanical faults, enabling engine safety monitoring; secondly, by controlling engine vibration levels to a low level, it extends the engine's service life and saves operating costs.

[0005] In vibration data analysis, digital filtering and spectrum analysis are important analytical methods. Discrete Fourier Transform (DFT) and Fast Fourier Transform (FFT) algorithms are fundamental tools for spectrum analysis of digital signals. Digital filtering yields a high signal-to-noise ratio (SNR) effective signal, and then DFT or FFT is used to obtain the signal's spectral distribution. Spectral analysis is then used to calculate key characteristic parameters of engine vibration data, which are then used as important parameters for engine condition monitoring in engineering applications, thereby enabling typical engine fault diagnosis and health assessment.

[0006] However, the existing vibration data analysis methods described above have the following problems and drawbacks:

[0007] The commonly used spectrum analysis methods in existing vibration data analysis methods (such as Discrete Fourier Transform or Fast Fourier Transform) are not flexible enough in signal processing. The Discrete Fourier Transform has an exponential increase in computational complexity as the amount of input signal data increases, while the Fast Fourier Transform requires the number of sampling points of the input signal to be an integer power of 2, which limits its application in certain situations.

[0008] Meanwhile, due to electromagnetic interference and vibration signal interference from non-monitored vibration sources during actual engine vibration data acquisition, the signal-to-noise ratio of the monitored object's vibration signal is poor. As a result, the calculation methods for key characteristic parameters of vibration data, such as the total vibration amount, in existing vibration data analysis methods are not scientific enough. Noise signals are easily superimposed into the calculation of the total vibration amount, resulting in a large calculation error when the signal-to-noise ratio is poor. Summary of the Invention

[0009] In view of the deficiencies of the prior art mentioned in the background, the purpose of this invention is to provide a vibration data analysis method and system based on narrowband spectral filtering.

[0010] To achieve the above objectives, in a first aspect, embodiments of the present invention provide a vibration data analysis method based on narrowband spectral filtering, comprising:

[0011] Import the raw vibration data and engine parameter data; the engine parameter data contains all the required engine parameters, such as Ng, T45, etc.

[0012] The engine operating condition is obtained by identifying the engine operating parameters.

[0013] Based on the engine operating conditions, the original vibration data is extracted and windowed to obtain the data to be analyzed.

[0014] The data to be analyzed was filtered and spectral analyzed using a bandpass filter and Cooley-Tukey FFT transform to obtain the spectral distribution map of the original vibration data.

[0015] The effective vibration components are obtained by screening the spectral distribution map.

[0016] The characteristic parameters are calculated based on the effective vibration components; the characteristic parameters include the total vibration of the engine, the fundamental frequency amplitude of the vibration of each key component, the first harmonic and the second harmonic;

[0017] The engine and its components are monitored for condition and diagnosed for faults based on the aforementioned characteristic parameters.

[0018] As one specific implementation of this application, the data to be analyzed is obtained as follows:

[0019] The original vibration data is segmented according to the engine operating conditions to obtain segmented vibration data.

[0020] The segmented vibration data is windowed to obtain the data to be analyzed; the window functions used in the windowing process include Hanning window, Hamming window and Blackman window.

[0021] As a specific implementation of this application, the effective vibration component is obtained as follows:

[0022] The spectrum distribution diagram is windowed and corrected. The frequency band range is selected from the spectrum distribution diagram according to the engine operating frequency. The maximum frequency amplitude value is selected in the corresponding frequency band range to obtain the effective vibration component of the operating frequency of each engine component, which is the effective vibration component.

[0023] As a specific implementation of this application, condition monitoring and fault diagnosis of the engine and its components are performed based on the aforementioned characteristic parameters, specifically as follows:

[0024] Determine feature thresholds and compare the feature parameters with the feature thresholds to achieve condition monitoring and fault diagnosis of the engine as a whole and its components.

[0025] Specifically, determining the feature threshold involves:

[0026] Features are selected from time-domain features, frequency-domain features, or time-domain synchronous average features. Based on the selected features, a feature threshold is established using the 3σ rule. The time-domain features include average value, peak-to-peak value, effective value, kurtosis, and crest factor. The frequency-domain features include first-order harmonics and second-order harmonics.

[0027] Optionally, as an alternative implementation of this application, condition monitoring and fault diagnosis of the engine and its components are performed based on the aforementioned characteristic parameters, specifically as follows:

[0028] Establish alarm rules; the alarm rules are to trigger an alarm if multiple current features exceed a feature threshold simultaneously, or if any current feature exceeds a feature threshold multiple times consecutively; the current features are the total vibration of the engine, the fundamental frequency amplitude of vibration of each key component, the first harmonic and the second harmonic;

[0029] Based on the aforementioned characteristic parameters and alarm rules, the condition monitoring and fault diagnosis of the entire engine and its components can be achieved.

[0030] Secondly, embodiments of the present invention also provide a vibration data analysis system based on narrowband spectral filtering, comprising:

[0031] The data acquisition module is used to import raw vibration data and engine parameter data, identify engine operating conditions through the engine parameter data to obtain engine operating conditions, and perform data truncation and windowing on the raw vibration data according to the engine operating conditions to obtain the data to be analyzed.

[0032] The spectrum analysis module is used to filter and perform spectrum analysis on the data to be analyzed using a bandpass filter and Cooley-Tukey FFT transform to obtain the spectrum distribution map of the original vibration data.

[0033] The screening module is used to screen the effective vibration components of the spectrum distribution map to obtain the effective vibration components;

[0034] The calculation module is used to calculate characteristic parameters based on the effective vibration components; the characteristic parameters include the total vibration of the engine, the fundamental frequency amplitude of vibration of each key component, the first harmonic and the second harmonic;

[0035] The condition monitoring and fault diagnosis module is used to perform condition monitoring and fault diagnosis on the engine and its components based on the characteristic parameters.

[0036] Thirdly, embodiments of the present invention also provide a vibration data analysis system based on narrowband spectrum filtering, including a processor, an input device, an output device, and a memory, wherein the processor, input device, output device, and memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the method described in the first aspect.

[0037] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0038] 1. When analyzing engine vibration data, the Cooley-Tukey frequency domain transform algorithm was used instead of the Discrete Fourier Transform or Fast Fourier Transform. The number of sampling points N of the input signal can be arbitrarily selected, and its computation speed has been optimized to reduce the number of operations from N... 2 Optimized to Nlog r (N); that is, the embodiments of the present invention can achieve a significant increase in the operation speed when the amount of input signal data increases, while having no requirement for the number of sampling points of the input signal, thus overcoming the shortcomings of discrete Fourier transform or fast Fourier transform.

[0039] 2. Combining bandpass digital filters and the Cooley-Tukey frequency domain transform algorithm can create a highly effective narrowband spectrum analysis system. Introducing this system into the analysis of actual vibration data allows for the demonstration of the feasibility of an engine health management system and vibration-based condition monitoring and fault diagnosis. This enables the application of vibration data analysis results to overall engine and component condition monitoring and fault diagnosis, supporting the gradual maturation and development of the engine health management system. Ultimately, this improves engine maintenance efficiency, reduces maintenance costs, and ensures safety. Furthermore, it allows for a shift from scheduled maintenance to condition-based maintenance, further enhancing engine readiness and mission success rates.

[0040] 3. For key characteristic parameters of vibration data, such as the calculation of total vibration, the method of filtering effective components of vibration signal and then weighting and averaging can effectively overcome the noise influence caused by electromagnetic interference and vibration signal interference from non-monitored vibration sources, and reduce the error in calculating total vibration. Attached Figure Description

[0041] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below.

[0042] Figure 1 This is a flowchart of a vibration data analysis method based on narrowband spectral filtering provided in an embodiment of the present invention;

[0043] Figure 2 This is a flowchart for calculating characteristic values ​​such as total engine vibration;

[0044] Figure 3 This is a schematic diagram of a digital narrowband spectrum analysis system;

[0045] Figure 4 This is a schematic diagram of a frequency decimation algorithm with a length of 8 and a radix-2.

[0046] Figure 5 This is a structural diagram of the vibration data analysis system based on narrowband spectral filtering provided in an embodiment of the present invention;

[0047] Figure 6 yes Figure 5 Another structural diagram of the system shown;

[0048] Figure 7 These are the raw vibration data of the engine platform;

[0049] Figure 8 This is a vibration spectrum diagram of the engine platform;

[0050] Figure 9 This is a graph showing the results of vibration data extraction and eigenvalue calculation. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0053] The inventive concept of this invention is as follows: First, a narrowband spectrum analysis system is constructed by combining a digital filter and the Cooley-Tukey FFT algorithm. The number of sampling points N of the input signal can be the transform length, and the number of calculations is significantly optimized. Second, a vibration data analysis method based on narrowband spectrum filtering is used to process the vibration data, extracting various characteristic values ​​of engine vibration. The frequency band range is selected according to the engine operating frequency to obtain the effective vibration components of the operating frequencies of each engine component, which can effectively improve the calculation accuracy of vibration amplitude. At the same time, based on the spectrum analysis results, characteristic parameters such as the first harmonic and second harmonic of the operating frequencies of each engine component can be calculated, realizing the status monitoring of the entire engine and its components.

[0054] Please refer to Figure 1 and Figure 2 The vibration data analysis method based on narrowband spectral filtering provided in this embodiment of the invention includes the following steps:

[0055] S1. Import the original vibration data and engine parameter data. Identify the engine operating condition using the engine parameter data to obtain the engine operating condition. Based on the engine operating condition, extract and window the original vibration data to obtain the data to be analyzed.

[0056] In practice, step S1 is as follows:

[0057] Engine operating conditions are identified based on engine parameter data; this data includes all necessary engine parameters, such as Ng and T45.

[0058] The original vibration data is segmented according to the engine operating conditions to obtain segmented vibration data.

[0059] The segmented vibration data is windowed to obtain the data to be analyzed.

[0060] It should be noted that windowing is used to reduce spectral leakage and ensure the accuracy of the calculation of each vibration component in the spectrum. The window functions used in this embodiment include: Hanning window, Hamming window, and Blackman window.

[0061] S2, use a bandpass filter and Cooley-Tukey FFT transform to filter and perform spectral analysis on the data to be analyzed, and obtain the spectral distribution map of the original vibration data.

[0062] In this embodiment, a bandpass filter is used to filter the vibration data (i.e., the aforementioned data to be analyzed). A reasonable cutoff frequency is set to effectively filter high-frequency and low-frequency noise signals, thereby extracting vibration data with a high signal-to-noise ratio. Spectral analysis is performed using Cooley-Tukey FFT transform, which can efficiently calculate the spectral distribution of the vibration data.

[0063] That is, embodiments of the present invention provide a digital narrowband spectrum analysis system, the data processing process of which is as follows: Figure 3 As shown in the figure. The parameters in the figure are explained as follows:

[0064] x(t): The input continuous-time signal;

[0065] x(n): The digital signal obtained by sampling x(t);

[0066] NBDF: Narrowband Digital Filter;

[0067] CT FFT: Cooley-Tukey Fast Spectrum Calculation Method;

[0068] FAP: Calculate the amplitude spectrum and phase spectrum;

[0069] x i (n): The i-th digital signal;

[0070] X i (k): Discrete spectrum data of the i-th path;

[0071] Ai(k): The frequency amplitude of the i-th signal;

[0072] φ i (k): The phase of the i-th signal.

[0073] As shown in the figure, when a continuous-time signal x(t) or a digital signal x(n) is input, the spectrum of each segment of the input signal is obtained at the output of each branch.

[0074] The Cooley-Tukey algorithm is the most versatile of all FFT algorithms because it allows for arbitrary factorization, meaning the transform length N can be chosen arbitrarily. This allows the use of N = r s The radix-r algorithm.

[0075] The Cooley-Tukey algorithm, where N = N1·N2 points, can be performed using the following steps:

[0076] (1) Calculate the index transformation of the input sequence;

[0077] (2) Calculate N2 DFTs of length N1;

[0078] (3) Apply a twitch factor to the output of the first transform stage;

[0079] (4) Calculate N1 DFTs of length N2;

[0080] (5) Calculate the index transformation of the output sequence.

[0081] Figure 4 A diagram of the 8-point Cooley-Tukey FFT is given. In the signal flow diagram, the radix-2 point DFT is drawn as a butterfly diagram. The signal flow diagram has been simplified so that all arrows pointing to a node represent addition, and constant coefficient multiplication is represented by adding a factor to the arrow. The radix-r algorithm has logarithmic power. r There are N levels, and each group has the same type of twitch factor.

[0082] As can be seen from the signal flow diagram, the storage locations used by the butterfly algorithm can be rewritten, and the data is no longer needed in the next calculation. Therefore, the Cooley-Tukey algorithm overcomes the limitation of the FFT algorithm where the number of input signal sampling points N must be an integer power of 2, while significantly reducing the number of calculations compared to the DFT, thus greatly improving computational efficiency.

[0083] S3. Screen the effective vibration components from the spectral distribution map to obtain the effective vibration components.

[0084] The main source of vibration in the engine is the turbine rotor, while the components in the gearbox and transmission of the accessory gearbox are also significant sources of vibration. Due to the numerous vibration sources, each has some impact on the sensors, resulting in data collected by the sensors containing frequency components from multiple components. Furthermore, the engine data acquisition system is also susceptible to interference from complex electromagnetic environments, frequently resulting in interference components mixed into the signal.

[0085] Based on this, this embodiment of the invention, building upon the frequency domain calculation method, extracts the characteristic spectrum and vibration amplitude values, first harmonics, second harmonics, and other parameters of key components, adds a filtering function for the effective components of each frequency, extracts the effective vibration components, and then calculates the total vibration value. Specifically, the spectrum distribution diagram is windowed, and frequency bands are filtered according to the engine operating frequency in the spectrum distribution diagram. The maximum amplitude value is selected within the corresponding frequency band range to obtain the effective vibration components of each engine component's operating frequency, which are the effective vibration components. The windowing process is described in step S1 and will not be repeated here.

[0086] It should be noted that the selection of effective components in the calculation method is a key step in the vibration total extraction algorithm. According to the spectral characteristics in Fourier transform, only the peak point in each frequency spike is the true frequency component value, while the other parts are spectral leakage results, and only the peak point is close to the true value.

[0087] Therefore, the purpose of screening is to extract the peak points of the operating frequency of each engine component. First, the frequency band range is screened according to the engine operating frequency, and then the maximum frequency amplitude value is selected in the corresponding frequency band range, which is the effective vibration component of the operating frequency of each engine component, so as to obtain the effective vibration component from the spectrum data.

[0088] S4, Calculate characteristic parameters based on the effective vibration components.

[0089] The characteristic parameters include the total vibration of the engine, the fundamental frequency amplitude of vibration of each key component, the first harmonic and the second harmonic, etc.

[0090] In practice, vibration acceleration sensors are installed at appropriate locations on the engine to collect its vibration information. The vibration data is processed, primarily through time-domain analysis, frequency analysis, and time-frequency analysis, to extract various characteristic indicators of engine vibration and monitor the condition of the entire engine and its components. Based on the effective vibration components in the frequency spectrum, the total vibration of the engine, the fundamental frequency amplitude of each key component, and characteristic parameters such as first and second harmonics can be calculated.

[0091] The total vibration is calculated by performing a digital integration of the vibration frequency amplitude from the 0th to the 2nd order according to the final required characteristic quantity form (acceleration, velocity or displacement) to obtain the total vibration.

[0092] Other common characteristic parameters include maximum value, minimum value, peak-to-peak value, kurtosis, crest factor, first harmonic, and second harmonic.

[0093] S5, perform condition monitoring and fault diagnosis on the engine and its components based on the aforementioned characteristic parameters.

[0094] For engine condition monitoring and fault diagnosis, commonly used features include time-domain features, frequency-domain features, and some features of time-domain synchronous averaging. Among them, time-domain features include average value, peak-to-peak value, RMS value, kurtosis, crest factor, etc., while frequency-domain features include first-order harmonics, second-order harmonics, etc.

[0095] To determine suitable features for this engine and reasonable thresholds for these features, these features need to be screened. After screening, feature thresholds for multiple states or a single state are established based on the screened features.

[0096] Engines typically operate in a non-defined state, especially with very few fault conditions, resulting in extremely limited data samples. For such samples, feature thresholds can be established using methods such as the 3σ rule. Exceeding these thresholds (red or yellow) requires alerts or warnings.

[0097] That is, condition monitoring and fault diagnosis of the engine and its components are performed based on characteristic parameters, specifically:

[0098] Features are selected from time-domain features, frequency-domain features, or time-domain synchronous average features. Based on the selected features, a feature threshold is established using the 3σ rule. The time-domain features include average value, peak-to-peak value, RMS value, kurtosis, and crest factor. The frequency-domain features include first-order harmonics and second-order harmonics.

[0099] The characteristic parameters and characteristic thresholds are compared to achieve condition monitoring and fault diagnosis of the engine as a whole and its components.

[0100] Furthermore, since the operating conditions of rotating machinery such as engines are constantly changing and new conditions may arise, reasonable alarm rules need to be established to reduce the false alarm rate. This can be approached from two aspects: first, combining several characteristics so that an alarm is triggered when they all exceed their limits simultaneously; second, for a specific characteristic, an alarm is triggered if it exceeds its limit three times consecutively. When alarms are triggered for the characteristics corresponding to certain components, it indicates that these components may have malfunctioned.

[0101] That is, alarm rules can also be established to perform condition monitoring and fault diagnosis on the engine as a whole and its components, specifically:

[0102] Establish alarm rules; the alarm rules are to trigger an alarm if multiple current features exceed a feature threshold simultaneously, or if any current feature exceeds a feature threshold multiple times consecutively; the current features are the total vibration of the engine, the fundamental frequency amplitude of vibration of each key component, the first harmonic and the second harmonic;

[0103] Based on the aforementioned characteristic parameters and alarm rules, the condition monitoring and fault diagnosis of the entire engine and its components can be achieved.

[0104] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0105] 1. When analyzing engine vibration data, the Cooley-Tukey frequency domain transform algorithm was used instead of the Discrete Fourier Transform or Fast Fourier Transform. The number of sampling points N of the input signal can be arbitrarily selected, and its computation speed has been optimized to reduce the number of operations from N... 2 Optimized to Nlog r (N); that is, the embodiments of the present invention can achieve a significant increase in the operation speed when the amount of input signal data increases, while having no requirement for the number of sampling points of the input signal, thus overcoming the shortcomings of discrete Fourier transform or fast Fourier transform.

[0106] 2. Combining bandpass digital filters and the Cooley-Tukey frequency domain transform algorithm can create a highly effective narrowband spectrum analysis system. Introducing this system into the analysis of actual vibration data allows for the demonstration of the feasibility of an engine health management system and vibration-based condition monitoring and fault diagnosis. This enables the application of vibration data analysis results to overall engine and component condition monitoring and fault diagnosis, supporting the gradual maturation and development of the engine health management system. Ultimately, this improves engine maintenance efficiency, reduces maintenance costs, and ensures safety. Furthermore, it allows for a shift from scheduled maintenance to condition-based maintenance, further enhancing engine readiness and mission success rates.

[0107] 3. For key characteristic parameters of vibration data, such as the calculation of total vibration, the method of filtering effective components of vibration signal and then weighting and averaging can effectively overcome the noise influence caused by electromagnetic interference and vibration signal interference from non-monitored vibration sources, and reduce the error in calculating total vibration.

[0108] Based on the same inventive concept, embodiments of the present invention also provide a vibration data analysis system based on narrowband spectral filtering, such as... Figure 5 As shown, it includes:

[0109] The data acquisition module is used to import raw vibration data and engine parameters, identify engine operating conditions through the engine parameters, obtain engine operating conditions, and perform data truncation and windowing on the raw vibration data based on the engine operating conditions to obtain the data to be analyzed.

[0110] The spectrum analysis module is used to filter and perform spectrum analysis on the data to be analyzed using a bandpass filter and Cooley-Tukey FFT transform to obtain the spectrum distribution map of the original vibration data.

[0111] The screening module is used to screen the effective vibration components of the spectrum distribution map to obtain the effective vibration components;

[0112] The calculation module is used to calculate characteristic parameters based on the effective vibration components; the characteristic parameters include the total vibration of the engine, the fundamental frequency amplitude of vibration of each key component, the first harmonic and the second harmonic;

[0113] The condition monitoring and fault diagnosis module is used to perform condition monitoring and fault diagnosis on the engine and its components based on the characteristic parameters.

[0114] Specifically, the data to be analyzed module is used for:

[0115] Engine operating conditions are identified based on engine parameters; these engine parameters include Ng, T45, etc.

[0116] The original vibration data is segmented according to the engine operating conditions to obtain segmented vibration data.

[0117] The segmented vibration data is windowed to obtain the data to be analyzed; the window functions used in the windowing process include Hanning window, Hamming window and Blackman window.

[0118] Furthermore, the filtering module is specifically used for:

[0119] The spectrum distribution diagram is windowed and corrected. The frequency band range is selected from the spectrum distribution diagram according to the engine operating frequency. The maximum frequency amplitude value is selected in the corresponding frequency band range to obtain the effective vibration component of the operating frequency of each engine component, which is the effective vibration component.

[0120] Furthermore, the condition monitoring and fault diagnosis module is specifically used for:

[0121] Features are selected from time-domain features, frequency-domain features, or time-domain synchronous average features. Based on the selected features, a feature threshold is established using the 3σ rule. The time-domain features include average value, peak-to-peak value, RMS value, kurtosis, and crest factor. The frequency-domain features include first-order harmonics and second-order harmonics.

[0122] The characteristic parameters and characteristic thresholds are compared to achieve condition monitoring and fault diagnosis of the engine as a whole and its components.

[0123] Optionally, the condition monitoring and fault diagnosis module can also be used for:

[0124] Establish alarm rules; the alarm rules are to trigger an alarm if multiple current features exceed a feature threshold simultaneously, or if any current feature exceeds a feature threshold multiple times consecutively; the current features are the total vibration of the engine, the fundamental frequency amplitude of vibration of each key component, the first harmonic and the second harmonic;

[0125] Based on the aforementioned characteristic parameters and alarm rules, the condition monitoring and fault diagnosis of the entire engine and its components can be achieved.

[0126] It should be noted that the specific workflow of this embodiment is described in the foregoing method embodiment section, and will not be repeated here.

[0127] Furthermore, another embodiment of the present invention also provides a vibration data analysis system based on narrowband spectral filtering. For example... Figure 6As shown, the system may include one or more processors 101, one or more input devices 102, one or more output devices 103, and a memory 104. The processors 101, input devices 102, output devices 103, and memory 104 are interconnected via a bus 105. The memory 104 stores a computer program, which includes program instructions. The processor 101 is configured to invoke the program instructions to execute the method described in the above-described method embodiment.

[0128] It should be understood that, in this embodiment of the invention, the processor 101 may be a central processing unit (CPU), but it may also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0129] Input device 102 may include a keyboard, etc., and output device 103 may include a display (LCD, etc.), a speaker, etc.

[0130] The memory 104 may include read-only memory and random access memory, and provides instructions and data to the processor 101. A portion of the memory 104 may also include non-volatile random access memory. For example, the memory 104 may also store device type information.

[0131] In specific implementations, the processor 101, input device 102, and output device 103 described in the embodiments of the present invention can execute the implementation methods described in the embodiments of the vibration data analysis method based on narrowband spectrum filtering provided in the embodiments of the present invention, which will not be repeated here.

[0132] To better understand the vibration data analysis method and system provided in the embodiments of the present invention, a more detailed description is given below using the vibration data of a certain type of turboshaft engine as an example:

[0133] The main vibration sources of a turboshaft engine are the gas turbine rotor and the free turbine rotor. Additionally, the components in the gearbox and transmission of the accessory gearbox are also significant vibration sources. The vibration data source for analysis includes data from two vibration sensors: Sensor 1 primarily collects vibration data from the accessory gearbox and gas turbine, while Sensor 2 primarily collects vibration data from the free turbine and gas turbine. Due to the numerous vibration sources, each has a certain impact on the sensors; therefore, the data collected by both sensors contains frequency components from multiple components. Furthermore, the engine data acquisition system is also subject to interference from a complex electromagnetic environment, frequently resulting in interference components in the signal. Figure 7 This is the raw data of engine platform vibration. Figure 8 This is a vibration spectrum diagram of the engine platform.

[0134] Digital narrowband spectrum filtering modules are used for digital filtering and spectrum analysis. Based on frequency domain calculation methods, characteristic spectra of key components and parameters such as vibration amplitude, first harmonic, and second harmonic are extracted. The module adds a filtering function for effective components at each frequency to extract effective vibration components, and then calculates the total vibration value. The vibration data extraction and characteristic value calculation results are as follows: Figure 9 As shown.

[0135] The data acquisition system collects and records vibration data, rotational speed, and other data during the engine bench test. The data can be monitored by the host computer software. After the data is unloaded, vibration data analysis software is used to perform data analysis. The analysis results of the three data segments are compared with the corresponding vibration data analysis results collected in the bench test laboratory. The Ng rotational speed corresponding to two data segments is 35000 r / min, and the Ng rotational speed corresponding to one data segment is 36500 r / min.

[0136] Using the spectral analysis results of vibration sensor data collected in the vehicle test laboratory as a control group, the results of spectral analysis of vibration data collected by the data acquisition system were compared. The average error of the fundamental frequency amplitude was 4.83%. As shown in the table below, Ng_PF represents the vehicle test system's Ng fundamental frequency amplitude, Np_PF represents the vehicle test system's Np fundamental frequency amplitude, Ng_DC represents the data acquisition system's Ng fundamental frequency amplitude, Np_DC represents the data acquisition system's Np fundamental frequency amplitude, error_ng represents the Ng fundamental frequency amplitude error, and error_np represents the Np fundamental frequency amplitude error. Please refer to Table 1.

[0137]

[0138] Table 1

[0139] During the engine test on the test bench, the data acquisition system collected vibration and speed data. Spectral analysis was performed on the vibration data collected by multiple data acquisition systems, and the results were compared with those from the vibration data collected in the test bench laboratory. Some differences were observed, but these were within the normal range. The trends in the total vibration and fundamental frequency amplitude as the gas turbine speed remained stable were consistent with the results of the test bench spectral analysis.

[0140] Based on engine vibration data acquired by the data acquisition system, a narrowband spectrum analysis system is used to process the vibration data, mainly including time-domain analysis, spectrum analysis, and time-frequency analysis. Multiple characteristic indication (CI) values ​​of engine vibration are extracted. Frequency bands are selected based on the engine's operating frequency, and then the maximum amplitude value within the corresponding frequency band is selected as the effective vibration component of each engine component's operating frequency. The total engine vibration is then calculated, effectively eliminating interference factors such as noise. Furthermore, characteristic parameters such as the first and second harmonics of each engine component's operating frequency can be calculated based on the spectrum data to be analyzed, enabling condition monitoring of the entire engine and its components. Finally, the analysis method for vibration data acquired in the vehicle test laboratory and the vibration data acquisition system are compared with the analysis results of 60 sets of vibration data acquired by the data acquisition system, demonstrating that the calculation results and accuracy of this vibration data analysis method meet the actual engineering needs.

[0141] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A vibration data analysis method based on narrowband spectral filtering, characterized in that, include: Import the raw vibration data and emission parameter data; Engine operating conditions are identified using the aforementioned engine parameter data; Based on the engine operating conditions, the original vibration data is truncated and windowed to obtain the data to be analyzed. The data to be analyzed was filtered and spectral analyzed using a bandpass filter and Cooley-Tukey FFT transform to obtain the spectral distribution map of the original vibration data. The effective vibration components are obtained by screening the spectral distribution map. The characteristic parameters are calculated based on the effective vibration components; the characteristic parameters include the total vibration of the engine, the fundamental frequency amplitude of the vibration of each key component, the first harmonic and the second harmonic; The engine and its components are monitored for condition and diagnosed for faults based on the aforementioned characteristic parameters.

2. The vibration data analysis method as described in claim 1, characterized in that, The specific data to be analyzed is as follows: The original vibration data is segmented according to the engine operating conditions to obtain segmented vibration data. The segmented vibration data is windowed to obtain the data to be analyzed; the window functions used in the windowing process include Hanning window, Hamming window and Blackman window.

3. The vibration data analysis method as described in claim 1, characterized in that, The effective vibration components are obtained as follows: The spectrum distribution diagram is windowed and corrected. The frequency band range is selected from the spectrum distribution diagram according to the engine operating frequency. The maximum frequency amplitude value is selected in the corresponding frequency band range to obtain the effective vibration component of the operating frequency of each engine component, which is the effective vibration component.

4. The vibration data analysis method as described in claim 1, characterized in that, Based on the aforementioned characteristic parameters, condition monitoring and fault diagnosis are performed on the engine as a whole and its components, specifically as follows: Determine feature thresholds and compare the feature parameters with the feature thresholds to achieve condition monitoring and fault diagnosis of the engine as a whole and its components.

5. The vibration data analysis method as described in claim 4, characterized in that, The specific steps for determining the feature threshold are as follows: Features are selected from time-domain features, frequency-domain features, or time-domain synchronous average features. Based on the selected features, a feature threshold is established using the 3σ rule. The time-domain features include average value, peak-to-peak value, effective value, kurtosis, and crest factor. The frequency-domain features include first-order harmonics and second-order harmonics.

6. The vibration data analysis method as described in claim 1, characterized in that, Based on the aforementioned characteristic parameters, condition monitoring and fault diagnosis are performed on the engine as a whole and its components, specifically as follows: Establish alarm rules; the alarm rules are to trigger an alarm if multiple current features exceed a feature threshold simultaneously, or if any current feature exceeds a feature threshold multiple times consecutively; the current features are the total vibration of the engine, the fundamental frequency amplitude of vibration of each key component, the first harmonic and the second harmonic; Based on the aforementioned characteristic parameters and alarm rules, the condition monitoring and fault diagnosis of the entire engine and its components can be achieved.

7. A vibration data analysis system based on narrowband spectral filtering, characterized in that, include: The data acquisition module is used to import raw vibration data and engine parameter data, identify engine operating conditions through the engine parameter data to obtain engine operating conditions, and perform data truncation and windowing on the raw vibration data according to the engine operating conditions to obtain the data to be analyzed. The spectrum analysis module is used to filter and perform spectrum analysis on the data to be analyzed using a bandpass filter and Cooley-Tukey FFT transform to obtain the spectrum distribution map of the original vibration data. The screening module is used to screen the effective vibration components of the spectrum distribution map to obtain the effective vibration components; The calculation module is used to calculate characteristic parameters based on the effective vibration components; The characteristic parameters include the total vibration of the engine, the fundamental frequency amplitude of vibration of each key component, the first harmonic and the second harmonic; The condition monitoring and fault diagnosis module is used to perform condition monitoring and fault diagnosis on the engine and its components based on the characteristic parameters.

8. The vibration data analysis system as described in claim 7, characterized in that, The data module to be analyzed is specifically used for: The original vibration data is segmented according to the engine operating conditions to obtain segmented vibration data. The segmented vibration data is windowed to obtain the data to be analyzed; the window functions used in the windowing process include Hanning window, Hamming window and Blackman window.

9. The vibration data analysis system as described in claim 7, characterized in that, The status monitoring and fault diagnosis module is specifically used for: Features are selected from time-domain features, frequency-domain features, or time-domain synchronous average features. Based on the selected features, a feature threshold is established using the 3σ rule. The time-domain features include average value, peak-to-peak value, RMS value, kurtosis, and crest factor. The frequency-domain features include first-order harmonics and second-order harmonics. The characteristic parameters and characteristic thresholds are compared to achieve condition monitoring and fault diagnosis of the engine as a whole and its components.

10. A vibration data analysis system based on narrowband spectral filtering, characterized in that, The system includes a processor, an input device, an output device, and a memory, which are interconnected. The memory is used to store a computer program, which includes program instructions. The processor is configured to invoke the program instructions to perform the method as described in any one of claims 1-6.

Images