Turbo mechanical blade resonance interval identification method, device and equipment and storage medium
By processing the tip pulse signal of turbine blades using adaptive ultrawavelet transform and undersampling principle, the resonance interval is accurately identified, solving the problem that traditional methods are difficult to separate resonance features at high speeds, and realizing accurate resonance point identification in complex environments.
Patent Information
- Application Number
- CN202511142466.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-15
- Publication Date
- 2025-11-18
AI Technical Summary
Traditional methods are difficult to effectively separate the resonance characteristics of turbine blades under high speed conditions, making it difficult to locate blade resonance events, especially under complex vibration signals and aliasing phenomena.
Adaptive ultrawavelet transform is used to process the tip pulse signal of turbine machinery. Noise is removed through preprocessing, the vibration displacement of the target blade is calculated, and the two-dimensional relationship between aliasing frequency and rotational frequency is identified by combining the undersampling principle, thereby accurately identifying the resonance interval.
In environments with weak resonance characteristics and strong noise, the resonance point of the blade can be accurately identified, which improves the accuracy of resonance event identification and avoids the problem of missed detection by traditional methods.
Smart Images
Figure CN120970805A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of blade resonance identification, and in particular to a turbine blade resonance interval identification method, device, equipment and storage medium. BACKGROUND
[0002] Turbines (such as steam turbines, compressors, etc.) have a wide range of applications in industrial production, and the health status of their blades directly affects the operating efficiency and safety of the equipment. Blade resonance is one of the main causes of blade fatigue and damage, so it is very important to locate the occurrence of resonance events. Traditional resonance event positioning often relies on experienced engineers, but under high speed conditions, the complexity and aliasing of blade vibration signals make it difficult for engineers to locate resonance events manually. In recent years, Blade Tip Timing (BTT) technology, as a non-contact measurement method, has received widespread attention as it can obtain real-time vibration information of rotating blades. However, traditional signal processing methods cannot effectively separate resonance characteristics, making it difficult for many resonance event-based algorithms to be truly applied in turbine blade health monitoring. Therefore, there is an urgent need for an efficient and accurate blade resonance event positioning method to solve the above problems. SUMMARY
[0003] Therefore, the present application aims to overcome the deficiencies in the prior art and provide a turbine blade resonance interval identification method, device, equipment and storage medium for adaptively processing BTT signals using superwavelet transform and accurately identifying blade resonance events by analyzing aliasing frequencies in the undersampled signal.
[0004] The present application provides the following technical solutions: In a first aspect, the present application provides a turbine blade resonance interval identification method, comprising: obtaining a blade tip pulse signal of a turbine; preprocessing the blade tip pulse signal and calculating a target blade vibration displacement based on the preprocessed blade tip pulse signal; performing segmented time-frequency analysis on the target blade vibration displacement based on adaptive superwavelet transform to obtain a two-dimensional relationship between aliasing frequencies and rotational frequencies; identifying a resonance interval based on the two-dimensional relationship between aliasing frequencies and rotational frequencies according to the undersampling principle.
[0005] In an embodiment, the preprocessing of the blade tip pulse signal comprises: removing narrow pulse noise signals in the blade tip pulse signal based on a preset sliding window length to obtain candidate pulse signals in each sliding window; calculating the time interval of adjacent candidate pulse signals in each sliding window; For each of the sliding windows, a time interval average value corresponding to each of the sliding windows is calculated according to the time interval; An interval reference threshold value corresponding to each of the sliding windows is determined according to a preset ratio and the time interval average value; For each of the sliding windows, the candidate pulse signal is linearly interpolated according to the time interval and the interval reference threshold value to obtain the preprocessed blade tip pulse signal.
[0006] In an embodiment, the target blade vibration displacement is calculated according to the preprocessed blade tip pulse signal, comprising: An initial blade vibration displacement is calculated according to the preprocessed blade tip pulse signal; The initial blade vibration displacement is subjected to trend item removal to obtain the target blade vibration displacement.
[0007] In an embodiment, the initial blade vibration displacement is calculated according to the preprocessed blade tip pulse signal, comprising: Based on a linear relationship between a blade theoretical reaching time and an arc corresponding to blade accumulation, a least square method linear fitting is solved according to the preprocessed blade tip pulse signal to obtain the blade theoretical reaching time; A blade rotation speed is calculated according to the blade theoretical reaching time and a rotating blade radius; The initial blade vibration displacement is calculated according to the blade theoretical reaching time and the blade rotation speed.
[0008] In an embodiment, the initial blade vibration displacement is subjected to trend item removal to obtain the target blade vibration displacement, comprising: A target trend item is obtained by fitting the initial blade vibration displacement with a trend item; The initial blade vibration displacement is subjected to trend item removal according to the target trend item to obtain the target blade vibration displacement.
[0009] In an embodiment, the time-frequency analysis of the target blade vibration displacement is segmented based on adaptive overwavelet transform to obtain a two-dimensional relationship between aliasing frequency and rotation frequency, comprising: An overwavelet base function is constructed; the target blade vibration displacement is divided into data processing units based on a preset fixed rotation frequency step; A center frequency corresponding to each of the data processing units is calculated, and an order corresponding to each of the data processing units is calculated according to each of the center frequencies, a preset order range and a preset center frequency range; The target blade vibration displacement corresponding to each data processing unit is convolved based on the superwavelet basis function according to the center frequency and the order, to obtain a time-frequency distribution of the target blade vibration displacement at the center frequency; The time-frequency distribution is converted based on a mapping relationship between time and rotational frequency, to obtain a two-dimensional relationship between the aliasing frequency and the rotational frequency.
[0010] In an embodiment, the two-dimensional relationship between the aliasing frequency and the rotational frequency is used to identify a resonance interval based on an undersampling principle, including: A frequency threshold is determined based on the undersampling principle; Based on the two-dimensional relationship between the aliasing frequency and the rotational frequency, an aliasing frequency interval in which the aliasing frequency is less than the frequency threshold is determined as a target interval; A rotational frequency interval corresponding to the target interval is determined as the resonance interval.
[0011] In a second aspect, the present application provides a turbine blade resonance interval identification device, including: An acquisition module is configured to acquire a blade tip pulse signal of a turbine; A processing module is configured to pre-process the blade tip pulse signal and calculate a target blade vibration displacement based on the pre-processed blade tip pulse signal; An analysis module is configured to perform segmented time-frequency analysis on the target blade vibration displacement based on adaptive superwavelet transform, to obtain a two-dimensional relationship between an aliasing frequency and a rotational frequency; An identification module is configured to identify a resonance interval based on an undersampling principle according to the two-dimensional relationship between the aliasing frequency and the rotational frequency.
[0012] In a third aspect, the present application provides a computer device including a memory and a processor, the memory storing a computer program, and the computer program being executed by the processor to implement the turbine blade resonance interval identification method according to the first aspect.
[0013] In a fourth aspect, the present application provides a computer readable storage medium storing a computer program, and the computer program being executed by a processor to implement the turbine blade resonance interval identification method according to the first aspect.
[0014] The turbine blade resonance interval identification method, device, equipment and storage medium disclosed by the application acquire a blade tip pulse signal of a turbine; the blade tip pulse signal is preprocessed, and target blade vibration displacement is calculated according to the preprocessed blade tip pulse signal; the target blade vibration displacement is subjected to segmented time-frequency analysis based on adaptive hyperwavelet transform, to obtain a two-dimensional relationship between aliasing frequency and rotational frequency; and the resonance interval is identified according to the two-dimensional relationship between aliasing frequency and rotational frequency based on the principle of undersampling. In this way, the blade tip pulse signal is preprocessed to obtain a clean blade tip pulse signal, target vibration displacement is acquired, adaptive hyperwavelet transform is further used to analyze the pulse signal in time-frequency domain, and the principle of undersampling is combined to accurately identify vibration amplitude and frequency in the scene of weak resonance characteristics and strong noise, and to capture resonance points that are easily missed by traditional blade tip vibration monitoring methods. BRIEF DESCRIPTION OF DRAWINGS
[0015] In order to more clearly illustrate the technical solutions of the present application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as limiting the scope of protection of the present application. In the various drawings, similar components are denoted by similar reference numerals.
[0016] Figure 1 A flowchart of the turbine blade resonance interval identification method proposed in the present embodiment is shown; Figure 2 Another flowchart of the turbine blade resonance interval identification method proposed in the present embodiment is shown; Figure 3 Still another flowchart of the turbine blade resonance interval identification method proposed in the present embodiment is shown; Figure 4 Yet another flowchart of the turbine blade resonance interval identification method proposed in the present embodiment is shown; Figure 5 Still another flowchart of the turbine blade resonance interval identification method proposed in the present embodiment is shown; Figure 6 A rotational speed-vibration displacement monitoring situation diagram proposed in the present embodiment is shown; Figure 7 A blade tip pulse signal diagram before and after preprocessing proposed in the present embodiment is shown; Figure 8 A resonance interval identification result diagram proposed in the present embodiment is shown; Figure 9 A structure diagram of the turbine blade resonance interval identification device proposed in the present embodiment is shown.
[0017] Explanation of the drawing: 900 - turbine blade resonance interval identification device; 901 - acquisition module; 902 - processing module; 903 - analysis module; 904 - identification module. DETAILED DESCRIPTION
[0018] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application.
[0019] The components of the embodiments of the present application generally described and illustrated in the accompanying drawings can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but only represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.
[0020] Hereinafter, the terms "include", "have", and their conjugates used in various embodiments of the present application are only intended to denote a certain characteristic, number, step, operation, element, component, or combination of the foregoing, and should not be construed as excluding the presence or addition of one or more other characteristics, numbers, steps, operations, elements, components, or combinations thereof.
[0021] In addition, the terms "first", "second", "third", and the like are only used to distinguish descriptions, and cannot be understood as indicating or implying relative importance.
[0022] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which various embodiments of the present application belong. The terms (such as those defined in commonly used dictionaries) will be interpreted as having a meaning that is the same as the contextual meaning in the relevant technical field and will not be interpreted as having an idealized or overly formal meaning unless clearly defined in various embodiments of the present application.
[0023] Embodiment 1 The embodiments of the present disclosure provide a turbine blade resonance interval identification method for adaptive wavelet transform processing of BTT signals, which accurately identifies the resonance event of the blade by analyzing the aliasing frequency in the undersampled signal.
[0024] Please refer to Figure 1 The turbine blade resonance interval identification method includes steps S101-S104, which will be described in detail below.
[0025] Step S101, obtaining a blade tip pulse signal of a turbomachinery.
[0026] In the embodiment, the blade tip pulse signal of the turbomachinery is obtained by a data acquisition device, and the blade tip pulse signal is a time sequence of blades passing through the data acquisition device . The data acquisition device can be a sensor.
[0027] Step S102, preprocessing the blade tip pulse signal, and calculating a target blade vibration displacement according to the preprocessed blade tip pulse signal.
[0028] In the embodiment, the blade tip pulse signal is preprocessed to obtain a preprocessed blade tip pulse signal after removing noise and linear interpolation, and a clean and accurate target blade vibration displacement is further calculated according to the preprocessed blade tip pulse signal.
[0029] Please refer to Figure 2 , in a specific embodiment, step S102 includes steps S201-S205, which will be described in detail below.
[0030] Step S201, removing a narrow pulse noise signal in the blade tip pulse signal based on a preset sliding window length to obtain a candidate pulse signal in each sliding window.
[0031] In the embodiment, the blade tip pulse signal is set with a sliding window with a length of W, and first, an IIR filter is used to remove the narrow pulse noise signal in the blade tip pulse signal in the sliding window, so that the real blade passing pulse is retained while the interference is filtered out, and a cleaner candidate pulse signal is obtained for subsequent analysis.
[0032] Step S202, calculating a time interval of adjacent candidate pulse signals in each sliding window.
[0033] In the embodiment, the time interval of adjacent candidate pulse signals in each sliding window is calculated , wherein , wherein is the i-th candidate pulse signal, is the time interval corresponding to the i-th candidate pulse signal.
[0034] Step S203, for each sliding window, calculating a time interval average value corresponding to each sliding window according to the time interval.
[0035] In the embodiment, for each sliding window, a time interval average value corresponding to each sliding window is calculated according to all time intervals in each sliding window .
[0036] Step S204, determining the interval reference threshold corresponding to each sliding window according to the preset proportion and the average value of each time interval.
[0037] In this embodiment, the product of the preset proportion and the average value of each time interval is calculated respectively as the interval reference threshold corresponding to each sliding window. Taking the preset proportion as 0.75 for example, the interval reference threshold can be 0.75 .
[0038] Step S205, for each sliding window, performing linear interpolation on the candidate pulse signals according to the time interval and the interval reference threshold to obtain the preprocessed blade tip pulse signal.
[0039] In this embodiment, for each sliding window, the candidate pulse signals with time intervals less than the interval reference threshold are determined as data missing, linear interpolation is performed at the data missing place, and adjacent correct time interval data are interpolated to obtain the preprocessed blade tip pulse signal, so as to dynamically utilize the sampling rule of the data itself to avoid misjudgment caused by artificial setting of a fixed threshold, and at the same time, the preset proportion and the time interval average value are utilized to determine the interval reference threshold, so as to avoid misjudgment of short delay as missing and reduce false positives.
[0040] Please refer to Figure 3 In a specific embodiment, step S102 includes steps S301-S302, which are described in detail as follows.
[0041] Step S301, calculating an initial blade vibration displacement according to the preprocessed blade tip pulse signal.
[0042] In this embodiment, the preprocessed blade tip pulse signal is quantified as an initial blade vibration displacement reflecting a real-time vibration amplitude.
[0043] In a specific embodiment, step S301 includes: based on a linear relationship between a blade theoretical reaching time and an accumulated radian corresponding to the blade, performing least square linear fitting according to the preprocessed blade tip pulse signal to obtain the blade theoretical reaching time; calculating a blade rotation speed according to the blade theoretical reaching time and a rotating blade radius; and calculating the initial blade vibration displacement according to the blade theoretical reaching time and the blade rotation speed.
[0044] In this embodiment, the rotation speed in a circle is considered to be constant, and the blade theoretical reaching time and the accumulated radian corresponding to the blade satisfy a linear relationship: , assuming that the number of blades in a circle is , the preprocessed blade tip pulse signal (actual reaching time data) of blades and the corresponding accumulated radian The least square linear fitting is performed to obtain , , , wherein represents a theoretical arrival time of the blade No. k in the i th cycle, represents an accumulated arc of the blade No. k in the i th cycle, represents a slope of the fitting line in the i th cycle, is an intercept of the fitting line in the i th cycle; the blade rotating speed in the i th cycle can be represented as , wherein , , is a rotating blade radius, is a rotating period of the i th cycle; the initial vibration displacement of the blade is calculated as , wherein represents the vibration displacement of the blade No. k in the i th cycle. , , , wherein represents the vibration displacement of the blade No. k in the i th cycle. Step S302, trend item removal is performed on the initial blade vibration displacement to obtain the target blade vibration displacement.
[0045] In the embodiment, the trend item removal is performed on the initial blade vibration displacement to obtain a cleaner target vibration displacement, so as to avoid inaccurate resonance identification caused by low-frequency interference.
[0046] In a specific embodiment, step S302 includes: trend item fitting is performed on the initial blade vibration displacement to obtain a target trend item; and trend item removal is performed on the initial blade vibration displacement according to the target trend item to obtain the target blade vibration displacement.
[0047] In the embodiment, all target vibration displacements of a single blade are recorded as ; a trend item curve fitting is performed on the initial vibration displacement by using an adaptive window SG filter, and the fitted target trend item is
[0048] ; and the target blade vibration displacement, which needs to be positioned for a resonance event, is obtained by further subtracting the target trend item from the initial blade vibration displacement . Step S103, time-frequency analysis is performed on the target blade vibration displacement by using adaptive hyperwavelet transform to obtain a two-dimensional relationship between the aliasing frequency and the rotating frequency.
[0049] Step S103, time-frequency analysis is performed on the target blade vibration displacement by using adaptive hyperwavelet transform to obtain a two-dimensional relationship between the aliasing frequency and the rotating frequency.
[0050] In the embodiment, the BTT signal usually has an undersampling problem, resulting in aliasing frequencies in the signal spectrum, so that the two-dimensional relationship between the aliasing frequency and the rotation frequency of each segment is obtained by performing segmented time-frequency analysis on the target blade vibration displacement based on adaptive wavelet transform.
[0051] See Figure 4 In an embodiment, step S103 includes steps S1031-S1034, which are described in detail below.
[0052] In step S1031, a wavelet basis function is constructed, and the target blade vibration displacement is divided into data processing units based on a preset fixed rotation frequency step.
[0053] In the embodiment, the wavelet basis function is wherein, , is the center frequency of the wavelet, is an extension parameter related to the period number and the center frequency, is a constant related to the wavelet period number, and a commonly used value is 5. A preset order range and a center frequency range are preset in advance, and the set contains wavelets with different period numbers .
[0054] Meanwhile, for the target vibration displacement , a data processing unit is set according to a preset fixed rotation frequency step .
[0055] In step S1032, the center frequency corresponding to each data processing unit is calculated, and the order corresponding to each data processing unit is calculated according to the center frequency, the preset order range, and the preset center frequency range.
[0056] In the embodiment, and are the minimum and maximum rotation frequencies in the data processing unit, in Hz. According to the Nyquist theorem, the frequency range that can be analyzed by the spectrum is , so the center frequency of the data processing unit is set to ; and the order of the data processing unit is .
[0057] In step S1033, for each processing unit, the target blade vibration displacement corresponding to the data processing unit is convolved based on the wavelet basis function according to the center frequency and the order, to obtain the time-frequency distribution of the target blade vibration displacement at the center frequency.
[0058] In the embodiment, the target vibration displacement data in the data processing unit is convolved with the corresponding wavelet basis function, and the target vibration displacement in the window is obtained by taking the geometric mean . The time-frequency distribution when the center frequency is .
[0059] Step S1034, converting the time-frequency distribution based on the mapping relationship between time and rotational frequency, to obtain a two-dimensional relationship between the aliasing frequency and the rotational frequency.
[0060] In the embodiment, the time t and the rotational frequency have a one-to-one mapping relationship, and the data of a plurality of center frequencies are respectively subjected to hyperwavelet transform ASLT, and the time-frequency distribution corresponding to each center frequency obtained after combination can be mapped into a two-dimensional relationship between the aliasing frequency and the rotational frequency : .
[0061] Step S104, identifying a resonance interval based on the two-dimensional relationship between the aliasing frequency and the rotational frequency according to the undersampling principle.
[0062] In the embodiment, the resonance interval is identified based on the two-dimensional relationship between the aliasing frequency and the rotational frequency according to the undersampling principle, so that the vibration amplitude and frequency can be accurately identified in the scene of weak resonance characteristics and strong noise, and the resonance point that is easy to be missed by the traditional blade tip vibration monitoring method can be captured.
[0063] Please refer to Figure 5 , in a specific embodiment, step S104 includes steps S1041-S1043, which will be described in detail below.
[0064] Step S1041, determining a frequency threshold based on the undersampling principle.
[0065] In the embodiment, the blade vibration frequency is set to Hz, and the decomposed blade vibration frequency is , where is an integer, is a remainder, and satisfies .
[0066] According to the Nyquist theorem, when undersampling, will be aliased to 0, and the maximum resolution of the frequency spectrum is , so the aliasing frequency can be calculated as .
[0067] When the sampling frequency of the blade vibration signal is in the under-sampling interval , the aliasing frequency can be calculated: .
[0068] When the blade resonates, the blade vibration frequency is equal to an integer multiple of the rotation frequency, i.e. . Therefore, the frequency threshold value can be set as the value "0", but the actual signal cannot be completely "0" frequency, but a frequency close to "0". Therefore, the frequency threshold value can also be a value close to "0".
[0069] In step S1042, based on the two-dimensional relationship between the aliasing frequency and the rotation frequency, the aliasing frequency interval in which the aliasing frequency is less than the frequency threshold value is determined as the target interval.
[0070] In this embodiment, in the two-dimensional relationship between the aliasing frequency and the rotation frequency, the aliasing frequency interval in which the aliasing frequency is less than the frequency threshold value is determined as the target interval.
[0071] In step S1043, the rotation frequency interval corresponding to the target interval is determined as the resonance interval.
[0072] In this embodiment, the rotation frequency interval corresponding to the target interval is determined as the resonance interval, so as to realize positioning of the resonance interval.
[0073] For example, an exemplary engine has 29 blades in the first stage with a radius of 104460 microns, and 32 blades in the second stage with a radius of 99780 microns. During monitoring, the engine is gradually accelerated to above 32000 RPM, then kept at a constant speed for a period of time, and then artificially induced to surge before gradually decelerating to stop. The BTT is disturbed during monitoring of the two stages of blades, resulting in failure of the monitoring. The rotation speed-vibration displacement monitoring of the first stage of blades is shown in Figure 6 . The first stage of blades is in a normal monitoring state at 0 to 29000 RPM, and the BTT sensor is affected before entering the constant speed state, and the entire vibration displacement-rotation speed curve becomes distorted. After decelerating to 24000 RPM, the monitoring state returns to normal. The second stage of blades is in a normal monitoring state at 0 to 26000 RPM, and is also affected when entering the 32000 RPM constant speed stage, and the monitoring curve is abnormal. However, after decelerating to 26000 RPM, the overall monitoring state returns to normal. In the subsequent analysis, it is found that the researchers need to open the air guide valve in the engine in advance to prepare for the subsequent experiment, and the sudden airflow causes the monitoring environment to be more polluted, so that the signal quality of the laser sensor is poor, and finally the BTT monitoring fails.
[0074] Taking the first stage impeller as an example, the collected blade tip pulse signals are set with a sliding window length of W, the first stage impeller has 29 blades, so W is set to 29, the IIR filter is used for the 29 blade tip pulse signals to remove possible narrow pulse signals, after filtering, if the data points are less than 29, the window is moved to make the data length 29, and the IIR digital filtering is performed again until the data in the window after filtering is 29.
[0075] Further, the adjacent time intervals of the 29 candidate pulse signals in the sliding window are calculated , and the mean value of is solved, when the candidate pulse signal in is less than 0.75 times the mean value, it is determined that the data is missing; for example is determined to be data missing, but is a normal interval, so it is necessary to interpolate data, and the interpolated data is . The results before and after the blade tip pulse signal preprocessing are shown in Figure 7 .
[0076] Further, the initial vibration displacement is calculated according to the preprocessed blade tip pulse signal, the initial vibration displacement is subjected to SG filtering curve fitting with a window length of 101 as a target trend item curve, and the target vibration displacement data for super wavelet transformation is obtained by subtracting the target trend item curve from the initial vibration displacement.
[0077] Further, the wavelet basis function order range is preset to [3, 12], and the center frequency range is [500, 2500], and a plurality of wavelet sets with different periods are generated according to these parameters; because different blades have different natural frequencies, before setting these two values, the approximate resonance center range is determined by a one-time speed up and speed down, generally, 0.5 times to 10 times of the nearest resonance center frequency at the maximum speed of the turbomachinery can be set.
[0078] Further, a fixed rotational frequency step of 2 Hz is taken as a unit of data processing, the target vibration displacement signal is divided into a plurality of intervals, super wavelet transformation is performed on these intervals, and they are spliced in the rotational frequency domain, the resonance event threshold (frequency threshold) is set to 2.5 Hz, taking blade No. 1 of the first stage impeller as an example, the resonance interval positioning is performed, and the results are shown in Figure 8 . The blade vibration natural frequency identification is performed on the identified 6 resonance intervals, and the results are shown in the following table.
[0079] It can be seen that the blade strain gauge output frequency is 1817.76 Hz, the maximum deviation of the blade natural frequency calculated by the 6 resonance intervals identified by ALST is 12.74 Hz compared with the calculation result of the strain gauge, and the maximum relative error is 0.7009%, which verifies that the occurrence of the blade resonance event is accurately identified in this embodiment.
[0080] The turbine blade resonance interval identification method provided in this embodiment acquires a blade tip pulse signal of a turbine; pre-processes the blade tip pulse signal, and calculates a target blade vibration displacement according to the pre-processed blade tip pulse signal; performs segmented time-frequency analysis on the target blade vibration displacement based on adaptive hyperwavelet transform to obtain a two-dimensional relationship between aliasing frequency and rotational frequency; and identifies a resonance interval according to the two-dimensional relationship between aliasing frequency and rotational frequency based on the principle of under-sampling. In this way, the blade tip pulse signal is pre-processed to obtain a clean blade tip pulse signal, and then the target vibration displacement is acquired, further, the adaptive hyperwavelet transform is used to perform time-frequency analysis on the pulse signal, and the principle of under-sampling is combined to accurately identify the vibration amplitude and frequency in the scene of weak resonance characteristics and strong noise, and capture the resonance points that are easily missed by the traditional blade tip vibration monitoring method.
[0081] Embodiment 2 In addition, the disclosure embodiment provides a turbine blade resonance interval identification device 900, please see Figure 9 , comprising: The acquisition module 901 is configured to acquire a blade tip pulse signal of a turbine; The processing module 902 is configured to pre-process the blade tip pulse signal, and calculate a target blade vibration displacement according to the pre-processed blade tip pulse signal; The analysis module 903 is configured to perform segmented time-frequency analysis on the target blade vibration displacement based on adaptive hyperwavelet transform to obtain a two-dimensional relationship between aliasing frequency and rotational frequency; The identification module 904 is configured to identify a resonance interval according to the two-dimensional relationship between aliasing frequency and rotational frequency based on the principle of under-sampling.
[0082] Optionally, the processing module 902 is further configured to remove narrow pulse noise signals in the blade tip pulse signal based on a preset sliding window length to obtain candidate pulse signals in each sliding window; calculate the time interval of adjacent candidate pulse signals in each sliding window; for each sliding window, calculate the time interval average corresponding to each sliding window according to the time interval; determine the interval reference threshold corresponding to each sliding window according to a preset ratio and each time interval average; and for each sliding window, perform linear interpolation on the candidate pulse signals according to the time interval and the interval reference threshold to obtain the pre-processed blade tip pulse signal.
[0083] Optionally, the processing module 902 is further configured to calculate an initial blade vibration displacement according to the preprocessed blade tip pulse signal; and remove a trend item from the initial blade vibration displacement to obtain the target blade vibration displacement.
[0084] Optionally, the processing module 902 is further configured to perform least square linear fitting according to the preprocessed blade tip pulse signal based on a linear relationship between a blade theoretical reaching time and an accumulated radian of the blade to obtain the blade theoretical reaching time; calculate a blade rotating speed according to the blade theoretical reaching time and a rotating blade radius; and calculate the initial blade vibration displacement according to the blade theoretical reaching time and the blade rotating speed.
[0085] Optionally, the processing module 902 is further configured to fit a target trend item from the initial blade vibration displacement; and remove the target trend item from the initial blade vibration displacement to obtain the target blade vibration displacement.
[0086] Optionally, the analysis module 903 is further configured to construct an overwave base function; divide the target blade vibration displacement into data processing units based on a preset fixed rotating frequency step; calculate a center frequency corresponding to each data processing unit, and calculate an order corresponding to each data processing unit according to each center frequency, a preset order range and a preset center frequency range; and for each data processing unit, perform convolution on a target blade vibration displacement corresponding to the data processing unit based on the overwave base function, the center frequency and the order to obtain a time-frequency distribution of the target blade vibration displacement at the center frequency; and convert the time-frequency distribution based on a mapping relationship between time and rotating frequency to obtain a two-dimensional relationship between the aliasing frequency and the rotating frequency.
[0087] Optionally, the identification module 904 is further configured to determine a frequency threshold based on the undersampling principle; determine an aliasing frequency interval in which an aliasing frequency is less than the frequency threshold as a target interval based on the two-dimensional relationship between the aliasing frequency and the rotating frequency; and determine a rotating frequency interval corresponding to the target interval as the resonance interval.
[0088] The apparatus provided in the embodiments of the present disclosure can perform the steps of the turbine blade resonance interval identification method provided in Embodiment 1, and thus will not be described again.
[0089] The turbine blade resonance interval recognition device provided in the embodiment obtains a blade tip pulse signal of a turbine; pre-processes the blade tip pulse signal, and calculates a target blade vibration displacement according to the pre-processed blade tip pulse signal; performs segmented time-frequency analysis on the target blade vibration displacement based on adaptive hyperwavelet transform, to obtain a two-dimensional relationship between aliasing frequency and rotational frequency; and identifies a resonance interval according to the two-dimensional relationship between aliasing frequency and rotational frequency based on an under-sampling principle. In this way, the blade tip pulse signal is pre-processed to obtain a clean blade tip pulse signal, and then the target vibration displacement is obtained. Further, adaptive hyperwavelet transform is used to perform time-frequency analysis on the pulse signal, and the under-sampling principle is combined to accurately identify the vibration amplitude and frequency in the scenario of weak resonance characteristics and strong noise, and to capture the resonance points that are easily missed by traditional blade tip vibration monitoring methods.
[0090] Embodiment 3 In addition, the embodiment of the present disclosure provides a computer device comprising a memory and a processor, the memory storing a computer program, and the computer program is executed by the processor to implement the turbine blade resonance interval recognition method of the embodiment 1.
[0091] The device provided in the embodiment of the present disclosure can execute the steps of the turbine blade resonance interval recognition method provided in the embodiment 1, and details are not repeated.
[0092] Embodiment 4 The embodiment of the present disclosure provides a computer readable storage medium storing a computer program, and the computer program is executed by a processor to implement the turbine blade resonance interval recognition method of the embodiment 1.
[0093] In the embodiment, the computer readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.
[0094] The computer readable storage medium provided in the embodiment can implement the turbine blade resonance interval recognition method provided in the embodiment 1, and details are not repeated.
[0095] In all examples shown and described herein, any specific values should be interpreted as merely exemplary and not as a limitation, and thus, other examples of the example embodiments can have different values.
[0096] It should be noted that: similar reference numerals and letters represent similar items in the following drawings, and thus, once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings.
[0097] The above embodiments only express several implementation manners of the present application, which are described in a more specific and detailed manner, but cannot be understood as a limitation on the scope of the present application. It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, which all belong to the protection scope of the present application.
Claims
1. A method for identifying a resonance region of a turbine blade, characterized by, The method comprises: obtaining a blade tip pulse signal of a turbine; preprocessing the blade tip pulse signal, and calculating a target blade vibration displacement according to the preprocessed blade tip pulse signal; performing segmented time-frequency analysis on the target blade vibration displacement based on adaptive superwavelet transform to obtain a two-dimensional relationship between aliasing frequency and rotational frequency; identifying a resonance interval according to the two-dimensional relationship between aliasing frequency and rotational frequency based on undersampling principle.
2. The turbomachinery blade resonance region identification method of claim 1, wherein, The preprocessing of the blade tip pulse signal comprises: removing narrow pulse noise signals in the blade tip pulse signal based on a preset sliding window length to obtain candidate pulse signals in each sliding window; calculating time intervals of adjacent candidate pulse signals in each sliding window; for each sliding window, calculating a time interval average value corresponding to the sliding window according to the time intervals; determining a reference interval threshold corresponding to each sliding window according to a preset ratio and the time interval average value; for each sliding window, performing linear interpolation on the candidate pulse signals according to the time intervals and the reference interval threshold to obtain the preprocessed blade tip pulse signal.
3. The turbomachinery blade resonance region identification method of claim 1, wherein The calculation of the target blade vibration displacement according to the preprocessed blade tip pulse signal comprises: calculating an initial blade vibration displacement according to the preprocessed blade tip pulse signal; performing trend item removal on the initial blade vibration displacement to obtain the target blade vibration displacement.
4. The turbomachinery blade resonance region identification method according to claim 3, characterized by, The calculation of the initial blade vibration displacement according to the preprocessed blade tip pulse signal comprises: based on a linear relationship between a blade theoretical reaching time and an arc corresponding to the blade accumulation, performing least square linear fitting on the preprocessed blade tip pulse signal to obtain the blade theoretical reaching time; calculating a blade rotation speed according to the blade theoretical reaching time and a rotating blade radius; calculating the initial blade vibration displacement according to the blade theoretical reaching time and the blade rotation speed.
5. The turbomachinery blade resonance region identification method of claim 3, wherein, The trend item removal on the initial blade vibration displacement to obtain the target blade vibration displacement comprises: performing trend item fitting on the initial blade vibration displacement to obtain a target trend item; performing trend item removal on the initial blade vibration displacement according to the target trend item to obtain the target blade vibration displacement.
6. The turbomachinery blade resonance region identification method of claim 1, wherein, The segmented time-frequency analysis on the target blade vibration displacement based on adaptive superwavelet transform to obtain the two-dimensional relationship between aliasing frequency and rotational frequency comprises: constructing a superwavelet base function; dividing the target blade vibration displacement into data processing units based on a preset fixed rotational frequency step; calculating a center frequency corresponding to each data processing unit, and calculating an order corresponding to each data processing unit according to each center frequency, a preset order range and a preset center frequency range; for each processing unit, performing convolution on a target blade vibration displacement corresponding to the data processing unit based on the superwavelet base function, the center frequency and the order to obtain a time-frequency distribution of the target blade vibration displacement at the center frequency; performing conversion on the time-frequency distribution based on a mapping relationship between time and rotational frequency to obtain the two-dimensional relationship between aliasing frequency and rotational frequency.
7. The turbomachinery blade resonance region identification method of claim 6, wherein, The resonance interval is identified based on the two-dimensional relationship between the aliasing frequency and the rotation frequency according to the undersampling principle, and the resonance interval identification method comprises the following steps: A frequency threshold is determined based on the undersampling principle; Based on the two-dimensional relationship between the aliasing frequency and the rotation frequency, it is determined that the aliasing frequency interval in which the aliasing frequency is less than the frequency threshold is a target interval; The rotation frequency interval corresponding to the target interval is determined as the resonance interval.
8. A device for identifying the resonance range of a turbine blade, characterized in that, The method comprises the following steps: An acquisition module is configured to acquire a blade tip pulse signal of a turbomachine; A processing module is configured to pre-process the blade tip pulse signal and calculate a target blade vibration displacement based on the pre-processed blade tip pulse signal; An analysis module is configured to perform segmented time-frequency analysis on the target blade vibration displacement based on adaptive hyperwavelet transform to obtain a two-dimensional relationship between an aliasing frequency and a rotation frequency; An identification module is configured to identify a resonance interval based on the two-dimensional relationship between the aliasing frequency and the rotation frequency according to the undersampling principle.
9. A computer device, comprising: The turbomachine blade resonance interval identification method comprises the following steps:
10. A computer-readable storage medium, characterized in that, The turbomachine blade resonance interval identification method comprises the following steps: The computer program is stored in the memory and is executed by the processor to implement the turbomachine blade resonance interval identification method according to any one of claims 1 to 7. The computer program is stored in the memory and is executed by the processor to implement the turbomachine blade resonance interval identification method according to any one of claims 1 to 7.