Shaft vibration signal noise reduction method and device, storage medium and electronic equipment

Abstract

The invention provides a shaft vibration signal noise reduction method and device, a storage medium and electronic equipment, and is applied to the technical field of signal noise reduction, and the method comprises the steps: decomposing an amplitude signal to be subjected to noise reduction, and obtaining a plurality of intrinsic mode functions and residual errors; processing each intrinsic mode function to obtain permutation entropy, and determining a signal dominant type of the intrinsic mode function by using the permutation entropy; and performing mathematical form filtering processing on each intrinsic mode function of which the signal dominant type is useful signal dominant, and then performing signal reconstruction by using the filtered intrinsic mode function to obtain a noise-reduced shaft vibration signal. According to the scheme, useful signals and noise can be separated from noise-containing shaft vibration signals in a high-noise environment, further noise reduction processing is performed on the useful signals, the shaft vibration signals after noise reduction are provided, and reliable data support is provided for fault diagnosis and early warning of a power station.

Description

Technical Field

[0001] The present invention relates to the technical field of signal noise reduction, and in particular to a method and device for reducing the noise of a shaft vibration signal, a storage medium and an electronic device. Background Art

[0002] The shaft vibration signal in a power station is one of the key parameters that reflects the operating status of the shaft. However, due to factors such as mechanical friction and wear and electromagnetic interference, the collected shaft vibration signal often contains a lot of background noise, which makes it difficult to accurately reflect the actual operating status of the shaft, increasing the difficulty of subsequent data processing and signal analysis.

[0003] Noise reduction of shaft vibration signals can provide reliable data support for power station fault diagnosis and early warning. However, traditional noise reduction methods are difficult to accurately separate useful signals and noise in high-noise environments, which affects the accuracy of subsequent power station fault diagnosis and early warning. Summary of the invention

[0004] In view of this, an embodiment of the present application provides a shaft vibration signal noise reduction method and device, a storage medium and an electronic device. By applying the scheme provided in the present application, the noisy shaft vibration signal can be separated into useful signals and noise, and the useful signal can be subjected to noise reduction processing to obtain a noise-reduced shaft vibration signal, thereby providing reliable data support for the power station.

[0005] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0006] A method for reducing noise of shaft vibration signals, comprising:

[0007] Decomposing the shaft vibration signal to be de-noised to obtain a function set and a residual, wherein the function set includes a plurality of intrinsic mode functions;

[0008] Processing each intrinsic mode function in the function set to obtain permutation entropy;

[0009] Determining a signal-dominant type of each intrinsic mode function in the function set according to the permutation entropy, wherein the signal-dominant type is one of useful signal-dominant and noise-dominant;

[0010] Mathematical morphology filtering is performed on each intrinsic mode function whose dominant signal type is dominated by useful signals, and signal reconstruction is performed on each intrinsic mode function after filtering to obtain a noise-reduced shaft vibration signal.

[0011] In the above method, optionally, the shaft vibration signal to be de-noised is decomposed to obtain a function set and a residual, including:

[0012] Determining the shaft vibration signal to be de-noised as a target signal;

[0013] Adding a preset adaptive Gaussian white noise to the target signal to construct a plurality of signals to be decomposed;

[0014] Decomposing each of the signals to be decomposed to obtain an intrinsic mode function, and storing the intrinsic mode function in a function set;

[0015] Determine a residual using the intrinsic mode function and the target signal, and update the number of iterations;

[0016] Based on the residual and the updated number of iterations, determining whether a preset stop condition is met;

[0017] When the stop condition is not met, the residual is used as a new target signal, and the step of adding a preset adaptive Gaussian white noise to the target signal to construct multiple signals to be decomposed is returned to the step, until the stop iteration condition is met, and the function set and the residual are output.

[0018] In the above method, optionally, the decomposing each of the signals to be decomposed to obtain an intrinsic mode function includes:

[0019] Performing empirical mode decomposition on each of the signals to be decomposed to obtain each signal component of each of the signals to be decomposed;

[0020] The first signal component of each of the signals to be decomposed is averaged, and the obtained average is used as the intrinsic mode function.

[0021] In the above method, optionally, judging whether a preset stop condition is satisfied based on the residual and the updated number of iterations includes:

[0022] When the residual is a monotonic function and / or the number of iterations reaches a preset maximum number of iterations, it is determined that the stopping condition is met; otherwise, it is determined that the stopping condition is not met.

[0023] In the above method, optionally, the processing of each intrinsic mode function in the function set to obtain the permutation entropy includes:

[0024] Reconstructing each of the intrinsic mode functions in phase space to obtain a phase space matrix;

[0025] Sorting each row in the phase space matrix to obtain a column index symbol sequence;

[0026] The column index symbol sequence is processed to obtain permutation entropy.

[0027] In the above method, optionally, the mathematical morphological filtering is performed on each intrinsic mode function whose dominant signal type is dominated by the useful signal, including:

[0028] Determine mathematical morphology operators constructed using the mean of opening and closing operations;

[0029] Based on the mathematical morphology operator, mathematical morphology filtering is performed on each intrinsic mode function whose dominant signal type is dominated by useful signals to obtain the intrinsic mode function after filtering.

[0030] In the above method, optionally, the signal reconstruction of each intrinsic mode function after filtering to obtain a noise-reduced shaft vibration signal includes:

[0031] The eigenmode functions after filtering are added together to obtain a noise-reduced shaft vibration signal.

[0032] A shaft vibration signal noise reduction device, comprising:

[0033] A decomposition unit, used for decomposing the shaft vibration signal to be de-noised to obtain a function set and a residual, wherein the function set includes a plurality of intrinsic mode functions;

[0034] A processing unit, used for processing each intrinsic mode function in the function set to obtain a permutation entropy;

[0035] A determination unit, configured to determine a signal-dominant type of each intrinsic mode function in the function set according to the permutation entropy, wherein the signal-dominant type is one of useful signal-dominant and noise-dominant;

[0036] The filtering unit is used to perform mathematical morphological filtering on each intrinsic mode function whose dominant signal type is dominated by the useful signal, and reconstruct the signal of each intrinsic mode function after the filtering process to obtain a noise-reduced shaft vibration signal.

[0037] In the above device, optionally, the decomposition unit performs a process of decomposing the shaft vibration signal to be de-noised to obtain a function set and a residual, including:

[0038] Determining the shaft vibration signal to be de-noised as a target signal;

[0039] Adding a preset adaptive Gaussian white noise to the target signal to construct a plurality of signals to be decomposed;

[0040] Decomposing each of the signals to be decomposed to obtain an intrinsic mode function, and storing the intrinsic mode function in a function set;

[0041] Determine a residual using the intrinsic mode function and the target signal, and update the number of iterations;

[0042] Based on the residual and the updated number of iterations, determining whether a preset stop condition is met;

[0043] When the stop condition is not met, the residual is used as a new target signal, and the step of adding a preset adaptive Gaussian white noise to the target signal to construct multiple signals to be decomposed is returned to the step, until the stop iteration condition is met, and the function set and the residual are output.

[0044] In the above device, optionally, the decomposition unit performs a process of decomposing each of the signals to be decomposed to obtain an intrinsic mode function, including:

[0045] Performing empirical mode decomposition on each of the signals to be decomposed to obtain each signal component of each of the signals to be decomposed;

[0046] The first signal component of each of the signals to be decomposed is averaged, and the obtained average is used as the intrinsic mode function.

[0047] In the above device, optionally, the decomposition unit performs a process of judging whether a preset stop condition is satisfied based on the residual and the updated number of iterations, including:

[0048] When the residual is a monotonic function and / or the number of iterations reaches a preset maximum number of iterations, it is determined that the stopping condition is met; otherwise, it is determined that the stopping condition is not met.

[0049] In the above device, optionally, the processing unit performs a process of processing each intrinsic mode function in the function set to obtain a permutation entropy, comprising:

[0050] Reconstructing each of the intrinsic mode functions in phase space to obtain a phase space matrix;

[0051] Sorting each row in the phase space matrix to obtain a column index symbol sequence;

[0052] The column index symbol sequence is processed to obtain permutation entropy.

[0053] In the above device, optionally, the filtering unit performs a process of performing mathematical morphological filtering on each intrinsic mode function whose dominant signal type is dominated by the useful signal, comprising:

[0054] Determine mathematical morphology operators constructed using the mean of opening and closing operations;

[0055] Based on the mathematical morphology operator, mathematical morphology filtering is performed on each intrinsic mode function whose dominant signal type is dominated by useful signals to obtain the intrinsic mode function after filtering.

[0056] In the above device, optionally, the filtering unit performs a process of reconstructing the signals of the filtered intrinsic mode functions to obtain a noise-reduced shaft vibration signal, including:

[0057] The eigenmode functions after filtering are added together to obtain a noise-reduced shaft vibration signal.

[0058] A storage medium includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the shaft vibration signal noise reduction method as described above.

[0059] An electronic device includes a memory and one or more instructions, wherein the one or more instructions are stored in the memory and are configured to be executed by one or more processors to perform the shaft vibration signal noise reduction method as described above.

[0060] Compared with the prior art, this application has the following advantages:

[0061] The present application provides a shaft vibration signal denoising method and device, storage medium and electronic device, including: decomposing the shaft vibration signal to be denoised to obtain a function set and residuals, the function set includes multiple intrinsic mode functions; processing each intrinsic mode function in the function set to obtain permutation entropy; determining the signal dominant type of each intrinsic mode function in the function set based on the permutation entropy, the signal dominant type is one of useful signal dominant and noise dominant; performing mathematical morphological filtering on each intrinsic mode function whose signal dominant type is useful signal dominant, and reconstructing the signal of each intrinsic mode function after filtering to obtain a denoised shaft vibration signal. The present application decomposes the amplitude signal to be denoised to obtain multiple intrinsic mode functions and residuals; processes each intrinsic mode function to obtain permutation entropy, and uses the permutation entropy to determine the signal dominant type of the intrinsic mode function; performs mathematical morphological filtering on each intrinsic mode function whose signal dominant type is useful signal dominant, and then uses the filtered intrinsic mode function to reconstruct the signal to obtain a denoised shaft vibration signal. In this way, a noise-reduced shaft vibration signal can be provided, providing reliable data support for fault diagnosis and early warning of the power station. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0063] Figure 1 A method flow chart of a shaft vibration signal noise reduction method provided in an embodiment of the present application;

[0064] Figure 2A flow chart of a method for decomposing a shaft vibration signal to be de-noised to obtain a function set and a residual provided in an embodiment of the present application;

[0065] Figure 3 A flow chart of a method for processing each intrinsic mode function in a function set to obtain permutation entropy provided in an embodiment of the present application;

[0066] Figure 4 Another flow chart of a method for reducing noise of a shaft vibration signal provided in an embodiment of the present application;

[0067] Figure 5 An example diagram of a time domain waveform of a simulation signal provided in an embodiment of the present application;

[0068] Figure 6 An example diagram of the analysis of a noisy shaft vibration signal by the CEEMDAN method provided in an embodiment of the present application;

[0069] Figure 7 An example diagram of the noise reduction result of the noisy shaft vibration signal provided in the embodiment of the present application after AVG processing;

[0070] Figure 8 A schematic diagram of the structure of a shaft vibration signal noise reduction device provided in an embodiment of the present application;

[0071] Fig. 9 This is an example diagram of results of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0072] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0073] In this application, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element.

[0074] Terminology explanation:

[0075] EMD: Empirical Mode Decomposition, is a new adaptive signal time-frequency processing method;

[0076] CEEMDAN: Complete Ensemble Empirical Mode Decomposition with AdaptiveNoise, a signal processing method based on empirical mode decomposition (EMD). CEEMDAN is an improved method developed on the basis of EMD.

[0077] PE: Permutation Entropy, is an average entropy function that measures the complexity of time series and detects dynamic mutations.

[0078] As can be seen from the background technology, traditional signal denoising methods are difficult to accurately separate useful signals and noise in high-noise environments, which affects the accuracy of subsequent fault diagnosis and early warning. In addition to this problem, traditional vibration signal denoising methods, such as empirical mode decomposition (EMD) and other technologies, are prone to problems such as modal aliasing and endpoint effects when processing signals with strong background noise, and are difficult to meet the signal denoising requirements in the complex operating environment of power stations.

[0079] In order to solve the above problems, the present application provides a shaft vibration signal noise reduction solution, which effectively classifies useful signals and noise signals based on fully adaptive noise set empirical mode decomposition, permutation entropy and mathematical morphology, and further performs noise reduction on useful signals. This solution effectively improves the noise reduction effect of the signal, provides reliable data support for power station fault diagnosis and early warning, and improves the operating efficiency and reliability of power station units.

[0080] The present application can be applied to the control system of a power station, and the control system can be formed by using a variety of general or special computing device environments or configurations, such as personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor devices, and distributed computing environments including any of the above devices or equipment.

[0081] Reference Figure 1 , is a method flow chart of a shaft vibration signal noise reduction method provided in an embodiment of the present application, and is specifically described as follows:

[0082] S101. Decompose the shaft vibration signal to be de-noised to obtain a function set and a residual, wherein the function set includes a plurality of intrinsic mode functions.

[0083] The shaft vibration signal is a signal collected from a power station. At this time, the shaft vibration signal to be de-noised is a shaft vibration signal containing noise.

[0084] The CEEMDAN method is used to decompose the shaft vibration signal to be denoised, thereby obtaining several intrinsic mode functions and residuals.

[0085] Preferably, several eigenmode functions are gathered in a function set.

[0086] Preferably, the amplitude signal to be denoised can be represented by V(t), and the intrinsic mode function is represented by IMF i Residual is represented by Res. Further, i=1,2,3,…,M, where M is the total number of IMFs.

[0087] S102, processing each intrinsic mode function in the function set to obtain permutation entropy.

[0088] The eigenmode functions obtained after decomposition using the CEEMDAN method are calculated to obtain the permutation entropy. Preferably, the permutation entropy is calculated using PE i express.

[0089] S103. Determine the signal-dominant type of each intrinsic mode function in the function set according to the permutation entropy, where the signal-dominant type is one of useful signal-dominant and noise-dominant.

[0090] According to the permutation entropy, the signal dominant type of each intrinsic mode function in the function set is determined. This step is to use the permutation entropy to identify useful intrinsic mode functions and useless intrinsic mode functions. The signal dominant type is divided into useful signal dominated and noise signal dominated, among which the intrinsic mode function dominated by the noise signal is a useless signal, and the intrinsic mode function dominated by the useful signal is a useful signal.

[0091] S104, performing mathematical morphological filtering processing on each intrinsic mode function whose dominant signal type is dominated by the useful signal, and reconstructing each intrinsic mode function after the filtering processing to obtain a noise-reduced shaft vibration signal.

[0092] Performing mathematical morphological filtering on each eigenmode function whose dominant signal type is dominated by user signals is essentially performing noise reduction processing, and the eigenmode function after filtering processing is the eigenmode function after noise reduction.

[0093] The intrinsic mode function after filtering is subjected to signal reconstruction processing to obtain a noise-reduced shaft vibration signal.

[0094] The present application provides an amplitude signal denoising method, which decomposes the shaft vibration signal to be denoised to obtain a function set and a residual, wherein the function set includes a plurality of intrinsic mode functions; processes each intrinsic mode function in the function set to obtain a permutation entropy; determines the signal dominant type of each intrinsic mode function in the function set based on the permutation entropy, wherein the signal dominant type is one of useful signal dominant and noise dominant; performs mathematical morphological filtering on each intrinsic mode function whose signal dominant type is useful signal dominant, and reconstructs the signal of each intrinsic mode function after filtering to obtain a denoised shaft vibration signal. The present application decomposes the amplitude signal to be denoised to obtain a plurality of intrinsic mode functions and a residual; processes each intrinsic mode function to obtain a permutation entropy, and uses the permutation entropy to determine the signal dominant type of the intrinsic mode function; performs mathematical morphological filtering on each intrinsic mode function whose signal dominant type is useful signal dominant, and then uses the filtered intrinsic mode function to reconstruct the signal to obtain a denoised shaft vibration signal. This solution can separate useful signals and noise from noisy shaft vibration signals in a high-noise environment, further reduce the noise of useful signals, and provide shaft vibration signals after noise reduction, thus providing reliable data support for fault diagnosis and early warning of power plants.

[0095] Reference Figure 2 , is a flow chart of a method for decomposing a shaft vibration signal to be de-noised to obtain a function set and a residual provided in an embodiment of the present application, and is specifically described as follows:

[0096] S201: Determine the shaft vibration signal to be de-noised as a target signal.

[0097] S202: Add preset adaptive Gaussian white noise to the target signal to construct multiple signals to be decomposed.

[0098] Preferably, the preset number of times of adding noise to the adaptive Gaussian white noise is K, and the value of K is a positive integer, and the specific value can be set according to actual needs. Further, the number of the multiple signals to be decomposed is also K.

[0099] S203 , performing decomposition processing on each signal to be decomposed to obtain an intrinsic mode function, and storing the intrinsic mode function in a function set.

[0100] Performing empirical mode decomposition on each signal to be decomposed to obtain each signal component of each signal to be decomposed; performing a mean operation on the first signal component of each signal to be decomposed, and taking the obtained mean as the intrinsic mode function.

[0101] In other words, the K signals to be decomposed are subjected to EMD decomposition respectively, and the first signal component of each signal to be decomposed is averaged, and the obtained average is used as the intrinsic mode function. Preferably, the intrinsic mode function uses IMFi Indicates; preferably, i can represent the number of iterations, and the value is a positive integer, for example: i = 1, 2, 3, ..., M, M is the total number of IMFs, for example, IMF1 represents the intrinsic mode function obtained in the first iteration.

[0102] S204: Use the intrinsic mode function and the target signal to determine the residual, and update the number of iterations.

[0103] The residual between the intrinsic mode function and the target signal is calculated, and then the number of iterations is updated. When the number of iterations is updated, the number of iterations may be increased by one. Preferably, the initial number of iterations is zero.

[0104] S205. Based on the residual and the updated number of iterations, determine whether a preset stop condition is met; when the stop condition is not met, execute S206; when the stop condition is met, execute S207.

[0105] When the residual is a monotonic function and / or the number of iterations reaches a preset maximum number of iterations, it is determined that the stopping condition is met; otherwise, it is determined that the stopping condition is not met.

[0106] The preset maximum number of iterations can be set according to actual needs.

[0107] When the stop condition is not met, S206 is executed; when the stop condition is met, S207 is executed.

[0108] S206: Use the residual as a new target signal and return to step S202.

[0109] S207. Output function set and residual.

[0110] Preferably, the function set output at this time includes multiple intrinsic mode functions, and the residual output at this time is the residual obtained in step S204 in the current iteration.

[0111] The present application uses adaptive Gaussian white noise to decompose the shaft vibration signal to be denoised, thereby obtaining a series of intrinsic mode functions and residuals, so as to facilitate the subsequent distinction between useful mode functions and useless mode functions from a series of intrinsic mode functions, thereby improving the subsequent signal noise reduction processing effect.

[0112] Reference Figure 3 , which is a flowchart of a method for processing each intrinsic mode function in a function set to obtain permutation entropy provided in an embodiment of the present application, and is specifically described as follows:

[0113] S301, reconstructing each intrinsic mode function in phase space to obtain a phase space matrix.

[0114] For each intrinsic mode function IMF, the phase space is reconstructed to obtain the phase space matrix. The specific expression of the phase space matrix is ​​as follows:

[0115]

[0116] Where N is the number of data points of IMFi, m is the embedding dimension, and τ is the time delay. The number of data points is the signal length, and the embedding dimension and time delay can be set according to experience or needs. For example, the embedding dimension can be 128 and the time delay can be 3.

[0117] S302. Sort each row in the phase space matrix to obtain a column index symbol sequence.

[0118] Sort each row in the phase space matrix to obtain the column index symbol sequence of each row. The expression of the column index symbol sequence is as follows: S(l) = (j1, j2, ..., j m ), where l is the column index symbol sequence number, l = 1, 2, 3, ..., k, k ≤ m! , m! is the total number of symbol sequences mapped in the m-dimensional phase space, j is the index of the element in each row; and m is the embedding dimension.

[0119] S303: Process the column index symbol sequence to obtain permutation entropy.

[0120] Calculate the probability of the column index symbol sequence appearing in S302. The probability can be expressed as: {P1, P2, P3, ..., P k}. The specific calculation formula of probability P is as follows:

[0121]

[0122] Where l = 1, 2, 3, …, k; τ is the time delay, N is the number of data points in IMFi, and m is the embedding dimension.

[0123] Based on Shannon's theory, the signal arrangement is defined, and the specific formula is:

[0124]

[0125] Where l is the column index symbol sequence number; P l is the probability of occurrence of the column index symbol sequence.

[0126] According to the maximum value of PE, PE is normalized. The specific formula is as follows:

[0127]

[0128] Where m is the embedding dimension, m! is the total number of symbol sequences mapped in the m-dimensional phase space; PE i is the permutation entropy of IMF.

[0129] From this, we can get the permutation entropy.

[0130] In the present application, after obtaining the permutation entropy, the signal dominant type of each intrinsic mode function is determined according to the permutation entropy, and the signal dominant type is one of useful signal dominant and noise dominant; the process of performing mathematical morphological filtering on the intrinsic mode function whose signal dominant type is useful signal dominant is as follows:

[0131] A mathematical morphology operator constructed by using the mean of an opening operation and a closing operation is determined; based on the mathematical morphology operator, each intrinsic mode function whose dominant signal type is dominated by a useful signal is subjected to mathematical morphology filtering processing to obtain the intrinsic mode function after filtering processing.

[0132] It should be noted that the mathematical morphology operator can be expressed using AVG. The specific expression of the mathematical morphology operator is: in, represents an open operation; represents a closed operation; g is a structural element, preferably, a flat structural element is used here, and the length is 20. IMF in the formula i is the eigenmode function dominated by the useful signal. Substituting the eigenmode function dominated by the useful signal into the mathematical morphology operator can obtain the eigenmode function after filtering. Furthermore, the eigenmode function after filtering is the eigenmode function after noise reduction.

[0133] After obtaining the filtered intrinsic mode function, each filtered intrinsic mode function is added together to obtain a reconstructed shaft vibration signal.

[0134] Reference Figure 4 , is another flow chart of a method for reducing noise of a shaft vibration signal provided in an embodiment of the present application, as shown in the figure:

[0135] S401, obtaining a noisy shaft vibration signal.

[0136] S402. Use the CEEMDAN method to process the noisy shaft vibration signal to obtain a series of IMFs and a residual.

[0137] S403 , processing a series of IMFs to obtain permutation entropy, and then using the permutation entropy to classify each IMF into a useful signal-dominated IMF and a noise-dominated IMF.

[0138] S404, performing mathematical morphological filtering on the dominant IMF of the useful signal, and then reconstructing the signal of the filtered IMF to obtain a shaft vibration signal after noise reduction.

[0139] Furthermore, the noisy shaft vibration signal is processed using the CEEMDAN method to obtain a series of IMFs and a residual, as shown in the following figure:

[0140] S501. Determine the noisy shaft vibration signal as a signal x(t).

[0141] S502: Set the number of noise additions K and the maximum number of iterations P.

[0142] The values ​​of K and P are both positive integers, and the specific values ​​can be set according to actual needs.

[0143] S503 , adding adaptive Gaussian white noise to the signal x(t) to obtain a signal to be decomposed, performing EMD decomposition on the signal to be decomposed to obtain a first signal component of the signal to be decomposed, and updating the number of noise adding iterations.

[0144] S504, determine whether the number of noise addition iterations reaches the number of noise additions, if not, return to execute S503; if reached, execute S505.

[0145] The number of noise additions here is the number of noise additions set in S502; the initial value of the number of noise addition iterations is 0, and the method for updating the number of noise addition iterations may be to perform a process of adding one.

[0146] S505 , perform a mean operation on each first signal component to obtain an IMF (equivalent to the intrinsic mode function mentioned above), determine the residual according to the IMF and the signal x(t), and update the number of iterations.

[0147] S506. Determine whether the preset stop iteration condition is met based on the updated number of iterations and the residual; if so, execute S507; if not, use the residual as the new signal x(t), and then return to execute step S503.

[0148] When the number of iterations reaches a maximum number of iterations, and / or the residual is a monotonic function, it is determined that the condition for stopping iteration is met.

[0149] S507. Obtain a series of IMFs and a residual.

[0150] This application introduces CEEMDAN technology to decompose the original signal, significantly suppressing the interference of noise on the signal IMF. Combined with PE analysis, it accurately distinguishes the noise-dominated and signal-dominated components, and uses mathematical morphology to further process the useful signal-dominated components to remove noise to the maximum extent and retain the main features of the signal; thus overcoming the limitations of traditional signal processing methods that rely on empirical mode decomposition or simple threshold denoising methods in terms of noise suppression and signal fidelity.

[0151] The noisy signal processed by the solution provided in this application has a higher signal-to-noise ratio (SNR), a lower root mean square error (RMSE) and a higher correlation coefficient (CC) compared with the original signal. By simulating shaft vibration signals under noises of different intensities, it is verified that the method has a certain applicability to signals with different noise intensities. Therefore, the present invention can provide important technical support for efficient and accurate noise reduction of shaft vibration signals in power plants, and lay a solid foundation for subsequent fault diagnosis and early warning work.

[0152] The following takes the analog signal as an example and describes it in detail in conjunction with the accompanying drawings. It should be emphasized that the following description is only exemplary and is not intended to limit the scope and application of the present invention.

[0153] This application proposes a denoising scheme based on fully adaptive noise set empirical mode decomposition, permutation entropy and mathematical morphology. First, CEEMDAN is used to decompose the noisy shaft vibration signal into several components. Secondly, the dominant component of the useful signal and the dominant component of the noise are distinguished through permutation entropy analysis. Then, mathematical morphology is used to denoise the dominant component of the useful signal. Finally, the dominant components of the useful signal after denoising are added to obtain the denoised shaft vibration signal.

[0154] A shaft vibration signal is simulated and experimentally verified based on the method proposed in this patent. The specific formula of the simulated signal is as follows:

[0155]

[0156] Among them, the characteristic frequency f1 represents the low frequency, which is generally 0.2-0.45 times the rotation frequency. Here we take 0.25f n The characteristic frequencies f1, f2 and f3 represent the 1-3 times of the rotation frequency, respectively. n , 2f n and 3f n Here, the speed is simulated at 500 rpm, f n It is about 8.33Hz. Generally speaking, the amplitude of the 1-fold frequency is the largest. Therefore, the vibration amplitudes corresponding to other characteristic frequencies are dimensionless. * 、A2 * 、A3 * and A4 * They are 0.25, 1, 0.25 and 0.1 respectively. The signal sampling frequency is 1000Hz and the sampling time is 2s.

[0157] In order to verify the noise reduction effect of this scheme, Gaussian white noise with signal-to-noise ratios of 5dB, 2.5dB, 0dB, -2.5dB and -5dB is introduced into the original shaft vibration signal to construct the noisy shaft vibration signal V1 * 、V2 * 、V3 * 、V4 * and V5 * The time domain waveform of the above analog signal is as follows: Figure 5 As shown, Figure 5 (a) is the noise-free shaft vibration signal V0*, Figure 5 (b) is the noisy shaft vibration signal V1 * , Figure 5 (c) is the noisy shaft vibration signal V2 * , Figure 5 (d) is the noisy shaft vibration signal V3 * , Figure 5 (e) is the noisy shaft vibration signal V4 * , Figure 5 (f) is the noisy shaft vibration signal V5 * .

[0158] Step 1: Decompose the noisy shaft vibration signal into several components by using CEEMDAN.

[0159] Take the noisy shaft vibration signal V1 * For example, Figure 6 Shows the noisy shaft vibration signal V1 * According to the analysis results of CEEMDAN method, Figure 6 (a) in the figure is the time domain waveform of each IMF. Figure 6 (b) is the FFT spectrum of each IMF. Noisy shaft vibration signal V1 * After CEEMDAN decomposition, 11 IMFs and 1 Res are obtained. Figure 6 In (b), IMF1 to IMF5 have no obvious characteristic frequencies and can be regarded as noise-dominated IMFs. IMF6 contains characteristic frequencies f2, f3, and f4, IMF7 contains characteristic frequency f2, and IMF8 and IMF9 both contain characteristic frequency f1. Therefore, CEEMDAN can effectively decompose the useful characteristic frequency components in the noisy shaft vibration signal.

[0160] Step 2: Distinguish the dominant component of useful signals from the dominant component of noise through permutation entropy analysis.

[0161] The step 2 includes the following sub-steps:

[0162] Sub-step A1: Perform PE analysis on each IMF and Res, and calculate the PE value of each IMF. The larger the PE value, the more noise content, and the more likely it is that the IMF is dominated by noise.

[0163] Sub-step A2: Set the PE threshold of IMF screening to 0.4 and perform screening based on this threshold. IMFs with PE values ​​higher than 0.4 are considered noise-dominated IMFs. These IMFs do not require further noise reduction and are directly removed. As shown in Table 1, IMF1 to IMF5 have PE values ​​greater than 0.4 and are considered noise-dominated IMFs and are directly removed, while IMF6 to IMF7 have PE values ​​greater than 0.4 and are considered noise-dominated IMFs and are directly removed. 11 If the PE value of Res is less than 0.4, it is considered that the useful signal dominates the IMF and further noise reduction processing is required.

[0164] Table 1 Decomposition based on CEEMDAN V1 * Get the permutation entropy of each IMF

[0165]

[0166] Step 3: Use the AVG method to reduce the noise of the IMF dominated by the useful signal.

[0167] For the selected IMF6 to IMF 11 AVG noise reduction is performed on the image and Res respectively.

[0168] Step 4: After AVG processing, all the useful signals are added and reconstructed to dominate the IMF to obtain the shaft vibration signal after noise reduction.

[0169] Noisy shaft vibration signal V1 * 、V2 * 、V3 * 、V4 * and V5 * The denoising results after AVG processing are as follows Figure 7 shown. Figure 7 (a) is the noise-free shaft vibration signal V0*, Figure 7 (b) is the noisy shaft vibration signal V1 * Noise reduction results, Figure 7 (c) is the noisy shaft vibration signal V2 * Noise reduction results, Figure 7 (d) is the noisy shaft vibration signal V3 * Noise reduction results, Figure 7 (e) is the noisy shaft vibration signal V4 * Noise reduction results, Figure 7 (f) is the noisy shaft vibration signal V5 *Denoising results. The three indicators of signal-to-noise ratio, root mean square error and correlation coefficient with the original shaft vibration signal without noise are introduced to evaluate the denoising effect. Table 2 shows the noisy shaft vibration signal V1 * 、V2 * 、V3 * and V4 * Evaluation index of denoising results processed by the method proposed in this patent.

[0170] Table 2 Simulated shaft vibration signal noise reduction index

[0171]

[0172] This application proposes a new noise reduction method based on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN), Permutation Entropy (PE) and mathematical morphology, aiming to improve the noise reduction effect of shaft vibration signals and ensure their accuracy and stability in complex noise environments. This method significantly improves the processing effect of shaft vibration signals in power plants and provides technical support for the operating efficiency and reliability of power plant equipment. By introducing CEEMDAN technology, the original signal is decomposed, which significantly suppresses the interference of noise on the signal IMF. Combined with PE analysis, the noise-dominated and signal-dominated components are accurately distinguished, and mathematical morphology is used to further process the dominant components of useful signals to remove noise to the maximum extent and retain the main features of the signal.

[0173] and Figure 1 Corresponding to the method shown in the figure, the present application also discloses a shaft vibration signal noise reduction device, which is used to support Figure 1 Implementation of the method shown.

[0174] Reference Figure 8 , is a schematic diagram of the structure of a shaft vibration signal noise reduction device provided in an embodiment of the present application, and is specifically described as follows:

[0175] A decomposition unit 801 is used to decompose the shaft vibration signal to be de-noised to obtain a function set and a residual, wherein the function set includes a plurality of intrinsic mode functions;

[0176] A processing unit 802 is used to process each intrinsic mode function in the function set to obtain a permutation entropy;

[0177] A determination unit 803 is used to determine a signal-dominant type of each intrinsic mode function in the function set according to the permutation entropy, wherein the signal-dominant type is one of useful signal-dominant and noise-dominant;

[0178] The filtering unit 804 is used to perform mathematical morphological filtering on each intrinsic mode function whose dominant signal type is dominated by the useful signal, and reconstruct the signal of each intrinsic mode function after the filtering process to obtain a noise-reduced shaft vibration signal.

[0179] The present application provides an amplitude signal denoising device, which decomposes the shaft vibration signal to be denoised to obtain a function set and a residual, wherein the function set includes a plurality of intrinsic mode functions; processes each intrinsic mode function in the function set to obtain a permutation entropy; determines the signal dominant type of each intrinsic mode function in the function set based on the permutation entropy, wherein the signal dominant type is one of useful signal dominant and noise dominant; performs mathematical morphological filtering on each intrinsic mode function whose signal dominant type is useful signal dominant, and reconstructs the signal of each intrinsic mode function after filtering to obtain a denoised shaft vibration signal. The present application decomposes the amplitude signal to be denoised to obtain a plurality of intrinsic mode functions and a residual; processes each intrinsic mode function to obtain a permutation entropy, and uses the permutation entropy to determine the signal dominant type of the intrinsic mode function; performs mathematical morphological filtering on each intrinsic mode function whose signal dominant type is useful signal dominant, and then uses the filtered intrinsic mode function to reconstruct the signal to obtain a denoised shaft vibration signal. In this way, a noise-reduced shaft vibration signal can be provided, providing reliable data support for fault diagnosis and early warning of the power station.

[0180] In another device provided by the present application, the decomposition unit 801 of the device performs a process of decomposing the shaft vibration signal to be de-noised to obtain a function set and a residual, including:

[0181] Determining the shaft vibration signal to be de-noised as a target signal;

[0182] Adding a preset adaptive Gaussian white noise to the target signal to construct a plurality of signals to be decomposed;

[0183] Decomposing each of the signals to be decomposed to obtain an intrinsic mode function, and storing the intrinsic mode function in a function set;

[0184] Determine a residual using the intrinsic mode function and the target signal, and update the number of iterations;

[0185] Based on the residual and the updated number of iterations, determining whether a preset stop condition is met;

[0186] When the stop condition is not met, the residual is used as a new target signal, and the step of adding a preset adaptive Gaussian white noise to the target signal to construct multiple signals to be decomposed is returned to the step, until the stop iteration condition is met, and the function set and the residual are output.

[0187] In another device provided by the present application, the decomposition unit 801 of the device performs decomposition processing on each of the signals to be decomposed to obtain an intrinsic mode function, including:

[0188] Performing empirical mode decomposition on each of the signals to be decomposed to obtain each signal component of each of the signals to be decomposed;

[0189] The first signal component of each of the signals to be decomposed is averaged, and the obtained average is used as the intrinsic mode function.

[0190] In another device provided by the present application, the decomposition unit 801 of the device performs a process of determining whether a preset stop condition is met based on the residual and the updated number of iterations, including:

[0191] When the residual is a monotonic function and / or the number of iterations reaches a preset maximum number of iterations, it is determined that the stopping condition is met; otherwise, it is determined that the stopping condition is not met.

[0192] In another device provided by the present application, the processing unit 802 of the device performs a process of processing each intrinsic mode function in the function set to obtain a permutation entropy, including:

[0193] Reconstructing each of the intrinsic mode functions in phase space to obtain a phase space matrix;

[0194] Sorting each row in the phase space matrix to obtain a column index symbol sequence;

[0195] The column index symbol sequence is processed to obtain permutation entropy.

[0196] In another device provided by the present application, the filtering unit 804 of the device performs a process of performing mathematical morphological filtering on each intrinsic mode function whose dominant signal type is dominated by the useful signal, including:

[0197] Determine mathematical morphology operators constructed using the mean of opening and closing operations;

[0198] Based on the mathematical morphology operator, mathematical morphology filtering is performed on each intrinsic mode function whose dominant signal type is dominated by useful signals to obtain the intrinsic mode function after filtering.

[0199] In another device provided by the present application, the filtering unit 804 of the device performs a process of reconstructing the signals of the filtered intrinsic mode functions to obtain a noise-reduced shaft vibration signal, including:

[0200] The eigenmode functions after filtering are added together to obtain a noise-reduced shaft vibration signal.

[0201] Although the present invention depicts operations in a particular order, this should not be understood as requiring that the operations be performed in the particular order shown or in a sequential order.Multitasking and parallel processing may be advantageous under certain circumstances.

[0202] It should be understood that the various steps described in the method embodiments of the present disclosure may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this respect.

[0203] An embodiment of the present invention further provides a storage medium, which includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the above-mentioned shaft vibration signal noise reduction method.

[0204] The embodiment of the present invention further provides an electronic device, the structural diagram of which is shown in FIG. Fig. 9 As shown, it specifically includes a memory 901 and one or more instructions 902, wherein the one or more instructions 902 are stored in the memory 901 and are configured to be executed by one or more processors 903 to perform the following operations:

[0205] Decomposing the shaft vibration signal to be de-noised to obtain a function set and a residual, wherein the function set includes a plurality of intrinsic mode functions;

[0206] Processing each intrinsic mode function in the function set to obtain permutation entropy;

[0207] Determining a signal-dominant type of each intrinsic mode function in the function set according to the permutation entropy, wherein the signal-dominant type is one of useful signal-dominant and noise-dominant;

[0208] Mathematical morphology filtering is performed on each intrinsic mode function whose dominant signal type is dominated by useful signals, and signal reconstruction is performed on each intrinsic mode function after filtering to obtain a noise-reduced shaft vibration signal.

[0209] It should be noted that the information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards in the relevant regions.

[0210] The specific implementation processes and derivative methods of the above-mentioned embodiments are all within the protection scope of the present invention.

[0211] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0212] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0213] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for reducing noise of shaft vibration signal, characterized in that: include: Decomposing the shaft vibration signal to be de-noised to obtain a function set and a residual, wherein the function set includes a plurality of intrinsic mode functions; Processing each intrinsic mode function in the function set to obtain permutation entropy; Determining a signal-dominant type of each intrinsic mode function in the function set according to the permutation entropy, wherein the signal-dominant type is one of useful signal-dominant and noise-dominant; Mathematical morphology filtering is performed on each intrinsic mode function whose dominant signal type is dominated by useful signals, and signal reconstruction is performed on each intrinsic mode function after filtering to obtain a noise-reduced shaft vibration signal.

2. The method according to claim 1, characterized in that The shaft vibration signal to be de-noised is decomposed to obtain a function set and a residual, including: Determining the shaft vibration signal to be de-noised as a target signal; Adding a preset adaptive Gaussian white noise to the target signal to construct a plurality of signals to be decomposed; Decomposing each of the signals to be decomposed to obtain an intrinsic mode function, and storing the intrinsic mode function in a function set; Determine a residual using the intrinsic mode function and the target signal, and update the number of iterations; Based on the residual and the updated number of iterations, determining whether a preset stop condition is met; When the stop condition is not met, the residual is used as a new target signal, and the step of adding a preset adaptive Gaussian white noise to the target signal to construct multiple signals to be decomposed is returned to the step, until the stop iteration condition is met, and the function set and the residual are output.

3. The method according to claim 2, characterized in that The decomposing and processing each of the signals to be decomposed to obtain an intrinsic mode function includes: Performing empirical mode decomposition on each of the signals to be decomposed to obtain each signal component of each of the signals to be decomposed; The first signal component of each of the signals to be decomposed is averaged, and the obtained average is used as the intrinsic mode function.

4. The method according to claim 2, characterized in that: The determining whether a preset stop condition is satisfied based on the residual and the updated number of iterations includes: When the residual is a monotonic function and / or the number of iterations reaches a preset maximum number of iterations, it is determined that the stopping condition is met; otherwise, it is determined that the stopping condition is not met.

5. The method according to claim 1, characterized in that The processing of each intrinsic mode function in the function set to obtain the permutation entropy includes: Reconstructing each of the intrinsic mode functions in phase space to obtain a phase space matrix; Sorting each row in the phase space matrix to obtain a column index symbol sequence; The column index symbol sequence is processed to obtain permutation entropy.

6. The method according to claim 1, characterized in that The mathematical morphological filtering is performed on each intrinsic mode function whose dominant signal type is dominated by the useful signal, including: Determine mathematical morphology operators constructed using the mean of opening and closing operations; Based on the mathematical morphology operator, mathematical morphology filtering is performed on each intrinsic mode function whose dominant signal type is dominated by useful signals to obtain the intrinsic mode function after filtering.

7. The method according to claim 1, characterized in that The signal reconstruction of each intrinsic mode function after filtering to obtain a noise-reduced shaft vibration signal includes: The eigenmode functions after filtering are added together to obtain a noise-reduced shaft vibration signal.

8. A shaft vibration signal noise reduction device, characterized in that: include: A decomposition unit, used for decomposing the shaft vibration signal to be de-noised to obtain a function set and a residual, wherein the function set includes a plurality of intrinsic mode functions; A processing unit, used for processing each intrinsic mode function in the function set to obtain a permutation entropy; A determination unit, configured to determine, based on the permutation entropy, a signal-dominant type of each intrinsic mode function in the function set, wherein the signal-dominant type is one of useful signal-dominant and noise-dominant; The filtering unit is used to perform mathematical morphological filtering on each intrinsic mode function whose dominant signal type is dominated by the useful signal, and reconstruct the signals of each intrinsic mode function after the filtering process to obtain a noise-reduced shaft vibration signal.

9. A storage medium, characterized in that: The storage medium includes stored instructions, wherein when the instructions are executed, the device where the storage medium is located is controlled to execute the shaft vibration signal noise reduction method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: It comprises a memory and one or more instructions, wherein the one or more instructions are stored in the memory and are configured to be executed by one or more processors to perform the shaft vibration signal noise reduction method as claimed in any one of claims 1 to 7.

Images