Electric gate valve fault diagnosis method, device and equipment based on strong noise background
By using the variational mode decomposition model optimized by the Blue Gardener optimization algorithm and the singular value decomposition method, combined with the bidirectional gated cyclic unit model of the particle swarm optimization algorithm, noise in the acceleration signal of the electric gate valve is effectively removed, and high-accuracy fault diagnosis is achieved under strong noise background.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HARBIN ENG UNIV
- Filing Date
- 2023-06-19
- Publication Date
- 2026-06-02
AI Technical Summary
In nuclear power plants, the numerous and highly coupled devices surrounding electric gate valves result in a large amount of noise in the collected acceleration signals, reducing the accuracy of fault diagnosis.
A variational mode decomposition model optimized by the Blue Gardener's optimization algorithm is adopted, combined with a bidirectional gated cyclic unit model of singular value decomposition and particle swarm optimization algorithm, to decompose and remove high-frequency and mid-to-low-frequency noise in the acceleration signal of electric gate valve, and extract fault signals for diagnosis.
Even in a noisy environment, it can accurately identify faults in electric gate valves, improving the accuracy of fault diagnosis and anti-interference ability, with a diagnostic accuracy rate of over 99%.
Smart Images
Figure CN116754213B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of fault diagnosis technology, and in particular to a method, apparatus and equipment for fault diagnosis of electric gate valves under strong noise background. Background Technology
[0002] Electric gate valves, as a crucial type of equipment in nuclear power plants, primarily function to open and close pipelines, control flow direction, and regulate and control the transported media. Due to the harsh working environment and high frequency of use in nuclear power plants, electric gate valves are more prone to failure compared to other equipment. A failure in an electric gate valve can disrupt the normal operation of the entire nuclear power plant and even increase the potential threat of radioactive leakage. Therefore, timely and accurate detection and identification of the fault type in the early stages of an electric gate valve failure are of paramount importance for improving the safety and reliability of the nuclear power plant.
[0003] Vibration signal analysis is widely used in the fault diagnosis of electric gate valves. It involves placing acceleration sensors near the electric gate valve to collect acceleration signals containing fault signals, and then processing these signals to analyze the fault type. However, due to the large number of devices in a nuclear power plant and the mutual coupling between them, the collected acceleration signals containing fault signals often contain a significant amount of noise. This noise may partially or completely mask the fault signals of the electric gate valve, thus reducing the accuracy of fault diagnosis. Summary of the Invention
[0004] This specification provides an embodiment of a method, apparatus, and equipment for diagnosing electric gate valve faults under strong noise conditions, in order to solve the problem of low accuracy in existing fault diagnosis methods.
[0005] To solve the above-mentioned technical problems, the embodiments in this specification are implemented as follows:
[0006] This specification provides an embodiment of a fault diagnosis method for electric gate valves under strong noise conditions, including:
[0007] Acquire the acceleration signal of an electric gate valve under strong noise background;
[0008] The first noise in the acceleration signal of the electric gate valve is determined using a variational mode decomposition model optimized by the Bluebird optimization algorithm.
[0009] The first noise is removed from the acceleration signal of the electric gate valve to obtain the first reconstructed acceleration signal of the electric gate valve.
[0010] The second noise is removed from the first reconstructed acceleration signal of the electric gate valve to obtain the second reconstructed acceleration signal of the electric gate valve; the frequency band of the second noise is smaller than the frequency band of the first noise.
[0011] Based on the second recombinant acceleration signal of the electric gate valve, fault diagnosis is performed on the electric gate valve.
[0012] Optionally, it also includes:
[0013] The number of decomposition levels and the penalty factor in the variational mode decomposition model are optimized based on the Blue Bowerbird optimization algorithm, resulting in the optimized variational mode decomposition model.
[0014] Optionally, determining the first noise in the electric gate valve acceleration signal using a variational mode decomposition model optimized based on the Bluebird Bobbin algorithm specifically includes:
[0015] The electric gate valve acceleration signal is decomposed using a variational mode decomposition model optimized by the Bluebird optimization algorithm to obtain multiple intrinsic mode functions of multiple frequency band signals containing the electric gate valve acceleration signal.
[0016] For any of the intrinsic mode functions, the frequency band of the signal contained in any of the intrinsic mode functions is determined by using the fast Fourier transform;
[0017] The first noise is determined based on the frequency band of the signal contained in any of the intrinsic mode functions and the frequency band where the first noise is located.
[0018] Optionally, the frequency band of the first noise is determined according to the following steps:
[0019] Mechanism analysis was performed on the electric gate valve to determine the frequency band of the electric gate valve's acceleration signal where the fault signal occurs when the electric gate valve malfunctions.
[0020] Based on the frequency band of the electric gate valve acceleration signal where the fault signal is located, the frequency band where the first noise is located is determined.
[0021] Optionally, before performing a mechanism analysis on the electric gate valve to determine the frequency band of the electric gate valve's acceleration signal where the fault signal occurs when the electric gate valve malfunctions, the method further includes:
[0022] Determine the fault characteristics of the electric gate valve when it malfunctions;
[0023] Based on the fault characteristics of the electric gate valve when it malfunctions, a fault condition for the electric gate valve is set.
[0024] An experimental platform was constructed for mechanistic analysis of the electric gate valve; the experimental platform included the electric gate valve fault and an acceleration sensor for acquiring the acceleration signal of the electric gate valve.
[0025] Optionally, removing the second noise from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve specifically includes:
[0026] Based on the singular value decomposition method, the second noise is removed from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve.
[0027] Optionally, the fault diagnosis of the electric gate valve based on the second recombined acceleration signal of the electric gate valve specifically includes:
[0028] Extract time-domain features from the second recombined acceleration signal of the electric gate valve;
[0029] Based on the extracted time-domain features, the electric gate valve is used for fault diagnosis using a bidirectional gated cyclic unit model optimized by the particle swarm optimization algorithm.
[0030] Optionally, the fault types of electric gate valves include at least one of the following: three-phase imbalance fault, damaged sealing packing fault, internal leakage fault, external leakage fault, valve stem lifting fault, transmission fault, and electric actuator fault.
[0031] This specification provides an embodiment of an electric gate valve fault diagnosis device based on a strong noise background, comprising:
[0032] The acquisition module is used to acquire the acceleration signal of the electric gate valve under strong noise background;
[0033] The determination module is used to determine the first noise in the acceleration signal of the electric gate valve using a variational mode decomposition model optimized based on the Blue Bowerbird optimization algorithm.
[0034] The first removal module is used to remove the first noise from the acceleration signal of the electric gate valve to obtain the first reconstructed acceleration signal of the electric gate valve.
[0035] The second removal module is used to remove the second noise from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve; the frequency band of the second noise is smaller than the frequency band of the first noise.
[0036] The diagnostic module is used to perform fault diagnosis on the electric gate valve based on the second recombinant acceleration signal of the electric gate valve.
[0037] This specification provides an embodiment of an electric gate valve fault diagnosis device based on a strong noise background, comprising:
[0038] At least one processor; and,
[0039] A memory communicatively connected to the at least one processor; wherein,
[0040] The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to:
[0041] Acquire the acceleration signal of an electric gate valve under strong noise background;
[0042] The first noise in the acceleration signal of the electric gate valve is determined using a variational mode decomposition model optimized by the Bluebird optimization algorithm.
[0043] The first noise is removed from the acceleration signal of the electric gate valve to obtain the first reconstructed acceleration signal of the electric gate valve.
[0044] The second noise is removed from the first reconstructed acceleration signal of the electric gate valve to obtain the second reconstructed acceleration signal of the electric gate valve; the frequency band of the second noise is smaller than the frequency band of the first noise.
[0045] Based on the second recombinant acceleration signal of the electric gate valve, fault diagnosis is performed on the electric gate valve.
[0046] One embodiment of this specification achieves the following beneficial effects:
[0047] This embodiment provides a method, apparatus, and equipment for fault diagnosis of electric gate valves under strong noise background. By using the variational mode decomposition model optimized by the Bluebird optimization algorithm, the first noise in the acceleration signal of the electric gate valve can be accurately determined, and most of the noise can be removed from the acceleration signal of the electric gate valve, thereby improving the accuracy of fault diagnosis of electric gate valves. Attached Figure Description
[0048] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0049] Figure 1 A flowchart illustrating a fault diagnosis method for an electric gate valve under strong noise conditions, provided as an embodiment of this specification;
[0050] Figure 2 A structural diagram of a bidirectional gated loop unit model provided in the embodiments of this specification;
[0051] Figure 3 A schematic diagram illustrating a fault diagnosis method for an electric gate valve under strong noise background, provided as an embodiment of this specification.
[0052] Figure 4 A schematic diagram of an electric gate valve fault diagnosis device based on a strong noise background, provided for an embodiment of this specification;
[0053] Figure 5 This is a schematic diagram of an electric gate valve fault diagnosis device based on a strong noise background, provided as an embodiment of this specification. Detailed Implementation
[0054] To make the objectives, technical solutions, and advantages of one or more embodiments of this specification clearer, the technical solutions of one or more embodiments of this specification will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of them. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of one or more embodiments of this specification.
[0055] The technical solutions provided in the various embodiments of this specification are described in detail below with reference to the accompanying drawings.
[0056] In existing technologies, vibration signal analysis is used to diagnose faults in electric gate valves. This involves placing accelerometers near the valve to collect acceleration signals containing fault signals, then processing these signals to analyze the fault type. However, due to the numerous devices in a nuclear power plant and their interconnections, the collected acceleration signals containing fault signals often contain significant noise. This noise can partially or completely mask the fault signals of the electric gate valve, thus reducing the accuracy of fault diagnosis.
[0057] To address the shortcomings of existing technologies, this solution provides the following embodiments:
[0058] Figure 1 A flowchart illustrating a fault diagnosis method for an electric gate valve under strong noise conditions, provided as an embodiment of this specification, is shown below. Figure 1 As shown, the method may include:
[0059] Step 102: Obtain the acceleration signal of the electric gate valve under strong noise background.
[0060] In this embodiment, the acquired electric gate valve acceleration signal includes noise and the electric gate valve's fault signal. Strong noise background refers to a situation where the noise portion of the electric gate valve's acceleration signal completely or entirely covers the electric gate valve's fault signal. In one specific embodiment, this is due to the presence of numerous devices in the electric gate valve's environment, and the mutual coupling between these devices.
[0061] Step 104: Determine the first noise in the acceleration signal of the electric gate valve using the variational mode decomposition model optimized by the Blue Gardener optimization algorithm.
[0062] In this embodiment, the first noise can be high-frequency noise, and the second noise can be mid-to-low-frequency noise.
[0063] Existing technologies for denoising the first type of noise, i.e., high-frequency noise, generally employ Empirical Mode Decomposition (EMD) and Wavelet Transform (WT) algorithms. However, both methods have limitations. EMD suffers from mode aliasing and endpoint effects, significantly impacting signal decomposition and feature parameter extraction. WT, on the other hand, requires pre-setting parameters such as wavelet basis functions, thresholds, and the number of decomposition levels, making adaptive decomposition difficult. Therefore, this embodiment employs a variational mode decomposition model—specifically, an optimized variational mode decomposition model—to determine the first type of noise in the electric gate valve's acceleration signal, thereby removing it from the signal.
[0064] In signal processing, Variational Mode Decomposition (VMD) is a signal decomposition and estimation method. This method determines the frequency center and bandwidth of each component by iteratively searching for the optimal solution of the variational model during the acquisition of decomposed components, thereby adaptively achieving frequency domain partitioning of the signal and effective separation of its components.
[0065] However, in variational mode decomposition, the number of modes needs to be manually set; too many or too few modes will affect the decomposition accuracy. Therefore, in this embodiment, an optimized variational mode decomposition model is used.
[0066] Specifically, this embodiment uses a variational mode decomposition model optimized by a genetic optimization algorithm. The genetic optimization algorithm is a method that searches for the optimal solution by simulating the natural evolutionary process. For an optimization problem, the genetic optimization algorithm iteratively performs operations such as selection, crossover, mutation, and evaluation on a population of a certain number of candidate solutions, causing the population to evolve towards a better solution.
[0067] Considering the applicability issues of different genetic optimization algorithms to different optimization objects, this application utilizes multiple genetic optimization algorithms, including the Wolf Pack Algorithm (WPA), Crow Search Algorithm (CSA), Salp Swarm Algorithm (SSA2017), Sine Cosine Algorithm (SCA), Satin Bowerbird Optimizer (SBO), Sparrow Search Algorithm (SSA2020), and Whale Optimization Algorithm (WOA), to optimize the variational mode decomposition model. Then, the optimal model is selected from the optimized variational mode decomposition models.
[0068] Specifically, the first noise in the acceleration signal of the electric gate valve is first determined by each optimized variational mode decomposition model, and then the first noise is removed from the acceleration signal of the electric gate valve to reduce the noise, thus obtaining the reconstructed acceleration signals after noise reduction.
[0069] In this implementation, the performance evaluation metric for each genetic optimization algorithm is the root mean square error (RMSE). Specifically, the RMS error is calculated by comparing the denoised reconstructed acceleration signal with the original electric gate valve acceleration signal. A smaller RMS error indicates that the variational mode decomposition model optimized by the genetic optimization algorithm can more accurately identify the first noise in the electric gate valve acceleration signal. The formula for calculating the RMS error is as follows:
[0070] (1)
[0071] In formula (1), The root mean square error, The number of reconstructed acceleration signals after noise reduction, original t For the t-th signal in the original electric gate valve acceleration signal, recommend t Let t be the t-th signal in the denoised reconstructed acceleration signal.
[0072] Finally, it was found that the smaller the root mean square error of the variational mode decomposition model optimized by the Satin Bowerbird Optimizer (SBO), the more accurately the first noise in the acceleration signal of the electric gate valve can be determined using the variational mode decomposition model optimized by the Satin Bowerbird Optimizer.
[0073] Because this embodiment utilizes a variational mode decomposition model optimized based on the Bluebird optimization algorithm, it can accurately determine the first noise, i.e., high-frequency noise, in the acceleration signal of the electric gate valve. Then, the determined noise is removed from the acceleration signal of the electric gate valve. In this way, most of the noise can be removed, thereby greatly preserving the fault signal in the acceleration signal of the electric gate valve and improving the accuracy of fault diagnosis of the electric gate valve.
[0074] Step 105: Remove the first noise from the acceleration signal of the electric gate valve to obtain the first recombined acceleration signal of the electric gate valve.
[0075] Once the first noise is determined, it can be removed from the acceleration signal of the electric gate valve to obtain the first recombined acceleration signal of the electric gate valve.
[0076] Step 108: Remove the second noise from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve; the frequency band of the second noise is smaller than the frequency band of the first noise.
[0077] After removing the first noise, which is high-frequency noise, the second noise, which is mid-to-low-frequency noise, can also be removed, thus obtaining the second recombined acceleration signal of the electric gate valve, which is the fault signal of the electric gate valve in the acceleration signal of the electric gate valve.
[0078] Step 110: Based on the second recombinant acceleration signal of the electric gate valve, perform fault diagnosis on the electric gate valve.
[0079] Based on the second recombined acceleration signal of the electric gate valve, that is, the fault signal of the electric gate valve in the acceleration signal of the electric gate valve, the fault diagnosis of the electric gate valve is performed.
[0080] It should be understood that the order of some steps in the methods described in one or more embodiments of this specification may be interchanged according to actual needs, or some steps may be omitted or deleted.
[0081] based on Figure 1 In addition to the method described herein, this specification also provides some specific implementation schemes of the method, which will be described below.
[0082] In this embodiment, before determining the first noise in the electric gate valve acceleration signal using the variational mode decomposition model optimized by the Bluebird optimization algorithm in step 104, the fault diagnosis method provided in this embodiment may further include:
[0083] The number of decomposition levels and the penalty factor in the variational mode decomposition model are optimized based on the Blue Bowerbird optimization algorithm, resulting in the optimized variational mode decomposition model.
[0084] Specifically, the decomposition process of the variational mode decomposition model is essentially the solution of a variational problem, that is, establishing constraint equations based on the narrowband conditions of the components, while simultaneously minimizing the sum of the estimated bandwidths of each mode. Its calculation process is shown in formulas (2) and (3):
[0085] (2)
[0086] (3)
[0087] In formulas (2) and (3), The input signal for the variational mode decomposition model is... For the k-th modal function component, Let be the center frequency of the k-th modal function. Here, δ represents the Dirac delta function, * represents the convolution operator, and δ represents the impulse function. For time, It is an imaginary number.
[0088] Then, the number of decomposition levels and the penalty factor in the variational mode decomposition model are optimized based on the Blue Bowerbird optimization algorithm, where the number of decomposition levels and the number of modes in the variational mode decomposition model are positively correlated. The penalty factor is then... and Lagrange operators By introducing the concept of transforming the solution of the constrained variational problem into the solution of the unconstrained variational problem:
[0089] (4)
[0090] In formula (4), the penalty factor It can guarantee the reconstruction accuracy of the signal under noisy conditions. It can make the constraints more stringent.
[0091] Then, using the Alternating Direction Method of Multipliers (ADMM) and Parseval's theorem, the solution is obtained. and The calculation process is as follows:
[0092] (5)
[0093] (6)
[0094] (7)
[0095] (8)
[0096] (9)
[0097] In formula (6), Noise tolerance.
[0098] The iterative solution process for the variational mode decomposition model is as follows:
[0099] (1) Initialization , , , and the maximum number of iterations N, where iteration n starts from 0;
[0100] (2) Update using formulas (7), (8) and (9) , and ;
[0101] (3) Determine whether the following inequalities are satisfied:
[0102]
[0103] in, To improve convergence accuracy, if the inequality is not satisfied, return to step (2); if the inequality is satisfied, proceed to step (4).
[0104] (4) When n≥N, the iteration is complete.
[0105] The variational mode decomposition (MODER) model is optimized based on the Satin Blue Bowerbird optimization algorithm. The objective parameters for optimization are the number of decomposition layers and the penalty factor, which are related to the envelope entropy value output by the variational mode decomposition model. The basic idea of envelope entropy is to extract signal features from the envelope signal obtained by Hilbert transform, combined with information entropy. After the signal is demodulated using Hilbert, the original signal sequence is transformed into a probability distribution sequence. The mathematical definition of the process of extracting signal features is:
[0106] (10)
[0107] (11)
[0108] (12)
[0109] In formulas (10), (11) and (11), Let be the envelope entropy. For signal The envelope signal sequence obtained by Hilbert demodulation; H is the Hilbert transform of the signal; for The normalized form, for The maximum value, This is the original signal.
[0110] In this embodiment, step 104, which uses the variational mode decomposition model optimized based on the Blue Bowerbird optimization algorithm to determine the first noise in the electric gate valve acceleration signal, may specifically include:
[0111] Step 1042: Decompose the electric gate valve acceleration signal using the variational mode decomposition model optimized by the Blue Gardener's optimization algorithm to obtain multiple intrinsic mode functions of multiple frequency band signals containing the electric gate valve acceleration signal.
[0112] Among them, the Intrinsic Mode Function (IMF) is a nonlinear function that can be used to simulate complex real-world models.
[0113] The intrinsic mode function can decompose the noise signal and fault signal of each frequency band in the acceleration signal of the electric gate valve.
[0114] Step 1044: For any of the intrinsic mode functions, determine the frequency band of the signal contained in any of the intrinsic mode functions using fast Fourier transform.
[0115] The Fast Fourier Transform (FFT) is an improved algorithm of the Discrete Fourier Transform (DFT). The calculation formula for the Discrete Fourier Transform is as follows:
[0116] (13)
[0117] In formula (12), Let j be the signal sequence in the time domain, where j is the imaginary unit.
[0118] The Fast Fourier Transform (FFT) offers higher computational efficiency compared to the Discrete Fourier Transform (DFT). Specifically, the FFT can decompose a signal of length N into two signals of length N / 2, as shown in the following formula:
[0119] (14)
[0120] In formula (13), EV(k) is the DFT of even-numbered sample points, obtained by performing a DFT on an odd-numbered sample point using an FFT, and Od(k) is the DFT of odd-numbered sample points, obtained by performing a DFT on an even-numbered sample point using an FFT.
[0121] , .
[0122] Step 1046: Determine the first noise based on the frequency band of the signal contained in any of the intrinsic mode functions and the frequency band where the first noise is located.
[0123] Based on the frequency band in which the first noise is located, the first noise can be determined from the signals contained in the intrinsic mode function.
[0124] In this embodiment, the frequency band of the first noise can be determined according to the following steps:
[0125] Mechanism analysis was performed on the electric gate valve to determine the frequency band of the electric gate valve's acceleration signal where the fault signal occurs when the electric gate valve malfunctions.
[0126] Based on the frequency band of the electric gate valve acceleration signal where the fault signal is located, the frequency band where the first noise is located is determined.
[0127] Furthermore, before performing a mechanism analysis on the electric gate valve to determine the frequency band of the electric gate valve's acceleration signal where the fault signal occurs when the electric gate valve malfunctions, the method provided in this embodiment may further include:
[0128] Determine the fault characteristics of the electric gate valve when it malfunctions;
[0129] Based on the fault characteristics of the electric gate valve when it malfunctions, a fault condition for the electric gate valve is set.
[0130] An experimental platform was constructed for mechanistic analysis of the electric gate valve; the experimental platform included the electric gate valve fault and an acceleration sensor for acquiring the acceleration signal of the electric gate valve.
[0131] Based on the constructed experimental platform, the mechanism of the electric gate valve can be analyzed, and the frequency band of the electric gate valve acceleration signal where the fault signal occurs can be quickly determined.
[0132] In this embodiment, step 108, which removes the second noise from the first reconstructed acceleration signal of the electric gate valve to obtain the second reconstructed acceleration signal of the electric gate valve, may specifically include:
[0133] Based on the singular value decomposition method, the second noise is removed from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve.
[0134] Compared to other methods, singular value decomposition is more effective at removing low- and mid-frequency noise from the acceleration signal of an electric gate valve.
[0135] The noise reduction process of Singular Value Decomposition (SVD) is as follows:
[0136] (1) Reconstruct the recombined acceleration signal s(t) into a Hankel matrix of order As shown below:
[0137] (15)
[0138] In formula (15), N is the length of the reconstructed signal s(t); 1 <m<N,1<n<N。
[0139] To make the Hankel matrix closer to a square matrix, the number of rows m and the number of columns n are selected according to formulas (16) and (17):
[0140] (16)
[0141] (17)
[0142] (2) For the Hankel matrix Singular value decomposition yields the following decomposition form:
[0143] (18)
[0144] In formula (18), It is Left orthogonal matrix, It is A right orthogonal matrix of order 1. It is a diagonal matrix, and its specific form is shown below:
[0145] (19)
[0146] In formula (19), Hankel matrix The singular values, and q represents the total number of singular values obtained through singular value decomposition.
[0147] (3) Select an effective singular value discrimination method and determine the number of effective singular values k.
[0148] (4) Preserve the diagonal matrix The first k valid singular values are used, and the remaining smaller singular values are set to zero to obtain a new singular value matrix. ,Will Substituting into equation (18), we obtain a new matrix A. Then, matrix A is converted into a one-dimensional signal through inverse operation. .
[0149] The calculation process of the singular value difference spectrum is as follows:
[0150] (20)
[0151] In formula (20), .
[0152] Let the difference spectrum be denoted as In D, the first k singular values are the singular values corresponding to the true components of the signal, and the (k+1)th singular value represents the singular values corresponding to the noise. The kth singular value is:
[0153] (twenty one)
[0154] In this embodiment, step 110, based on the second recombined acceleration signal of the electric gate valve, performs fault diagnosis on the electric gate valve, which may specifically include:
[0155] Extract time-domain features from the second recombined acceleration signal of the electric gate valve;
[0156] Based on the extracted time-domain features, the electric gate valve is used for fault diagnosis using a bidirectional gated cyclic unit model optimized by the particle swarm optimization algorithm.
[0157] In this embodiment, the extracted time-domain features include at least one of the following: mean, root mean square, peak value, peak index, impulse index, skewness index, kurtosis index, and margin index.
[0158] In this embodiment, a bidirectional gate recurrent unit (BGRU) model is used to diagnose faults in electric gate valves. The BGRU is an improvement upon the gate recurrent unit (GRU), meaning it's an enhanced version of the GRU model. Compared to the GRU model, the BGRU model can fully utilize both forward and backward data during training, resulting in higher recognition accuracy.
[0159] Figure 2 A structural diagram of a bidirectional gated loop unit model provided in the embodiments of this specification is shown below. Figure 2As shown, the bidirectional gated recurrent unit (GRU) model consists of two GRU models: a forward-gated GRU model for receiving forward input and a reverse-gated GRU model for learning reverse input. In the diagram, GRU2 represents the forward circuit, and GRU1 represents the reverse circuit. Indicates the time. The specific calculation process is as follows:
[0160] (twenty two)
[0161] (twenty three)
[0162] (twenty four)
[0163] Where, x t As input to the model, y t For model output, For positive output, This is the reverse output.
[0164] However, the parameters in the bidirectional gated loop unit model require manual adjustment, which is rarely optimal. Therefore, in this implementation, particle swarm optimization (PSO) can be used to optimize the parameters in the bidirectional gated loop unit model, and the optimized model can be used to diagnose electric gate valve faults. This allows for more accurate diagnosis of electric gate valve faults.
[0165] Figure 3 A schematic diagram illustrating a fault diagnosis method for an electric gate valve under strong noise conditions, provided in the embodiments of this specification, is shown below. Figure 3 As shown, the method may include:
[0166] Step 302: Set up various types of electric gate valve faults and build an experimental platform; at the same time, arrange acceleration sensors around the electric gate valve to collect the acceleration signal of the electric gate valve under strong noise background.
[0167] In this embodiment, the fault types of the electric gate valve include three-phase imbalance fault, sealing packing damage fault, and external leakage fault. In one embodiment, internal leakage fault, valve stem lifting fault, transmission fault, and electric actuator fault can also be set. Of course, other electric gate valve fault types can also be set, which are not listed here.
[0168] Step 304: Optimize the variational mode decomposition model using CSA, WPA, SBO, SCA, WOA, SSA2017, and SSA2020 algorithms respectively to obtain multiple optimized variational mode decomposition models.
[0169] Step 306: For any optimized variational mode decomposition model, use this optimized variational mode decomposition model to determine the high-frequency noise, and then remove the high-frequency noise to obtain the reconstructed acceleration signal after removing the high-frequency noise.
[0170] Step 308: Calculate the root mean square error (RMSE) between the reconstructed acceleration signal after removing high-frequency noise and the original electric gate valve acceleration signal.
[0171] By comparing the root mean square errors (RMSEs) of all optimized variational mode decomposition (MODE) models, the optimized MODE model with the smallest RMSE is designated as the optimal model. Calculations show that the MODE model optimized using the Bluebird optimization algorithm has the smallest RMSE. Therefore, high-frequency noise can be identified based on the MODE model optimized using the Bluebird optimization algorithm, and then removed to obtain the optimal reconstructed acceleration signal after removing the high-frequency noise.
[0172] Step 310: Remove low- and mid-frequency noise from the optimal recombined acceleration signal.
[0173] Specifically, the singular value decomposition method is used to remove low- and mid-frequency noise from the optimal reconstructed acceleration signal, resulting in a reconstructed acceleration signal after removing the low- and mid-frequency noise.
[0174] Step 312: Extract time-domain features from the reconstructed acceleration signal after removing low- and mid-frequency noise.
[0175] The extracted time-domain features include at least one of the following: mean, root mean square, peak, peak index, impulse index, skewness index, kurtosis index, and margin index.
[0176] Step 314: Based on the extracted time-domain features, perform fault diagnosis on the electric gate valve.
[0177] Specifically, a bidirectional gated loop unit model is first built, and then the PSO algorithm is used to optimize the bidirectional gated loop unit model. The extracted time-domain features and the optimized bidirectional gated loop unit model are used to diagnose the faults of the electric gate valve.
[0178] The fault diagnosis method for electric gate valves under strong noise background provided in this embodiment was verified using historical data. The experimental results show that the accuracy of fault diagnosis for various types of electric gate valves is above 99%.
[0179] This embodiment utilizes a variational mode decomposition model optimized based on the Bluebird optimization algorithm to accurately identify high-frequency noise in the acceleration signal of an electric gate valve. This high-frequency noise is then removed to obtain a reconstructed acceleration signal. Next, based on singular value decomposition, mid- and low-frequency noise is removed from the reconstructed acceleration signal to obtain a reconstructed acceleration signal with mid- and low-frequency noise removed. Finally, fault diagnosis is performed. Through these steps, most of the noise in the electric gate valve's acceleration signal can be removed while preserving the fault signal to a great extent. Compared to existing electric gate valve fault diagnosis methods, the fault diagnosis method provided in this embodiment has extremely strong anti-interference capabilities. That is, the fault diagnosis method in this embodiment can accurately diagnose electric gate valve faults even when noise completely covers the fault signal, i.e., under strong noise conditions.
[0180] Based on the same idea, embodiments of this specification also provide apparatus corresponding to the above methods. Figure 4 This is a schematic diagram of an electric gate valve fault diagnosis device based on a strong noise background, provided as an embodiment of this specification. Figure 5 As shown, the device may include:
[0181] Acquisition module 402 is used to acquire the acceleration signal of the electric gate valve under strong noise background;
[0182] The determination module 404 is used to determine the first noise in the electric gate valve acceleration signal using the variational mode decomposition model optimized based on the Blue Bowerbird optimization algorithm;
[0183] The first removal module 406 is used to remove the first noise from the electric gate valve acceleration signal to obtain the first reconstructed acceleration signal of the electric gate valve.
[0184] The second noise removal module 408 is used to remove the second noise from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve; the frequency band of the second noise is smaller than the frequency band of the first noise.
[0185] The diagnostic module 410 is used to perform fault diagnosis on the electric gate valve based on the second recombinant acceleration signal of the electric gate valve.
[0186] Furthermore, the device may also include:
[0187] The model acquisition module is used to optimize the number of decomposition layers and the penalty factor in the variational mode decomposition model based on the Blue Bowerbird Optimization Algorithm, so as to obtain the variational mode decomposition model optimized by the Blue Bowerbird Optimization Algorithm.
[0188] Furthermore, module 404 may specifically include:
[0189] The decomposition unit is used to decompose the electric gate valve acceleration signal using a variational mode decomposition model optimized based on the Blue Bowerbird optimization algorithm, to obtain multiple intrinsic mode functions of multiple frequency band signals containing the electric gate valve acceleration signal.
[0190] The first determining unit is used to determine the frequency band of the signal contained in any of the intrinsic mode functions using a fast Fourier transform.
[0191] The second determining unit is used to determine the first noise based on the frequency band of the signal contained in any of the intrinsic mode functions and the frequency band where the first noise is located.
[0192] Furthermore, the frequency band of the first noise can be determined according to the following steps:
[0193] Mechanism analysis was performed on the electric gate valve to determine the frequency band of the electric gate valve's acceleration signal where the fault signal occurs when the electric gate valve malfunctions.
[0194] Based on the frequency band of the electric gate valve acceleration signal where the fault signal is located, the frequency band where the first noise is located is determined.
[0195] Furthermore, before performing a mechanism analysis on the electric gate valve to determine the frequency band of the electric gate valve's acceleration signal where the fault signal occurs when the electric gate valve malfunctions, the process may further include:
[0196] Determine the fault characteristics of the electric gate valve when it malfunctions;
[0197] Based on the fault characteristics of the electric gate valve when it malfunctions, a fault condition for the electric gate valve is set.
[0198] An experimental platform was constructed for mechanistic analysis of the electric gate valve; the experimental platform included the electric gate valve fault and an acceleration sensor for acquiring the acceleration signal of the electric gate valve.
[0199] Furthermore, the second removal module 408 may specifically include:
[0200] The removal unit is used to remove the second noise from the first recombined acceleration signal of the electric gate valve based on the singular value decomposition method, so as to obtain the second recombined acceleration signal of the electric gate valve.
[0201] Furthermore, the diagnostic module 410 may specifically include:
[0202] Extraction unit, used to extract time-domain features from the second recombined acceleration signal of the electric gate valve;
[0203] The fault diagnosis unit is used to diagnose the electric gate valve based on the extracted time-domain features and using a bidirectional gated loop unit model optimized by the particle swarm optimization algorithm.
[0204] Furthermore, the types of electric gate valve failures include at least one of the following: three-phase imbalance failure, damaged sealing packing failure, internal leakage failure, external leakage failure, valve stem lifting failure, transmission failure, and electric actuator failure.
[0205] Based on the same idea, this specification also provides devices corresponding to the above methods in its embodiments.
[0206] Figure 5 This is a schematic diagram of an electric gate valve fault diagnosis device based on a strong noise background, provided as an embodiment of this specification. Figure 5 As shown, device 500 may include:
[0207] At least one processor 510; and,
[0208] Memory 530 communicatively connected to the at least one processor; wherein,
[0209] The memory 530 stores instructions 520 that can be executed by the at least one processor 510, the instructions being executed by the at least one processor 510 to enable the at least one processor 510 to:
[0210] Acquire the acceleration signal of an electric gate valve under strong noise background;
[0211] The first noise in the acceleration signal of the electric gate valve is determined using a variational mode decomposition model optimized by the Bluebird optimization algorithm.
[0212] The first noise is removed from the acceleration signal of the electric gate valve to obtain the first reconstructed acceleration signal of the electric gate valve.
[0213] The second noise is removed from the first reconstructed acceleration signal of the electric gate valve to obtain the second reconstructed acceleration signal of the electric gate valve; the frequency band of the second noise is smaller than the frequency band of the first noise.
[0214] Based on the second recombinant acceleration signal of the electric gate valve, fault diagnosis is performed on the electric gate valve.
[0215] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on its differences from other embodiments. In particular, for... Figure 5The electric gate valve fault diagnosis device shown is basically similar to the method embodiment, so the description is relatively simple. For relevant parts, please refer to the description of the method embodiment.
[0216] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.
[0217] For ease of description, the above devices are described separately by function as various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.
[0218] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0219] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.
[0220] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.
Claims
1. A fault diagnosis method for electric gate valves under strong noise background, characterized in that, include: Acquire the acceleration signal of an electric gate valve under strong noise background; The first noise in the electric gate valve acceleration signal is determined using a variational mode decomposition model optimized by the Bluebird optimization algorithm. Specifically, this includes: decomposing the electric gate valve acceleration signal using the variational mode decomposition model optimized by the Bluebird optimization algorithm to obtain multiple intrinsic mode functions of multiple frequency band signals containing the electric gate valve acceleration signal. For any of the intrinsic mode functions, the frequency band of the signal contained in any of the intrinsic mode functions is determined by using the fast Fourier transform; The first noise is determined based on the frequency band of the signal contained in any of the intrinsic mode functions and the frequency band where the first noise is located; The first noise is removed from the acceleration signal of the electric gate valve to obtain the first reconstructed acceleration signal of the electric gate valve. The second noise is removed from the first reconstructed acceleration signal of the electric gate valve to obtain the second reconstructed acceleration signal of the electric gate valve; the frequency band of the second noise is smaller than the frequency band of the first noise. Based on the second recombined acceleration signal of the electric gate valve, fault diagnosis is performed on the electric gate valve. The frequency band of the first noise is determined according to the following steps: performing a mechanism analysis on the electric gate valve to determine the frequency band of the electric gate valve acceleration signal where the fault signal when the electric gate valve malfunctions is located; Based on the frequency band of the electric gate valve acceleration signal where the fault signal is located, the frequency band where the first noise is located is determined; The step of removing the second noise from the first reconstructed acceleration signal of the electric gate valve to obtain the second reconstructed acceleration signal of the electric gate valve specifically includes: Based on the singular value decomposition method, the second noise is removed from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve.
2. The method according to claim 1, characterized in that, Also includes: The number of decomposition levels and the penalty factor in the variational mode decomposition model are optimized based on the Blue Bowerbird optimization algorithm, resulting in the optimized variational mode decomposition model.
3. The method according to claim 1, characterized in that, Before performing a mechanism analysis on the electric gate valve to determine the frequency band of the electric gate valve's acceleration signal where the fault signal occurs when the electric gate valve malfunctions, the method further includes: Determine the fault characteristics of the electric gate valve when it malfunctions; Based on the fault characteristics of the electric gate valve when it malfunctions, a fault condition for the electric gate valve is set. An experimental platform was constructed for mechanistic analysis of the electric gate valve; the experimental platform included the electric gate valve fault and an acceleration sensor for acquiring the acceleration signal of the electric gate valve.
4. The method according to claim 1, characterized in that, The fault diagnosis of the electric gate valve based on the second recombined acceleration signal of the electric gate valve specifically includes: Extract time-domain features from the second recombined acceleration signal of the electric gate valve; Based on the extracted time-domain features, the electric gate valve is used for fault diagnosis using a bidirectional gated cyclic unit model optimized by the particle swarm optimization algorithm.
5. The method according to claim 1, characterized in that, Electric gate valve failure types include at least one of the following: three-phase imbalance failure, damaged sealing packing failure, internal leakage failure, external leakage failure, valve stem lifting failure, transmission failure, and electric actuator failure.
6. A fault diagnosis device for electric gate valves under strong noise background, characterized in that, include: The acquisition module is used to acquire the acceleration signal of the electric gate valve under strong noise background; The determination module is used to determine the first noise in the electric gate valve acceleration signal using a variational mode decomposition model optimized based on the Bluebird optimization algorithm. Specifically, it includes: decomposing the electric gate valve acceleration signal using a variational mode decomposition model optimized based on the Bluebird optimization algorithm to obtain multiple intrinsic mode functions of multiple frequency band signals containing the electric gate valve acceleration signal. For any of the intrinsic mode functions, the frequency band of the signal contained in any of the intrinsic mode functions is determined by using the fast Fourier transform; The first noise is determined based on the frequency band of the signal contained in any of the intrinsic mode functions and the frequency band where the first noise is located; The first removal module is used to remove the first noise from the acceleration signal of the electric gate valve to obtain the first reconstructed acceleration signal of the electric gate valve. The second removal module is used to remove the second noise from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve; the frequency band of the second noise is smaller than the frequency band of the first noise. The diagnostic module is used to perform fault diagnosis on the electric gate valve based on the second recombined acceleration signal of the electric gate valve; The frequency band of the first noise is determined according to the following steps: performing a mechanism analysis on the electric gate valve to determine the frequency band of the electric gate valve acceleration signal where the fault signal when the electric gate valve malfunctions is located; Based on the frequency band of the electric gate valve acceleration signal where the fault signal is located, the frequency band where the first noise is located is determined; The step of removing the second noise from the first reconstructed acceleration signal of the electric gate valve to obtain the second reconstructed acceleration signal of the electric gate valve specifically includes: Based on the singular value decomposition method, the second noise is removed from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve.
7. A fault diagnosis device for electric gate valves under strong noise background, comprising: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, which, when executed by the at least one processor, enable the at least one processor to: Acquire the acceleration signal of an electric gate valve under strong noise background; The first noise in the electric gate valve acceleration signal is determined using a variational mode decomposition model optimized by the Bluebird optimization algorithm. Specifically, this includes: decomposing the electric gate valve acceleration signal using the variational mode decomposition model optimized by the Bluebird optimization algorithm to obtain multiple intrinsic mode functions of multiple frequency band signals containing the electric gate valve acceleration signal. For any of the intrinsic mode functions, the frequency band of the signal contained in any of the intrinsic mode functions is determined by using the fast Fourier transform; The first noise is determined based on the frequency band of the signal contained in any of the intrinsic mode functions and the frequency band where the first noise is located; The first noise is removed from the acceleration signal of the electric gate valve to obtain the first reconstructed acceleration signal of the electric gate valve. The second noise is removed from the first reconstructed acceleration signal of the electric gate valve to obtain the second reconstructed acceleration signal of the electric gate valve; the frequency band of the second noise is smaller than the frequency band of the first noise. Based on the second recombined acceleration signal of the electric gate valve, fault diagnosis is performed on the electric gate valve. The frequency band of the first noise is determined according to the following steps: performing a mechanism analysis on the electric gate valve to determine the frequency band of the electric gate valve acceleration signal where the fault signal when the electric gate valve malfunctions is located; Based on the frequency band of the electric gate valve acceleration signal where the fault signal is located, the frequency band where the first noise is located is determined; The step of removing the second noise from the first reconstructed acceleration signal of the electric gate valve to obtain the second reconstructed acceleration signal of the electric gate valve specifically includes: Based on the singular value decomposition method, the second noise is removed from the first recombined acceleration signal of the electric gate valve to obtain the second recombined acceleration signal of the electric gate valve.