Switch cabinet structural anomaly detection method and device

Abstract

The invention discloses a method and a device for detecting structural anomaly of a switch cabinet, which are used for solving the technical problem of non-ideal detection effect caused by the existing method for detecting the structural anomaly of the switch cabinet. The method comprises the following steps: preprocessing an obtained switch cabinet original vibration signal based on a plurality of preset white noise sequences, and outputting a normalized switch cabinet digital signal and a plurality of normalized switch cabinet disturbance signals; performing first-order mode successive extraction according to the normalized switch cabinet digital signal and the plurality of normalized switch cabinet disturbance signals based on an empirical mode decomposition algorithm, and outputting a total residual error and a plurality of initial adaptive intrinsic mode function components; screening each initial adaptive intrinsic mode function component, and outputting a plurality of target adaptive intrinsic mode function components; performing superposition reconstruction according to the multiple target adaptive intrinsic mode function components and the total residual error, and outputting a switch cabinet noise reduction signal; and performing structural anomaly detection on the switch cabinet based on the switch cabinet noise reduction signal, and outputting a target detection result.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electrical equipment detection, and in particular to a switch cabinet structural abnormality detection method and device. BACKGROUND

[0002] As a key primary equipment in the power system, the running state of the switch cabinet is directly related to the safety and stability of the power distribution system. As an important node of power transmission and distribution, the switch cabinet will be subjected to mechanical stress, electromagnetic force, and environmental temperature changes during long-term operation, which may cause changes in the mechanical structure state of the equipment.

[0003] Vibration signals, as one of the important physical characteristics of equipment state, can reflect potential abnormalities such as mechanical structure loosening, wear, deformation, or impact in the early stage, and are helpful for realizing non-invasive fault warning and state evaluation. Therefore, how to efficiently extract key vibration features of the switch cabinet during operation has become an important technical means to realize intelligent operation and maintenance and preventive maintenance.

[0004] Existing switch cabinet structural abnormality detection methods mostly process switch cabinet signals based on fast Fourier transform or multi-scale analysis based on fixed wavelet bases, and then complete abnormality detection based on the processed signals. However, when facing the strong nonlinearity and non-stationary characteristics of switch cabinet vibration signals, as well as the complex electromagnetic interference background in the field, there may be situations such as limited resolution and leakage of key feature information, resulting in unsatisfactory detection results. SUMMARY

[0005] The present application provides a switch cabinet structural abnormality detection method and device to solve the technical problem of unsatisfactory detection results caused by existing switch cabinet structural abnormality detection methods.

[0006] The first aspect of the present application provides a switch cabinet structural abnormality detection method, comprising:

[0007] Obtaining a switch cabinet original vibration signal, and preprocessing the switch cabinet original vibration signal based on a plurality of preset white noise sequences to output a normalized switch cabinet digital signal and a plurality of normalized switch cabinet disturbance signals;

[0008] Based on the empirical mode decomposition algorithm, a first-order modal is extracted from the normalized switch cabinet digital signal and a plurality of normalized switch cabinet disturbance signals, and a total residual error and a plurality of initial adaptive intrinsic mode function components are output;

[0009] Each of the initial adaptive intrinsic mode function components is screened to output a plurality of target adaptive intrinsic mode function components;

[0010] Superimposed reconstruction is performed according to the plurality of target adaptive eigenmode function components and the total residual error, and a switch cabinet noise reduction signal is output.

[0011] Based on the switch cabinet noise reduction signal, structural anomaly detection is performed on the switch cabinet, and a target detection result is output.

[0012] Optionally, the plurality of preset white noise sequences are used to preprocess the switch cabinet original vibration signal, and a normalized switch cabinet digital signal and a plurality of normalized switch cabinet disturbance signals are output, including:

[0013] Signal conversion is performed on the switch cabinet original vibration signal, and a switch cabinet digital signal is output.

[0014] The switch cabinet digital signal is normalized, and a normalized switch cabinet digital signal is output.

[0015] Each of the preset white noise sequences is used to disturb the normalized switch cabinet digital signal, and a plurality of normalized switch cabinet disturbance signals are generated.

[0016] Optionally, based on the empirical mode decomposition algorithm, first-order modal successive extraction is performed on the normalized switch cabinet digital signal and the plurality of normalized switch cabinet disturbance signals, and a total residual error and a plurality of initial adaptive eigenmode function components are output, including:

[0017] The empirical mode decomposition algorithm is used to decompose each of the normalized switch cabinet disturbance signals into an eigenmode function component, and the eigenmode function component corresponding to each of the normalized switch cabinet disturbance signals is output.

[0018] Each of the eigenmode function components is averaged, and an initial adaptive eigenmode function component is output.

[0019] The normalized switch cabinet digital signal and the initial adaptive eigenmode function component are used to determine a residual error signal.

[0020] It is determined whether the residual error signal is a monotonic function.

[0021] If yes, the energy density corresponding to the initial adaptive eigenmode function component is used to calculate a damping weight factor corresponding to the initial adaptive eigenmode function component.

[0022] The normalized switch cabinet digital signal, the initial adaptive eigenmode function component, and the damping weight factor corresponding to the initial adaptive eigenmode function component are used to calculate a total residual error.

[0023] If no, noise sequence decomposition is performed on each of the normalized switch cabinet disturbance signals, and a noise mode corresponding to each of the normalized switch cabinet disturbance signals is output.

[0024] Superimpose the noise mode corresponding to each of the normalized switch cabinet disturbance signals with the residual signal respectively to generate a plurality of new normalized switch cabinet disturbance signals, and jump to execute the step of performing intrinsic mode function component decomposition on each of the normalized switch cabinet disturbance signals using the empirical mode decomposition algorithm, and output the intrinsic mode function component corresponding to each of the normalized switch cabinet disturbance signals until the residual signal is the monotonic function.

[0025] Optionally, the step of screening each of the initial adaptive intrinsic mode function components and outputting a plurality of target adaptive intrinsic mode function components comprises:

[0026] An intrinsic mode function component correlation coefficient corresponding to each of the initial adaptive intrinsic mode function components is calculated using the energy density and average period corresponding to each of the initial adaptive intrinsic mode function components.

[0027] A main frequency band spectral entropy corresponding to each of the initial adaptive intrinsic mode function components is output by performing main frequency band spectral entropy calculation on each of the initial adaptive intrinsic mode function components.

[0028] Each of the intrinsic mode function component correlation coefficients is compared with a preset first threshold value, and each of the main frequency band spectral entropies is compared with a preset second threshold value.

[0029] The initial adaptive intrinsic mode function component corresponding to the intrinsic mode function component correlation coefficient greater than the preset first threshold value and the main frequency band spectral entropy less than the preset second threshold value is taken as a target adaptive intrinsic mode function component.

[0030] Optionally, the calculation formula of the total residual is specifically:

[0031] ;

[0032] wherein, is the total residual; is the normalized switch cabinet digital signal; is the total order, indicating the total number of times of empirical mode decomposition of the signal; is the damping weight factor corresponding to the jth-order initial adaptive intrinsic mode function component; is the jth-order initial adaptive intrinsic mode function component.

[0033] Optionally, the calculation formula of the intrinsic mode function component correlation coefficient is specifically:

[0034] ;

[0035] ;

[0036] wherein, is an intrinsic modal function component correlation coefficient corresponding to the initial adaptive intrinsic modal function component of the jth order; is a product of the energy density and the average period corresponding to the initial adaptive intrinsic modal function component of the jth order; is an average value of the product of the energy density and the average period corresponding to the initial adaptive intrinsic modal function component of the j-1th order; is an energy density corresponding to the initial adaptive intrinsic modal function component of the jth order; is an average period corresponding to the initial adaptive intrinsic modal function component of the jth order.

[0037] The second aspect of the present application provides a switch cabinet structural anomaly detection device, comprising:

[0038] An acquisition module is configured to acquire a switch cabinet original vibration signal, and pre-process the switch cabinet original vibration signal based on a plurality of preset white noise sequences, output a normalized switch cabinet digital signal and a plurality of normalized switch cabinet disturbance signals;

[0039] An output module is configured to perform first-order modal successive extraction based on an empirical mode decomposition algorithm according to the normalized switch cabinet digital signal and the plurality of normalized switch cabinet disturbance signals, and output a total residual error and a plurality of initial adaptive intrinsic modal function components;

[0040] A screening module is configured to screen each of the initial adaptive intrinsic modal function components, and output a plurality of target adaptive intrinsic modal function components;

[0041] A reconstruction module is configured to superimpose and reconstruct according to the plurality of target adaptive intrinsic modal function components and the total residual error, and output a switch cabinet noise reduction signal;

[0042] A detection module is configured to perform structural anomaly detection on a switch cabinet based on the switch cabinet noise reduction signal, and output a target detection result.

[0043] The third aspect of the present application provides a computer device, comprising a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the switch cabinet structural anomaly detection method according to any one of the above aspects.

[0044] The fourth aspect of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed to implement the steps of the switch cabinet structural anomaly detection method according to any one of the above aspects.

[0045] A fifth aspect of the present invention provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions, wherein when the program instructions are executed by a computer, the computer executes the steps of the switch cabinet structural abnormality detection method as described in any one of the above items.

[0046] It can be seen from the above technical solutions that the present invention has the following advantages:

[0047] The above technical solution of the present invention provides a method for detecting structural anomalies of a switch cabinet, which obtains the original vibration signal of the switch cabinet, preprocesses the original vibration signal of the switch cabinet based on multiple preset white noise sequences, and outputs a normalized switch cabinet digital signal and multiple normalized switch cabinet disturbance signals; based on the empirical mode decomposition algorithm, the first-order mode is extracted one by one according to the normalized switch cabinet digital signal and the multiple normalized switch cabinet disturbance signals, and the total residual and multiple initial adaptive intrinsic mode function components are output; each initial adaptive intrinsic mode function component is screened, and multiple target adaptive intrinsic mode function components are output. state function components; based on multiple target adaptive intrinsic mode function components and total residuals, superposition and reconstruction are performed to output the switch cabinet noise reduction signal; structural anomaly detection is performed on the switch cabinet based on the switch cabinet noise reduction signal, and the target detection result is output; based on the above scheme, the present invention performs first-order mode extraction on the multiple normalized switch cabinet disturbance signals obtained based on the empirical mode decomposition algorithm to realize adaptive extraction of modal components, and on this basis, reconstructs the filtered modes to obtain a high-fidelity switch cabinet noise reduction signal, thereby maximally retaining the structural fault characteristics and thus improving the detection effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0049] Figure 1 A flowchart of a method for detecting structural anomalies in a switch cabinet provided in the first embodiment of the present invention;

[0050] Figure 2 A schematic flow chart of a method for detecting structural anomalies in a switch cabinet provided in the first embodiment of the present invention;

[0051] Figure 3 This is a structural block diagram of a switch cabinet structural abnormality detection device provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0052] The embodiment of the present application provides a switch cabinet structural abnormality detection method and device, and aims to solve the technical problem that the existing switch cabinet structural abnormality detection method results in an unsatisfactory detection effect.

[0053] In order to make the invention purposes, features and advantages of the present application more obvious and easy to understand, the technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the following described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the present application.

[0054] Please refer to Figure 1 , Figure 1 The embodiment one of the present application provides a step flow chart of a switch cabinet structural abnormality detection method.

[0055] The embodiment of the present application provides a switch cabinet structural abnormality detection method, which comprises the following steps.

[0056] Step 101: acquiring a switch cabinet original vibration signal, and pre-processing the switch cabinet original vibration signal based on a plurality of preset white noise sequences to output a normalized switch cabinet digital signal and a plurality of normalized switch cabinet disturbance signals.

[0057] It should be noted that the high-sensitivity vibration sensor is used to collect the switch cabinet original vibration signal in a running state, i.e., the switch cabinet original vibration signal.

[0058] Further, step 101 can comprise the following sub-steps S11-S13.

[0059] Step S11: performing signal conversion on the switch cabinet original vibration signal to output a switch cabinet digital signal.

[0060] Step S12: normalizing the switch cabinet digital signal to output a normalized switch cabinet digital signal.

[0061] Step S13: using each preset white noise sequence to disturb the normalized switch cabinet digital signal respectively to generate a plurality of normalized switch cabinet disturbance signals.

[0062] It should be noted that the normalized switch cabinet digital signal x(t) is obtained by converting the switch cabinet digital signal into a digital signal (i.e., a switch cabinet digital signal) through an analog-to-digital conversion module, and adding a plurality of white noise sequences with different amplitudes (i.e., a plurality of preset white noise sequences) to the normalized switch cabinet digital signal to obtain a plurality of disturbance signals, i.e., a plurality of normalized switch cabinet disturbance signals. The noise adding times N and the noise amplitude coefficient ε are set. Due to the strong impact of the switch cabinet opening and closing action at the moment, the frequency span is large, the vibration signal generated has nonlinear and non-stationary characteristics, and the effective mode is usually concentrated between high frequencies IMF1-IMF3. Therefore, in order to avoid mode aliasing, ε should be appropriately increased and set within 0.2-0.4 to enhance the frequency spectrum expansion capability. N is set to 50-100 to maintain statistical average while reducing real-time processing burden. The noise is added to the original signal after normalization to obtain the reconstructed signal x i (t). i (t) is the i-th normalized switch cabinet disturbance signal.

[0063] Step 102: based on the empirical mode decomposition algorithm, first-order modal is extracted from the normalized switch cabinet digital signal and the plurality of normalized switch cabinet disturbance signals, and a total residual error and a plurality of initial adaptive intrinsic mode function components are output.

[0064] Specifically, step 102 can include the following sub-steps S21-S28:

[0065] Step S21: using the empirical mode decomposition algorithm, the intrinsic mode function component of each normalized switch cabinet disturbance signal is decomposed, and the intrinsic mode function component corresponding to each normalized switch cabinet disturbance signal is output.

[0066] Step S22: averaging each intrinsic mode function component, an initial adaptive intrinsic mode function component is output.

[0067] Step S23: determining a residual error signal based on the normalized switch cabinet digital signal and the initial adaptive intrinsic mode function component.

[0068] Step S24: determining whether the residual error signal is a monotonic function.

[0069] Step S25: if yes, the energy density corresponding to the initial adaptive intrinsic mode function component is used to calculate the damping weight factor corresponding to the initial adaptive intrinsic mode function component.

[0070] Step S26: calculating the total residual error based on the normalized switch cabinet digital signal, the initial adaptive intrinsic mode function component, and the damping weight factor corresponding to the initial adaptive intrinsic mode function component.

[0071] Step S27: If not, perform noise sequence decomposition on each normalized switchgear disturbance signal, and output the noise mode corresponding to each normalized switchgear disturbance signal;

[0072] Step S28: Superimpose the noise modes corresponding to each normalized switchgear disturbance signal with the residual signal to generate multiple new normalized switchgear disturbance signals, and jump to execute the step of using the empirical mode decomposition algorithm to decompose the intrinsic mode function components of each normalized switchgear disturbance signal, and output the intrinsic mode function components corresponding to each normalized switchgear disturbance signal until the residual signal is a monotonic function.

[0073] It should be noted that the present invention constructs a step-by-step adaptive decomposition process. First, the empirical mode decomposition algorithm is used to decompose only each group of disturbance signals (normalized switchgear disturbance signals) x i The first-order intrinsic mode function (IMF) component in (t) , that is, the first-order intrinsic mode function component corresponding to the i-th normalized switchgear disturbance signal, and then all the first-order IMF components Take average, , get the first-order adaptive IMF component, that is, the first-order initial adaptive eigenmode function component , then calculate the residual signal of the first-order mode : , determine whether the residual signal is a monotonic function. If not, perform EMD decomposition (Empirical Mode Decomposition) on the noise added to each set of disturbance signals to obtain the corresponding first-order noise mode. The noise mode includes the IMF sequence and residuals , the residual signal is superimposed with the noise mode corresponding to each group of disturbance signals, and multiple new normalized switch cabinet disturbance signals are output. The next round of decomposition is performed, and only the first-order mode is extracted to obtain the second-order IMF component. And so on. The order residual signal is iteratively processed. Specifically, after obtaining multiple new normalized switchgear disturbance signals, the currently obtained initial adaptive intrinsic mode function components are retained, and then the process jumps to step S21 until the residual signal is a monotonic function, and all the initial adaptive intrinsic mode function components retained when the residual signal is a monotonic function are output; wherein, is the j-th order IMF sequence corresponding to the i-th normalized switchgear disturbance signal, is the residual corresponding to the ith normalized switchgear disturbance signal, and t is the time.

[0074] Furthermore, the final total residual is expressed as , M is the number of times the decomposition is completed when the termination condition is reached, that is, the total number of times the signal undergoes empirical mode decomposition. The present invention introduces a damping weight factor in the calculation of the total residual. , this factor is assigned according to the energy density of each order IMF component (each order initial adaptive intrinsic mode function component) to reflect its structural contribution to the original signal. It is defined as , regulatory factor The value can be in the range of 0.8 to 1.2. is the energy density of the j-th order initial adaptive eigenmode function component, which is calculated as follows: , A is the length of the IMF component (initial adaptive intrinsic mode function component), wherein the criterion for the termination of the decomposition of the present invention is when the residual signal r j The decomposition stops when it becomes a monotonic function. Therefore, the total residual contains the low-frequency trend component in the signal that cannot be decomposed. In the subsequent reconstruction, it is superimposed with the filtered effective IMF to complete the signal reconstruction and ensure the integrity of the reconstructed signal.

[0075] It is worth mentioning that in the signal decomposition process, "the residual signal is a monotonic function" means that the residual signal remaining after multiple empirical mode decompositions (EMD) no longer contains any oscillatory components, and manifests as a single trend that continues to rise, fall, or remain constant over time. Specifically, a monotonic signal mathematically satisfies the property that its first-order derivative is always non-negative (monotonically increasing) or non-positive (monotonically decreasing), which means that the signal no longer has local extreme points or zero crossings and cannot be further decomposed into physically meaningful oscillatory modes (IMF components). In practical applications, this condition serves as the termination criterion for EMD decomposition. Its significance lies in: when the residual is monotonic, it indicates that all valid oscillation modes of the original signal have been completely extracted, and the remaining part is only the DC component or the slowly changing trend term. Continuing to decompose not only fails to obtain new useful information, but may also introduce false components. For example, in switchgear vibration signal analysis, if the final residual shows linear growth (e.g., due to equipment temperature drift), decomposition is considered complete. At this point, all oscillation modes reflecting fault characteristics such as mechanical shock and looseness are included in the extracted IMF components, while monotonic residuals can generally be attributed to environmental interference or long-term drift effects. This mechanism ensures the completeness of signal decomposition and its engineering practicality.

[0076] Step 103: Screen the initial adaptive intrinsic mode function components and output a plurality of target adaptive intrinsic mode function components.

[0077] Specifically, step 103 may include the following sub-steps:

[0078] The energy density and average period corresponding to each initial adaptive intrinsic mode function component are used to calculate the intrinsic mode function component correlation coefficient corresponding to each initial adaptive intrinsic mode function component;

[0079] Calculating the main frequency band spectral entropy of each initial adaptive intrinsic mode function component and outputting the main frequency band spectral entropy corresponding to each initial adaptive intrinsic mode function component;

[0080] Comparing the correlation coefficient of each intrinsic mode function component with a preset first threshold value, and comparing the spectral entropy of each main frequency band with a preset second threshold value;

[0081] An initial adaptive eigenmode function component corresponding to a main frequency band spectral entropy having an eigenmode function component correlation coefficient greater than a preset first threshold and less than a preset second threshold is used as a target adaptive eigenmode function component.

[0082] It should be noted that the present invention introduces a dual-characteristic entropy criterion to set the IMF validity screening mechanism to screen the initial adaptive intrinsic mode function components. Specifically, first use the formula , , calculate the energy density of the j-th order initial adaptive eigenmode function component and averaging period ,in, is the j-th order initial adaptive eigenmode function component The number of extreme points. The IMF correlation coefficient based on energy density and average period Defined as , It represents the product of the energy density of the j-th order initial adaptive eigenmode function component and the average period, that is, , Before The average value of the product of the j-1-order IMF, that is, the average value of the product of the energy density corresponding to the initial adaptive eigenmode function components of the first j-1 orders and the average period.

[0083] Furthermore, let the spectral density distribution of the initial adaptive eigenmode function component be , count the cumulative distribution of spectrum energy, start accumulating from the lowest frequency, and intercept the main frequency range F where the cumulative energy reaches the set proportion (such as 90%) main The spectrum energy is normalized only within the main frequency band, and the spectrum entropy is recalculated. The spectrum entropy of the main frequency band is: , f represents the frequency component, that is, the frequency coordinate in the Fourier transform result, the preset first threshold value is 1, the second threshold value is 0.8, when >1, <0.8 (in actual tests, the modal functions with structural periodic characteristics such as characteristic vibration signals caused by electromagnetic excitation or abnormal defects of the switch cabinet often show spectral concentration, and the spectral entropy is usually less than 0.8) indicates >1, The j-th order initial adaptive intrinsic modal function component corresponding to <0.8 contains more useful signal components and is retained, that is, the initial adaptive intrinsic modal function component corresponding to the main frequency band spectral entropy greater than the preset first threshold value and less than the preset second threshold value is taken as the target adaptive intrinsic modal function component, and the initial adaptive intrinsic modal function component corresponding to the main frequency band spectral entropy less than or equal to the preset first threshold value or greater than or equal to the preset second threshold value is removed. The j-th order initial adaptive intrinsic modal function component corresponding to <0.8 contains more useful signal components and is retained, that is, the initial adaptive intrinsic modal function component corresponding to the main frequency band spectral entropy greater than the preset first threshold value and less than the preset second threshold value is taken as the target adaptive intrinsic modal function component, and the initial adaptive intrinsic modal function component corresponding to the main frequency band spectral entropy less than or equal to the preset first threshold value or greater than or equal to the preset second threshold value is removed.

[0084] Step 104, superimposed reconstruction is performed according to the plurality of target adaptive intrinsic modal function components and the total residual error, and a switch cabinet denoising signal is output.

[0085] It should be noted that the effective intrinsic modal function components (target adaptive intrinsic modal function components) retained after screening are combined with the total residual error for superimposed reconstruction to obtain a denoised vibration signal (switch cabinet denoising signal).

[0086] Step 105, structural anomaly detection is performed on the switch cabinet based on the switch cabinet denoising signal, and a target detection result is output.

[0087] It should be noted that a series of characteristic parameters capable of reflecting the internal structural state of the switch cabinet are extracted based on the reconstructed signal (switch cabinet denoising signal), including time domain and frequency domain characteristics. In terms of time domain characteristics, the impact amplitude index is extracted by detecting the maximum amplitude change of the signal on the time axis, which is used to reflect the transient impact behavior caused by switch closing and other behaviors. At the same time, the envelope of the switch cabinet denoising signal is extracted by Hilbert transform, and the envelope energy is further calculated to identify the signal characteristics caused by mechanical looseness and component wear. In terms of frequency domain characteristics, the kurtosis coefficient of the signal can be used to represent the sharpness and non-Gaussian characteristics of the signal. In addition, the spectral complexity indicators such as the spectral entropy calculated in step five are used to quantitatively analyze the order of the signal. The above various characteristic parameters can be used to identify whether there is mechanical looseness, component wear or other structural anomaly problems in the switch cabinet.

[0088] As a comparison of technical effects, reference can be made in combination with the prior art. One of the traditional processing methods is variational modal decomposition, which decomposes the original vibration signal into a set of modal components with limited bandwidth and step-by-step optimized center frequency by constructing a variational model. In the decomposition process, each mode is regarded as a narrowband signal in different frequency bands, and the sum of the bandwidths of the modes is minimized through iterative optimization, so that components with good frequency domain separation characteristics are obtained. After completing the decomposition, the energy size or energy proportion of each modal component can be calculated respectively to analyze its trend under different working conditions, which is used to identify possible abnormal states of the equipment, such as impact or structural loosening. In order to extract high-frequency response representing fault features such as mechanical gap change, the envelope demodulation method is often combined, that is, the envelope curve is extracted by rectifying or filtering the components, and the frequency spectrum characteristics of the envelope signal are analyzed. Another commonly used method is wavelet packet decomposition, which is based on the standard wavelet transform and further divides the signal into more detailed frequency bands, that is, not only the approximate signal in the sub-band is further decomposed, but also the detail signal in the sub-band is recursively decomposed, so that more rich time-frequency information is obtained. Wavelet packet decomposition decomposes the original signal into a plurality of sub-signals with different center frequencies and bandwidths, each sub-band covers the entire frequency spectrum, and is suitable for analyzing the local frequency characteristics in non-stationary signals. In the feature extraction stage, the energy or energy entropy of each frequency band sub-signal is usually calculated, that is, by measuring the complexity of the energy distribution of each frequency band, the order of the signal structure and the information content are reflected. The energy entropy trend can be used to identify the differences between the running states of the equipment, such as whether the energy distribution characteristics change significantly between normal and fault states, so as to realize the preliminary identification of the fault type.

[0089] Based on the above foundation, traditional vibration signal processing methods such as fast Fourier transform and multi-scale analysis based on fixed wavelet basis often have problems such as limited resolution, leakage of key feature information, etc. when facing the strong nonlinearity, non-stationarity and complex electromagnetic interference background of the switch cabinet vibration signal. The methods of empirical mode decomposition and ensemble empirical mode decomposition appeared in recent years to some extent alleviate the modal aliasing phenomenon, but still face technical bottlenecks such as non-adaptive threshold selection, large cumulative error of decomposition residual, strong parameter dependence, etc. Although the diagnosis method based on machine learning has intelligence, it is strongly dependent on high-quality labeled samples, and it is difficult to obtain abnormal vibration samples in actual working conditions, which limits the promotion and application of the model.

[0090] In view of the strong nonlinearity, non-stationarity and high noise background of the switch cabinet vibration signal in actual operation, the traditional method has limitations such as insufficient resolution, serious modal aliasing and strong parameter dependence in noise reduction and feature extraction, and it is difficult to accurately identify potential structural faults of the equipment, etc. The present application proposes a switch cabinet structural anomaly detection method, please refer to Figure 2, by introducing white noise disturbance and first-order modal successive extraction mechanism, adaptive extraction of modal components and noise suppression are realized; the dual feature entropy criterion constructed by combining energy density and average period is used to screen effective modes and suppress background interference; on this basis, the screened modes are reconstructed to obtain high-fidelity feature signals, and time domain features and frequency domain features are further extracted to fully reflect the internal structure state of the switch cabinet, improve the accuracy and intelligent level of abnormal identification such as mechanical looseness, component wear and impact response, so as to realize the state monitoring and early fault warning of the switch cabinet.

[0091] In summary, the adaptive ensemble empirical mode decomposition structure is designed, the modal bandwidth and center frequency are dynamically adjusted by constructing multiple disturbance signals and extracting first-order modes in stages, the nonlinear and non-stationary characteristics of the switch cabinet vibration signal are adaptively decomposed, the modal aliasing is effectively suppressed and the decomposition stability is improved. Unlike fixed threshold or single entropy index screening, the energy-period correlation coefficient and spectral entropy dual criterion are introduced, the energy density and frequency spectrum order degree of each IMF are evaluated, the useful mode and noise mode can be distinguished, the reconstruction error is reduced and the structural fault characteristics are maximally retained. In addition, the time domain impact amplitude and envelope energy are extracted to reflect transient impact and mechanical looseness, and the frequency domain complexity indicators such as kurtosis coefficient and spectral entropy are combined to fully quantify the order degree of the reconstructed signal, which can more accurately extract the vibration characteristics of different operating states.

[0092] In the embodiment of the application, the switch cabinet structural anomaly detection method provided by the application comprises the following steps: obtaining a switch cabinet original vibration signal, and preprocessing the switch cabinet original vibration signal based on a plurality of preset white noise sequences to output a normalized switch cabinet digital signal and a plurality of normalized switch cabinet disturbance signals; performing first-order modal successive extraction based on an empirical mode decomposition algorithm according to the normalized switch cabinet digital signal and the plurality of normalized switch cabinet disturbance signals to output a total residual error and a plurality of initial adaptive intrinsic mode function components; screening the initial adaptive intrinsic mode function components to output a plurality of target adaptive intrinsic mode function components; superimposing and reconstructing the plurality of target adaptive intrinsic mode function components and the total residual error to output a switch cabinet denoising signal; and detecting the structure of the switch cabinet based on the switch cabinet denoising signal to output a target detection result. Based on the above scheme, the plurality of normalized switch cabinet disturbance signals obtained based on the empirical mode decomposition algorithm are subjected to first-order modal successive extraction, adaptive extraction of modal components is realized, and on this basis, the screened modes are reconstructed to obtain a high-fidelity switch cabinet denoising signal, thereby maximally retaining structural fault characteristics and further improving the detection effect.

[0093] Please refer to Figure 3 , Figure 3 The structural block diagram of a switch cabinet structural anomaly detection device provided by Embodiment Three of the application is shown in the figure.

[0094] The application provides a switch cabinet structural anomaly detection device, which comprises:

[0095] The acquisition module 301 is configured to acquire a switch cabinet original vibration signal, and preprocess the switch cabinet original vibration signal based on a plurality of preset white noise sequences, output a normalized switch cabinet digital signal and a plurality of normalized switch cabinet disturbance signals.

[0096] The output module 302 is configured to perform first-order modal successive extraction on the normalized switch cabinet digital signal and the plurality of normalized switch cabinet disturbance signals based on an empirical mode decomposition algorithm, and output a total residual error and a plurality of initial adaptive intrinsic mode function components.

[0097] The screening module 303 is configured to screen the initial adaptive intrinsic mode function components, and output a plurality of target adaptive intrinsic mode function components.

[0098] The reconstruction module 304 is configured to superimpose and reconstruct the plurality of target adaptive intrinsic mode function components and the total residual error, and output a switch cabinet denoising signal.

[0099] The detection module 305 is configured to perform structural anomaly detection on the switch cabinet based on the switch cabinet denoising signal, and output a target detection result.

[0100] Further, the acquisition module 301 is specifically configured to:

[0101] perform signal conversion on the switch cabinet original vibration signal, and output a switch cabinet digital signal;

[0102] normalize the switch cabinet digital signal, and output a normalized switch cabinet digital signal;

[0103] disturb the normalized switch cabinet digital signal using each preset white noise sequence, and generate a plurality of normalized switch cabinet disturbance signals.

[0104] Further, the output module 302 is specifically configured to:

[0105] perform intrinsic mode function component decomposition on each normalized switch cabinet disturbance signal using an empirical mode decomposition algorithm, and output an intrinsic mode function component corresponding to each normalized switch cabinet disturbance signal;

[0106] average each intrinsic mode function component, and output an initial adaptive intrinsic mode function component;

[0107] determine a residual error signal based on the normalized switch cabinet digital signal and the initial adaptive intrinsic mode function component;

[0108] determine whether the residual error signal is a monotonic function;

[0109] If yes, the energy density corresponding to the initial adaptive eigenmode function component is adopted to calculate the damping weight factor corresponding to the initial adaptive eigenmode function component;

[0110] According to the normalized switch cabinet digital signal, the initial adaptive eigenmode function component, and the damping weight factor corresponding to the initial adaptive eigenmode function component, the total residual error is calculated.

[0111] If no, the noise sequence decomposition is performed on each normalized switch cabinet disturbance signal respectively, and the noise mode corresponding to each normalized switch cabinet disturbance signal is output.

[0112] The noise mode corresponding to each normalized switch cabinet disturbance signal is superimposed with the residual error signal respectively to generate multiple new normalized switch cabinet disturbance signals, and the step of performing the empirical mode decomposition algorithm to decompose each normalized switch cabinet disturbance signal into an eigenmode function component is executed until the residual error signal is a monotonic function.

[0113] Further, the screening module 303 is specifically configured to:

[0114] The energy density and the average period corresponding to each initial adaptive eigenmode function component are adopted to calculate the eigenmode function component correlation coefficient corresponding to each initial adaptive eigenmode function component.

[0115] The main frequency band spectral entropy of each initial adaptive eigenmode function component is calculated, and the main frequency band spectral entropy corresponding to each initial adaptive eigenmode function component is output.

[0116] Each eigenmode function component correlation coefficient is compared with a preset first threshold value, and each main frequency band spectral entropy is compared with a preset second threshold value.

[0117] The initial adaptive eigenmode function component corresponding to the eigenmode function component correlation coefficient greater than the preset first threshold value and the main frequency band spectral entropy less than the preset second threshold value is taken as a target adaptive eigenmode function component.

[0118] Further, the calculation formula of the total residual error is specifically:

[0119]

[0120] wherein, is the total residual error; is the normalized switch cabinet digital signal; is the total order, indicating the total number of times of empirical mode decomposition of the signal; is the damping weight factor corresponding to the jth-order initial adaptive eigenmode function component; ​An initial adaptive eigenmode function component of the jth order.

[0121] Further, the calculation formula of the eigenmode function component correlation coefficient is specifically:

[0122]

[0123]

[0124] wherein, The eigenmode function component correlation coefficient corresponding to the jth initial adaptive eigenmode function component; The product of the energy density and the average period corresponding to the jth initial adaptive eigenmode function component; The average value of the product of the energy density and the average period corresponding to the first j-1 initial adaptive eigenmode function components; The energy density corresponding to the jth initial adaptive eigenmode function component; The average period corresponding to the jth initial adaptive eigenmode function component.

[0125] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described device and module can refer to the corresponding process in the foregoing method embodiments, and will not be described here.

[0126] The embodiment of the application further provides a computer device, including a memory and a processor, the memory stores a computer program; the computer program is executed by the processor, so that the processor executes the steps of the switch cabinet structural anomaly detection method according to any one of the above embodiments.

[0127] The embodiment of the application further provides a computer readable storage medium, which stores a computer program / instruction, and the computer program / instruction is executed by a processor to realize the steps of the switch cabinet structural anomaly detection method according to any one of the above embodiments.

[0128] The embodiment of the application further provides a computer program product, which includes a computer program / instruction, and the computer program / instruction is executed by a processor to realize the steps of the switch cabinet structural anomaly detection method according to any one of the above embodiments.

[0129] ​​In several embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other manners. For example, the division of the units is only a logical function division, and there can be another division manner for the actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between the units can be indirect couplings or communication connections through some interfaces, devices or units, and can be in electrical, mechanical or other forms.

[0130] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, can be located in one place, or can be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0131] The above description and the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for detecting structural anomalies in a switch cabinet, characterized in that: include: Acquire an original vibration signal of the switchgear, and preprocess the original vibration signal of the switchgear based on multiple preset white noise sequences, and output a normalized switchgear digital signal and multiple normalized switchgear disturbance signals; Performing first-order mode extraction one by one based on the normalized switch cabinet digital signal and the plurality of normalized switch cabinet disturbance signals based on an empirical mode decomposition algorithm, and outputting a total residual and a plurality of initial adaptive intrinsic mode function components; screening each of the initial adaptive intrinsic mode function components and outputting a plurality of target adaptive intrinsic mode function components; Performing superposition and reconstruction based on the plurality of target adaptive intrinsic mode function components and the total residual, and outputting a switch cabinet noise reduction signal; Structural anomaly detection is performed on the switch cabinet based on the switch cabinet noise reduction signal, and a target detection result is output.

2. The switch cabinet structural anomaly detection method according to claim 1, characterized in that: The preprocessing of the original switchgear vibration signal based on multiple preset white noise sequences to output a normalized switchgear digital signal and multiple normalized switchgear disturbance signals includes: Performing signal conversion on the original vibration signal of the switch cabinet and outputting a switch cabinet digital signal; Normalizing the switch cabinet digital signal and outputting a normalized switch cabinet digital signal; The preset white noise sequences are used to disturb the normalized switch cabinet digital signal respectively to generate a plurality of normalized switch cabinet disturbance signals.

3. The switch cabinet structural anomaly detection method according to claim 1, characterized in that: The empirical mode decomposition algorithm is based on the normalized switch cabinet digital signal and the multiple normalized switch cabinet disturbance signals to perform first-order mode extraction one by one, and output a total residual and multiple initial adaptive intrinsic mode function components, including: Decomposing each of the normalized switchgear disturbance signals into intrinsic mode function components using the empirical mode decomposition algorithm, and outputting the intrinsic mode function components corresponding to each of the normalized switchgear disturbance signals; Averaging the intrinsic mode function components and outputting an initial adaptive intrinsic mode function component; Determining a residual signal according to the normalized switchgear digital signal and the initial adaptive intrinsic mode function component; Determining whether the residual signal is a monotonic function; If so, the energy density corresponding to the initial adaptive intrinsic mode function component is used to calculate the damping weight factor corresponding to the initial adaptive intrinsic mode function component; Calculating a total residual according to the normalized switchgear digital signal, the initial adaptive intrinsic mode function component, and a damping weight factor corresponding to the initial adaptive intrinsic mode function component; If not, performing noise sequence decomposition on each of the normalized switch cabinet disturbance signals, and outputting the noise mode corresponding to each of the normalized switch cabinet disturbance signals; The noise modes corresponding to each of the normalized switch cabinet disturbance signals are superimposed on the residual signal to generate multiple new normalized switch cabinet disturbance signals, and the step of using the empirical mode decomposition algorithm to decompose the intrinsic mode function components of each of the normalized switch cabinet disturbance signals is jumped to execute, and the step of outputting the intrinsic mode function components corresponding to each of the normalized switch cabinet disturbance signals is output until the residual signal is the monotonic function.

4. The switch cabinet structural anomaly detection method according to claim 1, characterized in that: The screening of the initial adaptive intrinsic mode function components and outputting a plurality of target adaptive intrinsic mode function components includes: Calculating the intrinsic mode function component correlation coefficient corresponding to each of the initial adaptive intrinsic mode function components using the energy density and the average period corresponding to each of the initial adaptive intrinsic mode function components; Performing main frequency band spectral entropy calculation on each of the initial adaptive intrinsic mode function components, and outputting the main frequency band spectral entropy corresponding to each of the initial adaptive intrinsic mode function components; Comparing the correlation coefficient of each of the intrinsic mode function components with a preset first threshold value, and comparing the spectral entropy of each of the main frequency bands with a preset second threshold value; An initial adaptive eigenmode function component corresponding to a main frequency band spectral entropy having an eigenmode function component correlation coefficient greater than the preset first threshold and less than the preset second threshold is used as a target adaptive eigenmode function component.

5. The switch cabinet structural anomaly detection method according to claim 3, characterized in that: The calculation formula of the total residual is specifically: ； in, is the total residual; is the normalized switchgear digital signal; is the total order, indicating the total number of times the signal undergoes empirical mode decomposition; is the damping weight factor corresponding to the j-th order initial adaptive eigenmode function component; is the j-th order initial adaptive eigenmode function component.

6. The switch cabinet structural anomaly detection method according to claim 4, characterized in that: The calculation formula of the correlation coefficient of the intrinsic mode function component is specifically: ； ； in, is the correlation coefficient of the eigenmode function component corresponding to the initial adaptive eigenmode function component of the jth order; is the product of the energy density corresponding to the j-th order initial adaptive eigenmode function component and the average period; is the average value of the product of the energy density and the average period corresponding to the initial adaptive eigenmode function components of the first j-1 orders; is the energy density corresponding to the j-th order initial adaptive eigenmode function component; is the average period corresponding to the j-th order initial adaptive eigenmode function component.

7. A switch cabinet structural anomaly detection device, characterized in that: include: An acquisition module is used to acquire an original vibration signal of the switch cabinet, preprocess the original vibration signal of the switch cabinet based on multiple preset white noise sequences, and output a normalized switch cabinet digital signal and multiple normalized switch cabinet disturbance signals; An output module is used to extract first-order modes one by one according to the normalized switch cabinet digital signal and multiple normalized switch cabinet disturbance signals based on an empirical mode decomposition algorithm, and output a total residual and multiple initial adaptive intrinsic mode function components; A screening module, configured to screen the initial adaptive intrinsic mode function components and output a plurality of target adaptive intrinsic mode function components; a reconstruction module, configured to perform superposition reconstruction based on the plurality of target adaptive intrinsic mode function components and the total residual, and output a switch cabinet noise reduction signal; The detection module is used to perform structural anomaly detection on the switch cabinet based on the switch cabinet noise reduction signal and output a target detection result.

8. A computer device, characterized in that: The device comprises a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of the switch cabinet structural abnormality detection method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed, the switch cabinet structural abnormality detection method according to any one of claims 1 to 6 is implemented.

10. A computer program product, characterized in that The computer program product includes a computer program stored on a non-transitory computer-readable storage medium, wherein the computer program includes program instructions, wherein when the program instructions are executed by a computer, the computer is caused to execute the switch cabinet structural abnormality detection method according to any one of claims 1 to 6.

Images