74 results about "Time frequency decomposition" patented technology

The invention relates to the technical field of fault detection, and discloses a diesel generating set fault detection method and system based on deep learning, and the method comprises the steps: obtaining first vibration signal data, and carrying out the time-frequency decomposition, and obtaining a dynamic change feature; de-noising processing is carried out on the dynamic change features to obtain a time-frequency feature sequence; extracting a peak energy distribution data set, and calculating each frequency band entropy value to obtain a frequency band entropy value sequence; classifying the frequency band entropy sequence, determining a random fluctuation reference mode, and separating to obtain an abnormal frequency component; calculating a spectral line spacing and amplitude ratio, obtaining a spectral line feature data set, classifying the spectral line feature data set, and determining a fault classification result; obtaining current second vibration signal data, performing similarity calculation on the current second vibration signal data and a pre-established normal mode library, and outputting a fault feature vector; and verifying the fault feature vector to obtain a final fault detection result. According to the method, closed-loop diagnosis from signal acquisition to fault classification can be realized, and the fault detection precision of the diesel generating set is improved.

The invention provides a power transmission and distribution line suspended foreign matter detection method and system based on image processing and deep learning. Wherein a historical point cloud set on the surface of the power transmission tower is acquired, a deformation gradient map is generated through dynamic registration of curvature differences of adjacent point clouds, a super-threshold region is marked as a deformation correlation region, and foreign matter contour parameters are extracted; performing time-frequency decomposition on the node strain signal, retaining a steady-state component, and mapping the steady-state component to an association region to form a strain map; calculating the displacement offset of the deformation association interval to construct a deformation propagation network, and extracting load over-threshold nodes to construct a stress diffusion network; and superposing the network, fusing the displacement and the load value in the superposition area to generate a risk index, and outputting early warning in combination with the foreign matter coordinates and the risk index. According to the technical scheme provided by the invention, the cross-modal dynamic analysis of deformation propagation and stress diffusion paths is fused, and the early warning efficiency of foreign matter attachment and structure abnormity risks of the power transmission tower is improved.

The invention provides a power transmission network equipment insulation aging state evaluation and life prediction method and system, and relates to the technical field of power equipment state monitoring, and the method comprises the steps: obtaining multi-dimensional operation monitoring data, extracting features through multi-scale time-frequency decomposition, and carrying out the orthogonal transformation to obtain a decoupling feature vector; calculating a mahalanobis distance to obtain an aging state deviation metric value; constructing a stress accumulation damage factor to perform nonlinear modulation on the degradation rate; an aging state trajectory is predicted based on the modified degradation rate and a remaining life is determined. The method can accurately evaluate the insulation aging state and predict the residual life of the equipment, and improves the evaluation precision and reliability.

The invention, which relates to the technical field of education recommendation, discloses a student portrait-driven education agent recommendation system based on a cognitive diagnosis map, comprising a time-frequency decomposition processing module, a weight rhythm matching module, a time sequence coupling damping module, a dynamic difference readjustment module and a time-frequency domain self-balancing control module. And the time-frequency decomposition processing module is used for establishing a time-frequency decomposition processing layer based on drifting characteristics of student portrait parameters in a time dimension, and performing energy distribution analysis on multi-source dynamic data streams from learning behaviors, test performance and emotion feedback according to time slices. Through time-frequency decomposition and non-resonance rhythm matching, dynamic coordination of weight adjustment and student portrait drifting is realized, characteristic fluctuation amplification is prevented, and the stability of a learning path is guaranteed; and a dynamic closed loop is formed through time sequence coupling damping and time-frequency domain self-balancing control, stable convergence of cognitive features is realized, and the accuracy and continuity of educational agent recommendation are improved.

The invention provides a postoperative bleeding risk monitoring method and system for a hepatobiliary patient. The method comprises the following steps: collecting real-time multi-modal data; constructing an LSTM-CNN hybrid model by using the time-frequency decomposition features, and obtaining a local tissue hypoxia index and a vasomotor function anomaly probability; the low-frequency impedance change rate and the albumin level are fused through a random forest algorithm, and the ascites occurrence probability and the effusion amount predicted value are obtained; and generating a bleeding point positioning coordinate and a thermodynamic diagram risk grade. And calculating a comprehensive bleeding risk probability and positioning a bleeding area. And generating graded early warning signals and recommending personalized treatment schemes or nursing suggestions. According to the invention, the LSTM-CNN hybrid model, the random forest algorithm and the three-dimensional convolutional neural network are adopted to deeply extract different data features, so that the limitation of single index evaluation is avoided. A causal relationship model is established through the Bayesian network, and the comprehensive risk probability calculation preciseness is improved; the early warning threshold is dynamically adjusted by combining the individual characteristics of the patient, and the traditional problem of easy misjudgment is solved.

The invention belongs to the technical field of equipment state monitoring, and discloses a relay state monitoring and diagnosis method and system. Vibration time domain signals of all monitoring points in the working process of a relay are collected, a time delay characteristic matrix is constructed through segmented cross-correlation operation, and a near-field sound source candidate area is analyzed and recognized in combination with the spatial divergence of a time delay gradient vector field. A sound-vibration coupling propagation model is introduced, near-field sound source excitation and structure conduction vibration components are distinguished, multi-component signal decoupling is realized through spatial constraint and frequency domain sparsity constraint, and a relay body vibration signal is effectively separated. The method comprises the following steps: performing time-frequency decomposition on a body vibration signal, extracting a transient impact feature of a contact action and a steady-state vibration feature of coil excitation, constructing a multi-dimensional state feature vector, and realizing state diagnosis through similarity matching with a preset feature library. According to the invention, the early fault detection sensitivity and diagnosis accuracy of the relay are improved, and a reliable basis is provided for preventive maintenance.

The invention relates to the technical field of audio signal processing, and discloses an industrial noise identification method and system, and the method comprises the steps: obtaining an audio signal, and carrying out the time-frequency decomposition, and obtaining preliminary decomposition data; classifying the coupling features of the preliminary decomposition data to obtain a coupling region boundary value; if the boundary value of the coupling region exceeds a preset boundary threshold value, performing filtering processing to obtain separated noise spectrum data; calculating an overlapping proportion of the separated noise spectrum data and a preset target sound; if the overlapping proportion causes extraction deviation, extracting dynamic change characteristics of the separated noise spectrum data and fusing the dynamic change characteristics into a spectrum extraction process to obtain an accurate noise spectrum; calculating an intensity value according to the accurate noise spectrum, and determining an intensity evaluation result; and comparing the intensity evaluation result with a preset pollution threshold, calculating a noise pollution degree and outputting a noise identification result. The method is suitable for industrial environment noise monitoring, and can accurately extract the noise spectrum.

The invention discloses a porosity real-time prediction method and system based on alkane gas and carbon isotope characteristics thereof and a storage medium, relates to the technical field of oil and gas reservoir development, and aims to solve the problem of insufficient precision in a heterogeneous reservoir due to neglect of dynamic characteristics of alkane gas carbon isotope in an existing method. The method is characterized by comprising the following steps: S100, collecting alkane gas concentration and carbon isotope data thereof in real time, and performing time-frequency decomposition on the data to obtain characteristic data on different time and frequency scales; s200, performing multi-scale feature extraction on the data to obtain multi-scale features; s300, performing feature fusion on the multi-scale features; s400, constructing a GNN network model, introducing an adaptive attention mechanism, residual connection, jump aggregation and time sequence dynamic fusion, constructing a graph structure by taking a reservoir unit as a node, and mapping a fusion feature to a graph node attribute; and S500, constructing a reservoir space structure based on the GNN network model, and combining physical constraint optimization to realize real-time prediction of the porosity.

The invention provides a fault detection and early warning method and system for an insulating layer of an oil-immersed high-voltage wiring harness, and relates to the technical field of power equipment detection, and the method comprises the steps: obtaining a multi-physical field coupling signal, carrying out the time-frequency decomposition, obtaining degradation feature distribution, extracting a feature response amplitude, calculating a gradient change rate, and recognizing an abnormal feature cluster; deducing residual life distribution of the insulating medium according to the degradation rate and generating a risk index; according to the risk index, the early warning level and the intervention time window are determined, early detection and early warning of the fault of the insulating layer of the oil-immersed high-voltage wire harness can be realized, and the operation safety of power equipment is improved.

The application relates to the technical field of analytical chemistry detection, and specifically discloses a rare earth ion chromatographic online analysis and detection method and system, which adopts a chromatography-mass spectrometry combined system, realizes multi-scale time-frequency decomposition of signals by constructing an adaptive wavelet base function library, accurately identifies and classifies interference types in combination with a convolutional neural network and vacuum degree coupling analysis; a U-Net generator with an attention mechanism and a multi-scale discriminator are designed for signal repair aiming at repairable interference; a state transition model containing a mass-to-charge ratio database is established, empirical mode decomposition and adaptive filtering technology are used to realize dynamic calibration of a mass axis; quantitative accuracy is ensured through double internal standard correction and a triple verification mechanism; when unrepairable interference is detected, a hierarchical self-checking program is started, and fault diagnosis is carried out in combination with spectrum fingerprint analysis.

The invention discloses a converter transformer fault diagnosis method and system based on voiceprints and thermal signals, and the method comprises the steps: carrying out the time-frequency decomposition of each voiceprint original signal data and temperature thermal signal data of a converter transformer, and obtaining a sound time-frequency matrix and a temperature time-frequency matrix; the kurtosis and the sound pulse duration of the time-frequency matrix of the sound are calculated to serve as basic time-frequency characteristics of the sound; calculating an energy gradient and an entropy change rate of the time-frequency matrix of the temperature as basic time-frequency characteristics; coupling characteristics of the time-frequency matrixes of the sound and the temperature in the same frequency band are extracted; performing feature screening on all basic time-frequency features and coupling features by calculating a maximum mutual information coefficient, and outputting an optimal feature subset; and constructing a converter transformer fault diagnosis-oriented identification model fusing ensemble learning and random forest, and training the identification model according to the optimal feature subsets in different operation states to perform fault diagnosis. The fault risk sensing capability of the converter transformer can be effectively improved.

This invention discloses a bearing remaining life prediction method based on frequency domain degradation perception. It constructs an integrated prediction system through a time-frequency decomposition dual-branch feature extraction module, a frequency domain degradation perception weight allocation module, and a life prediction loss function optimization module. First, the time-frequency decomposition dual-branch feature extraction module decomposes the original time series into high and low frequencies and inputs it into an sLSTM and mLSTM dual-branch structure for feature extraction. Second, the frequency domain degradation perception weight allocation module performs degradation perception processing on the fused features, generating dynamic weights and modulating the features to obtain the final fused features, which are then mapped to the prediction space. Finally, the life prediction loss function optimization module calculates the loss, optimizes the model output, and obtains the bearing remaining life prediction result. This invention comprehensively covers the full-stage features of bearings from healthy to severely degraded, improves the sensitivity of early fault detection, and still possesses excellent generalization and robustness under complex operating conditions.

The invention discloses a bridge beam end track state evaluation method, system and equipment based on time-frequency decomposition and Shannon entropy measurement, and a medium, and the method comprises the steps: S1, obtaining track irregularity historical dynamic detection data of a long-span railway bridge, carrying out the alignment preprocessing, and building a beam end track irregularity data set; s2, using an improved VMD algorithm to decompose the beam end track irregularity data set to obtain a plurality of IMF components; s3, calculating the power spectrum density of each IMF component, and determining the maximum energy concentration wavelength L1 of each IMF component; s4, measuring the information density of different track irregularity data which can be accommodated by different fixed window lengths by using Shannon entropy, and determining the maximum information window length L2; s5, calculating a relative deviation between L1 and L2, and determining a fixed calculation window length L3; and S6, calculating a beam end track quality index based on the L3, and evaluating the smoothness state of the current bridge beam end track. According to the method, the time-frequency decomposition theory and the information entropy theory are combined, and the irregularity state of the long-span railway bridge beam end track is evaluated more accurately.

The invention discloses an AI large model-based atmospheric water vapor content prediction method, system and device. The method comprises the following steps of: obtaining atmospheric water vapor content data and related meteorological variable data; performing time-frequency decomposition on the atmospheric water vapor content data and the related meteorological variable data by using wavelet transform, and extracting a multi-scale feature subsequence; constructing an AI large model based on a long short-term memory network, and inputting the extracted multi-scale feature subsequences into the large model for multi-factor nonlinear training; and performing space-time prediction on the future atmospheric water vapor content by using the trained large model to obtain a prediction result. Through the WT-LSTM mixed large model, the interannual prediction R2 is improved by 36%, the RMSE is reduced by 56.52%, the nonlinear bottleneck of a traditional model is solved, and high-precision space-time prediction is provided.

The application provides a multi-directional stress prediction method, device and equipment in a milling cutter milling process and a storage medium. It relates to the field of numerical control machine tool processing digital twin technology. The method comprises: obtaining small sample experimental data based on an orthogonal test method, time-frequency decomposition of the milling force test signal, and extraction of multi-dimensional features of the milling force dynamic characteristics; analyzing the correlation between the processing parameters and the features, screening the key features, establishing a physical mapping model of the process parameters to the key features and solving the cutting coefficients; constructing a time-varying signal prediction model based on a recurrent neural network, predicting the multi-directional dynamic milling force of the milling cutter with the key features in the small sample test data; and based on the cutting coefficient, designing an adaptive filter to post-process and optimize the predicted signal and inverse normalize it, and output the final prediction value. Based on small sample data, the application can accurately predict the multi-directional dynamic stress of the milling cutter only with the process parameters, and effectively improve the virtual-real mapping and dynamic optimization capability of the processing process.

The application discloses a CNN-LAE-LSTM hybrid network centrifugal pump cavitation diagnosis method and device, comprising: collecting different degree cavitation data, including cavitation pressure signal and cavitation visual image; performing Kalman filtering and modal time-frequency decomposition on the cavitation pressure signal; performing gray scale processing on the cavitation visual image; combining the processed cavitation data to form a training set and a test set; constructing a CNN-LAE-LSTM-based centrifugal pump cavitation diagnosis model, wherein the CNN is used for extracting local key features of the cavitation pressure signal and the cavitation visual image; the LAE is used for feature dimension reduction, redundancy elimination and noise suppression; the LSTM is used for capturing signal time sequence correlation and adapting to pump operation dynamic characteristics; the training set is used for training the centrifugal pump cavitation diagnosis model, and the test set is used for verifying the model accuracy; and the to-be-tested data is input into the trained centrifugal pump cavitation diagnosis model to obtain a fault diagnosis result.

The invention provides a method and device for predicting multidirectional stress in the milling process of a milling cutter, equipment and a storage medium. Relates to the technical field of numerical control machine tool machining digital twinning. The method comprises the following steps: acquiring small sample experimental data based on an orthogonal test method, performing time-frequency decomposition on a milling force test signal, and extracting multi-dimensional characteristics of milling force dynamic characteristics; analyzing the correlation between the processing technological parameters and the features, screening key features, establishing a physical mapping model from the technological parameters to the key features, and solving a cutting coefficient; constructing a time-varying signal prediction model based on a recurrent neural network, and predicting the multidirectional dynamic milling force of the milling cutter under small sample test data by using the key features; and designing an adaptive filter based on cutting coefficient constraint to perform post-processing optimization and inverse normalization on the predicted signal, and outputting a final predicted value. According to the method, based on the small sample data, the multi-directional dynamic stress of the milling cutter is predicted in a high-precision mode only through the technological parameters, and the virtual-real mapping and dynamic optimization capacity in the machining process is effectively improved.

This invention discloses an activity pattern recognition method and system based on wavelet transform and CNN / Transformer. First, wavelet transform is used to perform multi-scale time-frequency decomposition on the original signal to enhance the perception of subtle local features and high-frequency components. Then, convolution operations are used to further extract and fuse effective features in the time-frequency domain. Finally, leveraging the self-attention mechanism of Transformer, global dependencies are captured in the enhanced feature sequence, thereby achieving efficient collaborative modeling of local information and global context. This invention overcomes the insensitivity of Transformer models to local details and high-frequency information, effectively avoiding the neglect or smoothing of crucial subtle action patterns and transiently changing high-frequency signals during recognition, significantly improving the model's recognition accuracy and enhancing the reliability and interpretability of the entire system in practical applications.

The application provides a brain-computer interface cognitive function decline monitoring system, method, device and medium, which can be applied to the field of electroencephalogram signal analysis. The brain-computer interface cognitive function decline monitoring system comprises a signal monitoring module, a signal processing module, a result processing module and a medium. The signal monitoring module is used for collecting multi-channel electroencephalogram signals in multiple time periods. The signal processing module is used for performing time-frequency decomposition and global field power determination on the multi-channel electroencephalogram signals. The probability of the multi-channel electroencephalogram signals belonging to each microstate mode is determined, and a microstate feature vector is generated. According to the phase locking value of the electroencephalogram signals between nodes and the similarity between the microstate feature vectors corresponding to the nodes, a topological heterogeneous brain network is generated. The topological heterogeneous brain network is subjected to graph feature extraction and state classification to obtain an analysis result. The result processing module is used for generating a brain load change trend according to multiple analysis results, and generating an adjustment suggestion based on the brain load change trend.

The application discloses a kind of acoustic signal feature extraction methods of gas insulated equipment, comprising: the sound signal collected is carried out noise reduction processing, and the mean and variance of signal are normalized processing;Using short-time Fourier transform STFT and wavelet transform WT, the sound signal of pre-processing is carried out time-frequency decomposition;Comprehensive MFCC, AWPC adaptive wavelet packet coefficient and instantaneous frequency feature IF are carried out comprehensive extraction to the feature of sound signal;The voiceprint feature extracted is carried out feature fusion, and MFCC, AWPC and IF feature are spliced, form a high-dimensional feature vector, then using principal component analysis PCA is carried out dimension reduction;The feature after dimension reduction is standardized, and correlation analysis method is further used to screen features.The application can extract effective voiceprint features from the sound signal generated by gas insulated equipment defects and reduce the influence of noise on detection, thereby improving the accuracy of detection.

The invention discloses a foam cleaning agent proportion control method based on online parameter regulation and control, and relates to the technical field of fine chemical process automatic control, and the method comprises the following steps: establishing a unified time base line in a feeding loop, and generating a phase reference field; low-amplitude calibration pulses are injected into all the feeding pumps, flow response is collected, inherent pulse characteristics are extracted, and a zero phase point is determined; and on the basis of the unified time base line and the zero phase point, constructing a pulsation tensor, and performing time-frequency decomposition on flow, pressure and valve position signals of each feeding pump to generate a synchronous phase track. According to the method, the time base line is constructed, the pump pulsation characteristics and the synchronous phase track are extracted, accurate recognition of the pump set interaction state is achieved, and the resonance event is judged in combination with disturbance verification; after high-risk judgment, a resonance path is cut off in a high-frequency fragmentation and time staggering mode, the matching stability is guaranteed, precise, automatic and intelligent control over the cleaning agent preparation process is achieved, and the product consistency and the system safety are improved.

The application discloses a power grid oscillation risk scene sample generation method, system, device and medium, and particularly relates to the technical field of oscillation scene generation, and the technical points are as follows: a time-frequency fusion generative adversarial network is constructed, the time-frequency fusion generative adversarial network comprises at least one generator and at least one discriminator; wherein the generator comprises a time-frequency decomposition module, a time domain generation module, a frequency domain generation module and a U-Net module; a random noise vector is input into the time-frequency decomposition module to obtain a time domain signal and a frequency domain signal; the time domain signal is input into the time domain generation module, and the frequency domain signal is input into the frequency domain generation module to generate a time domain waveform and a frequency domain feature respectively; the time domain waveform and the frequency domain feature are input into the U-Net module to generate a time-frequency fusion power grid oscillation signal; the generated power grid oscillation signal and an original real oscillation signal are input into the discriminator, and the power grid oscillation signal is iteratively optimized and trained in combination with a loss function to obtain a power grid oscillation scene sample.

A time-frequency aligned EEG signal analysis method, apparatus, device, and medium are disclosed, aiming to consider and reduce the temporal and frequency distribution differences of EEG signals, thereby more effectively improving the model's generalization ability across different subjects or experimental sessions. The method involves time-frequency decomposition of the EEG signal, calculating global statistics in each frequency band, and performing whitening transformation to calibrate the signal. Finally, the multi-channel EEG signal is reconstructed to obtain a time- and frequency-aligned multi-channel EEG signal. This invention's time-frequency alignment process comprehensively reduces the distribution differences between EEG data from different sources in both time and frequency dimensions. Compared to methods that only align the time domain, it more effectively improves the model's performance on new user or new session data, thus more effectively enhancing the model's generalization ability across different subjects or experimental sessions.

The invention discloses a UWB positioning precision optimization method based on IWPT-BILSTM, and relates to the field of UWB positioning. A lightweight attention module is introduced to improve a traditional WPT, the improved WPT (IWPT) has a self-adaptive sub-band screening and weighted reconstruction mechanism, accurate time-frequency decomposition and effective feature extraction are performed on original UWB signals by using the IWPT, noise in data is weakened to reserve more effective time-frequency signals, and the UWB signals are subjected to time-frequency decomposition and effective feature extraction. And then low-noise and high-validity data are sent to a BI-LSTM network for training, the BI-LSTM network is used for carrying out time sequence correlation and dynamic error correction on effective characteristic signals after IWPT processing, and finally nonlinear time sequence characteristics generated in the time sequence correlation process are accurately mapped into a UWB positioning result. By adopting the method, the positioning error can be obviously reduced, the positioning robustness and generalization ability in a complex environment can be improved, and the method has a relatively good engineering application value.

The invention discloses an industrial flexible load state space model parameter prediction method and system, and the method comprises the steps: constructing a time sequence segmentation index storage library based on load state segmentation, calculating a heterogeneous dynamic index through phase-space reconstruction, constructing a time-frequency recurrence spectrum through time-frequency analysis, calculating a frequency domain recurrence index, and carrying out the prediction of the parameters of an industrial flexible load state space model. And adaptively selecting a feature enhancement mode of an iterative feature screening and adversarial generation network or a dynamic time-frequency decomposition and variational coding network based on the comprehensive feature coefficient, and finally training a deep learning model to realize parameter prediction. Through adaptive feature processing driven by dynamic complexity, the accuracy and reliability of industrial load parameter analysis are improved, and the operation efficiency and stability of industrial equipment are optimized.

The invention relates to the field of reaction kettle stirring blade arrangement optimization, and discloses a reaction kettle stirring blade arrangement optimization method and system based on CFD fluid simulation, and the method comprises the steps: obtaining the structure and working condition data of a reaction kettle, carrying out the long-time-scale transient CFD simulation, separating a low-frequency flow field time sequence, obtaining a precession grid field, constructing a precession voxel octree, and carrying out the long-time-scale transient CFD simulation; paddle layer local vortex structure segmentation and time-frequency decomposition are executed based on a precession voxel octree to obtain a paddle layer local vortex shedding mode set, an interlayer coherent vortex atlas tree is constructed, and a stirring paddle arrangement multi-target optimization model is established based on the interlayer coherent vortex atlas tree to solve and obtain a candidate paddle arrangement solution set. Generating a stirring paddle mounting scheme, and performing high-fidelity verification and simulation in combination with a result to form an arrangement correction rule; according to the method, the problem that low-frequency large-scale circulation precession cannot be captured by the traditional steady-state CFD is solved, the arrangement optimization of the stirring blades is realized, and the batch stability and the product consistency in the reaction process are improved.

The invention discloses a lightning arrester resistive current prediction method and system, and relates to the technical field of lightning arrester current prediction, the system is composed of a plurality of function modules, and the system comprises a multi-source data acquisition module used for acquiring lightning arrester leakage current, environmental parameters and electrical parameters; wherein the environmental parameters comprise temperature, relative humidity and atmospheric pressure; the data processing and feature mining module is used for performing time-frequency decomposition on the leakage current signal by adopting an improved CEEMDAN algorithm, constructing a multi-physical field feature correlation matrix, acquiring the environmental parameters and the electrical parameters, screening out IMF components and environmental factors which are strongly coupled with resistive current, and verifying, correcting and outputting a high-quality feature set through a physical constraint isolated forest; according to the improved CEEMDAN algorithm, CEEMDAN is introduced into empirical mode decomposition, and adaptive Gaussian white noise is added into an original leakage current signal; and the hybrid model construction and optimization module is used for constructing a hybrid architecture to perform physical dimension pre-mapping on the input high-quality feature set.

The invention belongs to the field of fault diagnosis, and particularly relates to a bearing single-source-domain generalization fault diagnosis method based on wavelet adaptive causal decoupling, and the method comprises the steps: carrying out the time domain decomposition and adaptive feature extraction of a source domain sample based on a training set and a test set, and obtaining a global feature; in the second convolutional neural network, introducing a causal decoupling strategy based on a binary mask, and displaying and decoupling the extracted global features into causal features and non-causal features; and performing second convolutional neural network training, establishing a fault classifier, and determining causal purification loss based on the fault classifier. In the feature extraction stage, multi-scale time-frequency decomposition of mechanical vibration signals is achieved by adaptively optimizing wavelet filter parameters, low-frequency stable information and high-frequency impact features can be effectively separated, the stability and discrimination of feature expression can still be kept even in a strong-noise and multi-component coupling environment, and the method is suitable for large-scale popularization and application. Therefore, the accuracy and robustness of fault feature extraction are remarkably improved.

The invention provides a high-frequency transient response island detection method and device, equipment and a storage medium, and belongs to the field of island detection. The method comprises the following steps: acquiring high-frequency transient voltage and current waveform data, and performing time-frequency decomposition on the data to obtain a time-frequency coefficient matrix containing voltage and current wavelet packet coefficients; calculating a transient energy characteristic and a high-frequency transient impedance angle characteristic based on the matrix; inputting the transient energy characteristics into an energy identification model to obtain an energy classification confidence coefficient; inputting the high-frequency transient impedance angle characteristics into a transient impedance angle identification model to obtain an impedance angle classification confidence coefficient; and the decision model fuses the two confidence coefficients to generate a final island state judgment result and a power grid instability early warning signal. According to the invention, the detection blind area under power balance is overcome by using dual features. Multi-dimensional features are fused, and maloperation and refusal operation are avoided; and a passive detection method does not need to inject disturbance, so that the electric energy quality is guaranteed, and the safety and reliability of a power grid are improved.

The invention provides a sound source positioning method and system based on a multi-microphone array, and the method comprises the steps: carrying out the preprocessing of a first sound source signal collected by a microphone array, and outputting a second sound source signal and an acoustic physical parameter; performing time-frequency decomposition on the second sound source signal to obtain a plurality of first time-frequency regions; performing acoustic coherence analysis on the first time-frequency region according to the acoustic physical parameters to obtain a time-frequency coherence matrix; detecting a second time-frequency region dominated by a single source from the time-frequency coherence matrix based on sound wave multipath and diffraction characteristics, and constructing an output single-source time-frequency point index according to the second time-frequency region; generating a plurality of direction-of-arrival estimation values according to a preset frequency domain graph signal and the single-source time-frequency point index; and performing clustering and weighted fusion on the plurality of direction-of-arrival estimated values to obtain direction-of-arrival information of the sound source.