982 results about "Signal extraction" patented technology

The invention discloses a multi-sensor fusion heat pump full life cycle AI maintenance early warning system, and relates to the technical field of new energy utilization, and the early warning system comprises a data collection module which obtains operation parameters in a heat pump full life cycle based on a sensor array, and constructs a data set after preprocessing the parameters; the operation parameters comprise temperature, pressure, flow and micro vibration; the data fusion module is used for extracting trend correlation characteristics and parameter coupling characteristics from temperature, pressure and flow parameters by adopting a dynamic sliding window adaptive to a working condition, and preserving core nonlinear information through KPCA dimension reduction; the micro-vibration signal extraction comprises frequency domain and time domain features. According to the method, features are extracted through a working condition adaptive dynamic sliding window, then through cross-space mapping and a life cycle-working condition double-attention mechanism, the analysis and early warning module depends on a core feature mapping library and a two-dimensional dynamic baseline, through instantaneous and accumulated deviation double judgment, abnormal accurate recognition and stage division are achieved, and early warning perspectiveness is high.

The invention provides a wind turbine generator voiceprint fault recognition method, and relates to the technical field of wind turbine generator state monitoring and fault diagnosis, and the method comprises the steps: carrying out the noise reduction of an original audio signal through variational mode decomposition, screening a target mode of which the frequency, energy and kurtosis accord with features, and reconstructing the signal; extracting a Mel frequency cepstrum coefficient and a sensing noise robust coefficient, and generating multi-dimensional voiceprint data in combination with statistical characteristics such as a frequency spectrum gravity center, a spectrum entropy, energy, kurtosis and a zero-crossing rate; constructing a support set based on the prototype network, realizing small sample fault classification by calculating the Euclidean distance between the feature vector and the prototype vector, and outputting a preliminary result; judging whether the voiceprint is abnormal according to a preset threshold value, if so, storing the voiceprint into a dynamic abnormal voiceprint knowledge base; frequently occurring abnormal samples are manually labeled and added into a support set, the prototype network is retrained to update the model, and continuous optimization of the fault recognition capability is achieved.

The invention relates to the technical field of biomedical engineering and computer vision, in particular to a non-contact physiological signal extraction method and system based on frequency self-adaption and illumination noise perception.The method comprises the following steps of multi-mode video stream collection and spatio-temporal data preprocessing, illumination-noise perception mask generation and feature filtering, multi-mode video stream collection and spatio-temporal data preprocessing, illumination-noise perception mask generation and feature filtering, and non-contact physiological signal extraction. Frequency adaptive gating and frequency domain feature enhancement, depth time attention feature re-calibration, physiological signal regression and closed loop optimization; the method has the beneficial effects that a lightweight end-to-end deep learning network architecture is constructed by systematically fusing three core modules of illumination-noise perception mask, frequency adaptive gating and depth time attention, and the defects that a traditional physical model depends on artificial prior and is poor in anti-interference performance and high in reliability are overcome. And the one-sidedness caused by high calculation complexity and difficulty in distinguishing the signal and noise of the existing deep learning model is avoided, and the weak physiological signal can be recovered from the face video more accurately and robustly.

The invention relates to the technical field of cable positioning, in particular to a one-way grounding fault positioning method based on a signal injection method. The method comprises the following steps: installing an impedance detection unit at a grounding knife switch of a transformer substation, and injecting a controllable AC signal into a power transmission line through a built-in injection coil to enable the line to be in an open circuit state; collecting the injected current signal and performing frequency enhancement to extract the amplitude and phase of a target frequency component; zero-sequence impedance is calculated based on the injection voltage and current, and the zero-sequence impedance is compared with a preset impedance model to judge single-phase grounding; and when a fault is detected, determining a fault distance, a tower position and a grounding phase according to an impedance difference matching result, thereby realizing accurate positioning. According to the invention, by injecting the controllable AC signal into the power transmission line and combining frequency enhancement, phase self-calibration and impedance difference matching analysis, high-sensitivity detection, high-stability signal extraction and high-precision positioning of the single-phase earth fault are realized.

The invention relates to the technical field of safety monitoring, and provides an oil and gas pipeline leakage monitoring method and system, a medium and equipment, and the method comprises the steps: obtaining a leakage vibration signal of each measurement point of an oil and gas pipeline, and extracting a time domain feature, a frequency domain feature, a spatial-temporal feature and an initial distance from a leakage point to a starting end of an optical fiber; performing cross-correlation analysis on the leakage vibration signals of the adjacent measuring points to obtain a propagation time difference and a cross-correlation coefficient, and calculating a correction distance from the leakage point to the starting end of the optical fiber; if the leakage point is in the spiral section, calibrating the correction distance to obtain a calibration distance; calculating the difference value between the initial distance from the leakage point to the optical fiber starting end and the corrected distance to obtain a preliminary positioning error value; and based on the time domain feature, the frequency domain feature and the spatial-temporal feature, in combination with a cross correlation coefficient, a calibration distance, an amplitude enhancement coefficient and a preliminary positioning error value, obtaining a leakage medium type through a medium identification model. The method gives consideration to the positioning efficiency and precision, and greatly improves the accuracy and robustness of a leakage medium type recognition result.

The invention provides a method for extracting electrical characteristic signals of electrical equipment in a complex working condition environment, and belongs to the technical field of electrical equipment detection.The method comprises the steps that a multi-channel ultrasonic sensor array is arranged to collect partial discharge signals, background noise is eliminated through adaptive noise cancellation processing, and an ultra-sparse frequency band energy distribution vector is constructed; the method comprises the following steps: calling a self-adaptive time-frequency analysis model to extract instantaneous frequency, amplitude and phase parameters to form a micro-hour-frequency characteristic matrix, separating independent source signals through independent component analysis, calculating a kurtosis value and a skewness value, fusing multi-domain characteristics to construct a transient stationary comprehensive characteristic vector, matching with a standard discharge characteristic vector library to identify the discharge type and intensity, and calculating the discharge intensity. A corresponding prediction algorithm is selected according to the discharge mode, multi-parameter coupling optimization adjustment is started under a certain condition, an electrical characteristic signal description vector is finally constructed, and the technical problem that the partial discharge signal of the power equipment is difficult to accurately extract under a complex working condition environment is solved.

The invention relates to the technical field of information, and discloses a radar dynamic anti-interference method and system based on interference source positioning. The method comprises the following steps: acquiring current signal data and a historical signal sequence, and determining an initial deviation value; extracting a historical deviation sequence according to the initial deviation value to obtain a deviation change trend vector; calculating a trend slope and a fluctuation amplitude, and determining an adjustment trigger signal; extracting feature interaction influence according to the adjustment trigger signal, and determining a weight coefficient update increment; iterative optimization is carried out in combination with the convergence rate parameter, and an optimized convergence rate value is obtained; adjusting deviation compensation model parameters according to the convergence speed value, and outputting a deviation compensation result when the compensation residual error is lower than a preset threshold value; and carrying out positioning calculation according to a deviation compensation result, and determining a corrected positioning coordinate. The model is dynamically optimized through technologies such as a sliding window and gradient descent, the problem of positioning deviation caused by complex environment signal interference is solved, and the positioning precision and the response speed are improved.

The invention relates to the field of gas leakage identification, and discloses a gas leakage identification method and system based on sound localization, a medium and equipment, and the method comprises the steps: collecting a leakage sound wave signal through a microphone array, and extracting acoustic features in a complex background; simulating a diffusion process of gas in a turbulence environment after gas leakage through an established leakage gas convection-diffusion model, and combining gas concentration distribution in the simulated diffusion process with acoustic characteristics to predict sound source localization at a leakage position; modeling a sound source localization search process as POMDP, and iteratively updating a confidence state of a sound source position through a particle filter; spatial distribution features are extracted from the confidence state through DBSCAN clustering to serve as input of the LSTM-DQN network, spatial and temporal features are fused through the constructed LSTM-DQN network, cross-scene generalization is achieved through transfer learning, and sound source coordinates are dynamically optimized and output through the Q network. According to the invention, the positioning accuracy and real-time performance in a complex environment are improved.

The invention belongs to the technical field of sewage treatment and membrane separation, and provides an in-situ nondestructive membrane pollution detection system and method based on an ultrasonic transmission signal, and the system comprises an ultrasonic excitation module which is used for generating a sine wave excitation signal which penetrates through a membrane body to excite the ultrasonic transmission signal for pollution detection; the signal processing module is used for continuously collecting and preprocessing ultrasonic transmission signals of the clean membrane body and the membrane bodies with different pollution degrees, extracting ultrasonic transmission signal characteristics under different pollution degrees, and establishing a pollution thickness-signal characteristic mapping model; wherein the ultrasonic transmission signal features comprise a time-frequency domain energy gravity center feature and a frequency attenuation spectrum slope feature; and the pollution detection module is used for judging the pollution stage of the membrane body to be detected based on the pollution thickness-signal feature mapping model, and completing in-situ nondestructive membrane pollution detection. According to the technical scheme, the sludge pollution state in the ceramic membrane and flat membrane filtering process can be monitored in real time, and early warning and pollution dynamic tracking are achieved.

The invention provides a self-adaptive calibration system for an automobile fuel evaporation pressure sensor, which relates to the technical field of sensor control, and comprises a processing module used for collecting original pressure data of the fuel evaporation pressure sensor, temperature data of an engine compartment, environment humidity data and electromagnetic interference intensity data of surrounding vehicles in real time, filtering the original pressure data to obtain a processed pressure signal; and the extraction module is used for dividing the processed pressure signal into a first pressure section and a second pressure section according to a time sequence, and respectively selecting pressure data at three different moments in the second pressure section as a first group of characteristic pressure data set and a second group of characteristic pressure data set. According to the invention, the compensation amount is dynamically generated and optimized to correct the response delay of the sensor, and the calibration parameters are adaptively adjusted in combination with the electromagnetic interference intensity of surrounding vehicles, so that the accurate measurement of the fuel evaporation pressure is finally realized.

The invention provides a non-contact pressure and fatigue state identification method based on double-indication joint reasoning, and relates to the field of psychological and physiological analysis, and the method comprises the steps: collecting a face visible light video and a first EDA signal; extracting an rPPG signal and a first facial behavior indication sequence; processing the signal amplitudes of the first EDA signal and the first facial behavior indication sequence to obtain a target EDA signal and a target behavior indication sequence; inputting the rPPG signal and the target behavior indication sequence into a fatigue detection model to obtain fatigue information; inputting the rPPG signal and the target EDA signal into a psychological stress detection model to obtain psychological stress information; and synchronously outputting and displaying the fatigue information and the psychological pressure information based on a preset signal waveform comparison display area. The psychological stress state and the fatigue state are analyzed at the same time, double indications are used for comprehensive consideration, cross-scene state fluctuation is identified, and the precision and efficiency of psychological stress state and fatigue state analysis can be effectively improved.

The invention relates to the technical field of image processing and biological signal detection, discloses a facial physiological detection method and system based on signal quality driven ROI selection, and aims to solve the problem of signal degradation of a traditional fixed geometric ROI in a complex scene. The method comprises the following steps: collecting a user face video stream through a camera and preprocessing the user face video stream; detecting a face bounding box and dividing the face bounding box into a plurality of sub-regions; calculating the signal-to-noise ratio, the periodic intensity and the motion artifact interference degree of each sub-region; screening an optimal sub-region according to a weighted fusion formula to generate a dynamic ROI mask; extracting a pure rPPG signal from the dynamic ROI mask coverage area; detrending and band-pass filtering are carried out on the rPPG signals, and physiological parameters such as the heart rate and the blood oxygen saturation degree are extracted. The system comprises a face video acquisition module, a face region positioning and segmentation module, a signal quality evaluation module, a dynamic ROI selection module, an rPPG signal extraction module and a physiological parameter estimation module. According to the technical scheme, signal degradation caused by local shielding, illumination abrupt change or attitude offset can be effectively avoided, the signal-to-noise ratio and the stability of the rPPG signal are remarkably improved, and the universality and the robustness of the method are enhanced.

The invention relates to the technical field of hydrological monitoring, and discloses a regional groundwater reserve change dynamic monitoring method based on gravity satellite time series data analysis. According to the method, the multi-source data scale effect and observation noise are effectively eliminated, the physical reliability and resolution of groundwater signal extraction are improved, and the recognition precision and early warning perspectiveness of the abnormal loss trend are enhanced. A full-chain monitoring closed loop of multi-source satellite observation, high-precision physical inversion and dynamic statistical benchmark is constructed, multi-dimensional environment data is aligned, a component separation strategy based on multi-element physical mechanism constraint is established based on a water balance principle, and anomaly diagnosis is performed in combination with a self-adaptively updated long-time-sequence historical database.

The invention discloses an emotion recognition method based on an attention mechanism and a capsule network. The emotion recognition method specifically comprises the following steps: inputting multi-channel original EEG signals; performing data truncation, data filling and data standardization on the original EEG signal, unifying the format, and generating a preprocessed EEG signal; extracting space-time joint features from the preprocessed EEG signals by using a three-dimensional convolutional neural network; the construction of the graph structure comprises modeling a complex relationship between channels; introducing a graph attention mechanism to strengthen information expression of important channels and relationships; and high-dimensional vector modeling and final classification of the features are carried out through a capsule network, and capture of deeper space structure information is completed.

The invention provides a Rowland time delay signal prediction method and system, electronic equipment, a program product and a storage medium. The method comprises the following steps: acquiring an initial Rowland time delay signal; extracting a periodic term of the initial Rowland time delay signal, and inputting the periodic term to a constructed multi-periodic term and trend term model; calculating a residual signal between the observed value of the initial Rowland time delay signal and the output of the multicycle term and trend term model; wavelet decomposition and threshold denoising are carried out on the residual signals; fusing the output of the multi-cycle term and trend term model and the denoised residual signal through an adaptive weight mechanism to obtain a predicted Rowland time delay signal; the deterministic periodic term and the long-term trend drift of the Rowland time delay signal are captured through a multi-periodic term and trend term model, and non-stationary disturbance is processed by using wavelet denoising, so that the Rowland time delay prediction precision is remarkably improved, and the method is particularly suitable for a high-precision timing scene in a complex electromagnetic environment.

The invention discloses a summer rainfall sub-season prediction method and system fused with multi-scale deep learning, and the method comprises the steps: collecting multi-source weather forecast data and observation data, and carrying out the empirical orthogonal decomposition of the observation data, and obtaining a rainfall main mode and a mode sequence; performing multi-scale signal extraction on the observation data, and performing correlation analysis on the observation data and the modal sequence to obtain respective weight fields; constructing and training a deep learning model fusing a multi-pole attention mechanism and time sequence decomposition; inputting the forecast data into the trained model to carry out transfer learning, and optimizing the model; and substituting forecast data of preset time into the trained model to generate a high-quality summer rainfall sub-season forecast product. According to the method, the synergistic effect of sea, land and gas and the interaction of multi-scale signals are fully considered, the model is constructed based on an artificial intelligence method and a numerical model forecasting product, the sub-season forecasting skill of summer rainfall is effectively improved, and the method plays an important role in disaster prevention and reduction.

The invention relates to the technical field of data processing, and discloses a cable joint arc multi-physical field simulation method and system. The method comprises the following steps: establishing a defect-containing geometric model, judging according to a temperature gradient change rate and a current rising rate to obtain an adaptive time step length, and carrying out bidirectional coupling iteration on an arc temperature field and a material dielectric parameter based on the time step length to obtain an accumulated damage factor; arranging a virtual temperature detection point to convert the simulation temperature field into a virtual monitoring signal, extracting temperature characteristic parameters and ultrasonic spectrum characteristic parameters, and projecting the temperature characteristic parameters and the ultrasonic spectrum characteristic parameters to a two-dimensional characteristic space to obtain a fault mode fingerprint spectrum. According to the method, adaptive adjustment of the time step length is realized by establishing an arc evolution phase discrimination mechanism, and the problems of low calculation efficiency and insufficient transient characteristic capture caused by a fixed time step length are solved.

The invention relates to the technical field of data processing, and discloses an emotion recognition method and system based on multi-modal feature retrieval, a terminal and a storage medium, and the method comprises the steps: collecting a video signal and an audio signal of a subject, converting the audio signal into a text signal, and extracting a video feature, an audio feature and a text feature; retrieving in a single-mode feature retrieval library according to the video features, the audio features and the text features to obtain enhanced features; aligning the enhanced features to a unified feature space through a mapping module, inputting a double-branch structure, dynamically adjusting the weight of the enhanced features through a modal weight distributor to obtain weighted enhanced features, and querying in a multi-modal feature retrieval library according to the enhanced features to obtain multi-modal retrieval features; and fusing the weighted enhanced features and the multi-modal retrieval features, inputting the fused features into a multi-modal large model for processing, and outputting an emotion recognition result of the subject. According to the invention, accurate perception of the individual emotional state is realized.

The invention relates to the technical field of cable insulation detection, and discloses a power cable insulation state evaluation method and system based on multi-dimensional feature fusion, and the method comprises the steps: collecting a high-frequency current signal on a cable grounding wire, extracting a power frequency phase, a discharge amplitude and a pulse frequency band width, and carrying out the normalization processing, thereby obtaining a sample set; calculating a local sparseness adaptive factor of each sample point, and calculating an energy weighted local density of each sample point; calculating the relative distance of each sample point, constructing a decision value, and determining the sample point with the larger decision value as a clustering center; calculating the path transmission reliability of the non-clustering center sample points relative to each clustering center, distributing the sample points with the path transmission reliability meeting a preset condition to the corresponding clustering center, and if the preset condition is not met, marking the sample points as noise; and extracting statistical characteristics according to the separated discharge clusters, and evaluating the insulation state of the cable by using a preset fault diagnosis rule base. According to the invention, the accuracy of cable insulation state evaluation can be improved.

The invention discloses a communication base station spectrum signal anti-interference method and device, equipment and a medium, and relates to the technical field of data processing. The method comprises the following steps: acquiring radio frequency spectrum signals in a coverage range of a base station, extracting interference signal features to generate a multi-dimensional feature vector group, constructing an interference propagation prediction matrix, calculating an interference diffusion probability, and generating a prediction data set. And according to the prediction data set, calculating anti-interference parameters under different interference combinations and constructing a strategy preset library. And when a target interference signal is detected, feature parameters are extracted to be matched with the prediction data set, and a corresponding anti-interference parameter combination is selected for processing. In the processing process, the monitored signal-to-noise ratio and the bit error rate are compared with a reference threshold value, feedback evaluation data are obtained, and preset library parameter configuration is updated. By implementing the technical scheme provided by the invention, the communication accuracy of the communication base station can be improved.

The invention belongs to the technical field of artificial intelligence and physiological signal processing, and discloses an electroencephalogram emotion recognition method and system based on a multi-task self-supervision and dynamic graph fusion network, and the method comprises the steps: obtaining and preprocessing electroencephalogram and other physiological signals, and extracting multi-band energy features; constructing a dynamic graph structure, taking electrodes as nodes, taking frequency band energy as characteristics, and fusing spatial distance and functional connectivity to generate a dynamic adjacency matrix; designing multi-task self-supervised pre-training, including spatial jigsaw, frequency jigsaw and cross-modal contrast learning tasks, to learn general characterization; a dynamic graph fusion network is adopted to carry out end-to-end training, and a shared feature extraction module of the dynamic graph fusion network realizes adaptive fusion of multi-modal features by utilizing Chebyshev graph convolution and embedding a cross-modal attention mechanism; the classification module sets an independent classification head for each task, and optimization is carried out through a joint loss function. According to the method, the accuracy and generalization ability of electroencephalogram emotion recognition are remarkably improved.

The invention relates to the technical field of anti-seismic performance detection, in particular to an anti-seismic performance detection method and system for an assembled steel-wood composite structure, and the method comprises the steps: arranging a multidirectional acoustic emission sensor in a steel-wood structure node area, applying simulated seismic vibration, and synchronously collecting acoustic emission signals; the method comprises the following steps: preprocessing collected acoustic emission signals, dividing the acoustic emission signals into signals corresponding to microcracks and contact slippage, respectively calculating an energy attenuation rate and a phase offset, and constructing a matrix; micro-impact is applied to the steel-wood node through an electromagnetic excitation device, a response signal is collected, resonant frequency is extracted, the drift distance between the resonant frequency and a reference value is calculated, and the severity of interface damage is determined according to the value of the drift distance; the damage correlation index calculation takes the characteristic value of the interface wave propagation characteristic matrix and the resonance frequency drift distance as input, the damage correlation index is calculated through the correlation coefficient, and the damage grade is divided according to the correlation coefficient. And differential characterization of different damage types is realized through the interface wave propagation characteristic matrix.

The invention provides an error checking method and system of an intelligent water meter, and relates to the technical field of error checking of the intelligent water meter, and the method comprises the following steps: collecting multi-modal monitoring data of the intelligent water meter; performing signal preprocessing on the water flow acoustic signal and the mechanical vibration signal to generate an acoustic fusion signal, extracting time-frequency domain features of the acoustic fusion signal related to errors, and constructing an optimal error feature set; performing pattern recognition based on the optimal error feature set and a preset standard verification feature library to generate an error preliminary result; constructing an error prediction and verification model, verifying the error preliminary result in combination with the optimal error feature set and the flow data, and outputting an error verification result; when the error grade reaches a preset error grade threshold value, an early warning mechanism is triggered, and early warning information is pushed, so that the metering performance of the water meter can be evaluated in real time, the error change trend can be dynamically predicted, the timeliness and effectiveness of error verification are effectively guaranteed, and the dynamic error management and control requirements of the intelligent water meter are comprehensively met.

The invention discloses an intelligent diagnosis method for analyzing line hidden dangers based on dynamic electrical fingerprint characteristics, and belongs to the technical field of power line state monitoring and fault prediction. According to the method, a dynamic electrical fingerprint is constructed through high-frequency acquisition of line voltage and current instantaneous value signals and extraction of multi-dimensional features such as a time domain, a frequency domain and a time-frequency domain; establishing an electrical fingerprint database and a health model of different load working conditions in a line health state; during on-line monitoring, through three-level analysis of deviation comparison, state classification and trend prediction, graded early warning of line hidden dangers is realized, and abnormal feature items are output for auxiliary diagnosis. According to the invention, subtle changes of electrical parameters can be captured, the problems of low sensitivity, weak anti-interference capability and incapability of early warning in the prior art are solved, early recognition and advanced prediction of latent hidden dangers are realized, and the safety and reliability of a power system are significantly improved.

The invention belongs to the technical field of equipment intelligent fault diagnosis and predictive maintenance, and particularly relates to a bearing intelligent fault diagnosis method fusing sound and vibration signals, which is characterized in that rapid diagnosis of a bearing fault is realized based on analysis of the fused sound and vibration signals; comprising the steps of acoustic vibration signal acquisition, extraction and fusion, enhanced feature processing, model training and fault diagnosis result generation. The method has the advantages that the one-dimensional convolutional neural network is adopted for feature extraction, multi-modal attention (MMA) is utilized for feature fusion, and the complementary advantages of sound signals and vibration signals are fully utilized. And dynamically constructing a k-nearest neighbor (KNN) graph based on the fusion features, and applying a graph convolutional network (GCN). A multi-task learning (MTL) framework is introduced, fault classification is used as a main task, and modal classification is used as an auxiliary task. According to the method, more abundant bearing fault feature information can be obtained, the classification precision is high, and an effective solution is provided for bearing intelligent fault diagnosis in industrial application.

The invention relates to a non-contact pulse signal extraction method and system based on a face video, and the method comprises the steps: preprocessing face video stream data, constructing a pulse signal extraction model, inputting the preprocessed face video stream data into the pulse signal extraction model for training, and obtaining a trained pulse signal extraction model; performing non-contact pulse signal extraction by using the trained pulse signal extraction model; according to the invention, a wavelet difference fusion module is provided, so that the signal extraction stability and precision of the model in a complex dynamic scene are effectively improved; sensing local and global spatial dependence through a self-adaptive multi-scale expansion convolution module, and strengthening the response of a key physiological signal region; a linear Mama module is designed to achieve low computing resource overhead, meanwhile, a stable and extensible time sequence feature expression framework is constructed, high modeling capacity is kept, and meanwhile good generalization performance and actual deployment advantages are achieved.

The invention discloses a hydropower station unit state monitoring system based on deep learning, and relates to the technical field of hydropower station unit state monitoring and intelligent diagnosis. Comprising a multi-modal acquisition module, an electromagnetic common-mode texture anchor point generation module, an event transient skeleton generation module, a sparse anchor point index retrieval module, a continuous time joint inversion module and an analysis domain prior correction module. Generating an event stream from the mechanical side signal, and outputting an aligned candidate set to define a time mapping feasible region and a candidate time window; solving a unified time axis continuous potential signal under the constraint of missing mask information, and outputting an alignment uncertainty field; generating an alarm permission result based on the precedence relation and the delay window; and updating index parameters, threshold value strategies, priori condition parameters and iteration schedule parameters according to permission results to realize long-term stable monitoring and reliable alarm.

The invention relates to the technical field of energy storage battery health state evaluation, and provides a low-voltage station area battery capacity calibration and health factor extraction method, which comprises the following steps: acquiring battery end voltage, current, temperature and SOC data in real time; a complete charge-discharge cycle is automatically divided according to a preset rule, and the equivalent discharge capacity is calculated through an ampere-hour integral; constructing an exponential type working condition correction factor based on the circulating temperature and the root-mean-square current to correct the capacity, and obtaining a dynamic calibration capacity in a sliding window by adopting exponential weight or weighted least square fitting; meanwhile, multi-dimensional health factors are extracted from voltage, current, temperature and SOC signals, and comprehensive health characterization is formed through adaptive weighted fusion in a data driving mode. According to the method, online and dynamic capacity estimation and multi-dimensional health assessment of the energy storage battery in the transformer area under the complex and unsteady working condition are achieved, the capacity calibration precision and the life prediction reliability are remarkably improved, and intelligent operation and maintenance and risk early warning are facilitated.

The invention discloses a method for measuring and analyzing residual magnetism characteristics of an iron core. According to the method, a to-be-detected iron core is demagnetized, preset residual magnetism is set, direct-current voltage excitation with opposite polarities, the same amplitude and the same duration is applied to a primary winding of the to-be-detected iron core, and transient current response signals are synchronously collected; two groups of current under positive / negative polarity excitation are extracted, and the residual magnetism direction is judged by comparing the change rate of the current; a curve of equivalent resistance changing along with time is calculated based on forward current and voltage, a resistance value corresponding to a characteristic inflection point moment is selected, the resistance value is substituted into a pre-established residual magnetism-equivalent resistance empirical formula to deduce the magnitude of residual magnetism, and a matched iron core residual magnetism detection and demagnetization integrated system integrates information management, residual magnetism detection, demagnetization and display modules. And closed-loop automation of detection, analysis and demagnetization is realized. According to the method, a magnetic field sensor is not needed, the residual magnetism direction and size can be synchronously obtained in a high-precision and robust mode only through port electric signals, and the method is suitable for on-site off-line or live detection.

The invention provides a vehicle fault identification and driving behavior early warning method based on multi-modal feature fusion, and relates to the field of vehicle fault diagnosis and safe driving, and the method comprises the steps: collecting a vibration signal, an acoustic signal and a driving operation signal, extracting a fault feature and a behavior feature, calculating a time delay cross-correlation function to obtain a time delay influence spectrum, and carrying out the early warning of a driving behavior. And performing recursive calculation to obtain an evolution trajectory curve, calculating a fault-behavior coupling risk index by detecting the deviation of the evolution trajectory curve from a preset stable region, and generating a multi-modal early warning signal of a corresponding intensity level according to the risk index. The method has the advantages that correlation analysis of faults and driving behaviors is achieved, and the accuracy and timeliness of early warning are improved.