654 results about "Spatial correlation" patented technology

The invention discloses a Mama-based spectrum dynamic fusion and double-attention enhancement medical image segmentation method, which comprises the following steps of: firstly, constructing a Mama integrated spectrum domain and attention pyramid module, fusing spectrum dynamic characteristics and a self-attention pooling mechanism, and performing frequency domain information compensation and local characteristic enhancement to obtain a spectrum dynamic fusion image; the spatial correlation loss caused by image blocking processing is relieved; secondly, designing a layered enhanced U-shaped architecture, deploying an MISAP module in a shallow layer of an encoder to capture multi-scale global context features, introducing a bipolar routing attention mechanism in a deep layer, and dynamically allocating sparse attention weights to focus a key pathological region; according to the method, the segmentation precision of complex edge textures and tiny lesions in medical images can be remarkably improved, and the Dice coefficient in breast tumor, polyp and abdominal organ segmentation tasks is averagely improved by 6.5%.

The invention relates to the technical field of marine environment monitoring, and discloses an intelligent collection method and system for marine multi-dimensional information and a storage medium. The method comprises the following steps: synchronously acquiring environmental parameters at a plurality of measuring points and a depth layer, and constructing a hydrological characteristic data set; calculating spatial correlation and dividing a high variation region and a low variation region; constructing an equipment spacing model based on an information density classification result, and generating a point distribution scheme by adopting an optimization algorithm; combining real-time monitoring data to judge significant changes and dynamically updating layout parameters; and issuing an instruction to the equipment through the main and standby wireless channels to complete deployment. According to the invention, the adaptability and layout efficiency of marine environment perception are improved, and the method is suitable for intelligent monitoring application of complex and changeable sea areas.

The invention provides a cradle head inspection dynamic regulation and control method and system for a power transmission line, a storage medium and a program product, and relates to the technical field of power transmission line monitoring, and the method comprises the steps: obtaining the multi-source data of a target power transmission line, inputting the multi-source data into a preset dynamic risk assessment model, obtaining a dynamic risk thermodynamic diagram output by the preset dynamic risk assessment model; establishing a spatial characteristic model for a plurality of cradle head preset positions of the target power transmission line, and performing spatial correlation analysis on the spatial characteristic model and the dynamic risk thermodynamic diagram to obtain spatial correlation characteristic parameters of each cradle head preset position in the plurality of cradle head preset positions; determining an equipment state weight value of each piece of pan-tilt camera equipment in a plurality of pieces of pan-tilt camera equipment of the target power transmission line, and determining a comprehensive inspection weight value of each pan-tilt preset position according to the equipment state weight value and the spatial correlation characteristic parameters; and converting the comprehensive inspection weight value into a target inspection period of each cradle head preset position by using a first preset inverse proportional function.

The invention relates to the technical field of power grid monitoring, and discloses a voltage sensor multi-parameter real-time monitoring method and device in an Internet of Things environment. According to the method, real-time voltage fluctuation data of a target area is acquired by deploying a multi-channel voltage sensor array and is transmitted to an edge computing node; dynamic feature extraction is executed in the edge nodes, and a multi-dimensional feature matrix containing voltage fluctuation ratio, harmonic distortion and phase deviation features is generated; identifying voltage sag, overvoltage and harmonic resonance event characteristics based on an anomaly detection model; spatial correlation modeling is carried out in combination with power grid topological structure parameters, the coupling strength of adjacent sensor node parameters is calculated, and a multi-parameter correlation map containing event propagation paths and influence ranges is generated; inputting the voltage into a diagnosis model to generate a power grid state diagnosis report, and dynamically adjusting the sampling frequency and filtering parameters of the voltage sensor according to the power grid state diagnosis report. According to the method, real-time monitoring and analysis of multiple parameters of the power grid can be realized, and comprehensive support is provided for power grid state evaluation.

The invention relates to the technical field of power dispatching, in particular to a power distribution network photovoltaic energy storage collaborative optimization dispatching decision-making system based on big data and artificial intelligence. Comprising a source network load storage full-dimension data acquisition unit; a multi-source heterogeneous data fusion processing unit; the dynamic multi-target intelligent optimization decision-making unit is used for constructing a power distribution network photovoltaic energy storage collaborative scheduling strategy through an improved non-dominated sorting multi-target grey wolf optimization algorithm with adaptive weight adjustment; and a closed-loop control execution unit. According to the invention, photovoltaic, energy storage, load and power grid operation data in the power distribution network are comprehensively captured through the source-network-load-storage full-dimension data acquisition unit, and spatial correlation mapping of topological nodes of the power distribution network in a multi-source heterogeneous data fusion processing process is combined; and constraint conditions such as power balance and node voltage during dynamic multi-objective optimization decision making are included, so that neglect on topology and operation constraints of the power distribution network is effectively made up.

The invention provides an electricity load equipment identification method based on spatial-temporal feature fusion and an attention mechanism, and the method comprises the steps: carrying out the dual preprocessing of the data through the real-time collection of the electrical time sequence data, operation state labels, spatial layout information and environment parameters of target equipment, comprising the steps of dynamically aligning a time sequence curve by time warping and adaptively adjusting data distribution in a standardized manner. A spatial correlation graph is constructed by using a graph neural network, and space-time proximity is calculated to generate a weighted adjacency matrix. A multi-head attention mechanism is adopted to fuse electrical time sequence features and spatial topology features, and bidirectional asymmetric attention modulation network optimization feature fusion is constructed. Time and space dimension features are extracted through a time flow module and a space flow module respectively, the features are dynamically weighted and fused through a gating attention unit, and the identification model is input to output the equipment type and the working state. According to the method, the accuracy and robustness of electricity load equipment identification can be improved, and the method has good adaptability and expansibility.

The invention provides a dynamic spectrum prediction method based on a time sequence knowledge graph, a medium and equipment. The method comprises the following steps: constructing a communication spectrum coordination knowledge graph framework structure fusing static knowledge and dynamic knowledge; constructing a knowledge embedding model based on a cyclic evolution network, and realizing semantic representation of entities and relationships thereof in an electromagnetic spectrum space; through a knowledge fusion method, static knowledge embedding and dynamic knowledge embedding are fused with historical communication node frequency data so as to realize feature interaction of a static knowledge graph and a dynamic knowledge graph, and features of communication nodes in a frequency domain are extracted at the same time; constructing a space-time diagram convolutional neural network model, extracting space correlation characteristics between nodes, capturing a time evolution rule of node attributes, and dynamically predicting communication frequency; according to the method, the characteristics of the communication node frequency are effectively extracted from the communication semantic relationship, and the prediction precision of the communication frequency is remarkably improved.

The invention relates to an epilepsy prediction method based on adaptive sparse attention and a hierarchical graph convolution network, and the method comprises the steps: carrying out the time domain convolution, spectrum transformation and Haar wavelet down-sampling of an electroencephalogram signal, respectively generating time domain, spectral domain and fidelity down-sampling features, and fusing the features into a low-level feature set; on the basis of a sparse attention mechanism, constructing and applying a multi-level sparse mask to adaptively screen and weight-aggregate key discriminative features in the feature set to obtain screened features; on the basis of the feature, by constructing a local channel graph and a global frequency band graph and respectively executing graph convolution, capturing local spatial correlation of each channel in a single frequency band and global cross-frequency-band spatial dependence among different frequency bands, and fusing the local spatial correlation and the global cross-frequency-band spatial dependence into an embedded feature; and inputting the embedded features into a classifier to obtain a state probability, and triggering an alarm based on the state probability. Therefore, the problems of key information loss, insufficient time-space spectrum dependent modeling and feature redundancy are solved, and the accuracy, stability and real-time performance of epilepsy prediction are improved.

The invention discloses a mine earthquake intelligent identification method based on multi-mode deep learning and signal processing, and belongs to the technical field of mine earthquake identification. Aiming at the problems that a traditional mine earthquake signal processing method cannot fully utilize spatial relevance and frequency domain characteristics of signals in a sensor network, the automation level is relatively low, the response speed is slow, and the emergency processing effect is influenced, seismic waveform data and multi-modal characteristic input are adopted, so that the emergency processing effect is influenced. Based on text semantic features generated by a large language model (LLMs), frequency domain features extracted by Fourier transform and time sequence features after obspy processing, a dynamic graph recognition model of a fusion graph neural network (GNN) and Transform is constructed; graph structure edge weights are dynamically updated through cross correlation between signals, and the structure change of complex environments such as mine roadways is self-adapted.

The invention provides a deicing control method and system for a power transmission line in a cold region, and relates to the technical field of power system power transmission line operation and maintaining.The method comprises the steps that icing data of the power transmission line are collected, the icing data are multi-source data and comprise the ice layer thickness, the ice type category and the distribution characteristic, and the ice layer thickness, the ice type category and the distribution characteristic are obtained by combining environment temperature, humidity and geographical location information; dynamically selecting a plurality of spatial sampling positions on the target line to construct a spatial correlation area for feature analysis; and obtaining correction parameters based on morphological characteristics of the spatial correlation region, performing spatial superposition and feature fusion analysis on the multi-source data according to the correction parameters, and extracting an icing feature set which is dynamically corrected and reflects the overall icing state of the line. According to the invention, differentiated ice removal in different climate areas can be realized, and intelligent deicing control of accurate identification, mode adaptation and closed-loop adjustment is realized in combination with multi-source monitoring data.

The invention discloses a space-time sequence interpolation method and device oriented to heterogeneous deletion, and belongs to the technical field of space-time data processing. Aiming at random or continuous loss of the sensor network caused by faults and communication interruption, the method comprises the following steps: setting static space experts, dynamic space experts, short-term experts and long-term period experts in parallel in the same framework, and respectively capturing fixed geographical adjacency, time-varying space correlation, local continuous trend and long-period rules; spatial features are extracted through high-order diffusion diagram convolution and bidirectional gating circulation, time features are extracted through multi-layer space-time attention, a memory attention gating network is introduced to dynamically weight and fuse output of experts according to reconstruction errors, and node-level and time-step-level self-adaptive interpolation is achieved. Experiments show that compared with the prior art, the method has the advantages that under various real data sets and heterogeneous missing scenes, the precision and robustness are remarkably improved, and the method can be widely applied to scenes needing high-integrity spatio-temporal data, such as intelligent transportation, air quality monitoring and energy internet of things.

The invention discloses a multi-source environment monitoring data fusion analysis and abnormity early warning method based on deep learning. According to the method, the spatial attention weight is calculated by fusing the feature similarity and the geographic distance, and the modeling precision of the complex spatial dependency relationship between the environmental monitoring nodes is remarkably improved. Specifically, the model firstly divides a monitoring area into fine grid units, each unit serves as an independent node to participate in calculation, the direction consistency of node feature vectors is quantified through cosine similarity, and meanwhile, a geographic distance attenuation factor is introduced to dynamically adjust weight distribution. The double consideration enables nodes with similar environmental characteristics in adjacent areas to obtain higher attention, and effectively avoids the problem of spatial correlation misjudgment easily occurring in a complex terrain in a traditional fixed adjacent matrix. When local pollution diffusion occurs in a monitoring area, the model can quickly position a propagation path of a core pollution source and surrounding affected nodes, and the space correlation identification efficiency is improved.

The invention relates to the technical field of meteorological data processing, in particular to a training data acquisition method and system for a power transmission line fault prediction model. The method comprises the following steps: constructing a multi-source heterogeneous original data pool containing severe convection meteorological data, power transmission line operation and fault data and geographic space data; performing time-space reference unified processing and grid division on the data; mapping the line paragraph to a grid unit, identifying a severe convection event and dividing an influence area through a density clustering algorithm based on a multi-parameter collaborative threshold value, calculating a grid exposure degree, and combining a strength grade weight and a terrain correction weight to calculate a space correlation strength; setting a differentiation time association window to calculate a time factor according to the severe convection dominant type, calculating association credibility based on the space association strength and the time factor, and screening high-credibility association results; and performing attribution labeling on the fault information to obtain a training data set. The problem of low training data quality is solved.

The invention relates to the technical field of virtual power plants, and discloses a virtual power plant optimization control system based on source-grid-load-storage collaboration, which comprehensively acquires and preprocesses wind and light output, load, weather and power grid operation data through a multi-source data acquisition module, provides accurate original data for error analysis, and improves the accuracy of the system. The prediction error space-time correlation analysis module mines space correlation and time self-correlation characteristics of errors and generates an error scene set containing correlation characteristics, the source load scene library module combines with the error scene set to update and form a dynamic scene library containing risks, and the robust optimization scheduling module carries out robust optimization scheduling on the basis of the dynamic scene library. A scheduling scheme is solved by taking economy and robustness as targets, and a real-time scheduling execution module is linked with a system monitoring feedback module, so that error perceptibility, scene dynamic updating and scheduling adaptive adjustment are integrally realized, the ability of a virtual power plant to deal with uncertainty is improved, the economy and power grid stability are guaranteed, and the method is suitable for large-scale popularization and application. And the reliability and competitiveness of system operation are enhanced.

The invention discloses a data reading method and system based on an RFID reader-writer, particularly relates to the technical field of radio frequency identification data acquisition, and is used for solving the problem that the communication reliability of an existing RFID system is reduced due to carrier frequency deviation and multipath interference in a high-speed mobile environment. A phase change characteristic and an error rate index are obtained through double-antenna receiving signal analysis, when the error rate exceeds a limit, spatial correlation and error rate distribution difference are jointly analyzed to determine a multipath interference level, and meanwhile, a relative motion state is detected and judged through a phase track curvature characteristic. And dynamically selecting an optimal modulation coding mode to read data again according to the calculated motion speed estimated value and the interference level, thereby realizing the optimal balance between the communication rate and the reliability in the mobile environment.

The invention relates to the technical field of industrial automation quality control, in particular to a bicycle lifting seat production defect detection method and system based on machine vision, and the method comprises the steps: carrying out the illumination decomposition and highlight reconstruction of collected surface data, and generating a balanced texture map with uniform illumination through a gradient domain local repair algorithm and global histogram equalization. Then, the balanced texture map is input into a two-channel parallel analysis architecture comprising a linear flaw attention network and a regional heterogeneity analysis network, and probability maps of scratch defects and oxidation defects are extracted; then, performing spatial correlation intelligent arbitration and weighted fusion based on a local confidence mean value on the two paths of probability graphs so as to eliminate overlapping conflicts and background noise, and generating a fusion defect graph; and finally, carrying out topology and geometric constraint filtering on the fused defect graph, normalizing the defect form, and finally outputting a defect positioning mask representing the position and contour of the defect. According to the invention, identification and distinguishing of surface scratches and oxidation defects of the bicycle lifting seat are realized.

The invention discloses a combustion chamber and turbine guider integrated composite flow channel outlet flow field uniformity evaluation method, and belongs to the technical field of gas turbine pneumatic design. The method comprises the steps that a geometric model of a combustion chamber and turbine guider integrated composite flow channel is built, a combustion chamber function section and a guider pneumatic section are fused through integral pneumatic modeling to form a continuous flow channel, a front section adopts a diffusion structure to form a low-speed combustion area, and a rear section adopts a gradually-shrinking channel to achieve gas accelerated expansion; performing three-dimensional cold-state flow field numerical simulation by adopting a structured grid division and delay separation vortex simulation technology; flow parameters including total pressure, Mach number, airflow angle and vorticity are extracted from the cross section of a flow channel outlet, conversion from a Cartesian coordinate system to a polar coordinate system is carried out, and a uniformly distributed aerodynamic parameter field is reconstructed by using an interpolation algorithm; the directional coupling degree is calculated based on the spatial correlation of the pressure gradient vector and the vorticity vector, and the vortex-pressure coupling entropy production index representing the overall aerodynamic loss of the flow channel is established in combination with the pressure pulsation intensity and the vorticity concentration ratio. According to the method, a three-dimensional evaluation system is constructed by integrating circumferential and radial non-uniformity and entropy production parameters, and accurate quantitative evaluation of flow characteristics and uniformity of the combustion chamber and turbine integrated structure is realized.

The invention relates to the cross technical field of artificial intelligence and traffic infrastructure monitoring, discloses an intelligent health monitoring and prediction method and system for a road roadbed, and aims to solve the problems of insufficient monitoring coverage, shallow data mining, low prediction model precision and disjunction of operation and maintenance decisions in the prior art. The method comprises the following steps: collecting roadbed multi-dimensional physical field data through a multi-source sensor network; denoising, abnormity correction, time alignment and feature compression are carried out at the edge end; fusing multi-scale time sequence modeling and spatial correlation analysis to extract health features; predicting a future health state and a risk probability by using an LSTM-Attention model in combination with historical data and environment variables; and triggering graded early warning based on the dynamic threshold and generating a maintenance strategy. According to the technical scheme, high-precision and high-timeliness roadbed health perception and prediction can be realized, and the early warning response speed and the maintenance decision intelligent level are improved.

The invention relates to the technical field of digital twinning modeling and simulation, in particular to a train operation and maintenance method based on digital twinning. The method comprises the following steps: collecting vibration and acoustic sensing data from a wheel and track interaction area, and fusing train operation condition data to obtain wheel-track interaction data; carrying out adaptive filtering processing on the wheel-rail interaction data to obtain filtering contact features; analyzing the spatial correlation of vibration signals in the filtering contact features, and reconstructing a contact point distribution diagram; performing pressure field calculation according to the contact point distribution diagram to obtain a contact pressure distribution diagram; performing equivalent taper derivation according to the contact pressure distribution diagram and the contact point distribution diagram, constructing a wheel-rail contact digital twin model, and generating a contact geometric feature diagram; a potential wear pattern is analyzed based on the contact geometry map. According to the method, virtual-real mapping and closed-loop optimization of the wheels and the steel rails are achieved through the digital twin modeling and simulation technology, the train operation and maintenance period is prolonged, and operation safety is improved.

The invention provides a voice enhancement and high-precision recognition method and system in a complex environment, and relates to the technical field of voice processing, and the method comprises the steps: collecting a time domain signal in an off-road parking sentry box environment for preprocessing, detecting a mute segment signal in a standard time domain signal for noise power spectral density estimation, and obtaining a noise power spectral density value; a reverberation parameter is obtained by combining voice onset information and noise spatial correlation estimation, prediction is performed by using a deep neural network model, voice masking is applied to microphone array signals to perform enhancement processing, adaptive feature extraction is performed on time domain enhanced voice signals, and a voice signal is obtained. And performing high-precision recognition on the voice adaptive feature sequence based on an acoustic model and a language model, and outputting a target recognition text. The technical problems of poor voice signal quality and low recognition accuracy in a complex noise environment in the prior art are solved. The technical effects of improving the voice signal quality and the recognition accuracy and realizing clear, accurate and real-time voice interaction are achieved.

The invention discloses a GCN-LSTM-based tailing dam multi-point settlement deformation prediction method. The method comprises the steps of collecting and preprocessing original settlement monitoring data of monitoring points in a research area; constructing a graph structure for settlement data between all monitoring point pairs, setting a threshold value, connecting node pairs with significant correlation coefficient relationships by using edges, constructing a weighted undirected graph, and converting the weighted undirected graph into a normalized adjacent matrix as the input of a graph convolutional network GCN; extracting the spatial topology of each monitoring point by using a GCN; capturing long-term time dependence in the settlement process by using a gating mechanism of the LSTM network; and fusing the GCN and the LSTM network, and finally outputting a predicted value through a full connection layer. According to the method, the spatial topological graph among the monitoring points of the tailing dam is constructed, and the GCN and LSTM networks are fused, so that the spatial correlation among the monitoring points and the time dynamic characteristics of the settlement data are effectively captured, and high-precision tailing dam multipoint settlement prediction is realized.

The invention relates to the technical field of mine safety monitoring, and particularly discloses a metal mine safety risk avoiding multi-source data fusion monitoring method, which comprises the following steps of: receiving multi-source data streams from environment, equipment and personnel positioning and video monitoring subsystems under a unified time reference and roadway coordinate system; a subsystem identifier, a roadway node identifier and a timestamp are written into each record; constructing a space-time diagram containing an observation layer and a roadway layer, mapping multi-source data to observation layer nodes, and aggregating the multi-source data to corresponding roadway nodes according to time slices; in the sliding time window, credibility indexes are generated for observation layer nodes according to fault records, sampling intervals and spatial correlation, the credibility indexes are input into the credibility mapping model, the dynamic credibility of the observation layer nodes is obtained, and the dynamic credibility is summarized into node credibility according to roadway nodes. According to the invention, through the multi-source monitoring space-time diagram and the risk-avoiding decision model driven by the credibility, reliable evaluation of the monitoring result during the metal mine disaster and dynamic optimization of the risk-avoiding path are realized.

The invention relates to the technical field of supplementing missing measurement data of an anemometer tower, in particular to a method for supplementing missing measurement data of an anemometer tower. In the data preparation stage, historical observation data of a target anemometer tower needs to be collected, and missing time periods and missing features are determined, so that subsequent method selection is linked; meanwhile, data of surrounding meteorological stations in the same period are obtained, the spatial correlation between the data and the anemometer tower is verified, and a basis is provided for probability distribution mapping and quantile matching in subsequent CDF-t and QDM methods. The nonlinear correction method based on cumulative distribution function transformation and quantile increment mapping shows unique advantages in the field of climate downscaling, the CDF-t can effectively eliminate system deviation by establishing a probability distribution mapping relation between observation data and mode output, and the system performance is improved. The QDM maintains statistical characteristics of historical sequences while retaining climatic change signals through quantile matching, and new possibility is provided for long-period data reconstruction through combination of the two methods.

The invention discloses a short-term wind power prediction method based on dynamic graph and multi-scale space-time fusion. A dynamic space-time convolution attention network is constructed. Aiming at the dynamic topology and sequence complex characteristics of the wind power plant, a space-time topology module based on metric learning is designed, the physical distance and real-time characteristics are adaptively fused, and dynamic space correlation caused by the wind direction and wake flow effect is captured. The core ST-CAN module integrates graph convolution, time convolution and a self-attention mechanism, and through a mixed strategy of local convolution and global attention, modeling spatial dependence, local high-frequency fluctuation and a long-range period rule in a collaborative manner. According to the method, the prediction precision and generalization ability are remarkably improved, and power prediction challenges under complex wind conditions can be effectively coped with. According to the method, MAE, RMSE and MAPE indexes under different prediction step lengths are superior to those of a traditional method and an existing deep learning model, the prediction precision is high, the robustness is high, and the method is suitable for a complex wind field environment.

A metering equipment warehouse division demand prediction and scheduling method based on spatio-temporal feature fusion comprises the steps of firstly obtaining and collecting warehouse division data, then establishing a spatio-temporal feature fusion model based on the warehouse division data, and then establishing a spatio-temporal diagram convolution prediction network, the spatio-temporal diagram convolution comprises a diagram convolution layer and a time convolution layer, through extraction and fusion of the two layered features, dynamic weight fusion is obtained, joint modeling of space-time dynamics is realized, and finally, prediction and allocation decision are implemented based on the dynamic weight fusion. According to the method, the multi-modal graph structure fusing the space-time association and the replacement rule is constructed, and a space-time joint modeling architecture and a dynamic feedback mechanism are designed, so that high-precision demand prediction and global inventory optimization are realized; the method systematically solves the core problems of insufficient spatial correlation modeling, dynamic event response lagging, low efficiency in multi-source data utilization and the like of a traditional method, and provides an efficient solution for electric power material management.

The invention discloses a reconfigurable intelligent surface (RIS)-assisted multiple-input-multiple-output (MIMO) implicit channel estimation method based on a graph attention network, which is used for efficient downlink transmission in a multi-user scene. Firstly, a graph attention network is designed, user nodes and RIS nodes are modeled in a unified mode, received pilot signals serve as initial features, spatial feature expression is enhanced in combination with user three-dimensional position information, and therefore interference between users and a spatial correlation structure are accurately represented; secondly, end-to-end feature aggregation is achieved based on a message passing mechanism, a base station beam forming matrix and an RIS reflection coefficient are directly predicted under the condition that explicit channel estimation is not needed, and the total transmitting power constraint and the unit mode constraint are met through normalization processing so as to complete joint optimization; according to the method, the users and the speed of the system can be remarkably improved under limited pilot frequency overhead, and the method has excellent generalization performance and robustness in a multi-user complex propagation environment.

Embodiments of the present invention provide a DMRS port determining method and apparatus, a device, and a storage medium. The method comprises: receiving first indication information sent by a network device, the first indication information being used for indicating quasi co-location (QCL) relationship information and / or power control parameters corresponding to a physical uplink shared channel (PUSCH); and according to the first indication information, determining DMRS port(s) corresponding to the PUSCH, wherein the DMRS port(s) correspond(s) to one or multiple pieces of space related information, and the DMRS ports corresponding to different space related information among the multiple pieces of space related information are different.

The invention discloses an event-driven architecture-oriented spatial-temporal feature extraction method and system for multi-type fault power failure events, and belongs to the field of power system fault diagnosis. According to the method, historical fault data, equipment operation data and external environment data are collected through event monitoring triggering and multi-source data fusion; an XGBoost model based on probability distribution is adopted to carry out precise cleaning and complementing on the data, and a comprehensive full-scene fault sample library is constructed; and further, a time sequence convolutional network and a graph convolutional neural network are utilized to deeply mine and fuse spatio-temporal characteristics such as periodicity, high-incidence time periods, geographical distribution rules and spatial relevance of fault events from a sample library. According to the method, spatio-temporal evolution law analysis and key influence factor identification of fault events can be efficiently and intelligently realized, the accuracy and timeliness of power grid fault research and judgment are remarkably improved, a scientific basis is provided for fault positioning and emergency disposal, and construction of a novel power system is assisted.

The invention discloses a distributed optical fiber sensing system for water conservancy project structure health monitoring, relates to the technical field of water conservancy project safety monitoring, and provides the following scheme: the distributed optical fiber sensing system comprises a signal acquisition and preprocessing module which is used for dividing distributed optical fiber segments according to a fixed length, acquiring vibration signals of central points of the optical fiber segments in a fixed period, and sending the vibration signals to a data processing module; generating a sampling point sequence with uniform time intervals; and the feature extraction and criterion construction module is connected with the signal acquisition and preprocessing module and is used for detecting each sampling point in the sampling point sequence. According to the distributed optical fiber sensing system for water conservancy project structure health monitoring provided by the invention, through double-level criterion construction associated with frequency domain characteristics and spatial topology, a dynamic heat point classification mechanism based on signal propagation characteristics and collaborative optimization of a hierarchical transmission strategy; the problems of false signal flooding and transmission resource waste caused by multi-dimensional criterion isolation and spatial correlation deficiency in the prior art are effectively solved.

The invention discloses a BIM-based cable construction method. The method comprises the steps of obtaining BIM model data including a cable design path, a design bending radius and design stress; in the cable laying process, the optical fiber sensors are tightly attached and fixed to the surface of an outer sheath of the cable to be laid, and the optical fiber sensors are continuously distributed in the length direction of the cable; acquiring actual strain distribution data and actual bending radius distribution data of the cable measured by the optical fiber sensor in real time, and performing spatial position correlation comparison with design stress data and design bending radius data in the BIM model; if the actual strain value of any position point exceeds the strain threshold range corresponding to the design stress or the actual bending radius value is smaller than the design bending radius value, it is judged that construction deviation exists in the position point; and generating an adjustment instruction for subsequent cable laying operation based on the position point information with the construction deviation. The cable laying precision and the construction safety can be effectively improved.