225 results about "Conjoint analysis" patented technology

The invention relates to the technical field of electrical fault diagnosis, and discloses an electrical system fault diagnosis method based on multiple parameters. According to the method, multi-dimensional operation parameters of an electrical system are collected, wherein the multi-dimensional operation parameters comprise voltage waveform data, current waveform data and temperature distribution data; time-frequency conjoint analysis is carried out on the multi-dimensional operation parameters, and multi-scale electrical features are extracted and comprise steady-state feature components and transient feature components; based on a historical fault case library, performing mode matching on the multi-scale electrical characteristics to generate an initial fault type set; performing confidence evaluation on the fault types in the initial fault type set by using a dynamic weight distribution algorithm, and screening out high-confidence fault types; and according to the high-confidence fault type, a fault evolution path model is constructed, and the fault evolution path model is used for describing a time sequence of fault features.

The invention provides a method for reducing the multipath effect of a take-off and landing guide radar based on adaptive off-axis tracking, and the method comprises the steps: enabling the take-off and landing guide radar to transmit a signal, and transmitting the signal to a radar direction after receiving the signal; the radar receives the direct signal and the multi-path signal at the same time, and separates and establishes a spatial-temporal characteristic database of the multi-path signal; when the influence of the signal on the angle error exceeds a dynamic threshold value, starting an off-axis tracking mode; dynamically calculating the optimal off-axis angle, locking the elevation angle of the antenna, and performing closed-loop tracking of the azimuth; carrying out conjoint analysis, and identifying and eliminating false targets in real time; intelligent suppression is carried out on angle random errors; a real-time multi-path propagation model is constructed, and phase and amplitude compensation is carried out on a received signal. The method has excellent real-time performance and engineering realizability, and can meet strict requirements of aircraft precision approach guidance. The method has good expandability and adaptability, and provides reliable technical support for guaranteeing safe take-off and landing of the aircraft in severe weather and complex environments.

The invention relates to the technical field of medical information, in particular to a medical data conjoint analysis system based on medical knowledge graph driving. The system comprises a medical knowledge graph construction module, a medical data quality evaluation module, a medical feature engineering module and a medical knowledge driven analysis module, and heterogeneous knowledge graph modeling can be performed by integrating medical clinical guidelines, biomedical literatures, a drug database and historical medical data so as to generate a medical heterogeneous knowledge graph; obtaining new medical clinical test information and carrying out delayed contradictory learning update to generate a medical dynamic update knowledge graph; the method comprises the following steps: obtaining multi-modal medical data, carrying out quality verification evaluation and medical feature engineering analysis, carrying out knowledge path joint driving analysis at the same time, generating a decision support reasoning path corresponding to medical clinical knowledge, and outputting a corresponding medical knowledge path confidence coefficient. According to the method, fusion reasoning among cross-source data can be realized by constructing the multi-modal medical knowledge graph.

The embodiment of the invention relates to the technical field of artificial intelligence, and provides an apartment security linkage method and system based on artificial intelligence, and the method comprises the steps: enabling initial security data obtained in real time to form a security space data set matched with an apartment space according to a space-time association relation; performing conjoint analysis on the security space data set by adopting a security intelligent decision engine to generate a corresponding linkage strategy; sending the corresponding control instruction and the security control parameter to the corresponding target security equipment based on the linkage strategy; the linkage priority of the security and protection equipment is set according to the security and protection event type and the mapping relation between the security and protection equipment and the security and protection event emergency degree, and a control array containing all levels of target security and protection equipment is constructed according to the linkage priority in combination with the current security and protection event type. The control array preferentially triggers the core target security and protection equipment to respond to the indication of the linkage strategy, so that remote control and linkage response of the security and protection equipment are realized, and the intelligent level and the emergency processing capability of apartment security and protection are improved.

The invention discloses a multi-dimensional natural resource intelligent monitoring method and system based on big data analysis, and relates to the technical field of resource monitoring, and the method comprises the steps: calculating the information sharing degree between different data streams, constructing an abnormal feature credibility distribution map, and carrying out the calculation of the abnormal feature credibility distribution map; and inputting the abnormal feature credibility distribution graph and the multi-source auxiliary information into an integrated classifier, carrying out joint analysis and comprehensive research and judgment on multi-dimensional information to output a target list, constructing an event response relation graph reflecting a monitoring task priority and an execution logic relation according to the attribute features of each element in the target list, and carrying out event response analysis on the monitoring task priority and the execution logic relation. Wherein each node represents a to-be-monitored target, an edge represents a dependency relationship between tasks, and an optimal monitoring execution path is planned for various intelligent monitoring carriers by using an improved A algorithm integrated with a multi-factor cost function. Through natural resource monitoring of multi-modal data interference correction, intelligent classification and path optimization and dynamic response, the monitoring quality and the execution efficiency are effectively improved.

The embodiment of the invention discloses a feature recognition labeling method and system applied to three-dimensional tunnel modeling, and the method comprises the steps: firstly carrying out the preprocessing of an obtained tunnel construction design drawing and point cloud scanning data of a corresponding construction stage, and obtaining a three-dimensional tunnel model data set with a space coordinate alignment relation; the three-dimensional tunnel model data set comprises a multi-source data fused structural point cloud unit and a design drawing vectorized contour line unit; secondly, performing three-dimensional structure feature extraction on the data set to obtain geometric morphology features of a structure point cloud unit and structure correlation features of a contour line unit; then carrying out conjoint analysis on the two features through a deep learning network to generate a feature analysis result containing feature response distribution; and finally, based on the result, identifying and labeling a key structure region in the three-dimensional tunnel model data set, generating a feature identification and labeling result containing feature type labels and spatial position information, and providing more accurate analysis and management basis for tunnel construction.

The invention discloses a target aircraft flight trajectory correction method based on multi-source fusion navigation, and relates to the technical field of trajectory correction. According to the method, multi-source navigation information such as AI visual navigation, inertial measurement and barometric altitude can be fully fused, short-time disturbance and long-time deviation are effectively distinguished through conjoint analysis of short-time envelope and trend envelope, and the sensitivity and reliability of trajectory anomaly detection are remarkably improved; by setting a direction consistency screening and grading weight suppression mechanism, misjudgment caused by air pressure fluctuation or instantaneous noise in low-altitude flight can be suppressed, so that the false alarm rate is reduced; the method has the advantages that the method is simple and easy to implement, an envelope difference accumulated evidence and a threshold linkage triggering mechanism are utilized, early warning of trajectory deviation is achieved, the timeliness of abnormal response in a flight task is improved, stable and accurate trajectory correction can be dynamically generated in a changeable flight environment, and robustness and convergence of the correction process are guaranteed.

The invention relates to the technical field of artificial intelligence, in particular to an agent integration optimization method and system based on multi-modal interaction, and the method comprises the following steps: obtaining a time point statistical frequency difference value of a multi-modal event, comparing a trend with offset, screening out a non-linkage event, extracting a semantic comparison context, positioning a modal in-situ, and obtaining a multi-modal event time point statistical frequency difference value; and calculating the priority of a response offset sorting channel, judging interlaced and densely set interference marks, screening fast and undisturbed channel rebinding content, and obtaining an integrated optimization scheme. According to the method, abnormal modal recognition is achieved through modal frequency fluctuation and response offset trend conjoint analysis, semantic fragments are combined with context continuity and task density constraints to improve screening precision, channel response delay difference sorting is clear in priority relation, and a signal interference area is positioned through amplitude mutation and time intersection nodes. And channel path elimination and content remapping guarantee response smoothness, accurate matching between expression intention and channel response is realized, and modal collaboration consistency and task execution stability are enhanced.

The invention relates to the technical field of pipeline detection, and provides a non-contact pipeline fluid flow velocity measurement method based on a frequency wavenumber domain spatial spectrogram, and the method comprises the steps: collecting a turbulence signal in a pipeline through a piezoelectric film sensor; carrying out snapshot number segmentation on the turbulence signal to obtain a plurality of segmented turbulence time domain signals; fourier transform and narrowband signal processing are carried out on the segmented turbulence time domain signals, and a spatial spectrum corresponding to each narrowband signal component is obtained; constructing a three-dimensional frequency-wave number domain spectrogram according to the spatial spectrum function; and according to the three-dimensional frequency-wavenumber domain spectrogram, calculating the flow velocity of the pipeline fluid. According to the method, a three-dimensional spectrogram characteristic space is constructed through conjoint analysis of the frequency and the wave number, the angle, distance and frequency characteristics of the turbulence signals are decoupled through spatial spectrum estimation methods such as the MUSIC algorithm, multipath interference and noise are effectively restrained, the flow speed of fluid in a pipeline can be accurately estimated, and limitation in a traditional method is overcome.

The invention relates to an electric power information questioning system based on reference knowledge graph dynamic expansion. The system comprises a reference knowledge graph construction module, a multi-modal data fusion module, a questioning inertia library generation module and a user questioning analysis module. And deep upgrading is carried out through multi-source data fusion. And the multi-modal data fusion module dynamically weights different modal data through an attention mechanism, and realizes image-text-operation behavior conjoint analysis in combination with cross-modal transfer learning, so that the equipment fault identification efficiency is shortened from manual processing. Simulation fault data of novel equipment is generated through a data enhancement algorithm, a knowledge graph and a question inertia library are automatically updated in combination with incremental learning, a closed-loop evolutionary mechanism of data acquisition-evaluation-enhancement-application is formed, the adaptation period of a new technology is shortened, and the efficiency is improved. And the knowledge graph data quality is continuously optimized through a three-dimensional evaluation system (integrity, accuracy and consistency).

The invention provides an operation optimization method and system based on electric power data conjoint analysis, and the method comprises the steps: obtaining synchronous sampling data and power grid topological parameters of a wide-area electric power system; performing equipment state evaluation, broadband resonance traceability and harmonic responsibility quantification on the synchronous sampling data and the power grid topological parameters by using a conjoint analysis engine, and generating a power grid control strategy; and carrying out operation optimization on the power equipment in the wide-area power system based on the power grid control strategy. According to the method, conjoint analysis is carried out on equipment state evaluation, broadband resonance traceability and harmonic responsibility quantification of the wide-area power system, traditional static spectrum analysis is upgraded to dynamic behavior perception, the power grid analysis depth and accuracy can be improved, the high-frequency harmonic analysis requirement can be met through broadband resonance traceability, and the method is suitable for large-scale popularization and application. Intelligent analysis can be realized for different application scenes, the waveform analysis capability and the frequency response range are improved, and the analysis effect is improved.

The invention relates to a forklift examination anti-cheating method and system based on artificial intelligence and a quantification method. The anti-cheating method for the forklift examination comprises the following steps: S1, acquiring a limb action video V (t) of an examinee during the examination and pressure data P (t) of a forklift driving system; s2, extracting a human body skeleton graph Gt at the t-th moment from the V (t); extracting the Gt to obtain a spatial structure feature set St, and extracting a time sequence dynamic feature set Dt of the St; and fusing the St and Dt features to obtain a pose feature Zt. Visual sense and hydraulic sensing data are fused, association analysis is carried out on examinee actions and vehicle responses, and abnormities which are difficult to find through a single video or sensing means are detected. Through conjoint analysis of images and pressure data, hidden cheating means such as remote control and dangerous violation operation can be identified, and three-dimensional monitoring of the examination process is realized.

The invention discloses a power distribution area fluctuation load early warning method and system, and relates to the technical field of power distribution network intelligent monitoring and control. The method comprises the following steps: constructing a digital twinborn model synchronous with an operation state, carrying out time domain and frequency domain conjoint analysis on a digital twinborn data set, extracting a power amplitude change rate and a phase jump index, and generating a load fluctuation vector; determining an adaptive threshold based on the statistical characteristics of the load fluctuation vector and historical load distribution, and generating a primary early warning signal when the fluctuation exceeds the limit; using a distributed collaborative analysis algorithm to identify a dominant load source and an imbalance category; performing correlation analysis on the identification result of the dominant load source and the primary early warning signal to generate comprehensive early warning information; and the comprehensive early warning information is output to a power distribution area monitoring platform, so that dynamic visualization and intelligent regulation and control are realized. According to the method, real-time identification and active early warning of the fluctuation load of the power distribution area can be realized, and the safety and the intelligent level of system operation are improved.

The invention relates to the technical field of hyper-spectral satellite remote sensing, in particular to a hyper-spectral-VOCs underway combined tracing method based on deep learning, which screens abnormal concentration through underway monitoring, extracts multiple indexes to judge pollution phase change, identifies boundaries in combination with remote sensing images, extracts pixel identification components and calculates probability to screen a tracing emission source. According to the method, space-time linkage screening is carried out on VOCs concentration threshold exceeding behaviors by constructing a concentration abnormal response task flow, a high-sensitivity mark of pollution response is effectively formed, and on the basis of conjoint analysis of three groups of indexes of concentration time sequence difference, space standard deviation and local fluctuation ratio, by means of standardization processing and jump trend identification means, the VOCs concentration threshold exceeding behaviors are identified, and the VOCs concentration threshold exceeding behaviors are identified. And multi-dimensional dynamic discrimination of the pollution phase change phenomenon is realized, so that the response capability of sudden or hidden emission events is enhanced. The boundary of a suspected pollution area in a remote sensing image is finely depicted through a gray gradient change trend, and the stability of the boundary is judged in combination with time sequence displacement and local gradient density change, so that artifact disturbance factors are effectively eliminated.

The invention provides a remote maintenance and intelligent inspection early warning management method and system for a weight meter, and relates to the technical field of intelligent measurement and monitoring, and the method comprises the steps: arranging a multi-source sensor array on a weighing mechanism to collect detection data, carrying out the three-domain conjoint analysis through an edge calculation unit, and extracting multi-dimensional coupling features; the remote monitoring platform establishes a multi-scale damage evolution prediction model based on deep learning to realize the prediction of microscopic crack propagation and macroscopic fatigue damage; an intelligent material is adopted to realize self-adaptive active compensation, and prediction-compensation-evaluation-optimization closed-loop control is formed. The operation reliability and the metering precision of the weight meter can be improved, and the service life of equipment is prolonged.

The invention relates to the technical field of cell data analysis, and discloses a single-cell transcriptome data and text description conjoint analysis method based on a multi-modal language model, which comprises the following steps: acquiring a single-cell RNA sequencing expression matrix and a corresponding cell text description, preprocessing the single-cell RNA sequencing expression matrix and the corresponding cell text description, and analyzing the single-cell transcriptome data and the corresponding cell text description; according to the method, a multi-modal data set is constructed, deep fusion of gene expression data and text knowledge is realized by constructing a double-model and cross-modal projection module, limitation of a single mode is avoided, a gene expression value and an index sequence are reserved during preprocessing, a rough coding mode is changed, and the cell type identification accuracy is improved; based on a pre-training strategy of comparative learning, matching learning and a cross-modal projection module, fine-grained cross-modal information interaction and sharing are realized, and cross-modal task effects of text generation cells or cell generation texts and the like are optimized.

The invention relates to the technical field of meteorological data processing and artificial intelligence prediction, and discloses an extreme weather event prediction method based on an intelligent model. The method comprises the steps of collecting multi-source meteorological data, and performing preprocessing to unify time and space references of the data; extracting spatio-temporal features reflecting meteorological element change rules based on a preprocessing result, and forming feature vectors for training a deep learning model; inputting fusion data in the real-time platform according to a sliding time window for prediction, generating early warning information according to a prediction result and a set threshold value, and issuing the early warning information through multiple channels; and early warning and monitoring results are fed back to a feature extraction and model updating link to realize online updating and closed-loop processing. According to the method, the timeliness and accuracy of extreme weather event recognition can be improved on the basis of ensuring data consistency through multi-source meteorological data fusion and spatial-temporal feature modeling in combination with conjoint analysis of time and spatial information by deep learning.

The invention discloses a help-calling emergency response system based on an intelligent communication network, particularly relates to the field of intelligent communication, and comprises a multi-source data acquisition layer, a preprocessing and feature extraction layer, a progressive analysis engine, a decision fusion center and an execution control layer. According to the method, the data integrity and the anti-interference capability are remarkably improved, the time-space reference unification module adopts a PTP protocol and a coordinate system conversion technology, accurate alignment of multi-source data is achieved, a multi-dimensional feature matrix is generated through time-frequency conjoint analysis, and the problem that the feature extraction dimension of a traditional system is single is solved; the progressive analysis engine triggers a hierarchical response strategy based on dynamic fusion of a transmission robustness index, a space confidence coefficient and an environmental risk value, combines terrain refraction compensation and a timeliness attenuation mechanism to optimize decision logic, breaks through the limitation of a fixed threshold response mode, adapts to communication quality fluctuation and environmental sudden change, reduces rescue delay risk, and improves rescue efficiency. And the resource scheduling accuracy and the system intelligence level are improved.

The invention is suitable for the technical field of gas transmission and distribution and pipe network safety monitoring, and provides a gas pipe network remote valve regulation and control system and method based on intelligent Internet of Things, and the method comprises the steps: obtaining an initial backflow risk value of a target pipeline region and historical data of a downstream pipe section tail end valve corresponding to the target pipeline region, historical data are analyzed, and a plurality of valve turn-off action samples meeting stable working conditions are extracted; according to the method, a conjoint analysis mechanism based on the difference between the valve turn-off action duration change trend and the valve front pressure micro-drop value is constructed, and a correction factor generation mode of slope deviation judgment and the pressure micro-drop relative difference rate at a specific time node is introduced for the first time; and dynamic perception of the evolution state of the sealing performance of the tail end valve is realized.

The invention relates to the technical field of engineering monitoring, in particular to a large-tonnage pre-stressed anchor cable tensioning quality intelligent evaluation system which comprises a time sequence construction module, a track recognition module, a stability recognition module, a coordination analysis module and a quality evaluation module. According to the method, the trend direction and rate conjoint analysis is carried out on multiple types of response parameters in the anchor cable tensioning process, the response synchronism characteristics in the tensioning period and in the cross-period can be accurately recognized, the anchor cable response track with a time sequence structure is constructed on the basis, accurate recognition of the lag phenomenon and the trend closing relation is achieved, and the accuracy of the anchor cable tensioning process is improved. The method further extracts the dynamic correlation characteristics between the structural disturbance and the tensioning segments, enhances the substantive identification capability of the structural stability risk by means of the time domain coincidence analysis of the disturbance duration of the unstable segments and the falling time consumption, and improves the accuracy of the structural stability risk by normalizing the response rates of the segments with the time overlapping relationship. An index system reflecting the overall coordination level can be established in an anchor cable group.

The invention relates to a servo driver parameter self-tuning oscillation suppression compensation method and related equipment, and the method comprises the following steps: carrying out the multi-channel synchronous collection of an operation state signal of a servo motor, and obtaining a real-time dynamic response data flow; performing time-frequency domain conjoint analysis on the servo motor based on the real-time dynamic response data flow to obtain an oscillation characteristic parameter matrix; performing oscillation source identification on the servo motor based on the oscillation characteristic parameter matrix to obtain an oscillation excitation source classification matrix; performing compensation parameter calculation on the oscillation excitation source classification matrix to obtain a multi-dimensional compensation coefficient set; and performing real-time parameter updating adjustment on the servo driver based on the multi-dimensional compensation coefficient set to obtain a servo control output signal after oscillation suppression, thereby solving the technical problem that a traditional servo driver depends on a fixed parameter setting method, is difficult to adapt to a complex and changeable working condition environment, and cannot effectively suppress an oscillation phenomenon.

The invention provides a method and a device for configuring a motion intention decoding model based on electroencephalogram signal data, which are used for providing a set of novel training mechanism of the motion intention decoding model, and performing conjoint analysis on the electroencephalogram signal data and motion posture data obtained by analyzing videos in the training process. A model system can quickly capture association between neural signals and motion tracks and comprehensively capture motion intentions, the recognition precision of the model for various complex motion modes is remarkably improved, meanwhile, computing resources needed by model processing are reduced, the whole decoding process is more efficient, individual adaptability and generalization are considered, and the method is suitable for popularization and application. And the real-time application requirement of intrusive brain-computer interface research can be well met.

The invention discloses an expressway hot melt marking quality analysis method and system, and relates to the technical field of marking quality analysis. The invention discloses a quality analysis system for a hot-melt marking line of an expressway. The quality analysis system comprises a hot-melt marking line data modeling module and a hot-melt marking line quality analysis module, according to the method, independent tracking data is established for each section of hot melt marking in the same construction period, quality analysis of information such as construction time, space coordinates, raw material batches and technological parameters is ensured, data confusion under different construction conditions can be avoided, and therefore the marking quality analysis and management precision is remarkably improved; through conjoint analysis of three types of data of traffic section indexes, traffic flow indexes and raw material indexes, and by utilizing a full-connection neural network and a hierarchical task analysis model, shared features are extracted, and comprehensive effects of three major elements of environment, traffic and materials on marking quality are revealed.

The invention provides a remote rehabilitation monitoring method and system combined with an Internet of Things intelligent device, and relates to the technical field of visual detection, and the method comprises the steps: monitoring and obtaining a rehabilitation training video and training accompanying sign data of a user through an Internet of Things sensing module; performing risk interference compensation on the reference rehabilitation training video according to the user features, obtaining an optimized reference video, and performing multi-node attitude conjoint analysis on the user rehabilitation training video to obtain a first coefficient; adaptively correcting the first coefficient based on the sign correlation correction function and the training accompanying sign data to obtain a second coefficient; and if the second coefficient does not meet the standard reaching condition, performing feedback interaction optimization on the current scheme. The technical problems in the prior art that the training effect evaluation is not accurate enough and the training scheme is difficult to dynamically optimize and adjust due to the lack of individual difference adaptation ability of a rehabilitation training visual detection mechanism are solved, the accuracy of rehabilitation training visual detection evaluation is improved, and the dynamic optimization ability of the training scheme is enhanced.

The invention discloses a decision control system and method based on temperature field and atmosphere analysis, and relates to the technical field of intelligent control. A monitoring module in the system obtains a temperature set and a gas concentration set; the trend analysis module constructs a three-dimensional model and establishes a coordinate system, a variation function is combined with a Kriging equation to supplement a temperature set to generate an expanded temperature set, the three-dimensional model is discretized into a plurality of nodes and is matched with points in the expanded temperature set through nearest neighbor search, a heat conduction equation and a convection boundary are set, and a finite element matrix equation set is constructed; solving to obtain a temperature field model and carrying out statistics to generate a temperature trend curve; a decision module compares the temperature trend curve, the gas concentration set and a process standard to obtain an abnormal section number, and analyzes a section temperature difference and a concentration error vector of the abnormal section number by adopting a fuzzy PID control algorithm to generate an adjustment vector; and the execution module performs automatic adjustment based on the adjustment vector, so that intelligent decision adjustment based on temperature field and atmosphere conjoint analysis is realized.

The invention relates to the technical field of heterogeneous data analysis, and discloses a data joint debugging method and system based on a heterogeneous monitoring system, and the method comprises the steps: obtaining the original data of each data source in the heterogeneous monitoring system; each data source is used as a node in the graph, edges among the nodes represent association relationships among different data sources, and a knowledge graph of the heterogeneous monitoring system is constructed; by analyzing the dependency relationship between the edges of the nodes and adopting a graph time sequence conjoint analysis method, carrying out delay correlation analysis on the feature data of each node, so that the original data of each node is correlated; and triggering a joint debugging mechanism according to the attention weight and the data feature of the node to complete multi-source data collaborative joint debugging of the heterogeneous monitoring system. The time sequence consistency of multi-source monitoring data, the timeliness of abnormal response and the intelligent level of a joint debugging strategy are remarkably improved, and the method has good adaptability and expandability and can be widely applied to complex monitoring scenes such as industrial Internet of Things and smart cities.

The invention discloses a container resource prediction method and system based on a graph neural network, and the method comprises the steps: building a Kubernetes-based micro-service system, configuring a monitoring platform, and obtaining a log result through simulating the use of a real user, and constructing a data set; on the basis of the data set, constructing a time series prediction G-Times Net model based on a graph neural network, extracting interrelation characteristics between containers, and mining a periodic rule of container resource use; according to the predicted value of the G-Times Net model, carrying out elastic expansion and contraction judgment on the container resources; according to the method, the topological dependency relationship of real-time change among the containers is dynamically captured through the graph neural network, and comprehensive modeling of the container resource use mode is realized by combining the extraction capability of the TimesNet model on the multi-scale periodic features, so that the defects of a traditional method on spatial-temporal feature conjoint analysis are overcome, and the container resource prediction accuracy is remarkably improved.

The invention relates to a water pump scheduling method for estimating lift based on data driving, which comprises the following steps: calling structural information and sensing data, calculating friction loss and structural symmetry, extracting differential pressure data flow gradient, judging load matching degree, activating path nodes, identifying differential pressure data change sections, and mapping a calculation path. According to the method, dynamic correction and compensation of input parameters are achieved by means of conjoint analysis of the structural features of the water inlet and the water outlet and flow pressure difference data, and comprehensive judgment of the working condition matching degree and the periodic change state is combined, so that the flow resistance trend is judged, and the calculation result is corrected. The adaptability to cold load fluctuation and loop temperature difference change is enhanced, the response capability of lift calculation to water body characteristic change is improved by introducing flow resistance trend and viscous state factors, and the calculation stability of the model in different operation sections and the calculation accuracy of the model in complex working conditions are optimized by using a path activation configuration and output switching mechanism.

The invention provides a seabed free site accurate calibration method combining HVSR and 2D-GIT. According to the method, on the basis of submarine seismological observation data, HVSR and 2D-GIT are adopted to obtain submarine site transfer functions respectively, and a submarine free site is accurately calibrated through difference analysis of the site transfer functions; comprising the steps of seabed seismic data acquisition and preprocessing, HVSR spectrum ratio extraction, two-dimensional earth crust attenuation model modeling, 2D-GIT inversion and data fusion, conjoint analysis and environmental parameter correction, site type evaluation and dynamic parameter calibration. The characteristic that the HVSR technology is obviously influenced by local landforms when the horizontal transfer function of the seabed site is evaluated and the characteristic that the 2D-GIT technology can accurately evaluate the horizontal transfer function of each type of seabed site are jointly utilized, and analysis results of the two methods are compared. The method achieves the evaluation of the type of the seabed site and the more accurate and reliable positioning and characteristic evaluation of the free seabed site, effectively makes up for the limitation of a single technology, and provides a comprehensive and accurate type of the seabed site.

The invention relates to the technical field of multi-modal data, in particular to a multi-modal learning method which comprises the following steps: S1, acquiring multi-modal data from various sensors and operation logs through a data acquisition module; s2, preprocessing the multi-modal data through a data processing and fusion module, and performing multi-modal feature fusion by using a cross-modal attention mechanism to obtain fusion features; when the method is used, through joint feature representation, recognition of various faults is achieved, the multi-modal recognition performance is improved, then the accuracy is improved, the method adapts to the change of an operation environment through a dynamic threshold mechanism, misinformation and missing report are effectively reduced, the response speed of operation and maintenance personnel is improved through visual report and task distribution, the operation and maintenance efficiency is conveniently improved, and the operation and maintenance efficiency is improved. Through conjoint analysis of a thermodynamic diagram and stress data, local overheating is accurately detected, an overheating fault is conveniently recognized in real time, meanwhile, the specific position and time of an overheating area are marked in real time, and accurate fault positioning is facilitated.