187 results about "Causal analysis" patented technology

The invention belongs to the technical field of artificial intelligence and engineering, and particularly discloses a tunnel deformation prediction method and system based on causal and space-time mixture graph attention, and the method comprises the steps: receiving monitoring data of a tunnel section, carrying out the data preprocessing of the monitoring data, and obtaining a time sequence; fusing the spatial adjacency relation of the monitoring points and the causal analysis result of the time sequence, generating a graph structure containing physical association and causal dependence, and constructing a weighted adjacency matrix in combination with the geological similarity of the monitoring points; inputting the weighted adjacent matrix and the time sequence into the hybrid network model, extracting spatial features and time sequence features, splicing the spatial features and the time features, inputting the spliced features into a full connection layer, and outputting a prediction result; and carrying out interpretability analysis on a prediction result, dynamically adjusting an early warning threshold value based on statistical distribution of prediction errors, and triggering a graded early warning signal for prompting. According to the invention, the prediction precision of tunnel deformation can be improved.

The invention relates to the technical field of data visualization and integration, in particular to a power equipment fault positioning and maintenance decision-making method based on knowledge graph reasoning, which comprises the following steps of: acquiring a structure parameter generation node, constructing a causal path positioning source component and outputting a maintenance link structure. In the invention, the time attribute is introduced through the node set constructed by the structure number and the operation parameter to strengthen the time sequence characteristic of the power equipment state identification, and the logic response sequence between the nodes is sorted through the causal relationship path chain to improve the accuracy of fault chain reasoning. A multi-path screening mechanism effectively avoids the problems of large path hop count and inconsistent response directions, path dominance identification ensures convergence and pertinence of fault positioning, and key components and maintenance suggested nodes are spliced so that a link structure not only has traceability, but also has operability. An ordered closed loop from structure identification, causal analysis and maintenance output of the power equipment is realized, and the fault processing accuracy and the performability of the power equipment are improved.

The invention discloses an industrial data analysis system and method based on digital twinning and causal inference, and the system comprises a physical sensing layer which is used for collecting multi-source heterogeneous data of an industrial site; the digital twinborn platform layer is used for constructing and operating a virtual twinborn model corresponding to the physical entity; the intelligent analysis engine layer is integrated with a causal analysis module and a federal learning module which are associated; the application and interaction layer is used for visualizing the analysis result and issuing a control instruction; wherein the causal analysis module is used for constructing a causal graph based on the multi-source heterogeneous data and performing causal inference. Through federal learning, on the premise of protecting data privacy of all parties, cross-organization and cross-region collaborative modeling and knowledge sharing are realized.

The invention discloses a power system fault tracing method and system based on data analysis, and relates to the technical field of fault detection, and the method comprises the following steps: collecting multi-source operation data in a target power dispatching system, carrying out the abnormal mode preprocessing, and generating a local abnormal feature sequence of each subsystem; constructing an abnormal feature incidence matrix based on the local abnormal feature sequence, and identifying matrix features based on causal correlation analysis of multi-source operation data to distinguish homologous anomalies and pseudo-synchronization anomalies; based on the identification result, fault boundary identification and anomaly tracing are carried out, and the anomaly level of each subsystem and the corresponding fault source node are output; and generating fault response indication information for scheduling control according to the abnormal level and the fault source node. According to the method, fault traceability and response indication generation are realized through feature association and causal analysis of subsystem abnormity, so that the abnormity judgment precision is improved, and the problems of inaccurate traceability and untimely response of a traditional method are solved.

The invention relates to the technical field of intelligent mine and environmental protection crossing, and discloses a mine development ecological data fusion and causal mining method based on a double-lineage knowledge graph. The method comprises the steps of constructing a double-pedigree domain ontology model comprising a production activity pedigree and an ecological response pedigree, carrying out multi-modal data space-time alignment and standardization, constructing a dynamic map reflecting the real-time state of a mine, predicting future ecological indexes by using a time sequence diagram neural network introduced with physical constraints, and carrying out real-time monitoring on the real-time state of the mine. And calculating the contribution degree of each production link to abnormity by utilizing anti-fact reasoning so as to lock a disaster-causing source, and generating a production regulation and control instruction according to a causal analysis result, and feeding back and executing the production regulation and control instruction. According to the method and the system, full-process traceability and accurate regulation and control of mine environment problems such as surface deformation and water and soil pollution are realized, so that the problems that data multi-source heterogeneous is difficult to fuse and the causal relationship is difficult to confirm in mine ecological environment monitoring are solved.

The invention provides an intelligent identification and alarm method for respiratory suppression events in an anesthesia revival period, which comprises the following steps of: continuously acquiring high-frequency physiological data such as respiration, blood oxygen and electrocardio of a patient through multi-channel equipment, and establishing a dynamic causal network model fusing medical priori knowledge and clinical guidelines after standardized processing and feature extraction; a Granger causal test and a dynamic time warping algorithm are combined, a significant causal relationship among key physiological parameters is dynamically identified, a causal network structure is updated in real time, a causal analysis result is further input into a time sequence Bayesian network, calculation of a respiratory suppression event occurrence probability and reasoning path tracing are realized, and the probability of occurrence of a respiratory suppression event is calculated. According to the method and the system, the probability score is calculated, an interpretable medical logic evidence chain and thermodynamic diagram visualization are automatically generated, and if the probability score exceeds the limit, multi-mode alarm and data locking are synchronously triggered, so that the timeliness, intelligence and interpretability of respiratory suppression detection are improved, and clinical precise intervention is facilitated.

The invention discloses a multi-network cascading failure propagation prediction method based on a heterogeneous graph neural network under a disaster. The method comprises the following steps: acquiring multi-source data of disaster, electric power, communication and traffic networks; constructing a four-layer heterogeneous graph model, and defining multi-type interlayer edges; establishing a mapping relation between disaster physical quantities and physical node health states, and predicting an initial fault state at a disaster impact moment; defining a multi-scale and multi-mechanism propagation rule based on interlayer edges, and simulating a cascading failure process; a space-time heterogeneous graph neural network model is constructed, a heterogeneous graph neural network module and a gating circulation unit module are stacked in the model, and a double-end prediction layer is adopted to predict the continuous operation state and the discrete function state of nodes in an autoregression mode; and post-processing a state time sequence output by the model, and reconstructing a cross-domain cascading failure propagation path through a causal attribution algorithm. According to the method, the whole evolution process from disaster occurrence to multi-network cascading failure can be accurately simulated, and multi-scale failure prediction and causal analysis are realized.

The invention discloses an object full-cycle analysis method and system based on an artificial intelligence large model and a medium, and relates to the field of artificial intelligence, and the method comprises the steps: receiving identification information, and constructing an object digital twinborn body; obtaining and analyzing historical data to obtain an original design intention and a boundary condition of the to-be-analyzed object, and generating an analysis result; the analysis result is written into the historical background dimension of the object digital twinborn body, and provenance analysis is completed; acquiring real-time data and performing suitability comparison on the real-time data and the boundary conditions to obtain a suitability evaluation result; writing the suitability evaluation result into the object digital twinborn body to complete today analysis; generating a risk simulation scene based on the suitability evaluation result; simulation calculation is executed in the risk simulation scene, a deduction result is obtained, and future analysis is completed; the analysis results of the provenance analysis, the today analysis and the future analysis are fused, and a causal analysis chain is established according to the analysis results; and generating a full-cycle analysis report based on the causal analysis chain.

The invention relates to the field of multi-modal neural network model compression, and particularly discloses a multi-modal model compression and distillation method and system based on causal reasoning, and the method comprises the steps: constructing a comprehensive causal discovery module, identifying a causal dependency relationship among the multi-modal features through information theory measurement, Granger causal analysis and intervention-based verification; executing an adaptive compression engine, and performing pruning, mixing precision quantification and low-rank decomposition based on a causal relationship; a cross-modal distiller is applied, and multiple loss function combinations are adopted to maintain the relationship between modals; and implementing a dynamic optimizer to carry out hardware perception and context-sensitive reasoning optimization. According to the method, the compression decision is guided through causal reasoning, the high compression rate is achieved while the key causal path is kept, and the deployment problem of the multi-modal model in the resource-constrained environment is effectively solved.

The invention discloses a water body health assessment system based on big data, relates to the technical field of water body health assessment, and realizes time synchronization, space mapping and quality assessment and restoration of multi-source data, calculation of water quality semantic indexes, data fusion, health level output and sampling work order generation. According to the method, deep fusion and standardized processing of multi-source heterogeneous monitoring data are realized, data islands are broken through, through high-precision space-time alignment and intelligent quality control, data reliability is ensured, human activity events are creatively fused for causal analysis, result interpretation is enhanced, multi-scale dynamic evaluation and adaptive threshold determination are supported, accuracy is improved, and the method is suitable for popularization and application. A closed-loop mechanism of sampling verification is evaluated, the unmanned ship is used for actively checking a suspicious area, secondary judgment is triggered, the response speed and result reliability of emergencies are remarkably improved, the whole-process compliance recording and standardized interface guarantee process is traceable, results are easy to share, and key technical support is provided for fine management of a water body.

The invention discloses a mine gas monitoring and early warning method and system based on a large model, and the method comprises the steps: collecting the multi-modal data under a mine in real time, carrying out the cleaning, denoising, alignment and standardization processing of the multi-modal data, and obtaining the standard multi-modal data; establishing a mine safety knowledge graph, performing SFT supervised fine tuning and field adaptability pre-training by using the mine safety knowledge graph on the basis of the general large language model, and constructing a large model for mine field knowledge enhancement; standard multi-modal data are input into the large model, a multi-modal encoder is used for unified understanding and encoding to extract key information, and panoramic situation description is generated; and based on panoramic situation description, performing causal analysis and risk deduction by using a large model, and generating a mine safety early warning report in combination with a mine safety knowledge graph. The mine gas concentration can be accurately predicted, and the emergency disposal efficiency and effect are improved.

The invention relates to the field of constructional engineering, and discloses a dynamic risk prediction system, which generates space-time alignment input through multi-source data fusion, adopts tensor field modeling to embed contract constraint to construct a risk dynamic model, and solves and outputs a continuous risk field through a partial differential equation; a propagation path is analyzed in combination with asymmetric causal analysis, model parameters are adjusted in real time through a dynamic optimization algorithm, and closed-loop optimization of a risk field is achieved; and finally, through four-dimensional thermodynamic diagram interaction early warning and resource intelligent scheduling, a whole-process closed-loop system of data modeling-causal analysis-dynamic optimization-visual management and control is formed. According to the method, dynamic optimization of risk field parameters is realized through adjoint equation back propagation, and the modeling precision of a complex scene is improved; a four-dimensional space-time thermodynamic diagram rendering technology is innovated to solve the problem of fragmentation of multi-modal information expression, and risk disposal response is accelerated; key task resource supply is guaranteed by combining video memory preemption and containerization scheduling strategies, and the system stability bottleneck in a high-load scene is overcome.

The invention discloses a security event traceability and response management platform suitable for a wind and light storage station. The platform comprises a station edge layer and a cloud middle platform layer. The station edge layer comprises an acquisition and isolation unit, a time synchronization and evidence shaping unit and a credible evidence storage and equipment snapshot unit; the evidence storage module is used for underlying data acquisition, time synchronization, evidence shaping and credible evidence storage to form original event data and an encrypted evidence chain; the cloud middle platform layer comprises a cross-domain causal atlas and root cause analysis engine, a control physical consistency verification model, a hierarchical response and energy storage security state control unit, a drill and strategy verification unit and a redisk and adaptive optimization unit; the method is used for cross-domain causal analysis, control consistency verification, hierarchical response decision and strategy redisk optimization. Unified acquisition, credible recording and cross-domain traceability of security events are realized, event sources and propagation paths can be accurately identified in a complex operation environment, and closed-loop control of security response is supported.

Accuracy, reliability, and response speed improvements for software applications executed by a computing system or platform are provided herein. There are provided systems and methods for multidimensional error causal analysis for error intercorrelations that impact application availability. A service provider may utilize different computing services for data processing to provide different computing services to users, such as via websites and / or applications of the service provider. Due to errors, users may be unable to utilize applications or may face decreased performance and application availability. To improve application performance, error causal analysis may be performed that identifies error intercorrelations that impact application availability and other performance by identifying error effects on each other. Causal statements may be intelligently generated to then identify error intercorrelations. Once generated, these statements may be tested and verified to allow debugging teams and others to fix errors that reduce application performance and availability.

A causal inference stack implements a targeted maximum likelihood scheme to conduct causal analysis of the observational data. At a data-handling layer, the causal inference stack obtains one or more memory locations for a dataset and establishes analysis nodes to setup localized data handling for the memory locations. At a data classification layer, the causal inference stack characterizes the missingness of the dataset. At a pipeline layer, the causal inference stack obtains a data element dependency query from a user and sets up an end-to-end solution path to determine the presence of a causal relationship between data elements identified in the data element dependency query.

The invention relates to the technical field of fault prediction and health management, and discloses a pump station equipment risk intelligent prediction method based on big data analysis, and the method comprises the steps: collecting equipment operation and environment data; a hybrid intelligent causal analysis model for automatic screening is constructed to obtain correlation characteristics among equipment parameters; constructing a risk analysis model which combines graph topology analysis and time sequence characteristics and is calibrated by environmental data, and outputting a systematic operation risk index and a dynamic critical value; and constructing a long-time-sequence prediction model, carrying out long-term trend pre-judgment on the risk index, calling a causal model to carry out root cause tracing when the predicted risk reaches a critical value, and generating a prevention management scheme containing a clear early warning level and a specific resource demand. According to the method, the whole-process intelligentization from unified risk assessment and accurate prediction to closed-loop decision support is realized.

The application relates to the field of multi-modal neural network model compression, and specifically discloses a multi-modal model compression and distillation method and system based on causal reasoning, which comprises the following steps: constructing a comprehensive causal discovery module to identify the causal dependence relationship among multi-modal features through information theory measurement, Granger causality analysis and intervention-based verification; performing an adaptive compression engine to perform pruning, mixed precision quantization and low-rank decomposition based on the causal relationship; applying a cross-modal distiller to maintain the relationship among modes by using a variety of loss function combinations; and implementing a dynamic optimizer to perform hardware perception and context-sensitive reasoning optimization. The application guides the compression decision through causal reasoning, realizes high compression rate while maintaining key causal paths, and effectively solves the deployment problem of multi-modal models in a resource-limited environment.

The invention relates to the technical field of road traffic monitoring and operation and maintenance, in particular to a toll station cloud duty monitoring method and system. The method comprises the following steps: collecting multi-modal data at a toll station, carrying out time sequence alignment, feature extraction and semantic mapping at an edge end to generate event semantic fingerprints, compressing and storing the event semantic fingerprints, and completing anomaly measurement and priority judgment at the same time; the event priority is calculated based on a multi-dimensional scoring mechanism in combination with the anomaly degree, the influence range and the frequency, preliminary anomaly grading is carried out through semantic similarity matching, and an intelligent transmission strategy is formulated; after receiving the high-priority events, the cloud integrates an event set, disassembles semantic fingerprints to construct a cross-modal causal incidence matrix, identifies a causal chain and extracts a high-causal-strength event combination to generate a synthetic event vector for anomaly verification; and feeding back a cloud causal analysis result to an edge end, adaptively adjusting a screening weight threshold by the edge, and checking an iterative optimization strategy through cloud edge consistency. According to the invention, abnormal real-time response and resource optimization are realized.

The invention belongs to the technical field of large language models, and discloses an improved semantic unit-based heterogeneous government affair knowledge graph construction and multi-dimensional causal retrieval enhancement method, which comprises the following steps of: extracting an entity and relationship triple from an unstructured text through a thinking chain CoT technology guided by a large language model LLM; a semantic unit is introduced to serve as a fine-grained knowledge carrier, and a double-track heterogeneous graph structure is constructed; the method comprises the following steps: constructing three semantic dimensions of an entity, a theme and a global, respectively designing adaptive retrieval strategies, executing retrieval and matching in parallel, and realizing accurate fusion and efficient retrieval of multi-source knowledge through a multi-channel concurrent mechanism; and a causal analysis report mechanism is further introduced, semantic refining and redundancy filtering are performed on a fusion result, and the accuracy and interpretability of a question and answer result are improved.

A distributed photovoltaic grid-connected point electric energy quality dynamic monitoring system relates to the field of distributed photovoltaic grid-connected points, and comprises a knowledge graph construction module used for constructing a dynamic knowledge graph based on a dynamic database and updating the dynamic knowledge graph by using real-time data and integrated dynamic knowledge; the causal analysis module is used for establishing a causal relationship model of the power quality problem based on the dynamic knowledge graph; and the backtracking and positioning module is used for performing causal chain backtracking based on the causal relationship model, positioning a root cause influencing the power quality and generating an analysis report. According to the method, the causal chain backtracking is carried out along the weighted directed acyclic graph by taking the abnormal index as a starting point based on the causal relationship model, so that the root cause node, such as a specific equipment fault or health deterioration, which causes the power quality problem can be quickly and automatically positioned, and the automation degree and the positioning precision of fault diagnosis are greatly improved.

The invention relates to the technical field of elevator equipment detection, and discloses an elevator traveling cable dynamic fault detection method, system and device and a medium, and the method comprises the following steps: obtaining multi-mode state data of each line of a traveling cable and elevator operation state data; inputting the operation state data into a preset electromechanical coupling model to obtain a corresponding multi-dimensional health reference signal vector; performing time alignment on the multi-modal state data and the reference vector, and performing subtraction to obtain a multi-dimensional signal residual sequence; inputting the residual error sequence into a pre-training deep learning classification model to obtain a preliminary fault judgment result; a final result is confirmed through causal analysis; and if the fault exists, target fault information is obtained through the time domain reflectometer, and visual information is generated and sent to the human-computer interaction interface. According to the embodiment of the invention, the electrical characteristics of the cable and the mechanical motion state of the elevator are dynamically coupled, and high-precision and high-efficiency fault early warning and diagnosis are realized.

The invention relates to the technical field of computers, and provides a causal relationship analysis method and device, equipment and a medium, and the method comprises the steps: generating an initial causal network structure according to an association relationship between variables in observation data; based on the number of shared adjacent nodes of the variable nodes in the initial causal network structure in the domain knowledge graph and the path length between the nodes, determining the semantic association degree between the variable nodes; according to the semantic association degree, adjusting the confidence degree of causal edges in the initial causal network structure, and generating a target causal graph; and calculating causal effect intensity among the variable nodes based on the target causal graph, and generating a causal analysis result by using the causal effect intensity. According to the method, the deviation of the initial causal network structure is corrected by utilizing a mechanism of fusing the observation data and the knowledge graph, and the causal effect intensity is accurately calculated based on the target causal graph, so that a causal analysis result containing an accurate causal relationship and a quantitative influence degree can be provided for a user.

The invention provides a rolling bearing fault fusion diagnosis method based on tendency score matching causal inference, which comprises the following steps: firstly, acquiring running state time sequence data of a rolling bearing in different states by a multi-source sensor, performing filtering denoising and feature extraction, and then forming a causal analysis data feature matrix by using features; calculating tendency scores of different state samples of the rolling bearing in the causal analysis data feature matrix; matching the fault state sample and the normal state sample by using the tendency score, and carrying out multi-dimensional balance verification on a matched pair; training a rolling bearing fault recognition model by taking the verified rolling bearing state sample as a training sample; and finally, after feature extraction is carried out on actually-collected rolling bearing data, obtained feature data are input into the trained rolling bearing fault recognition model, and a fault recognition result is obtained. According to the invention, the interpretability and adaptability of the fault identification model under complex working conditions are improved.

The invention belongs to the technical field of intelligent fault diagnosis, and particularly discloses a composite fault detection and diagnosis system and method based on a cloud-side-end collaborative AI intelligent agent.The composite fault detection and diagnosis system comprises a terminal sensing layer, an edge operation layer and a cloud brain layer, and a sensing module of the terminal layer collects and preprocesses multi-source industrial signals; an edge layer is embedded into a distributed cooperative Transform which is innovatively designed and is used for extracting features of multiple subsystems and reducing signal interference among the subsystems; the cloud layer takes a large language model as a core and undertakes the functions of reasoning decision, task scheduling and man-machine interaction. The method can dynamically adapt to various tasks such as fault detection, diagnosis and man-machine interaction, the generalization ability of unseen composite fault combinations is improved, the communication cost is reduced, meanwhile, a comprehensive diagnosis report containing causal analysis, risk grading and maintenance suggestions is output, and the method is suitable for system-level composite fault detection and diagnosis of a complex mechanical system.

The invention provides a brain function network causal analysis method based on phase-space reconstruction and unified GCA, and relates to the field of functional brain network analys.The method comprises the steps that fMRI data are collected and preprocessed, and a time sequence of interested nodes is extracted from the preprocessed fMRI data; for extracting time sequences X and Y of any two to-be-analyzed interested nodes, constructing a variable time delay unified Granger causal model based on phase space reconstruction; and traversing all to-be-analyzed node pairs of interest, calculating the causal direction and strength between each pair of nodes to construct a whole-brain directed causal connection matrix, and performing network metric attribute analysis. According to the method, phase-space reconstruction is taken as a core, a causal analysis framework is provided by unifying GCA, and end-to-end modeling is realized. The final target is to generate a high-fidelity fMRI data model, so that the causal connection relationship is closer to a brain real neural mechanism, and the reliability and the application value of functional brain network research are improved.

The invention relates to the technical field of aluminum ash recycling digital twinning, and discloses an aluminum ash recycling treatment method based on a high-aluminum material. According to the method, a digital twinborn body in a processing process is constructed, and a virtual probe array corresponding to a key reaction site of a real device is set in the digital twinborn body, so that dynamic evolution information of aluminum ash processing is continuously captured. An evolution chain of a physical field and a material flow is constructed based on the information, then a processing path is deduced reversely, and key evolution nodes and constraint conditions are identified. The nodes and conditions are converted into a control strategy set, and a real processing system is driven to execute closed-loop regulation and control. According to the method, deep perspective of the internal evolution mechanism in the aluminum ash treatment process is achieved, accurate regulation and control are achieved based on causal analysis, and the stability and the resource recovery rate of the treatment process are improved.

The invention discloses an emotion recognition method and system based on semantic analysis, and relates to the technical field of semantic data processing.The emotion recognition method comprises the steps that a semantic database for emotion recognition analysis is established, and semantic data is obtained, marked and stored according to the semantic database; a semantic database is established, a mode library based on external emotion connection and an event library of activities arranged by psychological rehabilitators are constructed, different semantic data are continuously brought into the semantic database, and an emotion expression model is constructed to predict and understand complex emotion expression according to changes of external emotion connection and campus activity scenes. According to the method, features related to emotions in semantic data are captured more accurately, emotional expressions and new emotional situations are adapted, texts are analyzed into event triples, a common sense causal rule base is matched to generate an emotional decision chain, explainable reasoning of knowledge is achieved, and a dynamic emotional thermodynamic diagram is generated through emotional state evolution tracking. And the causal analysis is converted into an early warning decision.

The invention discloses a feed pump energy-saving optimization method and system based on variable frequency speed regulation, and relates to the technical field of industrial process control and energy-saving optimization, and the method comprises the steps: constructing a causal analysis model by extracting residual statistical characteristics, recognizing the main drive type of carbon emission change, quantifying the causal effect, and estimating the life loss in combination with sensitivity; energy consumption and service life target weights are dynamically adjusted according to causal labels, a dual-objective optimization function is constructed, a multi-objective genetic algorithm is adopted to determine control parameters of frequency adjustment, and frequency self-adaptive adjustment is achieved. The offset abnormal driving reason is effectively identified, and the accuracy of energy-saving control and the equipment service life management and control capability are improved.

The invention discloses a power plant automatic operation and maintenance management system based on big data analysis, and the system comprises the following steps: the system constructs a standardized time sequence through collecting equipment operation, environment and operation and maintenance record data, updates an operation state feature structure in real time through employing a subspace tracking algorithm, recognizes the operation characteristic offset, and carries out the operation and maintenance of a power plant. In combination with Granger causal analysis, a causal weight matrix between time sequence variables is constructed, a modeling process is embedded, dynamic subspace structure adjustment is realized, the system further identifies a state abnormal section based on subspace angle change, and a fault link is constructed in combination with a causal path. Accurate identification of the operation and maintenance state of the power plant and intelligent presentation of a potential problem chain are realized. The method is suitable for an operation and maintenance management scene of a power plant complex system.

This invention relates to the field of data analysis technology, specifically to a precise analysis method and system for urban health checks based on AI-driven multimodal data collaboration. The method includes the following steps: constructing an adaptive fractal spatiotemporal grid to map urban multimodal data sources into spatiotemporal encoded vectors with fractal dimensions; establishing a physics-driven co-resonance network to extract abnormal co-resonance patterns exceeding normal resonance thresholds; calculating modal entropy chain values ​​based on the real-time metabolic rate of the urban system and generating an optimal weight matrix using a non-equilibrium thermodynamic model; and fusing abnormal co-resonance patterns and the dynamic weight matrix to generate a spatiotemporal causal graph displaying the origin and propagation path of faults using a causal discovery algorithm. This invention, utilizing multimodal data collaboration and causal analysis technology, enhances the emergency response capability of urban systems in the face of sudden faults, ensuring the stability of urban operations. The spatiotemporal causal graph provides comprehensive visualization information for the fault propagation process.