Dynamic risk analysis method and system based on knowledge graph and digital twin model

Abstract

The application relates to the technical field of risk analysis, in particular to a dynamic risk analysis method and system based on a knowledge graph and a digital twin model. The method comprises the following steps: obtaining a preset three-dimensional twin model of a hydrogen cylinder; determining a key weak area of the hydrogen cylinder according to the preset three-dimensional twin model; predicting a potential failure mode of the hydrogen cylinder based on a preset knowledge graph and the key weak area; simulating a risk evolution state of the hydrogen cylinder according to the failure mode, and determining a transportation strategy according to the simulation result. The method avoids data fragmentation, lays a foundation for the risk evolution state, ensures accurate and timely input data, reduces the cost of comprehensive scanning, ensures that knowledge graph queries and risk simulation are concentrated on high-probability failure points, improves overall analysis efficiency, provides operable input for risk evolution state simulation, enhances the reliability and pertinence of risk identification, realizes the dynamic correlation of risk prevention and control and transportation decision, and improves the initiative of safety management.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of risk analysis, in particular to a dynamic risk analysis method and system based on a knowledge graph and a digital twin model. BACKGROUND

[0002] Hydrogen, as a clean and efficient energy carrier, is increasingly widely used in the field of transportation. However, hydrogen cylinders, as key components for storing and transporting hydrogen, face various failure risks under complex working conditions, including hydrogen embrittlement, delamination of composite materials, and bulging and collapse of plastic liners.

[0003] However, the existing hydrogen cylinder risk analysis method can only predict the approximate failure area based on a knowledge graph. This approach can lead to a rough strategy when facing complex scenarios, which cannot support real-time decision-making and increases the risk of safety accidents. SUMMARY

[0004] The application provides a dynamic risk analysis method and system based on a knowledge graph and a digital twin model to solve the above problems.

[0005] In a first aspect, the application provides a dynamic risk analysis method based on a knowledge graph and a digital twin model, which comprises:

[0006] acquiring a preset three-dimensional twin model of a hydrogen cylinder;

[0007] determining a key weak area of the hydrogen cylinder according to the preset three-dimensional twin model;

[0008] predicting a potential failure mode of the hydrogen cylinder based on a preset knowledge graph according to the key weak area;

[0009] simulating a risk evolution state of the hydrogen cylinder according to the failure mode, and determining a transportation strategy according to the simulation result.

[0010] According to the preset three-dimensional twin model of the hydrogen cylinder, the unified digital representation is ensured, the data fragmentation is avoided, and the foundation for the risk evolution state is laid, ensuring that the input data is accurate and timely. According to the preset three-dimensional twin model, the key weak area of the hydrogen cylinder is determined, the overhead of comprehensive scanning is reduced, the knowledge graph query and risk simulation are concentrated on the high-probability failure point, and the overall analysis efficiency is improved. Based on the preset knowledge graph, the potential failure mode of the hydrogen cylinder is predicted according to the key weak area, which provides an operable input for the risk evolution state simulation and enhances the reliability and pertinence of risk identification. According to the failure mode, the risk evolution state of the hydrogen cylinder is simulated, and the transportation strategy is determined according to the simulation result, realizing the dynamic correlation of risk prevention and control and transportation decision-making, and improving the initiative of safety management.

[0011] Optionally, the preset three-dimensional twin model of the hydrogen cylinder is obtained, comprising:

[0012] Component information of the hydrogen cylinder is obtained; the component information is analyzed to determine component geometric parameters and component material properties;

[0013] According to the component geometric parameters and the component material properties, a stress-strain curve under a hydrogen environment is constructed;

[0014] According to the component material properties, fiber orientation and matrix performance of a composite material are determined;

[0015] According to the component geometric parameters and the component material properties, the thickness and hydrogen permeation coefficient of the plastic liner are determined;

[0016] According to the stress-strain curve, the fiber orientation, the matrix performance, the thickness and the hydrogen permeation coefficient, the preset three-dimensional twin model is generated.

[0017] By this scheme, the component information of the hydrogen cylinder is obtained, ensuring that the preset three-dimensional twin model construction process has complete initial information support. The component information is analyzed to determine the component geometric parameters and the component material properties, providing accurate input parameters for constructing the stress-strain curve, eliminating data ambiguity and supporting parameterized modeling. According to the component geometric parameters and the component material properties, the stress-strain curve under the hydrogen environment is constructed, reflecting the actual deformation characteristics under the hydrogen environment. According to the component material properties, the fiber orientation and the matrix performance of the composite material are determined, supporting accurate characterization of the failure behavior of the composite material. According to the component geometric parameters and the component material properties, the thickness and hydrogen permeation coefficient of the plastic liner are determined, ensuring that the preset three-dimensional twin model accurately simulates the physical and chemical behavior of the liner. According to the stress-strain curve, the fiber orientation, the matrix performance, the thickness and the hydrogen permeation coefficient, the preset three-dimensional twin model is generated, supporting failure simulation and strategy optimization.

[0018] Optionally, the preset three-dimensional twin model is generated according to the stress-strain curve, the fiber orientation, the matrix performance, the thickness and the hydrogen permeation coefficient, comprising:

[0019] According to the stress-strain curve, hydrogen embrittlement sensitivity is determined;

[0020] Guided wave monitoring data under cyclic load is obtained;

[0021] Based on the fiber orientation and the matrix performance, the guided wave monitoring data under the cyclic load is used to determine the debonding probability of the composite layer;

[0022] Acoustic emission signals of the plastic liner are obtained;

[0023] analyze the acoustic emission signal based on the thickness and the hydrogen permeation coefficient to determine a bulge collapse critical value;

[0024] generate the preset three-dimensional twin model according to the hydrogen embrittlement sensitivity, the debonding probability and the bulge collapse critical value.

[0025] According to the stress-strain curve, the hydrogen embrittlement sensitivity is determined, and the prediction accuracy of hydrogen embrittlement failure is improved. The guided wave monitoring data under cyclic load is obtained to ensure the response to dynamic working condition changes. Based on the fiber orientation and the matrix performance, the debonding probability of the composite layer is determined according to the guided wave monitoring data under cyclic load, and the dynamic capture ability of the delamination failure is enhanced. The acoustic emission signal of the plastic liner is obtained to reflect the actual degradation process of the plastic liner. Based on the thickness and the hydrogen permeation coefficient, the acoustic emission signal is analyzed to determine the bulge collapse critical value, and the prediction reliability of the instability of the liner caused by hydrogen permeation is improved. According to the hydrogen embrittlement sensitivity, the debonding probability and the bulge collapse critical value, the preset three-dimensional twin model is generated, the defects of insufficient modeling integrity and weak decision support are eliminated, and the failure mode detail simulation and real-time strategy optimization are supported.

[0026] Optionally, the potential failure mode of the hydrogen cylinder is predicted according to the key weak area based on a preset knowledge graph, including:

[0027] Obtain historical failure cases;

[0028] Analyze the historical failure cases to determine historical damage modes;

[0029] Analyze the preset knowledge graph to determine node correlation;

[0030] According to the node correlation, determine the coupling effect between the composite layer and the metal layer;

[0031] Obtain visual inspection data, analyze the visual inspection data, and determine the wear condition of the key weak area;

[0032] According to the wear condition, the coupling effect and the historical damage mode, the potential failure mode of the hydrogen cylinder is predicted.

[0033] By the scheme, historical failure cases are obtained, and blind reasoning is avoided, so that the empirical nature and relevance of the prediction are enhanced. By analyzing the historical failure cases and determining the historical damage mode, the accuracy and consistency of the prediction are improved. By analyzing the preset knowledge graph and determining the node correlation, it is ensured that the hidden failure correlation path is captured in the prediction process. According to the node correlation, the coupling effect between the composite layer and the metal layer is determined, and the cross-material failure transmission path is determined. The visual detection data is obtained, the visual detection data is analyzed, the wear condition of the key weak area is determined, and the contribution of the actual use trace to the failure is reflected. According to the wear condition, the coupling effect and the historical damage mode, the potential failure mode of the hydrogen cylinder is predicted, and is associated with the operable failure type and position.

[0034] Optionally, the determining the transportation strategy according to the simulation result comprises:

[0035] Obtaining vibration spectrum data of a current transportation route;

[0036] Analyzing the vibration spectrum data to determine a resonance frequency coincidence degree of the vibration spectrum and the hydrogen embrittlement sensitivity;

[0037] According to the resonance frequency coincidence degree, constructing a detour instruction, and determining the detour instruction as the transportation strategy.

[0038] By the scheme, the vibration spectrum data of the current transportation route is obtained, the potential vibration source causing the hydrogen embrittlement of the metal part is identified, and the defect that the decision support is weak is eliminated. The vibration spectrum data is analyzed to determine the resonance frequency coincidence degree of the vibration spectrum and the hydrogen embrittlement sensitivity, in response to the deficiency that the vibration spectrum data cannot be dynamically associated with the hydrogen embrittlement sensitivity, the quantitative identification of the resonance risk is realized, and an operable basis for generating the detour instruction is provided. According to the resonance frequency coincidence degree, the detour instruction is constructed, and the detour instruction is determined as the transportation strategy, thereby actively avoiding the high-risk scene caused by vibration resonance, eliminating the problem that the high-risk scene cannot be actively avoided in the transportation process, and improving the initiative of risk prevention and control.

[0039] Optionally, the constructing the detour instruction according to the resonance frequency coincidence degree comprises:

[0040] Obtaining a transportation task of the hydrogen cylinder;

[0041] Analyzing the transportation task to determine a backup route;

[0042] Obtaining route information of the backup route; analyzing the route information to determine road flatness and slope data;

[0043] Analyzing the road flatness and the slope data to determine a real-time influence on the hydrogen cylinder;

[0044] According to the real-time influence and the resonance frequency coincidence degree, constructing a detour instruction.

[0045] By this scheme, the transportation task of the hydrogen cylinder is obtained, and the risk assessment process is started based on the actual logistics demand. The transportation task is analyzed, the backup route is determined, the feasible alternative path options are provided, and the flexibility of the transportation strategy is ensured. The route information of the backup route is obtained, the route data is enriched, and the digital twin model accurately maps the influence of the external environment on the hydrogen cylinder. The route information is analyzed, the road flatness and slope data are determined, and the potential effect of the road condition on the hydrogen cylinder is evaluated. The real-time influence of the road flatness and slope data on the hydrogen cylinder is analyzed to provide a basis for decision-making, and the risk prediction is associated with the current route condition in real time. According to the real-time influence and the resonance frequency coincidence degree, the detour instruction is constructed, the dynamic optimization of the transportation strategy is realized, and high-risk scenarios are actively avoided.

[0046] Optionally, the analysis of the preset knowledge graph to determine the node association includes:

[0047] Based on the historical failure cases, the preset knowledge graph is analyzed to determine the associated nodes;

[0048] Stress distribution data of the hydrogen cylinder is obtained;

[0049] The stress distribution data is matched with the historical failure cases, and the load similarity is determined according to the matching result;

[0050] According to the similarity, the association weight between any associated nodes is determined;

[0051] According to the association weight, the node association is determined.

[0052] Through this scheme, based on historical failure cases, the preset knowledge graph is analyzed to determine associated nodes, eliminating the defect of incomplete modeling integrity, ensuring that metal component hydrogen embrittlement sensitivity, composite layer risk, and other nodes are included in the analysis. The stress distribution data of the hydrogen cylinder is obtained, which makes up for the defect of being unable to accurately capture thermal stress failure. The stress distribution data is matched with the historical failure cases, and the load similarity is determined according to the matching result, eliminating the defect of lacking accurate simulation of failure details. According to the similarity, the association weight between any associated nodes is determined, eliminating the problem of missing group hazard state prediction, supporting group risk modeling. According to the association weight, the node association is determined, responding to the defect of weak decision support, and providing the basis for the association of the transportation strategy.

[0053] Optionally, after the detour instruction is constructed according to the real-time influence and the resonance frequency coincidence degree, it further includes:

[0054] The transportation task is analyzed to determine the task timeliness requirement;

[0055] analyze the bypass instruction, determine a predicted delay time of the bypass path;

[0056] determine a safety gain according to the real-time influence;

[0057] determine a bypass necessity according to the predicted delay time and the safety gain based on the time limit requirement of the task.

[0058] According to the scheme, the transportation task is analyzed, the time limit requirement of the task is determined, and the evaluation is ensured to be within the time limit of the task, so as to avoid decision deviation caused by ignoring the time limit constraint. The bypass instruction is analyzed, the predicted delay time of the bypass path is determined, the time cost of the bypass strategy is quantified, and the safety improvement degree of the bypass strategy is provided. According to the real-time influence, the safety gain is determined, the safety improvement degree of the bypass strategy is reflected, and the risk relief benefit of the decision is provided. Based on the time limit requirement of the task, the predicted delay time and the safety gain are determined, the bypass necessity is determined, the dynamic optimization of the transportation strategy is realized, and the high-risk scene is actively avoided under the premise of meeting the time limit.

[0059] Optionally, after the analysis of the acoustic emission signal based on the thickness and the hydrogen permeation coefficient, the method further comprises:

[0060] acquiring real-time temperature data of the plastic liner;

[0061] analyzing the temperature data to determine a permeation influence of temperature on the hydrogen permeation coefficient;

[0062] updating the bubble collapse critical value according to the permeation influence.

[0063] According to the scheme, the real-time temperature data of the plastic liner is acquired, and the correction of the bubble collapse critical value is ensured to be based on the actual environmental conditions. The temperature data is analyzed, the permeation influence of temperature on the hydrogen permeation coefficient is determined, and the permeation rate calculation deviation caused by temperature fluctuation is eliminated. According to the permeation influence, the bubble collapse critical value is updated, and the prediction accuracy of the plastic liner bubble collapse is improved.

[0064] In a second aspect, the application provides a dynamic risk analysis system based on a knowledge graph and a digital twin model, the system comprising:

[0065] a model acquisition module configured to acquire a preset three-dimensional twin model of a hydrogen cylinder;

[0066] a region determination module configured to determine a key weak region of the hydrogen cylinder according to the preset three-dimensional twin model;

[0067] a mode prediction module configured to predict a potential failure mode of the hydrogen cylinder based on a preset knowledge graph and the key weak region;

[0068] The transport determination module is configured to simulate a risk evolution state of the hydrogen cylinder according to the failure mode, and determine a transport strategy according to a simulation result.

[0069] Optionally, when the model acquisition module acquires the preset three-dimensional twin model of the hydrogen cylinder, the model acquisition module is configured to:

[0070] acquire component information of the hydrogen cylinder, analyze the component information to determine component geometric parameters and component material properties;

[0071] construct a stress-strain curve in a hydrogen environment according to the component geometric parameters and the component material properties;

[0072] determine fiber orientation and matrix performance of a composite material according to the component material properties;

[0073] determine a thickness of a plastic liner and a hydrogen permeation coefficient according to the component geometric parameters and the component material properties;

[0074] generate the preset three-dimensional twin model according to the stress-strain curve, the fiber orientation, the matrix performance, the thickness, and the hydrogen permeation coefficient.

[0075] Optionally, when the model acquisition module generates the preset three-dimensional twin model according to the stress-strain curve, the fiber orientation, the matrix performance, the thickness, and the hydrogen permeation coefficient, the model acquisition module is configured to:

[0076] determine hydrogen embrittlement sensitivity according to the stress-strain curve;

[0077] acquire guided wave monitoring data under cyclic loading;

[0078] determine a debonding probability of a composite layer according to the guided wave monitoring data under the cyclic loading based on the fiber orientation and the matrix performance;

[0079] acquire an acoustic emission signal of the plastic liner;

[0080] analyze the acoustic emission signal to determine a bulge collapse critical value based on the thickness and the hydrogen permeation coefficient;

[0081] generate the preset three-dimensional twin model according to the hydrogen embrittlement sensitivity, the debonding probability, and the bulge collapse critical value.

[0082] Optionally, when the mode prediction module predicts a potential failure mode of the hydrogen cylinder based on a preset knowledge graph according to the key weak area, the mode prediction module is configured to:

[0083] acquire historical failure cases;

[0084] analyze the historical failure cases to determine historical damage modes;

[0085] analyze the preset knowledge graph to determine node correlation;

[0086] determine the coupling effect between the composite layer and the metal layer according to the node correlation;

[0087] obtain visual inspection data, analyze the visual inspection data, and determine the wear condition of the key weak area;

[0088] According to the wear condition, the coupling effect and the historical damage mode, the potential failure mode of the hydrogen cylinder is predicted.

[0089] Optionally, when the transport determination module determines the transport strategy according to the simulation result, it is used for:

[0090] Obtain the vibration spectrum data of the current transportation route;

[0091] Analyze the vibration spectrum data to determine the resonance frequency coincidence degree of the vibration spectrum and the hydrogen embrittlement sensitivity;

[0092] According to the resonance frequency coincidence degree, the detour instruction is constructed, and the detour instruction is determined as the transport strategy.

[0093] Optionally, when the transport determination module constructs the detour instruction according to the resonance frequency coincidence degree, it is used for:

[0094] Obtain the transportation task of the hydrogen cylinder;

[0095] Analyze the transportation task to determine the backup route;

[0096] Obtain the route information of the backup route; analyze the route information to determine the road flatness and slope data;

[0097] Analyze the road flatness and the slope data to determine the real-time influence on the hydrogen cylinder;

[0098] According to the real-time influence and the resonance frequency coincidence degree, the detour instruction is constructed.

[0099] Optionally, when the mode prediction module analyzes the preset knowledge graph to determine the node correlation, it is used for:

[0100] Based on the historical failure cases, analyze the preset knowledge graph to determine the associated nodes;

[0101] Obtain the stress distribution data of the hydrogen cylinder;

[0102] Match the stress distribution data with historical failure cases, and determine the load similarity according to the matching result;

[0103] According to the similarity, determine the association weight between any associated nodes;

[0104] According to the association weight, determine the node association.

[0105] Optionally, the dynamic risk analysis system based on the knowledge graph and the digital twin model further comprises a necessity determination module configured to:

[0106] Analyze the transportation task to determine a task timeliness requirement;

[0107] Analyze the detour instruction to determine a predicted delay time of the detour path;

[0108] According to the real-time influence, determine a safety gain;

[0109] Based on the task timeliness requirement, determine a detour necessity according to the predicted delay time and the safety gain.

[0110] Optionally, the dynamic risk analysis system based on the knowledge graph and the digital twin model further comprises a critical value updating module configured to:

[0111] Obtain real-time temperature data of the plastic liner;

[0112] Analyze the temperature data to determine a permeation influence of temperature on the hydrogen permeation coefficient;

[0113] According to the permeation influence, update the bulge collapse critical value. BRIEF DESCRIPTION OF DRAWINGS

[0114] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and those skilled in the art can also obtain other drawings according to these drawings without creative labor.

[0115] Figure 1 An application scenario schematic diagram provided by an embodiment of the present application;

[0116] Figure 2 A flowchart of a dynamic risk analysis method based on a knowledge graph and a digital twin model provided by an embodiment of the present application;

[0117] Figure 3 A dynamic risk analysis system structure schematic diagram based on a knowledge graph and a digital twin model provided by an embodiment of the present application. DETAILED DESCRIPTION

[0118] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0119] In addition, the term "and / or" in the present application is only used to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can represent the following three cases: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in the present application generally represents an "or" relationship between the associated objects, unless otherwise specified.

[0120] The embodiments of the present application will be further described below with reference to the drawings of the specification.

[0121] The existing hydrogen cylinder risk analysis method can only predict the approximate failure area based on the knowledge graph. This way will lead to a rough strategy when facing complex scenarios, which cannot support real-time decision-making and increase the risk of safety accidents.

[0122] Therefore, the present application provides a dynamic risk analysis method and system based on a knowledge graph and a digital twin model. The preset three-dimensional twin model of the hydrogen cylinder is obtained to ensure a unified digital representation, avoid data fragmentation, lay a foundation for risk evolution state, and ensure accurate and timely input data. According to the preset three-dimensional twin model, the key weak area of the hydrogen cylinder is determined to reduce the overhead of comprehensive scanning, ensure that the knowledge graph query and risk simulation are concentrated on the high-probability failure point, and improve the overall analysis efficiency. Based on the preset knowledge graph, the potential failure mode of the hydrogen cylinder is predicted according to the key weak area to provide an operable input for risk evolution state simulation and enhance the reliability and pertinence of risk identification. According to the failure mode, the risk evolution state of the hydrogen cylinder is simulated, and the transportation strategy is determined according to the simulation result to realize the dynamic correlation of risk prevention and control and transportation decision-making and improve the initiative of safety management.

[0123] Figure 1 An application scenario diagram provided by the present application is applied when performing dynamic risk analysis.

[0124] Specifically, the method provided by the application is applied to any server, the server interacts with a three-dimensional modeling tool, and a preset three-dimensional twin model of the hydrogen cylinder is constructed through the three-dimensional modeling tool. According to the preset three-dimensional twin model, the key weak area of the hydrogen cylinder is determined. Based on the preset knowledge graph, the potential failure mode of the hydrogen cylinder is predicted according to the key weak area, an operable input is provided for risk evolution state simulation, and the reliability and pertinence of risk identification are enhanced. According to the failure mode, the risk evolution state of the hydrogen cylinder is simulated, and according to the simulation result, the transportation strategy is determined, the dynamic correlation of risk prevention and control and transportation decision is realized, and the initiative of safety management is improved. The specific implementation mode can refer to the following embodiments.

[0125] Figure 2 A flowchart of a dynamic risk analysis method based on a knowledge graph and a digital twin model is provided for an embodiment of the application. The method of the embodiment can be applied to the server in the above scenario. As shown in the figure, the method comprises: Figure 2

[0126] S201, obtaining a preset three-dimensional twin model of a hydrogen cylinder;

[0127] The hydrogen cylinder can be a pressure container for storing and transporting hydrogen.

[0128] The preset three-dimensional twin model can be a pre-constructed virtual mapping model of the hydrogen cylinder. It is pre-stored in the server and called when used.

[0129] Specifically, based on the design data and historical operation data of the hydrogen cylinder, the geometric data of the hydrogen cylinder is imported through a three-dimensional modeling tool, and the material attribute parameters are input. Then, the model state is dynamically updated in combination with real-time sensor data, so as to construct the preset three-dimensional twin model of the hydrogen cylinder.

[0130] S202, determining the key weak area of the hydrogen cylinder according to the preset three-dimensional twin model;

[0131] The key weak area can be a physical part of the hydrogen cylinder prone to failure.

[0132] Specifically, stress-strain simulation is performed on the preset three-dimensional twin model by using finite element analysis, stress distribution under different working conditions is calculated, and high stress areas are identified. Then, based on reliability engineering theory, typical failure cases accumulated for a long time are stored in a structured manner to generate a preset database. Then, the historical failure records in the preset database are queried, and similar structural features in the preset three-dimensional twin model are matched. Finally, the analysis results are summarized and analyzed according to the clustering algorithm, and the key weak area of the hydrogen cylinder is determined.

[0133] S203, predicting the potential failure mode of the hydrogen cylinder according to the key weak area based on the preset knowledge graph; ​

[0134] The preset knowledge graph can be a structured knowledge base pre-constructed to organize entities and associated relationships in historical failure cases in a graph structure. The pre-stored in the server is called when used.

[0135] The potential failure mode can be a failure type that the hydrogen cylinder can occur.

[0136] Specifically, based on the key weak area, according to the ontology and semantic network theory, the entity relationship modeling of the historical failure case is carried out, so as to construct the preset knowledge graph; further, the related failure mode in the preset knowledge graph is searched; then, the rule reasoning engine is applied, and the real-time environmental data collected by the environmental sensor is combined to predict the potential failure mode of the hydrogen cylinder.

[0137] S204, according to the failure mode, simulate the risk evolution state of the hydrogen cylinder, and determine the transportation strategy according to the simulation result.

[0138] The risk evolution state can be a dynamic development process of the failure behavior.

[0139] The simulation result can be the output data of the risk evolution state simulation.

[0140] The transportation strategy can be an executable optimization instruction generated according to the simulation result.

[0141] Specifically, the risk evolution state of the failure mode is simulated by using a computer simulation device: first, according to the failure mode, the failure position and size are determined; then, combined with the real-time environmental data, the failure direction, rate and damage range are calculated. Based on the simulation result, according to the real-time transportation data collected by the vehicle-mounted sensor and positioning device, combined with the application of heuristic algorithm, the transportation strategy is generated.

[0142] Through the scheme, the preset three-dimensional twin model of the hydrogen cylinder is obtained, ensuring that the unified digital representation is based on, avoiding data fragmentation, and at the same time laying a foundation for the risk evolution state, ensuring that the input data is accurate and timely. According to the preset three-dimensional twin model, the key weak area of the hydrogen cylinder is determined, the overhead of comprehensive scanning is reduced, the knowledge graph query and risk simulation are concentrated on the high probability failure point, and the overall analysis efficiency is improved. Based on the preset knowledge graph, according to the key weak area, the potential failure mode of the hydrogen cylinder is predicted, which provides an operable input for the risk evolution state simulation, enhances the reliability and pertinence of risk identification. According to the failure mode, the risk evolution state of the hydrogen cylinder is simulated, and the transportation strategy is determined according to the simulation result, realizing the dynamic association of risk prevention and control and transportation decision, and improving the initiative of safety management.

[0143] In some embodiments, component information of the hydrogen cylinder is acquired; the component information is parsed to determine component geometric parameters and component material properties; a stress-strain curve under a hydrogen environment is constructed according to the component geometric parameters and the component material properties; fiber orientation and matrix performance of a composite material are determined according to the component material properties; thickness and a hydrogen permeation coefficient of a plastic liner are determined according to the component geometric parameters and the component material properties; and a preset three-dimensional twin model is generated according to the stress-strain curve, the fiber orientation, the matrix performance, the thickness, and the hydrogen permeation coefficient.

[0144] The component information can be descriptive data of each component of the hydrogen cylinder.

[0145] The component geometric parameters can be size, shape, and position parameters of the components of the hydrogen cylinder.

[0146] The component material properties can be material characteristic parameters of the components of the hydrogen cylinder.

[0147] The stress-strain curve can be a curve describing the deformation behavior of a material when subjected to stress.

[0148] The composite material can be a material structure composed of fibers and a matrix.

[0149] The fiber orientation can be the arrangement direction of the fibers in the composite material.

[0150] The matrix performance can be the performance characteristics of the matrix material in the composite material.

[0151] The plastic liner can be a plastic layer inside the hydrogen cylinder.

[0152] The thickness can be the thickness dimension of the plastic liner.

[0153] The hydrogen permeation coefficient can be a parameter describing the rate of hydrogen penetration through a material.

[0154] Specifically, the component information of the hydrogen cylinder is extracted by calling the data access interface. Then, the component information is parsed by using regular expression matching to determine the geometric parameters and component material properties. Subsequently, based on the geometric parameters and component material properties, a material mechanics model constructed by applying classical material mechanics theory under the condition of simulating a hydrogen environment is applied; further, combined with the hydrogen environment influence factors extracted from the standard material manual; then, according to the interpolation method, the stress-strain curve is generated. Subsequently, the component material properties are parsed to identify the attribute items of the composite material; then, the material test report is queried to extract the fiber orientation and matrix performance. Subsequently, based on the geometric parameters and component material properties, the material test report is queried, and the thickness and hydrogen permeation coefficient of the plastic liner are determined through data calculation. Further, based on the geometric parameters, the three-dimensional geometric structure of the hydrogen cylinder is constructed according to the parameterized modeling device; finally, the stress-strain curve is applied to set the mechanical behavior, the fiber orientation and matrix performance are applied to set the behavior of the composite material layer, and the thickness and hydrogen permeation coefficient are applied to set the behavior of the plastic liner, and the simulation engine is integrated to generate the preset three-dimensional twin model.

[0155] By the scheme, the component information of the hydrogen cylinder is obtained, ensuring that the preset three-dimensional twin model construction process has complete initial information support. The component information is parsed to determine the component geometric parameters and component material properties, providing accurate input parameters for constructing the stress-strain curve, eliminating data ambiguity and supporting parameterized modeling. According to the component geometric parameters and component material properties, the stress-strain curve under the hydrogen environment is constructed to reflect the actual deformation characteristics under the hydrogen environment. According to the component material properties, the fiber orientation and matrix performance of the composite material are determined to support accurate characterization of the failure behavior of the composite material. According to the component geometric parameters and component material properties, the thickness and hydrogen permeation coefficient of the plastic liner are determined to ensure that the preset three-dimensional twin model accurately simulates the physical and chemical behavior of the liner. According to the stress-strain curve, the fiber orientation, the matrix performance, the thickness, and the hydrogen permeation coefficient, the preset three-dimensional twin model is generated to support failure simulation and strategy optimization.

[0156] In some embodiments, according to the stress-strain curve, the hydrogen embrittlement sensitivity is determined; the guided wave monitoring data under cyclic loading is obtained; based on the fiber orientation and the matrix performance, the debonding probability of the composite layer is determined according to the guided wave monitoring data under cyclic loading; the acoustic emission signal of the plastic liner is obtained; based on the thickness and the hydrogen permeation coefficient, the acoustic emission signal is analyzed to determine the bulging collapse critical value; according to the hydrogen embrittlement sensitivity, the debonding probability, and the bulging collapse critical value, the preset three-dimensional twin model is generated.

[0157] The hydrogen embrittlement sensitivity can be a numerical index quantifying the susceptibility of a metal component to hydrogen embrittlement failure in a hydrogen environment.

[0158] The cyclic load can be a load condition of repeatedly applying mechanical stress.

[0159] Guided wave monitoring data can include signal data such as wave velocity, attenuation coefficient, and reflection characteristics.

[0160] The composite layer can be a composite material layer structure found in hydrogen cylinders.

[0161] The debonding probability can be a value representing the probability of interlayer separation occurring in the composite layer.

[0162] Acoustic emission signals can be acoustic signal data such as event counts, amplitudes, and frequency spectra.

[0163] The bulging and collapse threshold can be the critical pressure threshold at which the plastic inner liner bulges or collapses and fails.

[0164] Specifically, based on the stress-strain curve, mechanical parameters such as yield strength and elongation at break are extracted. Then, the deformation characteristics of the stress-strain curve under hydrogen environment are calculated to determine the hydrogen embrittlement sensitivity. Subsequently, guided wave signal data is collected in real time under cyclic loading using guided wave sensors on the hydrogen cylinder composite layer. Then, based on fiber orientation and matrix properties, combined with guided wave monitoring data under cyclic loading, regression analysis is used to calculate the debonding probability of the composite layer. Next, acoustic emission signals of the plastic liner are collected in real time under the operating state of the hydrogen cylinder using acoustic emission sensors on the surface of the plastic liner. Subsequently, signal analysis algorithms are used to extract features from the acoustic emission signals, and the critical stress threshold is calculated by combining the thickness and hydrogen permeability coefficient; based on the critical stress threshold, the critical value for bulging collapse is determined. Finally, a preset three-dimensional twin model is automatically generated by integrating the hydrogen embrittlement sensitivity, debonding probability, and bulging collapse critical value using a simulation engine.

[0165] This scheme determines hydrogen embrittlement sensitivity based on stress-strain curves, improving the prediction accuracy of hydrogen embrittlement failure. Guided wave monitoring data under cyclic loading is acquired to ensure responsiveness to dynamic operating conditions. Based on fiber orientation and matrix properties, the debonding probability of the composite layer is determined using guided wave monitoring data under cyclic loading, enhancing the dynamic capture capability of delamination failure. Acoustic emission signals from the plastic liner are acquired to reflect the actual degradation process of the plastic liner. Based on thickness and hydrogen permeability coefficient, the acoustic emission signals are analyzed to determine the critical value for bulging and collapse, improving the predictive reliability of liner instability caused by hydrogen permeation. Based on hydrogen embrittlement sensitivity, debonding probability, and bulging and collapse critical value, a pre-defined three-dimensional twin model is generated, eliminating deficiencies in modeling completeness and weak decision support, supporting detailed simulation of failure modes and real-time strategy optimization.

[0166] In some embodiments, historical failure cases are acquired; the historical failure cases are parsed to determine historical damage patterns; a preset knowledge graph is analyzed to determine node relevance; the node relevance is used to determine coupling effects between the composite layer and the metal layer; visual detection data are acquired and analyzed to determine wear conditions of key weak areas; and the wear conditions, the coupling effects, and the historical damage patterns are used to predict potential failure modes of the hydrogen cylinder.

[0167] The historical failure cases can be records of failure events of the hydrogen cylinder during historical transportation.

[0168] The historical damage patterns can be structured patterns formed by failure feature classification.

[0169] The node relevance can be a quantitative indicator of mutual influence strength between nodes.

[0170] The metal layer can be a metal component in the hydrogen cylinder.

[0171] The coupling effects can be dynamic behavior descriptions of physical interactions between the composite layer and the metal layer.

[0172] The visual detection data can be surface image data of the hydrogen cylinder.

[0173] The wear conditions can be quantitative descriptions of surface damage states of the key weak areas.

[0174] Specifically, historical failure cases are retrieved from a pre-stored database. Then, the historical failure case data are parsed to extract failure features such as failure types, positions, and severities; further, the failure features are integrated to determine historical damage patterns. Subsequently, based on a preset knowledge graph, a graph traversal algorithm is executed to analyze relationships between nodes, thereby determining node relevance. Then, the node relevance is used to identify the relevance between the composite layer and the metal layer; further, according to the relevance, physical interactions are deduced to determine coupling effects between the composite layer and the metal layer. Then, visual detection data are acquired by a visual sensor; further, the visual detection data are image-processed to identify visual features of key weak areas; subsequently, according to the visual features, wear conditions of the key weak areas are quantified. Finally, a reasoning engine of the preset knowledge graph is applied to match the wear conditions, the coupling effects, and the historical damage patterns; then, failure evolution is simulated to predict potential failure modes of the hydrogen cylinder.

[0175] By the scheme, historical failure cases are obtained to avoid reasoning by empty, thereby enhancing the positivism and relevance of the prediction. The historical failure cases are analyzed to determine the historical damage mode, thereby improving the accuracy and consistency of the prediction. The preset knowledge graph is analyzed to determine the node relevance, thereby ensuring that the prediction process captures the hidden failure correlation path. According to the node relevance, the coupling effect between the composite layer and the metal layer is determined to clearly show the cross-material failure transmission path. The visual detection data is obtained and analyzed to determine the wear condition of the key weak area, thereby reflecting the contribution of the actual use trace to the failure. According to the wear condition, the coupling effect and the historical damage mode, the potential failure mode of the hydrogen cylinder is predicted, and is associated to the operable failure type and position.

[0176] In some embodiments, vibration spectrum data of the current transportation route is obtained; the vibration spectrum data is analyzed to determine the resonance frequency coincidence degree of the vibration spectrum and the hydrogen embrittlement sensitivity; and the bypass instruction is constructed according to the resonance frequency coincidence degree, and the bypass instruction is determined as the transportation strategy.

[0177] The current transportation route can be the path actually traveled by the hydrogen cylinder transportation vehicle.

[0178] The vibration spectrum data can be data representing the vibration characteristics of the transportation route.

[0179] The vibration spectrum can be a frequency distribution feature for describing the core frequency component of the vibration.

[0180] The resonance frequency coincidence degree can be a matching degree quantification index between the vibration spectrum and the resonance frequency of the hydrogen embrittlement sensitivity.

[0181] The bypass instruction can be an action command for avoiding a high-risk section or selecting an alternative path.

[0182] Specifically, the vibration spectrum data of the current transportation route is collected by a data collection device. Then, the vibration spectrum of the vibration spectrum data is compared with the resonance frequency of the hydrogen embrittlement sensitivity. Subsequently, the matching degree of the vibration spectrum and the hydrogen embrittlement sensitivity is calculated. Then, the resonance frequency and the matching degree are integrated to determine the resonance frequency coincidence degree. Then, a preset risk threshold is set according to the statistical analysis of the historical failure cases. Subsequently, it is judged whether the resonance frequency coincidence degree exceeds the preset risk threshold. If the resonance frequency coincidence degree exceeds the preset risk threshold, the bypass instruction is generated. Finally, the bypass instruction is transmitted to the transportation management device to generate the transportation strategy.

[0183] By the scheme, the vibration spectrum data of the current transportation route is obtained, the potential vibration source causing hydrogen embrittlement of the metal part is identified, and the defect of weak decision support is eliminated. The vibration spectrum data is analyzed to determine the resonance frequency coincidence degree of the vibration spectrum and the hydrogen embrittlement sensitivity, in response to the deficiency that the vibration spectrum data cannot be dynamically associated with the hydrogen embrittlement sensitivity, the resonance risk is quantitatively identified, and an operable basis is provided for generating the bypass instruction. According to the resonance frequency coincidence degree, the bypass instruction is constructed, the bypass instruction is determined as the transportation strategy, the high-risk scene caused by vibration resonance is actively avoided, the problem that the high-risk scene cannot be actively avoided in the transportation process is eliminated, and the initiative of risk prevention and control is improved.

[0184] In some embodiments, a transportation task of the hydrogen cylinder is obtained; the transportation task is analyzed to determine a backup route; route information of the backup route is obtained; the route information is parsed to determine road flatness and slope data; the road flatness and slope data are analyzed to determine real-time influence on the hydrogen cylinder; and a bypass instruction is constructed according to the real-time influence and the resonance frequency coincidence degree.

[0185] The transportation task can be a transportation plan of the hydrogen cylinder from a starting point to an ending point.

[0186] The backup route can be a candidate alternative path for avoiding a high-risk area.

[0187] The route information can be attribute data of the backup route, such as road type, length, real-time traffic state, historical accident data, and environmental factors.

[0188] The road flatness can be a quantitative indicator of the roughness of the road surface.

[0189] The slope data can be the inclination of the road.

[0190] The real-time influence can be the dynamic effect of the road flatness and slope data on the hydrogen cylinder.

[0191] Specifically, the transportation task of the hydrogen cylinder is received through an application programming interface of a transportation management device. Then, the transportation task is processed using a path planning algorithm, and historical route data in a preset knowledge graph is combined to generate a backup route. Subsequently, route information of the backup route is queried through a map service API. Then, based on the route information, the road flatness is determined by analyzing historical vibration spectrum data; at the same time, the slope data is obtained based on laser scanning data. Then, based on real-time sensor data, a physical rule model constructed based on the material properties of the hydrogen cylinder is applied to analyze the vibration amplitude caused by the road flatness and the pressure change caused by the slope data, thereby determining the real-time influence on the hydrogen cylinder. Then, the real-time influence and the resonance frequency coincidence degree are integrated; finally, a decision logic model constructed by a rule engine of the preset knowledge graph is used to determine whether the integrated value exceeds a preset risk threshold; if the preset risk threshold is exceeded, a bypass instruction is generated

[0192] By this scheme, the transportation task of the hydrogen cylinder is obtained, and the risk assessment process is started based on the actual logistics demand. The transportation task is analyzed, the backup route is determined, the feasible alternative path options are provided, and the flexibility of the transportation strategy is ensured. The route information of the backup route is obtained, the route data is enriched, and the digital twin model accurately maps the influence of the external environment on the hydrogen cylinder. The route information is analyzed, the road flatness and slope data are determined, and the potential effect of the road conditions on the hydrogen cylinder is evaluated. The real-time influence of the road flatness and slope data on the hydrogen cylinder is analyzed to provide a basis for decision-making, and the risk prediction is associated with the current route conditions in real time. According to the real-time influence and the resonance frequency coincidence degree, the detour instructions are constructed, the dynamic optimization of the transportation strategy is realized, and high-risk scenarios are actively avoided.

[0193] In some embodiments, based on historical failure cases, the preset knowledge graph is analyzed to determine associated nodes; stress distribution data of the hydrogen cylinder is obtained; the stress distribution data is matched with the historical failure cases, and based on the matching result, a load similarity is determined; based on the similarity, an association weight between any associated nodes is determined; and based on the association weight, node association is determined.

[0194] The associated nodes can be entity nodes and associated relationships related to the historical failure cases in the preset knowledge graph.

[0195] The stress distribution data can be stress value distribution data of each part of the hydrogen cylinder.

[0196] The matching result can be a similarity score output after matching the stress distribution data with the historical failure cases.

[0197] The load similarity can be a numerical indicator quantifying the similarity degree of the current stress load and the stress load of the historical failure cases.

[0198] The association weight can be a numerical weight representing the relationship strength between any associated nodes.

[0199] Specifically, according to the type of the historical failure cases, a graph traversal algorithm is used to match the nodes stored in the preset knowledge graph, and entity nodes and associated relationships are extracted to determine the associated nodes. Then, the application programming interface of the digital twin model constructed by the virtual mapping theory of physical entities reads the stress distribution data of the hydrogen cylinder in real time. Subsequently, the stress feature vector of the historical failure cases is extracted, and a similarity calculation algorithm is used to match the stress distribution data with the historical failure cases; then, based on the matching result, the load similarity is determined. Further, based on the load similarity, a weight update rule is applied to adjust the association weight of the edges between any associated nodes in the preset knowledge graph. Finally, the association weight is processed by a weighted average algorithm to determine the node association.

[0200] By this scheme, based on historical failure cases, the preset knowledge graph is analyzed, the associated nodes are determined, the defects of insufficient modeling integrity are eliminated, and the nodes such as hydrogen embrittlement sensitivity of metal parts and risk of composite layer are ensured to be included in the analysis. The stress distribution data of the hydrogen cylinder is obtained, and the defect of being unable to accurately capture the thermal stress failure is remedied. The stress distribution data is matched with the historical failure cases, and according to the matching result, the load similarity is determined, and the defect of lacking accurate simulation of failure details is eliminated. According to the similarity, the association weight between any associated nodes is determined, the problem of missing group risk modeling is eliminated, and the group risk modeling is supported. According to the association weight, the node association is determined, the defect of weak decision support is responded, and the association basis is provided for the transportation strategy.

[0201] In some embodiments, the transportation task is analyzed to determine a task timeliness requirement; the detour instruction is analyzed to determine a predicted delay time of the detour path; a safety gain is determined according to real-time influence; and the detour necessity is determined based on the task timeliness requirement, the predicted delay time and the safety gain.

[0202] The task timeliness requirement can be a time constraint requirement of the transportation task.

[0203] The detour path can be an alternative transportation route recommended to avoid risks.

[0204] The predicted delay time can be a predicted delay time relative to the originally planned path.

[0205] The safety gain can be a degree of risk reduction or a quantitative value of safety improvement.

[0206] The detour necessity can be a decision conclusion of whether to execute the detour instruction.

[0207] Specifically, the database of the transportation task is accessed, and the planned completion time or the deadline of the transportation task is extracted as the task timeliness requirement. Then, the travel time of the new path is estimated by time calculation according to the path coordinates and road condition parameters in the detour instruction; subsequently, the predicted delay time of the detour path is determined by comparing with the reference time of the original path. Based on real-time influence, the safety gain is calculated by a safety evaluation algorithm. Finally, the task timeliness requirement, the predicted delay time and the safety gain are integrated, and the detour necessity is evaluated using a decision algorithm; if the detour necessity is “necessary”, the detour instruction is determined as the transportation strategy.

[0208] By the scheme, the transportation task is analyzed, the task time limit requirement is determined, the evaluation is ensured to be carried out within the task time limit, and the decision deviation caused by ignoring the time limit constraint is avoided. The bypass instruction is analyzed, the predicted delay time of the bypass path is determined, the time cost of the bypass strategy is quantified, and the basis for weighing the safety benefit and time loss is provided. According to the real-time influence, the safety gain is determined, the safety improvement degree of the bypass strategy is reflected, and the measurable risk relief benefit for decision-making is provided. Based on the task time limit requirement, the predicted delay time and the safety gain are determined, and the necessity of bypass is determined, so as to realize the dynamic optimization of the transportation strategy and ensure that the high-risk scene is actively avoided under the premise of meeting the time limit.

[0209] In some embodiments, real-time temperature data of the plastic liner is acquired; the temperature data is analyzed to determine the permeation influence of temperature on the hydrogen permeation coefficient; and the drum collapse critical value is updated according to the permeation influence.

[0210] The real-time temperature data can be a current temperature value collected in real time on the surface of the plastic liner of the hydrogen cylinder.

[0211] The permeation influence can be a quantitative effect of temperature change on the hydrogen permeation coefficient.

[0212] Specifically, the temperature sensor on the surface of the plastic liner of the hydrogen cylinder is used for real-time collection; then, the real-time temperature data of the plastic liner is acquired through the data transmission interface. Subsequently, the pre-stored hydrogen permeation coefficient-temperature relationship model of the plastic liner material is called based on the experimental data, the real-time temperature data is input into the hydrogen permeation coefficient-temperature relationship model, and the permeation influence of the hydrogen permeation coefficient at the current temperature is calculated. Further, the hydrogen permeation coefficient is adjusted according to the permeation influence; then, the thickness of the plastic liner and the acoustic emission signal are combined; further, the drum collapse critical value calculation model is constructed based on the preset knowledge graph, the physical parameters mapped by the digital twin model, and the theoretical reasoning of the historical failure cases; finally, the adjusted hydrogen permeation coefficient, the thickness, and the acoustic emission signal are input into the drum collapse critical value calculation model to update the drum collapse critical value.

[0213] By the scheme, the real-time temperature data of the plastic liner is acquired, and the correction of the drum collapse critical value is ensured to be based on the actual environmental conditions. The temperature data is analyzed to determine the permeation influence of temperature on the hydrogen permeation coefficient, and the permeation rate calculation deviation caused by temperature fluctuation is eliminated. The drum collapse critical value is updated according to the permeation influence, and the prediction accuracy of the plastic liner drum collapse is improved.

[0214] Figure 3 The structure schematic diagram of a dynamic risk analysis system based on a knowledge graph and a digital twin model provided by an embodiment of the present application is as shown in Figure 3As shown, the dynamic risk analysis system 300 based on the knowledge graph and the digital twin model of the embodiment comprises a model acquisition module 301, a region determination module 302, a mode prediction module 303, and a transportation determination module 304.

[0215] The model acquisition module 301 is configured to acquire a preset three-dimensional twin model of a hydrogen cylinder.

[0216] The region determination module 302 is configured to determine a key weak area of the hydrogen cylinder according to the preset three-dimensional twin model.

[0217] The mode prediction module 303 is configured to predict a potential failure mode of the hydrogen cylinder based on a preset knowledge graph and the key weak area.

[0218] The transportation determination module 304 is configured to simulate a risk evolution state of the hydrogen cylinder according to the failure mode, and determine a transportation strategy according to a simulation result.

[0219] Optionally, when the model acquisition module 301 acquires the preset three-dimensional twin model of the hydrogen cylinder, the model acquisition module 301 is configured to:

[0220] acquire component information of the hydrogen cylinder, analyze the component information to determine component geometric parameters and component material properties;

[0221] construct a stress-strain curve under a hydrogen environment according to the component geometric parameters and the component material properties;

[0222] determine fiber orientation and matrix performance of a composite material according to the component material properties;

[0223] determine a thickness of a plastic liner and a hydrogen permeation coefficient according to the component geometric parameters and the component material properties;

[0224] generate the preset three-dimensional twin model according to the stress-strain curve, the fiber orientation, the matrix performance, the thickness, and the hydrogen permeation coefficient.

[0225] Optionally, when the model acquisition module 301 generates the preset three-dimensional twin model according to the stress-strain curve, the fiber orientation, the matrix performance, the thickness, and the hydrogen permeation coefficient, the model acquisition module 301 is configured to:

[0226] determine hydrogen embrittlement sensitivity according to the stress-strain curve;

[0227] acquire guided wave monitoring data under cyclic load;

[0228] determine a debonding probability of a composite layer according to the guided wave monitoring data under the cyclic load based on the fiber orientation and the matrix performance;

[0229] acquire an acoustic emission signal of the plastic liner;

[0230] analyze the acoustic emission signal based on the thickness and the hydrogen permeation coefficient to determine a bulge collapse critical value;

[0231] generate the preset three-dimensional twin model according to the hydrogen embrittlement sensitivity, the debonding probability, and the bulge collapse critical value.

[0232] Optionally, when the mode prediction module 303 predicts the potential failure mode of the hydrogen cylinder based on the preset knowledge graph and according to the key weak area, it is used for:

[0233] acquire historical failure cases;

[0234] analyze the historical failure cases to determine historical damage modes;

[0235] analyze the preset knowledge graph to determine node relevance;

[0236] determine the coupling effect between the composite layer and the metal layer according to the node relevance;

[0237] acquire visual inspection data, analyze the visual inspection data, and determine the wear condition of the key weak area;

[0238] predict the potential failure mode of the hydrogen cylinder according to the wear condition, the coupling effect, and the historical damage mode.

[0239] Optionally, when the transportation determination module 304 determines the transportation strategy according to the simulation result, it is used for:

[0240] acquire vibration spectrum data of the current transportation route;

[0241] analyze the vibration spectrum data to determine the resonance frequency coincidence degree of the vibration spectrum and the hydrogen embrittlement sensitivity;

[0242] construct a detour instruction according to the resonance frequency coincidence degree, and determine the detour instruction as the transportation strategy.

[0243] Optionally, when the transportation determination module 304 constructs the detour instruction according to the resonance frequency coincidence degree, it is used for:

[0244] acquire a transportation task of the hydrogen cylinder;

[0245] analyze the transportation task to determine a backup route;

[0246] acquire route information of the backup route; analyze the route information to determine road flatness and slope data;

[0247] Analyze the road flatness and the slope data to determine a real-time influence on the hydrogen cylinder;

[0248] Construct a detour instruction according to the real-time influence and the resonance frequency coincidence degree.

[0249] Optionally, when the mode prediction module 303 analyzes the preset knowledge graph to determine the node correlation, it is used for:

[0250] Based on the historical failure cases, analyze the preset knowledge graph to determine the associated nodes;

[0251] Obtain stress distribution data of the hydrogen cylinder;

[0252] Match the stress distribution data with historical failure cases, and determine the load similarity according to the matching result;

[0253] According to the similarity, determine the correlation weight between any associated nodes;

[0254] According to the correlation weight, determine the node correlation.

[0255] Optionally, the dynamic risk analysis system based on knowledge graph and digital twin model further includes a necessity determination module 305, which is used for:

[0256] Analyze the transportation task to determine the task time requirement;

[0257] Analyze the detour instruction to determine the predicted delay time of the detour path;

[0258] According to the real-time influence, determine the safety gain;

[0259] Based on the task time requirement, determine the necessity of detour according to the predicted delay time and the safety gain.

[0260] Optionally, the dynamic risk analysis system based on knowledge graph and digital twin model further includes a critical value updating module 306, which is used for:

[0261] Obtain real-time temperature data of the plastic liner;

[0262] Analyze the temperature data to determine the permeation influence of temperature on the hydrogen permeation coefficient;

[0263] According to the permeation influence, update the bulge collapse critical value.

[0264] The system of the embodiment can be used to execute the method of any of the above embodiments, and has similar implementation principles and technical effects, which will not be described here.

Claims

1. A dynamic risk analysis method based on a knowledge graph and a digital twin model, characterized in that, The method comprises the following steps: acquire a preset three-dimensional twin model of a hydrogen cylinder, including: acquire component information of the hydrogen cylinder; analyze the component information to determine component geometric parameters and component material properties; construct a stress-strain curve under a hydrogen environment according to the component geometric parameters and the component material properties; determine the fiber orientation and matrix performance of the composite material according to the component material properties; determine the thickness of the plastic liner and the hydrogen permeation coefficient according to the component geometric parameters and the component material properties; generate the preset three-dimensional twin model according to the stress-strain curve, the fiber orientation, the matrix performance, the thickness, and the hydrogen permeation coefficient, including: determine the hydrogen embrittlement sensitivity according to the stress-strain curve; acquire guided wave monitoring data under cyclic loading; determine the debonding probability of the composite layer based on the fiber orientation and the matrix performance according to the guided wave monitoring data under cyclic loading; acquire acoustic emission signals of the plastic liner; analyze the acoustic emission signals based on the thickness and the hydrogen permeation coefficient to determine the blister collapse critical value; generate the preset three-dimensional twin model according to the hydrogen embrittlement sensitivity, the debonding probability, and the blister collapse critical value; determine the key weak area of the hydrogen cylinder according to the preset three-dimensional twin model; predict the potential failure mode of the hydrogen cylinder based on a preset knowledge graph according to the key weak area; simulate the risk evolution state of the hydrogen cylinder according to the failure mode, and determine the transportation strategy according to the simulation result, including: acquire vibration spectrum data of the current transportation route; analyze the vibration spectrum data to determine the resonance frequency coincidence degree of the vibration spectrum and the hydrogen embrittlement sensitivity; the resonance frequency coincidence degree is a matching degree quantitative index between the vibration spectrum and the resonance frequency of the hydrogen embrittlement sensitivity; construct a detour instruction according to the resonance frequency coincidence degree, and determine the detour instruction as the transportation strategy.

2. The method of claim 1, wherein, The method of predicting the potential failure mode of the hydrogen cylinder based on a preset knowledge graph according to the key weak area comprises: acquire historical failure cases; analyze the historical failure cases to determine historical damage modes; analyze the preset knowledge graph to determine node correlation; determine the coupling effect between the composite layer and the metal layer according to the node correlation; acquire visual inspection data, analyze the visual inspection data, and determine the wear condition of the key weak area; predict the potential failure mode of the hydrogen cylinder according to the wear condition, the coupling effect, and the historical damage mode.

3. The method of claim 2, wherein, The method of constructing a detour instruction according to the resonance frequency coincidence degree comprises: acquire a transportation task of the hydrogen cylinder; analyze the transportation task to determine a backup route; acquire route information of the backup route; analyze the route information to determine road flatness and slope data; analyze the road flatness and the slope data to determine the real-time influence on the hydrogen cylinder; construct a detour instruction according to the real-time influence and the resonance frequency coincidence degree.

4. The method of claim 3, wherein, The method of analyzing the preset knowledge graph to determine node correlation comprises: analyze the preset knowledge graph based on the historical failure cases to determine associated nodes; Obtain stress distribution data of the hydrogen cylinder; Match the stress distribution data with historical failure cases, and determine the load similarity according to the matching result; Determine the association weight between any associated nodes according to the similarity; Determine the node association according to the association weight.

5. The method of claim 4, wherein, After constructing the bypass instruction according to the real-time influence and the resonance frequency coincidence degree, it further includes: Analyze the transportation task to determine the task time limit requirement; Analyze the bypass instruction to determine the expected delay time of the bypass path; Determine the safety gain according to the real-time influence; Determine the necessity of bypassing based on the task time limit requirement, the expected delay time and the safety gain.

6. The method of claim 1, wherein, After analyzing the acoustic emission signal based on the thickness and the hydrogen permeation coefficient to determine the blister collapse critical value, it further includes: Obtain real-time temperature data of the plastic liner; Analyze the temperature data to determine the permeation influence of temperature on the hydrogen permeation coefficient; Update the blister collapse critical value according to the permeation influence.

7. A dynamic risk analysis system based on a knowledge graph and a digital twin model, characterized in that, Applied to the method of any one of claims 1-6, comprising: A model acquisition module is configured to obtain a preset three-dimensional twin model of a hydrogen cylinder; A region determination module is configured to determine the key weak region of the hydrogen cylinder based on the preset three-dimensional twin model; A mode prediction module is configured to predict the potential failure mode of the hydrogen cylinder based on the key weak region and a preset knowledge graph; A transportation determination module is configured to simulate the risk evolution state of the hydrogen cylinder based on the failure mode, and determine a transportation strategy based on the simulation result.

Images