A confined space hydrogen system cascade accident monitoring method and device

Abstract

The application discloses a kind of restricted space hydrogen system cascade accident monitoring method and device, the method includes: establishing hydrogen electricity coupling cascade accident whole-process fast prediction model;According to the evolution law of cable fire temperature field and the hydrogen system combustion risk threshold, the safe interval of cable fire and the safe operation interval of hydrogen system are quantified respectively, to form the double-dimension coupling safety boundary of electricity side-hydrogen side;Combining the risk area distribution result of hydrogen electricity coupling cascade accident whole-process fast prediction model, the collaborative optimization arrangement scheme of heat sensor and hydrogen concentration sensor in restricted space is determined;According to hydrogen electricity coupling cascade accident whole-process fast prediction model, the double-dimension coupling safety boundary of electricity side-hydrogen side and collaborative optimization arrangement scheme, build restricted space hydrogen system cascade accident analysis system, carry out real-time monitoring, risk pre-judgment, consequence fast prediction and early warning of hydrogen electricity coupling cascade accident, realize the fast and accurate prediction of cascade accident consequence.

Description

Technical Field

[0001] This invention belongs to the field of hydrogen safety and industrial fire prevention technology, specifically a method and device for monitoring cascade accidents in confined space hydrogen systems. Background Technology

[0002] With the rapid development of the hydrogen energy industry, hydrogen production, storage, and utilization equipment are often integrated into confined spaces, such as hydrogen refueling station buildings, hydrogen energy workshops, and vehicle-mounted hydrogen system cabins. Hydrogen leaks within confined spaces can easily accumulate to the explosive limit, potentially triggering a combustion and explosion upon contact with an ignition source. The resulting overpressure and thermal radiation can further impact surrounding hydrogen-related equipment, triggering a domino effect and causing catastrophic consequences.

[0003] Current research on the domino effect of hydrogen leakage primarily focuses on hydrogen leakage as the initial accident source, emphasizing the simulation and prediction of hydrogen leakage diffusion and explosion consequences. However, it neglects the most common initial hazard in confined spaces—electrical cable fires. In actual engineering scenarios, most industrial fires in confined spaces are caused by electrical wire fires. Since the electrical systems associated with hydrogen-related equipment are distributed throughout the confined space, a complete hydrogen-electric coupling cascade accident chain is formed: "electrical wire fire → combustion heat release → sudden temperature rise in confined space → thermal structural deformation of hydrogen-related equipment → sealing failure and hydrogen leakage → hydrogen explosion → overpressure and thermal radiation triggering more hydrogen equipment failures → multi-stage domino accident."

[0004] Current research on this coupling chain faces significant technological gaps: First, there is a lack of a full-chain numerical simulation method that combines FDS modeling of electrical wire fires with hydrogen system combustion and explosion simulation, making it impossible to fully reveal the evolution of hydrogen-electric coupling cascade accidents. Second, existing prediction models do not consider the hydrogen leakage process caused by thermally induced structural deformation, making it impossible to achieve rapid prediction of the entire process from initial cable fire to the final domino accident. Third, there is a lack of a two-dimensional safety interval quantification method for hydrogen-electric coupling risks, as well as a sensor collaborative deployment scheme adapted to the accident evolution characteristics, making it impossible for monitoring systems to achieve early warning. Fourth, there is a lack of integrated systems and devices that integrate hazard identification, accident prediction, real-time monitoring, and early warning and prevention, making it difficult to meet the actual safety management needs of engineering sites.

[0005] Therefore, developing a method for monitoring cascading accidents in confined space hydrogen systems is of great significance for improving the inherent safety level of confined space hydrogen systems and preventing major hydrogen energy safety accidents. Summary of the Invention

[0006] To address the technical deficiencies of the existing technologies, this invention provides a method and device for monitoring cascading accidents in confined space hydrogen systems. It comprehensively covers the entire accident chain of "cable fire → thermal deformation → hydrogen leakage → explosion → domino effect" through coupled cable fire FDS simulation, hydrogen-related equipment thermal-structural coupling simulation, hydrogen leakage and combustion-explosion fluid dynamics simulation, and artificial neural network technology, enabling rapid and accurate prediction of cascading accident consequences. Simultaneously, it quantifies the dual-dimensional safety range of the electrical and hydrogen sides, proposes a sensor collaborative optimization layout scheme, and builds an integrated monitoring device and analysis system to achieve real-time monitoring, early warning, and proactive prevention of hydrogen-electric coupling cascading accidents in confined spaces.

[0007] Therefore, the present invention adopts the following technical solution.

[0008] In a first aspect, the present invention provides a method for monitoring cascade accidents in a confined space hydrogen system, comprising: Based on the cable ignition temperature field dataset, confined space thermal hydrogen leakage and diffusion dataset, single combustion and explosion accident consequence dataset, and hydrogen leakage domino effect secondary accident dataset, we trained and constructed cable ignition temperature field prediction model, thermal hydrogen leakage and diffusion prediction model, single combustion and explosion accident consequence prediction model, and secondary accident consequence prediction model based on artificial neural network, respectively. Through multi-model coupling, we established a rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents. Based on the evolution law of the temperature field of cable ignition and the risk threshold of hydrogen system combustion and explosion, the safe range of cable ignition and the safe operation range of hydrogen system are quantified respectively, forming a two-dimensional coupled safety boundary between the electric side and the hydrogen side; combined with the risk area distribution results of the rapid prediction model of the whole process of hydrogen-electric coupling cascade accident, the collaborative optimization arrangement scheme of thermal sensor and hydrogen concentration sensor in confined space is determined. Based on the rapid prediction model of the entire process of hydrogen-electric coupling cascade accidents, the dual-dimensional coupling safety boundary of the electric side and the hydrogen side, and the collaborative optimization layout scheme, a confined space hydrogen system cascade accident analysis system is constructed. This analysis system is used for real-time monitoring, risk prediction, rapid consequence prediction, and early warning of hydrogen-electric coupling cascade accidents.

[0009] Furthermore, the confined space hydrogen system cascade accident monitoring method also includes: Failure mode analysis was used to identify high-risk locations for cable fires and high-risk units for hydrogen leakage. Fault tree analysis was used to complete the spatial matching and association of dangerous units on the electrical side and dangerous units on the hydrogen side. Based on the identification results and the actual layout of the confined space, a three-dimensional full-size model of the confined space hydrogen-electric coupling system was constructed.

[0010] Furthermore, the process of establishing the cable fire temperature field dataset includes: constructing an FDS simulation model for cable fire based on the identified cable fire risk points, conducting numerical simulations of cable combustion for different fire scenarios, obtaining the temperature field distribution and time-series evolution characteristics in the confined space, performing cable fire safety range analysis, and establishing the cable fire temperature field dataset.

[0011] Furthermore, the process of establishing the confined space thermally induced hydrogen leakage and diffusion dataset includes: based on the simulation results of the cable combustion temperature field, performing thermal structural deformation analysis of hydrogen-related equipment and systems, quantifying the equipment deformation and sealing failure probability under temperature load, establishing the mapping relationship between equipment deformation and hydrogen leakage rate and leakage scale, performing deformation-induced hydrogen leakage characteristic analysis, and establishing a confined space thermally induced hydrogen leakage and diffusion dataset.

[0012] Furthermore, the process of establishing the single combustion and explosion accident consequence dataset includes: performing hydrodynamic simulation analysis of hydrogen combustion and explosion accidents based on the thermally induced hydrogen leakage and diffusion dataset, obtaining the distribution and evolution characteristics of overpressure field and temperature field under different single combustion and explosion accidents, performing physical explosion limit analysis, and establishing a single combustion and explosion accident consequence dataset; the hydrogen combustion and explosion accidents include jet flame, flash fire and physical explosion.

[0013] Furthermore, the process of establishing the hydrogen leakage domino effect secondary accident dataset includes: based on the single combustion and explosion accident consequence dataset, combined with the domino effect triggering threshold criterion, judging the risk of secondary accident triggering, constructing a cascade accident scenario of "initial cable fire → hydrogen leakage → primary combustion and explosion → secondary / multi-stage hydrogen equipment combustion and explosion", conducting full-chain fluid dynamics simulation, obtaining the evolution law of temperature field and overpressure field under domino effect, and establishing the hydrogen leakage domino effect secondary accident dataset.

[0014] Furthermore, the process of establishing the rapid prediction model for the entire process of the hydrogen-electric coupling cascade accident includes: Based on the cable fire temperature field dataset, the fire location, fire power and fire time are used as input features, and the time series data of the temperature field of the whole area in the corresponding scenario are used as output labels. The training set, validation set and test set are divided to train the cable fire temperature field prediction model. Based on a dataset of thermally induced hydrogen leakage and diffusion in confined spaces, a thermally induced hydrogen leakage and diffusion prediction model is trained by using temperature field distribution, leakage location and leakage scale as input features and spatiotemporal distribution data of hydrogen concentration in the corresponding scenario as output labels. Based on a single combustion and explosion accident consequence dataset, the ignition location, ignition time, and hydrogen concentration field distribution are used as input features, and the time series data of overpressure field and temperature field under the corresponding scenario are used as output labels to train a single combustion and explosion accident consequence prediction model. Based on the secondary accident dataset of hydrogen leakage domino effect, the overpressure field and temperature field distribution of the primary combustion and explosion accident are used as input features, and the overpressure field and temperature field evolution data of the corresponding secondary / multi-stage accidents are used as output labels to train a secondary accident consequence prediction model. By coupling the above multiple prediction models, a cascaded architecture is constructed, which includes "prediction of cable ignition temperature field → prediction of thermal hydrogen leakage and diffusion → prediction of consequences of single combustion and explosion accidents → risk assessment of domino accidents → prediction of consequences of secondary accidents", forming a rapid prediction model for the entire process of hydrogen-electricity coupled cascaded accidents.

[0015] Furthermore, the process of determining the collaborative optimization layout scheme includes: The safe range for cable fire is quantified with the boundary that the cable does not catch fire and the heat radiation after the fire does not cause the hydrogen-related equipment to fail; the safe operating range of the hydrogen system is quantified with the boundary that the hydrogen concentration does not reach the explosion limit and the combustion and explosion overpressure does not trigger a secondary accident, thus establishing a two-dimensional coupled safety boundary between the electrical side and the hydrogen side. Based on the high-risk area, temperature field and hydrogen concentration field gradient distribution characteristics output by the rapid prediction model of the whole process of hydrogen-electric coupling cascade accident, combined with the electrical-hydrogen dual-dimensional coupled safety boundary, a spatial deployment scheme for thermal sensors and hydrogen concentration sensors is formed. For the spatial deployment scheme of sensors, the effectiveness of response is simulated and verified for different accident scenarios. The number of sensors, installation positions and sampling frequencies are optimized to form the final collaborative optimization layout scheme.

[0016] Furthermore, the confined space hydrogen system cascade accident analysis system includes: The data acquisition unit includes multiple thermal sensors and hydrogen concentration sensors deployed according to the collaborative optimization scheme, as well as electrical circuit voltage and current monitoring units and video monitoring units, which are used to collect temperature data, hydrogen concentration data, electrical operating parameters and on-site images in the confined space in real time. The edge computing unit has a built-in rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents. It is used to receive real-time data, conduct real-time risk prediction and rapid consequence prediction of accidents, and at the same time complete data preprocessing, local storage and abnormal data identification. The early warning output unit communicates with the edge computing unit, sets multiple early warning thresholds, and outputs corresponding audible and visual early warnings, equipment linkage control signals, and emergency response prompts. The cloud-based management and control unit is wirelessly connected to the edge computing unit to enable historical data backtracking, model iteration and updates, remote management and control of multiple sites, and emergency response plan push.

[0017] In a second aspect, the present invention provides a confined space hydrogen system cascade accident monitoring device for implementing the above-mentioned confined space hydrogen system cascade accident monitoring method, comprising: The cascading accident full-process rapid prediction model building unit is used to train and build cable fire temperature field prediction model, thermal hydrogen leakage and diffusion prediction model, single combustion and explosion accident consequence prediction model and secondary accident consequence prediction model based on artificial neural network according to cable fire temperature field dataset, confined space thermal hydrogen leakage and diffusion dataset, single combustion and explosion accident consequence prediction model and secondary accident consequence prediction model, respectively. Through multi-model coupling, a hydrogen-electric coupling cascading accident full-process rapid prediction model is established. The collaborative optimization layout scheme determination unit is used to quantify the safe range of cable ignition and the safe operation range of hydrogen system according to the evolution law of cable ignition temperature field and the explosion risk threshold of hydrogen system, forming a two-dimensional coupled safety boundary between the electric side and the hydrogen side; combined with the risk area distribution results of the rapid prediction model of the whole process of hydrogen-electric coupling cascade accident, the collaborative optimization layout scheme of thermal sensor and hydrogen concentration sensor in confined space is determined. The cascading accident analysis system construction unit is used to construct a confined space hydrogen system cascading accident analysis system based on the rapid prediction model of the entire cascading accident process, the dual-dimensional coupled safety boundary of the electric side and the hydrogen side, and the collaborative optimization layout scheme.

[0018] The present invention has the following beneficial effects: 1. This invention fully covers the entire chain of hydrogen-electric coupling cascade accidents, namely "electrical wire fire → thermally induced structural deformation → hydrogen leakage → combustion and explosion → domino effect". It includes FDS simulation of coupling cable fire, thermal-structural coupling simulation of hydrogen-related equipment (thermal structural deformation analysis of hydrogen-related equipment and systems), and hydrodynamic simulation of hydrogen leakage and combustion and explosion, which fully reveals the evolution law of hydrogen-electric coupling cascade accidents in confined spaces.

[0019] 2. This invention constructs a multi-stage cascaded artificial neural network prediction model, which realizes rapid prediction of the entire process of hydrogen-electric coupling cascade accidents. Compared with the traditional field-by-field simulation method, the prediction efficiency is greatly improved while ensuring prediction accuracy, which can meet the needs of real-time risk assessment at engineering sites.

[0020] 3. This invention quantifies the safe operating range of cable fire and hydrogen system in two dimensions, and proposes a collaborative optimization arrangement scheme for thermal sensors and hydrogen concentration sensors, which solves the problems of blind layout and delayed early warning in existing monitoring systems, and can realize early warning of accidents.

[0021] 4. This invention can be directly applied to various hydrogen-related scenarios in confined spaces, such as hydrogen refueling stations, hydrogen energy workshops, and vehicle-mounted hydrogen systems. Attached Figure Description

[0022] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are one embodiment of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of the confined space hydrogen system cascade accident monitoring method of the present invention; Figure 2 This is a schematic diagram illustrating the hydrogen leakage and diffusion evolution of the hydrogen-related equipment of the present invention. Figure 3 This is a flowchart of the multi-stage artificial neural network model training process of the present invention; Figure 4 This is a structural diagram of the rapid prediction model for cascade accidents in confined space hydrogen systems according to the present invention. Figure 5 This is a schematic diagram of the confined space hydrogen-electric coupling cascade accident monitoring device of the present invention. Detailed Implementation

[0024] The following examples illustrate specific implementations of the present invention. These examples are carried out based on the solution described in the present invention, and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following examples.

[0025] Example 1 This embodiment describes a method for monitoring cascade accidents in a confined space hydrogen system, such as... Figure 1 As shown, the steps are as follows: S1. High-risk locations for cable fires and high-risk units for hydrogen leakage are identified through failure mode analysis. Spatial matching and association of electrical and hydrogen-side hazardous units are completed through fault tree analysis. Based on the identification results and the actual layout of the confined space, a three-dimensional full-size model of the confined space hydrogen-electric coupling system is constructed. S2. Based on the identified cable fire risk points, construct the FDS simulation model for cable fire, conduct numerical simulation of cable combustion for different fire scenarios, obtain the temperature field distribution and time-series evolution characteristics in the confined space, conduct cable fire safety range analysis, and establish a cable fire temperature field dataset. S3. Based on the simulation results of the cable combustion temperature field, conduct thermal structural deformation analysis of hydrogen-related equipment and systems, quantify the equipment deformation and sealing failure probability under temperature load, establish the mapping relationship between equipment deformation and hydrogen leakage rate and leakage scale, conduct deformation-induced hydrogen leakage characteristic analysis, and establish a confined space thermal-induced hydrogen leakage diffusion dataset. S4. Based on the thermally induced hydrogen leakage and diffusion dataset, perform hydrodynamic simulation analysis of hydrogen combustion and explosion accidents. The hydrogen combustion and explosion accidents include jet flame, flash fire and physical explosion. Obtain the distribution and evolution characteristics of overpressure field and temperature field under different single combustion and explosion accidents, perform physical explosion limit analysis, and establish a single combustion and explosion accident consequence dataset. S5. Based on the single combustion and explosion accident consequence dataset and combined with the domino effect triggering threshold criterion, determine the risk of secondary accident triggering, construct a cascade accident scenario of "initial cable fire → hydrogen leakage → primary combustion and explosion → secondary / multi-stage hydrogen equipment combustion and explosion", conduct full-chain fluid dynamics simulation, obtain the evolution law of temperature field and overpressure field under domino effect, and establish a hydrogen leakage domino effect secondary accident dataset. S6. Based on the datasets established in steps S2 to S5, train and construct a cable fire temperature field prediction model, a thermal hydrogen leakage and diffusion prediction model, a single combustion and explosion accident consequence prediction model, and a secondary accident consequence prediction model based on artificial neural networks. Through multi-model coupling, establish a rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents. S7. Based on the evolution law of the temperature field of cable ignition and the explosion risk threshold of hydrogen system, the safe range of cable ignition and the safe operation range of hydrogen system are quantified respectively to form a two-dimensional coupled safety boundary between the electric side and the hydrogen side; combined with the risk area distribution results of the cascade accident rapid prediction model, the collaborative optimization arrangement scheme of thermal sensors and hydrogen concentration sensors in the confined space is determined. S8. Based on the rapid prediction model of the entire cascading accident process, the dual-dimensional coupling safety boundary of the electric side and the hydrogen side, and the collaborative optimization layout scheme, a confined space hydrogen system cascading accident analysis system is constructed. This analysis system is used for real-time monitoring, risk prediction, rapid consequence prediction, and early warning of hydrogen-electric coupling cascading accidents.

[0026] Specifically, the steps for constructing the three-dimensional full-size model of the confined space hydrogen-electric coupling system in step S1 are as follows: A. Using Failure Mode and Effects Analysis (FMEA), we conducted failure mode analyses on electrical cable systems for short circuits, overloads, and poor contact to identify high-risk locations for cable fires; and conducted failure mode analyses on hydrogen-related equipment for valve failures, tank ruptures, and seal failures to identify high-risk units for hydrogen leakage. B. Using fault tree analysis, a top-event fault tree for “cable fire → hydrogen system explosion” is constructed to identify the basic hazard sources and key risk nodes of the hydrogen-electric coupling system, and to complete the spatial matching and association division of electrical-side hazardous units and hydrogen-side hazardous units. C. Based on the identification results and the actual layout of the confined space, a three-dimensional full-size model of the confined space hydrogen-electric coupling system, including electrical cables, hydrogen production devices, hydrogen storage systems, and hydrogen use equipment, is established to provide a geometric basis for subsequent numerical simulation.

[0027] Specifically, the process of establishing the cable fire temperature field dataset in step S2 is as follows: A. Based on the identified cable fire risk points, cable material, specifications, and laying method, a numerical model of cable fire using the Fire Dynamics Simulator (FDS) is established. This model solves the Navier-Stokes equations for low Mach number compressible flow based on the FDS, simulating the thermal convection, thermal radiation, and smoke component transport processes during cable combustion. The core conservation equations are as follows: 1. Mass conservation equation

[0028] 2. Momentum Conservation Equation

[0029] 3. Energy Conservation Equation

[0030] 4. Component Conservation Equation

[0031] In the formula, This indicates the density of the flue gas, expressed in kg / m³. 3 t represents time, in seconds (s). Represents the Hamiltonian operator; This represents a velocity vector, with units of m / s; This indicates the mass generation rate per unit volume of cable combustion, in units of... ; The static pressure is expressed in Pa; g represents the gravitational acceleration vector, with units of m / s². 2 ; This represents the viscous stress tensor, with units of Pa. This represents a volume force vector, with units of N / m. 3 ; This indicates the specific enthalpy of flue gas, expressed in J / kg. , These represent the heat release rate and radiative heat transfer rate per unit volume, respectively, in W / m². 3 ; This represents the thermal conductivity of flue gas, with units of . ; Temperature is expressed in units of . ; This represents the mass fraction of the i-th component; This represents the diffusion coefficient of the i-th component, expressed in m. 2 / s; This represents the unit volume production / consumption rate of the i-th component, in units of... .

[0032] Simulation scenarios with different ignition power, ignition location, and ventilation conditions were set up. Temperature monitoring points were evenly distributed in a grid within the confined space. Numerical simulation of cable combustion was carried out to obtain time-series distribution data of the temperature field in the entire confined space under different scenarios. B. Conduct cable fire combustion experiments, measure temperature change data under different fire scenarios, verify the accuracy of FDS simulation results, and include the simulation results with an accuracy of ≥80% in the cable fire temperature field dataset. C. Based on simulation and experimental data, using the highest withstand temperature of hydrogen-related equipment and the critical temperature of hydrogen combustion and explosion as thresholds, the safe time interval, safe distance interval, and safe power interval for cable ignition are divided, and a quantitative analysis of the safe interval for cable ignition is completed to establish a cable ignition temperature field dataset.

[0033] Specifically, the process of establishing the confined space thermally induced hydrogen leakage and diffusion dataset in step S3 is as follows: A. The simulation results of the cable combustion temperature field are used as thermal loads and imported into the structural mechanics model of hydrogen-related equipment. Thermal structural deformation analysis of hydrogen-related equipment and system is carried out (i.e., thermal-structural coupling simulation, simulation of thermal stress caused by temperature rise, which causes thermal expansion and deformation of components). The deformation, stress distribution and sealing surface failure probability of hydrogen storage tank, valve, and connecting pipeline under different temperature loads are obtained. B. Establish a quantitative mapping relationship between equipment deformation and hydrogen leakage aperture, leakage rate, and leakage scale, and classify leakage scale into small leakage, medium leakage, and large leakage, corresponding to 1%, 10%, and 100% of the leakage flow area, respectively; C. Based on different leakage scale parameters, conduct hydrogen leakage diffusion hydrodynamic simulation, deploy hydrogen concentration monitoring points in the gas phase region at high density, obtain the spatiotemporal distribution characteristics of hydrogen concentration under different scenarios, and establish a confined space thermally induced hydrogen leakage diffusion dataset after experimental verification.

[0034] Specifically, the process of establishing a single combustion and explosion accident consequence dataset in step S4 is as follows: Based on the thermally induced hydrogen leakage and diffusion dataset, and combined with different ignition locations, ignition times, and hydrogen concentration distributions, three types of combustion and explosion accident scenarios—jet flame, flash fire, and physical explosion—are constructed. Fluid dynamics simulations are conducted to obtain the distribution and time-series evolution data of the overpressure field and temperature field across the entire region under different single combustion and explosion accidents. Physical explosion limit analysis is performed to establish a dataset of consequences for single combustion and explosion accidents.

[0035] Specifically, the process of establishing the secondary accident dataset for the hydrogen leakage domino effect in step S5 is as follows: A. Based on a single combustion and explosion accident consequence dataset, extract the maximum overpressure and highest temperature at each hydrogen-side hazardous unit location. Combine the equipment pressure resistance threshold, heat resistance threshold, and domino effect triggering accident standard to determine the risk of secondary accident triggering and construct a multi-level cascading accident scenario of "initial cable fire → hydrogen leakage → primary combustion and explosion → secondary / multi-level hydrogen equipment combustion and explosion". B. Conduct full-chain hydrodynamic simulation of cascading accidents to obtain the propagation and evolution laws of overpressure and temperature fields under the domino effect, complete the characteristic analysis of the domino effect caused by physical explosion, and establish a dataset of secondary accidents caused by hydrogen leakage domino effect.

[0036] Specifically, the process of establishing a rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents in step S6 is as follows: A. Based on the cable fire temperature field dataset, the fire location, fire power, and fire time are used as input features, and the time series data of the temperature field of the whole area in the corresponding scenario are used as output labels. The dataset is divided into training set, validation set, and test set to train and build a cable fire temperature field prediction model. B. Based on a dataset of thermally induced hydrogen leakage and diffusion in confined spaces, a thermally induced hydrogen leakage and diffusion prediction model is trained and constructed by using temperature field distribution, leakage location, and leakage scale as input features and spatiotemporal distribution data of hydrogen concentration in the corresponding scenario as output labels.

[0037] 1. Subsonic flow conditions ( )

[0038] 2. Sonic Flow Condition ( )

[0039] 3. Critical pressure ratio

[0040] In the formula, This indicates the mass flow rate of hydrogen leakage, expressed in kg / s. This represents the flow coefficient of the leakage orifice, typically taken as 0.6 to 0.8; This indicates the effective area of ​​the leakage hole, in meters (m²). 2 ; This represents the adiabatic index of hydrogen, taken as 1.41; This indicates the internal pressure of the hydrogen storage device, expressed in Pa. This indicates the density of hydrogen gas inside the equipment, in kg / m³. 3 ; This indicates environmental pressure, expressed in Pa. This represents the hydrogen gas constant, taken as 4124. ; This indicates the temperature of the hydrogen gas inside the equipment, in units of... ; This indicates the critical pressure ratio; for hydrogen, it is approximately 0.528.

[0041] C. Based on a single combustion and explosion accident consequence dataset, the ignition location, ignition time, and hydrogen concentration field distribution are used as input features, and the overpressure field and temperature field time series data under the corresponding scenario are used as output labels to train and construct a single combustion and explosion accident consequence prediction model. D. Based on the secondary accident dataset of hydrogen leakage domino effect, the overpressure field and temperature field distribution of the primary combustion and explosion accident are used as input features, and the overpressure field and temperature field evolution data of the corresponding secondary / multi-stage accidents are used as output labels to train and construct a secondary accident consequence prediction model. E. Couple the above multiple artificial neural network models to construct a cascaded architecture of "prediction of cable fire temperature field → prediction of thermal hydrogen leakage and diffusion → prediction of single combustion and explosion accident consequences → assessment of domino accident risk → prediction of secondary accident consequences", forming a rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents.

[0042] Specifically, the process for determining the coordinated optimization arrangement scheme of the thermal sensor and the hydrogen concentration sensor within the confined space in step S7 is as follows: A. Quantify the safe range for cable fires, with the boundary being that the cable does not catch fire and the heat radiation after the fire does not cause hydrogen-related equipment to fail; quantify the safe operating range of the hydrogen system, with the boundary being that the hydrogen concentration does not reach the explosion limit and the combustion and explosion overpressure does not trigger a secondary accident, and establish a two-dimensional coupled safety boundary between the electrical side and the hydrogen side. B. Based on the gradient distribution characteristics of high-risk areas, temperature fields and hydrogen concentration fields output by the rapid prediction model of the whole process of hydrogen-electric coupling cascade accidents, combined with the dual-dimensional coupled safety boundary, a spatial deployment scheme for thermal sensors and hydrogen concentration sensors is formed based on the principles of "full area coverage, high-risk area densification, rapid anomaly response and controllable data redundancy". C. For the spatial deployment scheme of sensors, conduct simulation verification of response effectiveness for different accident scenarios, optimize the number of sensors, installation positions and sampling frequency, and form the final collaborative optimization layout scheme.

[0043] Specifically, the confined space hydrogen system cascade accident analysis system mentioned in step S8 includes: The data acquisition unit includes multiple thermal sensors and hydrogen concentration sensors deployed according to the collaborative optimization scheme, as well as electrical circuit voltage and current monitoring units and video monitoring units, which are used to collect temperature data, hydrogen concentration data, electrical operating parameters and on-site images in the confined space in real time. The edge computing unit has a built-in cascaded accident full-process rapid prediction model, which is used to receive the collected real-time data, carry out real-time prediction of accident risks and rapid prediction of consequences, and at the same time complete data preprocessing, local storage and abnormal data identification. The early warning output unit communicates with the edge computing unit, sets multiple early warning thresholds, and outputs corresponding audible and visual early warnings, equipment linkage control signals, and emergency response prompts. The cloud-based management and control unit is wirelessly connected to the edge computing unit to enable historical data backtracking, model iteration and updates, remote management and control of multiple sites, and emergency response plan push.

[0044] Example 2 This embodiment uses the method described in Embodiment 1 for the following application.

[0045] Taking a confined space hydrogen production-storage integrated equipment room as the implementation object, the equipment room has a size of 12m×8m×3.5m. It is equipped with an alkaline electrolysis hydrogen production device, a high-pressure gaseous hydrogen storage cylinder group, a fuel cell system, and is equipped with power cables, control cables, power distribution cabinets and other electrical systems, which is a typical confined space hydrogen-related scenario.

[0046] The confined space hydrogen system cascade accident monitoring method adopts the following steps: 1) Based on the layout of hydrogen-related equipment systems in confined spaces, the electrical cable layout scheme and equipment failure modes, conduct hazard source identification of hydrogen-electric coupling systems, divide the electrical and hydrogen side hazard units, determine the cable fire risk points and high-risk hydrogen leakage points, and construct a three-dimensional full-size model of the confined space hydrogen-electric coupling system.

[0047] Using the FMEA method, failure mode analysis was conducted on the electrical system, identifying three high-risk points for cable fires: the incoming line terminal of the distribution cabinet, the solenoid valve terminal of the hydrogen storage cylinder group, and the corner of the cable tray. Failure mode analysis was also conducted on the hydrogen-related equipment, identifying three high-risk units for hydrogen leakage: the valve at the mouth of the hydrogen storage cylinder, the connection point of the high-pressure pipeline, and the gas-liquid separator of the electrolyzer. This completed the division of hazardous units on the electrical and hydrogen sides.

[0048] Using fault tree analysis, a top-event fault tree for a "hydrogen-electric coupling cascade combustion and explosion accident" was constructed, identifying 12 basic hazard sources and 8 key risk nodes, with cable ignition identified as the core initial hazard source. Based on the actual layout and dimensions of the equipment room, a 1:1 three-dimensional full-size model was established, including electrical cables, solid-state hydrogen production unit, gas-liquid separation unit, purification unit, hydrogen pump, electrolyzer, hydrogen storage cylinder group, and fuel cell system, providing a geometric basis for subsequent simulations.

[0049] 2) Based on the identified cable fire risk points, establish an FDS simulation model for cable fire, conduct numerical simulations of cable combustion under different fire scenarios, obtain the temperature field distribution and time-series evolution characteristics in the confined space, complete the analysis of the safe range for cable fire, and establish a cable fire temperature field dataset.

[0050] Based on the identified three cable fire risk points, a numerical model of cable fire ignition (FDS) was established. The cable material was YJV cross-linked polyethylene cable. Three ignition powers (5kW, 10kW, and 20kW) were set to correspond to small, medium, and large fire scenarios, respectively. A control scenario with different ventilation conditions was also included. Temperature monitoring points were deployed in a uniform grid of 0.5m × 0.5m × 0.5m within the equipment room, covering the entire gas phase region. Multi-scenario cable combustion numerical simulations were conducted to obtain the spatiotemporal distribution time-series data of the temperature field within the equipment room under different scenarios.

[0051] A full-scale cable fire combustion experiment was conducted in the equipment room. Thermocouples were placed at the same locations, and temperature change data under different fire scenarios were measured to verify the accuracy of the FDS simulation results. In this embodiment, the average accuracy of the simulation results met the accuracy requirements, and all valid simulation data were included in the cable fire temperature field dataset. Based on the dataset, the safe time interval and safe distance interval for cable fire under different fire powers were quantified, completing the cable fire safety interval analysis.

[0052] 3) Based on the simulation results of the temperature field of cable combustion, conduct thermal structural deformation analysis of hydrogen-related equipment and systems, quantify the deformation of equipment and the probability of sealing failure under temperature load, establish the mapping relationship between deformation and hydrogen leakage rate and leakage scale, complete the characteristic analysis of hydrogen leakage caused by deformation, and establish a dataset of thermally induced hydrogen leakage diffusion in confined space.

[0053] The simulation results of the cable combustion temperature field were used as a thermal load and imported into the structural mechanics model of the hydrogen storage cylinder group and high-pressure pipeline to perform thermo-structural coupling simulation. This yielded the deformation, stress distribution, and seal failure probability at the valve sealing surface and pipeline connections of the hydrogen storage cylinder under different temperature loads. A quantitative mapping relationship between deformation and hydrogen leakage aperture was established. Leakage was classified into small, medium, and large scales based on 1%, 10%, and 100% of the leakage flow area, corresponding to the hydrogen leakage rate and total leakage volume under different deformation amounts.

[0054] Based on different leakage locations and scale parameters, hydrogen leakage diffusion hydrodynamic simulations were performed. The monitoring point array and temperature field monitoring points maintained spatial topological consistency to obtain the spatiotemporal distribution characteristics of hydrogen concentration under different scenarios. Figure 2 As shown, a hydrogen leakage diffusion experiment was conducted, and the simulation accuracy was verified by actual data from a hydrogen concentration sensor. The average accuracy of the simulation results met the accuracy requirements, and the effective data were included in the confined space thermally induced hydrogen leakage diffusion dataset.

[0055] 4) Based on the thermally induced hydrogen leakage and diffusion dataset, conduct hydrodynamic simulation analysis of hydrogen combustion and explosion accidents, covering three accident types: jet flame, flash fire, and physical explosion. Obtain the distribution and evolution characteristics of overpressure field and temperature field under different single combustion and explosion accidents, complete the physical explosion limit analysis, and establish a dataset of consequences for single combustion and explosion accidents.

[0056] Based on the thermally induced hydrogen leakage and diffusion dataset, and combined with different ignition locations, ignition times, and hydrogen concentration distributions, three types of combustion and explosion accident scenarios—jet flame, flash fire, and physical explosion—were constructed. Fluid dynamics simulations were performed to obtain the time-series evolution data of the overpressure field and temperature field within the equipment room under different single combustion and explosion accidents. Physical explosion limit analysis was completed, and it was determined that the peak overpressure of hydrogen physical explosion in this embodiment scenario can reach 1.2 MPa. A dataset of consequences for a single combustion and explosion accident was established.

[0057] 5) Based on a single combustion and explosion accident consequence dataset and combined with the domino effect triggering threshold standard, the risk of secondary accident triggering is judged. A cascading accident scenario of "initial cable fire → hydrogen leakage → primary combustion and explosion → secondary / multi-stage hydrogen equipment combustion and explosion" is constructed. Full-chain fluid dynamics simulation is carried out to obtain the evolution law of temperature field and overpressure field under domino effect and establish a secondary accident dataset of hydrogen leakage domino effect.

[0058] Based on a single combustion and explosion accident consequence dataset, the maximum overpressure and highest temperature at each hydrogen-side hazardous unit location are extracted. Combined with equipment pressure and heat resistance thresholds, and domino effect triggering accident criteria, it is determined that a combustion and explosion accident caused by a cable fire exceeding 10kW will 100% trigger a secondary explosion in the hydrogen storage cylinder group, forming a domino effect. Based on this, a multi-level cascading accident scenario is constructed, and full-chain fluid dynamics simulation is conducted to obtain the propagation and evolution laws of the overpressure and temperature fields under the domino effect, establishing a dataset of secondary accidents caused by the hydrogen leakage domino effect.

[0059] 6) Based on the above multi-dimensional datasets, train and construct artificial neural network-based models for predicting cable fire temperature fields, thermally induced hydrogen leakage and diffusion, single combustion and explosion accident consequences, and domino effect secondary accident consequences, respectively. Figure 3 As shown; through multi-model coupling, a rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents is established, such as... Figure 4 As shown.

[0060] Based on the above multi-dimensional dataset, four artificial neural network models were trained respectively: Cable fire temperature field prediction model: Taking the fire location, fire power, and fire time as inputs, and the time series data of the temperature field of the whole area as outputs, the dataset is divided into training set, validation set, and test set in an 8:1:1 ratio to complete the model training and calibration. The prediction accuracy of the test set reaches 92.3%. A thermally induced hydrogen leakage diffusion prediction model: taking temperature field distribution, leakage location, and leakage scale as inputs, and spatiotemporal distribution data of hydrogen concentration as outputs, the prediction accuracy on the test set reached 90.1% after training. A single combustion and explosion accident consequence prediction model: taking ignition location, ignition time, and hydrogen concentration field distribution as inputs, and overpressure field and temperature field time series data as outputs, the prediction accuracy on the test set reached 88.7% after training. Secondary accident consequence prediction model: Taking the overpressure field and temperature field distribution of the primary combustion and explosion accident as input, and the overpressure and temperature evolution data of the secondary accident as output, the prediction accuracy of the test set after training is 87.5%.

[0061] By coupling the above four models, a cascaded architecture is constructed, consisting of "prediction of cable ignition temperature field → prediction of thermally induced hydrogen leakage and diffusion → prediction of consequences of a single combustion and explosion accident → assessment of domino accident risks → prediction of consequences of secondary accidents," forming a rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents. Testing showed that this model can complete the prediction of the entire process from initial cable ignition to multi-stage domino accidents within 30 seconds.

[0062] 7) Based on the evolution law of the temperature field of cable ignition and the risk threshold of hydrogen system combustion and explosion, the safe range of cable ignition and the safe operation range of hydrogen system are quantified respectively to form a two-dimensional coupled safety boundary; combined with the risk area distribution results of the accident rapid prediction model, the collaborative optimization arrangement scheme of thermal sensors and hydrogen concentration sensors in the confined space is determined.

[0063] Based on simulation and experimental data, the safe operating range of the cable fire and the safe operating range of the hydrogen system were quantified, establishing a dual-dimensional coupled safety boundary between the electrical and hydrogen sides, and clarifying the safe operating thresholds under different scenarios. Combining the gradient distribution of high-risk areas, temperature fields, and hydrogen concentration fields output by the rapid accident prediction model, and adhering to the principles of "full area coverage, high-risk area densification, rapid anomaly response, and controllable data redundancy," a sensor collaborative optimization deployment scheme was adopted: thermal sensors were densely deployed around three high-risk cable fire locations, hydrogen concentration sensors were densely deployed around three high-risk hydrogen leakage units, and basic monitoring points were evenly deployed at 2-meter intervals within the equipment room. Ultimately, 16 thermal sensors and 12 hydrogen concentration sensors were deployed, along with supporting electrical parameter monitoring units and video monitoring units. Simulation verification showed that this deployment scheme can detect temperature anomalies within 3 seconds of cable fire and concentration anomalies within 5 seconds of hydrogen leakage, achieving early warning at the initial stage of an accident.

[0064] 8) Based on a multi-model coupled cascading accident rapid prediction model, a two-dimensional safety zone, and a sensor collaborative optimization arrangement scheme, a confined space hydrogen system cascading accident analysis system is built. This analysis system is used for real-time monitoring, risk prediction, rapid consequence prediction, and early warning of hydrogen-electric coupling cascading accidents.

[0065] Example 3 This embodiment is a confined space hydrogen system cascade accident monitoring device, such as... Figure 5 As shown, it consists of a system three-dimensional full-size model construction unit, a cable ignition temperature field dataset establishment unit, a thermally induced hydrogen leakage diffusion dataset establishment unit, a single combustion and explosion accident consequence dataset establishment unit, a hydrogen leakage domino effect secondary accident dataset establishment unit, a cascade accident full-process rapid prediction model construction unit, a collaborative optimization layout scheme determination unit, and a cascade accident analysis system construction unit.

[0066] The system's three-dimensional full-size model construction unit is used to identify high-risk locations for cable fires and high-risk units for hydrogen leakage through failure mode analysis. It completes the spatial matching and association of electrical-side and hydrogen-side hazardous units through fault tree analysis. Based on the identification results and the actual layout of the confined space, it constructs a three-dimensional full-size model of the confined space hydrogen-electric coupling system.

[0067] The cable fire temperature field dataset establishment unit is used to construct an FDS simulation model of cable fire based on the identified cable fire risk points, conduct numerical simulation of cable combustion for different fire scenarios, obtain the temperature field distribution and time series evolution characteristics in the confined space, conduct cable fire safety range analysis, and establish a cable fire temperature field dataset.

[0068] The thermal hydrogen leakage and diffusion dataset establishment unit is used to conduct thermal structural deformation analysis of hydrogen-related equipment and systems based on the simulation results of cable combustion temperature field, quantify the deformation of equipment and the probability of seal failure under temperature load, establish the mapping relationship between equipment deformation and hydrogen leakage rate and leakage scale, conduct deformation-induced hydrogen leakage characteristic analysis, and establish a confined space thermal hydrogen leakage and diffusion dataset.

[0069] The single combustion and explosion accident consequence dataset establishment unit is used to perform hydrodynamic simulation analysis of hydrogen combustion and explosion accidents based on the thermally induced hydrogen leakage and diffusion dataset. The hydrogen combustion and explosion accidents include jet flame, flash fire and physical explosion. The unit obtains the distribution and evolution characteristics of overpressure field and temperature field under different single combustion and explosion accidents, performs physical explosion limit analysis, and establishes a single combustion and explosion accident consequence dataset.

[0070] The hydrogen leakage domino effect secondary accident dataset establishment unit is used to determine the risk of secondary accident triggering based on the single combustion and explosion accident consequence dataset and the domino effect triggering threshold criterion. It constructs a cascading accident scenario of "initial cable fire → hydrogen leakage → primary combustion and explosion → secondary / multi-stage hydrogen equipment combustion and explosion", conducts full-chain fluid dynamics simulation, obtains the evolution law of temperature field and overpressure field under domino effect, and establishes a hydrogen leakage domino effect secondary accident dataset.

[0071] The aforementioned rapid prediction model building unit for the entire process of cascading accidents is used to train and build, respectively, a cable fire temperature field prediction model, a thermally induced hydrogen leakage and diffusion prediction model, a single combustion and explosion accident consequence prediction model, and a secondary accident consequence prediction model based on an artificial neural network, based on a cable fire temperature field dataset, a confined space thermally induced hydrogen leakage and diffusion dataset, a single combustion and explosion accident consequence prediction model, and a secondary accident consequence prediction model, based on a hydrogen leakage domino effect secondary accident dataset. Through multi-model coupling, a rapid prediction model for the entire process of hydrogen-electric coupling cascading accidents is established.

[0072] The collaborative optimization layout scheme determination unit is used to quantify the cable fire safety zone and the hydrogen system safe operation zone according to the evolution law of the cable fire temperature field and the hydrogen system combustion and explosion risk threshold, forming a two-dimensional coupled safety boundary; combined with the risk area distribution results of the cascade accident full-process rapid prediction model, it determines the collaborative optimization layout scheme of thermal sensors and hydrogen concentration sensors in the confined space.

[0073] The aforementioned cascading accident analysis system construction unit is used to construct a confined space hydrogen system cascading accident analysis system based on the cascading accident full-process rapid prediction model, the dual-dimensional coupling safety boundary of the electric side and the hydrogen side, and the collaborative optimization layout scheme. This analysis system performs real-time monitoring, risk prediction, rapid consequence prediction, and early warning of hydrogen-electric coupling cascading accidents.

[0074] It should be noted that each unit in the aforementioned confined space hydrogen system cascade accident monitoring device can be implemented entirely or partially through software, hardware, or a combination thereof. These units can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module. For specific limitations regarding the confined space hydrogen system cascade accident monitoring device, please refer to the limitations of the confined space hydrogen system cascade accident monitoring method described above; both have the same function and role, and will not be repeated here.

[0075] The above description of the embodiments is provided to enable those skilled in the art to understand and use the invention. It will be apparent to those skilled in the art that various modifications can be made to these embodiments, and the general principles described herein can be applied to other embodiments without inventive effort. Therefore, the present invention is not limited to the above embodiments, and any improvements and modifications made by those skilled in the art based on the disclosure of the present invention without departing from the scope of the invention should be within the protection scope of the present invention.

Claims

1. A method for monitoring cascade accidents in a confined space hydrogen system, characterized in that, include: Based on the cable ignition temperature field dataset, confined space thermal hydrogen leakage and diffusion dataset, single combustion and explosion accident consequence dataset, and hydrogen leakage domino effect secondary accident dataset, we trained and constructed cable ignition temperature field prediction model, thermal hydrogen leakage and diffusion prediction model, single combustion and explosion accident consequence prediction model, and secondary accident consequence prediction model based on artificial neural network, respectively. Through multi-model coupling, we established a rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents. Based on the evolution law of the temperature field of cable ignition and the risk threshold of hydrogen system combustion and explosion, the safe range of cable ignition and the safe operation range of hydrogen system are quantified respectively, forming a two-dimensional coupled safety boundary between the electric side and the hydrogen side; combined with the risk area distribution results of the rapid prediction model of the whole process of hydrogen-electric coupling cascade accident, the collaborative optimization arrangement scheme of thermal sensor and hydrogen concentration sensor in confined space is determined. Based on the rapid prediction model of the entire process of hydrogen-electric coupling cascade accidents, the dual-dimensional coupling safety boundary of the electric side and the hydrogen side, and the collaborative optimization layout scheme, a confined space hydrogen system cascade accident analysis system is constructed. This analysis system is used for real-time monitoring, risk prediction, rapid consequence prediction, and early warning of hydrogen-electric coupling cascade accidents.

2. The method for monitoring cascading accidents in a confined space hydrogen system according to claim 1, characterized in that, Also includes: Failure mode analysis was used to identify high-risk locations for cable fires and high-risk units for hydrogen leakage. Fault tree analysis was used to complete the spatial matching and association of dangerous units on the electrical side and dangerous units on the hydrogen side. Based on the identification results and the actual layout of the confined space, a three-dimensional full-size model of the confined space hydrogen-electric coupling system was constructed.

3. The method for monitoring cascade accidents in a confined space hydrogen system according to claim 1 or 2, characterized in that, The process of establishing the cable fire temperature field dataset includes: constructing an FDS simulation model of cable fire based on the identified cable fire risk points, conducting numerical simulation of cable combustion for different fire scenarios, obtaining the temperature field distribution and time series evolution characteristics in the confined space, conducting cable fire safety range analysis, and establishing the cable fire temperature field dataset.

4. The method for monitoring cascade accidents in a confined space hydrogen system according to claim 1 or 2, characterized in that, The process of establishing the confined space thermally induced hydrogen leakage and diffusion dataset includes: based on the simulation results of the cable combustion temperature field, performing thermal structural deformation analysis of hydrogen-related equipment and systems, quantifying the equipment deformation and sealing failure probability under temperature load, establishing the mapping relationship between equipment deformation and hydrogen leakage rate and leakage scale, performing deformation-induced hydrogen leakage characteristic analysis, and establishing the confined space thermally induced hydrogen leakage and diffusion dataset.

5. The method for monitoring cascade accidents in a confined space hydrogen system according to claim 1 or 2, characterized in that, The process of establishing the single combustion and explosion accident consequence dataset includes: based on the thermally induced hydrogen leakage and diffusion dataset, performing hydrodynamic simulation analysis of hydrogen combustion and explosion accidents, obtaining the distribution and evolution characteristics of overpressure field and temperature field under different single combustion and explosion accidents, performing physical explosion limit analysis, and establishing a single combustion and explosion accident consequence dataset; the hydrogen combustion and explosion accidents include jet flame, flash fire and physical explosion.

6. The method for monitoring cascade accidents in a confined space hydrogen system according to claim 1 or 2, characterized in that, The process of establishing the hydrogen leakage domino effect secondary accident dataset includes: based on the single combustion and explosion accident consequence dataset, combined with the domino effect triggering threshold criterion, judging the risk of secondary accident triggering, constructing a cascade accident scenario of "initial cable fire → hydrogen leakage → primary combustion and explosion → secondary / multi-stage hydrogen equipment combustion and explosion", conducting full-chain fluid dynamics simulation, obtaining the evolution law of temperature field and overpressure field under domino effect, and establishing the hydrogen leakage domino effect secondary accident dataset.

7. The method for monitoring cascade accidents in a confined space hydrogen system according to claim 1 or 2, characterized in that, The process of establishing the rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents includes: Based on the cable fire temperature field dataset, the fire location, fire power and fire time are used as input features, and the time series data of the temperature field of the whole area in the corresponding scenario are used as output labels. The training set, validation set and test set are divided to train the cable fire temperature field prediction model. Based on a dataset of thermally induced hydrogen leakage and diffusion in confined spaces, a thermally induced hydrogen leakage and diffusion prediction model is trained by using temperature field distribution, leakage location and leakage scale as input features and spatiotemporal distribution data of hydrogen concentration in the corresponding scenario as output labels. Based on a single combustion and explosion accident consequence dataset, the ignition location, ignition time, and hydrogen concentration field distribution are used as input features, and the time series data of overpressure field and temperature field under the corresponding scenario are used as output labels to train a single combustion and explosion accident consequence prediction model. Based on the secondary accident dataset of hydrogen leakage domino effect, the overpressure field and temperature field distribution of the primary combustion and explosion accident are used as input features, and the overpressure field and temperature field evolution data of the corresponding secondary / multi-stage accidents are used as output labels to train a secondary accident consequence prediction model. By coupling the above multiple prediction models, a cascaded architecture is constructed, which includes "prediction of cable ignition temperature field → prediction of thermal hydrogen leakage and diffusion → prediction of consequences of single combustion and explosion accidents → risk assessment of domino accidents → prediction of consequences of secondary accidents", forming a rapid prediction model for the entire process of hydrogen-electricity coupled cascaded accidents.

8. The method for monitoring cascade accidents in a confined space hydrogen system according to claim 1 or 2, characterized in that, The process of determining the collaborative optimization layout scheme includes: The safe range for cable fire is quantified with the boundary that the cable does not catch fire and the heat radiation after the fire does not cause the hydrogen-related equipment to fail; the safe operating range of the hydrogen system is quantified with the boundary that the hydrogen concentration does not reach the explosion limit and the combustion and explosion overpressure does not trigger a secondary accident, thus establishing a two-dimensional coupled safety boundary between the electrical side and the hydrogen side. Based on the high-risk area, temperature field and hydrogen concentration field gradient distribution characteristics output by the rapid prediction model of the whole process of hydrogen-electric coupling cascade accident, combined with the electrical-hydrogen dual-dimensional coupled safety boundary, a spatial deployment scheme for thermal sensors and hydrogen concentration sensors is formed. For the spatial deployment scheme of sensors, the effectiveness of response is simulated and verified for different accident scenarios. The number of sensors, installation positions and sampling frequencies are optimized to form the final collaborative optimization layout scheme.

9. The method for monitoring cascade accidents in a confined space hydrogen system according to claim 1 or 2, characterized in that, The confined space hydrogen system cascade accident analysis system includes: The data acquisition unit includes multiple thermal sensors and hydrogen concentration sensors deployed according to the collaborative optimization scheme, as well as electrical circuit voltage and current monitoring units and video monitoring units, which are used to collect temperature data, hydrogen concentration data, electrical operating parameters and on-site images in the confined space in real time. The edge computing unit has a built-in rapid prediction model for the entire process of hydrogen-electric coupling cascade accidents. It is used to receive real-time data, conduct real-time risk prediction and rapid consequence prediction of accidents, and at the same time complete data preprocessing, local storage and abnormal data identification. The early warning output unit communicates with the edge computing unit, sets multiple early warning thresholds, and outputs corresponding audible and visual early warnings, equipment linkage control signals, and emergency response prompts. The cloud-based management and control unit is wirelessly connected to the edge computing unit to enable historical data backtracking, model iteration and updates, remote management and control of multiple sites, and emergency response plan push.

10. A confined space hydrogen system cascade accident monitoring device, used to implement the confined space hydrogen system cascade accident monitoring method according to any one of claims 1-9, characterized in that, include: The cascading accident full-process rapid prediction model building unit is used to train and build cable fire temperature field prediction model, thermal hydrogen leakage and diffusion prediction model, single combustion and explosion accident consequence prediction model and secondary accident consequence prediction model based on artificial neural network according to cable fire temperature field dataset, confined space thermal hydrogen leakage and diffusion dataset, single combustion and explosion accident consequence prediction model and secondary accident consequence prediction model, respectively. Through multi-model coupling, a hydrogen-electric coupling cascading accident full-process rapid prediction model is established. The collaborative optimization layout scheme determination unit is used to quantify the safe range of cable ignition and the safe operation range of hydrogen system according to the evolution law of cable ignition temperature field and the explosion risk threshold of hydrogen system, forming a two-dimensional coupled safety boundary between the electric side and the hydrogen side; combined with the risk area distribution results of the rapid prediction model of the whole process of hydrogen-electric coupling cascade accident, the collaborative optimization layout scheme of thermal sensor and hydrogen concentration sensor in confined space is determined. The cascading accident analysis system construction unit is used to construct a confined space hydrogen system cascading accident analysis system based on the rapid prediction model of the entire process of hydrogen-electric coupling cascading accidents, the dual-dimensional coupling safety boundary of the electric side and the hydrogen side, and the collaborative optimization layout scheme.

Images