A hydrogen leakage situation prediction model modeling method and device

Abstract

The application discloses a hydrogen leakage situation prediction model modeling method and device, and the scheme is characterized in that: first, the measurement data of historical hydrogen leakage events, preset leakage risk index information and hierarchical system information between different leakage risk indexes are acquired; the first weight and the second weight of each leakage risk index information are calculated by using the entropy weight method and the analytic hierarchy process respectively; then, based on the first weight and the second weight, the relative distance degree between each leakage risk index and the worst solution of the index is determined by using the superior-inferior solution distance method, and a relative risk identification result is obtained; the highest risk event of hydrogen leakage in the relative risk identification result is subjected to fault tree decomposition, a plurality of basic component failure events of the highest risk event are obtained, and model training is performed by taking the measurement data as a training set, so that a hydrogen leakage situation prediction model is obtained, and the technical problem that the existing hydrogen leakage prevention and control cannot perform situation awareness on the space-time evolution of hydrogen leakage events is solved.

Description

Technical Field

[0001] This application relates to the field of hydrogen energy safety and prevention technology, and in particular to a modeling method and apparatus for predicting hydrogen leakage situations. Background Technology

[0002] Mobile hydrogen power generation devices offer advantages such as zero carbon emissions, low noise, and high efficiency, making them a promising direction for future emergency power supply development. However, current safety monitoring and control measures for hydrogen leakage from mobile hydrogen power generation devices in complex scenarios are lacking. The risk of gas cloud combustion and explosion caused by hydrogen storage tank leaks is difficult to assess, and corresponding prevention and control strategies and methods are inadequate, becoming a bottleneck restricting the large-scale application of mobile hydrogen power generation devices in the future.

[0003] Existing hydrogen leak prevention systems primarily rely on installing a certain number of hydrogen concentration sensors and hydrogen leak early warning systems within a defined space to prevent leaks after they occur. However, this approach can only accurately display the real-time hydrogen leak situation based on sensor data after an incident has taken place. It cannot effectively simulate the spatiotemporal evolution of a hydrogen leak event, hindering the development of more targeted hydrogen leak prevention strategies and limiting further improvements in hydrogen leak prevention safety. Summary of the Invention

[0004] This application provides a modeling method and apparatus for predicting hydrogen leakage situations, which addresses the technical problem that existing hydrogen leakage prevention and control measures cannot effectively simulate the spatiotemporal evolution of hydrogen leakage events, thus limiting further improvements in the safety of hydrogen leakage prevention and control.

[0005] To address the aforementioned technical problems, the first aspect of this application provides a modeling method for predicting hydrogen leakage situations, comprising:

[0006] Acquire measurement data of historical hydrogen leak events, preset leak risk index information, and hierarchical information on the information between different leak risk indicators;

[0007] Based on the measurement data, calculate the first weight of each leakage risk indicator;

[0008] Based on the hierarchical system information, an indicator judgment matrix corresponding to each level is constructed, wherein the indicator judgment matrix contains the comparison results of the importance of any indicator in the risk indicator set;

[0009] Based on the aforementioned indicator judgment matrix, calculate the second weight of each leakage risk indicator;

[0010] Based on the risk indicator data matrix and the first weight, the relative distance between each leakage risk indicator and the worst solution of the indicator is determined, and the relative risk identification result is obtained based on the relative distance and the second weight.

[0011] Based on the relative risk identification results and the measurement data, a preset fuzzy neural network is trained to obtain a hydrogen leakage situation prediction model.

[0012] Preferably, the step of calculating the first weight of each leakage risk indicator based on the risk indicator data matrix using the entropy weight method specifically includes:

[0013] The elements of the risk indicator data matrix are normalized.

[0014] Based on the normalized element values, combined with the preset index entropy value calculation formula, the entropy value of each leakage risk index information is calculated.

[0015] Based on the entropy values ​​of the various leakage risk indicators, and combined with the preset entropy weight calculation formula, the first weight corresponding to each leakage risk indicator is obtained.

[0016] Preferably, before calculating the second weight of each leakage risk indicator information using the analytic hierarchy process based on the indicator judgment matrix, the method further includes:

[0017] Based on the maximum eigenvalue and matrix order of the indicator judgment matrix, and combined with the preset consistency ratio calculation formula, the consistency verification result of the indicator judgment matrix is ​​obtained. If the consistency verification result passes, the second weight of each leakage risk indicator information is calculated according to the indicator judgment matrix using the analytic hierarchy process.

[0018] Preferably, the step of determining the relative distance between each leakage risk indicator and the worst solution of the indicator based on the standardized matrix and the first weight, using the good-or-bad solution distance method, specifically includes:

[0019] Based on each element in the standardized matrix, determine the optimal and worst solutions for each leakage risk indicator.

[0020] Based on the optimal and worst solutions, and combined with the first weight, the distance values ​​between each indicator data of the same leakage risk indicator information and the optimal and worst solutions are calculated, and the optimal solution distance and the worst solution distance are obtained respectively.

[0021] The relative distance between the leakage risk index and the worst solution is determined based on the ratio of the worst solution distance to the sum of the optimal solution distance and the worst solution distance.

[0022] Preferably, the step of obtaining the hydrogen leakage situation prediction result by combining the hydrogen leakage situation prediction model with the measured data obtained from on-site measurements specifically includes:

[0023] Obtain on-site measured data, input the measured data into the hydrogen leakage situation prediction model, and obtain the hydrogen leakage situation prediction result through the calculation of the hydrogen leakage situation prediction model.

[0024] Preferably, the step of calculating the first weight of each leakage risk indicator based on the measurement data specifically includes:

[0025] Based on the measurement data and the leakage risk index information, a risk index data matrix is ​​constructed;

[0026] Based on the risk indicator data matrix, the first weight of each leakage risk indicator is calculated using the entropy weight method.

[0027] Preferably, constructing the indicator judgment matrix corresponding to each level based on the hierarchical system information specifically includes:

[0028] Based on the hierarchical system information, risk indicators at the same level are grouped into the same risk indicator set to obtain several risk indicator sets, and then indicator judgment matrices corresponding to each risk indicator set are constructed.

[0029] Preferably, determining the relative distance between each leakage risk indicator and its worst-case solution based on the risk indicator data matrix and the first weight, and obtaining the relative risk identification result based on the relative distance and the second weight specifically includes:

[0030] The risk indicator data matrix is ​​standardized to obtain a standardized matrix. Then, based on the standardized matrix and the first weight, the relative distance between each leakage risk indicator and the worst solution of the indicator is determined by the best-to-worst solution distance method. The relative risk identification result is obtained by multiplying the relative distance with the second weight.

[0031] Preferably, the step of training a preset fuzzy neural network based on the relative risk identification results and the measurement data to obtain a hydrogen leakage situation prediction model specifically includes:

[0032] Based on the highest risk event of hydrogen leakage in the relative risk identification results, fault tree decomposition is performed to obtain several basic component failure events of the highest risk event. Then, the measurement data, the basic component failure events, and the highest risk event of hydrogen leakage are used as training sets to train a preset fuzzy neural network to obtain a hydrogen leakage situation prediction model that includes the correlation between basic component failure events and measurement data, as well as the correlation between measurement data and hydrogen leakage risk events.

[0033] Meanwhile, a second aspect of this application provides a modeling device for predicting hydrogen leakage situations, comprising:

[0034] The historical data acquisition unit is used to acquire measurement data of historical hydrogen leak events, preset leak risk index information, and hierarchical system information between different leak risk indicators.

[0035] The first weight calculation unit is used to calculate the first weight of each leakage risk indicator based on the measurement data.

[0036] The second matrix construction unit is used to construct an indicator judgment matrix corresponding to each level based on the hierarchical system information, wherein the indicator judgment matrix contains the comparison results of the importance of any indicators in the risk indicator set.

[0037] The second weight calculation unit is used to calculate the second weight of each leakage risk indicator information according to the indicator judgment matrix.

[0038] The relative risk identification unit is used to determine the relative distance between each leakage risk indicator and the worst solution of the indicator based on the risk indicator data matrix and the first weight, and to obtain the relative risk identification result based on the relative distance and the second weight.

[0039] The prediction model generation unit is used to train a preset fuzzy neural network based on the relative risk identification results and the measurement data to obtain a hydrogen leakage situation prediction model.

[0040] As can be seen from the above technical solutions, this application has the following advantages:

[0041] The proposed solution first acquires measurement data of historical hydrogen leak events, pre-defined leak risk indicators, and hierarchical information among different leak risk indicators. Then, based on the measurement data and leak risk indicator information, a risk indicator data matrix is ​​constructed. The entropy weight method is used to calculate the first weight of each leak risk indicator. Next, based on the hierarchical information, risk indicators at the same level are grouped into the same risk indicator set, thereby constructing an indicator judgment matrix corresponding to each risk indicator set. The analytic hierarchy process (AHP) is then used to calculate the second weight of each leak risk indicator. Based on the first and second weights, the relative distance between each leak risk indicator and its worst solution is determined using the best-case scenario distance method. The product of the relative distance and the second weight yields the relative risk identification result. Fault tree decomposition is performed on the highest-risk hydrogen leak event from the relative risk identification result to obtain several basic component failure events of the highest-risk event. The measurement data, basic component failure events, and the highest-risk hydrogen leak event are then used as a training set to train a pre-defined fuzzy neural network, resulting in a hydrogen leak situation prediction model that includes the correlation between basic component failure events and measurement data, as well as the correlation between measurement data and hydrogen leak risk events. In practical applications, it is only necessary to use on-site sensors to input the collected measured data into the established hydrogen leakage situation prediction model. By calculating the prior knowledge in the hydrogen leakage situation prediction model, the hydrogen leakage situation prediction result can be obtained. This solves the technical problem that existing hydrogen leakage prevention and control measures cannot effectively simulate the spatiotemporal evolution of hydrogen concentration after a leak occurs, thus limiting the further improvement of the safety of hydrogen leakage prevention and control. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart illustrating an embodiment of a hydrogen leakage situation prediction modeling method provided in this application.

[0044] Figure 2 A logical flowchart illustrating the overall process of a hydrogen leakage situation prediction modeling method embodiment provided in this application.

[0045] Figure 3 This is information on the risk hierarchy among hydrogen leakage risk indicators.

[0046] Figure 4This is a schematic diagram of an embodiment of a hydrogen leakage situation prediction modeling device provided in this application. Detailed Implementation

[0047] This application provides a modeling method and apparatus for predicting hydrogen leakage situations, which addresses the technical problem that existing hydrogen leakage prevention and control measures cannot effectively simulate the spatiotemporal evolution of hydrogen leakage events, thus limiting further improvements in the safety of hydrogen leakage prevention and control.

[0048] To make the inventive objectives, features, and advantages of this application more apparent and understandable, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0049] First, a detailed description of an embodiment of the hydrogen leakage situation prediction modeling method provided in the first aspect of this application is as follows:

[0050] Please see Figures 1 to 3 This embodiment provides a modeling method for predicting hydrogen leakage situations, including:

[0051] Step 101: Obtain measurement data of historical hydrogen leak events, preset leak risk index information, and hierarchical information of information between different leak risk indicators.

[0052] It should be noted that, firstly, the risk events, systems, and risk indicators to be assessed should be identified, such as the hydrogen leakage risk event database for hydrogen fuel cell emergency power vehicles, the risk system and its corresponding scenarios and failure categories. Then, a risk matrix model should be established. The description of the risk matrix varies depending on the risk, and the scores of the two indicators of the probability of risk occurrence and the severity of risk are used.

[0053] Step 102: Based on the measurement data, calculate the first weight of each leakage risk indicator.

[0054] It should be noted that the constructed risk indicator data matrix can be represented by the following equation (1).

[0055] (1)

[0056] Among them, matrix The number of rows m represents the number of risk events, and the number of columns n represents the number of indicators in the risk matrix, which is usually 2, namely the severity of the risk and the probability of occurrence.

[0057] Based on the obtained risk indicator data matrix, the first weight of each leakage risk indicator is calculated using the entropy weight method. .

[0058] More specifically, using equation (2) on the matrix The data is normalized.

[0059] (2)

[0060] In the formula, Let be the value of the j-th evaluation indicator for the i-th evaluated risk event. for The normalized value.

[0061] Then, the entropy value of the j-th evaluation index is calculated using equation (3). ,have

[0062]

[0063] Finally, the weight of the j-th indicator is calculated using equation (4). ,Right now

[0064]

[0065] Step 103: Based on the hierarchical system information, construct the indicator judgment matrix corresponding to each level.

[0066] The indicator judgment matrix contains the results of comparing the importance of any indicators in the risk indicator set.

[0067] It should be noted that after using the entropy weight method to assign weights to the risk matrix indicators, the analytic hierarchy process (AHP) is then used to assign weights to the risk indicators of the risk system.

[0068] (1) First, classify and summarize the indicators of the risk system to form multi-level indicators. Then, compare the indicators of the same level in pairs to form a judgment matrix. As shown in equation (5) below.

[0069]

[0070] In the formula, n risk indicators representing the same level, Indicators relative to indicators The degree of importance. Its value is usually determined using a nine-level scale, as shown in Table 1.

[0071] Table 1. Nine-level scale and its meaning

[0072]

[0073] In some embodiments, a judgment matrix is ​​obtained. Afterwards, the consistency of the judgment matrix can be checked to avoid logical errors in comparison. The consistency check is made by comparing the consistency ratio CR with 0.1. If CR < 0.1, the consistency check passes; otherwise, the judgment matrix needs to be readjusted. The calculation method of CR value is shown in equation (6).

[0074]

[0075] Wherein, CI is the consistency index, and its value can be obtained from equation (7). RI is the average random consistency index, and its values ​​are commonly found in Table 2.

[0076]

[0077] In the formula, To determine the matrix The largest eigenvalue, where n is the number of indices and also the matrix order.

[0078] Table 2. Values ​​of RI

[0079]

[0080] Once the consistency check of the judgment matrix passes, you can proceed to step 105 to continue the subsequent weight calculation.

[0081] Step 104: Calculate the second weight of each leakage risk indicator based on the indicator judgment matrix.

[0082] It should be noted that the geometric mean method is used for calculation, and the calculation formula is shown in equation (8).

[0083]

[0084] in, This represents the indicator weight for this level. After obtaining the weights, to facilitate data processing, the weights are normalized using equation (9) to obtain the normalized weight values. .

[0085]

[0086] Finally, if there are multiple levels of indicators, the total weight is calculated by formula (10) after calculating the weight of each level of indicator. .

[0087]

[0088] in, The weight of the primary indicator relative to the target layer. The weight of the secondary indicator relative to the primary indicator.

[0089] Step 105: Based on the risk indicator data matrix and the first weight, determine the relative distance between each leakage risk indicator and the worst solution of the indicator, and obtain the relative risk identification result based on the relative distance and the second weight.

[0090] It should be noted that the original data first needs to be normalized, i.e., forwarded. The methods for forwarding have been described in detail in relevant literature and will not be repeated here. From the perspective of risk grading, the indicator data of the risk matrix are obviously very large-scale homogeneous data. Therefore, the data is forwarded according to equation (11). After obtaining the normalized data, it needs to be standardized according to equation (12) to construct a standardized matrix. .

[0091]

[0092] In the formula, and These represent the minimum and maximum values ​​of the data under the j-th indicator, respectively.

[0093]

[0094] After obtaining the standardized matrix, the data scores are sorted by using the superior-inferior solution distance method and equations (13)-(15). However, the traditional method only considers the weight of the indicator in which the data itself is located. Corresponding to the risk matrix, it only considers the weight of the two indicators of the risk matrix and does not consider the weight of the risk system indicators. Therefore, this embodiment improves it by adding the risk system weight to improve the calculation method. The improved method is shown in equation (16).

[0095]

[0096]

[0097]

[0098] In the formula, and These represent the distances between the evaluated object and the worst and best solutions, respectively. In the risk matrix, these distances correspond to the minimum and maximum values ​​of the two indicator data. Let j represent the risk matrix indicator. and These represent the minimum and maximum values ​​of the risk data under the j-th indicator, respectively. The value represents the relative distance from the worst solution. The larger the value, the farther the research object is from the worst solution. In the risk matrix, it can be understood as the relative distance from the minimum value of the risk index data. The larger the value, the higher the risk level.

[0099]

[0100] Step 106: Based on the relative risk identification results and measurement data, train the preset fuzzy neural network to obtain a hydrogen leakage situation prediction model.

[0101] It should be noted that, based on the relative risk identification results obtained in the previous step, the highest relative risk event is identified. Using the hydrogen leak-induced combustion and explosion event (i.e., the highest relative risk event identified by the improved risk matrix method) as the top event, fault tree analysis is performed on the hydrogen leak combustion and explosion event. By consulting literature and using a dynamic Bayesian network, the key basic events (component failure, etc.) of the hydrogen leak combustion and explosion event are identified. Mechanism analysis is then performed on the key basic events (key components), a key component failure model is established, the key state parameters leading to key component failure are determined, and a correlation model between the key component failure state parameters and hydrogen leak is established, thus obtaining a hydrogen leak situation prediction model. This hydrogen leak situation prediction model can be embedded in the field terminal. Corresponding sensors are then deployed on-site to measure relevant parameters, and the measured parameters are input into the failure model to determine the component failure state. This completes the construction of the hydrogen leak risk identification module.

[0102] The following is a specific scenario example provided in this application: Taking the event of metal corrosion inside the battery as an example, firstly, the weight shift of its severity and probability is determined by the entropy weight method (EWM) part in the improved risk matrix method. Then, the relative weight of the power generation scenario is determined by the tomographic analysis method (AHP). Finally, combined with the determined index weight values, the risk score is calculated by the superior-inferior solution distance method (TOPSIS). After comparing the scores, it is determined that its risk level is the highest.

[0103] Then, the failure causes of metal corrosion inside the battery were analyzed, the failure characteristics of the two components were extracted, and a fuzzy neural network was trained based on the relationship between component failure and hydrogen leakage.

[0104] The trained model was then deployed on-site to monitor hydrogen leakage in hydrogen fuel cell application scenarios.

[0105] The hydrogen fuel cell system consists of hydrogen storage tanks, valves, hydrogen pipelines and hydrogen fuel cells. An improved risk matrix was used to determine that valves and pipelines are most prone to hydrogen leakage.

[0106] Then, the causes of valve and pipeline failures are analyzed, the failure characteristics of these two components are extracted, and a fuzzy network is trained to study the correlation between component failures and hydrogen leaks.

[0107] The trained model was then deployed on-site to monitor hydrogen leakage in hydrogen fuel cell application scenarios.

[0108] Numerical simulation analysis of the spatiotemporal evolution of hydrogen concentration under various hydrogen leakage scenarios was performed in a computer.

[0109] We used some numerical simulation results to train graph convolutional neural networks and PINN neural networks on hydrogen leakage in hydrogen-related scenarios, and used other numerical simulation results as a test set to improve the prediction speed and accuracy of the model.

[0110] The trained artificial intelligence model was deployed at hydrogen-related sites;

[0111] At hydrogen-related sites, hydrogen concentration sensors, temperature and humidity sensors, and wind speed and direction sensors are used to measure flow field information, enabling visualization of the spatiotemporal evolution of hydrogen concentration after hydrogen leakage is detected.

[0112] Complete the spatiotemporal evolution model of hydrogen concentration monitoring;

[0113] By inputting information from multiple sensors into a graph neural network, the complex relationships between various environmental factors can be fully utilized to improve the accuracy and reliability of predictions. For example, a graph neural network can identify characteristic patterns in hydrogen concentration changes under specific wind speed and temperature conditions, providing early warnings of potential leakage risks. Simultaneously, the graph neural network model can dynamically adjust its predictions of hydrogen diffusion behavior under different environmental conditions, providing a scientific basis for emergency response.

[0114] Hydrogen leak sensors are placed at the hydrogen storage tank opening, passenger compartment, and fuel cell engine system—areas prone to hydrogen accumulation and leakage—to monitor the hydrogen content inside the vehicle in real time. In the event of a hydrogen leak, immediate response measures are taken to ensure passenger safety. Furthermore, when any sensor detects a hydrogen volume fraction exceeding the lower explosive limit (4% hydrogen volume in air) by 10%, 25%, and 50%, the monitor will issue Level I, Level II, and Level III audible and visual alarm signals, respectively. Specific control measures are shown in Table 3.

[0115] Table 3 Safety Status Levels for Hydrogen Leakage

[0116]

[0117] Based on the national emergency response plan for public emergencies and considering factors such as hydrogen leaks and the external environment, the security level of network security is divided into 5 levels, and the weight and characteristics of each level are quantitatively described using values ​​in the range [0,1].

[0118] Leveraging the strong learning ability and fuzzy information processing capabilities of fuzzy neural networks, this method is employed to effectively identify hydrogen leaks at different concentrations. An FNN model is established, comprising an input layer, a fuzzy layer, a fuzzy inference layer, a compensation operation layer, and an output layer. The number of nodes in the input layer is determined by the hydrogen leak concentration data detected by the hydrogen sensor under different environments, with each node corresponding to a specific concentration. The number of nodes in the fuzzy layer is determined by the number of input variables and their fuzzy subsets. The number of nodes in both the fuzzy inference layer and the compensation operation layer is determined by the number of rules, m. The output layer has only one node.

[0119] Based on the hydrogen leakage concentration data detected by hydrogen sensors under different environments, a safety situation assessment and intelligent prediction are performed. Using numerical simulation models under fault conditions, a system state database is simulated under different fault modes, fault locations, and fault severity. Based on hydrogen leakage characteristics and hydrogen equipment safety prediction mechanisms, a hydrogen safety situation awareness system is constructed.

[0120] The above is a detailed description of an embodiment of a hydrogen leakage situation prediction modeling method provided in this application. The following is a detailed description of an embodiment of a hydrogen leakage situation prediction modeling device provided in this application.

[0121] Please see Figure 4 This embodiment provides a hydrogen leakage situation prediction modeling device, comprising:

[0122] The historical data acquisition unit 201 is used to acquire measurement data of historical hydrogen leak events, preset leak risk index information, and hierarchical system information between different leak risk indicators.

[0123] The first weight calculation unit 202 is used to calculate the first weight of each leakage risk indicator based on the measurement data.

[0124] The second matrix construction unit 203 is used to construct an indicator judgment matrix corresponding to each level based on the hierarchical system information. The indicator judgment matrix contains the comparison results of the importance of any indicator in the risk indicator set.

[0125] The second weight calculation unit 204 is used to calculate the second weight of each leakage risk indicator information according to the indicator judgment matrix.

[0126] The relative risk identification unit 205 is used to determine the relative distance between each leakage risk indicator and the worst solution of the indicator based on the risk indicator data matrix and the first weight, and to obtain the relative risk identification result based on the relative distance and the second weight.

[0127] The prediction model generation unit 206 is used to train a preset fuzzy neural network based on the relative risk identification results and measurement data to obtain a hydrogen leakage situation prediction model.

[0128] Preferably, the first weight calculation unit 202 is specifically used for:

[0129] The elements of the risk indicator data matrix are normalized.

[0130] Based on the normalized element values, combined with the preset index entropy value calculation formula, the entropy value of each leakage risk index information is calculated.

[0131] Based on the entropy values ​​of various leakage risk indicators, and combined with the preset entropy weight calculation formula, the first weight corresponding to each leakage risk indicator is obtained.

[0132] Preferably, it further includes:

[0133] The consistency verification unit 2040 is used to obtain the consistency verification result of the indicator judgment matrix based on the maximum eigenvalue and matrix order of the indicator judgment matrix, combined with the preset consistency ratio calculation formula. If the consistency verification result passes, the second weight of each leakage risk indicator information is calculated according to the indicator judgment matrix through the analytic hierarchy process.

[0134] Preferably, the relative risk identification unit 205 is specifically used for:

[0135] The relative risk identification unit is used to perform data standardization processing on the risk indicator data matrix to obtain a standardized matrix.

[0136] Based on the elements in the standardized matrix, determine the optimal and worst solutions for each leakage risk indicator;

[0137] Based on the optimal and worst solutions, and combined with the first weight, the distance values ​​between the data of each indicator of the same leakage risk indicator and the optimal and worst solutions are calculated, and the distances of the optimal and worst solutions are obtained respectively.

[0138] The relative distance between the leakage risk index and the worst solution is determined by the ratio of the distance to the worst solution to the distance to the optimal solution and the sum of the distances to the worst solution.

[0139] Preferably, the hydrogen leakage situation prediction results obtained by combining the hydrogen leakage situation prediction model with the measured data obtained from on-site measurements specifically include:

[0140] Obtain on-site measured data, input the measured data into the hydrogen leakage situation prediction model, and obtain the hydrogen leakage situation prediction result through the calculation of the hydrogen leakage situation prediction model.

[0141] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the terminals, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0142] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0143] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0144] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0145] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0146] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0147] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0148] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A modeling method for predicting hydrogen leakage situations, characterized in that, include: Acquire measurement data of historical hydrogen leak events, preset leak risk index information, and hierarchical information on the information between different leak risk indicators; Based on the measurement data and the leakage risk index information, a risk index data matrix is ​​constructed; Based on the aforementioned risk indicator data matrix, the first weight of each leakage risk indicator is calculated using the entropy weight method. Based on the hierarchical system information, an indicator judgment matrix corresponding to each level is constructed, wherein the indicator judgment matrix contains the comparison results of the importance of any indicator in the risk indicator set; Based on the aforementioned indicator judgment matrix, the second weight of each leakage risk indicator is calculated using the analytic hierarchy process (AHP). Based on the risk indicator data matrix and the first weight, the relative distance between each leakage risk indicator and the worst solution of the indicator is determined, and the relative risk identification result is obtained by multiplying the relative distance and the second weight. Based on the highest risk event of hydrogen leakage in the relative risk identification results, fault tree decomposition is performed to obtain several basic component failure events of the highest risk event. Then, the measurement data, the basic component failure events, and the highest risk event of hydrogen leakage are used as training sets to train a preset fuzzy neural network to obtain a hydrogen leakage situation prediction model that includes the correlation between basic component failure events and measurement data, as well as the correlation between measurement data and hydrogen leakage risk events.

2. The modeling method for predicting hydrogen leakage situation according to claim 1, characterized in that, The step of calculating the first weight of each leakage risk indicator based on the risk indicator data matrix using the entropy weight method specifically includes: The elements of the risk indicator data matrix are normalized. Based on the normalized element values, combined with the preset index entropy value calculation formula, the entropy value of each leakage risk index information is calculated. Based on the entropy values ​​of the various leakage risk indicators, and combined with the preset entropy weight calculation formula, the first weight corresponding to each leakage risk indicator is obtained.

3. The modeling method for predicting hydrogen leakage situation according to claim 1, characterized in that, Before calculating the second weights of each leakage risk indicator based on the indicator judgment matrix using the analytic hierarchy process, the process also includes: Based on the maximum eigenvalue and matrix order of the indicator judgment matrix, and combined with the preset consistency ratio calculation formula, the consistency verification result of the indicator judgment matrix is ​​obtained. If the consistency verification result is passed, the process jumps to the step of calculating the second weight of each leakage risk indicator information according to the indicator judgment matrix using the analytic hierarchy process.

4. The modeling method for predicting hydrogen leakage situation according to claim 1, characterized in that, Based on the risk indicator data matrix and the first weight, the relative distance between each leakage risk indicator and the worst-case solution of the indicator is determined, and the relative risk identification result is obtained by multiplying the relative distance and the second weight. Specifically, this includes: The risk indicator data matrix is ​​standardized to obtain a standardized matrix. Then, based on the standardized matrix and the first weight, the relative distance between each leakage risk indicator and the worst solution of the indicator is determined by the best-to-worst solution distance method. The relative risk identification result is obtained by multiplying the relative distance with the second weight.

5. The modeling method for predicting hydrogen leakage situation according to claim 1, characterized in that, After obtaining the hydrogen leakage situation prediction model, the following steps are also included: Obtain on-site measured data, input the measured data into the hydrogen leakage situation prediction model, and obtain the hydrogen leakage situation prediction result through the calculation of the hydrogen leakage situation prediction model.

6. The modeling method for predicting hydrogen leakage situation according to claim 1, characterized in that, The specific steps of constructing the indicator judgment matrix corresponding to each level based on the hierarchical system information include: Based on the hierarchical system information, risk indicators at the same level are grouped into the same risk indicator set to obtain several risk indicator sets, and then indicator judgment matrices corresponding to each risk indicator set are constructed.

7. The modeling method for predicting hydrogen leakage situation according to claim 4, characterized in that, The step of determining the relative distance between each leakage risk indicator and the worst solution of the indicator based on the standardized matrix and the first weight, using the best-to-worst solution distance method, specifically includes: Based on each element in the standardized matrix, determine the optimal and worst solutions for each leakage risk indicator. Based on the optimal and worst solutions, and combined with the first weight, the distance values ​​between each indicator data of the same leakage risk indicator information and the optimal and worst solutions are calculated, and the optimal solution distance and the worst solution distance are obtained respectively. The relative distance between the leakage risk index and the worst solution is determined based on the ratio of the worst solution distance to the sum of the optimal solution distance and the worst solution distance.

8. A modeling device for predicting hydrogen leakage situations, characterized in that, include: The historical data acquisition unit is used to acquire measurement data of historical hydrogen leak events, preset leak risk index information, and hierarchical system information between different leak risk indicators. The first weight calculation unit is used to construct a risk indicator data matrix based on the measurement data and the leakage risk indicator information, and to calculate the first weight of each leakage risk indicator information based on the risk indicator data matrix using the entropy weight method. The second matrix construction unit is used to construct an indicator judgment matrix corresponding to each level based on the hierarchical system information, wherein the indicator judgment matrix contains the comparison results of the importance of any indicators in the risk indicator set. The second weight calculation unit is used to calculate the second weight of each leakage risk indicator information according to the indicator judgment matrix and the analytic hierarchy process. The relative risk identification unit is used to determine the relative distance between each leakage risk indicator and the worst solution of the indicator based on the risk indicator data matrix and the first weight, and to obtain the relative risk identification result based on the product of the relative distance and the second weight. The prediction model generation unit is used to perform fault tree decomposition on the highest risk event of hydrogen leakage in the relative risk identification results to obtain several basic component failure events of the highest risk event. Then, the measurement data, the basic component failure events and the highest risk event of hydrogen leakage are used as training sets to train a preset fuzzy neural network to obtain a hydrogen leakage situation prediction model that includes the correlation between basic component failure events and measurement data and the correlation between measurement data and hydrogen leakage risk events.

Images