Chain risk propagation quantification method for water resource system, electronic equipment and storage medium

By building a risk multi-link network of water resources systems and analyzing node characteristics, the problem that traditional technology is difficult to analyze dynamic risk transmission of complex water resources systems is solved, and systematic and diversified analysis of water resources system risks is realized, providing science risk assessment and efficient governance strategies.

CN118966742APending Publication Date: 2024-11-15HOHAI UNIV
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202410878121.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-02
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

Traditional risk analysis techniques are difficult to effectively analyze the dynamically changing risk transmission process in complex water resource systems, resulting in challenges in risk management of water resource systems.

Method used

By dividing the water resource system into subsystems, identifying the causal relationship of related elements, building a risk multi-link network, calculating node characteristics and network indicators, and dividing into the stages of start-up, propagation and elimination, analyzing the risk propagation process, and selecting it from the network perspective Prevention and control strategies.

Benefits of technology

It realizes a systematic and diversified analysis of the risk transmission process of water resources system, provides scientific risk assessment methods and efficient governance strategies, highlighting the periodicity and complexity of risk transmission.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118966742A_ABST
    Figure CN118966742A_ABST
Patent Text Reader

Abstract

The invention provides a chain risk propagation quantification method for a water resource system, electronic equipment and a storage medium. Firstly, a causal relationship of associated elements is identified; then, the causal relationship among the partition associated elements is collected to form a subunit network, the subunit networks are integrated through the hydraulic relationship among the regions, a watershed water resource system risk multi-link network is formed, the degree and betweenness centrality of each node are calculated, and the degree and betweenness centrality of each node are calculated; in the starting stage, the importance level, the safety bearing threshold value, the risk sensitivity and the initial load of the nodes are calculated according to the node features calculated by the risk multi-link network; in the propagation stage, the probability of risk propagation is calculated by utilizing node network characteristics, and propagation load is calculated; in the fading stage, a fading coefficient is defined to reduce node loads; and the change condition of the risk in the propagation process is analyzed from the node, the link and the network. And finally, providing a risk prevention and control strategy through node features in the network. According to the invention, a water resource system risk analysis technology under a complex condition can be perfected.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of risk analysis and relates to a water resource system chain risk propagation quantification method, electronic equipment and storage medium. Background Art

[0002] As a typical complex system, the water resources system in the river basin is affected by meteorological, hydrological, social-economic, ecological and other complex systems. There are complex relationships between different systems, such as information barriers, time-delay transmission, and oscillation differentiation. Under the combined influence of global climate change and human activities, the water cycle process in the river basin has changed significantly, and the risk of water crisis has become increasingly prominent. Therefore, how to analyze the characteristics of complex water resources systems under changing conditions is a major scientific problem that needs to be solved urgently. Most traditional risk analysis techniques are applied to static qualitative or quantitative risk assessment methods, but risks often change dynamically, and are usually propagated and distributed in a certain form among system elements, resulting in water problems in the river basin that are characterized by a single hair moving the whole body, which brings severe challenges to the development and utilization of water resources systems and risk management. Therefore, in the analysis of the characteristics of complex water resources systems, more attention should be paid to the risk propagation process of related elements. Summary of the invention

[0003] In response to the problems existing in the prior art, the present invention provides a method, device and system for quantifying chain risk propagation in a water resource system. The causal relationship of related elements in the water resource system is identified through a causal diagnosis method, and a multi-link network of water resource system risks is constructed in combination with hydraulic connections. The risk propagation mechanism in the network is described based on the characteristics of the related elements in the network. The purpose is to reveal the dynamic risk propagation process of multidimensional elements in the water resource system, provide a scientific basis for the risk analysis method of the water resource system, better improve the risk assessment technology of the water resource system, and adopt more efficient governance methods for the comprehensive management and risk control of the river basin.

[0004] In order to solve the above technical problems, the present invention provides the following technical solutions: a water resource system chain risk propagation quantification method, comprising the following steps:

[0005] S1. Divide the water resource system into several subsystems and determine the set of associated elements of each subsystem; based on the collected data of associated elements of the target area, identify the causal relationship of the associated elements of each subsystem and form an associated element link;

[0006] S2. For each sub-region of the target area, the associated element links are collected to form a sub-unit network. Then, the sub-unit networks are integrated into a multi-link network of watershed water resource system risks through hydraulic connections between regions, and redundant links of the multi-link network are deleted.

[0007] S3. Treat the multi-link network of water resource system risk as a complex network, the associated elements as nodes, and the causal relationship between the associated elements as edges, and deduce the complex network indicators; calculate the importance level, safety carrying threshold, risk status, and risk sensitivity of the nodes based on the node characteristics derived from the risk multi-link network;

[0008] S4. Carry out the risk communication process: start-up phase, communication phase, and decline phase;

[0009] S5. Construct analysis indicators from the three perspectives of nodes, links, and networks, analyze the changes in risks during the propagation process, and then select prevention and control strategies from the perspective of reducing the degree of risk spread and reducing the number of risk nodes.

[0010] Furthermore, the aforementioned step S1 specifically includes: dividing the water resource system into four major subsystems, namely meteorology, hydrology, socio-economics, and ecology, and using a causal diagnosis method to identify the causal relationship of related elements.

[0011] Furthermore, in the aforementioned step S2, the redundant links of the multi-link network are deleted as follows:

[0012] S2.1. Use the transfer entropy method to determine the information transmission direction of the associated element X and the associated element Y to delete redundant links.

[0013]

[0014] Where H(x) is the information entropy of the associated element X, and H(x,y) is the joint information entropy of the associated elements X and Y.

[0015] H(X t-τ >Y t )=H(X t-τ ,Y t-τ )+H(Y t ,Y t-τ )-H(Y t-τ )-H(X t-τ ,Y t-τ ,Y t ) (25)

[0016] In the formula, H(X>Y) is the transfer entropy of correlation factor X to correlation factor Y, X t-τ , Y t-τ is the time series of the associated elements X and Y at time t-τ;

[0017] S2.2. Comparison of H(X t-τ >Y t ) and H(Y t-τ >X t )’s information transfer entropy value, if H(X t-τ >Yt )>H(Y t-τ >X t ), then it is considered that there is a causal relationship X→Y; if H(Y t-τ >X t )>H(X t-τ >Y t ), then there is a causal relationship Y→X; if the two are the same, then it is believed that there is no causal relationship between X and Y.

[0018] Furthermore, in the aforementioned step S3, the causal relationship between the associated elements is regarded as an edge, and the complex network indicators including the degree centrality and betweenness centrality of each node are calculated; specifically:

[0019] The degree centrality is calculated as follows:

[0020]

[0021] In the formula, k i is the degree of node i; and are the in-degree and out-degree of node i respectively; N is the total number of nodes; DC i is the degree centrality of node i;

[0022] The betweenness centrality is calculated as follows:

[0023]

[0024] In the formula, BC i is the betweenness centrality of the node; σ jk is the number of shortest paths from node j to node k; σ jk (i) refers to the number of shortest paths from node j to node k that pass through node i.

[0025] Furthermore, in the aforementioned step S3, the importance level, safety carrying threshold, and risk sensitivity of the node are calculated, specifically:

[0026] The importance level is used to reflect the importance of a node in the network. The calculation formula is:

[0027]

[0028] In the formula, λ i is the importance level of node i; n is the number of network nodes; N is the set of network nodes; k i is the degree of node i;

[0029] The safety carrying threshold is the critical value of the element to resist risk and ensure the safety of the water resources system. The calculation formula is:

[0030]

[0031] In the formula, C i is the safety load threshold of node i; is the out-degree of node i; is the parent node set of node i; For collection The number of elements in ;

[0032] Risk state. In the mth node risk propagation process, the state of node i at time t is recorded as If the load exceeds the safe carrying threshold, it is a risk state. It is recorded as a risk node; otherwise, it is a safe state Recorded as a safe node;

[0033]

[0034] In the formula, is the load of node i in the mth node risk propagation process at time t; T is the single node risk propagation time;

[0035] The risk sensitivity calculation formula is as follows:

[0036]

[0037] In the formula, φ i is the risk sensitivity of node i; BC i is the betweenness centrality of the node.

[0038] Furthermore, the aforementioned step S4 includes the following sub-steps:

[0039] S4.1. In the startup phase, the initial load refers to the load that the node has at the beginning of risk propagation, which is calculated through indicators such as the safety carrying threshold and importance level. The initial load is calculated as follows:

[0040]

[0041] Where, t f is the moment when a node propagation phase ends;

[0042] S4.2, in the propagation stage, the probability of risk propagation is calculated using the node network characteristics, and then the propagation load is calculated by combining the causal strength between the elements.

[0043] The formula for calculating the probability of risk propagation from node i to node j is:

[0044]

[0045] Where P ij∈[0,1] represents the risk propagation probability of node i to node j; θ i is the set of child nodes of node i;

[0046] Taking into account the influence of all parent nodes of node j, the risk propagation probability formula of node j is calculated as follows:

[0047]

[0048] In the formula, is the probability that node j becomes a risk node at time t+1 in the risk propagation process of the mth node; f i mT (t) is the probability that the parent node i of node j becomes a risk node at time t in the risk propagation process of the mth node. The transfer entropy relationship is used to quantitatively evaluate the causal strength between the parent and child nodes. The calculation formula is:

[0049]

[0050] In the formula, u ij is the causal strength of node i to node j; h ij is the transfer entropy from node i to node j;

[0051] At this time, if the load of parent node i Does not exceed its safe carrying threshold C i , then no load change and propagation will occur, Otherwise, if More than C i When , the load of parent node i that exceeds the safe carrying threshold will be propagated to child node j, and the load of the safe carrying threshold is recorded as the overload load, which is calculated as follows:

[0052]

[0053] In the formula, is the overload load propagated from parent node i to child node j at time t in the risk propagation process of the mth node; is the total load of the parent node i at time t in the risk propagation process of the kth node;

[0054] After receiving the overload loads of all parent nodes, the total load of child node j is calculated as follows:

[0055]

[0056] After load transfer occurs, after node i propagates the corresponding overload load to all child nodes, its total load becomes:

[0057]

[0058] In the formula, It represents the load of node i at time t+1 during the risk propagation process of the mth node; υ is the load retention coefficient, which reflects the node's extreme load. In particular, if υ=0, it means that the node load is reduced to the safe load threshold after the propagation is completed; if υ=1, it means that the parent node load remains unchanged before and after the risk propagation;

[0059] In the S4.3 fading stage, the total load of node i is calculated by introducing the fading coefficient as follows:

[0060]

[0061] In the formula, is the ideal load of node i; e is the decay coefficient.

[0062] Further, in the aforementioned step S5, the analysis indicators include the node risk sensitivity index, the link carrying risk amount, the network risk diffusion index, and the network structure vulnerability index; specifically as follows:

[0063] The calculation formula of node risk sensitivity index is:

[0064]

[0065] Where SI i is the risk sensitivity index of node i.

[0066] The calculation formula of link carrying risk is:

[0067] LCR ij =ΔL ij ·P ij (42)

[0068] Where LCR ij is the link risk of link i→j;

[0069] The calculation formula of network risk diffusion index is:

[0070]

[0071] Where RDI is the network risk diffusion index;

[0072] The calculation formula of network structure vulnerability index is:

[0073]

[0074] Where h is the total number of nodes in the largest semi-connected subgraph in the network. If there is a causal relationship between node i and node j, then node i and node j are semi-connected; if there is a semi-connected relationship between any two nodes, then the network is a semi-connected graph, n is the total number of nodes; d is the number of nodes at risk.

[0075] Furthermore, in the aforementioned step S5, four prevention and control strategies, namely, random prevention and control, degree prevention and control, betweenness prevention and control, and SI prevention and control, are selected from the perspectives of reducing the degree of risk diffusion and reducing the number of risk nodes.

[0076] Another aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any one of the methods described in the present invention when executing the computer program.

[0077] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of any one of the methods described in the present invention when executed by a processor.

[0078] Compared with the prior art, the beneficial technical effects of the present invention using the above technical solution are as follows:

[0079] 1. The present invention coordinates the multi-dimensional elements in the water resources system and constructs a multi-link network of water resources system risks, highlighting the systematic and diverse nature of risk analysis.

[0080] 2. This invention proposes a risk propagation mechanism from a network perspective. Based on the risk definition, it combines the relationship between load, probability and risk, and divides the risk propagation process into the start-up phase, propagation phase and decay phase, highlighting the periodicity and complexity of risk propagation.

[0081] 3. Based on the network characteristic indicators of correlation factors, the present invention combines the risk analysis methods in the fields of economics and computer science to propose risk prevention and control strategies, highlighting the interdisciplinary and innovative nature of risk transmission research. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] Figure 1 The figure is a flow chart of the method of the present invention.

[0083] Figure 2 This is a multi-link network diagram of the risk of the water resources system in the Yellow River Basin.

[0084] Figure 3 This is a graph showing the changes in node risk sensitivity index under different increased loads.

[0085] Figure 4 The following is a graph showing the change in link risk under different load increases.

[0086] Figure 5 This is a graph showing the changes in the network risk diffusion index.

[0087] Figure 6 This is a graph showing the change process of the network structure vulnerability index.

[0088] Figure 7 This is a graph showing the changes in the network risk diffusion index under different prevention and control strategies.

[0089] Figure 8 This is a graph showing the changes in the network structure vulnerability index under different prevention and control strategies. DETAILED DESCRIPTION

[0090] In order to better understand the technical content of the present invention, specific embodiments are described below in conjunction with the accompanying drawings.

[0091] Various aspects of the invention are described herein with reference to the accompanying drawings, in which many illustrative embodiments are shown. The embodiments of the invention are not limited to those described in the accompanying drawings. It should be understood that the invention is implemented by any of the various concepts and embodiments described above, as well as the concepts and embodiments described in detail below, because the concepts and embodiments disclosed in the invention are not limited to any implementation. In addition, some aspects disclosed in the invention may be used alone or in any appropriate combination with other aspects disclosed in the invention.

[0092] refer to Figure 1 The present invention provides a water resource system chain risk transmission quantification method, comprising the following steps:

[0093] S1. Divide the water resource system into several subsystems and determine the set of associated elements of each subsystem; based on the collected data of associated elements of the target area, identify the causal relationship of the associated elements of each subsystem and form an associated element link;

[0094] S2. For each sub-region of the target area, the links of associated elements are collected to form a sub-unit network. Then, the sub-unit networks are integrated into a multi-link network of water resources system risk in the basin through the hydraulic connection between regions, and the redundant links of the multi-link network are deleted. S3. The multi-link network of water resources system risk is regarded as a complex network, the associated elements are regarded as nodes, and the causal relationship between the associated elements is regarded as edges, and the complex network indicators are calculated. The importance level, safety carrying threshold, risk status, and risk sensitivity of the nodes are calculated according to the node characteristics calculated by the risk multi-link network.

[0095] S4. Carry out the risk communication process: start-up phase, communication phase, and decline phase;

[0096] S5. Construct analysis indicators from the three perspectives of nodes, links, and networks, analyze the changes in risks during the propagation process, and then select prevention and control strategies from the perspective of reducing the degree of risk spread and reducing the number of risk nodes.

[0097] The meteorological, hydrological, socio-economic, and ecological historical data of seven secondary water resource areas along the mainstream of the Yellow River Basin were selected to determine the set of associated elements as shown in Table 1. The method proposed in the present invention was used to identify the causal relationship of the associated elements and establish a multi-link network of sub-regional water resource system risks. The multi-link network of the Yellow River Basin water resource system risks was constructed by linking the seven secondary water resource areas using hydraulic connections, as shown in Table 1. Figure 2 , infer network node indicators. First, the importance level, safety carrying threshold, and risk sensitivity of the node are calculated according to the inferred node characteristics; in the risk propagation startup phase, the initial node load is calculated, and the parent node where the risk propagation will occur is identified. When the first node risk propagation begins (t=T), the parent node of the propagated risk is the risk source node, and it is determined whether the parent node load exceeds its safety carrying threshold. If it does not exceed the safety carrying threshold, it enters the retreat phase; if it exceeds the safety carrying threshold, the probability of the child node becoming a risk node is calculated according to the node characteristics, and the load increment allocated to each child node is calculated in combination with the transfer entropy. At the same time, the parent node retains part of the load after the transfer. Then, after the retreat phase, the load of all nodes is updated, and the next node risk propagation is carried out (t→t+1). The child node that received the risk in the previous phase is used as the parent node of the next phase of risk propagation. According to the above steps, risk propagation is carried out until the risk is propagated to all nodes, which is recorded as a complete network risk propagation.

[0098] Table 1 Correlation factor system of water resources zoning subsystems in the Yellow River Basin

[0099]

[0100] Step S1 specifically includes: dividing the water resources system into four major subsystems, namely meteorology, hydrology, socio-economics, and ecology, and using the PCMCI (Condition Selection and Momentary Conditional Independence) causal diagnosis method to identify the causal relationship of related elements.

[0101] In step S2, for each sub-area of ​​the target area, the associated element links are collected to form a sub-unit network, and then the sub-unit networks are integrated through the hydraulic connections between regions (such as rainfall → runoff, upstream section inflow → surface water resources, etc.) to form a multi-link network of watershed water resources system risks. Redundant links such as repeated edges and circular edges in the network, such as node X pointing to itself (i.e., X→X) or nodes X and Y pointing to each other (i.e., X→Y, Y→X), will cause the network to generate circular feedback, making it difficult to parse the link propagation relationship. The transfer entropy method is used to determine the information transmission direction of X and Y to delete redundant links.

[0102]

[0103] Where H(x) is the information entropy of the associated element X, and H(x,y) is the joint information entropy of the associated elements X and Y.

[0104] H(X t-τ >Y t )=H(X t-τ ,Y t-τ )+H(Y t ,Y t-τ )-H(Y t-τ )-H(X t-τ ,Y t-τ ,Y t ) (47)

[0105] In the formula, H(X>Y) is the transfer entropy of correlation factor X to correlation factor Y, X t-τ , Y t-τ is the time series of the associated elements X and Y at time t-τ.

[0106] Then compare H(X t-τ >Y t ) and H(Y t-τ >X t )’s information transfer entropy value, if H(X t-τ >Y t )>H(Y t-τ >X t ), then it is considered that there is a causal relationship X→Y; if H(Y t-τ >X t )>H(X t-τ >Y t ), then there is a causal relationship Y→X; if the two are the same, then it is believed that there is no causal relationship between X and Y.

[0107] In step S3, the multi-link network of water resource system risk is regarded as a complex network, the associated elements are regarded as nodes, the causal relationship (link) between the associated elements is regarded as an edge, and the complex network indicators such as the degree and betweenness centrality of each node are calculated.

[0108] (I) Degree refers to the total number of nodes connected to the target node, which is used to describe the degree of association between variables. Degree centrality is a standardized indicator of node degree. Degree centrality is calculated as follows:

[0109]

[0110] In the formula, k i is the degree of node i; and are the in-degree and out-degree of node i respectively; N is the total number of nodes; DC i is the degree centrality of node i;

[0111] (ii) Betweenness centrality is a standardized indicator that describes the target node as an intermediary connecting other nodes. The larger its value, the greater the role of the node as a "bridge" in the network, and the more likely it is to bear more risk transmission processes.

[0112] The betweenness centrality is calculated as follows:

[0113]

[0114] In the formula, BC i is the betweenness centrality of the node; σ jk is the number of shortest paths from node j to node k; σ jk (i) refers to the number of shortest paths passing through node i among all links from node j to node k. There may be multiple paths between different nodes, and the shortest path refers to the shortest one or several paths connecting two nodes. In the shortest path, the number of edges is called the distance between two nodes.

[0115] In step S3, the importance level, safety carrying threshold, and risk sensitivity of the node are calculated, specifically:

[0116] In the water resources system, some related elements (such as rainfall and cross-sectional flow) have close causal relationships with other nodes and are more prone to risks, so it is necessary to strengthen the protection of such elements. Therefore, the importance level can effectively reflect the importance of the node in the network. The importance level is used to reflect the importance of the node in the network. The calculation formula is:

[0117]

[0118] In the formula, λ i is the importance level of node i; n is the number of network nodes; N is the set of network nodes; k i is the degree of node i;

[0119] When the node index exceeds a certain threshold, the internal structure or external characteristics will change, thus causing risk propagation. Therefore, the safety carrying threshold is the critical value for the element to resist risk and ensure the safety of the water resources system. The calculation formula is:

[0120]

[0121] In the formula, C i is the safety load threshold of node i; is the out-degree of node i; is the parent node set of node i; For collection The number of elements in ;

[0122] Risk state. In the mth node risk propagation process, the state of node i at time t is recorded as If the load exceeds the safe carrying threshold, it is a risk state. It is recorded as a risk node; otherwise, it is a safe state Recorded as a safe node;

[0123]

[0124] In the formula, is the load of node i in the mth node risk propagation process at time t; T is the single node risk propagation time;

[0125] In a complex network, some nodes play the role of key bridges and hubs, which have an important impact on the risk propagation process of the water resources system. In other words, when the node load changes, it is more likely to affect subsequent nodes. The risk sensitivity calculation formula is as follows:

[0126]

[0127] In the formula, φ i is the risk sensitivity of node i; BC i is the betweenness centrality of the node.

[0128] Step S4 includes the following sub-steps:

[0129] S4.1. In the startup phase, the initial load refers to the load that the node has at the beginning of risk propagation, which can be calculated through indicators such as the safety carrying threshold and importance level. The initial load is calculated as follows:

[0130]

[0131] Where, t f It is the time when a node propagation phase ends.

[0132] The parent node of the complex network that performs the first propagation in the startup phase is the risk source node of the network (which may be one or more), and the parent node of the subsequent risk propagation is the child node that received the load during the last propagation process.

[0133] S4.2, in the propagation stage, the probability of risk propagation is calculated using the node network characteristics, and then the propagation load is calculated based on the causal strength between the elements.

[0134] The probability of node j turning into a risk node is usually determined by its own characteristics and associated nodes: when the node itself has a greater risk sensitivity or is closely connected with other nodes, both will lead to more frequent risk propagation.

[0135] The formula for calculating the probability of risk propagation from node i to node j is:

[0136]

[0137] Where P ij ∈[0,1] represents the risk propagation probability of node i to node j; θ i is the set of child nodes of node i.

[0138] For any node j, all parent nodes connected to it may cause risk to node j. Therefore, the risk propagation probability needs to take into account the influence of all parent nodes of node j. The formula for calculating the risk propagation probability of node j is as follows:

[0139]

[0140] In the formula, is the probability that node j becomes a risk node at time t+1 in the risk propagation process of the mth node; f i mT (t) is the probability that the parent node i of node j becomes a risk node at time t in the risk propagation process of the mth node.

[0141] In complex network theory, when the parent node load exceeds its safe carrying threshold, it will propagate its own overload to all connected child nodes. The total amount of load received by each child node from the parent node is closely related to the causal strength between the two. The present invention uses the transfer entropy relationship to quantitatively evaluate the causal strength between the parent and child nodes, that is, the load distribution ratio. The calculation formula is:

[0142]

[0143] In the formula, u ij is the causal strength of node i to node j; h ij is the transfer entropy from node i to node j.

[0144] At this time, if the load of parent node i Does not exceed its safe carrying threshold C i , then no load change and propagation will occur, Otherwise, if More than C i When , the load of parent node i that exceeds the safe carrying threshold will be propagated to child node j, and the load of the safe carrying threshold is recorded as the overload load, which is calculated as follows:

[0145]

[0146] In the formula, is the overload load propagated from parent node i to child node j at time t in the risk propagation process of the mth node; is the total load of the parent node i at time t in the risk propagation process of the kth node.

[0147] After receiving the overload loads of all parent nodes, the total load of child node j is calculated as follows:

[0148]

[0149] After the load transfer occurs, the load of the parent node will be lost, but it usually will not disappear completely immediately. Therefore, this paper believes that after the risk node propagation is completed, a certain proportion of overload load will still remain; after node i propagates the corresponding overload load to all child nodes, its total load becomes:

[0150]

[0151] In the formula, It represents the load of node i at time t+1 during the risk propagation process of the mth node; υ is the load retention coefficient, which reflects the node's extreme load. In particular, if υ=0, it means that the node load is reduced to the safe load threshold after the propagation is completed; if υ=1, it means that the parent node load remains unchanged before and after the risk propagation;

[0152] In the S4.3 fading stage, the total load of node i is calculated by introducing the fading coefficient as follows:

[0153]

[0154] In the formula, is the ideal load of node i; e is the decay coefficient. In particular, if the decay coefficient e = 0, the node has no decay stage, and the total load of adjacent stages remains unchanged.

[0155] In step S5, the analysis indicators include node risk sensitivity index, link risk load, network risk diffusion index, and network structure vulnerability index; specifically, they are as follows:

[0156] The node risk sensitivity index is used to evaluate the impact of node load changes on other nodes. The larger the value, the greater the impact. It indicates that fine-tuning the node load will also cause risk propagation and have a significant impact on other nodes. The node risk sensitivity index calculation formula is:

[0157]

[0158] Where SI i is the risk sensitivity index of node i.

[0159] The link risk load is used to evaluate the risk transmission capability of an edge in a complex network. The larger the value, the more load the link transmits. The calculation formula for link risk load is:

[0160] LCR ij =ΔL ij ·P ij(60)

[0161] Where LCR ij is the link risk of link i→j;

[0162] The network risk diffusion index is used to evaluate the risk diffusion capacity of a node, which is usually proportional to the total number of propagation paths of the node, that is, the greater the betweenness centrality, the stronger the propagation capacity. The network risk diffusion index calculation formula is:

[0163]

[0164] Where RDI is the network risk diffusion index;

[0165] The network structure vulnerability index reflects the impact of risk transmission on the overall structure of the network. The larger the value, the more risk nodes there are, the more fragile the network is, and the easier it is to spread risks. The network structure vulnerability index calculation formula is:

[0166]

[0167] Where h is the total number of nodes in the largest semi-connected subgraph in the network. If there is a causal relationship between node i and node j, then node i and node j are semi-connected; if any two nodes are semi-connected, then the network is a semi-connected graph, n is the total number of nodes; d is the number of nodes with risks. If any two nodes are semi-connected, then the network is a semi-connected graph. Obviously, the largest semi-connected subgraph is a subnetwork where any two nodes are semi-connected.

[0168] In step S5, four prevention and control strategies, namely, random prevention and control, degree prevention and control, betweenness prevention and control, and SI prevention and control, are selected from the perspectives of reducing the degree of risk diffusion and reducing the number of risk nodes, and the changes in the risk propagation index under different prevention and control strategies are analyzed. In the four prevention and control strategies, 10% (20 nodes) of all nodes are selected for prevention and control measures. The prevention and control strategy is to control the selected nodes after each propagation, and reduce their load to r1 times of the original load. In the random prevention and control strategy, the operation is repeated 100 times, and 20 nodes are randomly selected from all nodes each time to ensure the reliability of the calculation results and the randomness of the network nodes, and the average value of the risk propagation index is taken as the final result; in the degree value prevention and control strategy, all nodes are arranged in descending order according to the degree centrality index, and the first 20 nodes with larger degree centrality are selected as the prevention and control nodes; in the betweenness prevention and control strategy, all nodes are arranged in descending order according to the betweenness centrality, and the first 20 nodes with larger betweenness centrality are taken as the prevention and control nodes; in the SI prevention and control strategy, all nodes are arranged in descending order according to the node risk sensitivity index, and the first 20 nodes with larger risk sensitivity index are taken as the prevention and control nodes to calculate the risk propagation index; in addition, a blank control test is added, that is, the risk propagation index without a prevention and control strategy is used as a comparison.

[0169] The present invention starts from the risk source node and regards the process of load traversing all nodes as a complete network risk propagation process. After analysis, it is found that: "The multi-link network of the risk of the water resources system in the Yellow River Basin starts from all risk sources in turn, and after 6 node risk propagations, the risk can be propagated to all nodes, indicating that a complete network risk propagation process involves 6 node risk propagation processes. "In order to compare the effects of load, load retention coefficient and decay coefficient, the present invention sets the initial load increase ratio to: -50%, 0, 50%, 100%, 200%, 300%, 400%, 500%; the load retention coefficient is set to: 0, 0.1, 0.2, 1; the decay coefficient is set to: 0, 0.02, 0.05. The present invention sets the initial probability of the risk source to f i (0) = 0.5, ideal node load L * =0, and then after a complete network risk propagation process, the network risk propagation index is analyzed. The present invention first calculates the network characteristics of the complex water resource system, such as degree and betweenness centrality, and then calculates the safety carrying threshold, node sensitivity, importance level, initial load and other indicators of each node; on this basis, the initial load of the node is increased by -50% to 500%, and different decay coefficients e and load retention coefficients υ are set. The corresponding network risk propagation process is simulated under different parameter combination scenarios, so as to calculate the risk propagation related indicators.

[0170] 1) In order to fully reflect the impact of node load changes, this section sets e = 0, υ = 1, constructs risk scenarios under different node loads, and calculates the risk sensitivity index of each node. The results are shown in Figure 3 .

[0171] 2) For the analysis of the link risk, refer to the steps in Section 1) to construct risk scenarios under different risk source loads, set e = 0.02, υ = 0.1, and calculate the link risk under different propagation times. The results are as follows: Figure 4 .

[0172] 3) For the analysis of the network risk diffusion index, refer to the steps in section 1), construct risk scenarios under different risk source loads, set different parameters e and υ, and calculate the network risk diffusion index under different propagation times. The results are as follows: Figure 5 .

[0173] 4) For the analysis of network structure vulnerability index, refer to the steps in section 1), construct risk scenarios under different risk source loads, set different parameters e and υ, and calculate the network structure vulnerability index under different propagation times. The results are as follows: Figure 6 .

[0174] Determine the parameter r1 = 0.98, select the scenario where the risk source load is increased by 300% to simulate the prevention and control strategy, and obtain 5 groups of network risk diffusion index and network structure vulnerability index. The simulation methods and control test results of the four prevention and control strategies are shown in Figure 2. Figure 7 and Figure 8 .

[0175] Another aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any one of the methods described in the present invention when executing the computer program.

[0176] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of any one of the methods described in the present invention when executed by a processor.

[0177] Although the present invention has been described above with preferred embodiments, it is not intended to limit the present invention. A person skilled in the art of the present invention may make various modifications and improvements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the definition of the claims.

Claims

1. A quantitative method for chain risk propagation in water resource systems, characterized in that: The steps include: S1. Divide the water resource system into several subsystems and determine the set of associated elements of each subsystem; based on the collected data of associated elements of the target area, identify the causal relationship of the associated elements of each subsystem and form an associated element link; S2. For each sub-region of the target area, the associated element links are collected to form a sub-unit network. Then, the sub-unit networks are integrated into a multi-link network of watershed water resource system risks through hydraulic connections between regions, and redundant links of the multi-link network are deleted. S3. Consider the multi-link network of water resource system risk as a complex network, the associated elements as nodes, the causal relationship between the associated elements as edges, and deduce the complex network indicators; According to the node characteristics inferred from the risk multi-link network, the importance level, safety carrying threshold, risk status and risk sensitivity of the node are calculated respectively; S4. Carry out the risk communication process: start-up phase, communication phase, and decline phase; S5. Construct analysis indicators from the three perspectives of nodes, links, and networks, analyze the changes in risks during the propagation process, and then select prevention and control strategies from the perspective of reducing the degree of risk spread and reducing the number of risk nodes.

2. A water resource system chain risk propagation quantification method according to claim 1, characterized in that: Step S1 specifically includes: dividing the water resource system into four major subsystems, namely meteorology, hydrology, socio-economics, and ecology, and using a causal diagnosis method to identify the causal relationship of related elements.

3. A water resource system chain risk propagation quantification method according to claim 1, characterized in that: In step S2, the redundant links of the multi-link network are deleted as follows: S2.

1. Use the transfer entropy method to determine the information transmission direction of the associated element X and the associated element Y to delete redundant links. Where H(x) is the information entropy of the associated element X, and H(x,y) is the joint information entropy of the associated elements X and Y. H(X t-τ >Y t )=H(X t-τ ,Y t-τ )+H(Y t ,Y t-τ )-H(Y t-τ )-H(X t-τ ,Y t-τ ,Y t ) (3) In the formula, H(X>Y) is the transfer entropy of correlation factor X to correlation factor Y, X t-τ , Y t-τ is the time series of the associated elements X and Y at time t-τ; S2.

2. Comparison of H(X t-τ >Y t ) and H(Y t-τ >X t )’s information transfer entropy value, if H(X t-τ >Y t )>H(Y t-τ >X t ), then it is considered that there is a causal relationship X→Y; if H(Y t-τ >X t )>H(X t-τ >Y t ), then there is a causal relationship Y→X; if the two are the same, then it is believed that there is no causal relationship between X and Y.

4. A water resource system chain risk propagation quantification method according to claim 1, characterized in that: In step S3, the causal relationship between the associated elements is regarded as an edge, and the complex network indicators including the degree centrality and betweenness centrality of each node are calculated; specifically: The degree centrality is calculated as follows: In the formula, k i is the degree of node i; and are the in-degree and out-degree of node i respectively; N is the total number of nodes; DC i is the degree centrality of node i; The betweenness centrality is calculated as follows: In the formula, BC i is the betweenness centrality of the node; σ jk is the number of shortest paths from node j to node k; σ jk (i) refers to the number of shortest paths from node j to node k that pass through node i.

5. A water resource system chain risk propagation quantification method according to claim 1, characterized in that: In step S3, the importance level, safety carrying threshold, and risk sensitivity of the node are calculated, specifically: The importance level is used to reflect the importance of a node in the network. The calculation formula is: In the formula, λ i is the importance level of node i; n is the number of network nodes; N is the set of network nodes; k i is the degree of node i; The safety carrying threshold is the critical value of the element to resist risk and ensure the safety of the water resources system. The calculation formula is: In the formula, C i is the safety load threshold of node i; is the out-degree of node i; is the parent node set of node i; For collection The number of elements in ; Risk state. In the mth node risk propagation process, the state of node i at time t is recorded as If the load exceeds the safe carrying threshold, it is a risk state. It is recorded as a risk node; otherwise, it is a safe state Recorded as a safe node; In the formula, is the load of node i in the mth node risk propagation process at time t; T is the single node risk propagation time; The risk sensitivity calculation formula is as follows: In the formula, φ i is the risk sensitivity of node i; BC i is the betweenness centrality of the node.

6. A water resource system chain risk propagation quantification method according to claim 5, characterized in that: Step S4 includes the following sub-steps: S4.

1. In the startup phase, the initial load refers to the load that the node has at the beginning of risk propagation, which is calculated through indicators such as the safety carrying threshold and importance level. The initial load is calculated as follows: Where, t f is the moment when a node propagation phase ends; S4.2, in the propagation stage, the probability of risk propagation is calculated using the node network characteristics, and then the propagation load is calculated by combining the causal strength between the elements. The formula for calculating the probability of risk propagation from node i to node j is: Where P ij ∈[0,1] represents the risk propagation probability of node i to node j; θ i is the set of child nodes of node i; Taking into account the influence of all parent nodes of node j, the risk propagation probability formula of node j is calculated as follows: In the formula, is the probability that node j becomes a risk node at time t+1 in the risk propagation process of the mth node; f i mT (t) is the probability that the parent node i of node j becomes a risk node at time t in the risk propagation process of the mth node; The transfer entropy relationship is used to quantitatively evaluate the causal strength between parent and child nodes. The calculation formula is: In the formula, u ij is the causal strength of node i to node j; h ij is the transfer entropy from node i to node j; At this time, if the load of parent node i Does not exceed its safe carrying threshold C i , then no load change and propagation will occur, Otherwise, if More than C i When , the load of parent node i that exceeds the safe carrying threshold will be propagated to child node j, and the load of the safe carrying threshold is recorded as the overload load, which is calculated as follows: In the formula, is the overload load propagated from parent node i to child node j at time t in the risk propagation process of the mth node; is the total load of the parent node i at time t in the risk propagation process of the kth node; After receiving the overload loads of all parent nodes, the total load of child node j is calculated as follows: After load transfer occurs, after node i propagates the corresponding overload load to all child nodes, its total load becomes: In the formula, represents the load of node i in the risk propagation process of the mth node at time t+1; υ is the load retention coefficient, which reflects the node's extreme load. In particular, if υ=0, it means that the node load is reduced to the safe load threshold after the propagation is completed; if υ=1, it means that the parent node load remains unchanged before and after the risk propagation; In the S4.3 fading stage, the total load of node i is calculated by introducing the fading coefficient as follows: In the formula, is the ideal load of node i; e is the decay coefficient.

7. A water resource system chain risk propagation quantification method according to claim 6, characterized in that: In step S5, the analysis indicators include node risk sensitivity index, link risk load, network risk diffusion index, and network structure vulnerability index; specifically, they are as follows: The calculation formula of node risk sensitivity index is: Where SI i is the risk sensitivity index of node i. The calculation formula of link carrying risk is: LCR ij =ΔL ij ·P ij (20) Where LCR ij is the link risk of link i→j; The calculation formula of network risk diffusion index is: Where RDI is the network risk diffusion index; The calculation formula of network structure vulnerability index is: Where h is the total number of nodes in the largest semi-connected subgraph in the network. If there is a causal relationship between node i and node j, then node i and node j are semi-connected; if there is a semi-connected relationship between any two nodes, then the network is a semi-connected graph, n is the total number of nodes; d is the number of nodes at risk.

8. A water resource system chain risk propagation quantification method according to claim 7, characterized in that: In step S5, four prevention and control strategies, namely, random prevention and control, degree prevention and control, betweenness prevention and control, and SI prevention and control, are selected from the perspectives of reducing the degree of risk diffusion and reducing the number of risk nodes.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Citation Information

Cited By

  • Critical path and intermediate node identification method for implied water shortage risk transmission based on intermediary centrality algorithm

    CN119807658A

  • Production line design change propagation regulation and control method based on hypergraph theory

    CN121882507A