Harbor district accident risk visualization analysis method based on network topology structure

Abstract

The invention discloses a harbor district accident risk visualization analysis method based on a network topology structure, and the method comprises the steps: taking a key facility and an important region as network nodes in a harbor district, determining a connection relation to obtain a network topology structure, analyzing historical cases to obtain a determined risk factor set, obtaining an uncertain factor set through edge calculation, and carrying out the calculation of the uncertain factor set. The method comprises the following steps: constructing an accident tree to carry out quantitative evaluation on network nodes and connection risks, constructing a risk evolution model, extracting an accident cause chain and calculating a risk influence range, visualizing a network topology structure, carrying out color mapping on the accident cause chain and the risk influence range, and then adding the accident cause chain and the risk influence range to obtain a three-dimensional topology view. According to the method, the network topology structure visualization analysis of the accident risk of the harbor district can be realized, the risk is comprehensively evaluated, the subsequent management decision of the harbor district is optimized, and meanwhile, the method has relatively good interpretability.

Description

Technical Field

[0001] The invention relates to the field of port engineering management, and in particular to a port accident risk visualization analysis method based on a network topology structure. Background Art

[0002] With the continuous development and expansion of port business, the safety management of port areas has become crucial. In modern port areas, various equipment, facilities and personnel activities are intertwined, forming a complex network topology.

[0003] Traditional methods of analyzing accident risks in port areas often rely on empirical judgment and simple data statistics, lacking in-depth understanding and visual presentation of the overall network structure of the port area. This makes it difficult to fully and accurately grasp the distribution and propagation path of risks in accident risk assessment and management. At the same time, with the rapid development of information technology, network topology analysis technology has been widely used in various fields. By analyzing the network structure, the relationship and mutual influence between the nodes in the system can be revealed, providing strong support for risk assessment and decision-making.

[0004] A visualization analysis method for port accident risks based on network topology structure uses network topology analysis and combines the actual situation of the port to conduct a comprehensive analysis of accident risks. It also intuitively displays the distribution and propagation path of risks through visualization, providing a scientific basis and decision-making support for port safety management. Summary of the invention

[0005] The purpose of the present invention is to provide a visual analysis method for port accident risks based on network topology.

[0006] To achieve the above object, the present invention is implemented according to the following technical solutions:

[0007] The first aspect of the present invention provides a method for visualizing and analyzing port accident risks based on a network topology structure, comprising:

[0008] S100 sets the key facilities and important areas in the port area as network nodes, determines the connection relationship between network nodes based on actual physical connections and business processes, and obtains the network topology of the port area;

[0009] S200 analyzes historical accident cases of the network topology to obtain a set of definite risk factors; performs edge computing on the network topology to obtain a set of uncertain risk factors;

[0010] S300 constructs an accident tree based on the determined risk factor set and the uncertain risk factor set, and performs quantitative assessment on the risk of the network topology structure based on the accident tree;

[0011] S400 constructs a risk evolution model, extracts the accident cause chain from the quantitative assessment results through the risk evolution model, and calculates the risk impact range according to the accident cause chain;

[0012] S500 visualizes the network topology structure, performs color mapping on the accident cause chain and the risk impact range, and adds them to the visualization result to obtain a three-dimensional topology view.

[0013] As a further method, the method of analyzing historical accident cases of the network topology structure to obtain a method for determining a set of risk factors includes:

[0014] Obtain accident records of historical accident cases, collect network topology data related to the accident, including node information, connection relationships, and edge flow data, and identify key factors that led to the accident;

[0015] Calculate the correlation coefficient between key factors and accident loss degree, the expression is:

[0016]

[0017] Where m is the number of historical accident cases, F ik Denoted as the key factor F in the kth case i The value of The key factor F i The mean value, L k is the accident loss degree of the kth case, is the mean of the accident loss extent of all cases;

[0018] The key factors whose correlation coefficients with the accident loss degree are greater than 0.01 are taken as the determined risk factors to obtain the determined risk factor set.

[0019] As a further method, the method of performing edge computing on the network topology structure to obtain a set of uncertain risk factors includes:

[0020] Perform edge computing on each network node to obtain the state variables of each network node;

[0021] The state variables between network nodes are combined arbitrarily, and the correlation coefficient matrix is ​​obtained by calculating the Pearson correlation coefficient;

[0022] The factors corresponding to the state variables greater than a preset threshold in the correlation coefficient matrix are obtained, and the factors repeated in the determined risk factor set are removed to obtain the uncertain risk factor set.

[0023] As a further method, the method for constructing an accident tree based on the determined risk factor set and the uncertain risk factor set includes: taking the risk existing in the port area as the top event, taking each factor in the determined risk factor set and the uncertain risk factor set as a basic event, and according to the logical relationship, using logical gates such as AND gates and OR gates to connect the basic events to obtain intermediate events and form an accident tree.

[0024] As a further method, the method for quantitatively evaluating the risk of the network topology structure based on the fault tree includes:

[0025] S21 uses the Bayesian network to calculate the risk probability of the corresponding basic events in the risk factor set, specifically:

[0026] The basic events in the accident tree are used as nodes of the Bayesian network to construct the Bayesian network;

[0027] The kernel function is established using the hyperparameter method to modify the Bayesian network. The expression of the kernel function is:

[0028]

[0029] Among them, α is the global scaling factor, x and y are the state variable and feature variable of the event in the Bayesian network, σ is the bandwidth of the Gaussian radial basis function, T represents the transposition operation, M is the normalization factor, d is the degree of the polynomial kernel, β and b are the slope and bias of the Sigmoid kernel, and p is the nonlinear exponent of the kernel function;

[0030] The probability distribution in the Bayesian network is updated by maximizing expectation. The updated expression is:

[0031]

[0032] Among them, θ is the parameter set in the Bayesian network, θ t and θ (t+1) are the estimates of θ in the tth iteration and the t+1th iteration respectively, arg maxθ is the maximization operator, is the expectation operator, P(X hidden |D,θ (t) ) is given the observed data D and parameter estimates θ (t) Under the condition that the hidden node X hidden The posterior probability distribution of is the log-likelihood function of the joint probability distribution, n is the number of nodes in the Bayesian network, P(X i |pa(X i );θ) represents the estimated node X that depends on the parameter θ i Given its parent node pa(X i)’s conditional probability;

[0033] The probability distribution in the Bayesian network is used as the risk probability of the basic event;

[0034] S22 uses fuzzy logic to calculate the risk probability of the corresponding basic events in the uncertain factor set, specifically:

[0035] The expression of the membership function is:

[0036]

[0037] Among them, λ i is the i-th uncertain risk factor, c i For x i The central value of is i The standard deviation of

[0038] The membership function is defuzzified by the centroid method to obtain the risk probability of the basic event;

[0039] S23 calculates the risk probability of network nodes and top events through the accident tree according to the risk probability of basic events. The expression is:

[0040]

[0041] Among them, E is a basic event or a complex event composed of basic events, P(ξ i ) is the basic event ξ i The probability of risk;

[0042] S24 takes the risk probability of each network node and top event as a quantitative assessment result.

[0043] As a further method, the method of constructing a risk evolution model and extracting the accident cause chain from the quantitative assessment results through the risk evolution model includes:

[0044] The risk evolution model is constructed using the Markov chain algorithm, where the expression of the state transition function is:

[0045]

[0046] Among them, P(X n+1 =j|X n =i) is the probability of transferring to the target state j at time step n+1 under the condition that the state is in state i at time step n, S is the set of all states, U is the set of determined risk factors, and p u (n) is the probability distribution of the determined risk factor u at time step n, V is the set of uncertain risk factors, q v(n) is the probability distribution of the uncertain risk factor v at time step n, p kj is the transition probability from the intermediate state k to the target state j;

[0047] Obtain historical accident cause chain data from historical accident cases, divide them into training set, test set and validation set in a ratio of 8:1:1 to train the risk evolution model, take the risk probability of each network node as input, and obtain the accident cause chain;

[0048] Calculate the causal strength between each network node and the risk source node, the expression is:

[0049]

[0050] Among them, ω1, ω2, ω3, ω4 and ω5 are the weight coefficients of degree centrality, betweenness centrality, closeness centrality, eigenvector centrality and clustering coefficient, respectively. D(v), B(v), C(v), E(v) and T(v) are the measurement values ​​of degree centrality, betweenness centrality, closeness centrality, eigenvector centrality and clustering coefficient, respectively. + (v) and deg - (v) represents the in-degree and out-degree of node v, N is the total number of nodes in the topological network structure, σ st (v) and σ st are the number of shortest paths through node v and the total number of shortest paths from node s to node t, d(v,t) is the length of the shortest path from node v to node t, λ is a normalization constant, and A vt is the connection weight between node v and node t, E(t) is the feature vector center score of node t, T v is the number of edges actually existing between neighbor nodes of node v, and deg(v) is the degree of node v.

[0051] As a further method, the method for calculating the risk impact range according to the accident cause chain includes:

[0052] Taking the risk source node as the center, the radius of the risk source's impact range is calculated by quantitative assessment results. The expression is:

[0053]

[0054] Among them, N i is the risk source node, P(T|N i ) is node N i The risk probability of is, η is the attenuation coefficient of risk transmission, and τ is the minimum acceptable level of risk impact;

[0055] Based on the causal strength between each network node and the risk source node, the influence radius of the remaining nodes on the causal chain is calculated. The expression is:

[0056] R i =ψ i ×R0

[0057] Among them, R i is the influence range of the ith network node, ψ i is the causal strength between the i-th network node and the risk source node.

[0058] As a further method, the method of visualizing the network topology structure and color-mapping the accident cause chain and the risk impact range and adding them to the visualization result includes:

[0059] The risk probability of each network node in the quantitative assessment results is taken as the attribute of the node, and the risk probability of the top event is taken as the overall attribute, and the risk probability of each network node and the top event is graded;

[0060] Create a three-dimensional view of the port area network topology structure, use different colors and depths to represent the risk probability of each network node and top event, and map the accident cause chain and risk impact range to the three-dimensional view through color;

[0061] Add interactive controls that allow users to filter and view data based on risk probability level or information along the incident causal chain.

[0062] In a second aspect, an embodiment of the present application further provides an electronic device, comprising: a processor; and a memory arranged to store computer executable instructions, wherein when the executable instructions are executed, the processor executes the method steps described in the first aspect.

[0063] In a third aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores one or more programs. When the one or more programs are executed by an electronic device including multiple applications, the electronic device executes the method steps described in the first aspect.

[0064] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0065] (1) The present invention constructs an accident tree based on a set of definite risk factors and a set of uncertain risk factors for quantitative assessment, and uses Bayesian network and fuzzy logic methods to assign specific numerical values ​​to risks, providing an accurate basis for decision-making.

[0066] (2) The present invention establishes a risk evolution model and extracts the accident cause chain from the quantitative assessment results, which helps to dynamically analyze the development and propagation path of risks, warn of the spread of potential risks in advance, and provide a basis for formulating effective risk control strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 The present invention is a flowchart of the steps of a method for visualizing the risk of port accidents based on a network topology structure in an embodiment of the present invention.

[0068] Figure 2 It is a schematic diagram of the structure of an electronic device in an embodiment of this specification. DETAILED DESCRIPTION

[0069] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0070] Reference Figure 1 As shown, the present invention provides a method for visualizing and analyzing port accident risks based on a network topology structure, comprising:

[0071] S100 sets the key facilities and important areas in the port area as network nodes, determines the connection relationship between network nodes based on actual physical connections and business processes, and obtains the network topology of the port area;

[0072] Take a medium-sized comprehensive port as an example. The port mainly operates container transportation, bulk cargo handling, and oil storage.

[0073] In the actual evaluation, the network nodes are determined: the three container yards of the port are set as nodes A1, A2, and A3 respectively; the two bulk cargo terminals are set as nodes B1 and B2; the four oil storage tank areas are set as nodes C1, C2, C3, and C4; the key position of the port's main channel is set as node D1; according to the actual transportation channels, pipelines, and business processes, there is a commonly used freight channel from the container yard A1 to the bulk cargo terminal B1, so a connection is established between A1 and B1, the oil storage tank area C1 is connected to the oil loading and unloading terminal in the port through a pipeline, and a connection is established between C1 and the corresponding terminal, and finally the network topology of the port is constructed.

[0074] S200 analyzes historical accident cases of the network topology to obtain a set of definite risk factors; performs edge computing on the network topology to obtain a set of uncertain risk factors;

[0075] It should be explained that the set of uncertain risk factors is those risk factors discovered through edge computing that have not appeared in historical accident cases or are difficult to predict in advance. These factors may be caused by new technology applications, changes in the external environment or other unknown factors, such as sudden equipment failures, abnormal fluctuations in network traffic, etc. There may be no similar cases in historical accidents, but they can be detected at the moment they occur through edge computing; it reflects the risk level caused by uncertain factors in the port area network topology. Together with the first risk indicator, the second risk indicator provides us with richer information for comprehensively evaluating the risk status of the port area network topology;

[0076] In the actual assessment, the accidents in the past five years were sorted out, and a total of 15 representative accidents were found, including cargo falling due to failure of container handling equipment, dust explosion during bulk cargo loading and unloading, oil leakage and other accidents; risk factors were extracted, specifically: containers were stacked too high, bulk cargo dust concentration was not properly monitored, oil storage tanks were aging, and waterway depth measurement was inaccurate; taking an oil leakage accident as an example, the network topology data related to the accident was collected, and the node information was determined to be that the accident occurred in the oil storage tank area C2, and the connection relationship involved pipeline connections with surrounding oil loading and unloading terminals and other storage tank areas, and the flow of edges The dynamic data shows that the oil flow in the pipeline is abnormal when the leak occurs. At the same time, the safety equipment status and staff operation records in the C2 area when the accident occurs are checked. For oil leakage accidents, the key factors are determined to be the aging degree of the storage tank, the reliability of the safety equipment, and the standardization of the staff's operation. After calculation, the correlation coefficient between the aging degree of the storage tank and the degree of accident loss is 0.024; the correlation coefficient of the reliability of the safety equipment is 0.021, and the correlation coefficient of the staff's operation standardization is 0.026. The aging degree of the storage tank, the reliability of the safety equipment and the standardization of the staff's operation are obtained as the set of determined risk factors.

[0077] In the actual evaluation, edge computing is performed on each network node. In the container yard A1, the calculated state variables include container stacking height, handling equipment operating time, and traffic flow around the yard. In the bulk cargo terminal B1, the state variables include bulk cargo stacking height and loading and unloading equipment maintenance status. In the oil storage tank area C3, the state variables include tank pressure value and temperature sensor reading. The state variables between network nodes are arbitrarily combined, and the Pearson correlation coefficient is calculated to obtain the factors corresponding to the state variables greater than the preset threshold (0.7) in the correlation coefficient matrix. The factors that are repeated in the determined risk factor set are removed, and it is found that the abnormal temperature fluctuation in the oil storage tank area (corresponding to the temperature sensor reading related factor) meets the conditions in the correlation coefficient matrix and is not repeated in the determined risk factor set. It is used as a factor in the uncertain risk factor set. Finally, it is determined that the uncertain risk factor set of the port includes abnormal temperature fluctuation in the oil storage tank area, abnormal liquid level change rate, valve sealing, loading and unloading efficiency of loading and unloading equipment, and other factors.

[0078] S300 constructs an accident tree based on the determined risk factor set and the uncertain risk factor set, and performs quantitative assessment on the risk of the network topology structure based on the accident tree;

[0079] It should be explained that, by using the logical structure of the accident tree and the probability of occurrence of each event, the probability of the top event (accident) can be calculated through probability calculation methods (such as Boolean algebra method, minimum cut set method, etc.). This probability reflects the possibility of the accident under the current combination of risk factors. According to the probability of the top event and the contribution of each basic event and intermediate event to the top event, the risk of each network node and connection can be quantitatively evaluated, which can help us determine which nodes and connections are high-risk areas and need to be focused on and risk control measures should be taken;

[0080] It should be understood that if the top event is composed of multiple complex events, including both AND gates and OR gates, it is necessary to use recursive expressions until all sub-events involved in the calculation are decomposed into basic events;

[0081] In the actual assessment, the risk of the medium-sized comprehensive port is taken as the top event, and the aging degree of the storage tanks and the standardization of the staff's operation in the risk factor set, as well as the abnormal temperature fluctuation and abnormal liquid level change rate in the oil storage tank area in the uncertain risk factor set are determined as basic events. The logical relationship construction includes: if the storage tank is seriously aged and the staff's operation is not standardized, resulting in major accidents such as oil leakage, these two basic events are connected with the gate; for uncertain risk factors such as abnormal temperature fluctuation and abnormal liquid level change rate, if they appear alone or in combination with certain risk factors, the probability of accidents will increase. According to the actual situation, the or gate connection is used. Among them, the temperature fluctuation is abnormal and the liquid level change rate is abnormal. At the same time, the aging degree of the storage tank is high, which will greatly increase the risk of oil leakage. At this time, the three basic events are connected with the gate to gradually construct the accident tree;

[0082] In the actual assessment, the Bayesian network is used to calculate the risk probability of a set of risk factors, the fuzzy logic is used to calculate the risk probability of a set of uncertain risk factors, the risk probability of network nodes and top events is calculated through the accident tree, and the risk probability of each network node and top event is used as a quantitative assessment result. Specifically, the probability of serious aging of storage tanks is 0.22, the probability of improper operation of staff is 0.15, the probability of abnormal temperature fluctuation risk is 0.18, the probability of abnormal liquid level change rate is 0.16, the probability of risk of nodes in oil storage tank areas is 0.28, and the probability of top events (risks in ports) is 0.35.

[0083] S400 constructs a risk evolution model, extracts the accident cause chain from the quantitative assessment results through the risk evolution model, and calculates the risk impact range according to the accident cause chain;

[0084] It should be explained that by constructing a risk evolution model, we can simulate the changes of risks over time and various factors, predict the development trend of potential risks in advance, and extract the cause chain of accidents to help clarify the root causes and paths of accidents, providing direction for taking targeted preventive measures. Calculating the scope of risk impact can enable managers to clearly understand the areas and extent that accidents may affect, so as to reasonably allocate resources for emergency response and risk control.

[0085] In the actual assessment, a Markov chain is used to construct a risk evolution model. The state set is {safe, low risk, medium risk, high risk}, the probability distribution of the determined risk factor set is {0.45, 0.35, 0.15, 0.05}, and the uncertain risk factor set is {0.42, 0.32, 0.2, 0.06}. Data is obtained from historical cases to divide the training set and other training models. The node risk probability is input into the model. The node probability of the oil storage tank area is 0.25, and the node probability of the container yard is 0.3. The accident cause chain is obtained, and the risk impact range is calculated. The probability of the risk source node is 0.35, the attenuation coefficient is 0.8, the minimum acceptable level is 0.1, and the radius is 0.65, of which the radius of the impact range of a certain node is 0.26.

[0086] S500 visualizes the network topology structure, performs color mapping on the accident cause chain and the risk impact range, and adds them to the visualization result to obtain a three-dimensional topology view.

[0087] In the actual assessment, the risk probability of each network node is graded into low (0-0.2), medium (0.2-0.4), and high (above 0.4) risk, and a three-dimensional view is created. Low-risk nodes are light blue, medium-risk nodes are moderate yellow, and high-risk nodes are prominent red. The accident causal chain is displayed with blue luminous lines, and the risk impact range is calculated based on translucent circles of different radii. Interactive controls are added, and users can click on nodes to view risk probabilities, select causal chain nodes to focus on and view related regional risks, and intuitively present the risk status of the port area.

[0088] In this embodiment, the method of analyzing historical accident cases of the network topology structure to obtain a set of determined risk factors includes:

[0089] Obtain accident records of historical accident cases, collect network topology data related to the accident, including node information, connection relationships, and edge flow data, and identify key factors that led to the accident;

[0090] Calculate the correlation coefficient between key factors and accident loss degree, the expression is:

[0091]

[0092] Where m is the number of historical accident cases, F ik Denoted as the key factor F in the kth case i The value of The key factor F i The mean value, L k is the accident loss degree of the kth case, is the mean of the accident loss extent of all cases;

[0093] The key factors whose correlation coefficients with the accident loss degree are greater than 0.01 are taken as the determined risk factors to obtain the determined risk factor set.

[0094] In this embodiment, the method of performing edge computing on the network topology structure to obtain a set of uncertain risk factors includes:

[0095] Perform edge computing on each network node to obtain the state variables of each network node;

[0096] The state variables between network nodes are combined arbitrarily, and the correlation coefficient matrix is ​​obtained by calculating the Pearson correlation coefficient;

[0097] The factors corresponding to the state variables greater than a preset threshold in the correlation coefficient matrix are obtained, and the factors repeated in the determined risk factor set are removed to obtain the uncertain risk factor set.

[0098] In this embodiment, the method for constructing an accident tree based on the determined risk factor set and the uncertain risk factor set includes: taking the risk existing in the port area as the top event, taking each factor in the determined risk factor set and the uncertain risk factor set as the basic event, and according to the logical relationship, using logical gates such as AND gates and OR gates to connect the basic events to obtain intermediate events and form an accident tree.

[0099] In this embodiment, the method for quantitatively evaluating the risk of a network topology structure based on the fault tree includes:

[0100] S21 uses the Bayesian network to calculate the risk probability of the corresponding basic events in the risk factor set, specifically:

[0101] The basic events in the accident tree are used as nodes of the Bayesian network to construct the Bayesian network;

[0102] The kernel function is established using the hyperparameter method to modify the Bayesian network. The expression of the kernel function is:

[0103]

[0104] Among them, α is the global scaling factor, x and y are the state variable and feature variable of the event in the Bayesian network, σ is the bandwidth of the Gaussian radial basis function, T represents the transposition operation, M is the normalization factor, d is the degree of the polynomial kernel, β and b are the slope and bias of the Sigmoid kernel, and p is the nonlinear exponent of the kernel function;

[0105] The probability distribution in the Bayesian network is updated by maximizing expectation. The updated expression is:

[0106]

[0107] Among them, θ is the parameter set in the Bayesian network, θ t and θ (t+1) are the estimates of θ in the tth iteration and the t+1th iteration respectively, arg maxθ is the maximization operator, is the expectation operator, P(X hidden |D,θ (t) ) is given the observed data D and parameter estimates θ (t) Under the condition that the hidden node X hidden The posterior probability distribution of is the log-likelihood function of the joint probability distribution, n is the number of nodes in the Bayesian network, P(X i |pa(X i );θ) represents the estimated node X that depends on the parameter θ i Given its parent node pa(X i )’s conditional probability;

[0108] The probability distribution in the Bayesian network is used as the risk probability of the basic event;

[0109] S22 uses fuzzy logic to calculate the risk probability of the corresponding basic events in the uncertain factor set, specifically:

[0110] The expression of the membership function is:

[0111]

[0112] Among them, λ i is the i-th uncertain risk factor, c i For x i The central value of is i The standard deviation of

[0113] The membership function is defuzzified by the centroid method to obtain the risk probability of the basic event;

[0114] S23 calculates the risk probability of network nodes and top events through the accident tree according to the risk probability of basic events. The expression is:

[0115]

[0116] Among them, E is a basic event or a complex event composed of basic events, P(ξ i ) is the basic event ξ i The probability of risk;

[0117] S24 takes the risk probability of each network node and top event as a quantitative assessment result.

[0118] In this embodiment, the method of constructing a risk evolution model and extracting an accident cause chain from a quantitative assessment result through the risk evolution model includes:

[0119] The risk evolution model is constructed using the Markov chain algorithm, where the expression of the state transition function is:

[0120]

[0121] Among them, P(X n+1 =j|X n =i) is the probability of transferring to the target state j at time step n+1 under the condition that the state is in state i at time step n, S is the set of all states, U is the set of determined risk factors, and p u (n) is the probability distribution of the determined risk factor u at time step n, V is the set of uncertain risk factors, q v (n) is the probability distribution of the uncertain risk factor v at time step n, p kj is the transition probability from the intermediate state k to the target state j;

[0122] Obtain historical accident cause chain data from historical accident cases, divide them into training set, test set and validation set in a ratio of 8:1:1 to train the risk evolution model, take the risk probability of each network node as input, and obtain the accident cause chain;

[0123] Calculate the causal strength between each network node and the risk source node, the expression is:

[0124]

[0125] Among them, ω1, ω2, ω3, ω4 and ω5 are the weight coefficients of degree centrality, betweenness centrality, closeness centrality, eigenvector centrality and clustering coefficient, respectively. D(v), B(v), C(v), E(v) and T(v) are the measurement values ​​of degree centrality, betweenness centrality, closeness centrality, eigenvector centrality and clustering coefficient, respectively. + (v) and deg - (v) represents the in-degree and out-degree of node v, N is the total number of nodes in the topological network structure, σ st (v) and σ st are the number of shortest paths through node v and the total number of shortest paths from node s to node t, d(v,t) is the length of the shortest path from node v to node t, λ is a normalization constant, and A vt is the connection weight between node v and node t, E(t) is the feature vector center score of node t, T v is the number of edges actually existing between neighbor nodes of node v, and deg(v) is the degree of node v.

[0126] In this embodiment, the method for calculating the risk impact range according to the accident cause chain includes:

[0127] Taking the risk source node as the center, the radius of the risk source's impact range is calculated by quantitative assessment results. The expression is:

[0128]

[0129] Among them, N i is the risk source node, P(T|N i ) is node N i The risk probability of is, η is the attenuation coefficient of risk transmission, and τ is the minimum acceptable level of risk impact;

[0130] Based on the causal strength between each network node and the risk source node, the influence radius of the remaining nodes on the causal chain is calculated. The expression is:

[0131] R i =ψ i ×R0

[0132] Among them, R i is the influence range of the ith network node, ψ i is the causal strength between the i-th network node and the risk source node.

[0133] In this embodiment, the method of visualizing the network topology structure, color mapping the accident cause chain and the risk impact range, and adding them to the visualization result includes:

[0134] The risk probability of each network node in the quantitative assessment results is taken as the attribute of the node, and the risk probability of the top event is taken as the overall attribute, and the risk probability of each network node and the top event is graded;

[0135] Create a three-dimensional view of the port area network topology structure, use different colors and depths to represent the risk probability of each network node and top event, and map the accident cause chain and risk impact range to the three-dimensional view through color;

[0136] Add interactive controls that allow users to filter and view data based on risk probability level or information along the incident causal chain.

[0137] Figure 2 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present application. Figure 2At the hardware level, the electronic device includes a processor, and optionally also includes an internal bus, a network interface, and a memory. The memory may include a memory, such as a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage. Of course, the electronic device may also include hardware required for other services.

[0138] The processor, network interface and memory can be interconnected through an internal bus, which can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc.

[0139] The memory is used to store the program. Specifically, the program may include a program code, and the program code includes a computer operation instruction. The memory may include a memory and a non-volatile memory, and provides instructions and data to the processor.

[0140] The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it, forming a visual analysis device for port accident risk based on the network topology structure at the logical level. The processor executes the program stored in the memory and is specifically used to execute any of the aforementioned visual analysis methods for port accident risk based on the network topology structure.

[0141] The present invention can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in the processor or an instruction in the form of software. The above processor can be a general-purpose processor, including a central processing unit, a network processor (Network Processor, NP), etc.; it can also be a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application can be directly embodied as a hardware decoding processor to be executed, or the hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0142] An embodiment of the present application also proposes a computer-readable storage medium, which stores one or more programs, and the one or more programs include instructions. When the instructions are executed by an electronic device including multiple applications, any one of the aforementioned port accident risk visualization analysis methods based on network topology structure is executed.

[0143] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A visual analysis method for port accident risk based on network topology, characterized in that: The following steps are involved: The key facilities and important areas in the port area are set as network nodes, and the connection relationship between network nodes is determined according to the actual physical connection and business process to obtain the network topology structure of the port area; Analyze historical accident cases of network topology to obtain a set of determined risk factors; Perform edge computing on the network topology to obtain a set of uncertain risk factors; Constructing an accident tree based on the determined risk factor set and the uncertain risk factor set, and quantitatively evaluating the risk of the network topology structure based on the accident tree; Constructing a risk evolution model, extracting the accident cause chain from the quantitative assessment results through the risk evolution model, and calculating the risk impact range according to the accident cause chain; The network topology is visualized, and the accident cause chain and the risk impact range are color mapped and added to the visualization result to obtain a three-dimensional topology view.

2. According to claim 1, a method for visualizing the risk of port accidents based on network topology structure is characterized in that: The method of analyzing historical accident cases of network topology structure to obtain a set of determined risk factors includes: Obtain accident records of historical accident cases, collect network topology data related to the accident, including node information, connection relationships, and edge flow data, and identify key factors that led to the accident; Calculate the correlation coefficient between key factors and accident loss degree, the expression is: Where m is the number of historical accident cases, F ik Denoted as the key factor F in the kth case i The value of The key factor F i The mean value, L k is the accident loss degree of the kth case, is the mean of the accident loss extent of all cases; The key factors whose correlation coefficients with the accident loss degree are greater than 0.01 are taken as the determined risk factors to obtain the determined risk factor set.

3. According to claim 1, a method for visualizing the risk of port accidents based on network topology structure is characterized in that: The method of performing edge computing on a network topology structure to obtain a set of uncertain risk factors includes: Perform edge computing on each network node to obtain the state variables of each network node; The state variables between network nodes are combined arbitrarily, and the correlation coefficient matrix is ​​obtained by calculating the Pearson correlation coefficient; The factors corresponding to the state variables greater than a preset threshold in the correlation coefficient matrix are obtained, and the factors repeated in the determined risk factor set are removed to obtain the uncertain risk factor set.

4. According to claim 1, a method for visualizing the risk of port accidents based on network topology structure is characterized in that: The method for constructing an accident tree based on the determined risk factor set and the uncertain risk factor set includes: taking the risk existing in the port area as the top event, taking each factor in the determined risk factor set and the uncertain risk factor set as the basic event, and according to the logical relationship, using the logical gates of AND gates and OR gates to connect the basic events to obtain intermediate events and form an accident tree.

5. According to claim 1, a method for visualizing and analyzing port accident risks based on network topology structure is characterized in that: The method for quantitatively evaluating the risk of a network topology structure based on the fault tree comprises: S21 uses the Bayesian network to calculate the risk probability of the corresponding basic events in the risk factor set, specifically: The basic events in the accident tree are used as nodes of the Bayesian network to construct the Bayesian network; The kernel function is established using the hyperparameter method to modify the Bayesian network. The expression of the kernel function is: Among them, α is the global scaling factor, x and y are the state variable and feature variable of the event in the Bayesian network, σ is the bandwidth of the Gaussian radial basis function, T represents the transposition operation, M is the normalization factor, d is the degree of the polynomial kernel, β and b are the slope and bias of the Sigmoid kernel, and p is the nonlinear exponent of the kernel function; The probability distribution in the Bayesian network is updated by maximizing expectation. The updated expression is: Among them, θ is the parameter set in the Bayesian network, θ t and θ (t+1) are the estimates of θ in the tth iteration and the t+1th iteration respectively, arg maxθ is the maximization operator, is the expectation operator, P(X hidden |D,θ (t) ) is given the observed data D and parameter estimates θ (t) Under the condition that the hidden node X hidden The posterior probability distribution of is the log-likelihood function of the joint probability distribution, n is the number of nodes in the Bayesian network, P(X i |pa(X i );θ) represents the estimated node X that depends on the parameter θ i Given its parent node pa(X i )’s conditional probability; The probability distribution in the Bayesian network is used as the risk probability of the basic event; S22 uses fuzzy logic to calculate the risk probability of the corresponding basic events in the uncertain factor set, specifically: The expression of the membership function is: Among them, λ i is the i-th uncertain risk factor, c i For x i The central value of is i The standard deviation of The membership function is defuzzified by the centroid method to obtain the risk probability of the basic event; S23 calculates the risk probability of network nodes and top events through the accident tree according to the risk probability of basic events. The expression is: Where E is a basic event or a complex event composed of basic events, and P(ξi) is the risk probability of the basic event ξi; S24 takes the risk probability of each network node and top event as a quantitative assessment result.

6. The method for visualizing the risk of port accidents based on network topology structure according to claim 1 is characterized in that: The method of constructing a risk evolution model and extracting an accident cause chain from a quantitative assessment result through the risk evolution model includes: The risk evolution model is constructed using the Markov chain algorithm, where the expression of the state transition function is: P(X n+1 =j|X n =i)=Σ k∈s (Π u∈U p u (n)(Π h∈H p h (n))p kj Among them, P(X n+1 =j|X n =i) is the probability of transferring to the target state j at time step n+1 under the condition that the state is in state i at time step n, S is the set of all states, U is the set of determined risk factors, and p u (n) is the probability distribution of the determined risk factor u at time step n, H is the set of uncertain risk factors, and p h (n) is the probability distribution of the uncertain risk factor h at time step n, pk j is the transition probability from the intermediate state k to the target state j; Obtain historical accident cause chain data from historical accident cases, divide them into training set, test set and validation set in a ratio of 8:1:1 to train the risk evolution model, take the risk probability of each network node as input, and obtain the accident cause chain; Calculate the causal strength between each network node and the risk source node, the expression is: Among them, ω1, ω2, ω3, ω4 and ω5 are the weight coefficients of degree centrality, betweenness centrality, closeness centrality, eigenvector centrality and clustering coefficient, respectively. D(v), B(v), C(v), E(v) and T(v) are the measurement values ​​of degree centrality, betweenness centrality, closeness centrality, eigenvector centrality and clustering coefficient, respectively. + (v) and deg - (v) represents the in-degree and out-degree of node v, N is the total number of nodes in the topological network structure, σ st (v) and σ st are the number of shortest paths through node v and the total number of shortest paths from node s to node t, d(v,t) is the length of the shortest path from node v to node t, λ is a normalization constant, and A vt is the connection weight between node v and node t, E(t) is the feature vector center score of node t, T v is the number of edges actually existing between neighbor nodes of node v, and deg(v) is the degree of node v.

7. The method for visualizing the risk of port accidents based on network topology structure according to claim 1 is characterized in that: The method for calculating the risk impact scope according to the accident cause chain includes: Taking the risk source node as the center, the radius of the risk source's impact range is calculated by quantitative assessment results. The expression is: Among them, N i is the risk source node, P(T|N i ) is node N i The risk probability of is, η is the attenuation coefficient of risk transmission, and τ is the minimum acceptable level of risk impact; Based on the causal strength between each network node and the risk source node, the influence radius of the remaining nodes on the causal chain is calculated. The expression is: R i =ψ i ×R0 Among them, R i is the influence range of the ith network node, ψ i is the causal strength between the i-th network node and the risk source node.

8. The method for visualizing the risk of port accidents based on network topology structure according to claim 1 is characterized in that: The method of visualizing the network topology structure, color mapping the accident cause chain and the risk impact range, and adding them to the visualization result includes: The risk probability of each network node in the quantitative assessment results is taken as the attribute of the node, and the risk probability of the top event is taken as the overall attribute, and the risk probability of each network node and the top event is graded; Create a three-dimensional view of the port area network topology structure, use different colors and depths to represent the risk probability of each network node and top event, and map the accident cause chain and risk impact range to the three-dimensional view through color; Add interactive controls that allow users to filter and view data based on risk probability level or information along the incident causal chain.

9. An electronic device, comprising: processor; as well as A memory arranged to store computer executable instructions, which when executed cause the processor to perform the method according to any one of claims 1 to 8.

10. A computer-readable storage medium storing one or more programs, which, when executed by an electronic device including a plurality of application programs, enables the electronic device to execute the method according to any one of claims 1 to 8.

Images