A method for early warning of quality risks in international railway freight transportation based on blockchain

Abstract

The present invention discloses a method for early warning of the quality risk of international railway freight transportation based on blockchain. The present invention includes the following steps: Step A: Identify the risk sources of international railway freight transportation and clarify the risk factors of international railway freight transportation; Step B: Establish a Bayesian network-based early warning model for the combined transportation risk of international railway freight transportation based on the Bayesian network to evaluate the level of transportation quality risk; Step C: According to the requirements of each participant in international railway freight transportation, construct and design a blockchain-based early warning framework for the quality risk of freight transportation and the corresponding functional modules. By constructing a blockchain-based early warning mechanism for the combined transportation risk of international railway freight transportation, the present invention can identify the quality risk factors of freight transportation, and at the same time realize the timely sharing of freight transportation information, so as to break the existing information sharing barriers in international railway freight transportation, improve the freight transportation safety risk early warning mechanism, and improve the service quality of international railway freight transportation.

Description

Technical Field

[0001] The present invention relates to the field of railway transportation, and particularly to a method for warning of quality risks in international railway freight transportation based on blockchain. Background Art

[0002] International railway cross-border transportation is a transportation mode second only to sea transportation in international trade. Its greatest advantages lie in its relatively large transportation volume, relatively fast speed, significantly lower transportation risks than sea transportation, and the ability to maintain punctual operation throughout the year. As a special case of railway transportation, international railway cross-border transportation currently has the following problems:

[0003] (1) The legal norms and freight transportation regulations along the international railway transportation routes of various countries are different, and there are significant obstacles to the sharing of freight transportation information between countries.

[0004] (2) The use of paper waybills makes it difficult to ensure the integrity of in-transit information of freight transportation.

[0005] (3) There is a lack of an efficient and perfect information co-construction and sharing mechanism, and it is impossible to achieve timely and effective tracking of the whole process of international railway freight transportation.

[0006] (4) It is impossible to timely warn of abnormal conditions in the quality of international railway freight transportation.

[0007] (5) The international railway freight transportation has a long distance span, crosses many countries and regions, involves many participating parties, and it is difficult to trace and investigate the key information of accidents in a timely manner after a transportation accident occurs.

[0008] The above existing problems in international railway freight transportation indicate that it is urgent to solve the problems of untimely information transmission, incomplete information sharing, lagging risk warning, and difficult accident backtracking in international railway freight transportation. Summary of the Invention

[0009] In view of the above deficiencies of the prior art, the present invention provides a method for warning of quality risks in international railway freight transportation based on blockchain.

[0010] To achieve the above invention object, the technical solution adopted by the present invention is as follows:

[0011] It includes the following steps:

[0012] Step A: Identify the risk sources of international railway freight transportation and clarify the risk factors of international railway freight transportation;

[0013] Determine the quality risks of international railway freight transportation, and identify the risk factors of the key paths and key links in the whole process of international railway freight transportation from the participating parties and links of international railway freight transportation by using the HAZOP method, network planning and interpretive structural modeling;

[0014] The quality risk sources of goods transportation include goods safety risk factors and train delay risk factors;

[0015] The HAZOP method is used to identify the risks of port operations and in-transit transportation for goods safety risk factors respectively;

[0016] The network planning diagram and Interpretive Structural Modeling (ISM) are used to identify the risks of port operations for train delay risk factors, and the network planning diagram, Interpretive Structural Modeling (ISM) and HAZOP method are used to identify the risks of in-transit transportation for train delay risk factors;

[0017] Step B: Hierarchically classify the risk factors, construct a Bayesian network for the quality risk of international railway transportation based on the relationships among the risk factors, and establish an early warning model for the combined transportation risk of international railway goods transportation based on the Bayesian network;

[0018] Step C: According to the needs of all parties involved in international railway goods transportation for applying blockchain technology to train transportation risk early warning, based on the Bayesian network of the quality risk of international railway transportation, construct and design a blockchain-based quality risk early warning framework and corresponding functional modules for goods transportation, and establish a blockchain-based quality risk early warning platform for international railway goods transportation;

[0019] Select the Fabric framework and the PoA consensus algorithm, design a blockchain-based quality risk early warning platform for international railway goods transportation, and design the functional modules.

[0020] Furthermore, the risk factors are classified according to the parties involved. For the consignee, there is container damage; for the carrier, there are inadequate equipment management, insufficient personnel management, imperfect rules and regulations, imperfect supervision facilities and equipment, abnormal car body circulation management, inflexible route selection, inadequate vehicle management, imperfect station operation supervision system, imperfect station security supervision system, imperfect railway risk prevention supervision, imperfect supervision facilities and equipment, imperfect production safety supervision, inadequate station supervision, insufficient transshipment capacity, failure to add epidemic prevention and control measures, low operation efficiency, inadequate safety awareness, unstable political situation, container damage, improper human operation, railway suspension or construction, and damaged freight car components; for the customs, there is overtime operation time; for others, there are natural disasters.

[0021] Furthermore, Step B also includes the following steps:

[0022] Step B1: Divide the levels of the transportation quality risk system

[0023] Hierarchically classify the risk factors from five aspects: organizational management, risk supervision, risk incentives, direct causes, and transportation quality risks;

[0024] Step B2: Determine the Bayesian network structure of transportation quality risk

[0025] Analyze the mutual relationship between risk factors to obtain the structure of the Bayesian network. Through the mutual relationship of the internal factors of the data, represent the relationship between nodes with an adjacency matrix Q.

[0026]

[0027] Among them, q ij represents the relationship of the i-th risk factor to the j-th risk factor. When q ij =1, the i-th risk factor has an impact on the j-th risk factor. When q ij =0, there is no impact.

[0028] Draw the Bayesian network of transportation quality risk according to the adjacency matrix.

[0029] Step B3: Determine the transportation quality risk data set

[0030] Perform hierarchical extraction of transportation quality risk factors on the collected data to obtain a data set related to transportation quality risk accidents and train delays.

[0031] Let be the data set, where D i is the set of risk factor states associated with the i-th row of data.

[0032] D i =[u i1 u i2 ... u in

[0033] If the i-th row of data is associated with risk factor A1, then u i1 =risk factor A1, otherwise u i1 =risk factor A2; n represents the number of risk factors extracted from the i-th row of data.

[0034] For the D data set, it is represented by u i1 as:

[0035]

[0036] Step B4: Divide the hazard levels of transportation quality risks

[0037] Since there are corresponding probability relationships for all nodes in the Bayesian network, comprehensively consider the sensitivity level of the node relative to the final node and the network level where the node risk itself is located. The sensitivity level calculation formula is as follows:

[0038] S i =5Pe i ​

[0039] Perform weighted calculation and round up to obtain the risk hazard level of the node:

[0040] E i = [αS i + βC i

[0041] where S i is the sensitivity level, Pe i is the proportion of the sensitivity of the transportation quality risk (F4) to risk factor i to the sensitivity of the transportation quality risk (F4) to itself, E i is the risk hazard level, C i is the level of the Bayesian network where risk factor i is located; since the influence of the node itself on the probability of transportation quality risk is weaker than the direct correlation of the consequences caused by the node risk itself, so take α = 0.3 and β = 0.7;

[0042] Step B5: Calculate the risk values of each risk factor

[0043] Take the product of the risk occurrence probability and the risk hazard level as the risk value. Let E i be the risk hazard level of risk factor i, and P i be the risk occurrence probability level of risk factor i. Then the risk value calculation formula is as follows:

[0044] R i = E i P i

[0045] Step B6: Determine the transportation quality risk warning level

[0046] Use the risk occurrence probability level P i and the risk hazard level E i to establish a risk matrix diagram, and use the risk matrix diagram to divide the risk value into five warning levels including high, relatively high, medium, relatively low, and low; when the risk value R i ∈ {1, 2}, risk factor i is in a low-risk state; R i ∈ {3, 4, 6}, risk factor i is in a relatively low-risk state; R i ∈ {5, 8, 9}, risk factor i is in a medium-risk state; R i ∈ {10, 12, 15, 16}, risk factor i is in a relatively high-risk state; R i ∈ {20, 25}, risk factor i is in a high-risk state.

[0047] Step B7: Complete the establishment of the international railway freight transport through train risk warning model based on the Bayesian network.

[0048] ​Furthermore, organizational management includes subsystems such as container management, personnel management, rules and regulations, facility and equipment management, car body management, line management, vehicle management, etc.; each subsystem corresponds to nodes O1, O2, O3, O4, O5, O6, O7 respectively; node O1 corresponds to A1 with in-place equipment management and A2 with inadequate equipment management; node O2 corresponds to A1 with normal personnel management and A2 with insufficient personnel management; node O3 corresponds to A1 with perfect rules and regulations and A2 with imperfect rules and regulations; node O4 corresponds to A2 with perfect supervision facilities and equipment and A2 with imperfect supervision facilities and equipment; node O5 corresponds to A1 with normal car body circulation management and A2 with abnormal car body circulation management; node O6 corresponds to A1 with flexible line selection and A2 with inflexible line selection; node O7 corresponds to A1 with in-place vehicle management and A2 with inadequate vehicle management;

[0049] Risk supervision includes subsystems such as operation supervision, security supervision, risk prevention supervision, supervision facilities, production safety supervision, etc.; each subsystem corresponds to nodes M1, M2, M3, M4, M5 respectively; node M1 corresponds to A1 with a perfect station operation supervision system and A2 with an imperfect station operation supervision system; node M2 corresponds to A1 with a perfect station security supervision system and A2 with an imperfect station security supervision system; node M3 corresponds to A1 with imperfect railway risk prevention supervision and A2 with perfect railway risk prevention supervision; node M4 corresponds to A1 with perfect supervision facilities and equipment and A2 with imperfect supervision facilities and equipment; node M5 corresponds to A1 with perfect production safety supervision and A2 with imperfect production safety supervision;

[0050] Risk incentives include subsystems such as natural environment, operation environment, equipment capacity, epidemic prevention and control, operation efficiency, safety awareness, political factors, etc.; each subsystem corresponds to nodes S1, S2, S3, S4, S5, S6, S7 respectively; node S1 corresponds to A1 with normal conditions and A2 with natural disasters occurring; node S2 corresponds to A1 with in-place station supervision and A2 with inadequate station supervision; node S3 corresponds to A1 with insufficient transshipment capacity and A2 with insufficient transshipment capacity; node S4 corresponds to A1 with no additional epidemic prevention measures and A2 with additional epidemic prevention measures; node S5 corresponds to A1 with normal operation efficiency and A2 with low operation efficiency; node S6 corresponds to A1 with in-place safety awareness and A2 with inadequate safety awareness; node S7 corresponds to A1 with stable political situation and A2 with unstable political situation;

[0051] The direct causes include subsystems such as container status, human operation, railway suspension or construction, operation time, and freight car component status; each subsystem corresponds to nodes Z1, Z2, Z3, Z4, and Z5; node Z1 corresponds to A1 normal container and A2 damaged container; node Z2 corresponds to A1 no improper human operation and A2 improper human operation; node Z3 corresponds to A1 no railway suspension or construction and A2 railway suspension or construction; node Z4 corresponds to A1 normal operation time and A2 overtime operation time; node Z5 corresponds to A1 normal freight car components and A2 damaged freight car components.

[0052] The transport quality risks include subsystems such as accidents, unexpected events other than accidents, train delays, and transport quality risks; each subsystem corresponds to nodes F1, F2, F3, and F4 respectively.

[0053] The beneficial effects of the present invention are as follows:

[0054] 1. By using the constructed international railway freight transport quality risk early warning mechanism, the gaps in the existing information mechanism are filled, and the degree of information sharing among countries is improved.

[0055] 2. The complete traceability of freight transport data and the whole-process tracking of the goods status are realized.

[0056] 3. Certain safety pre-management of risks can be carried out to reduce the probability of risk occurrence.

[0057] 4. Timely emergency responses are made to the risks that have occurred to avoid greater transport losses.

[0058] 5. The responsibility division and accountability process for transport accidents, etc. are unified among countries along the international railway.

[0059] 6. By constructing an international railway freight transport intermodal risk early warning mechanism based on blockchain, the factors of freight transport quality risks are identified, and at the same time, the timely sharing of freight transport information is realized, so as to break the existing information sharing barriers in international railway freight transport, improve the freight transport safety risk early warning mechanism, and improve the quality of international railway freight transport services. Description of the Drawings

[0060] Figure 1 It is a flowchart for identifying risk factors of key paths and key links in the whole process of international railway freight transport;

[0061] Figure 2 It is a flowchart of the HAZOP method identification process;

[0062] Figure 3 It is a Bayesian network diagram of transport quality risks;

[0063] Figure 4 It is an architecture diagram of the international railway freight transport blockchain;

[0064] Figure 5 is a risk warning flow chart based on blockchain;

[0065] Figure 6 is a goods tracking flow chart based on blockchain;

[0066] Figure 7 is an insured value insurance claim settlement flow chart based on blockchain;

[0067] Figure 8 is a risk matrix diagram. Specific implementation manners

[0068] The specific implementation manners of the present invention are described below to facilitate the understanding of the present invention by those skilled in the art of the present technology. However, it should be clear that the present invention is not limited to the scope of the specific implementation manners. For those of ordinary skill in the art of the present technology, as long as various changes are within the spirit and scope of the present invention defined and determined by the appended claims, these changes are obvious, and all inventions created using the concept of the present invention are within the scope of protection.

[0069] Step A: Identify the risk sources of international railway freight transportation and clarify the risk factors of international railway freight transportation;

[0070] Identify and sort out the risk sources of freight transportation quality from the participants and links in the freight transportation process. The risk sources of freight transportation quality include cargo safety risk factors and train delay risk factors;

[0071] Use the HAZOP method to identify the risks of port operations and risks during transit for cargo safety risk factors respectively;

[0072] Use the network planning diagram and interpretive structural modeling (ISM) to identify the risks of port operations for train delay risk factors, and use the network planning diagram, interpretive structural modeling (ISM) and HAZOP method to identify the risks during transit for train delay risk factors.

[0073] Use the network planning diagram, interpretive structural modeling (ISM) and HAZOP method to analyze and determine the basic operation links of international railway freight transportation. At the same time, consider the impact of path operation time and the overall structure of the process on the operation process efficiency, calculate the vulnerability of each link to obtain the key links of train delay, and then conduct parameter analysis on the key links in the order of operations, so as to identify the risks of port operations for train delay risk factors; Use the network planning diagram, interpretive structural modeling (ISM) and HAZOP method to sort out the relevant processes of each basic operation of international railway freight transportation, identify and analyze the key process nodes and deviations affecting cargo safety, and then conduct parameter analysis on the key process nodes in the order of the process, so as to identify the risks during transit for train delay risk factors.

[0074] A1: Use the HAZOP method to identify risk factors. The specific main steps are as follows;

[0075] A11: Sort out the operation process; A12: Identify key nodes; A13: Analyze key node parameters; A14: Analyze parameter deviations; A15: If all parameter analyses are completed, proceed to the next step; otherwise, return to step A14; A16: If all key node analyses are completed, proceed to the next step; otherwise, return to step A13; A17: The identification of risk factors is completed.

[0076] A2: Use the network planning diagram combined with GTFN to identify the key path of railway port station operations

[0077] The key path is a path in the operation process that takes the longest time from start to end and is composed of operation links in the entire plan; assume that P is an arbitrary path from node 1 to node n in network G, and p n is the set of all paths from node 1 to node n, and x - y is any operation link in the network planning diagram. Then the label value of the end point of path P is denoted as If Then P * is the key path in the network planning diagram, and the duration of the entire network plan is

[0078] A21: Identify the key path of railway port station operations based on the network planning diagram. The specific steps are as follows:

[0079] A211: Let Pred(1) = 0,

[0080] A212: Along the arrow direction, from node 2 to node n, calculate respectively If node i is the node with the largest label value pointing to node j, then sour(j) = i;

[0081] A213: If j = n, then p * = {n, Sour(n), Sour(Sour(n)),..., 1}, otherwise continue with step 2 to calculate the label value of the next node;

[0082] In the above formula, j: any node in the network planning diagram; i: one of the nodes in the network planning diagram that points to node j; Pred(j): the node pointing to j, j ∈ (1, n]; Sour(j): the previous adjacent node of node j in the key path; m: the number of nodes pointing to j, m ∈ [1, j); The label value of the kth node pointing to j, k ∈ [1, m];

[0083] A22: Considering the randomness of the operation time in the actual operation process, the method of the present invention selects a random distribution to represent the operation time of each link during the operation time investigation. According to the data characteristics, the normal distribution function and the uniform distribution function are used to represent the operation time; for the convenience of calculation, the distribution is approximately transformed into a generalized fuzzy number;

[0084] A23: Combining the network planning diagram and the generalized fuzzy number, calculate the label value of the network node to obtain the network critical path;

[0085] A3: Construct an interpretive structural model and adjust the vulnerability of the key links (the degree of influence of the change of local operation links on the overall operation)

[0086] A31: The main steps of constructing the interpretive structural model are as follows:

[0087] A311: Operation link analysis; analyze the operation process to determine the basic operation links {E1, E2...} in the process;

[0088] A312: Construct a directed graph; analyze the relationship between the operation links in the operation process and construct a directed graph;

[0089] A313: Establish an adjacency matrix; use a Boolean matrix to represent the directed relationship between operation links. Let the adjacency matrix be A. When the operation link Ei has a direct influence on the operation link Ej, the matrix element a ij takes 1, otherwise takes zero;

[0090] A314: Solve the reachability matrix; add the adjacency matrix A to the identity matrix I, and multiply the matrix continuously until the matrix does not change to obtain the reachability matrix R;

[0091] (A + I) k-1 ≠(A + I) k =(A + I) k+1 =R

[0092] That is, B k-1 ≠B k =B k+1 =R

[0093] A315: According to the obtained reachability matrix, perform regional decomposition and hierarchical decomposition on the operation process; construct an interpretive structural model;

[0094] A32: Calculate the critical vulnerability of the operation link;

[0095] Merge the operation links in the critical path with the links in the interpretive structural model where the number of levels is greater than or equal to 2. Calculate the similarity between the operation time of these links and the ideal information vector, and finally sort them from largest to smallest according to the similarity. The sorting number of the operation link i-j is denoted as x ij , make x ij dimensionless to obtain the critical vulnerability O of the operation link i-j ij ;

[0096]

[0097] Where: i-j is the operation link belonging to the critical path;

[0098] The number of operation links at the same level of the link i-j is denoted as Num i-j ;

[0099]

[0100] Where: when the link i-j and the link k-m are at the same level, x ijkm = 1, otherwise it is equal to 0. Perform dimensionless processing on the obtained Num i-j for each operation link to obtain the hierarchical vulnerability I of each operation link i-j ;

[0101]

[0102] A33: Obtain the comprehensive vulnerability of each link by weighting;

[0103] Take the weight of the hierarchical vulnerability as 0.3 and the critical vulnerability as 0.7, then obtain the comprehensive vulnerability N of each link i-j , as follows:

[0104] N i-j = 0.3I i-j + 0.7O i-j

[0105] A4: Classify the cargo safety risk factors and the train delay risk factors according to the involved parties of the risk, as shown in the following table:

[0106] Risk factors classified by involved parties

[0107]

[0108] Step B: Construct an early warning model for the quality risk of international railway freight transportation based on the Bayesian network

[0109] Step B1: Divide the levels of risk factors

[0110] Considering that there are risks of cargo transportation safety and train delay in international railway freight, and there are non-linear relationships and probability problems among the risk factors, the risk factors are systematically hierarchically divided as shown in Table 2 below;

[0111] Table 2 Risk System Stratification

[0112]

[0113]

[0114]

[0115] Step B2: Construct the structure of the Bayesian network for the quality risk of international railway transportation

[0116] According to the mutual relationship among the risk factors, establish the structure of the Bayesian network, and represent the relationship between all nodes in the Bayesian network with an adjacency matrix Q;

[0117]

[0118] Where q ij represents the relationship of the i-th risk factor to the j-th risk factor. When q ij =1, the i-th risk factor has an impact on the j-th risk factor. When q ij =0, there is no impact;

[0119] Draw the Bayesian network for the transportation quality risk according to the adjacency matrix Q, as Figure 3 shown.

[0120] Step B3: Determine the dataset of the transportation quality risk

[0121] Extract the stratified transportation quality risk factors from the collected transportation quality risk data to obtain the dataset D of the accidents related to the transportation quality risk and the train delays;

[0122]

[0123] Where, D i is the set of the states of the risk factors associated with the data in the i-th row;

[0124] D i =[u i1 u i2 ... u in

[0125] If the data in the i-th row is associated with the risk factor A1, then u i1 = the risk factor A1, otherwise u i1 = the risk factor A2; n represents the number of risk factors extracted from the data in the i-th row;​

[0126] For the dataset D, use u i1 to represent as:

[0127]

[0128] The method in the present invention considers that there is some unknown data in the dataset, and uses the EM learning algorithm to input the processed data into the Bayesian network structure to find the maximum likelihood Bayesian network; it is necessary to perform Bayesian network parameter learning for transportation quality risk;

[0129] EM learning is carried out through an iterative process. Starting from a candidate network, its log likelihood is reported, and then it is used to process the entire case set to find a better network; this process is continuously repeated until the log likelihood value no longer gets sufficient improvement or reaches the required number of iterations, then the iteration ends.

[0130] Step B4: Divide the hazard levels of transportation quality risks

[0131] Since there are corresponding probability relationships for all nodes in the Bayesian network, comprehensively considering the sensitivity level of the node relative to the final node and the network level where the node risk itself is located, calculate the sensitivity level S of all nodes in the Bayesian network i :

[0132] S i = 5Pe i

[0133] Perform weighted calculation on the sensitivity level S i and round up to obtain the risk hazard level E of the node i :

[0134] E i = [αS i + βC i

[0135] E i = [αS i + βC i

[0136] where, Pe i is the proportion of the sensitivity level of transportation quality risk F4 to risk factor i in the sensitivity level of transportation quality risk F4 to itself, E i is the risk hazard level, C i is the level of the Bayesian network where risk factor i is located; α is the influencing factor of transportation quality risk probability on the risk hazard level, β is the influencing factor of the consequences caused by the node risk itself on the risk hazard level, and the influence of α is weaker than that of β, so take α = 0.3 and β = 0.7; ​​

[0137] Step B5: Calculate the risk values of each risk factor

[0138] Take the product of the risk occurrence probability and the risk hazard level as the risk value, P i is the risk occurrence probability level of risk factor i, then the risk value R i The calculation formula is as follows:

[0139] R i = E i P i

[0140] Step B6: Determine the risk warning level of transportation quality

[0141] Utilize the risk occurrence probability level P i and the risk hazard level E i to establish a risk matrix diagram, and use the risk matrix diagram to divide the risk values into five warning levels: high, relatively high, medium, relatively low, and low; when the risk value R i ∈ {1, 2}, risk factor i is in a low-risk state; R i ∈ {3, 4, 6}, risk factor i is in a relatively low-risk state; R i ∈ {5, 8, 9}, risk factor i is in a medium-risk state; R i ∈ {10, 12, 15, 16}, risk factor i is in a relatively high-risk state; R i ∈ {20, 25}, risk factor i is in a high-risk state. Risk assessment is usually presented in the form of a risk matrix diagram, and the risk level, urgency, or actions to be taken for each cell are defined, and are represented by red, orange, or green according to the overall combination of risks and consequences; the risk matrix diagram is as Figure 8 shown.

[0142] Step B7: Complete the risk warning of transportation quality for international railway freight transportation based on the Bayesian network.

[0143] Step C: According to the needs of each participant in international railway freight transportation for the risk warning of applying blockchain technology to train transportation, on the basis of the Bayesian network of international railway freight transportation, construct and design the blockchain freight transportation quality risk warning framework and corresponding functional modules;

[0144] Select the Fabric framework and the PoA consensus algorithm, design the risk warning mechanism for international railway freight transportation applying blockchain, and design the functional modules.

[0145] Step C1: Design the blockchain framework, Figure 4 for the international railway freight transportation blockchain architecture

[0146] Network layer. Each participant in international railway freight transportation acts as a node to form a distributed network of the blockchain.

[0147] Consensus layer. Considering the consensus efficiency and the characteristics of each participant in international railway freight transportation belonging to an industry alliance, the PoA consensus mechanism is selected as the protocol for the consensus layer.

[0148] Application layer. Considering the safety requirements of international railway freight transportation, functional modules such as accident liability traceability, cargo tracking, transportation quality risk warning, risk closed-loop management, and automatic insurance compensation are designed. The application layer includes functions of accident liability traceability, risk closed-loop management, transportation quality risk warning, cargo tracking, and automatic compensation.

[0149] Step C2: Design functions

[0150] Design the function of warning against transportation quality risks in international railway freight transportation

[0151] The method of the present invention is based on the constructed Bayesian network model, introduces blockchain technology, and proposes a transportation quality risk warning process based on the blockchain. The specific process is as follows Figure 5 as shown. The main processes include a Bayesian network model, calculation of the risk values of each risk factor, and identification of high-risk factors.

[0152] It also includes the design of the function of tracking international railway cargo and the design of the function of automatic compensation for international railway cargo transportation insurance.

[0153] Design of the function of tracking international railway cargo: The design of this functional module is to upload the cargo location information and status information to the chain, and utilize the immutable feature of the blockchain to ensure the reliability of the cargo status information. The implementation of its process is mainly to send a request for retrieving information to the full nodes (nodes that synchronously store the complete ledger), and prove the existence of the retrieved data in the blockchain by obtaining Merkle proofs, so as to realize the reliable tracking of the cargo status. Figure 6 is the cargo tracking process based on the blockchain.

[0154] Design of the function of automatic compensation for international railway cargo transportation insurance: Figure 7 For the specific process, during the international railway cargo transportation process, if insurance compensation is required, the historical records of this batch of cargo can be traced. In the case of a clear chain of responsibility determination, the responsibility is automatically determined through a smart contract for insurance compensation; in the case of unclear responsibility, the status information of each section in the blockchain is retrieved and collected, submitted to the corresponding negotiation agency for responsibility division negotiation, and then compensation is made. The specific process is as Figure 7 shown.

Claims

1. A method for warning of quality risks in international railway freight transportation based on blockchain, characterized in that, It includes the following steps: Step A: Identify the risk sources of international railway freight transportation to obtain the risk factors of international railway freight transportation; Determine the quality risks of international railway freight transportation. From the participants and links of international railway freight transportation, use the HAZOP method, network planning diagram, and interpretive structural model to identify the risk factors of the key paths and key links in the whole process of international railway freight transportation; The quality risks of freight transportation include cargo safety risk factors and train delay risk factors; Use the HAZOP method to identify the port operation risks and in-transit transportation risks of cargo safety risk factors respectively; Use the network planning diagram and interpretive structural model (ISM) to identify the port operation risks of train delay risk factors, and use the network planning diagram, interpretive structural model (ISM), and HAZOP method to identify the in-transit transportation risks of train delay risk factors; Step B: Classify the risk factors hierarchically. According to the relationships between the risk factors, construct a Bayesian network for the transportation quality risks of international railways, establish an early warning model for the combined transportation risks of international railway freight transportation based on the Bayesian network, and evaluate the level of transportation quality risks; Step C: According to the needs of each participant in international railway freight transportation for applying blockchain technology to train transportation risk early warning, on the basis of the Bayesian network of international railway transportation quality risks, construct and design the blockchain cargo transportation quality risk early warning framework and functional modules, and establish a blockchain-based international railway cargo transportation quality risk early warning platform; Based on the Fabric framework and the PoA consensus algorithm, design an international railway cargo transportation quality risk early warning platform applying blockchain, and design the functional modules; The functional modules include: international railway cargo transportation quality risk early warning function, international railway cargo tracking function, and automatic compensation function for international railway cargo transportation insurance.

2. The method for warning of quality risks in international railway freight transportation based on blockchain according to claim 1, wherein The risk factors are classified according to the participants. For the consignee, there is container damage; for the carrier, there are inadequate equipment management, insufficient personnel management, imperfect rules and regulations, imperfect supervision facilities and equipment, abnormal car body cycle management, inflexible route selection, inadequate vehicle management, imperfect station operation supervision system, imperfect station security supervision system, imperfect railway risk prevention supervision, imperfect supervision facilities and equipment, imperfect production safety supervision, inadequate station supervision, insufficient transshipment capacity, failure to increase epidemic prevention and control measures, low operation efficiency, lack of safety awareness, unstable political situation, container damage, improper human operation, railway suspension or construction, and damaged freight car parts; for the customs, there is overtime operation time; for others, there are natural disasters.

3. The method for warning of quality risks in international railway freight transportation based on blockchain according to claim 1, wherein Step B also includes the following steps: Step B1: Classify the risk factors hierarchically Classify the risk factors hierarchically from five aspects: organizational management, risk supervision, risk incentives, direct causes, and transportation quality risks; Step B2: Construct the structure of the Bayesian network for international railway transportation quality risks According to the mutual relationships between the risk factors, establish the structure of the Bayesian network, and represent the relationships between all nodes in the Bayesian network with an adjacency matrix Q; where q ij represents the relationship of the i-th risk factor to the j-th risk factor. When q ij = 1, the i-th risk factor has an impact on the j-th risk factor. When q ij = 0, there is no impact; Establish a Bayesian network for transportation quality risk based on the adjacency matrix Q; Step B3: Determine the dataset of transportation quality risk Stratify and extract the transportation quality risk factors from the collected transportation quality risk data to obtain the dataset D of relevant accidents and train delays of transportation quality risk; where D i is the set of risk factor states associated with the data of the i-th row; D i = [u i1 u i2 ... u in ​ If the data in the i-th row is associated with risk factor A1, then u i1 = Risk factor A1, otherwise u i1 = Risk factor A2; n represents the number of risk factors extracted from the data in the i-th row; For the dataset D, use u i1 to denote as: Step B4: Classify the hazard levels of transportation quality risk Calculate the sensitivity level S of all nodes in the Bayesian network i : S i = 5Pe i Perform weighted calculation on the sensitivity level S i and round up to obtain the risk hazard level E of the node i : E i = [αS i + βC i ​ Among them, Pe i is the proportion of the sensitivity level of the transportation quality risk F4 to the risk factor i to the sensitivity level of the transportation quality risk F4 to itself, E i is the risk hazard level, C i is the level of the Bayesian network where the risk factor i is located; α is the influencing factor of the transportation quality risk probability on the risk hazard level, β is the influencing factor of the consequence caused by the node risk itself on the risk hazard level, and the influence of α is weaker than that of β, so α = 0.3 and β = 0.7 are taken; Step B5: Calculate the risk values of each risk factor Take the product of the risk occurrence probability and the risk hazard level as the risk value, P i is the risk occurrence probability level of risk factor i, then the risk value R i The calculation formula is as follows: R i = E i P i Step B6: Determine the warning levels of transportation quality risk Using the risk occurrence probability level P i and the risk hazard level E i A risk matrix diagram is established, and the risk matrix diagram is used to divide the risk value into five warning levels, including high, relatively high, medium, relatively low, and low; when the risk value R i ∈{1, 2}, the risk factor i is in a low-risk state; R i ∈{3, 4, 6}, the risk factor i is in a relatively low-risk state; R i ∈{5, 8, 9}, the risk factor i is in a medium-risk state; R i ∈{10, 12, 15, 16}, the risk factor i is in a relatively high-risk state; R i ∈{20, 25}, the risk factor i is in a high-risk state; Step B7: Complete the warning of transportation quality risk for international railway freight transportation based on the Bayesian network.

4. The method for warning of quality risks in international railway freight transportation based on blockchain according to claim 3, wherein The organizational management includes container management, personnel management, rules and regulations, facility and equipment management, car body management, line management, and vehicle management; container management, personnel management, rules and regulations, facility and equipment management, car body management, line management, and vehicle management correspond to nodes O1, O2, O3, O4, O5, O6, and O7 respectively; the O1 node corresponds to A1 proper equipment management and A2 improper equipment management; the O2 node corresponds to A1 normal personnel management and A2 insufficient personnel management; the O3 node corresponds to A1 perfect rules and regulations and A2 imperfect rules and regulations; the O4 node corresponds to A1 perfect supervision facilities and equipment and A2 imperfect supervision facilities and equipment; the O5 node corresponds to A1 normal car body circulation management and A2 abnormal car body circulation management; the O6 node corresponds to A1 flexible line selection and A2 inflexible line selection; the O7 node corresponds to A1 proper vehicle management and A2 improper vehicle management; The risk supervision includes operation supervision, security supervision, risk prevention supervision, supervision facilities, and production safety supervision; operation supervision, security supervision, risk prevention supervision, supervision facilities, and production safety supervision correspond to nodes M1, M2, M3, M4, and M5 respectively; the M1 node corresponds to A1 perfect station operation supervision system and A2 imperfect station operation supervision system; the M2 node corresponds to A1 perfect station security supervision system and A2 imperfect station security supervision system; the M3 node corresponds to A1 imperfect railway risk prevention supervision and A2 perfect railway risk prevention supervision; the M4 node corresponds to A1 perfect supervision facilities and equipment and A2 imperfect supervision facilities and equipment; the M5 node corresponds to A1 perfect production safety supervision and A2 imperfect production safety supervision; Risk incentives include natural environment, operation environment, equipment capacity, epidemic prevention and control, operation efficiency, safety awareness, and political factors; natural environment, operation environment, equipment capacity, epidemic prevention and control, operation efficiency, safety awareness, and political factors correspond to nodes S1, S2, S3, S4, S5, S6, and S7 respectively; the S1 node corresponds to A1 normal and A2 occurrence of natural disasters; the S2 node corresponds to A1 proper station supervision and A2 improper station supervision; the S3 node corresponds to A1 insufficient transshipment capacity and A2 insufficient transshipment capacity; the S4 node corresponds to A1 no additional epidemic prevention measures and A2 additional epidemic prevention measures; the S5 node corresponds to A1 normal operation efficiency and A2 low operation efficiency; the S6 node corresponds to A1 proper safety awareness and A2 improper safety awareness; the S7 node corresponds to A1 stable political situation and A2 unstable political situation; The direct causes include container status, human operation, railway suspension or construction, operation time, and wagon component status; the container status, human operation, railway suspension or construction, operation time, and wagon component status correspond to nodes Z1, Z2, Z3, Z4, and Z5 respectively; node Z1 corresponds to A1 normal container and A2 damaged container; node Z2 corresponds to A1 no improper human operation and A2 improper human operation; node Z3 corresponds to A1 no railway suspension or construction and A2 railway suspension or construction; node Z4 corresponds to A1 normal operation time and A2 overtime operation time; node Z5 corresponds to A1 normal wagon components and A2 damaged wagon components; The transportation quality risks include accidents, unexpected events other than accidents, train delay, and transportation quality risks; accidents, unexpected events other than accidents, train delay, and transportation quality risks correspond to nodes F1, F2, F3, and F4 respectively.

Images