Method for Constructing Dependent Network Model for Industrial Cyber-Physical System

Abstract

The present invention discloses a method for constructing a dependency network model for an industrial cyber-physical system, belonging to the field of industrial cyber-physical system security research. The method includes: respectively determining a structural dependency index and a service dependency index according to the structure and service characteristics of the industrial cyber-physical system; respectively calculating a structural dependency relationship evaluation value and a service dependency relationship evaluation value according to the structural dependency index and the service dependency index, and weighting the two to obtain the dependency relationship between the information node and the physical node. By analyzing the system topology structure and service characteristics, the present invention clarifies the interaction relationship between the information domain and the physical domain of the system, and comprehensively reveals the complex interaction relationship between multiple domains of the system from the perspectives of structure and service characteristics; the analytic hierarchy process is used to quantify the dependency relationship, solving the problem that it is difficult to quantify due to too many influencing factors in the process of evaluating the dependency relationship, realizing the quantitative evaluation of the dependency relationship, and being able to more comprehensively and truly reflect the complex interaction relationship between the information domain and the physical domain.

Description

Technical Field

[0001] The present invention belongs to the technical field of industrial cyber-physical system security research, and more specifically, relates to a dependency network modeling method for industrial cyber-physical systems. Background Art

[0002] Industrial cyber-physical systems are an important foundation for intelligent manufacturing and "Industry 4.0", and are widely used in core industrial fields such as intelligent manufacturing, smart grids, water conservancy, and energy. Compared with traditional industrial control systems, although the deep coupling of the industrial information domain and the physical domain improves the effectiveness and real-time nature of information space in managing physical processes and enables scientific decision-making in production scheduling, it also blurs the system security boundary, exposes more attack surfaces, and is extremely prone to security threat events. To evaluate the vulnerability of industrial cyber-physical systems, it is necessary to combine the inter-domain dependency relationship to evaluate the cross-domain impact of system vulnerability. Currently, existing dependency network modeling methods usually set the dependency relationship between the information domain and the physical domain as an assumed value, or only quantify the dependency relationship between nodes from a single perspective. However, there is a highly coupled relationship between the industrial information domain and the physical domain, and the system exhibits characteristics such as a flat structure and integrated interaction. Therefore, the existing dependency network modeling methods cannot comprehensively reflect the interaction characteristics between multiple domains of the system, and it is necessary to describe the complex interaction characteristics of both topological interaction and business dependency between the information-physical domains in combination with the system operation characteristics to further improve the accuracy of the dependency network modeling method. Summary of the Invention

[0003] In view of the above defects or improvement requirements of the prior art, the present invention provides a dependency network modeling method for industrial cyber-physical systems, and the purpose of this invention is to provide a more comprehensive and realistic dependency relationship evaluation method for industrial cyber-physical systems.

[0004] To achieve the above object, the present invention provides a method for constructing a dependency network model for industrial cyber-physical systems, including:

[0005] S1. Determine the structural dependency index and the business dependency index respectively according to the structure and business characteristics of the industrial cyber-physical system;

[0006] S2. Calculate the structural dependency relationship evaluation value and the business dependency relationship evaluation value respectively according to the structural dependency index and the business dependency index, and weight the two to obtain the dependency relationship between the information node and the physical node.

[0007] Further, the calculation formula for the structural dependency relationship evaluation value is:

[0008]

[0009] Among them, RS(i,j) represents the structural dependence metric value between information node i and physical node j; RS D (i,j) and RS B (i,j) are the structural dependence metric values based on "degree-degree" and "betweenness-betweenness" respectively; RS′ D (i,j) and RS′ B (i,j) are the results after normalizing RS D (i,j) and RS B (i,j); α is the index weight; respectively represent the degrees of information node i and physical node j; represents the probability that an information node with degree is interdependent with a physical node with degree ; respectively represent the probabilities that the degrees of nodes in the interdependent network are ; are the betweennesses of information node i and physical node j respectively, represents the probability that an information node with betweenness is interdependent with a physical node with betweenness ; respectively represent the probabilities that the betweenness of the information node is and the betweenness of the physical node is .

[0010] Furthermore, the business dependence index specifically includes the amount of status information data, the amount of control command data, and the maximum value of the communication traffic per unit time between information nodes and physical nodes in an industrial cyber-physical system during a production cycle.

[0011] Furthermore, the amount of control command data δ(i,j) between information node i and physical node j in an industrial cyber-physical system during a production cycle, and the specific formula is:

[0012]

[0013] Among them, k r (j) represents the data size of the control instruction r required for the normal operation of physical node j, and E r (j) represents the number of times the control instruction r is issued during a production cycle; k b (j) represents the data volume size of the fault maintenance instruction for node j, and E b (j) represents the average number of faults of node j during a production cycle.

[0014] Furthermore, the amount of status information data γ(j,i) between physical node j and information node i in an industrial cyber-physical system during a production cycle, and the specific formula is:

[0015]

[0016] Among them, h represents the type of status information periodically uploaded by the physical node, and t r1 ~t r2 represents the survival time of the r-th type of data within a production cycle, and f r represents the acquisition frequency of the r-th type of data, and k r represents the size of the r-th type of data.

[0017] Furthermore, the dependency relationship F C-P (i, j) between the information node i and the physical node j is calculated by the formula:

[0018]

[0019] Among them, ω S 、ω L are the weights of the structural dependency relationship evaluation value and the business dependency relationship evaluation value respectively, and ω S +ω L =1; RS′(i, j) is the result after normalizing the structural dependency relationship evaluation value; δ′(i, j), δ‘ m (i, j) are the results after normalizing the control command data volume and the maximum control command data volume per unit time between the information node i and the physical node j within a production cycle; are the weights of the two evaluation indicators of the business dependency relationship respectively, and

[0020] Furthermore, the dependency relationship F P-C (j, i) between the physical node j and the information node i is calculated by the formula:

[0021]

[0022] Among them, ω S 、ω L are the weights of the structural dependency relationship evaluation value and the business dependency relationship evaluation value respectively, and ω S +ω L =1; RS′(j, i) is the result after normalizing the structural dependency relationship evaluation value; γ′(j, i), γ’ m (j, i) are the results after normalizing the status information data volume and the maximum status information data volume per unit time between the physical node j and the information node i within a production cycle; are the weights of the two evaluation indicators of the business dependency relationship respectively, and

[0023] Generally speaking, compared with the prior art, the above technical solutions conceived by the present invention can achieve the following beneficial effects.

[0024] The present invention provides a method for constructing a dependency network model for an industrial cyber-physical system. By analyzing the topological structure and service characteristics of the system, the interaction relationship between the information domain and the physical domain of the system is clarified, and the complex interaction relationship between multiple domains of the system is comprehensively revealed from the perspectives of structure and service characteristics; and the analytic hierarchy process is used to quantify the dependency relationship, solving the problem that it is difficult to quantify due to too many influencing factors in the process of evaluating the dependency relationship, and realizing the quantitative evaluation of the dependency relationship. Brief Description of the Drawings

[0025] Figure 1 is a schematic diagram of the overall process for evaluating the dependency relationship of an industrial cyber-physical system provided by an embodiment of the present invention;

[0026] Figure 2 is a schematic diagram of a workpiece production intelligent manufacturing system model provided by an embodiment;

[0027] Figure 3 is Figure 2 a schematic diagram of the production workshop business process of the workpiece production intelligent manufacturing system shown;

[0028] Figure 4 is the interaction characteristic of the workpiece production intelligent manufacturing system provided by an embodiment of the present invention;

[0029] Figure 5 is a schematic diagram of the dependency network model of the workpiece production intelligent manufacturing system provided by an embodiment of the present invention;

[0030] Figure 6 is a schematic diagram of the hierarchical structure model for evaluating the dependency relationship of an industrial cyber-physical system provided by an embodiment of the present invention;

[0031] Figure 7 is a schematic diagram for verifying the evaluation result of the dependency relationship provided by an embodiment of the present invention. Detailed Embodiments

[0032] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0033] A method for constructing a dependency network model for an industrial cyber-physical system provided by the present invention includes:

[0034] S1. Determine the structure - dependent index and the service - dependent index respectively according to the structure and service characteristics of the cyber - physical system;

[0035] S2. Calculate the structure - dependent relationship evaluation value and the service - dependent relationship evaluation value respectively according to the structure - dependent index and the service - dependent index, and weight the two to obtain the dependence relationship between the physical node and the information node.

[0036] The specific implementation process is as follows:

[0037] Analyze the types and functions of each node according to the business process of the industrial cyber - physical system, and identify the topological interaction mode of the system operation; determine the composition of the control flow and data flow between the information node and the physical node in combination with the operation process of the ICPS, and analyze the dynamic interaction relationship between the data flow and the control flow in the system.

[0038] Combined with the structure and topological interaction characteristics of the system, divide the ICPS into three parts: the information layer (C), the physical layer (P), and the coupling layer (C - P); based on the complex network theory, construct topological structure models for the industrial information layer and the physical layer respectively, and represent the system as a set of nodes and connection edges, that is, G=(V, E); where V = {1, 2,... S} represents all nodes in the complex network; E = {e ij , 1≤i, j≤S} represents all connection edges in the network. Represent the topological connection relationship of the subsystem complex network as: where a ij represents the connection relationship between node i and node j. On this basis, obtain the topological model G C / G P and the adjacency matrix A C / A P ; analyze the coupling mode between the information node and the physical node based on the topological interaction relationship of the industrial cyber - physical system, so as to determine the dependent edges in the network, construct the topological graph G C-P of the coupling layer network and determine the neighbor matrix of the coupling layer On the basis of constructing the topological models of the information layer, the physical layer, and the coupling layer, comprehensively determine the topological structure and the adjacency matrix of the industrial cyber - physical system, and its adjacency matrix is specifically described as:

[0039]

[0040] The present invention uses the analytic hierarchy process to quantify the dependence relationship, specifically including:

[0041] Construction of the hierarchical model for the dependency relationship evaluation metrics. The dependency relationship evaluation problem is decomposed into topology-based dependency relationship evaluation and business-based dependency relationship evaluation, the influencing metrics of the structural dependency relationship and the business dependency relationship are determined, the interconnections between various factors are analyzed, and a hierarchical structure model for the dependency relationship evaluation is established.

[0042] Calculate the degree and betweenness of each node in the dependency network, and evaluate the structural dependency relationship between information nodes and physical nodes. The specific formula is:

[0043]

[0044] Among them, RS(i,j) represents the structural dependency metric value between information node i and physical node j; RS D (i,j), RS B (i,j) are the structural dependency metric values based on "degree-degree" and "betweenness-betweenness" respectively; RS′ D (i,j), RS′ B (i,j) are the results after normalizing RS D (i,j) and RS B (i,j) respectively; α is the index weight, which is calculated by the method of comprehensive subjective and objective weight assignment. This method can be used to balance the subjective and objective weight values and make the weight calculation result more reasonable; respectively represent the degrees of information node i and physical node j; represents the probability that an information node with degree and a physical node with degree are interdependent; respectively represent the probabilities that the degrees of the nodes in the dependency network are ; are the betweennesses of information node i and physical node j respectively, represents the probability that an information node with betweenness and a physical node with betweenness are interdependent; respectively represent the probabilities that the betweenness of the information node is and the betweenness of the physical node is ;

[0045] Existing research generally quantifies the structural dependence relationship by combining the degree index of nodes. However, the present invention simultaneously considers the influence of node degree and betweenness on the structural dependence relationship. According to the dependence network theory, the more important a node in the network is, the greater the impact of the failure of this node on its dependent nodes, indicating that the dependence relationship between these two nodes is stronger. The degree index can only reflect the importance of a node in the local network, while the betweenness index can reflect the importance of a node in the global network. Compared with the prior art, the present invention can more comprehensively reflect the structural dependence degree between information nodes and physical nodes.

[0046] Calculate the amount of state information data γ(i, j) of physical node j and information node i within a production cycle in combination with the communication service type of the system, which can be specifically expressed as:

[0047]

[0048] Among them, h represents the types of state information periodically uploaded by physical nodes; t r1 ~t r2 represents the survival time of the r-th type of data within a production cycle; f r represents the acquisition frequency of the r-th type of data; k r represents the size of the r-th type of data.

[0049] Analyze and calculate the amount of control command data δ(i, j) between information node i and physical node j within a production cycle in combination with the communication service type of the system, which can be specifically expressed as:

[0050]

[0051] Among them, k r (j) represents the data size of the control instruction r required for the normal operation in physical node j; E r (j) represents the number of times the control instruction r is issued within a production cycle; k b (j) represents the data volume size of the fault maintenance instruction in node j; E b (j) represents the average number of faults of node j within a production cycle, and the specific calculation formula is as follows:

[0052]

[0053]

[0054] Among them, λ(t) is the device failure probability function obtained by combining the Weibull distribution; T is the production cycle of the industrial cyber-physical system; the parameters β and η are obtained by solving the parameters of the linear regression equation according to the least squares method.

[0055] The method based on the calculation of subjective and objective weights assigns weights to each index. First, the subjective weights of each index are determined based on the analytic hierarchy process, specifically expressed as ω z ={ω z1 ,ω z2 ...ω zm}; Then, the objective weights of each index are calculated using the entropy method, specifically expressed as ω k ={ω k1 ,ω k2 ...ω km}; Finally, based on the principle of minimum information discrimination, the subjective and objective weight values are balanced, and the comprehensive weight ω j of each index is:

[0056]

[0057] where ω j is the comprehensive weight of the dependency relationship index j; ω zj is the subjective weight of index j; ω kj is the objective weight of index j; and j = 1, 2,..., m; m is the total number of indexes.

[0058] Based on the quantification of the dependency relationship influence index and the calculation of the index weights, the dependency relationship between the information node and the physical node is comprehensively evaluated. The dependency relationship F C-P (i, j) between the information node i and the physical node j is used to characterize the degree of dependence of the physical node j on the information node i, and the specific calculation formula is:

[0059]

[0060] where ω S , ω L are the weights of the structural dependency relationship evaluation value and the business dependency relationship evaluation value respectively, and ω S +ω L = 1; RS′(i, j) is the result of normalizing the structural dependency relationship evaluation value; δ′(i, j), δ‘ m (i, j) are the results of normalizing the control command data volume and the maximum control command data volume per unit time between the information node i and the physical node j within a production cycle; are the weights of the two evaluation indexes of the business dependency relationship respectively, and

[0061] The dependency relationship F P-C (j, i) between the physical node j and the information node i characterizes the degree of dependence of the information node i on the physical node j, and the specific calculation formula is:

[0062]

[0063] Among them, ω S and ω L are the weights of the structure dependence relationship evaluation value and the business dependence relationship evaluation value respectively, and ω S + ω L = 1; RS′(j,i) is the result after normalizing the structure dependence relationship evaluation value; γ′(j,i) and γ’ m (j,i) are the results after normalizing the amount of state information data and the maximum amount of state information data per unit time between the physical node j and the information node i within a production cycle, respectively; are the weights of the two evaluation indicators of the business dependence relationship respectively, and

[0064] The overall process schematic diagram of the construction of the dependence network model for the industrial cyber-physical system provided by the embodiments of the present invention is as shown in Figure 1 and includes the construction of the system topology model and the evaluation of the dependence relationship. The following combines with Figure 2 the workpiece production intelligent manufacturing system shown in the figure to specifically elaborate on the application of the construction of this dependence network model provided by the embodiment.

[0065] Figure 2 The basic structure diagram of the workpiece production intelligent manufacturing system shown in the figure is based on an environment including a physical system, a communication network, and a control center. Among them, the physical system is composed of terminal physical devices and is mainly responsible for the processing, assembly, transportation, and storage of workpieces; the communication network is supported by an industrial fieldbus and is responsible for the information interaction between the physical system and the control center; the information domain is composed of an upper-layer information system and is responsible for monitoring the workshop status and remotely guiding the control to ensure the safe and orderly operation of the workshop. The construction of the dependence network model for the industrial cyber-physical system will first clarify the system structure and business process, and analyze the information-physical interaction characteristics of the system; then, according to the system structure characteristics, build a system topology model based on the dependence network theory; finally, combine the system structure characteristics and business characteristics, and use the analytic hierarchy process to comprehensively evaluate the dependence relationship between the information node and the physical node. Based on the construction of the topology model and the evaluation of the dependence relationship, a complete dependence network model of the system is established. Specifically as follows:

[0066] Step 1: Combine the structure characteristics and business process of the workpiece production intelligent manufacturing system to analyze the information-physical interaction characteristics of the system;

[0067] Step 1.1: Analyze the system structure characteristics and identify the key interaction topologies in the system;

[0068] In this embodiment, there is no direct information interaction between the control devices in the information layer and the physical layer devices. The control nodes are autonomous nodes decoupled from the physical layer, and only the communication devices have a direct interaction relationship with the physical devices. A communication node can communicate with multiple physical nodes simultaneously to ensure the safe operation of multiple physical devices. At the same time, a physical node only needs to accept the instructions of one communication node to ensure the stable operation of the physical node.

[0069] Step 1.2: Determine the composition of the control flow and data flow between the information nodes and the physical nodes in combination with the business process of the system, and analyze the dynamic interaction relationship between the data flow and the control flow in the system.

[0070] In this embodiment, the business process of the system is as Figure 3 shown. Through analysis, it can be seen that the interaction relationship of the system is as Figure 4 shown.

[0071] Step 2: Abstract the system topology structure in the form of nodes - edges in combination with the structural characteristics and topological interaction characteristics of the intelligent manufacturing system for workpiece production, and realize the topological modeling of the system.

[0072] In this embodiment, the dependency network topology model of the system is as Figure 5 shown.

[0073] Step 3: Evaluate the dependency relationship between the nodes in the information domain and the physical domain based on the topology and business characteristics, calculate the evaluation indicators of the dependency relationship, and quantify the dependency relationship using the analytic hierarchy process. Specifically, it includes the following sub - steps:

[0074] Step 3.1: Construct the hierarchical model of the dependency relationship evaluation indicators.

[0075] In this embodiment, the hierarchical structure model of the dependency relationship evaluation is as Figure 6 shown.

[0076] Step 3.2: On the basis of constructing the dependency network topology model, evaluate the structural dependency relationship between the information nodes and the physical nodes.

[0077] Step 3.3: Determine the business data - related information between the nodes in combination with the business characteristics of the system, and establish the data information tables between different nodes.

[0078] In this embodiment, the business data - related information uploaded by the physical nodes is shown in Table 1.

[0079] Table 1 Business data - related information between physical - information nodes

[0080]

[0081] Step 3.4: Calculate the total communication service volume and the maximum communication service volume per unit time by combining the service data-related information between physical-information nodes, and use the subjective and objective comprehensive weighting method to assign weights to each index, so as to comprehensively evaluate the dependency relationship F based on topology and service C-P (i,j), F P-C (i,j).

[0082] To verify the accuracy of the evaluation results of the dependency relationship based on topology and service in the present invention, the present invention takes the calculation result F of the information-physical node dependency relationship C-P as an example, and respectively remove each dependency edge in descending order of the evaluation values obtained by the degree-degree dependency, service dependency, and the evaluation method based on the dependency relationship between topology and service, and count the change of the service volume of the workpiece production intelligent manufacturing system in the case of the functional failure of the dependency edge. The results are as Figure 7 shown. The total communication service volume carried by the nodes connected by the dependency edge is related to the node's own function and the position of the node in the dependency network; the greater the impact of the failure of the dependency edge on the system service volume, the stronger the mutual dependency relationship between the information-physical nodes. From Figure 7 it can be seen that the change trends of the system service volume under the three evaluation methods are the same, indicating that the evaluation method based on the dependency relationship between topology and service proposed by the present invention has a certain degree of rationality. Since the evaluation of the degree-degree dependency relationship only considers the topological interaction relationship of the system, but a larger node degree does not mean that the communication service volume it carries is also larger, and this method does not consider the service interaction relationship between the dependent nodes. Therefore, when removing each dependency edge according to the evaluation result of the degree-degree dependency relationship, the system service volume decreases at the slowest speed, indicating that this method is not accurate enough in evaluating the dependency relationship. When removing each dependency edge according to the evaluation results of the service dependency relationship and the dependency relationship between topology and service respectively, the coincidence degree of the change of the system service volume is relatively high, but the system service volume decreases faster under the evaluation of the dependency relationship between topology and service. It can be seen from this that the dependency relationship evaluation method in the present invention can better reflect the dependency relationship between information-physical nodes and has certain advantages in the evaluation of the dependency relationship.

[0083] Those skilled in the art can easily understand that the above description is only a preferred example of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for constructing a dependent network model for industrial cyber-physical systems, characterized in that Including: S1. Determine the structure-dependent index and business-dependent index respectively according to the structure and business characteristics of the industrial cyber-physical system; the business-dependent index specifically includes the amount of status information data, the amount of control command data, and the maximum value of communication traffic per unit time between information nodes and physical nodes in an industrial cyber-physical system within a production cycle. S2. Calculate the structure-dependent relationship evaluation value and the business-dependent relationship evaluation value respectively according to the structure-dependent index and the business-dependent index, and weight the two to obtain the dependence relationship between the information node and the physical node. The calculation formula for the structure-dependent relationship evaluation value is: Among them, RS(i,j) represents the structural dependence metric value between information node i and physical node j; RS D (i,j) and RS B (i,j) are the structural dependence metric values based on "degree-degree" and "betweenness-betweenness" respectively; RS' D (i,j) and RS' B (i,j) are the results after normalizing RS D (i,j) and RS B (i,j) respectively; α is the index weight; respectively represent the degrees of information node i and physical node j; represents the probability that an information node with degree is interdependent with a physical node with degree ; respectively represent the probabilities that the degrees of nodes in the interdependent network are ; are the betweennesses of information node i and physical node j respectively, represents the probability that an information node with betweenness is interdependent with a physical node with betweenness ; respectively represent the probabilities that the betweenness of the information node is and the betweenness of the physical node is ; Dependency relationship F between information node i and physical node j C-P (i, j) The calculation formula is as follows: Among them, ω S , ω L are the weights of the structure dependence relationship evaluation value and the business dependence relationship evaluation value respectively, and ω S +ω L = 1; RS'(i,j) is the result after normalizing the structure dependence relationship evaluation value; δ'(i,j), δ‘ m (i,j) are the results after normalizing the control command data volume of information node i and physical node j within a production cycle and the maximum control command data volume per unit time respectively; are the weights of the two evaluation indexes of the business dependence relationship respectively, and Dependency relationship F between physical node j and information node i P-C (j, i) The calculation formula is as follows: Among them, ω S and ω L are the weights of the structure dependency evaluation value and the business dependency evaluation value respectively, and ω S +ω L = 1; RS'(j,i) is the result after normalizing the structure dependency evaluation value; γ'(j,i) and γ’ m (j,i) are the results after normalizing the amount of state information data and the maximum amount of state information data per unit time between physical node j and information node i within a production cycle respectively; are the weights of the two evaluation indicators of business dependency respectively, and 2. The method for constructing a dependency network model for industrial cyber-physical systems according to claim 1, wherein The amount of control command data δ(i, j) between the information node i and the physical node j in an industrial cyber-physical system within a production cycle, and the specific formula is: where k r (j) represents the data size of the control instruction r required for the normal operation in the physical node j, E r (j) represents the number of times the control instruction r is issued in one production cycle; k b (j) represents the data volume size of the fault maintenance instruction in the node j, E b (j) represents the average number of faults in the node j in one production cycle.

3. A method for constructing a dependency network model for an industrial cyber-physical system according to claim 1, characterized in that, The amount of status information data γ(j, i) between the physical node j and the information node i in an industrial cyber-physical system within a production cycle, and the specific formula is: Among them, h represents the types of status information periodically uploaded by physical nodes, and t r1 ~t r2 represents the survival time of the r-th type of data within a production cycle, and f r represents the acquisition frequency of the r-th type of data, and k r represents the size of the r-th type of data.

Images