Multi-level fault diagnosis method and system based on dynamic sequential network, storage medium and equipment

Abstract

The invention discloses a multi-level fault diagnosis method and system based on a dynamic sequential network, a storage medium and equipment, and belongs to the field of fault diagnos.The method comprises the steps that S1, target system information is collected according to the given frequency, and the dynamic sequential network is established; s2, multi-layer structure parameters of the dynamic sequential network topology structure are calculated, and a multi-layer structure parameter sequence is established; s3, based on a control chart theory, performing multi-level preliminary fault diagnosis by using the multi-level structure parameter sequence to obtain a plurality of preliminary fault diagnosis results; and S4, fusing the plurality of preliminary fault diagnosis results by adopting a dynamic integration method. According to the method, dynamic behaviors of a real system are considered, a network model capable of reflecting a system operation topological structure is established based on a dynamic sequential network, preliminary fault diagnosis is realized from multiple levels, comprehensive fault diagnosis is realized based on a dynamic selection method, simple and feasible online fault diagnosis can be realized, and the fault diagnosis efficiency is improved. And health monitoring requirements of a complex system are met.

Description

Technical Field

[0001] The present invention relates to the field of fault diagnosis, and in particular to a multi-level fault diagnosis method, system, storage medium and device based on a dynamic sequential network. Background Art

[0002] Fault diagnosis refers to the process of determining system faults by detecting, identifying and judging abnormal conditions. It is of great significance for ensuring the normal operation of the system, reducing maintenance costs and optimizing system design. Fault diagnosis can be further divided into different tasks such as fault detection, fault classification, fault location and fault recovery. Among them, fault detection aims to determine whether the system has a fault or is about to fail. It can timely discover and solve potential problems and prevent system failures or crashes, thereby improving the reliability and stability of the system. This is especially important for key fields such as aerospace, because the failure of these systems may lead to catastrophic consequences. However, with the continuous development of technology, the scale of complex systems such as aerospace and power grids is getting larger and larger, and the complexity of the systems is getting higher and higher, making the systems difficult to understand and maintain, and also bringing huge challenges to fault diagnosis.

[0003] Complex network theory ignores the minute details of complex systems and highlights the universal laws. It is considered to be an effective means of describing complex systems. In this context, complex network theory has gradually been applied to the field of fault diagnosis. Among them, the fault diagnosis method based on community structure analysis is based on the complex network modeling of the system, combined with community structure analysis, network feature extraction, etc. to achieve fault diagnosis. However, its network model is static and cannot reflect the dynamic behavior of real complex systems. Specifically, during the operation of the system, its internal connection and other states are often constantly changing. The network model established on the basis of static structural analysis of the system cannot actually reflect the real-time changes in the system state, that is, it cannot reflect the dynamic behavior of complex systems in the real world. At the same time, during the dynamic operation of the system, its system topology will also change accordingly. The change in the topology reflects the operating state of the system. If the system topology can be monitored, system fault diagnosis can be expected. Summary of the invention

[0004] The purpose of the present invention is to overcome the problems existing in the prior art and provide a multi-level fault diagnosis solution based on a dynamic timing network. The solution establishes a dynamic timing network of the target system and calculates multi-level structural parameters based on the network topology. At the same time, the control chart is used to analyze the multi-level parameter sequence, judge the system status, and integrate the preliminary diagnosis results to achieve multi-level fault diagnosis.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] In a first aspect, a multi-level fault diagnosis method based on a dynamic sequential network is provided, comprising the following steps:

[0007] S1. Collect target system information according to given frequency and establish dynamic timing network;

[0008] S2, calculating the multi-level structural parameters of the dynamic time series network topology structure and establishing a multi-level structural parameter sequence;

[0009] S3. Based on the control chart theory, a multi-level structural parameter sequence is used to perform multi-level preliminary fault diagnosis and obtain multiple preliminary fault diagnosis results;

[0010] S4. Adopting a dynamic integration method to fuse the multiple preliminary fault diagnosis results.

[0011] In some embodiments, step S1 specifically includes:

[0012] At a given sampling frequency fre, each sampling time t i Get a set of system architecture data and convert t i The set of system entities obtained at the moment is recorded as n is the number of entities; the set of dependencies between entities is recorded as m is the number of dependencies between entities;

[0013] Based on t i System architecture data at the moment and Build i Complex network model of target system at each moment

[0014]

[0015] A series of network models established at all sampling moments constitute a dynamic timing network Where l represents the number of sampling times.

[0016] In some embodiments, step S2 specifically includes:

[0017] The macro-structure parameters of the network are calculated to establish a macro-level structure parameter sequence; the meso-structure parameters of the network are calculated to establish a meso-level structure parameter sequence; the micro-structure parameters of the network are calculated to establish a micro-level structure parameter sequence; among them, the macro-structure parameters refer to the parameters that characterize the structure of the entire network, the meso-structure parameters refer to the parameters that characterize the structure of some node sets in the network, and the micro-structure parameters refer to the parameters that characterize the local micro-structure in the network.

[0018] Preferably, the macroscopic structural parameters include average degree, average shortest path length, average clustering coefficient, network diameter, number of network subclusters and ratio of the maximum connected subclusters; the mesoscopic structural parameters include number of communities, modularity and community overlap; the microscopic structural parameters include number of 3-node motif types, number of isolated nodes and motif overlap.

[0019] In some embodiments, step S3 specifically includes:

[0020] Calculate the mean and standard deviation of the macro-level structure parameter sequence, the meso-level structure parameter sequence, and the micro-level structure parameter sequence respectively, and calculate the center line, upper control limit, and lower control limit of each sequence;

[0021] Draw the control chart of each parameter sequence, regard the control chart of each parameter sequence as a classifier, and give the preliminary fault diagnosis result corresponding to each parameter sequence according to the judgment criteria of the control chart theory.

[0022] In some embodiments, step S4 specifically includes:

[0023] In each round of diagnosis, the classifier whose diagnosis result is not adopted will not participate in the result fusion in the next round of diagnosis, thus realizing the dynamic selection of classifiers.

[0024] In some embodiments, the dynamic selection of the classifier specifically includes:

[0025] In t i At time t, for each classifier i The preliminary fault diagnosis result at time t is obtained. The number of votes for faults and normal conditions is counted by voting. The one with the most votes is taken as i Final diagnosis result at any time;

[0026] In t i+1 time, for t i The classifier whose output result is a minority class at time t only records its diagnosis result but does not participate in the voting. i The output result at the moment is the majority class classifier, which votes to get t i+1 Fault diagnosis results at the moment;

[0027] In t i+2 time, by t i+1 The output result at the moment is the majority class classifier, which votes to get t i+2 Fault diagnosis results at the moment;

[0028] And so on, dynamic integration of multi-level fault diagnosis results is achieved.

[0029] In a second aspect, a multi-level fault diagnosis system based on a dynamic sequential network is provided, comprising:

[0030] A dynamic timing network building module is used to collect target system information according to a given frequency and build a dynamic timing network;

[0031] A multi-level structure parameter sequence establishing module, used for calculating the multi-level structure parameters of the dynamic time series network topology structure and establishing a multi-level structure parameter sequence;

[0032] A preliminary fault diagnosis module is used to perform multi-level preliminary fault diagnosis based on control chart theory using a multi-level structure parameter sequence to obtain multiple preliminary fault diagnosis results;

[0033] The diagnosis fusion module is used to fuse the multiple preliminary fault diagnosis results by adopting a dynamic integration method.

[0034] In a third aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, any one of the multi-level fault diagnosis methods described above is implemented.

[0035] In a fourth aspect, an electronic device is provided, comprising a memory and a processor, wherein the memory stores computer instructions executable on the processor, and the processor executes any one of the multi-level fault diagnosis methods described above when executing the computer instructions.

[0036] It should be further explained that the technical features corresponding to the above options can be combined or replaced with each other to form a new technical solution if there is no conflict.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The present invention first obtains the system architecture information of the target system at different time points according to a certain frequency analysis, considers the dynamic behavior of the actual system, and establishes a network model that can reflect the system operation topology structure based on the dynamic timing network. Then, based on the network topology structure of the dynamic timing network, its structural parameters are calculated from multiple levels to obtain a multi-level parameter sequence, and preliminary fault diagnosis is achieved from multiple levels. Comprehensive fault diagnosis is achieved based on a dynamic selection method, which can realize simple and easy online fault diagnosis and meet the health monitoring needs of complex systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A flow chart of a multi-level fault diagnosis method based on a dynamic timing network shown in an embodiment of the present invention;

[0040] Figure 2 The specific process of multi-level fault diagnosis shown in the embodiment of the present invention;

[0041] Figure 3An example of a dynamic timing network shown in an embodiment of the present invention;

[0042] Figure 4 The embodiment of the present invention shows t i Example of a time-average control chart;

[0043] Figure 5 The embodiment of the present invention shows t i+1 Example of a Moment Mean Degree control chart. DETAILED DESCRIPTION

[0044] The technical solution of the present invention is clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. The components of the embodiments of the present application described and shown in the accompanying drawings can be arranged and designed in various configurations. 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.

[0045] It should be noted that the defects existing in the solutions in the above-mentioned prior art are the results obtained by the inventor after practice and careful research. Therefore, the discovery process of the above-mentioned problems and the solutions proposed in the embodiments of the present application for the above-mentioned problems below should all be the contributions made by the inventor to the present application in the process of invention and creation, and should not be understood as technical contents known to technical personnel in this field.

[0046] In view of the technical problems pointed out in the background technology, the embodiments provided by the present invention are as follows:

[0047] Example 1

[0048] Reference Figure 1 In an exemplary embodiment, a multi-level fault diagnosis method based on a dynamic sequential network is provided, comprising the following steps:

[0049] S1. Collect target system information according to given frequency and establish dynamic timing network;

[0050] S2, calculating the multi-level structural parameters of the dynamic time series network topology structure and establishing a multi-level structural parameter sequence;

[0051] S3. Based on the control chart theory, a multi-level structural parameter sequence is used to perform multi-level preliminary fault diagnosis and obtain multiple preliminary fault diagnosis results;

[0052] S4. Adopting a dynamic integration method to fuse the multiple preliminary fault diagnosis results.

[0053] Specifically, step S1 includes:

[0054] The sampling frequency fre is determined based on the monitoring requirements, cost and technical level of the target system, and then the architecture information of the target system is collected according to the given sampling frequency fre. The system architecture mainly includes the entities at a certain level of the system and the interactive dependencies between entities. Among them, the entities can be the software and hardware components that constitute the system, and the level of the entities is determined according to the needs of fault diagnosis. The relationship between entities can be the interactive dependencies such as data flow and control flow between entities.

[0055] At a given sampling frequency fre, each sampling time t i A set of system architecture data can be obtained, and t i The set of system entities obtained at the moment is recorded as n is the number of entities; the set of dependencies between entities is recorded as m is the number of dependencies between entities. i System architecture data at the moment and Then we can establish t i Complex network model of target system at each moment A series of network models established at all sampling moments constitute a dynamic timing network l represents the number of sampling times, which is also equal to the number of complex network models contained in the dynamic timing network.

[0056] Step S2 specifically includes:

[0057] (1) Calculate the macro-structure parameters of the network and establish a macro-level structure parameter sequence: For each network in the dynamic time series network, calculate its macro-level topological structure parameters. Macro-structure parameters refer to parameters that characterize the entire network structure. The macro-structure parameters determined by the present invention include common parameters such as average degree Deg, average shortest path length Pat, average clustering coefficient Coe, network diameter Dia, number of network sub-clusters Clu, and maximum connected sub-cluster size ratio Max. The present invention defines the maximum connected sub-cluster size ratio Max as:

[0058]

[0059] Among them, Nodes mcg Nodes refers to the number of nodes in the largest connected subcluster of the network. all is the total number of network nodes. Therefore, Max refers to the ratio of the number of nodes in the largest connected sub-cluster of the network to the total number of network nodes.

[0060] At this point, we can establish the macro-hierarchical parameter sequence of the network, including the average degree sequence Average shortest path length sequence Average clustering coefficient series Network diameter sequence Network subcluster number sequence And the maximum connected subcluster size ratio sequence All sequences have a length of l, which is the number of samples.

[0061] (2) Calculate the network meso-structure parameters to establish a meso-level structure parameter sequence: For each network in the dynamic time series network, calculate the meso-level topological structure parameters. Meso-structure parameters refer to parameters that characterize the structure of a part of the node set in the network. The meso-structure parameters determined by the present invention include common parameters such as the number of communities Com, modularity Mod, and community overlap Cov. The present invention defines the community overlap Cov as:

[0062]

[0063] in, Refers to the number of community shared nodes in the network. Community shared nodes are nodes that belong to two or more communities at the same time. all is the total number of network nodes.

[0064] At this point, we can establish the network's meso-level structure parameter sequence, including the community number sequence Modularity sequence Community overlap sequence All sequences have a length of l, which is the number of samples.

[0065] (3) Calculate the network microstructure parameters to establish a micro-level structure parameter sequence: For each network in the dynamic time series network, calculate its micro-level topological structure parameters. Microstructure parameters refer to parameters that characterize the local microstructure in the network. The microstructure parameters determined by the present invention include the number of 3-node motif types Mot, the number of isolated nodes Ino, and the motif overlap Mov. The present invention defines the motif overlap Mov as:

[0066]

[0067] in, Refers to the number of motif-sharing nodes in the network. Motif-sharing nodes are nodes that belong to two or more motif types at the same time. all Also the total number of network nodes.

[0068] At this point, the micro-hierarchical structure parameter sequence of the network can be established, including the number sequence of 3-node motif types Sequence of number of isolated nodes Motif overlap sequence All sequences have a length of l, which is the number of samples.

[0069] Step S3 specifically includes:

[0070] First, we calculate the macro-level structure parameter sequence Deg t Pat t , Coe t 、Dia t 、Clu t , Max t , the meso-level structure parameter sequence Com t Mod t , Cov t , and the microscopic hierarchical structure parameter sequence Mot t 、Ino t 、Mov t The mean μ and standard deviation σ of the series are then calculated, and the center line CL, upper control limit UCL and lower control limit LCL of each series are calculated:

[0071]

[0072] Among them, L is the control limit coefficient, which can be selected according to needs, and generally L=3 can be selected.

[0073] After determining the center line CL, upper control limit UCL and lower control limit LCL of the multi-level structure parameter sequence, the control chart of each parameter sequence can be drawn, and each structure parameter sequence control chart can be regarded as a classifier. Then, according to the abnormality judgment criteria of the control chart theory, a preliminary conclusion is given from the perspective of the multi-level structure parameter sequence whether there is an abnormality in the dynamic timing network, that is, the preliminary fault diagnosis result, including normal or faulty.

[0074] During the continuous monitoring process, the length of the control chart, that is, the sample size n used to draw the control chart, can be determined as needed. The purpose is to perform fault diagnosis based on the data obtained at the nth sampling moment before the current moment.

[0075] Step S4 specifically includes:

[0076] Based on the control chart theory, a dynamic integration method is used to fuse multiple preliminary fault diagnosis results to improve the diagnostic accuracy. The dynamic integration method of the present invention means that in each round of diagnosis, the classifier whose diagnosis result is not adopted does not participate in the result fusion in the next round of diagnosis. The dynamic selection of the classifier is realized through the above process, which is as follows:

[0077] ①At t i At time t, for each classifier i The preliminary diagnosis result at time t is obtained. The number of votes for faults and normal conditions is counted by voting. The one with the most votes is taken as i Final diagnosis result at any time;

[0078] ②At ti+1 time, for t i The classifier whose output result is a minority class at time t only records its diagnosis result but does not participate in the voting. i The output result at the moment is the majority class classifier, which votes to get t i+1 Fault diagnosis results at the moment;

[0079] ③In t i+2 time, by t i+1 The output result at the moment is the majority class classifier, which votes to get t i+2 The subsequent process is similar to this, realizing the dynamic integration of multi-level fault diagnosis results.

[0080] Example 2

[0081] Based on the inventive concept of Example 1, this example takes a certain aerospace product as an example to illustrate the implementation process of using the present invention to perform fault diagnosis. Figure 2 As shown, the specific steps are as follows:

[0082] Step 1: Dynamic timing network modeling

[0083] According to the system health monitoring requirements, the sampling frequency fre is determined to be 1Hz, that is, the system architecture information is collected once per second. Based on the system architecture data at each sampling moment and Build a dynamic timing network for the system The visualization results are as follows Figure 3 shown.

[0084] Steps 2 to 4: Multi-level parameter sequence calculation

[0085] In dynamic timing networks Based on this, NetworkX is used to calculate the corresponding macro-level structure parameter sequence Deg t Pat t , Coe t 、Dia t 、Clu t , Max t , the meso-level structure parameter sequence Com t Mod t , Cov t , and the microscopic hierarchical structure parameter sequence Mot t 、Ino t 、Mov t .

[0086] Steps 5-6: Multi-level fault diagnosis

[0087] First, a control chart is drawn for each structural parameter sequence to perform preliminary fault diagnosis. t For example, in this embodiment, the control limit coefficient L=3 is taken, the mean-standard deviation control chart of the average parameter sequence is established, and the control chart sample size n=15, t i The average degree control diagram at each moment is as follows: Figure 4 As shown. According to the control chart abnormality criterion, from the perspective of the average degree parameter sequence, the system is in a "normal" state. Similarly, Pat t , Coe t 、Dia t 、Clu t , Max t 、Com t Mod t , Cov t 、Mot t 、Ino t 、Mov t Equi-parameter sequences can also give their own preliminary fault diagnosis results.

[0088] Subsequently, the dynamic integration method is used to fuse the preliminary fault diagnosis results. In this embodiment, t i Moment Deg t Pat t , Coe t 、Dia t 、Clu t , Max t 、Com t Mod t , Cov t 、Mot t 、Ino t 、Mov t The preliminary diagnosis result of the structural parameter sequence is [normal, normal, fault, normal, fault, normal, normal, normal, fault, normal, normal, normal], so t i The final diagnosis result at this moment is "normal".

[0089] t i+1 The average degree control diagram at each moment is as follows: Figure 5 As shown in Figure 2, according to the control chart exception criteria, one observation falls outside the upper control limit, so from the perspective of the average parameter sequence, the system is in a "fault" state. i+1 Moment Deg t Pat t , Coe t 、Dia t 、Clu t , Max t 、Comt Mod t , Cov t 、Mot t 、Ino t 、Mov t The preliminary diagnosis result of the parameter sequence is [fault, normal, fault, fault, fault, normal, fault, fault, fault, normal, fault, fault]. In addition, according to the dynamic integration rule of the present invention, since at t i Moment Coe t 、Clu t , Cov t The initial diagnosis result is the minority class, so at t i+1 When the results are integrated at all times, Coe t 、Clu t , Cov t The preliminary diagnosis result of does not participate in the result fusion, and is only fused by the preliminary diagnosis results of the remaining parameter sequences, i.e., [fault, normal, fault, normal, fault, fault, normal, fault, fault], and the final diagnosis result is “fault”.

[0090] In t i+2 time, due to t i+1 Moment Pat t , Max t 、Mot t The diagnosis result of is the minority class. Therefore, at t i+2 Moment, Pat t , Max t 、Mot t The preliminary diagnosis results of do not participate in the result fusion, and the preliminary diagnosis results of the remaining parameter sequences are used for fusion. And so on, the integration of fault diagnosis at all sampling moments is completed.

[0091] Example 3

[0092] Based on the same inventive concept as that of Example 1, a multi-level fault diagnosis system based on a dynamic sequential network is provided, comprising:

[0093] A dynamic timing network building module is used to collect target system information according to a given frequency and build a dynamic timing network;

[0094] A multi-level structure parameter sequence establishing module, used for calculating the multi-level structure parameters of the dynamic time series network topology structure and establishing a multi-level structure parameter sequence;

[0095] A preliminary fault diagnosis module is used to perform multi-level preliminary fault diagnosis based on control chart theory using a multi-level structure parameter sequence to obtain multiple preliminary fault diagnosis results;

[0096] The diagnosis fusion module is used to fuse the multiple preliminary fault diagnosis results by adopting a dynamic integration method.

[0097] It should be noted that each module of the system integrates the software and hardware corresponding to the functions of each step in the implementation method to achieve the same function, and the specific implementation method of each module will not be repeated here.

[0098] Example 4

[0099] Based on the same inventive concept as Example 1, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the multi-level fault diagnosis method based on a dynamic timing network provided by the embodiment of the present invention is implemented. Based on such an understanding, the technical solution of this embodiment is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes various media that can store program codes, such as a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0100] Example 5

[0101] Based on the same inventive concept as Example 1, an electronic device is provided, including a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, and when the processor runs the computer instructions, the multi-level fault diagnosis method based on the dynamic timing network provided in the embodiment of the present invention is executed.

[0102] The processor may be a single-core or multi-core central processing unit or a specific integrated circuit, or one or more integrated circuits configured to implement the present invention.

[0103] Embodiments of the subject matter and functional operations described in this specification may be implemented in: tangibly embodied computer software or firmware, computer hardware including the structures disclosed in this specification and their structural equivalents, or a combination of one or more of them. Embodiments of the subject matter described in this specification may be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible, non-transitory program carrier to be executed by a data processing device or to control the operation of the data processing device. Alternatively or additionally, the program instructions may be encoded on an artificially generated propagated signal, such as a machine-generated electrical, optical, or electromagnetic signal, which is generated to encode and transmit information to a suitable receiver device for execution by the data processing device.

[0104] The processes and logic flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform corresponding functions by operating on input data and generating output. The processes and logic flows can also be performed by special purpose logic circuits, such as FPGAs (field programmable gate arrays) or ASICs (application-specific integrated circuits), and the apparatus can also be implemented as special purpose logic circuits.

[0105] Processors suitable for executing computer programs include, for example, general and / or special microprocessors, or any other type of central processing unit. Typically, the central processing unit will receive instructions and data from a read-only memory and / or a random access memory. The basic components of a computer include a central processing unit for implementing or executing instructions and one or more memory devices for storing instructions and data. Typically, the computer will also include one or more large-capacity storage devices for storing data, such as magnetic disks, magneto-optical disks, or optical disks, or the computer will be operably coupled to this large-capacity storage device to receive data from it or to transmit data to it, or both. However, the computer does not necessarily have such a device. In addition, the computer can be embedded in another device, such as a mobile phone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a global positioning system (GPS) receiver, or a portable storage device such as a universal serial bus (USB) flash drive, just to name a few.

[0106] It should be understood that each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code comprises one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

[0107] The above specific implementation methods are detailed descriptions of the present invention. It cannot be determined that the specific implementation methods of the present invention are limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions and substitutions can be made without departing from the concept of the present invention, which should be regarded as belonging to the protection scope of the present invention.

Claims

1. A multi-level fault diagnosis method based on dynamic sequential network, characterized in that: The following steps are involved: S1. Collect target system information according to given frequency and establish dynamic timing network; S2, calculating the multi-level structural parameters of the dynamic time series network topology structure and establishing a multi-level structural parameter sequence; S3. Based on the control chart theory, a multi-level structural parameter sequence is used to perform multi-level preliminary fault diagnosis and obtain multiple preliminary fault diagnosis results; S4. Adopting a dynamic integration method to fuse the multiple preliminary fault diagnosis results.

2. The multi-level fault diagnosis method based on dynamic sequential network according to claim 1 is characterized in that: The step S1 specifically includes: At a given sampling frequency fre, each sampling time t i Get a set of system architecture data and convert t i The set of system entities obtained at the moment is recorded as n is the number of entities; the set of dependencies between entities is recorded as m is the number of dependencies between entities; Based on t i System architecture data at the moment and Build i Complex network model of target system at each moment A series of network models established at all sampling moments constitute a dynamic timing network Where l represents the number of sampling times.

3. The multi-level fault diagnosis method based on dynamic sequential network according to claim 1 is characterized in that: The step S2 specifically includes: The macro-structure parameters of the network are calculated to establish a macro-level structure parameter sequence; the meso-structure parameters of the network are calculated to establish a meso-level structure parameter sequence; the micro-structure parameters of the network are calculated to establish a micro-level structure parameter sequence; among them, the macro-structure parameters refer to the parameters that characterize the structure of the entire network, the meso-structure parameters refer to the parameters that characterize the structure of some node sets in the network, and the micro-structure parameters refer to the parameters that characterize the local micro-structure in the network.

4. The multi-level fault diagnosis method based on dynamic sequential network according to claim 3 is characterized in that: The macro-structure parameters include average degree, average shortest path length, average clustering coefficient, network diameter, number of network subclusters and ratio of the maximum connected subclusters; the meso-structure parameters include number of communities, modularity and community overlap; the micro-structure parameters include number of 3-node motif types, number of isolated nodes and motif overlap.

5. The multi-level fault diagnosis method based on dynamic sequential network according to claim 3 is characterized in that: The step S3 specifically includes: Calculate the mean and standard deviation of the macro-level structure parameter sequence, the meso-level structure parameter sequence, and the micro-level structure parameter sequence respectively, and calculate the center line, upper control limit, and lower control limit of each sequence; Draw the control chart of each parameter sequence, regard the control chart of each parameter sequence as a classifier, and give the preliminary fault diagnosis result corresponding to each parameter sequence according to the judgment criteria of the control chart theory.

6. The multi-level fault diagnosis method based on dynamic sequential network according to claim 5 is characterized in that: The step S4 specifically includes: In each round of diagnosis, the classifier whose diagnosis result is not adopted will not participate in the result fusion in the next round of diagnosis, thus realizing the dynamic selection of classifiers.

7. The multi-level fault diagnosis method based on dynamic sequential network according to claim 6 is characterized in that: The dynamic selection of the classifier specifically includes: In t i At time t, for each classifier i The preliminary fault diagnosis result at time t is obtained. The number of votes for faults and normal conditions is counted by voting. The one with the most votes is taken as i Final diagnosis result at any time; In t i+1 time, for t i The classifier whose output result is a minority class at time t only records its diagnosis result but does not participate in the voting. i The output result at the moment is the majority class classifier, which votes to get t i+1 Fault diagnosis results at the moment; In t i+2 time, by t i+1 The output result at the moment is the majority class classifier, which votes to get t i+2 Fault diagnosis results at the moment; And so on, dynamic integration of multi-level fault diagnosis results is achieved.

8. A multi-level fault diagnosis system based on a dynamic sequential network, characterized in that: include: A dynamic timing network building module is used to collect target system information according to a given frequency and build a dynamic timing network; A multi-level structure parameter sequence establishing module, used for calculating the multi-level structure parameters of the dynamic time series network topology structure and establishing a multi-level structure parameter sequence; A preliminary fault diagnosis module is used to perform multi-level preliminary fault diagnosis based on control chart theory using a multi-level structure parameter sequence to obtain multiple preliminary fault diagnosis results; The diagnosis fusion module is used to fuse the multiple preliminary fault diagnosis results by adopting a dynamic integration method.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the multi-level fault diagnosis method described in any one of claims 1 to 7 is implemented.

10. An electronic device comprising a memory and a processor, wherein the memory stores computer instructions that can be executed on the processor, wherein: When the processor runs the computer instructions, it executes the multi-level fault diagnosis method described in any one of claims 1-7.

Images