Complex system causal graph construction and synthesis method and system, storage medium and electronic equipment

Abstract

The invention relates to a complex system causal graph construction and synthesis method and system, a storage medium and electronic equipment. The method comprises the following steps: collecting comprehensive data and constructing a plurality of subsystem causal graphs based on the comprehensive data; the comprehensive data comprises internal / external events, RCA analysis data, experience feedback data, RCM maintenance guide rule data and expert experience data; judging whether common nodes exist among the multiple subsystem causal diagrams or not; if no common node exists among the plurality of subsystem causal diagrams, synthesizing the plurality of subsystem causal diagrams based on a first synthesis mode to obtain a synthesized causal diagram of the complex system; and if the common nodes exist among the plurality of subsystem causal graphs, synthesizing the plurality of subsystem causal graphs based on a second synthesis mode to obtain a synthesized causal graph of the complex system. The complex system causal graph constructed by field experts is synthesized, so that the synthesized causal graph can truly reflect reality, deviation is reduced, and accuracy and reliability are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment monitoring and diagnosis, and more specifically, to a complex system causal graph construction and synthesis method, system, storage medium and electronic device. Background Art

[0002] Once a failure occurs in important sensitive equipment in a nuclear power plant, it will have a great impact on the safety of the power plant system. Quickly and accurately locating the cause of the failure in the early stage is an effective means to effectively avoid the deterioration of the failure and improve the safety of power plant operation. The large amount of historical data, operation and maintenance experience, maintenance reports and other data and information stored in the operation of the power plant contain rich value, but due to the lack of effective technical means, it is impossible to provide guidance for the operation and maintenance process of complex systems from the data and knowledge level.

[0003] A complex system is a system composed of a large number of interacting units. The research content of complexity is to study how complex systems produce organized behaviors under certain rules and the behaviors that emerge from the evolution of the system. Quantitative modeling of complex systems is a difficult and hot topic in the field of complex system research. Quantitative modeling refers to analyzing the input and output behaviors of the system based on the knowledge of the complex system (incomplete or complete), realizing the quantitative representation and quantitative control of the complex system.

[0004] The causal relationship between things is a hot topic in the field of artificial intelligence. Today, the theory of causal graphs has been greatly developed and is mainly used in the field of fault diagnosis. In particular, causal graphs have powerful model representation capabilities and can represent the complex causal relationships between units in the entire complex system, providing an effective way to build an expert knowledge base. However, in actual situations, the causal graphs of complex systems constructed by domain experts are somewhat subjective. How to make the constructed causal graphs truly reflect reality and reduce deviations is a problem worth studying. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a method, system, storage medium and electronic device for constructing and synthesizing a causal graph of a complex system in response to the problems existing in the prior art.

[0006] The technical solution adopted by the present invention to solve the technical problem is: constructing a complex system causal graph construction and synthesis method, including the following steps:

[0007] Collecting comprehensive data and constructing multiple subsystem cause-and-effect diagrams based on the comprehensive data; the comprehensive data includes: internal / external events, RCA analysis data, experience feedback data, RCM maintenance guideline data, and expert experience data;

[0008] Determine whether there are common nodes among the multiple subsystem cause-and-effect graphs;

[0009] If there are no common nodes between the multiple subsystem cause-and-effect graphs, synthesizing the multiple subsystem cause-and-effect graphs based on the first synthesis method to obtain a synthesized cause-and-effect graph of the complex system;

[0010] If there are common nodes between the multiple subsystem cause-effect graphs, the multiple subsystem cause-effect graphs are synthesized based on the second synthesis method to obtain a synthesized cause-effect graph of the complex system.

[0011] In the complex system causal graph construction and synthesis method of the present invention, the collecting of comprehensive data and constructing multiple subsystem causal graphs based on the comprehensive data includes:

[0012] Performing comprehensive data analysis on the comprehensive data to obtain comprehensive data analysis results;

[0013] Determine the failure mode of the complex system and the alarm signal that triggers the failure reasoning based on the comprehensive analysis results of the data;

[0014] The subsystem cause-and-effect diagram is constructed based on the failure mode or the alarm signal.

[0015] In the complex system causal graph construction and synthesis method of the present invention, synthesizing the multiple subsystem causal graphs based on the first synthesis method to obtain the synthesized causal graph of the complex system includes:

[0016] Obtaining an initial probability vector synthesis formula and a causal strength matrix synthesis formula;

[0017] Synthesize the multiple subsystem causal graphs according to the initial probability vector synthesis formula and the causal strength matrix synthesis formula to obtain a synthetic causal graph of the complex system;

[0018] The initial probability vector synthesis formula is:

[0019] PV=(PV1,PV2,...PV m );

[0020] The causal intensity matrix synthesis formula is:

[0021]

[0022] Among them, PV1, PV2, ...PV m are the initial probability vectors corresponding to the causal graphs of m subsystems; CS1, CS2, ...CS m are the causal intensity matrices corresponding to the causal graphs of m subsystems.

[0023] In the complex system causal graph construction and synthesis method of the present invention, synthesizing the multiple subsystem causal graphs based on the second synthesis method to obtain the synthesized causal graph of the complex system includes:

[0024] Normalizing the initial probability vectors and causal strength matrices of the plurality of subsystem causal graphs;

[0025] After completing the normalization, calculating the deviation degree of the cause-effect diagrams of the plurality of subsystems;

[0026] The synthesized causal graph is obtained by synthesizing the causal graphs of the multiple subsystems based on the deviation degrees of the causal graphs.

[0027] In the complex system causal graph construction and synthesis method of the present invention, the synthesis based on the deviation degree of the multiple subsystem causal graphs and the deviation degree of the synthetic causal graph to obtain the synthetic causal graph includes:

[0028] Screening based on the deviations of the plurality of subsystem cause-effect diagrams to determine unqualified subsystem cause-effect diagrams;

[0029] Deleting the unqualified subsystem cause-and-effect diagram and retaining the qualified subsystem cause-and-effect diagram;

[0030] The qualified subsystem cause-effect graph is synthesized by using a distance-based weight determination algorithm to obtain the synthesized cause-effect graph.

[0031] In the complex system causal graph construction and synthesis method of the present invention, the deviation degree of the subsystem causal graph includes: the deviation degree of the initial probability vector of the subsystem causal graph and the deviation degree of the causal strength matrix;

[0032] The screening based on the deviation degree of the plurality of subsystem cause-effect diagrams to determine unqualified subsystem cause-effect diagrams includes:

[0033] Calculate the degree of deviation of the initial probability vector of each subsystem causal graph;

[0034] Calculate the degree of deviation of the causal strength matrix of each subsystem causal graph;

[0035] The deviation degree of each subsystem causal graph is obtained by calculating the deviation degree of the initial probability vector of each subsystem causal graph and the deviation degree of the causal strength matrix;

[0036] Screening is performed based on the degree of deviation of each subsystem cause-effect diagram to determine unqualified subsystem cause-effect diagrams.

[0037] In the complex system cause-effect graph construction and synthesis method of the present invention, the screening according to the deviation degree of each subsystem cause-effect graph to determine the unqualified subsystem cause-effect graph includes:

[0038] comparing the degree of deviation of each subsystem cause-effect diagram with a reference value;

[0039] If the deviation of the subsystem cause-effect diagram is greater than the reference value, the subsystem cause-effect diagram is determined to be an unqualified subsystem cause-effect diagram, and the unqualified subsystem cause-effect diagram is screened out.

[0040] The present invention also provides a complex system causal graph construction and synthesis system, comprising:

[0041] A data collection unit, used to collect comprehensive data and construct multiple subsystem cause-and-effect diagrams based on the comprehensive data; the comprehensive data includes: internal / external events, RCA analysis data, experience feedback data, RCM maintenance guideline data and expert experience data;

[0042] A node analysis unit, used to determine whether there are common nodes between the multiple subsystem cause-and-effect graphs;

[0043] A first synthesis unit, configured to synthesize the plurality of subsystem causal graphs based on a first synthesis method to obtain a synthesized causal graph of a complex system if there are no common nodes between the plurality of subsystem causal graphs;

[0044] The second synthesis unit is used to synthesize the multiple subsystem cause-effect graphs based on a second synthesis method if there are common nodes between the multiple subsystem cause-effect graphs to obtain a synthesized cause-effect graph of the complex system.

[0045] The present invention also provides a storage medium storing a computer program, wherein the computer program is suitable for being loaded by a processor to execute the steps of the complex system causal graph construction and synthesis method as described above.

[0046] The present invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the complex system causal graph construction and synthesis method as described above by calling the computer program stored in the memory.

[0047] The complex system causal graph construction and synthesis method, system, storage medium and electronic device of the present invention have the following beneficial effects: including the following steps: collecting comprehensive data and constructing multiple subsystem causal graphs based on the comprehensive data; the comprehensive data includes: internal / external events, RCA analysis data, experience feedback data, RCM maintenance guide data and expert experience data; judging whether there are common nodes between multiple subsystem causal graphs; if there are no common nodes between multiple subsystem causal graphs, synthesizing multiple subsystem causal graphs based on the first synthesis method to obtain a synthetic causal graph of the complex system; if there are common nodes between multiple subsystem causal graphs, synthesizing multiple subsystem causal graphs based on the second synthesis method to obtain a synthetic causal graph of the complex system. The present invention synthesizes the complex system causal graph constructed by domain experts, so that the synthesized causal graph can truly reflect the reality, reduce deviations, and improve accuracy and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0049] Figure 1 It is a flow chart of the complex system cause-effect diagram construction and synthesis method provided by the present invention;

[0050] Figure 2 It is a schematic diagram of the process of constructing a subsystem cause-effect graph provided by the present invention;

[0051] Figure 3 It is a schematic diagram of the process of synthesizing the subsystem cause-effect graph provided by the present invention;

[0052] Figure 4 It is a principle block diagram of a complex system cause-effect diagram construction and synthesis system provided by the present invention;

[0053] Figure 5 is a schematic diagram of a subsystem causal graph D1 provided by the present invention;

[0054] Figure 6 is a schematic diagram of a subsystem causal graph D2 provided by the present invention;

[0055] Figure 7 is a schematic diagram of a subsystem causal graph D3 provided by the present invention;

[0056] Figure 8 is a schematic diagram of a subsystem causal graph D4 provided by the present invention;

[0057] Fig. 9 It is a schematic diagram of the synthetic causal graph D provided by the present invention. DETAILED DESCRIPTION

[0058] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] refer to Figure 1 , Figure 1 A schematic diagram of the process of constructing and synthesizing a complex system causal graph provided by the present invention.

[0060] Specifically, Figure 1 As shown, the complex system causal graph construction and synthesis method includes the following steps:

[0061] Step S101: Collect comprehensive data and construct multiple subsystem cause-and-effect diagrams based on the comprehensive data.

[0062] The comprehensive data includes: internal / external events, RCA (Root Cause Analysis) analysis data, experience feedback data, RCM (Reliability-centered Maintenance) maintenance guideline data and expert experience data. The RCM (Reliability-centered Maintenance) analysis guideline is a programmatic document for RCM analysis of nuclear power plant equipment, which defines the RCM analysis process and the criteria for critical judgment of failure consequences. RCA (Root Cause Analysis) is to find out the failure mechanism and root cause of equipment failure through standardized analysis methods.

[0063] In some embodiments, Figure 2 As shown, collecting comprehensive data and constructing multiple subsystem cause-and-effect diagrams based on the comprehensive data includes: performing comprehensive data analysis on the comprehensive data to obtain comprehensive data analysis results; determining the failure mode of the complex system and the alarm signal that triggers failure reasoning based on the comprehensive data analysis results; and constructing a subsystem cause-and-effect diagram based on the failure mode or the alarm signal.

[0064] Specifically, the comprehensive analysis of the comprehensive data mainly analyzes the monitoring parameters and failure modes of the equipment. Among them, the monitoring parameter analysis mainly focuses on the selection of parameters and the arrangement of measurement points; the failure mode analysis mainly focuses on common failure mechanisms. Furthermore, it is necessary to analyze the specific failure instances of the equipment to analyze its failure mechanism, cause and impact. This analysis follows the basic process of FMEA analysis. Through FMEA analysis, basic information such as the function, functional failure, failure mode, and failure impact of the equipment is obtained; events that significantly affect the performance, economy, reliability, safety or other required characteristics of the equipment are summarized and defined. Such events are usually the core of the fault tree analysis, that is, the top event; the characteristic signs of the equipment failure mode are analyzed. The characteristic signs of the failure mode specifically refer to the signs before the failure occurs obtained through secondary analysis of the monitoring parameters or subjective experience judgment. Domain experts can quantify the failure signs according to conditions or known quantities, and analyze the quantification results to obtain the failure signs corresponding to common failure modes. Usually, the judgment of the failure mode corresponds to the combined analysis of multiple signs; the association between the equipment failure mode and the monitoring parameters is analyzed. Based on in-depth research on equipment failure mechanisms, the failure mode groups and monitoring parameter groups of each device are summarized; and correlation analysis between failure modes and monitoring parameters is carried out.

[0065] After completing the comprehensive analysis of the data, construct the subsystem cause-and-effect diagram. The specific steps are as follows:

[0066] 1) Determine the list of {X, B} type variables of the complex system;

[0067] 2) Divide the states of {X, B} type variables and give the membership interval;

[0068] 3) Determine the initial probability parameters of the abnormal state of the B type variable, including the failure rate λ of the initial event and the failure probability of the non-initial event, where the dimension of λ must be unified;

[0069] 4) The specific steps of constructing a cause-effect diagram according to the alarm signal are:

[0070] a) Select the variable X in the subsystem n ;

[0071] b) Determine X from the defined {X, B} type variables n The parent variable of

[0072] c) Use action variables → or conditional action variables Combine the parent variable with the child variable X n connected, among which

[0073] Logic gate variables may be used, and the logic gate variables must be written in LGS table;

[0074] d) Add default variable D if necessary n As Xn The parent variable of X is linked to it using a scoped variable or a conditional scoped variable. n The default variable is usually connected to the X variable by a conditional event variable (dashed arrow), and should be used as little as possible;

[0075] For each action variable and conditional action variable, determine its causal strength p ij .

[0076] Among them, X: represents the result or effect event or variable, drawn as a circle, and it can also be a cause event or variable; B: represents the basic or root event or variable, drawn as a rectangle, and it can only be an independent cause; G: represents the logic gate event or variable. A logic gate can be the input variable of another logic gate, and multiple logic gates can be the parent variable of the same child variable; P: represents the connection event or variable; Initial probability: also known as failure rate or incidence rate. It is the reciprocal of the mean fault-free continuous working time MTBF, which can be obtained statistically, or can be obtained from the power plant, or estimated. The probability parameter values ​​involved in this article are set by experts based on experience and field experience feedback; Causal strength: expresses the possibility of a certain state of the parent variable and a certain state of the child variable. The setting of this value is set according to the lag principle of causal transmission.

[0077] Step S102: Determine whether there are common nodes among multiple subsystem cause-and-effect graphs.

[0078] Step S103: If there are no common nodes between the multiple subsystem cause-effect graphs, the multiple subsystem cause-effect graphs are synthesized based on the first synthesis method to obtain a synthesized cause-effect graph of the complex system.

[0079] In some embodiments, synthesizing multiple subsystem causal graphs based on a first synthesis method to obtain a synthetic causal graph of a complex system includes: obtaining an initial probability vector synthesis formula and a causal strength matrix synthesis formula; synthesizing multiple subsystem causal graphs according to the initial probability vector synthesis formula and the causal strength matrix synthesis formula to obtain a synthetic causal graph of the complex system.

[0080] Among them, the initial probability vector synthesis formula is:

[0081] PV=(PV1,PV2,...PV m ) (1)

[0082] The synthesis formula of the causal intensity matrix is:

[0083]

[0084] Among them, PV1, PV2, ...PV m are the initial probability vectors corresponding to the causal graphs of m subsystems; CS1, CS2, ...CS mare the causal intensity matrices corresponding to the causal graphs of m subsystems.

[0085] Specifically, connecting different causal graphs constructed for the same system into one causal graph is called the synthesis of causal graphs, which is defined as a synthetic causal graph.

[0086] Assume that there are n nodes involved in the system to be studied, and m experts on this system propose m subsystem causal graphs D1, D2, ..., D m , the corresponding initial probability vectors are PV1, PV2, ..., PV m , the causal strength matrix is ​​CS1,CS2,...,CS m , then the initial probability vector of the system causal graph D synthesized by these m causal graphs is PV, and the causal intensity matrix is ​​CS.

[0087] Corollary: Assume that there are n nodes involved in the system to be studied, and construct m sub-causal graphs D1, D2, ...D for this system. m , the corresponding initial probability vectors are PV1, PV2, ...PV m , the corresponding causal strength matrix is ​​CS1, CS2, ...CS m , then the initial probability vector of the causal graph D synthesized by m sub-causal graphs is PV, and the causal strength matrix is ​​CS.

[0088] If there are no common nodes between the causal graphs of the m subsystems, and their causal strength matrices and initial probability vectors are not standardized, the above equations (1) and (2) are used to synthesize the causal graphs.

[0089] Step S104: If there are common nodes between the multiple subsystem cause-effect graphs, the multiple subsystem cause-effect graphs are synthesized based on the second synthesis method to obtain a synthesized cause-effect graph of the complex system.

[0090] In some embodiments, multiple subsystem causal graphs are synthesized based on the second synthesis method to obtain a synthetic causal graph of a complex system, including: standardizing the initial probability vectors and causal strength matrices of the multiple subsystem causal graphs; after completing the standardization, calculating the deviation degrees of the multiple subsystem causal graphs; and synthesizing based on the deviation degrees of the multiple subsystem causal graphs to obtain a synthetic causal graph.

[0091] Among them, synthesizing based on the deviation degrees of multiple subsystem causal graphs and the deviation degrees of the synthetic causal graph to obtain the synthetic causal graph includes: screening based on the deviation degrees of multiple subsystem causal graphs to determine unqualified subsystem causal graphs; deleting unqualified subsystem causal graphs and retaining qualified subsystem causal graphs; synthesizing qualified subsystem causal graphs using a distance-based weight determination algorithm to obtain a synthetic causal graph.

[0092] In some embodiments, the deviation degree of the subsystem causal graph includes: the deviation degree of the initial probability vector of the subsystem causal graph and the deviation degree of the causal strength matrix. Wherein, screening based on the deviation degrees of multiple subsystem causal graphs to determine unqualified subsystem causal graphs includes: calculating the deviation degree of the initial probability vector of each subsystem causal graph; calculating the deviation degree of the causal strength matrix of each subsystem causal graph; calculating according to the deviation degree of the initial probability vector of each subsystem causal graph and the deviation degree of the causal strength matrix to obtain the deviation degree of each subsystem causal graph; screening according to the deviation degree of each subsystem causal graph to determine unqualified subsystem causal graphs.

[0093] Among them, screening is performed according to the deviation degree of each subsystem cause-effect diagram, and determining unqualified subsystem cause-effect diagrams includes: comparing the deviation degree of each subsystem cause-effect diagram with a reference value; if the deviation degree of the subsystem cause-effect diagram is greater than the reference value, the subsystem cause-effect diagram is determined to be an unqualified subsystem cause-effect diagram, and the unqualified subsystem cause-effect diagram is screened out.

[0094] Specifically, the standardization of the initial probability vector is as follows: suppose there are n nodes involved in the system to be studied, and m subsystem causal graphs are constructed for this system. Among them, the number of nodes involved in some subsystem causal graphs is less than n. The initial probability vectors of these m subsystem causal graphs are expanded into n-dimensional column vectors, and the probabilities corresponding to the intermediate events (referring to other result events except the top event, which are located between the top event and the bottom event) and the logic gate event nodes are set to 0. Fill the positions of the nodes that do not appear in the subsystem causal graph with 0, and arrange the m initial probability vectors in a uniform node order. This process is called the standardization of the initial probability vectors of the subsystem causal graph.

[0095] The standardization of the causal strength matrix is ​​as follows: suppose there are n nodes involved in the system to be studied, and m subsystem causal graphs are proposed for this system, some of which involve less than n nodes. The causal strength matrices of these m subsystem causal graphs are expanded to n-order square matrices, and the rows and columns of the nodes that do not appear in the causal graphs are filled with 0, and the rows and columns of these m causal strength matrices are arranged in a unified node order (consistent with the arrangement in the standardized initial probability vector). This process is called the standardization of the causal strength matrix.

[0096] The synthesis of different subsystem causal diagrams needs to eliminate unreasonable components. It is assumed that the experts meet the following conditions: 1) The experts have subjective consistency requirements for the mutual influence and causal relationship between things, but there are inconsistencies objectively; 2) The experts' understanding of the mutual influence and causal relationship between the same things objectively has a consistent trend and satisfies the normal distribution. On the premise of meeting the above assumptions, let's discuss how to integrate the knowledge of experts in the subsystem causal diagram, eliminate noise, and delete unreasonable components, so that the system causal diagram (i.e., synthetic causal diagram) can truly reflect the system.

[0097] Specifically, let the mean of the initial probability vector be: PV av , which can be expressed as:

[0098]

[0099] The mean of the causal strength matrix is: CS av , which can be expressed as:

[0100]

[0101] set up in Then it is called is the deviation of the initial probability vector of the subsystem causal graph constructed by expert h; where v i is the causal vector, av i is a matrix vector, ε ij Bias vector.

[0102] in Then it is called is the deviation of the causal strength matrix of the subsystem causal graph constructed by expert h.

[0103] Specifically, the deviation of the initial probability vector is Then the deviation degree of the initial probability direction of the subsystem causal graph constructed by expert h is: DPV h Among them, DPV h It can be expressed as:

[0104] The deviation of the causal strength matrix is: Then the deviation degree of the causal intensity matrix of the subsystem causal graph constructed by expert h is: DCS h Among them, DCS h It can be expressed as,

[0105] is the degree of deviation of the initial probability vector of the system causal graph constructed by m experts.

[0106] is the degree of deviation of the causal strength matrix of the system causal graph constructed by m experts.

[0107] CT h =DPV h *DCS h , is the deviation degree of the subsystem causal graph constructed by expert h.

[0108] is the degree of deviation of the system causal graph constructed by m experts.

[0109] Subsystem Cause and Effect Diagram D h The deviation reflects the difference between the causal graph constructed by expert h and the average causal graph constructed by all experts. The larger the deviation, the greater the difference between the causal graph constructed by expert h and the average causal graph constructed by all experts. This causal graph is less likely to reflect reality.

[0110] If the subsystem causal diagram D h If the deviation is close to 0, it means that expert h and other experts have a similar understanding of the system. If the deviation is equal to 0, it means that expert h and other experts have a completely consistent understanding of the system.

[0111] In general, the degree of deviation cannot be 0. A very small δ (0<δ<1) can be taken. If the degree of deviation is less than δ, it is considered that the experts have a completely consistent understanding of the system. Among them, δ is the reference value.

[0112] Specifically, when each subsystem cause-effect diagram is screened through the degree of deviation, unqualified subsystem cause-effect diagrams can be determined, and then the unqualified subsystem cause-effect diagrams are deleted, and finally the qualified subsystem cause-effect diagrams are synthesized and connected into a synthetic cause-effect diagram.

[0113] Specifically, Figure 3 As shown, k represents the number of qualified subsystem causal graphs.

[0114] Let PV = (pv i ) n×1 , CS=(cs ij ) n×n , In order to ensure the reliability of the synthetic causal graph, the present invention proposes a "distance-based weight determination algorithm" for synthesis. The following takes the initial probability vector as an example for explanation:

[0115] a) Definition: D i The initial probability vector PV iThe total distance to the initial probability vector of the causal graph of other subsystems is:

[0116]

[0117] According to the Euclidean distance formula:

[0118]

[0119] It can be seen that d i The smaller it is, the smaller the difference between the initial probability values ​​of the sub-causal graph and other causal graphs is, the higher the support from other experts is, and the greater the impact on the synthesized system causal graph is. i and m j is the distance vector, a ij and b ij are distance vector elements.

[0120] b) Assume d * =min(d i )i=1,2,…nie d * is the minimum value in the total distance. Here, d * As a key parameter, it has the greatest impact on the synthesized system cause-effect diagram.

[0121] c) The subjective evaluation method was used before, that is, the higher the authority of the expert, the higher the VS i The larger the weight value assigned.

[0122] Assume the subjective weight value is m i ;

[0123] Definition: PV i The relative objective weight of is:

[0124]

[0125] At this time, re-adjust PV i The weight value p i .Right now

[0126] p i =n i ×m i ;

[0127] After adjustment, the credibility of expert opinions that are far away from and have large differences with other initial probability vectors will be reduced.

[0128] Finally, the synthesized causal graph is obtained by synthesizing according to the readjusted weight values.

[0129] refer to Figure 4 , Figure 4 A principle block diagram of the complex system causal graph construction and synthesis system provided by the present invention.

[0130] Specifically, Figure 4 As shown, the complex system causal graph construction and synthesis system includes:

[0131] The data collection unit 401 is used to collect comprehensive data and construct multiple subsystem cause-and-effect diagrams based on the comprehensive data. The comprehensive data includes: internal / external events, RCA analysis data, experience feedback data, RCM maintenance guideline data and expert experience data.

[0132] The node analysis unit 402 is used to determine whether there are common nodes between multiple subsystem cause-effect graphs.

[0133] The first synthesis unit 403 is used to synthesize the multiple subsystem cause-effect graphs based on the first synthesis method to obtain a synthesized cause-effect graph of the complex system if there are no common nodes between the multiple subsystem cause-effect graphs.

[0134] The second synthesis unit 404 is used to synthesize the multiple subsystem cause-effect graphs based on the second synthesis method if there are common nodes between the multiple subsystem cause-effect graphs to obtain a synthesized cause-effect graph of the complex system.

[0135] Specifically, the specific coordination operation process between the units in the complex system causal graph construction and synthesis system can refer to the above-mentioned complex system causal graph construction and synthesis method, which will not be repeated here.

[0136] The following is an explanation of the generator hydrogen cooling system (GRH) of a nuclear power plant. Specifically, the function of the GRH is to use the water of the conventional island closed-circuit cooling water system (SRI) to cool the hydrogen circulating in the generator and the air circulating in the exciter.

[0137] The stator core and rotor of the invented machine are cooled by hydrogen. GRH is used to ensure the operation of the generator within the rated power range, and specifically completes the following functions: directly cool the stator core and rotor winding of the generator; fill the generator with hydrogen, nitrogen or air; maintain the rated hydrogen pressure during the operation of the generator and compensate for leakage. During normal operation, the hydrogen pressure in the generator is maintained between 0.48 and 0.52 MPa, and the pressure difference between the hydrogen pressure and the water pressure of the generator is maintained at 0.02 to 0.05 MPa to prevent the cooling water medium in the generator stator from leaking into the internal chamber of the generator, causing the insulation of the generator stator and rotor to decrease, grounding, short circuit and other faults, thereby damaging the generator body; the purity of hydrogen in the generator is required to be greater than 98%, the humidity is not greater than 25%, and the oxygen content in hydrogen should not exceed 0.8%.

[0138] 1) Taking the generator hydrogen cooling system as the research object, the causal diagram of the four subsystems constructed is as follows Figures 5 to 8 shown.

[0139] in, Figures 5 to 8 middle:

[0140] B1: Scaling on the cooling water side of the hydrogen cooler (101RF); B2: The cooling water inlet pipe of the hydrogen cooler is broken (101RF); B3: The cooling water inlet valve of the hydrogen cooler is mistakenly closed; B4: The cooling water outlet valve of the hydrogen cooler is mistakenly closed;

[0141] X1: 101 hydrogen cooler cold hydrogen temperature GRH411MT_XQ01; X2: 101 hydrogen cooler cold hydrogen temperature high alarm GRH411MT_XH01; X3: Generator 39 core temperature GRH039MT_XQ01; X4: Generator 40 core temperature GRH040MT_XQ01; X5: Generator 41 core temperature GRH041MT_XQ01; X6: Generator 42 core temperature GRH042MT_XQ01; X7: Generator 39 core temperature high alarm GRH039MT_XH01; X8: Generator 40 core temperature high alarm GRH040MT_XH01; X9: Generator 41 core temperature high alarm GRH041MT_XH01; X 10 :Generator 42 core temperature high alarm GRH042MT_XH01; X 11 :Generator 47 core temperature GRH047MT_XQ01; X 12 :Generator No. 48 core temperature GRH048MT_XQ01; X 13 :Generator 47 core temperature high alarm GRH047MT_XH01; X 14 :Generator 48 core temperature high alarm GRH048MT_XH01; X 15 :Generator 56 core temperature GRH056MT_XQ01; X 16 :Generator 57 core temperature GRH057MT_XQ01; X 17 :Generator 56 core temperature high alarm GRH056MT_XH01; X 18 : Generator No. 57 core high temperature alarm GRH057MT_XH01.

[0142] 2) Normalize the initial probability vectors and causal strength matrices of D1, D2, D3, and D4:

[0143] The node n of D1 is 6. After calculation, we can get:

[0144] PV1=(0.05,0.02,0.01,0.01,0,0);

[0145]

[0146] The node n of D2 is 10, and after calculation, we can get:

[0147] PV2=(0.055,0.025,0,0,0,0,0,0,0,0);

[0148]

[0149] The node n of D3 is 8, and after calculation, we can get:

[0150] PV3=(0.045,0.015,0,0,0,0,0,0);

[0151]

[0152] The node n of D4 is 6, and after calculation, we can get:

[0153] PV4 = (0.025, 0.02, 0, 0, 0, 0);

[0154]

[0155] standardization:

[0156] N = 22, after standardization we can get:

[0157]

[0158]

[0159]

[0160]

[0161] Similarly, the normalization of the causal strength matrix is ​​the same as the normalization principle of the initial probability vector. 3) Calculate the deviation degree of D1, D2, D3 and D4 and the deviation degree of the synthetic causal graph D.

[0162] Mean:

[0163]

[0164] deviation:

[0165]

[0166]

[0167]

[0168]

[0169] Deviation: 2 norm formula calculation:

[0170] DPV1 = 0.0081;

[0171] DPV2 = 0.0167;

[0172] DPV3 = 0.020;

[0173] DPV4 = 0.047;

[0174] VD=0.02295.

[0175] 4) Qualified subsystem cause-effect diagram:

[0176] From the above calculation, we can see that the deviation of the causal graphs of the four subsystems is less than δ (0<δ<1), so the causal graphs of the four subsystems are all qualified. Therefore, D1, D2, D3, and D4 are synthesized.

[0177] 5) Synthesis:

[0178] The “distance-based weight determination method” is used to calculate the initial probability vector PV and causal strength matrix CS of D.

[0179] The initial probability vector PV of D is calculated as follows:

[0180] d1=0.055;

[0181] d2=0.106;

[0182] d3=0.111;

[0183] d4=0.168;

[0184] Then d * =0.055;

[0185] The original subjective weight values ​​of PV1, PV2, PV3, and PV4 are:

[0186] m1=0.5, m2=0.2, m3=0.2, m4=0.1;

[0187] PV i The relative objective weight value of is:

[0188]

[0189]

[0190]

[0191]

[0192] Then PV i The weight value is:

[0193] p1=n1×m1=0.5;

[0194] p2 = n2 × m2 = 0.104;

[0195] p3 = n3 × m3 = 0.1;

[0196] p4 = n4 × m4 = 0.051;

[0197] Initial probability of B1 = 0.03522;

[0198] Initial probability of B2 = 0.013875;

[0199] Initial probability of B3 = 0.0065;

[0200] Initial probability of B4 = 0.00602;

[0201] Then the initial probability vector of the system causal diagram is:

[0202]

[0203] The calculation of the causal strength matrix CS of D is similar.

[0204] The final synthetic causal diagram is as follows Fig. 9 shown.

[0205] The complex system causal graph construction and synthesis method of the present invention synthesizes the subsystem causal graph, which has the following advantages compared with the fault tree synthesis:

[0206] The fault tree is static, representing a tree structure with certain relationships; while the cause-effect diagram is dynamic, representing a network structure with uncertain relationships. For large and complex systems such as nuclear power plants, there is usually no simple one-to-one correspondence between the cause of the fault and the abnormal signal, but often has characteristics such as uncertainty and delay. Therefore, the cause-effect diagram is more suitable for the cause-effect logic representation of complex systems;

[0207] The fault tree starts from the top event and finds the cause of the top event step by step, thus forming an inverted tree-like relationship diagram; the causal diagram does not restrict the topological structure of the network in terms of expression, and allows the appearance of directed loops in the network, which can express causal structures such as negative feedback and positive feedback. In this way, when constructing causal diagrams for complex systems, experts can use the method of tracing the cause from the effect or the method of tracing the effect from the cause. For situations with causal loops, multi-connected situations, and signal jumps, the causal diagram can also be used to fully and accurately express them;

[0208] All causal relationships in the fault tree are considered to be inevitable; the causal relationships in the causal diagram are represented by causal strength variables, which correspond to the knowledge structure in the expert's mind, and the expert can then give a practical setting value. Referring to our above-mentioned algorithm, combined with objective weights, the strength of the synthesized causal diagram is evaluated for reliability to ensure that the causal strength of the complex system is more in line with reality;

[0209] Fault trees use binary assumptions, but causal diagrams are not limited to binary values ​​and can be multi-valued. For complex systems, variables often have more than two states, and causal diagrams can fully and accurately describe such multi-valued variables.

[0210] Furthermore, the present invention adopts many different subsystem causal graphs, integrates expert opinions, and concentrates expert wisdom and experience data to provide conditions for accurately and comprehensively constructing system causal graphs. Through deviation calculation, unqualified sub-causal graphs are eliminated to ensure the accuracy of the synthesized system causal graph; the distance-based weight determination algorithm is used to calculate the initial probability vector and causal strength of the system causal graph. It has been verified that the calculated parameter values ​​are more in line with reality, providing sufficient conditions for subsequent fault diagnosis applications.

[0211] In addition, the present invention can also optimize and adjust the causal graph (including causal relationships, probability parameter optimization, etc.) based on information such as experience feedback from field experts who pay attention to power plants, and realize dynamic maintenance of the causal graph.

[0212] In addition, an electronic device of the present invention includes a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program to implement a complex system causal diagram construction and synthesis method such as any one of the above. Specifically, according to an embodiment of the present invention, the process described with reference to the flowchart above can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed by an electronic device and when executed, it performs the above functions defined in the method of the embodiment of the present invention. The electronic device in the present invention can be a terminal such as a notebook, a desktop, a tablet computer, a smart phone, or a server.

[0213] In addition, a storage medium of the present invention stores a computer program, and when the computer program is executed by a processor, it implements any of the above complex system cause-and-effect diagram construction and synthesis methods. Specifically, it should be noted that the above storage medium of the present invention can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer readable signal media may also be any computer readable medium other than computer readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0214] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0215] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0216] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0217] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented directly using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0218] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot limit the scope of protection of the present invention. All equivalent changes and modifications made to the scope of the claims of the present invention should fall within the scope of the claims of the present invention.

Claims

1. A method for constructing and synthesizing a causal graph of a complex system, characterized in that: The following steps are involved: Collecting comprehensive data and constructing multiple subsystem cause-and-effect diagrams based on the comprehensive data; the comprehensive data includes: internal / external events, RCA analysis data, experience feedback data, RCM maintenance guideline data, and expert experience data; Determine whether there are common nodes among the multiple subsystem cause-and-effect graphs; If there are no common nodes between the multiple subsystem cause-and-effect graphs, synthesizing the multiple subsystem cause-and-effect graphs based on the first synthesis method to obtain a synthesized cause-and-effect graph of the complex system; If there are common nodes between the multiple subsystem cause-effect graphs, the multiple subsystem cause-effect graphs are synthesized based on the second synthesis method to obtain a synthesized cause-effect graph of the complex system.

2. The complex system causal graph construction and synthesis method according to claim 1, characterized in that: The collecting of comprehensive data and constructing multiple subsystem cause-and-effect diagrams based on the comprehensive data includes: Performing comprehensive data analysis on the comprehensive data to obtain comprehensive data analysis results; Determine the failure mode of the complex system and the alarm signal that triggers the failure reasoning based on the comprehensive analysis results of the data; The subsystem cause-and-effect diagram is constructed based on the failure mode or the alarm signal.

3. The complex system causal graph construction and synthesis method according to claim 1, characterized in that: The synthesizing the plurality of subsystem cause-effect graphs based on the first synthesis method to obtain a synthesized cause-effect graph of the complex system includes: Obtaining an initial probability vector synthesis formula and a causal strength matrix synthesis formula; Synthesize the multiple subsystem causal graphs according to the initial probability vector synthesis formula and the causal strength matrix synthesis formula to obtain a synthetic causal graph of the complex system; The initial probability vector synthesis formula is: PV=(PV1,PV2,...PV m ); The causal intensity matrix synthesis formula is: Among them, PV1, PV2, ...PV m are the initial probability vectors corresponding to the causal graphs of m subsystems; CS1, CS2, ...CS m are the causal intensity matrices corresponding to the causal graphs of m subsystems.

4. The complex system causal graph construction and synthesis method according to claim 1, characterized in that: The synthesizing the plurality of subsystem cause-effect graphs based on the second synthesis method to obtain the synthesized cause-effect graph of the complex system comprises: Normalizing the initial probability vectors and causal strength matrices of the plurality of subsystem causal graphs; After completing the normalization, calculating the deviation degree of the cause-effect diagrams of the plurality of subsystems; The synthesized causal graph is obtained by synthesizing the causal graphs of the multiple subsystems based on the deviation degrees of the causal graphs.

5. The complex system causal graph construction and synthesis method according to claim 4, characterized in that: The synthesizing based on the deviation degrees of the plurality of subsystem causal graphs and the deviation degree of the synthesized causal graph to obtain the synthesized causal graph comprises: Screening based on the deviations of the plurality of subsystem cause-effect diagrams to determine unqualified subsystem cause-effect diagrams; Deleting the unqualified subsystem cause-and-effect diagram and retaining the qualified subsystem cause-and-effect diagram; The qualified subsystem cause-effect graph is synthesized by using a distance-based weight determination algorithm to obtain the synthesized cause-effect graph.

6. The complex system causal graph construction and synthesis method according to claim 5, characterized in that: The deviation degree of the subsystem causal graph includes: the deviation degree of the initial probability vector of the subsystem causal graph and the deviation degree of the causal strength matrix; The screening based on the deviation degree of the plurality of subsystem cause-effect diagrams to determine unqualified subsystem cause-effect diagrams includes: Calculate the degree of deviation of the initial probability vector of each subsystem causal graph; Calculate the degree of deviation of the causal strength matrix of each subsystem causal graph; The deviation degree of each subsystem causal graph is obtained by calculating the deviation degree of the initial probability vector of each subsystem causal graph and the deviation degree of the causal strength matrix; Screening is performed based on the degree of deviation of each subsystem cause-effect diagram to determine unqualified subsystem cause-effect diagrams.

7. The complex system causal graph construction and synthesis method according to claim 6, characterized in that: The screening according to the deviation degree of each subsystem cause-effect graph to determine the unqualified subsystem cause-effect graph comprises: comparing the degree of deviation of each subsystem cause-effect diagram with a reference value; If the deviation of the subsystem cause-effect diagram is greater than the reference value, the subsystem cause-effect diagram is determined to be an unqualified subsystem cause-effect diagram, and the unqualified subsystem cause-effect diagram is screened out.

8. A complex system causal graph construction and synthesis system, characterized by: include: A data collection unit, used for collecting comprehensive data and constructing a plurality of subsystem cause-and-effect diagrams based on the comprehensive data; The comprehensive data includes: internal / external events, RCA analysis data, experience feedback data, RCM maintenance guideline data and expert experience data; A node analysis unit, used to determine whether there are common nodes between the multiple subsystem cause-and-effect graphs; A first synthesis unit, configured to synthesize the plurality of subsystem causal graphs based on a first synthesis method to obtain a synthesized causal graph of a complex system if there are no common nodes between the plurality of subsystem causal graphs; The second synthesis unit is used to synthesize the multiple subsystem cause-effect graphs based on a second synthesis method if there are common nodes between the multiple subsystem cause-effect graphs to obtain a synthesized cause-effect graph of the complex system.

9. A storage medium, characterized in that: The storage medium stores a computer program, which is suitable for being loaded by a processor to execute the steps of the complex system causal graph construction and synthesis method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that: It comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the steps of the complex system causal graph construction and synthesis method as claimed in any one of claims 1 to 7 by calling the computer program stored in the memory.

Images