A method for complex equipment fault diagnosis based on causal belief rule base

Abstract

The application discloses a complex equipment fault diagnosis method based on a causal confidence rule base, relates to the technical field of complex equipment fault diagnosis, and comprises the following steps: acquiring a causal relationship between faults and indexes in complex equipment; constructing a causal graph based on the causal relationship between the faults and the indexes in the complex equipment; constructing rules and initializing confidence based on the causal graph to obtain a causal confidence rule base; and performing fault diagnosis on the complex equipment based on the causal confidence rule base; and the application improves the confidence rule base based on the causal relationship, thereby improving the accuracy of complex equipment fault diagnosis.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of complex equipment fault diagnosis, and in particular to a complex equipment fault diagnosis method based on a causal belief rule base. BACKGROUND

[0002] A belief rule base (BRB) is a rule base modeling method based on fuzzy logic. In 2006, Yang et al. proposed a belief rule base inference methodology using the evidential reasoning approach (RIMER), which introduced a belief framework into traditional production rules. The RIMER method is based on the Dempster-Shafer evidence theory, decision theory, fuzzy theory, and traditional production rules. As an important part of the RIMER method, the BRB is a generalization of the traditional fuzzy rule base (FRB) and can provide a more reliable description of the knowledge in engineering systems. At present, the BRB model has developed rapidly and has been applied in many fields such as fault detection, medicine, engineering system safety assessment, and complex system modeling. A BRB system is composed of a series of rules, and its essence is a kind of expert system that can effectively utilize various types of information to establish a nonlinear model between inputs and outputs. A basic rule base can be composed of many simple IF-THEN rules, and a belief rule can be obtained by adding a belief degree to the result part while considering the premise attribute weight and rule weight. A series of belief rules are combined to form a belief rule base.

[0003] The belief rule base has been widely applied in the field of complex equipment fault diagnosis. However, in the process of constructing the BRB, some rules may have pseudo-correlation due to the neglect of the influence of potential third-party factors. Specifically, the limitations of expert experience, the lack of mechanism knowledge, and the randomness of sample data can lead to pseudo-correlation between the premise attributes and the conclusions of the BRB rules, thereby to some extent, destroying the explainability of the BRB rules. The existing belief rule base itself may lead to pseudo-correlation between the indicators and the faults only due to data fitting during complex equipment fault diagnosis, resulting in low precision of complex equipment fault diagnosis. SUMMARY

[0004] The purpose of the present application is to provide a complex equipment fault diagnosis method based on a causal belief rule base, which improves the belief rule base based on causal relationships, thereby improving the precision of complex equipment fault diagnosis.

[0005] To achieve the above object, the application provides the following scheme.

[0006] The application provides a complex equipment fault diagnosis method based on a causal confidence rule base, comprising:

[0007] Obtaining a causal relationship between faults and indexes in complex equipment;

[0008] Constructing a causal graph based on the causal relationship between faults and indexes in complex equipment;

[0009] Constructing rules and initializing confidence based on the causal graph to obtain a causal confidence rule base;

[0010] Performing fault diagnosis on complex equipment based on the causal confidence rule base.

[0011] Optionally, the causal relationship comprises causal relationships corresponding to a plurality of fault types.

[0012] The causal relationship corresponding to any fault type comprises a plurality of cause indexes and a plurality of effect indexes corresponding to the fault type; the change of the cause indexes leads to the occurrence of the corresponding fault; the occurrence of the fault leads to the change of the corresponding plurality of effect indexes.

[0013] Optionally, performing fault diagnosis on complex equipment based on the causal confidence rule base comprises:

[0014] Let the first iteration number i be 1;

[0015] Determine that all rules corresponding to the i th premise attribute in the causal confidence rule base are current rules; the premise attribute is a cause index or an effect index;

[0016] Obtain the number M of current rules;

[0017] Let the second iteration number m be 1;

[0018] Determine the matching degree of the i th premise attribute in the m th current rule based on the reference value of the i th premise attribute;

[0019] Increase the value of the second iteration number m by 1, and return to the step of determining the matching degree of the i th premise attribute in the m th current rule based on the reference value of the i th premise attribute until the value of the second iteration number m is equal to the number M of current rules, to obtain the matching degree of the i th premise attribute in each corresponding rule;

[0020] Increase the value of the first iteration number i by 1, and return to the step of determining that all rules corresponding to the i th premise attribute in the causal confidence rule base are current rules until all premise attributes are traversed, to obtain the matching degree of different premise attributes in each corresponding rule;

[0021] An activation weight of each rule is determined based on a matching degree of each premise attribute in the corresponding rule;

[0022] A confidence of each conclusion is determined based on the activation weight of different rules; the conclusion corresponds to a fault type one by one;

[0023] A fault diagnosis result is determined based on the confidence of different conclusions.

[0024] Optionally, the matching degree is:

[0025] wherein, is a matching degree of the ith premise attribute in the mth current rule; is a reference value of the ith premise attribute in the n+1th rule; is a reference value of the ith premise attribute in the nth rule;x i is the ith premise attribute.

[0026] Optionally, the activation weight is:

[0027] wherein, w m is the activation weight of the mth rule; θ m is a rule weight of the mth rule; M1 is a number of premise attributes; δ i is a weight of the ith premise attribute; N is a number of rules; θ n is a rule weight of the nth rule; is a matching degree of the ith premise attribute in the nth current rule.

[0028] Optionally, the confidence of the conclusion is:

[0029]

[0030] B=D-E;

[0031]

[0032] wherein, β l is the confidence corresponding to the lth conclusion; A, B, C, D, E are all intermediate parameters; L is a number of conclusions; β l,m is the confidence corresponding to the lth conclusion of the mth rule; β i,m is the confidence corresponding to the ith conclusion of the mth rule.

[0033] Optionally, the fault diagnosis result is:

[0034]

[0035] wherein, Output is the fault diagnosis result; D l is the lth conclusion.

[0036] According to the specific embodiments provided in the application, the following technical effects are disclosed in the application:

[0037] The application provides a complex equipment fault diagnosis method based on a causal belief rule base. The object and index are determined, a causal graph is constructed, the rule and initial confidence are constructed through the causal graph, the matching degree of the premise attribute and reference value is calculated, the activation weight of the rule is calculated, the confidence of the final evaluation result is calculated by using an ER analysis algorithm, the final result is obtained, and a causal machine is added in the BRB. The causal machine filters the rules by using the causal graph to obtain the rules with causal relationship, effectively reduces the number of rules in the model, reduces the complexity of the model, improves the performance of the model by reasonably improving the reference level of the premise attribute, reasonably reduces the number of rules to prevent combination explosion, and improves the interpretability of the belief rule base. The new belief rule base with causal relationship is obtained by constructing the causal graph, the interpretability of the model is improved, the performance is not affected when the rules are filtered, and the precision, interpretability and efficiency of the complex equipment fault diagnosis are improved. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can be obtained based on these drawings without creative labor.

[0039] Figure 1 FIG. 1 is a flow chart of a complex equipment fault diagnosis method based on a causal belief rule base according to an embodiment of the application;

[0040] Figure 2 FIG. 2 is a basic structure diagram of a traditional belief rule base according to an embodiment of the application;

[0041] Figure 3 FIG. 3 is a basic structure diagram of a causal belief rule base according to an embodiment of the application;

[0042] Figure 4 FIG. 4 is a schematic diagram of a reasoning process of a causal belief rule base according to an embodiment of the application. DETAILED DESCRIPTION

[0043] With reference to the drawings and embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0044] The above-mentioned purposes, features and advantages of the present application will be more apparent and understandable, and the present application will be described in further detail below with reference to the drawings and specific embodiments.

[0045] In an exemplary embodiment, as shown in Figure 1 a complex equipment fault diagnosis method based on a causal confidence rule base is provided, comprising:

[0046] Step 101: Obtain the causal relationship between faults and indicators in complex equipment. The causal relationship includes the causal relationship corresponding to multiple fault types. The causal relationship corresponding to any fault type includes multiple cause indicators and multiple effect indicators corresponding to the fault type. The change of the cause indicator leads to the occurrence of the corresponding fault. The occurrence of the fault leads to the change of the corresponding multiple effect indicators.

[0047] Identify the object and the indicator: Before diagnosing the fault of the complex equipment, the specific equipment, i.e., the object, needs to be identified first, and then the indicators related to the occurrence of the fault or the indicators changed due to the occurrence of the fault are selected. Before building the rule base, the specific problem needs to be analyzed first, for example, when diagnosing the fault of the loosening of the fiber flange interface, the indicators that may cause the fault are determined first, or the indicators that change due to the occurrence of the fault are determined, so as to facilitate subsequent analysis and processing.

[0048] Step 102: Based on the causal relationship between faults and indicators in complex equipment, build a causal graph.

[0049] After the object and the indicator are identified, the relationship between the indicator and the fault is established. The causal relationship between the indicator and the fault is determined by analyzing the mechanism of the object and the indicators obtained in step 1, and the causal graph is initially built by using expert knowledge, and then the causal graph is analyzed and improved.

[0050] Step 103: Build rules and initialize confidence based on the causal graph, and obtain a causal confidence rule base. The rules are built according to the indicators selected by the causal graph, and the initial confidence is given. In the selection process of the causal graph, the causal strength between certain premise attributes and conclusions is less than a given threshold, and such rules are deleted, and the selected rule base (i.e., the causal confidence rule base) is obtained. The basic structure of the confidence rule base before selection (i.e., the traditional confidence rule base) is as shown in Figure 2 ; and the basic structure of the causal confidence rule base is as shown in Figure 3As shown, the reasoning process of the causal confidence rule base is as follows: Figure 4 As shown.

[0051] Figure 2 In the given text, x1 is the first prerequisite attribute, x2 is the second prerequisite attribute, and x... M Let δ1 be the weight of the first premise attribute, δ2 be the weight of the second premise attribute, and δ... M Let R1 be the weight of the Mth prerequisite attribute, R2 be the weight of the first rule, and R be the weight of the second rule. k For the k-th rule, R L Let L be the Lth rule, θ1 be the rule weight of the 1st rule, θ2 be the rule weight of the 2nd rule, and θ... k Let θ be the rule weight of the k-th rule. L Let D be the rule weight of the Lth rule, D1 be the first conclusion, D2 be the second conclusion, and D... N This is the Nth conclusion. In a causal machine, each node represents a premise attribute / evaluation result or a reference level of a premise attribute / evaluation result. Each directed edge represents a causal relationship. The arrow indicates the direction of the causal relationship, that is, the node pointed to by the arrow represents the result of the causal relationship, the node from which the arrow originates represents the cause of the causal relationship, and the weight on the edge represents the causal strength. Figure 3 In the diagram, R′1 represents the first rule with a causal relationship after being filtered by the causal machine, and R′2 represents the second rule with a causal relationship after being filtered by the causal machine. j Let R′ represent the j-th rule with causal relationship after causal machine filtering. p Let θ' represent the rule weight of the p-th rule with a causal relationship after being filtered by the causal machine; θ'1 represents the rule weight of the first rule with a causal relationship after being filtered by the causal machine; θ'2 represents the rule weight of the second rule with a causal relationship after being filtered by the causal machine; θ' j Let θ′ represent the rule weight of the j-th rule with causal relationship after causal machine filtering; p Let p represent the rule weight of the p-th rule that has a causal relationship after being filtered by the causal machine; Figure 4 In the diagram, c1 represents the first dependent attribute (dependent indicator), c2 represents the third dependent attribute, and c... m Let e1 represent the m-th cause attribute, e2 represent the 1st effect attribute (effect index), and e3 represent the 3rd effect attribute. n p represents the nth fruit attribute. 11 p represents the causal strength between the first cause attribute and the first effect attribute. 12 p represents the causal strength between the first causal attribute and the second effect attribute. 1nPi,j represents the causal strength between the jth cause attribute and the ith effect attribute, p 21 Pi,j represents the causal strength between the jth cause attribute and the ith effect attribute, p 22 Pi,j represents the causal strength between the jth cause attribute and the ith effect attribute, p 2n Pi,j represents the causal strength between the jth cause attribute and the ith effect attribute, p m1 Pi,j represents the causal strength between the jth cause attribute and the ith effect attribute, p m2 Pi,j represents the causal strength between the jth cause attribute and the ith effect attribute, p mn Pi,j represents the causal strength between the jth cause attribute and the ith effect attribute, p

[0052] Step 104: based on the causal belief rule base, the complex equipment is diagnosed.

[0053] Step 104, comprising:

[0054] Step 104-1: let the first iteration number i = 1.

[0055] Step 104-2: determine that all rules corresponding to the ith premise attribute in the causal belief rule base are the current rules. The premise attribute is the cause index or the effect index. The premise attribute corresponds to different cases. If it is diagnosis, the conclusion is the dependent variable, if it is prediction, the premise attribute is the independent variable: in diagnosis, because of the fault, the index changes, so the conclusion corresponds to the fault as the cause, and the index change corresponding to the premise attribute as the effect. If it is prediction, it is to judge what fault may occur through the index.

[0056] Step 104-3: get the number M of the current rules.

[0057] Step 104-4: let the second iteration number m = 1.

[0058] Step 104-5: based on the reference value of the ith premise attribute, determine the matching degree of the ith premise attribute in the mth current rule.

[0059] Step 104-6: let the value of the second iteration number m increase by 1, and return to step 104-5 until the value of the second iteration number m is equal to the number M of the current rules, to obtain the matching degree of the ith premise attribute in each corresponding rule.

[0060] Step 104-7: let the value of the first iteration number i increase by 1, and return to step 104-2 until all premise attributes are traversed, to obtain the matching degree of different premise attributes in each corresponding rule. The matching degree is:

[0061] wherein, is the matching degree of the i-th premise attribute in the m-th current rule. is the reference value of the i-th premise attribute in the n+1-th rule. is the reference value of the i-th premise attribute in the n-th rule. i is the i-th premise attribute.

[0062] Step 104-8: Based on the matching degrees of different premise attributes in each corresponding rule, the activation weight of each rule is determined respectively. The activation weight is:

[0063] wherein, w m is the activation weight of the m-th rule. m is the rule weight of the m-th rule. M1 is the number of premise attributes. i is the weight of the i-th premise attribute. N is the number of rules. n is the rule weight of the n-th rule. is the matching degree of the i-th premise attribute in the n-th current rule.

[0064] Step 104-9: Based on the activation weights of different rules, the confidence of each conclusion is determined respectively. The conclusion corresponds to the fault type one-to-one. The confidence of the conclusion is:

[0065]

[0066] B = D - E.

[0067]

[0068] wherein, β l is the confidence corresponding to the l-th conclusion. A, B, C, D, E are all intermediate parameters. L is the number of conclusions. l,m is the confidence corresponding to the l-th conclusion of the m-th rule. i,m is the confidence corresponding to the i-th conclusion of the m-th rule.

[0069] Step 104-10: Based on the confidence of different conclusions, the fault diagnosis result is determined. The fault diagnosis result is:

[0070]

[0071] wherein, Output is the fault diagnosis result. l is the l-th conclusion.

[0072] The application adds a causal machine into a belief rule base (BRB) fault diagnosis model, the causal machine adopts a causal graph to filter rules, and rules with a causal relationship are obtained, so that the number of rules is reasonably reduced to prevent combination explosion, and the interpretability of the belief rule base is improved. Advantages are that the number of rules in the model is effectively reduced, the complexity of the model is reduced, the performance of the model is improved by reasonably improving the reference level of the premise attribute, new belief rules with a causal relationship are obtained by constructing a causal graph, the interpretability of the model is improved, and the performance of the rules is not affected when the rules are filtered.

[0073] In an exemplary embodiment, a computer device, which can be a server or a terminal, is provided. The computer device includes a processor, a memory, an input / output interface (I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for running the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is configured to exchange information between the processor and external devices. The communication interface of the computer device is configured to communicate with external terminals through a network connection. The computer program is executed by the processor to implement a complex equipment fault diagnosis method based on a causal belief rule base.

[0074] In an exemplary embodiment, a computer readable storage medium storing a computer program is provided. The computer program is executed by a processor to implement the steps in the above method embodiments.

[0075] In an exemplary embodiment, a computer program product including a computer program is provided. The computer program is executed by a processor to implement the steps in the above method embodiments.

[0076] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0077] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0078] The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a blockchain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0079] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described, but as long as the combinations of the technical features do not exist, they should be considered as the scope of the present application.

[0080] The principles and implementation modes of the present application are described by applying specific examples herein. The above description of the embodiments is only used to help understand the method and its core idea of the present application; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range can be changed. In conclusion, the content of the present application should not be understood as a limitation.

Claims

1. A complex equipment failure diagnosis method of a causal belief rule base, characterized by, The method comprises the following steps: Obtaining the causal relationship between the faults and the indexes in the complex equipment; The causal relationship comprises the causal relationship corresponding to a plurality of fault types; The causal relationship corresponding to any fault type comprises a plurality of cause indexes and a plurality of effect indexes corresponding to the fault type; the change of the cause indexes leads to the occurrence of the corresponding fault; the occurrence of the fault leads to the change of the corresponding effect indexes; Constructing a causal graph based on the causal relationship between the faults and the indexes in the complex equipment; Constructing rules and initializing the confidence based on the causal graph to obtain a causal confidence rule base; constructing rules based on the indexes screened by the causal graph and giving the initial confidence; deleting the rules with the causal strength between the premise attributes and the conclusions less than a given threshold to obtain the causal confidence rule base; Performing fault diagnosis on the complex equipment based on the causal confidence rule base; The fault diagnosis on the complex equipment based on the causal confidence rule base comprises the following steps: Setting the first iteration number i to 1; Determining all the rules corresponding to the ith premise attribute in the causal confidence rule base as the current rules; the premise attribute is a cause index or an effect index; Obtaining the number M of the current rules; Setting the second iteration number m to 1; Determining the matching degree of the ith premise attribute in the mth current rule based on the reference value of the ith premise attribute; Increasing the value of the second iteration number m by 1 and returning to the step of determining the matching degree of the ith premise attribute in the mth current rule based on the reference value of the ith premise attribute until the value of the second iteration number m is equal to the number M of the current rules, thereby obtaining the matching degree of the ith premise attribute in each corresponding rule; Increasing the value of the first iteration number i by 1 and returning to the step of determining all the rules corresponding to the ith premise attribute in the causal confidence rule base as the current rules until all the premise attributes are traversed, thereby obtaining the matching degree of different premise attributes in each corresponding rule; Determining the activation weight of each rule based on the matching degree of different premise attributes in each corresponding rule; Determining the confidence of each conclusion based on the activation weight of different rules; the conclusion corresponds to a fault type one by one; Determining the fault diagnosis result based on the confidence of different conclusions.

2. The method of claim 1, wherein the method is characterized by, The matching degree is: ； wherein, is a matching degree of the i-th premise attribute in the m-th current rule; is a reference value of the i-th premise attribute in the m-th current rule; is a reference value of the i-th premise attribute in the m-th current rule; is a reference value of the i-th premise attribute in the m-th current rule; is a reference value of the i-th premise attribute in the m-th current rule; is the i-th premise attribute.

3. The method of claim 2, wherein the method further comprises: The activation weight is: ； wherein, is an activation weight for the mth rule; is a rule weight for the mth rule; is the number of premise attributes; is the weight of the ith premise attribute; N is the number of rules; is a rule weight for the nth rule; is the degree of match of the ith premise attribute in the nth current rule.

4. The method of claim 3, wherein the method further comprises: The confidence of the conclusion is: ； ； ； ； - ； ； in, For the first The confidence level corresponding to each conclusion; , , , , All are intermediate parameters; L is the number of conclusions; For the m-th rule The confidence level corresponding to each conclusion; For the m-th rule The confidence level corresponding to each conclusion.

5. The method of claim 4, wherein the method further comprises: The fault diagnosis result is: ； wherein, is a fault diagnosis result; is a first conclusion.

Images