阅读说明：本技术 一种维修辅助装备故障诊断方法及系统 (Fault diagnosis method and system for maintenance auxiliary equipment ) 是由 李英顺 周通 王德彪 郭占男 张杨 赵玉鑫 刘海洋 隋欢欢 于 2021-08-06 设计创作，主要内容包括：本发明公开了一种维修辅助装备故障诊断方法及系统。所述方法包括：获取待诊断装备的状态特征；由待诊断装备的状态特征确定待诊断装备产生各个故障征兆的可信度；由待诊断装备产生各个故障征兆的可信度、案例库中各个案例的故障征兆的可信度以及修正因子,计算待诊断装备与各案例的相似度；将相似度最大的案例对应的排故方案确定为待诊断装备的排故方案。本发明能提高故障诊断的准确性。(The invention discloses a fault diagnosis method and system for maintenance auxiliary equipment. The method comprises the following steps: acquiring state characteristics of equipment to be diagnosed; determining the credibility of each fault symptom generated by the equipment to be diagnosed according to the state characteristics of the equipment to be diagnosed; generating the credibility of each fault symptom, the credibility of the fault symptom of each case in the case library and a correction factor by the equipment to be diagnosed, and calculating the similarity of the equipment to be diagnosed and each case; and determining the fault elimination scheme corresponding to the case with the maximum similarity as the fault elimination scheme of the equipment to be diagnosed. The invention can improve the accuracy of fault diagnosis.)

1. A method for diagnosing a fault of a maintenance auxiliary equipment, comprising:

acquiring state characteristics of equipment to be diagnosed;

determining the credibility of each fault symptom generated by the equipment to be diagnosed according to the state characteristics of the equipment to be diagnosed;

generating the credibility of each fault symptom, the credibility of the fault symptom of each case in the case library and a correction factor by the equipment to be diagnosed, and calculating the similarity between the equipment to be diagnosed and each case;

and determining the fault elimination scheme corresponding to the case with the maximum similarity as the fault elimination scheme of the equipment to be diagnosed.

2. The method as claimed in claim 1, wherein the calculating the similarity between the equipment to be diagnosed and each case includes:

determining the reliability of the fault symptom of each case in the case base;

for case C in case base_pConfidence level, case C, of the ith fault sign generated by the equipment to be diagnosed_pThe reliability and the correction factor of the ith fault symptom are calculated, and the ith fault symptom and the case C generated by the equipment to be diagnosed are calculated_pSimilarity of the ith symptom of (1);

calculating the equipment to be diagnosed and the case C according to the similarity of all fault symptoms_pThe similarity of (c).

3. The method as claimed in claim 2, wherein the equipment to be diagnosed generates ith fault symptom and case C_pThe calculation formula of the similarity of the ith fault symptom is as follows:

D_si(C_ni,C_pi)＝β_i(1-|v(C_ni)-v(C_pi)|)；

wherein D is_si(C_ni,C_pi) For awaiting diagnosis equipment C_nGenerating ith fault symptom and case C_pSimilarity of the ith symptom of (1); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor;

4. the method as claimed in claim 2, wherein the equipment to be diagnosed and case C are used as the diagnosis tool_pThe calculation formula of the similarity is as follows:

wherein D is_s(C_n,C_p) For awaiting diagnosis equipment C_nAnd case C_pThe similarity of (2); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor; w_iFor awaiting diagnosis equipment C_nA weighting factor for the ith symptom of failure; m is the total number of fault symptoms;

5. a repair assistance equipment fault diagnosis system, comprising:

the data acquisition module is used for acquiring the state characteristics of the equipment to be diagnosed;

the reliability determining module is used for determining the reliability of each fault symptom generated by the equipment to be diagnosed according to the state characteristics of the equipment to be diagnosed;

the similarity calculation module is used for calculating the similarity between the equipment to be diagnosed and each case according to the credibility of each fault symptom generated by the equipment to be diagnosed, the credibility of the fault symptom of each case in the case library and the correction factor;

and the fault elimination scheme determination module is used for determining the fault elimination scheme corresponding to the case with the maximum similarity as the fault elimination scheme of the equipment to be diagnosed.

6. The system of claim 5, wherein the similarity calculation module specifically comprises:

the reliability determining unit is used for determining the reliability of the fault symptom of each case in the case base;

a symptom similarity calculation unit for case C in case base_pConfidence level, case C, of the ith fault sign generated by the equipment to be diagnosed_pThe reliability and the correction factor of the ith fault symptom are calculated, and the ith fault symptom and the case C generated by the equipment to be diagnosed are calculated_pSimilarity of the ith symptom of (1);

a case similarity calculation unit for calculating the equipment to be diagnosed and the case C according to the similarity of all fault symptoms_pThe similarity of (c).

7. The system of claim 6, wherein in the symptom similarity calculation unit, the equipment to be diagnosed generates the ith symptom and case C_pThe calculation formula of the similarity of the ith fault symptom is as follows:

D_si(C_ni,C_pi)＝β_i(1-|v(C_ni)-v(C_pi)|)；

wherein D is_si(C_ni,C_pi) For awaiting diagnosis equipment C_nGenerating ith fault symptom and case C_pSimilarity of the ith symptom of (1); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor;

8. a service assistance equipment fault diagnosis system according to claim 6 characterised in thatIn the case similarity calculation unit, the equipment to be diagnosed and case C_pThe calculation formula of the similarity is as follows:

wherein D is_s(C_n,C_p) For awaiting diagnosis equipment C_nAnd case C_pThe similarity of (2); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor; w_iFor awaiting diagnosis equipment C_nA weighting factor for the ith symptom of failure; m is the total number of fault symptoms;

Technical Field

The invention relates to the field of fault diagnosis, in particular to a fault diagnosis method and system for maintenance auxiliary equipment.

Background

Along with the development of equipment, the paper technical data manual increases the guarantee cost of the equipment, reduces the guarantee capability of the equipment, and is more and more difficult to meet the requirement of new equipment on maintenance guarantee. Portable Maintenance auxiliary equipment (PMA) not only can receive technical data information and Maintenance instruction, but also can provide fault information, spare part demand and the like for managers. The PMA realizes the connection between a maintenance site and a network and weaponry, and solves the problems.

At present, the fault diagnosis method of PMA has the following problems: when the difference between the number of the fault symptoms of the cases in the case base and the number of the fault symptoms of the cases to be tested is large and the number of the similar symptoms is small, the occurrence situation that the surfaces are dissimilar and actually similar or the surfaces are similar and actually dissimilar occurs. For example, when the reliability of the same fault symptom of the case to be detected and the case in the case base is 0, the similarity between the fault symptom of the case to be detected and the case in the case base obtained by the conventional fault diagnosis method is 1, which is obviously wrong. Therefore, the conventional fault diagnosis method has a problem of low accuracy.

Disclosure of Invention

Therefore, a need exists for a method and a system for fault diagnosis of a maintenance auxiliary device, so as to improve the accuracy of fault diagnosis.

In order to achieve the purpose, the invention provides the following scheme:

a method of service assistance equipment fault diagnosis, comprising:

acquiring state characteristics of equipment to be diagnosed;

determining the credibility of each fault symptom generated by the equipment to be diagnosed according to the state characteristics of the equipment to be diagnosed;

generating the credibility of each fault symptom, the credibility of the fault symptom of each case in the case library and a correction factor by the equipment to be diagnosed, and calculating the similarity between the equipment to be diagnosed and each case;

and determining the fault elimination scheme corresponding to the case with the maximum similarity as the fault elimination scheme of the equipment to be diagnosed.

Optionally, the calculating the similarity between the equipment to be diagnosed and each case by using the reliability of each fault symptom generated by the equipment to be diagnosed, the reliability of each fault symptom of each case in the case library, and the correction factor specifically includes:

determining the reliability of the fault symptom of each case in the case base;

for case C in case base_pConfidence level, case C, of the ith fault sign generated by the equipment to be diagnosed_pThe reliability and the correction factor of the ith fault symptom are calculated, and the ith fault symptom and the case C generated by the equipment to be diagnosed are calculated_pSimilarity of the ith symptom of (1);

calculating the equipment to be diagnosed and the case C according to the similarity of all fault symptoms_pThe similarity of (c).

Optionally, the equipment to be diagnosed generates the ith fault symptom and case C_pThe calculation formula of the similarity of the ith fault symptom is as follows:

D_si(C_ni,C_pi)＝β_i(1-|v(C_ni)-v(C_pi)|)；

wherein D is_si(C_ni,C_pi) For awaiting diagnosis equipment C_nGenerating ith fault symptom and case C_pSimilarity of the ith symptom of (1); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor;

optionally, the equipment to be diagnosed and case C_pThe calculation formula of the similarity is as follows:

wherein D is_s(C_n,C_p) For awaiting diagnosis equipment C_nAnd case C_pThe similarity of (2); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor; w_iFor awaiting diagnosis equipment C_nA weighting factor for the ith symptom of failure; m is the total number of fault symptoms;

the invention also provides a maintenance auxiliary equipment fault diagnosis system, which comprises:

the data acquisition module is used for acquiring the state characteristics of the equipment to be diagnosed;

the reliability determining module is used for determining the reliability of each fault symptom generated by the equipment to be diagnosed according to the state characteristics of the equipment to be diagnosed;

the similarity calculation module is used for calculating the similarity between the equipment to be diagnosed and each case according to the credibility of each fault symptom generated by the equipment to be diagnosed, the credibility of the fault symptom of each case in the case library and the correction factor;

and the fault elimination scheme determination module is used for determining the fault elimination scheme corresponding to the case with the maximum similarity as the fault elimination scheme of the equipment to be diagnosed.

Optionally, the similarity calculation module specifically includes:

the reliability determining unit is used for determining the reliability of the fault symptom of each case in the case base;

a symptom similarity calculation unit for case C in case base_pConfidence level, case C, of the ith fault sign generated by the equipment to be diagnosed_pThe reliability and the correction factor of the ith fault symptom are calculated, and the ith fault symptom and the case C generated by the equipment to be diagnosed are calculated_pSimilarity of the ith symptom of (1);

a case similarity calculation unit for calculating the equipment to be diagnosed and the case C according to the similarity of all fault symptoms_pThe similarity of (c).

Optionally, in the symptom similarity calculation unit, the equipment to be diagnosed generates the ith fault symptom and case C_pThe calculation formula of the similarity of the ith fault symptom is as follows:

D_si(C_ni,C_pi)＝β_i(1-|v(C_ni)-v(C_pi)|)；

wherein D is_si(C_ni,C_pi) For awaiting diagnosis equipment C_nGenerating ith fault symptom and case C_pSimilarity of the ith symptom of (1); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor;

optionallyIn the case similarity calculation unit, the equipment to be diagnosed and case C_pThe calculation formula of the similarity is as follows:

wherein D is_s(C_n,C_p) For awaiting diagnosis equipment C_nAnd case C_pThe similarity of (2); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor; w_iFor awaiting diagnosis equipment C_nA weighting factor for the ith symptom of failure; m is the total number of fault symptoms;

compared with the prior art, the invention has the beneficial effects that:

the embodiment of the invention provides a maintenance auxiliary equipment fault diagnosis method and system, wherein the reliability of each fault symptom, the reliability of each fault symptom of each case in a case library and a correction factor are generated by equipment to be diagnosed, the similarity between the equipment to be diagnosed and each case is calculated, and the troubleshooting scheme corresponding to the case with the maximum similarity is determined as the troubleshooting scheme of the equipment to be diagnosed.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.

FIG. 1 is a hardware framework diagram of a PMA;

FIG. 2 is a system management application software architecture diagram for a PMA;

FIG. 3 is a flow diagram of a prior art diagnostic process for a PMA;

FIG. 4 is a flow chart of a method for diagnosing a fault of a maintenance assistance device according to an embodiment of the present invention;

FIG. 5 is a diagram illustrating a primitive Bayesian network model according to an embodiment of the present invention;

FIG. 6 is a Bayesian network model diagram of fault information of an oil supply system according to an embodiment of the present invention;

FIG. 7 is a schematic diagram of an inference network architecture provided in an embodiment of the present invention;

fig. 8 is a schematic diagram of a specific process of matching the device to be diagnosed with the case library according to the embodiment of the present invention;

fig. 9 is a block diagram of a maintenance assistance equipment fault diagnosis system according to an embodiment of the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments.

After comprehensively considering the fault level, the complexity of the vehicle-mounted equipment subsystem and other conditions, the invention is explained based on the engine oil supply system of the vehicle-mounted equipment as an example of a diagnosis object of fault diagnosis. The process of diagnosing faults of complex equipment is actually a process of reasoning and obtaining a certain fault reason or certain fault reasons according to certain fault symptoms. Therefore, when the method is used for diagnosing faults of other systems of vehicle-mounted equipment, the fault diagnosis and the prediction of fault symptoms of other systems can be realized only by designing according to the steps of the fault diagnosis model construction of the engine oil supply system.

All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.

As a tool with interactive capability and automatic data processing capability, the portable maintenance auxiliary equipment (PMA) not only can meet the requirements of maintenance training, maintenance guidance and maintenance management capability of information equipment in different environments, but also can realize quick repair and remote technical support in battlefield environments, and the PMA is applied to greatly improve the guarantee capability of weapon equipment and has great application prospect in the field of complex equipment maintenance.

The hardware framework of the PMA is: a small screen display terminal, voice and video encryption, network reception, a large screen display terminal, a display driver, a computer motherboard, a basic interface, an adapter and the like, as shown in fig. 1.

As shown in fig. 2, the PMA system management application software mainly includes user basic information management, an Interactive Electronic Technical Manual (IETM) query system, a service technician basic information management system, a spare part tool query application system, a service technical document real-time recording and analyzing system, a service monitoring system, a remote technical support expert system, and a common technical fault diagnosis system.

Currently, the diagnostic procedure for PMA is as follows:

(1) the system transmits test and diagnosis information with the tested equipment through a special interface adapter, and the test information is interactively processed by an IETM software platform matched in the system to form a fault diagnosis strategy;

(2) automatically matching the obtained fault diagnosis strategy, matching the existing cases for maintenance and troubleshooting, and searching similar cases according to fault phenomena under the condition that the same cases are not matched, so as to provide one or more solutions for maintenance personnel;

(3) if the existing cases which are clearly indicated in the IETM are matched, the maintenance personnel are guided to carry out the next operation according to the maintenance steps of the existing maintenance cases;

(4) and if the fault information is more complex, case retrieval is carried out by using Bayesian network diagnosis. The collected original data information is classified, corresponding feature extraction rules are customized by using expert system knowledge, and feature vector extraction is carried out on the collected data information to generate a corresponding feature sample library. Extracting the test data information through the characteristic vector to serve as input data information of each fault diagnosis model, calculating the similarity of the input characteristic vector in different fault classification models, searching the case most similar to the new fault problem according to the calculated similarity of all cases, and extracting the cases to generate a preliminary troubleshooting scheme if the case similarity is greater than a certain threshold value. If there are no matching cases, the idea base is directly called to generate a preliminary troubleshooting plan. And finally, processing the fault diagnosis solution on the basis of the establishment of the preliminary fault diagnosis solution by a maintenance worker to obtain a final fault diagnosis solution. The diagnostic process for a PMA is shown in figure 3.

The fault diagnosis system in the IETM software platform mainly comprises three components: the system comprises a maintenance guide mode of an expert system, a fault automatic diagnosis mode and a remote maintenance cooperative support system.

The first important component of the expert system is a knowledge base which stores the expertise obtained from the expert in a certain field; the second component of the expert system is a reasoning machine, which has the capability of reasoning according to a certain strategy, i.e. can deduce conclusions according to knowledge, and can also guide maintenance personnel to perform troubleshooting according to the existing cases.

In the diagnosis process of PMA, the calculation method of case similarity is as follows:

the formula for calculating the similarity between the fault symptoms is as follows:

D_si(C_ni,C_pi)＝1-|v(C_ni)-v(C_pi)|；

wherein D is_si(C_ni,C_pi) For awaiting diagnosis equipment C_nGenerating ith fault symptom and case C_pSimilarity of the ith symptom of (1); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_nThe reliability of the ith symptom of failure.

The calculation formula of case similarity is as follows:

wherein D is_s(C_n,C_p) For awaiting diagnosis equipment C_nAnd case C_pThe similarity of (2); w_iFor awaiting diagnosis equipment C_nA weighting factor for the ith symptom of failure; m is the total number of the fault symptoms,

and combining the two formulas to obtain a calculation formula of case similarity:

the result obtained by the calculation formula of case similarity represents case C when the result is equal to 1.0_nAnd case C_pComplete matching; when the result is equal to 0.0, case C is represented_nAnd case C_pThere is no match at all.

For a case after matching, a similarity threshold lambda needs to be set according to expert experience, and when the similarity of the case is smaller than the similarity threshold lambda, the case in the case library is not enough to provide useful reference for a new case, so that the case does not need to be output to a user for reference.

The above formula for calculating case similarity has the following defects:

suppose that the equipment C to be diagnosed_nThe number of fault symptoms of (1) is M, case C_pWith a number of fault symptoms of N, equipment C to be diagnosed_nAnd case C_pThe number of symptoms with similarity in (1) is K, and there are always: k is less than or equal to MIN (M, N).

When K is equal to M and K is less than or equal to N, the equipment C to be diagnosed is represented_nAll the fault signs can be found in case C_pWhere similar fault symptoms are found, this may be case C_pIncludes a device C to be diagnosed_n. When K is equal to M and K is more than or equal to N, the equipment C to be diagnosed is represented_pAll the fault signs can be found in case C_nWhere similar fault symptoms are found, this may be case C_nIncludes a device C to be diagnosed_p. Therefore, when the number of symptoms is relatively large and the number of similar symptoms is relatively small, the similarity should be specifically analyzed to prevent the occurrence of the situation that the surfaces are not similar but actually similar or the situation that the surfaces are similar but actually dissimilar occurs.

The calculation formula of the similarity between the fault symptoms and the calculation formula of the case similarity are rewritten into the following form:

D_si(C_ni,C_pi)＝1-|v(C_ni)-v(C_pi)|；

as can be seen from the rewritten formula, the equipment C to be diagnosed_nAnd case C_pWhen the confidence levels of the same fault symptoms are all 0, the similarity between the fault symptoms and the cases is all 1, which is obviously wrong. Since the fault symptom itself is not reliable when the reliability of the fault symptom is 0 or less, an erroneous conclusion is necessarily caused. Therefore, in case matching, necessary corrections should be made to the calculation of the similarity.

Aiming at the defects, the invention introduces the correction factor into the similarity calculation process, solves the problems that the surfaces are dissimilar and actually similar or the surfaces are similar and actually dissimilar when the number of fault symptoms is larger and the similar symptoms are less, and improves the accuracy of fault diagnosis.

Fig. 4 is a flowchart of a method for diagnosing a fault of a maintenance auxiliary device according to an embodiment of the present invention.

Referring to fig. 4, the method for diagnosing a fault of a repair assist device of the present embodiment includes:

step 101: and acquiring the state characteristics of the equipment to be diagnosed.

The working principle of the engine oil supply subsystem is that gasoline which is filtered is continuously conveyed, and combustible mixed gas with a certain quantity and concentration is prepared according to the requirements of various working conditions of the engine and is supplied to a cylinder for combustion and work. Common fault phenomena of an oil supply system include engine shake, black smoke emission, insufficient engine power, abnormal starting and the like.

Therefore, in the present embodiment, the status characteristics regarding whether the engine is faulty or not can be acquired as follows: the engine has the disadvantages of shaking idling, poor acceleration, incapability of starting, difficulty in starting, easiness in flameout, reduction in the maximum speed per hour of the vehicle, high fuel consumption of the vehicle, black smoke emission of an exhaust pipe and the like.

The reason for the occurrence of the idling shake of the engine is that the single-cylinder non-oil-injection air filter is blocked due to the damage of the single-cylinder oil injector; the cause of poor engine acceleration may be a variety of conditions including: the single-cylinder non-oil injection air filter is blocked, the oil injection quantity is less, the oil injection quantity is not generated, the oil injection quantity is more, and the like; the reason for the inability to start is that there is no amount of fuel injected due to a failure of an Electronic Control Unit (ECU) or no pumping of oil by an oil pump; the reasons for the starting difficulties are mainly that the fuel injection quantity is too large or too small; the highest speed per hour of the vehicle is reduced because the single-cylinder air filter which does not spray oil is blocked or the fuel injection quantity is less, so that the power of the engine is insufficient; the reason why the exhaust pipe emits black smoke is that the amount of fuel injection is excessive.

According to the analysis of some statistical data, a common fault table of an engine oil supply system is summarized, and is shown in a table 1. The state feature vectors that can be extracted according to the state features include: the single-cylinder non-oil injection air filter is blocked, the oil injection quantity is less, the oil injection quantity is not generated, the oil injection quantity is more, and the like.

TABLE 1 Fault Meter for engine fuel supply system

Step 102: and determining the credibility of the equipment to be diagnosed for generating each fault symptom according to the state characteristics of the equipment to be diagnosed. The fault symptom is a fault symptom which can be generated by the equipment to be diagnosed, whether one fault symptom can be generated or not can be determined empirically, and the relation between the fault type and the fault symptom and the reliability of the fault type are determined based on a Bayesian network model.

First, a bayesian network model for fault diagnosis is introduced.

The primitive bayesian network structure is used as a bayesian network model, i.e., a "fault-symptom" model, as shown in fig. 5. The upper layer of the primitive Bayesian network model in FIG. 5 represents a fault type, the lower layer represents a fault symptom, nodes with dependency relationship between the fault type and the symptom are connected by a directed arc, then the conditional probability between two nodes with dependency relationship and the prior probability of various faults of the equipment are given according to expert knowledge, and then the Bayesian network model is established according to the corresponding judgment criteria by using the collected sample data.

The model consists of three parts, namely a part state layer, a fault layer and a fault symptom layer: upper layer C_iIs a sample set of fault cases expressed by the state layer of the part, namely the cause of the fault, the middle layer S₈～S₁₁Indicating fault level status, lower level S₁₂～S₁₆It is the fault symptom layer that represents the symptom sample set, namely the fault phenomenon. From the engine oil supply system fault information, a bayesian network model can be constructed, as shown in fig. 6, C₁For ECU failure, C₂For single cylinder injector damage, C₃For oil pumps not pumping oil, C₄In order to block the gasoline filter, the gasoline filter is blocked,C₅for oil leakage of oil pipe C₆For the vacuum tube of the fuel pressure regulator to come off C₇For fuel pressure regulator spring fatigue, S₈For single cylinder, without injection of oil, air filter clogging, S₉For the injection of a small quantity of fuel, S₁₀For no injection quantity, S₁₁For a large amount of fuel injected, S₁₂For engine shake, S₁₃For the engine to emit black smoke, S₁₄For engine power shortage, S₁₅For the engine to be unable to start, S₁₆It is difficult to start the engine.

As can be seen from fig. 6, the direct causal relationship between the cause of the fault case and the sign of the fault is clear, and due to the complexity and uncertainty of the fault of the oil supply system, there is a many-to-one, many-to-many relationship between the cause of the fault case and the sign of the fault. For example, S₉"less fuel injection amount" corresponds to C₂"Single cylinder injector damaged", C₄Clogging of gasoline filter C₅"oil pipe leaking" and C₇'gas pressure regulator spring fatigue' 4 father nodes, and the next level of child node is influenced by the child node with a problem₁₄"Engine Power deficiency" and S₁₆The "difficulty in starting the engine" occurs. Likewise, most parent nodes also have complex causal relationships with multiple child nodes.

The prior probability value required by the fault diagnosis network calculation is generally obtained by carrying out statistical analysis on the system fault on the basis of expert experience guidance. The prior probability is the probability of occurrence of each event determined based on historical data or subjective judgment. The prior probability is generally divided into two types, namely objective prior probability which is calculated by using historical data in the past; the second is subjective prior probability, which is the probability obtained by judging only according to the subjective experience of people when no history data or incomplete history data exists. In the absence of any evidence information, the prior probability and the reliability of the fault occurrence of each fault type node are shown in table 2.

TABLE 2 prior probability and confidence level of each fault type

Tables 3 and 4 list conditional probability tables for the network model part sub-nodes. Wherein Table 3 illustrates node S₉At its parent node C₇、C₅、C₄And C₂The conditional probability values of the various combined value states of (1), table 4 illustrates the node S₁₄At its parent node S₉、S₈Conditional probability values in various combined value states. Conditional probability values of the remaining nodes are obtained in the same manner, and only a part of the conditional probability tables of the nodes are listed here.

TABLE 3 node S₉Conditional probability value and reliability of

TABLE 4 node S₁₄Conditional probability value and reliability of

Based on the bayesian network model introduced above, step 102 specifically: in the MYCIN system, rules are determined based on an inference network, and the credibility of each fault symptom is obtained based on the rules according to an inference sequence.

In practical applications, the measure of the fault symptom of the fault phenomenon is measured by the rule strength, and the concept of the assertion strength is provided, wherein the credibility is refined into the rule strength and the assertion strength. The calculation process of the credibility is a process of propagating the rule strength to the assertion strength.

The rules are made up of assertions, and the initial conditions and final results of the inference are represented by assertions, such as the following rules:

(CS；(S,S(X),ctS))

here, CS is a failure layer (condition) in a layer in the bayesian network, S is a failure sign (conclusion), S (x) is a probability of the failure sign, and ctS is a rule strength, and the reasons why the failure sign S is concluded have uncertainty are:

(1) the corresponding relation in the condition CS may be given by a user, and the user does not have very accurate qualitative effect on a certain phenomenon, so that the conclusion S also has inaccuracy;

(2) the rules themselves have uncertainty (in terms of rule strength), resulting in the conclusion S also having uncertainty;

(3) conditional CS may be the conclusive part of other rules by which the process of reasoning for CS entails inaccuracies in CS.

It is therefore necessary to add some confidence to the conclusions to accurately reflect the true degree of reasoning about the symptoms of the fault. Where conclusion S is relative, it may appear as a condition of another rule by matching it during the inference process, but either the condition or conclusion, they are assertions. Some confidence is added to characterize the truth of an assertion, referred to herein as assertion strength.

In the invention, the credibility is calculated by using a deterministic theory in a MYCIN system, wherein the knowledge is expressed by a rule and is generally expressed in the form that:

if CS₁ and CS₂ and … CS_n then CF(S(X))

wherein CS_i(i ═ 1,2, …, n) is fault layer evidence, S is a fault sign conclusion, S (x) is the probability of a fault sign, and CF (S (x)) is the confidence level of the occurrence of a fault sign (also referred to as the confidence level of the fault sign probability). The rule indicates that: when evidence CS₁To CS_nIf the probability of the symptom S appearing S (x) is positive, the confidence level of the probability of the symptom S appearing S (x) is CF (S (x)). S (X) is the probability of the rule producing a symptom of a fault, given by a domain expert.

In MYCIN, its rules are connected into inference networks through inference strategies, and an imprecise inference model is used for inference:

(1) evidence is a single condition: when the rule if CS then CF (s (x)) is used, the confidence CF (s (x)) of the conclusion s (x) depends not only on the confidence CF (s (x), CS) of the rule but also on the confidence CF (CS) of the evidence CS:

CF(S(X))＝CF(S(X),CS)·max{0,CF(CS)}。

(2) evidence is a logical combination of conditions: evidence is and connection, and the credibility after logical combination is as follows:

CF(CS)＝CF(CS₁ and CS₂ and … CS_n)＝min{CF(CS₁),CF(CS₂),…,CF(CS_n)}

evidence is that the or is connected, and the credibility after the logic combination is as follows:

CF(CS)＝CF(CS₁ or CS₂ or … CS_n)＝max{CF(CS₁),CF(CS₂),…,CF(CS_n)}

(3) the two rules have the same conclusions: the following two rules are set:

rule 1:if CS₁ then CF[S(X),CS₁]

rule 2:if CS₂ then CF[S(X),CS₂]

firstly, respectively calculating the credibility of two rule conclusions:

CF₁(S(X))＝CF(S(X),CS₁)·max{0,CF(CS₁)}

CF₂(S(X))＝CF(S(X),CS₂)·max{0,CF(CS₂)}

the confidence of the conclusion is then calculated using the following equation:

CF₁₂(S(X))＝CF₁(S(X))+CF₂(S(X))

suppose a fault symptom and a fault layer symptom CS₁,CS₂,CS₃,CS₄,CS₅,CS₆In connection with this, the fault level symptom of one of the cases C is CS₁,CS₂The fault level indication for another case C is CS₃,CS₄,CS₅,CS₆The reliability CF (S (X)) of S (X) is obtained by the inference network shown in FIG. 7 (a) and (b), and the concrete steps are as follows：

From the inference network shown in fig. 7, the following rules can be derived:

rule 1:if CS₁ and CS₂ then CF(S(X))；

rule 2:if CS₃ and CS₄ and CS₅ and CS₆ then CF(S(X)))；

in the inference process, the probability S (X) of the symptom needs to be obtained from expert experience, and a final inference result is obtained by adopting forward inference. The reasoning order is as follows:

(1) according to CS₁And CS₂The confidence level of the conclusion S (X) of rule 1 is obtained;

in rule 1 there is CS₁Probability value CS of fault layer symptom₁(X)、CS₂Probability value CS of fault layer symptom₂(X) and weighting factor omega for fault symptom layer probability_i(i＝1,2,3…,n)，ω_iGiven by an expert. For the preconditions:

C＝CS₁(X,ω₁)and CS₂(X,ω₂)and … and CS_n(X,ω_n)

the calculation formula of the reliability is as follows:

the confidence level of the rule 1 conclusion S (X) can be found.

(2) In the same way according to CS₃,CS₄,CS₅,CS₆The confidence of the conclusion S (X) of rule 2 is obtained;

(3) and (3) for two rules with the same conclusion, the final credibility of the fault symptom probability S (X) is obtained through the summation obtained in the steps (1) and (2).

Step 103: and generating the credibility of each fault symptom, the credibility of the fault symptom of each case in the case library and a correction factor by the equipment to be diagnosed, and calculating the similarity between the equipment to be diagnosed and each case.

(1) And determining the credibility of the fault symptom of each case in the case base.

The reliability of the fault symptom of each case in the case base is obtained by performing statistical analysis on the system fault on the basis of expert experience guidance, and taking the reliability of the fault symptom of the engine oil supply system as an example, as shown in table 5.

TABLE 5 engine oil system fault symptom credibility

(2) For case C in case base_pConfidence level, case C, of the ith fault sign generated by the equipment to be diagnosed_pThe reliability and the correction factor of the ith fault symptom are calculated, and the ith fault symptom and the case C generated by the equipment to be diagnosed are calculated_pSimilarity of the ith symptom of (1). The specific calculation formula is as follows:

D_si(C_ni,C_pi)＝β_i(1-|v(C_ni)-v(C_pi)|)；

wherein D is_si(C_ni,C_pi) For awaiting diagnosis equipment C_nGenerating ith fault symptom and case C_pSimilarity of the ith symptom of (1); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor;

(3) calculating the equipment to be diagnosed and the case C according to the similarity of all fault symptoms_pPhase ofSimilarity. The specific calculation formula is as follows:

wherein D is_s(C_n,C_p) For awaiting diagnosis equipment C_nAnd case C_pThe similarity of (2); w_iFor awaiting diagnosis equipment C_nA weighting factor for the ith symptom of failure; m is the total number of symptoms of the fault.

Step 104: and determining the fault elimination scheme corresponding to the case with the maximum similarity as the fault elimination scheme of the equipment to be diagnosed.

Step 103 and step 104 are processes of matching the equipment to be diagnosed with a case base (fault case base), that is, finding a case similar to the equipment to be diagnosed in the case base. In the searching process, many cases similar to the target case are encountered, and are similar in some attributes although not the most similar, so that the final result and the similar cases can be regarded as a set of case sets with similar attributes. If a concept of threshold value is introduced into the system, the similarity between the target case and the cases in the case base is considered to be basically similar if the similarity is larger than the threshold value phi, case matching information with the similarity larger than the threshold value is obtained, all similarity calculation results are sorted, and the case with the largest matching degree is selected as a preliminary troubleshooting scheme. A specific implementation process of the process is shown in fig. 8, and the largest difference between the diagnosis process of the existing PMA provided in fig. 8 and that provided in fig. 3 is that fig. 3 ends after giving a preliminary troubleshooting proposal when matching fails, and in fig. 8, when there are unmatched cases, reliability is obtained based on network rules, and finally the pushing of the troubleshooting plan is implemented by using an improved similarity calculation method (a similarity calculation formula in step 103) based on the reliability, so as to improve the accuracy of fault diagnosis.

In the maintenance auxiliary equipment fault diagnosis provided by the embodiment, the calculation of case similarity is based on the attributes of the cases and mainly aims at the fault symptom attributes. By utilizing an improved similarity calculation method (introducing a correction factor into similarity calculation), the problem that the surfaces are dissimilar and actually similar or the surfaces are similar and actually dissimilar when the number of symptoms is larger and the number of similar symptoms is less can be solved, and the accuracy of fault diagnosis is improved.

The invention also provides a maintenance auxiliary equipment fault diagnosis system, and fig. 9 is a structural diagram of the maintenance auxiliary equipment fault diagnosis system provided by the embodiment of the invention.

Referring to fig. 9, the system comprises:

the data acquisition module 201 is configured to acquire a status characteristic of the equipment to be diagnosed.

A reliability determining module 202, configured to determine, from the status characteristics of the equipment to be diagnosed, a reliability of the equipment to be diagnosed for generating each fault symptom.

And the similarity calculation module 203 is used for calculating the similarity between the equipment to be diagnosed and each case according to the credibility of each fault symptom generated by the equipment to be diagnosed, the credibility of the fault symptom of each case in the case library and the correction factor.

A troubleshooting plan determining module 204, configured to determine the troubleshooting plan corresponding to the case with the largest similarity as the troubleshooting plan of the equipment to be diagnosed.

As an optional implementation manner, the similarity calculation module 203 specifically includes:

and the reliability determining unit is used for determining the reliability of the fault symptom of each case in the case base.

A symptom similarity calculation unit for case C in case base_pConfidence level, case C, of the ith fault sign generated by the equipment to be diagnosed_pThe reliability and the correction factor of the ith fault symptom are calculated, and the ith fault symptom and the case C generated by the equipment to be diagnosed are calculated_pSimilarity of the ith symptom of (1).

A case similarity calculation unit for calculating the equipment to be diagnosed and the case C according to the similarity of all fault symptoms_pThe similarity of (c).

As an optional implementation manner, in the symptom similarity calculation unit, the candidate is to be calculatedThe diagnostic equipment generates the ith fault symptom and case C_pThe calculation formula of the similarity of the ith fault symptom is as follows:

D_si(C_ni,C_pi)＝β_i(1-|v(C_ni)-v(C_pi)|)；

wherein D is_si(C_ni,C_pi) For awaiting diagnosis equipment C_nGenerating ith fault symptom and case C_pSimilarity of the ith symptom of (1); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor;

as an optional implementation manner, in the case similarity calculation unit, the equipment to be diagnosed and the case C_pThe calculation formula of the similarity is as follows:

wherein D is_s(C_n,C_p) For awaiting diagnosis equipment C_nAnd case C_pThe similarity of (2); c_niFor awaiting diagnosis equipment C_nThe ith symptom of failure; c_piCase C_pThe ith symptom of failure; v (C)_ni) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; v (C)_pi) For awaiting diagnosis equipment C_n(ii) the confidence level of the ith symptom of failure; beta is a_iIs a correction factor; w_iFor awaiting diagnosis equipment C_nA weighting factor for the ith symptom of failure; m is the total number of fault symptoms;

the embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.

The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.