CN111105152B - Train key component identification method based on accumulated prospect theory and fuzzy VIKOR theory - Google Patents

Train key component identification method based on accumulated prospect theory and fuzzy VIKOR theory Download PDF

Info

Publication number
CN111105152B
CN111105152B CN201911263409.1A CN201911263409A CN111105152B CN 111105152 B CN111105152 B CN 111105152B CN 201911263409 A CN201911263409 A CN 201911263409A CN 111105152 B CN111105152 B CN 111105152B
Authority
CN
China
Prior art keywords
different
accumulated
decision
foreground
value
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201911263409.1A
Other languages
Chinese (zh)
Other versions
CN111105152A (en
Inventor
秦勇
付勇
汪伟忠
贾利民
王志鹏
程晓卿
叶萌
李想
李明高
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Jiaotong University
Original Assignee
Beijing Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Jiaotong University filed Critical Beijing Jiaotong University
Priority to CN201911263409.1A priority Critical patent/CN111105152B/en
Publication of CN111105152A publication Critical patent/CN111105152A/en
Application granted granted Critical
Publication of CN111105152B publication Critical patent/CN111105152B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0635—Risk analysis of enterprise or organisation activities
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
    • G06Q50/40—Business processes related to the transportation industry

Landscapes

  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Educational Administration (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Tourism & Hospitality (AREA)
  • Physics & Mathematics (AREA)
  • Quality & Reliability (AREA)
  • Operations Research (AREA)
  • Game Theory and Decision Science (AREA)
  • Health & Medical Sciences (AREA)
  • General Health & Medical Sciences (AREA)
  • Primary Health Care (AREA)
  • Train Traffic Observation, Control, And Security (AREA)

Abstract

The invention discloses a train key component identification method based on an accumulated prospect theory and a fuzzy VIKOR theory, which comprises the following steps: firstly, extracting relevant components of a rail train and potential fault modes of the relevant components, and grading different fault modes based on two-type intuition fuzzy semantics; secondly, constructing a value prospect function of a fault mode based on an accumulated prospect theory, and calculating accumulated prospect values of the components under different indexes; and finally, fusing the accumulated prospect values of the components under different indexes by a VIKOR method to obtain a risk sequencing result of the components of the rail train system, and identifying key components of the system. The method is based on the failure mode, influence and hazard analysis of the train system, and carries out risk analysis and key component identification of the train system based on a method of an accumulated prospect theory, thereby providing theoretical support for key maintenance tasks of operation and maintenance personnel on a rail transit field.

Description

Train key component identification method based on accumulated prospect theory and fuzzy VIKOR theory
Technical Field
The invention relates to the field of rail train system risk analysis and key component identification, in particular to a rail train key component identification method based on an accumulated prospect theory and a fuzzy VIKOR theory.
Background
As a reliability Analysis and Risk management technology commonly used in rail transit sites, failure mode, influence and Criticality Analysis (FMECA) obtains a Risk Priority Number (RPN) of a component failure mode through fusion calculation of occurrence degree (O), Criticality degree (S) and detection degree (D) of each component failure mode, and then identifies a key component through fusion of failure mode Risk Priority numbers.
Although the RPN method is a very simple and easy to implement method in practical applications in the field of railways, it still has considerable associated problems and drawbacks. First, in the conventional FMECA analysis, the RPN method multiplies the O, S and D precision values, but in the practical application of the railway, the specific values of O, S and D are difficult to calibrate with the precision data. Secondly, in the conventional way of calculating the RPN by O, S and D, the assumed O, S and D factors have the same importance weight, while in the actual application of the railway, the importance of each factor is different. Finally, the simple multiplication calculation method enables O, S and D factors with different importance degrees to obtain the same RPN result, and certain misleading is generated on decision analysis. Meanwhile, in the field application of rail transit, the degree of harm of key components, equipment and the whole system is more needed, so that the components and the equipment are optimized and improved in a key mode, and the reliability of the system is improved.
In the FMECA-related literature analysis, scholars consider that the process of obtaining RPN results through fusion calculation of O, S and D is a problem of Multi-criterion Decision Making (MCDM). Among the multi-attribute decision methods, a sorting method (Technique for Order Preference by Similarity to an Ideal Solution, TOPSIS) and a multi-criteria compromise Solution sorting method (vlse krierriuska opticicicija I Komoromisno respenje, VIKOR) which approximate to an Ideal point are the most commonly used methods. However, it has been found from numerous documents that the TOPSIS method has certain limitations, and the TOPSIS method only compares the deviation of the alternative from the ideal scheme and the non-ideal scheme respectively, neglects the relative importance of the distance between the alternative and the ideal scheme and the non-ideal scheme, and the obtained result is not necessarily accurate. Compared with the TOPSIS method, the VIKOR method utilizes the proximity of alternative schemes and ideal schemes, and simultaneously considers group effectiveness and individual regressions for analysis, so that the obtained result has higher stability and reliability and is more widely applied to multi-objective decision making.
However, the O, S and D factors need to be scored by experts, and a plurality of uncertainty problems are caused to the analysis process. In related research, fuzzy theories such as grey theory, fuzzy c-means clustering method, triangular fuzzy number method, trapezoidal fuzzy number, Pythagorean fuzzy number and intuitive fuzzy number can be combined with the VIKOR method to solve the problem of uncertainty in the analysis process. Through comparison of documents, the Intuitive Fuzzy Set (IFS) can simultaneously consider information of membership, non-membership and hesitation, has stronger flexibility in describing the attributes of objects, and can better deal with the problems of uncertainty and ambiguity.
However, in practical applications, due to complexity and uncertainty of objective things, membership and non-membership of an intuitive fuzzy number are often difficult to express by an accurate real number, and thus, some scholars propose two types of intuitive fuzzy numbers to solve the problem. In contrast to the type one intuitive fuzzy number, the membership and non-membership of the type two intuitive fuzzy number are re-fuzzified on the type one fuzzy, so that a new model and formula are provided for representing multiple uncertainties and fuzziness. The document describes the membership degree and the non-membership degree by using interval numbers in a type I intuition fuzzy number, and the document is applied to a plurality of problems of uncertainty and ambiguity. However, when using the interval table, sometimes the two end points of the interval may need to be large or small in order to cover the whole range. After performing the operation of multiple intervals, the interval range may be further enlarged. Therefore, some scholars propose that the membership and non-membership values of the type-I intuitive fuzzy number are more suitable to adopt triangular fuzzy numbers between [0,1 ].
In the analysis, the VIKOR method based on the triangle fuzzy number intuitive fuzzy number can well overcome the defects of the existing FMECA method, but the method is premised on the 'rationality' when the FMECA expert makes a decision result, and the 'rationality' decision makes the obtained result slightly deviate from the actual result. Therefore, the scholars propose a "rational-limited" decision method. "rational limited" is a state between rational and irrational, and non-rational. The decision based on the 'limited rationality' considers that in the actual situation, a decision maker cannot accurately and objectively evaluate the possible utility of all decisions in the decision making process and cannot master all effective information, so that the complete rationality and the maximum utility cannot be achieved. Currently, the most representative "limited rationality" decision theory is the foreground theory.
The foreground theory can well explain the phenomenon that a decision maker presents risk avoidance when the decision making process faces obtaining and presents risk pursuit when the decision making process faces loss. For example, when the FMECA expert scores (1-10 points), the score tends to increase in the face of failure modes that occur frequently or have serious consequences (the higher the score, the more serious the consequences), which is a risk avoidance phenomenon; failure modes with low probability of occurrence or high reliability tend to have a lower score (the lower the score, the safer the outcome), which is a risk pursuit phenomenon. However, more and more scholars find that the prospect theory violates the random dominance theory. Subsequently, Kahneman and Tverseky introduce a grade dependence utility theory, a symbol marking dependence utility theory and a Choquet capacity function, provide an accumulation prospect theory, can well explain the advantages, and greatly broaden the application and research field of the theory.
Disclosure of Invention
The invention aims to overcome the hypothesis of 'complete rationality' existing in the traditional analysis process, and carry out risk analysis and key component identification on a rail train system based on an accumulated prospect theory method containing 'limited rationality'.
The purpose of the invention can be realized by the following technical method:
A rail train key component identification method based on an accumulated foreground theory and a fuzzy VIKOR theory comprises the following steps:
the method comprises the following steps:
(1) analyzing fault record data of a rail train system, extracting rail train components and potential fault modes thereof, and grading different fault modes based on two-type intuitionistic fuzzy semantics;
(2) on the basis of obtaining different fault mode scoring information, calculating foreground reference points of different indexes based on an accumulated foreground theory, constructing a value foreground function of a fault mode, and calculating accumulated foreground values of components under different indexes;
(3) and fusing the accumulated prospect values of the components under different indexes by a VIKOR method, calculating to obtain a risk sequencing result of the components of the railway train system, and identifying key components of the system.
Preferably, the specific way of scoring the different failure modes in the step (1) is
The form of index evaluation value is
Figure BDA0002312202810000031
Wherein,
Figure BDA0002312202810000032
is a triangular fuzzy number intuitive fuzzy number,
Figure BDA0002312202810000033
is that
Figure BDA0002312202810000034
The degree of membership of (a) is,
Figure BDA0002312202810000035
is that
Figure BDA0002312202810000036
The degree of membership of the fee of (c),
Figure BDA0002312202810000037
and
Figure BDA0002312202810000038
composed of triangular fuzzy numbers;
let the conversion number of the evaluation value given by the expert be
Figure BDA0002312202810000039
The evaluation information aggregation factor of each expert
Figure BDA00023122028100000310
Can be expressed as:
Figure BDA0002312202810000041
The evaluation information entropy of different experts can be expressed as:
Figure BDA0002312202810000042
the expert weights are expressed as:
Figure BDA0002312202810000043
aggregating foreground information r of different failure modes given by different experts based on expert weightisj。
Preferably, in the step (2), r is setjA selected reference point; d (r)isj,rj) Is r ofisjAnd a reference point rjHamming distance between them, the foreground cost function of different failure modes of the same component is
Figure BDA0002312202810000044
Wherein, alpha and beta (0< alpha, beta <1) represent different risk sensitivity coefficients, the larger the value of the coefficient is, the more sensitive the decision maker is to the risk, the smaller the value is, the less sensitive the decision maker is to the risk; lambda represents a loss avoidance coefficient of the decision maker, and the larger the value of the lambda represents, the larger the avoidance degree of the decision maker to the loss is;
constructing component failure mode prospect value matrix
Figure BDA0002312202810000045
Wherein, the train component is set as alternative Ai(i＝1,2,…m),AiE is A; different failure modes of the same component are set to different performance states FM of the componentis(is＝1,2,…t),FMiE is the FM; the failure modes of the components have 3 evaluation indexes Cj(j＝1,2,…n,n＝3)；
Calculating the cumulative foreground decision weight of different failure modes of the same part:
when the prospect value for the component failure mode is positive,
Figure BDA0002312202810000051
when the prospect value for the component failure mode is negative,
Figure BDA0002312202810000052
wherein p represents the probability of the failure mode occurring; γ and δ reflect the degree of curvature of the decision weight;
Calculating the accumulated foreground value of the component based on the foreground value matrix and the accumulated foreground decision weight of the component failure mode
Figure BDA0002312202810000053
Figure BDA0002312202810000054
Preferably, in the step (3), the cumulative foreground values of the components under different index conditions are defined as
Figure BDA0002312202810000055
The mean value of the accumulated foreground values of the components under different index conditions
Figure BDA0002312202810000056
Is shown as
Figure BDA0002312202810000057
The evaluation information entropy of different indexes is expressed as
Figure BDA0002312202810000058
The evaluation index weight is expressed as
Figure BDA0002312202810000059
Positive ideal solution for determining evaluation values under different evaluation indexes
Figure BDA00023122028100000510
Sum negative ideal solution
Figure BDA00023122028100000511
Figure BDA00023122028100000512
Maximum population utility:
Figure BDA00023122028100000513
minimal individuals regret:
Figure BDA0002312202810000061
definition of
Figure BDA0002312202810000062
The combined index of the different components is
Figure BDA0002312202810000063
Wherein v is a decision coefficient, and v >0.5 represents that the decision is more focused on maximizing the group effect, which indicates that the decision is made according to the opinions of most people; v <0.5 indicates that the decision is more heavily leaned on individuals, indicating that the decision is made based on the person who is rejected; v-0.5, indicating agreement is agreed upon by expert negotiation, indicating that the decision is made by an agreeing person.
Drawings
FIG. 1 is a two-type intuitive fuzzy semantic graph evaluating an index detection degree (D).
Figure 2 rail train bogie system component risk ranking.
Fig. 3 is a flow chart of the method of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation manner and a specific operation process are given, but the scope of the present invention is not limited to the following embodiments.
(1) Analyzing fault record data of a rail train system, extracting relevant components and potential fault modes of the rail train, and grading different fault modes under different professional rail experts based on two-type intuitionistic fuzzy semantics.
Treating rail train components as alternative Ai(i＝1,2,…m),AiE is A; different failure modes of the same component are considered to be different performance states FM of the componentis(is＝1,2,…t),FMiE is the FM; there are 3 evaluation indexes C for failure modes of components in FMECA tablej(j-1, 2, … n, n-3): degree of occurrence (O), degree of hazard (S) and degree of detection (D); special for FMECA (flexible rule and accounting) compiling and preparing groupHome DMkGroup (k 1,2, … l) is composed of train design and manufacture group, driver and train crew group, and maintenance group) 3 group members, and each group of experts scores different failure modes of different parts based on evaluation indexes.
1) Conversion of type II intuitive fuzzy semantics
Since the evaluation values of the occurrence degree (O), the hazard degree (S) and the detection degree (D) given by experts in the FMECA table are not accurate numbers, the fuzzy theory is required to be used for characterization and replacement. The triangular fuzzy number intuition fuzzy set is a fuzzy, namely type two fuzzy, because the membership degree and the non-membership degree of the triangular fuzzy number intuition fuzzy set are formed by triangular fuzzy numbers, and can more accurately represent the uncertainty and the hesitation of experts. The form of the index evaluation value obtained is defined as
Figure BDA0002312202810000071
Wherein,
Figure BDA0002312202810000072
is an intuitive fuzzy number of a triangular fuzzy number,
Figure BDA0002312202810000073
is that
Figure BDA0002312202810000074
The degree of membership of (a) to (b),
Figure BDA0002312202810000075
is that
Figure BDA0002312202810000076
The degree of membership of the fee of (c),
Figure BDA0002312202810000077
and
Figure BDA0002312202810000078
formed by triangular fuzzy numbers, i.e.
Figure BDA0002312202810000079
Figure BDA00023122028100000710
2) Solution of expert weights
Let the conversion number of the evaluation value given by the expert be
Figure BDA00023122028100000711
The evaluation information aggregation factor of each expert
Figure BDA00023122028100000712
Can be expressed as:
Figure BDA00023122028100000713
the evaluation information entropy of different experts can be expressed as:
Figure BDA00023122028100000714
thus, the expert weight can be expressed as:
Figure BDA00023122028100000715
aggregating foreground information r of different failure modes given by different experts based on expert weightisj。
(2) On the basis of obtaining the fault mode grading information under different professional rail experts, calculating foreground reference points of different indexes based on an accumulated foreground theory, constructing a value foreground function of a fault mode, and calculating accumulated foreground values of components under different indexes.
Set rjIs a selected reference point; d (r)isj,rj) To evaluate the value risjAnd a reference point rjHamming distance between. Thus, the prospect merit function of different failure modes of the same component is
Figure BDA0002312202810000081
Wherein, alpha and beta (0< alpha, beta <1) represent different risk sensitivity coefficients, the larger the value of the coefficient is, the more sensitive the decision maker is to the risk, the smaller the value is, the less sensitive the decision maker is to the risk; and lambda (lambda >1) represents a loss avoidance coefficient of the decision maker, and the larger the value of the loss avoidance coefficient, the larger the avoidance degree of the decision maker on the loss. When α is 0.88, β is 0.88, and λ is 2.25, it best fits the psychology of the decision maker.
In addition, for the comparison of the magnitude between the triangle fuzzy number intuitive fuzzy numbers, the triangle fuzzy number intuitive fuzzy number is defined
Figure BDA0002312202810000082
Score function of (2)
Figure BDA0002312202810000083
And variation function
Figure BDA0002312202810000084
Are respectively as
Figure BDA0002312202810000085
Figure BDA0002312202810000086
Wherein,
Figure BDA0002312202810000087
and
Figure BDA0002312202810000088
as a triangular fuzzy number
Figure BDA0002312202810000089
And
Figure BDA00023122028100000810
the average value of (a) of (b),
Figure BDA00023122028100000811
and
Figure BDA00023122028100000812
as a triangular fuzzy number
Figure BDA00023122028100000813
And
Figure BDA00023122028100000814
standard deviation of (2).
Figure BDA00023122028100000815
Figure BDA00023122028100000816
Defining two triangular fuzzy numbers
Figure BDA00023122028100000817
And
Figure BDA00023122028100000818
then
Figure BDA00023122028100000819
And
Figure BDA00023122028100000820
may be in the size relationship of
1) If it is
Figure BDA00023122028100000821
Then there is
Figure BDA00023122028100000822
2) If it is
Figure BDA00023122028100000823
And is
Figure BDA00023122028100000824
Then there is
Figure BDA00023122028100000825
3) If it is
Figure BDA00023122028100000826
And is
Figure BDA00023122028100000827
Then there is
Figure BDA00023122028100000828
Considering that the evaluation value is a triangular fuzzy number intuitive fuzzy number, the reference point adopts the design idea of fuzzy weighted aggregation factor of the evaluation value. Under the same index, the evaluation value risjIs risj＝((alisj,amisj.arisj)，(blisj,bmisj.brisj) For example, the reference point r of the index evaluation value is ((al, am, ar), (bl, bm, br)) is ═ am
Figure BDA0002312202810000091
Defining the selected reference point as rj＝((alj,amj,arj),(blj,bmj,brj) R) after the expert information is aggregated, the evaluation value is risj＝((alisj,amisj,arisj),(blisj,bmisj,brisj) Then the hamming distance between the two triangular blur numbers is:
Figure BDA0002312202810000092
on the basis, a component failure mode foreground value matrix is constructed
Figure BDA0002312202810000093
Wherein train components are considered as alternative Ai(i＝1,2,…m),AiE is A; different failure modes of the same component are considered to be different performance states FM of the componentis(is＝1,2,…t),FMiE is the FM; there are 3 evaluation indexes C for failure modes of components in FMECA tablej(j-1, 2, … n, n-3): degree of occurrence (O), degree of damage (S) and degree of detection (D).
After the foreground value matrix of the component failure mode is obtained, the accumulated foreground decision weights of different failure modes of the same component need to be calculated. The decision weight of the fault mode of the rail train part is a certain subjective judgment made by an evaluation expert according to the actual probability p which possibly occurs in the actual fault record data, the decision weight is actually not probability and does not meet the probability axiom sigmaipi1, it can be considered as evaluating the probability of psychological prediction made by experts in the analysis process, for different failure modes of the same component, based on decision weights of the original prospect theory:
when the prospect value of the failure mode of the part is a positive value, the following steps are provided:
Figure BDA0002312202810000094
when the prospect value of the component failure mode is a negative value, the following steps are provided:
Figure BDA0002312202810000101
wherein p represents the probability of the failure mode occurring; gamma and delta reflect the bending degree of decision weight and also reflect the deviation degree of subjective estimated probability and actual probability of a decision maker, and the function shape is more bent when the value is smaller. Through the calibration of Kahneman and Tverseky, the value ratio of gamma to delta is 0.61 to 0.69.
In the accumulated foreground theory, the modified decision weight function is called as an accumulated decision weight function, and the problem of ordering dependence of various possible result occurrence probabilities of alternative solutions is considered. The contribution of the cumulative prospect theory is to employ cumulative probability weights instead of individual probability weights and thus can be used to describe infinite or even continuous alternative scenarios.
A certain component i is defined to consist of t failure modes. Under a certain evaluation index j, the probability of the fault mode is sorted in an increasing mode according to the size of the obtained evaluation value, and then the probability of the occurrence of the second fault mode of the component is pisAnd is n is defined as-m. If the foreground values of different fault modes of the same part have positive values and negative values, the same part takes the fault mode decision weight of the positive value
Figure BDA0002312202810000102
And negative failure mode decision weight
Figure BDA0002312202810000103
Are respectively as
Figure BDA0002312202810000104
Wherein is 0-n, and
Figure BDA0002312202810000105
Figure BDA0002312202810000106
wherein-m is ≦ 0, and
Figure BDA0002312202810000107
therefore, based on the foreground value matrix and the accumulated foreground decision weights of the component failure modes,calculating cumulative foreground values for a component
Figure BDA0002312202810000108
Figure BDA0002312202810000109
(3) And fusing the accumulated prospect values of the components under different indexes by a VIKOR method, calculating to obtain a risk sequencing result of the components of the railway train system, and identifying key components of the system.
Firstly, calculating the weight of different evaluation indexes by an entropy weight method. Defining the accumulated foreground value of the part under different index conditions as
Figure BDA00023122028100001010
The mean value of the accumulated foreground values of the components under different index conditions
Figure BDA00023122028100001011
Can be expressed as
Figure BDA0002312202810000111
The evaluation information entropy of different indexes can be expressed as
Figure BDA0002312202810000112
Therefore, the evaluation index weight can be expressed as
Figure BDA0002312202810000113
Before calculating the comprehensive risk value of different components based on the VIKOR method, the positive ideal solution of evaluation values under different evaluation indexes needs to be determined
Figure BDA0002312202810000114
Sum negative ideal solution
Figure BDA0002312202810000115
Figure BDA0002312202810000116
Maximum population utility:
Figure BDA0002312202810000117
minimal individuals regret:
Figure BDA0002312202810000118
definition of
Figure BDA0002312202810000119
The combined index of the different components is
Figure BDA00023122028100001110
Wherein v is a decision coefficient, and v >0.5 represents that the decision is more focused on maximizing the group effect, which indicates that the decision is made according to the opinions of most people; v <0.5 indicates that the decision is more heavily leaned on individuals, indicating that the decision is made based on the person who is rejected; v-0.5, indicating agreement is agreed upon by expert negotiation, indicating that the decision is made by an agreeing person.
The method implementation of the embodiment specifically comprises the following steps:
s01: the method provided by the invention is verified by collecting and organizing fault data of a rail train bogie system and extracting 28 fault modes (shown in table 1) of bogie system components, and taking the fault modes as a case. The invention considers the rail train part as an alternative partyTable Ai(i＝1,2,…28),AiE is A; different failure modes of the same component are considered to be different performance states FM of the componentis(is＝1,2,…t),FMiE is the FM; there are 3 evaluation indexes C for failure modes of components in FMECA tablej(j-1, 2, … n, n-3): degree of occurrence (O), degree of hazard (S) and degree of detection (D); expert DM for FMECA compiling group kThe (k-1, 2, … 3) group is composed of train design and manufacture group, driver and train crew group, and maintenance group) 3 group members, and each group of experts scores different failure modes of different parts based on evaluation indexes.
Tables 2 to 4 show two types of intuitive fuzzy semantic tables for evaluating the occurrence degree (O), the hazard degree (S) and the detection degree (D) of indexes, and based on the fuzzy semantic tables, the invention converts the character evaluation result given by experts into two types of intuitive fuzzy numbers. The weights of the experts under the index occurrence degree (O) are calculated to be 0.2822, 0.3268 and 0.3910 respectively; the weights of the experts under the index hazard degree (S) are 0.2455, 0.2406 and 0.5139 respectively; the weights of the experts under the index detection degree (D) are 0.3678, 0.3538, and 0.2784, respectively. Based on the expert weight, foreground information r of different failure modes given by different aggregated experts is obtainedisj. The foreground information of the failure modes for the partial failure modes is shown in table 5.
S02: by calculation, the reference points at the index occurrence degree (O) were ((0.21,0.25,0.37), (0.74,1.0,1.0)), the reference points at the index harmfulness degree (S) were ((0.45,0.55,0.65), (0.45,0.55,0.65)), and the reference points at the index detection degree (D) were ((0.33,0.42,0.52), (0.58,1.0, 1.0)).
Calculating foreground information r of different failure modesisjAnd a reference point rjThe hamming distance between the components is the parameter of alpha 0.88, beta 0.88 and lambda 2.25, and the foreground value matrix of the failure mode of the component is constructed. The foreground cost matrix for a partial failure mode is shown in table 6.
The method is based on the prospect values of different fault modes of the same component, carries out sequential arrangement from small to large, and calculates to obtain the accumulated prospect decision weight of different fault modes of the same component. The cumulative foreground decision weights for the partial failure modes are shown in table 7. Based on partsForeground value matrix and accumulated foreground decision weight of fault mode, and accumulated foreground value of calculating component
Figure BDA0002312202810000121
The cumulative foreground values for some of the components are shown in table 8.
S03: by the entropy weight method, the weights of the evaluation index occurrence degree (O), the criticality degree (S) and the detection degree (D) can be calculated to be 0.25, 0.18 and 0.57 respectively. Finally, the maximum population utility, minimum individual regret and composite index of each part were calculated by the VIKOR method, as shown in table 9. Sorting according to the small to large part comprehensive indexes, and the sorting result is shown in figure 2.
The risk importance degree of the rail train bogie system components is ranked in the first ten components, namely a frame assembly, an axle, a traction rod, wheels, an axle box bearing, a central shaft, a parking brake unit, an air spring, a gear box assembly and an anti-rolling torsion bar. The frame assembly places a first place of importance in these 28 components because the frame assembly is the main part of a rail train bogie and the occurrence of any one failure mode can cause derailment of the train resulting in an irreparable hazard. The failure of the axle and the traction rod to break can also lead to the occurrence of derailment of the train, while the failure of the axle and the traction rod to crack can lead to the reduction of the running stability of the train, possibly endangering the safety of the train and the personnel, and the rubber wear or aging of the traction rod is a failure mode which accounts for a large proportion of the failure mode, leads to the reduction of the running stability of the train and has a great influence on the safety of the train and the personnel, so that the axle and the traction rod are arranged at the second and third important positions. In addition, the parking brake unit is an important part for braking the train and preventing the train from sliding after the train is parked, and has great relation with passengers to get on or off the train. If the parking brake unit fails, the train can slide, personnel can be injured and killed, the panic of passengers can be caused, and therefore the expert gives a higher harmfulness score when scoring, and the final sequencing result is also in the seventh position in 28 components.
TABLE 1 summary table of bogie components of railway train system of a certain model
Numbering Component part Numbering Component part Numbering Component part Numbering Component part
1 Frame assembly 8 Transverse buffer stop 15 Coupling joint 22 Tread brake unit
2 Axle shaft 9 Air spring 16 Traction motor 23 Parking brake unit
3 Wheel of vehicle 10 Height adjusting device 17 Draw bar 24 Brake pad
4 Axle box body 11 Differential pressure valve 18 Traction frame assembly 25 Wheel rim lubricating device
5 Axle box bearing 12 Transverse shock absorber 19 Center shaft 26 Grounding brush
6 First spring 13 Anti-side rolling torsion bar 20 Center pin 27 Radio frequency identificationFastening device
7 Vertical shock absorber 14 Gear box assembly 21 Transverse buffer device 28 Temperature sensor
TABLE 2 two-type intuitive fuzzy semantic Table for evaluating index degree of occurrence (O)
Expert scoring Numerical value of credit Type II intuitive fuzzy semantic value
Very low frequency of occurrence 1,2 ((0.1,0.1,0.2),(0.9,1.0,1.0))
The frequency of occurrence is low 3,4 ((0.2,0.3,0.4),(0.7,0.8,0.9))
Moderate frequency of occurrence 5,6 ((0.4,0.5,0.6),(0.5,0.6,0.7))
The frequency of occurrence is higher 7,8 ((0.6,0.7,0.8),(0.3,0.4,0.5))
The frequency of occurrence is very high 9,10 ((0.8,0.9,1.0),(0.1,0.2,0.3))
TABLE 3 evaluation index haziness (S) two-type intuitive fuzzy semantic table
Expert scoring Numerical value of credit Type II intuitive fuzzy semantic value
The consequence is very light 1,2 ((0.1,0.1,0.2),(0.9,1.0,1.0))
With less consequence 3,4 ((0.2,0.3,0.4),(0.7,0.8,0.9))
Moderate consequences 5,6 ((0.4,0.5,0.6),(0.5,0.6,0.7))
The consequence is more serious 7,8 ((0.6,0.7,0.8),(0.3,0.4,0.5))
The consequences are very serious 9,10 ((0.8,0.9,1.0),(0.1,0.2,0.3))
TABLE 4 two-type intuitive fuzzy semantic Table for evaluating index detection metric (D)
Expert scoring Numerical value of credit Type II intuitive fuzzy semantic value
The detection mode is very easy 1,2 ((0.1,0.1,0.2),(0.9,1.0,1.0))
The detection mode is easier 3,4 ((0.2,0.3,0.4),(0.7,0.8,0.9))
Moderate difficulty of detection mode 5,6 ((0.4,0.5,0.6),(0.5,0.6,0.7))
The detection mode is difficult 7,8 ((0.6,0.7,0.8),(0.3,0.4,0.5))
The detection method is very difficult 9,10 ((0.8,0.9,1.0),(0.1,0.2,0.3))
TABLE 5 Foreground information summary of failure modes for partial failure modes
Figure BDA0002312202810000141
Figure BDA0002312202810000151
6 partial failure mode foreground value matrix
Component part Failure mode O S D
Axle shaft Small cracks 0.38 0.19 0.37
Axle shaft Fracture of -0.28 0.41 0.44
Wheel of vehicle Scuffing of the tread 0.45 -0.03 -0.69
Wheel of vehicle Tread stripping 0.17 0.40 -0.69
Wheel of vehicle Crack of tread -0.20 0.08 0.38
Wheel of vehicle Rim wear 0.71 -0.14 -0.51
Wheel of vehicle Fracture of -0.28 0.40 0.56
Table 7 summary table of cumulative foreground decision weights for partial failure modes
Component part Failure mode O S D
Axle shaft Small cracks 0.71 0.81 0.81
Axle shaft Fracture of 0.17 0.19 0.19
Wheel of vehicle Scuffing of the tread 0.05 0.06 0.11
Wheel of vehicle Tread stripping 0.07 0.09 0.06
Wheel of vehicle Crack of tread 0.09 0.05 0.09
Wheel of vehicle Rim wear 0.64 0.71 0.66
Wheel of vehicle Fracture of 0.04 0.05 0.06
TABLE 8 accumulated foreground values for partial components
Component part O S D
Frame assembly 0.468144 -0.04211 -0.39385
Axle shaft 0.298719 -0.09364 -0.60525
Wheel of vehicle 0.281249 -0.22827 -0.42217
Axle box body 0.336425 -0.1147 -0.5351
Axle box bearing 0.030094 -0.15011 -0.5314
Primary spring 0.052312 -0.0584 -0.46359
Vertical shock absorber 0.098971 -0.40377 -0.41591
Transverse buffer stop 0.468144 -0.04211 -0.39385
Air spring 0.298719 -0.09364 -0.60525
TABLE 9 maximum population utility, minimum individual regret and composite index for part of the components
Component part Si Ri Qi
Frame assembly 0.086112 0.086112 0.006349
Axle shaft 0.162393 0.080131 0.045295
Wheel of vehicle 0.487848 0.425123 0.604797
Axle box body 0.642597 0.518317 0.795623
Axle box bearing 0.593685 0.437605 0.680894
Primary spring 0.604372 0.487389 0.740092
Vertical shock absorber 0.702672 0.485762 0.796734
Transverse buffer stop 0.652077 0.536024 0.820049
Air spring 0.680781 0.434846 0.729682

Claims (2)

1. A train key component identification method based on an accumulated prospect theory and a fuzzy VIKOR theory is characterized by comprising the following steps:
(1) Analyzing fault record data of a rail train system, extracting rail train components and potential fault modes thereof, and grading different fault modes based on two-type intuitionistic fuzzy semantics;
(2) on the basis of obtaining different fault mode scoring information, calculating foreground reference points of different indexes based on an accumulated foreground theory, constructing a value foreground function of a fault mode, and calculating accumulated foreground values of components under different indexes;
in the step (2), r is setjTo select a reference point, risjForeground information of different failure modes is given for different experts; d (r)isj,rj) Is risjAnd a reference point rjHamming distance between them, the foreground cost function of different failure modes of the same component is
Figure FDA0003629979740000011
Wherein, α, β; 0< α, β < 1; different risk sensitivity coefficients are represented, the larger the value of the risk sensitivity coefficient is, the more sensitive the decision maker is to the risk, the smaller the value is, and the less sensitive the decision maker is to the risk; lambda represents a loss avoidance coefficient of the decision maker, and the larger the value of the lambda represents, the larger the avoidance degree of the decision maker to the loss is;
constructing component failure mode prospect value matrix
Figure FDA0003629979740000012
Wherein, the train component is set as alternative Ai，i＝1,2,…m,AiE is A; the failure modes of the components have 3 evaluation indexes Cj，j＝1,2,…n,n＝3；
Calculating the cumulative foreground decision weight of different failure modes of the same part:
When the prospect value is positive for the component failure mode,
Figure FDA0003629979740000013
when the prospect value for the component failure mode is negative,
Figure FDA0003629979740000021
wherein p represents the probability of the failure mode occurring; γ and δ reflect the degree of curvature of the decision weight;
defining a certain component i to be composed of t fault modes, and sequencing the probability of the fault mode in an increasing mode according to the magnitude of the obtained evaluation value under a certain evaluation index j, wherein the probability of the occurrence of the second fault mode of the component is pisDefining-m is ≦ n, then the same component takes the positive value of the failure mode decision weight
Figure FDA0003629979740000022
And negative failure mode decision weight
Figure FDA0003629979740000023
Are respectively as
Figure FDA0003629979740000024
Wherein is 0-n, and
Figure FDA0003629979740000025
Figure FDA0003629979740000026
wherein-m is ≦ 0, and
Figure FDA0003629979740000027
calculating the accumulated foreground value of the component under different index conditions based on the foreground value matrix and the accumulated foreground decision weight of the component fault mode
Figure FDA0003629979740000028
Figure FDA0003629979740000029
(3) Integrating the accumulated prospect values of the components under different indexes by a VIKOR method, calculating to obtain a risk sorting result of the components of the rail train system, and identifying key components of the system;
in the step (3), the accumulated foreground values of the components under different index conditions are defined as
Figure FDA00036299797400000210
The mean value of the accumulated foreground values of the components under different index conditions
Figure FDA00036299797400000211
Is shown as
Figure FDA00036299797400000212
The evaluation information entropy of different indexes is expressed as
Figure FDA00036299797400000213
The evaluation index weight is expressed as
Figure FDA0003629979740000031
Positive ideal solution for determining evaluation values under different evaluation indexes
Figure FDA0003629979740000032
Sum negative ideal solution
Figure FDA0003629979740000033
Figure FDA0003629979740000034
Maximum population utility:
Figure FDA0003629979740000035
minimal individuals regret:
Figure FDA0003629979740000036
definition of
Figure FDA0003629979740000037
The combined index of the different components is
Figure FDA0003629979740000038
Wherein v is a decision coefficient, and v >0.5 represents that the decision is more focused on maximizing the group effect, which indicates that the decision is made according to the opinions of most people; v <0.5 indicates that the decision is more heavily leaned on individuals, indicating that the decision is made based on the person who is rejected; v-0.5, indicating agreement is agreed upon by expert negotiation, indicating that the decision is made by an agreeing person.
2. The method as claimed in claim 1, wherein the method for identifying train key components based on the accumulated foreground theory and the fuzzy VIKOR theory,
the specific mode of scoring different fault modes in the step (1) is
The form of index evaluation value is
Figure FDA0003629979740000039
Wherein,
Figure FDA00036299797400000310
is a triangular fuzzy number intuitive fuzzy number,
Figure FDA00036299797400000311
is that
Figure FDA00036299797400000312
The degree of membership of (a) is,
Figure FDA00036299797400000313
is that
Figure FDA00036299797400000314
The degree of membership of the fee of (c),
Figure FDA00036299797400000315
and
Figure FDA00036299797400000316
composed of triangular fuzzy numbers;
let the conversion number of the evaluation value given by the expert be
Figure FDA00036299797400000317
The evaluation information aggregation factor of each expert
Figure FDA0003629979740000041
Expressed as:
Figure FDA0003629979740000042
the evaluation information entropy of different experts is expressed as:
Figure FDA0003629979740000043
the expert weights are expressed as:
Figure FDA0003629979740000044
Aggregating foreground information r of different failure modes given by different experts based on expert weightisj。
CN201911263409.1A 2019-12-11 2019-12-11 Train key component identification method based on accumulated prospect theory and fuzzy VIKOR theory Active CN111105152B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911263409.1A CN111105152B (en) 2019-12-11 2019-12-11 Train key component identification method based on accumulated prospect theory and fuzzy VIKOR theory

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911263409.1A CN111105152B (en) 2019-12-11 2019-12-11 Train key component identification method based on accumulated prospect theory and fuzzy VIKOR theory

Publications (2)

Publication Number Publication Date
CN111105152A CN111105152A (en) 2020-05-05
CN111105152B true CN111105152B (en) 2022-06-10

Family

ID=70422975

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911263409.1A Active CN111105152B (en) 2019-12-11 2019-12-11 Train key component identification method based on accumulated prospect theory and fuzzy VIKOR theory

Country Status (1)

Country Link
CN (1) CN111105152B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112365121A (en) * 2020-10-09 2021-02-12 天津大学 Steady state power quality comprehensive evaluation method based on interval mathematics
CN116956646B (en) * 2023-09-20 2023-12-08 国家机床质量监督检验中心 Reliability analysis method and device for numerical control machine tool, computer equipment and storage medium

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103927448A (en) * 2014-04-18 2014-07-16 南京理工大学 Determining method of criticality of fault mode of rail transit vehicle component
CN109118037A (en) * 2018-06-26 2019-01-01 桂林电子科技大学 A kind of choice for traveling method based on accumulation prospect theory

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20150009331A1 (en) * 2012-02-17 2015-01-08 Balaji Venkatraman Real time railway disaster vulnerability assessment and rescue guidance system using multi-layered video computational analytics

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103927448A (en) * 2014-04-18 2014-07-16 南京理工大学 Determining method of criticality of fault mode of rail transit vehicle component
CN109118037A (en) * 2018-06-26 2019-01-01 桂林电子科技大学 A kind of choice for traveling method based on accumulation prospect theory

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
Operational risk assessment of railway train based on type-2 intuitionistic fuzzy set and dynamic VIKOR approach;Fu Yong 等;《Journal of Transportation Safety & Security》;20190715;1024-1046 *

Also Published As

Publication number Publication date
CN111105152A (en) 2020-05-05

Similar Documents

Publication Publication Date Title
CN112508392B (en) Dynamic evaluation method for traffic conflict risk of hidden danger road section of mountain area double-lane highway
WO2020244288A1 (en) Method and apparatus for evaluating truck driving behaviour based on gps trajectory data
CN111105152B (en) Train key component identification method based on accumulated prospect theory and fuzzy VIKOR theory
CN107153914B (en) System and method for evaluating automobile operation risk
CN112201038A (en) Road network risk assessment method based on risk of bad driving behavior of single vehicle
CN104809877A (en) Expressway site traffic state estimation method based on feature parameter weighted GEFCM algorithm
CN109635852B (en) User portrait construction and clustering method based on multi-dimensional attributes
CN101826183A (en) Intelligent car evaluation method and system
CN103678924A (en) Method for determining failure mode risk priority numbers of key component of rail transit vehicle
CN110570655B (en) Vehicle feature evaluation method based on hierarchical clustering and decision tree
CN107918826B (en) Driver evaluation and scheduling method for driving environment perception
CN109887292B (en) Vehicle type identification method and system
CN111832730B (en) Reliability characterization and state identification method for uncertain oil state
CN107169576A (en) The maintenance decision system and method for a kind of commerial vehicle maintenance period
CN115545101A (en) High-speed train bogie fault diagnosis method based on residual error neural network
CN115311858A (en) Urban road section grading control method based on traffic flow toughness
CN102207930A (en) Method for calculating passenger flow security level of comprehensive transport hub
CN111861765B (en) Intelligent anti-fraud method for vehicle insurance claim settlement
CN113920723B (en) Identification method for mountainous area expressway accident-prone road section
CN113240268A (en) Network taxi appointment supervision evaluation method based on AHP and fuzzy evaluation method
Yuan et al. A gray prediction method for economic loss of road traffic accidents based on Logistic model.
CN109118034B (en) Three-dimensional quantitative comprehensive evaluation method for driving behaviors
CN113112803A (en) Urban traffic road traffic flow data acquisition and analysis processing system based on video monitoring
Wang et al. A deep learning fault diagnosis method for metro on-board detection on rail corrugation
CN113657555B (en) Improved semi-supervised clustering-based ice and snow environment driving experience evaluation method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant