CN103902845A - Method for evaluating risk of simulation system based on fuzzy FMEA - Google Patents

Abstract

The invention relates to a method for evaluating the risk of a simulation system based on fuzzy FMEA, in particular to a method for evaluating the risk of the simulation system through fuzzy FMEA. The method aims to solve the problems that an existing method for evaluating the risk is unreasonable in cost allocation and evaluation, an evaluation result cannot meet requirements, and qualitative evaluation is based on opinions and judgment of people. The method includes the steps that firstly, the risk of the whole simulation system is preliminarily evaluated based on fuzzy FMEA, and then a fuzzy result is obtained; secondly, evaluation cost of the whole simulation system is estimated; thirdly, a calculation model of the risk and the cost of the whole simulation system is built; fourthly, the relation between fuzzy linear cost and the risk of the whole simulation system is evaluated, a fuzzy linear programming model of an evaluation solution is built, and then the final whole risk evaluation result is obtained. The method is applied to the field of risk evaluation.

Description

Analogue system methods of risk assessment based on fuzzy FMEA

Technical field

The present invention relates to risk assessment field, be specifically related to utilize the method for fuzzy FMEA to carry out the risk assessment of analogue system.

Background technology

Since nearly over half a century, modeling and simulation technology is under the promotion and promotion of various application program technology and related discipline, develop into the professional knowledge of a relatively complete system, as a versatility, Strategic Technology, at present, modeling and simulation technology forward " network virtualization, intelligent, Collaborative, extensiveization " feature of modernization development direction, it successful Application in Aero-Space, information, material, the energy, new and high technology and industry that manufacture etc. are advanced, business, agricultural, education, traffic, society, economical, medical science, life, the various fields such as service for life.It is believed that, modeling and simulation technology, together with high-performance calculation, has become after theoretical and experimental study, the third understanding, the important means of transforming the objective world.

The essence of emulation is a kind of process of knowledge processing, common system emulation process comprising: set up system model, set up realistic model, design and simulation software, carry out (seeing Fig. 1) such as emulation experiment and data analyses, it relates to the knowledge of multiple subject multiple fields.Along with the high speed development of Modern New technology, emulation technology has also obtained development at a high speed, application on the field of civil and military is also constantly to multi-direction expansion, mathematical, and in 21 century, the research of emulation technology will obtain larger development with application.

The application of analogue system is extensive gradually, and its reliability disadvantages also more and more causes everybody concern.Confidence of simulation system is to utilize assessment to ensure, how to ensure that by necessary assessment the confidence level of emulation involves the design problem of evaluation scheme.Generally in the ideal case, the mainly demand of the consideration based on technological layer and confidence level of the design of evaluation scheme.But, the fact is really not so, the design of evaluation scheme conventionally to a great extent, also can be assessed cost (the assessed cost meaning by complete assessment activity the summation of the cost paid of necessity) restriction, so can not reflect the demand of confidence of simulation system completely.So we need to use more the effectively optimal design mode of evaluation scheme, obtain a more reasonably evaluation scheme, enable cost and effectively utilize the assessment of limited analogue system, complete assessment.But, the Optimization Design of science, objective assessment scheme not at present, the problem that this situation causes the evaluation work of current complex simulation system to exist is a lot, and as unreasonable in absorption of costs assessment, outcome evaluation does not meet and meets the demands etc.In order to solve these current problems, key is to formulate the evaluation scheme Optimization Design of the analogue system of actual a, science.

Analogue system uncertainty of evaluation of prior art, when availability and some special datas, use qualitative evaluation, although in the past, VV & A utilization frequency is the highest, but their use is not very good, does not reach the effect of expectation yet, these all show, qualitative evaluation is with people's suggestion and be judged as foundation, but under normal circumstances, participates in qualitative evaluation personnel's judgement and assesses the regular process of all neither one sequencing.

For the current this irrational state of qualitative evaluation, there are 2 reasons here, the first, we are few regular and strict process in the time of qualitative evaluation; The second, those experts go back process or the program that neither one can be used for following in the time carrying out the risk assessment of analogue system, and this has just limited the validity of our current analogue system risk assessment.Go back at present neither one science, objective assessment scheme optimization method for designing, this situation causes current complex simulation system evaluation work to have a lot of problems, for example, assessed cost unreasonable distribution, assessment result can not meet acceptance requirement etc.

In order to improve this present situation, the present invention carries out analysis and the research of system to the problem such as risk assessment and fuzzy FMEA of analogue system.

For current deficiency, we will manage to realize the standard of system input, in the time that input need to be revised, for set fuzzy rule, we have the output of standard, this needs the true code of fuzzy inference rule, the validity that this just can ensure whole qualitative evaluation process, makes it more independent, stable.The realization of description and system is separated.We also will find a reliable assessment criteria to be used as the reference of fuzzy rule for the risk assessment of analogue system, and that is exactly fuzzy FMEA.

Summary of the invention

The present invention will solve that availability risk appraisal procedure absorption of costs assessment is unreasonable, outcome evaluation does not meet and meets the demands, qualitative evaluation is with people's suggestion and is judged as the problem of foundation, and the methods of risk assessment of the analogue system based on fuzzy FMEA is provided.

Analogue system methods of risk assessment based on fuzzy FMEA is realized according to the following steps:

One, based on fuzzy FMEA, the risk of whole analogue system is carried out to entry evaluation, obtain the result of obfuscation: very low VL, low L, middle M, high H and very high VL;

Two, adopt $V_j=(V_{\mathrm{pj}},V_{\mathrm{mj}},V_{\mathrm{oj}})=N\·\mathrm{CP}\·T_j=(\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{pji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{mji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{oji}})$ Assessed cost to whole analogue system is estimated;

Three, set up the computation model R=R of the risk and cost of whole analogue system _sim-cv;

Four, utilize $\mathrm{Maximize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\left(\frac{(c_{\mathrm{pj}}+4c_{\mathrm{mj}}+c_{\mathrm{oj}})}{6}{v}_j\right)$ Constraint condition: $\left\{\begin{array}{c}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{v}_j\≤V_t\\ v_{\mathrm{pj}}\≤v_j\≤v_{\mathrm{oj}}\end{array}\right.$ Evaluate the relation between the Fuzzy Linear cost and risk of whole analogue system, set up the fuzzy linear programming model of assessment solution, obtain the result of final whole risk assessment, completed the analogue system methods of risk assessment based on fuzzy FMEA.

Brief description of the drawings

Fig. 1 is system emulation procedure chart in background technology of the present invention;

Fig. 2 assesses risk RPN process flow diagram with fuzzy logic and fuzzy reasoning;

Fig. 3 is the graph of a relation of embodiment four risks and cost;

Fig. 4 is that in emulation experiment, GUI editing interface figure rolls bar input figure;

Fig. 5 is GUI editing interface form input figure in emulation experiment;

Fig. 6 is that in emulation experiment, GUI runnable interface rolls bar input figure;

Fig. 7 is that in emulation experiment, GUI runnable interface operation result 1 rolls bar input figure;

Fig. 8 is that in emulation experiment, GUI runnable interface operation result 2 rolls bar input figure;

Fig. 9 is GUI runnable interface form input figure in emulation experiment.

Embodiment

Embodiment one: the analogue system methods of risk assessment based on fuzzy FMEA of present embodiment is realized according to the following steps:

One, based on fuzzy FMEA, the risk of whole analogue system is carried out to entry evaluation, obtain the result of obfuscation: very low VL, low L, middle M, high H and very high VL;

Two, adopt $V_j=(V_{\mathrm{pj}},V_{\mathrm{mj}},V_{\mathrm{oj}})=N\·\mathrm{CP}\·T_j=(\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{pji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{mji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{oji}})$ Assessed cost to whole analogue system is estimated;

Three, set up the computation model R=R of the risk and cost of whole analogue system _sim-cv;

Four, utilize $\mathrm{Maximize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\left(\frac{(c_{\mathrm{pj}}+4c_{\mathrm{mj}}+c_{\mathrm{oj}})}{6}{v}_j\right)$ Constraint condition: $\left\{\begin{array}{c}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{v}_j\≤V_t\\ v_{\mathrm{pj}}\≤v_j\≤v_{\mathrm{oj}}\end{array}\right.$ Evaluate the relation between the Fuzzy Linear cost and risk of whole analogue system, set up the fuzzy linear programming model of assessment solution, obtain the result of final whole risk assessment, completed the analogue system methods of risk assessment based on fuzzy FMEA.

Present embodiment is by fuzzy FMEA (Failure Modes and Effect Analysis, be fault modes and effect analysis) method analyze the risk in analogue system and it assessed, analogue system assessment based on fuzzy FMEA has been proposed, fuzzy reasoning and fault analysis pattern are carried out to combination, complete the risk assessment to analogue system, and set up the mathematical model of risk and assessed cost, and be optimized calculating, obtain last risk and cost optimization result.

Carry out risk assessment by the method for fuzzy FMEA, this is that one combines with fuzzy reasoning the methods of risk assessment carrying out by FMEA (fault modes and effect analysis), be how to organically combine with fuzzy reasoning at this by the explanation basic thought of fuzzy reasoning and the thought of FMEA, thereby carry out the risk assessment of analogue system.

FMEA a kind ofly can analyze and identify various potential fault modes in product, service or technological process, determine their priority level, and weak link wherein and key project are taked to the systematic analysis instrument of innovative approach, identify and remove known or incipient fault, thereby for decision maker formulates and provides the important technology FMEA of information, general application risk Ser.No. (risk priority number, RPN) carries out the sequence of risk to goal systems, RPN is expressed as:

RPN＝S×O×D

In formula, the order of severity of S---fault;

The occurrence frequency of O---fault;

The complexity of D---fault finding.

Be similar to other system, analogue system carried out to FMEA and conventionally follow following steps:

(1) identification of system and function,

(2) identification of the defect mode of system,

(3) determining of the impact of fault mode,

(4) cause the identification of the reason that defect is possible,

(5) calculating of risk priority number RPN.

The fuzzy FMEA of present embodiment is specially:

In conventional FMEA method, to severity, occurrence frequency, be difficult for detection degree and carry out grade and comment, regularly, more uses be the linguistic form that " very high ", " medium ", " very low ", " very small " etc. meet custom, " height " here, " in ", " low ", " little " be all fuzzy message.Conventional FMEA has certain limitation and deficiency processing that these are uncertain, when inaccurate fuzzy message.For this limitation, here we use the way of FMEA as the criterion of the risk in evaluation simulation system, but in common situation, every evaluation index that our expert provides is all qualitatively, this is just particularly suitable for processing by the mode of fuzzy reasoning the input of FMEA, in fact namely here, core is one and builds the three-dimensional fuzzy reasoning system of fuzzy rule taking FMEA as instructing, then the process of the fuzzy reasoning elaborating in conjunction with present embodiment we just can obtain the process of fuzzy FMEA.We are by the S of risk (order of severity of fault), O (occurrence frequency of fault), D (complexity of fault finding), with in the present embodiment, used 5 fuzzy words: very low (VL), low (L), in (M), high (H) and very high (VL).Therefore, define word collection S={VL here, L, M, H, VH}.Subordinate function is selected triangular membership functions, fuzzy rule is by tri-factor S of FMEA (order of severity of fault), O (occurrence frequency of fault), D (complexity of fault finding) decides ultimate risk effect in actual conditions.

Consider the risk assessment of analogue system is related to a large amount of qualitative or uncertain comments, in the present embodiment, evaluate the risk of analogue system with fuzzy FMEA.Generally speaking, assess risk RPN with fuzzy logic and fuzzy reasoning, its process is as Fig. 2.Start describe the frequency of generation and the risk of the order of severity and be easy to detection level as input language variable, then use the obfuscation of input membership function, adopt fuzzy rule base to obtain a fuzzy risk.Finally, de-fuzzy fuzzy risk evaluating value.

Embodiment two: present embodiment is different from embodiment one: based on fuzzy FMEA, whole analogue system is carried out to entry evaluation in described step 1 and be specially:

(1) fault mode of identification analogue system, arranges and is made as R1, R2....Rn by contingent fault;

(2) the occurrence frequency O of the order of severity S to each contingent fault, fault and the complexity D of fault finding carry out expert assessment and evaluation classification grading, obtain the result of obfuscation; Wherein, the result of described obfuscation comprises very low VL, low L, middle M, high H and very high VL, therefore, and definition word collection S={VL, L, M, H, VH};

(3) adopt the arithmetic mean of multiple expert assessments and evaluations be multiple experts' comprehensive evaluation value as the risk evaluation result of each fault, its formula is as follows:

$F_j=\left(\frac{1}{m}\mathrm{Fij}\right)$

In formula, m---expert's number;

F _ij---order of severity S, the occurrence frequency O of fault and the complexity D of the fault finding classification grading of expert to fault;

F _j---the arithmetic mean of trying to achieve;

(4) application fuzzy inference system obtains the risk evaluation result of each fault according to multiple experts' comprehensive evaluation value, and the risk evaluation result of each fault is S1, S2...Sn, O1, O2...On, D1, D2...Dn.

(1) fault mode of identification analogue system, arranges and is made as R1, R2....Rn by contingent fault; Wherein, error, systematic error, electromagnetic interference (EMI) and other uncontrollable factors of described contingent fault bag enlarging mould;

(2) the occurrence frequency O of the order of severity S to each contingent fault, fault and the complexity D of fault finding carry out expert assessment and evaluation classification grading, obtain the result of obfuscation; Wherein, the result of described obfuscation comprises very low VL, low L, middle M, high H and very high VL, therefore, and definition word collection S={VL, L, M, H, VH};

(3) obtaining of multiple experts' comprehensive evaluation value:

Adopt arithmetic mean, its formula is as follows:

$F_j=\left(\frac{1}{m}\mathrm{Fij}\right)$

In formula, m---expert's number;

F _ij---order of severity S, the occurrence frequency O of fault and the complexity D of the fault finding classification grading of expert to fault;

F _j---the arithmetic mean of trying to achieve;

(4) application fuzzy inference system obtains the risk evaluation result of each fault according to multiple experts' comprehensive evaluation value, and the risk evaluation result of each fault is S1, S2...Sn, O1, O2...On, D1, D2...Dn.

In present embodiment, failure definition pattern is one of problem more difficult in the venture analysis of analogue system.The existing achievement in research of simulation analysis, the fault mode existing in discovery system as far as possible.For example, for software systems, the fault mode that may exist comprises: data fault, calculating fault, logic fault, interface fault and environmental bug.Fault mode associated with the data may comprise: data are incorrect, loss of data, data redundancy and data time sequence mistake.

Owing to being qualitative analysis, the data set that we obtain needs obfuscation to process again, we set the concrete criterion of processing the qualitative analysis as the case may be, this is for we set up fuzzy membership functions and fuzzy rule base has been made important reference, the process of venture analysis that Here it is afterwards.

Subordinate function is selected triangular membership functions, fuzzy rule is by tri-factor S of FMEA (order of severity of fault), O (occurrence frequency of fault), D (complexity of fault finding) decides ultimate risk effect in actual conditions.

Other step and parameter are identical with embodiment one.

Embodiment three: present embodiment is different from embodiment one or two: the assessed cost to whole analogue system in described step 2 is estimated:

The first step, according to the risk evaluation result of each fault of the analogue system in step 1, determines the assessment activity of analogue system;

Second step, according to the acceptable greateset risk R of the risk evaluation result of each fault of the analogue system in step 1 and identification _sim, require the possible assessment activity of expert advice;

The 3rd step, to each possible assessment activity, expert guarantees possible assessed cost; Wherein, described assessed cost is a function about the time, is a tlv triple T=(T by time representation _p, T _m, T _o); Wherein, described T _pfor the time of pessimistic value, T _ofor most probable time, T _mthe time of the value of showing optimism;

The 4th step, according to the expert of the result verification assessed cost of historical information, draws the result relatively reliably;

The 5th step, obtains evaluating the estimation of integrated cost;

Calculate the assessment of evaluating cost with identical arithmetic mean:

$V_j=(V_{\mathrm{pj}},V_{\mathrm{mj}},V_{\mathrm{oj}})=N\·\mathrm{CP}\·T_j=(\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{pji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{mji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{oji}})$

In formula, m---the expert's that estimation survey fee is used number;

N---carry out assessment activity expert's number;

CP---unit interval expert's expense;

T _j---the evaluation time;

V _j---the arithmetic mean of assessment of failure cost;

V _pj---the pessimistic arithmetic mean of assessment of failure cost;

V _mj---the optimistic arithmetic mean of assessment of failure cost;

V _oj---the most probable arithmetic mean of assessment of failure cost;

T _pji---the evaluation time of pessimistic value;

T _mji---the evaluation time of optimistic value;

T _oji---the evaluation time of most probable value.

In present embodiment, assessed cost has referred to the total cost that assessment action need pays.Assessed cost can be divided into two parts: fixed cost (C _f) and variable cost (C _c).Fixed cost mainly comprises perms's cost, equipment (as computer, the printer etc.) cost of buying, purchase instrument (as software, testing tool etc.) cost etc.Variable cost comprises the cost of transient worker, leased equipment, lease instrument cost and other costs (as travel expenses etc.).The computing formula of cost is multiplied by long-run cost rate service time (T) for (C/U).Because fixed cost is constant, therefore in the process of evaluation of programme design, attention be variable cost, set up the relation between risk and variable cost.In the ensuing part of present embodiment, if not otherwise specified, assessed cost refers to variable cost.

At conceptual model Qualify Phase, for example, need assessment activity may comprise examination, checks, desk checking [software] etc.; Need to be in result verification, assessment activity may comprise statistical testing of business cycles, sensitivity analysis, surface test etc.It is different that these assessment mobilities and the assessed cost needing reduce risk.The analogue system of each part, in order to apply risk to acceptable level, can have multiple assessment activities available.

Other step and parameter are identical with embodiment one or two.

Embodiment four: present embodiment is different from one of embodiment one to three: the computation model of setting up the risk and cost of whole analogue system in described step 3 is specially:

Application fuzzy inference system obtains according to the regression model of the Fuzzy Linear of use Triangular Fuzzy Number, and the relation table of the risk and cost of each fault is shown following mathematical model:

R＝R _sim-c·v

In formula, c=(c _p, c _m, c _o)---Triangular Fuzzy Number; Wherein, c _prepresent the Triangular Fuzzy Number of pessimistic value, c _mthe Triangular Fuzzy Number of the value of showing optimism, c _orepresent the Triangular Fuzzy Number of most probable value;

R---fuzzy risk variable;

V---survey fee variable;

In formula, Triangular Fuzzy Number c=(c _p, c _m, c _o) determined by equation below:

R′＝R _sim-u·v

Make (R ', v) equal respectively (R _max, V _p), (R _max, V _m) and (R _max, V _o), can obtain respectively u=c _p, u=c _mand u=c _o.

Other step and parameter are identical with one of embodiment one to three.

Embodiment five: one of present embodiment and embodiment one to four are different: evaluate the relation between the Fuzzy Linear cost and risk of whole analogue system in described step 4, the fuzzy linear programming model of setting up assessment solution, the result that obtains final whole risk assessment is specially:

By the relation between the Fuzzy Linear cost and risk of the whole analogue system of evaluation, the problem of setting up the fuzzy linear programming model of assessment solution is converted into the problem of the least risk of asking for whole analogue system:

$\mathrm{Minimize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{R}_j=\mathrm{Minimize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}(R_{\mathrm{simj}}-c_j{v}_j)$

In formula, R _j---the risk variable of j fault;

R _simj---the risk of j fault obtaining in venture analysis;

C _j---Triangular Fuzzy Number;

V _j---the cost variations of j fault;

Due to R _simjconstant, so the problems referred to above can be converted into problem below:

Objective function: $\mathrm{Maximize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\left(c_j{v}_j\right)$

Constraint condition: $\left\{\begin{array}{c}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{v}_j\≤V_t\\ v_{\mathrm{pj}}\≤v_j\≤v_{\mathrm{oj}}\end{array}\right.$

In formula, V _trepresent given assessed cost, v _pjrepresent the cost variations of j fault pessimistic value, v _ojrepresent the cost variations of j fault optimistic value;

Fuzzy Linear Programming Problems is solved by maximum-likelihood method;

Objective function: $\mathrm{Maximize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\left(\frac{(c_{\mathrm{pj}}+4c_{\mathrm{mj}}+c_{\mathrm{oj}})}{6}{v}_j\right)$

Constraint condition: $\left\{\begin{array}{c}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{v}_j\≤V_t\\ v_{\mathrm{pj}}\≤v_j\≤v_{\mathrm{oj}}\end{array}\right.,$

By obtaining the assessed cost of every fault and the objective function of risk in system, complete the analogue system methods of risk assessment based on fuzzy FMEA.

Other step and parameter are identical with one of embodiment one to four.

Emulation experiment:

One, the design of software frame

Consider functional and every demand at our interface, in design, adopt scroll bar input and two kinds of input modes of form input, built a gui interface, and its interpolation assembly is completed to the function that we want.The basic framework of GUI is as Fig. 4;

This emulation experiment has added three scroll bars to be S (order of severity of fault), O (occurrence frequency of fault), D (complexity of fault finding) is three inputs of fuzzy reasoning, and R (the risk assessment value in analogue system) is output.At this moment our backstage will enter the logical operation of fuzzy FMEA, completes the calculating of last R (the risk assessment value in analogue system) above by each process of the FIS that narrates.

Due to the output also having assessed cost, in our design, if in the time starting to calculate, do not exceed our default maximum risk value if we obtain the risk assessment value of last analogue system, that is to say that so existing risk, in we stand scope, does not spend and detecting in this respect, so at this moment, our assessed cost will can not exist, and namely now no matter our detection time assessed value is how many, and the value of assessed cost is 0.

According to said above, we have adopted the way of two kinds of inputs in the time of design, not only independent but also be related between these two kinds of input methods, the basic framework of our GUI user interface while being list input below, except input mode changes, the function of other assemblies and put basic identical, as Fig. 5.

Two, the function of user interface realizes

It is below the postrun interfacial effect of gui interface, due to represent some core function tool boxes M file also do not add, temporary transient only can reaching inputted function into this evaluating system due to S in the fuzzy FMEA of this subject by data, O, tri-data of D are all in 1 to 5 five grade, so interface is selected to slidably input between 1-5 and is chosen, as Fig. 6;

The main interface of GUI, after operation GUI, complete the typing of input data, definite threshold, the functions such as assessment result wherein, S (order of severity of fault), O (occurrence frequency of fault), the acquiescence default value of three inputs that D (complexity of fault finding) is fuzzy reasoning is 3.Manually input Rmax, default value is 1.Manually input evaluation time default value is 0.Before starting calculating, our all outputs are 0.

Touch the button " starting to calculate " to inputted data process processing assessment, comprising S (order of severity of fault), O (occurrence frequency of fault), the R (the risk assessment value in analogue system) that three inputs that D (complexity of fault finding) is fuzzy reasoning infer through fuzzy FMEA, as Fig. 7.

And " R " that calculate compared with the greateset risk Rmax value in input frame now, if the value-at-risk R obtaining is less than greateset risk, our assessed cost will be 0 so, because now do not need, this risk is detected; As Fig. 8;

Otherwise, but the analogue system risk assessment value that we obtain is greater than greateset risk, maximum time, expected time, these group three item numbers certificates of minimum time that system will be inputted according to us are carried out fuzzy linear programming calculating, show that one group of respective value of our last assessed cost is to obtain optimum solution.

Other subsidiary functions: click " help " link paper; Click " closing interface " and close GUI; Click " switching to form input " and be just switched to form input.In the inputs in the face of the fuzzy FMEA of many groups, the advantage of sheet format input is very clear, and our any one group of input of amendment that can be at any time.We are known in above-mentioned process.In this part, the function of our assembly in GUI user interface is substantially identical with the function of tumbling-type input above with position.As Fig. 9;

Click " switch to and roll bar input ", be switched to and roll bar input type interface, namely the initial interface of our whole GUI user interface.

It is exactly more than the repertoire that our GUI user interface realizes analogue system risk assessment.

Utilize the GUI graphical interfaces of MATLAB, what the tool boxes such as FIS fuzzy reasoning editing machine realized that we mention carries out risk assessment based on fuzzy FMEA to the process of modeling and simulation, complete human nature close friend's inputting interface, realize our desirable Simulation Evaluation, the function of assessment of cost, for the theoretical research of whole embodiment be a kind of practice be also to verify in one.

Claims (5)

1. the analogue system methods of risk assessment based on fuzzy FMEA, is characterized in that the analogue system methods of risk assessment based on fuzzy FMEA is realized according to the following steps:

One, based on fuzzy FMEA, the risk of whole analogue system is carried out to entry evaluation, obtain the result of obfuscation: very low VL, low L, middle M, high H and very high VL;

Two, adopt $V_j=(V_{\mathrm{pj}},V_{\mathrm{mj}},V_{\mathrm{oj}})=N\·\mathrm{CP}\·T_j=(\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{pji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{mji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{oji}})$ Assessed cost to whole analogue system is estimated;

Three, set up the computation model R=R of the risk and cost of whole analogue system _sim-cv;

Four, utilize $\mathrm{Maximize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\left(\frac{(c_{\mathrm{pj}}+4c_{\mathrm{mj}}+c_{\mathrm{oj}})}{6}{v}_j\right)$ Constraint condition: $\left\{\begin{array}{c}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{v}_j\≤V_t\\ v_{\mathrm{pj}}\≤v_j\≤v_{\mathrm{oj}}\end{array}\right.$ Evaluate the relation between the Fuzzy Linear cost and risk of whole analogue system, set up the fuzzy linear programming model of assessment solution, obtain the result of final whole risk assessment, completed the analogue system methods of risk assessment based on fuzzy FMEA.

2. the analogue system methods of risk assessment based on fuzzy FMEA according to claim 1, is characterized in that in described step 1, based on fuzzy FMEA, whole analogue system being carried out to entry evaluation is specially:

(1) fault mode of identification analogue system, arranges and is made as R1, R2....Rn by contingent fault;

(2) the occurrence frequency O of the order of severity S to each contingent fault, fault and the complexity D of fault finding carry out expert assessment and evaluation classification grading, obtain the result of obfuscation; Wherein, the result of described obfuscation comprises very low VL, low L, middle M, high H and very high VL, therefore, and definition word collection S={VL, L, M, H, VH};

(3) adopt the arithmetic mean of multiple expert assessments and evaluations be multiple experts' comprehensive evaluation value as the risk evaluation result of each fault, its formula is as follows:

$F_j=\left(\frac{1}{m}\mathrm{Fij}\right)$

In formula, m---expert's number;

F _ij---order of severity S, the occurrence frequency O of fault and the complexity D of the fault finding classification grading of expert to fault;

F _j---the arithmetic mean of trying to achieve;

(4) application fuzzy inference system obtains the risk evaluation result of each fault according to multiple experts' comprehensive evaluation value, and the risk evaluation result of each fault is S1, S2...Sn, O1, O2...On, D1, D2...Dn.

3. the analogue system methods of risk assessment based on fuzzy FMEA according to claim 2, is characterized in that in described step 2, the assessed cost to whole analogue system is estimated:

The first step, according to the risk evaluation result of each fault of the analogue system in step 1, determines the assessment activity of analogue system;

Second step, according to the acceptable greateset risk R of the risk evaluation result of each fault of the analogue system in step 1 and identification _sim, require the possible assessment activity of expert advice;

The 3rd step, to each possible assessment activity, expert guarantees possible assessed cost; Wherein, described assessed cost is a function about the time, is a tlv triple T=(T by time representation _p, T _m, T _o); Wherein, described T _pfor the time of pessimistic value, T _ofor most probable time, T _mthe time of the value of showing optimism;

The 4th step, according to the expert of the result verification assessed cost of historical information, draws the result relatively reliably;

The 5th step, obtains evaluating the estimation of integrated cost;

Calculate the assessment of evaluating cost with identical arithmetic mean:

$V_j=(V_{\mathrm{pj}},V_{\mathrm{mj}},V_{\mathrm{oj}})=N\·\mathrm{CP}\·T_j=(\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{pji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{mji}},\frac{N\·\mathrm{CP}}{m}\underset{i=1}{\overset{m}{\mathrm{\Σ}}}{T}_{\mathrm{oji}})$

In formula, m---the expert's that estimation survey fee is used number;

N---carry out assessment activity expert's number;

CP---unit interval expert's expense;

T _j---the evaluation time;

V _j---the arithmetic mean of assessment of failure cost;

V _pj---the pessimistic arithmetic mean of assessment of failure cost;

V _mj---the optimistic arithmetic mean of assessment of failure cost;

V _oj---the most probable arithmetic mean of assessment of failure cost;

T _pji---the evaluation time of pessimistic value;

T _mji---the evaluation time of optimistic value;

T _oji---the evaluation time of most probable value.

4. the analogue system methods of risk assessment based on fuzzy FMEA according to claim 3, is characterized in that the computation model of setting up the risk and cost of whole analogue system in described step 3 is specially:

Application fuzzy inference system obtains according to the regression model of the Fuzzy Linear of use Triangular Fuzzy Number, and the relation table of the risk and cost of each fault is shown following mathematical model:

R＝R _sim-c·v

In formula, c=(c _p, c _m, c _o)---Triangular Fuzzy Number; Wherein, c _prepresent the Triangular Fuzzy Number of pessimistic value, c _mthe Triangular Fuzzy Number of the value of showing optimism, c _orepresent the Triangular Fuzzy Number of most probable value;

R---fuzzy risk variable;

V---survey fee variable;

In formula, Triangular Fuzzy Number c=(c _p, c _m, c _o) determined by equation below:

R′＝R _sim-u·v

Make (R ', v) equal respectively (R _max, V _p), (R _max, V _m) and (R _max, V _o), can obtain respectively u=c _p, u=c _mand u=c _o.

5. the analogue system methods of risk assessment based on fuzzy FMEA according to claim 4, it is characterized in that evaluating in described step 4 the relation between the Fuzzy Linear cost and risk of whole analogue system, the fuzzy linear programming model of setting up assessment solution, the result that obtains final whole risk assessment is specially:

By the relation between the Fuzzy Linear cost and risk of the whole analogue system of evaluation, the problem of setting up the fuzzy linear programming model of assessment solution is converted into the problem of the least risk of asking for whole analogue system:

$\mathrm{Minimize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{R}_j=\mathrm{Minimize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}(R_{\mathrm{simj}}-c_j{v}_j)$

In formula, R _j---the risk variable of j fault;

R _simj---the risk of j fault obtaining in venture analysis;

C _j---Triangular Fuzzy Number;

V _j---the cost variations of j fault;

Due to R _simjconstant, so the problems referred to above can be converted into problem below:

Objective function: $\mathrm{Maximize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\left(c_j{v}_j\right)$

Constraint condition: $\left\{\begin{array}{c}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{v}_j\≤V_t\\ v_{\mathrm{pj}}\≤v_j\≤v_{\mathrm{oj}}\end{array}\right.$

In formula, V _trepresent given assessed cost, v _pjrepresent the cost variations of j fault pessimistic value, v _ojrepresent the cost variations of j fault optimistic value;

Fuzzy Linear Programming Problems is solved by maximum-likelihood method;

Objective function: $\mathrm{Maximize}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}\left(\frac{(c_{\mathrm{pj}}+4c_{\mathrm{mj}}+c_{\mathrm{oj}})}{6}{v}_j\right)$

Constraint condition: $\left\{\begin{array}{c}\underset{j=1}{\overset{n}{\mathrm{\Σ}}}{v}_j\≤V_t\\ v_{\mathrm{pj}}\≤v_j\≤v_{\mathrm{oj}}\end{array}\right.,$

By obtaining the assessed cost of every fault and the objective function of risk in system, complete the analogue system methods of risk assessment based on fuzzy FMEA.

Images