CN109612173A - A kind of assessment of fault and diagnostic method of vapor cycle refrigeration system - Google Patents
A kind of assessment of fault and diagnostic method of vapor cycle refrigeration system Download PDFInfo
- Publication number
- CN109612173A CN109612173A CN201811358466.3A CN201811358466A CN109612173A CN 109612173 A CN109612173 A CN 109612173A CN 201811358466 A CN201811358466 A CN 201811358466A CN 109612173 A CN109612173 A CN 109612173A
- Authority
- CN
- China
- Prior art keywords
- fault
- transition
- library
- failure
- individual
- 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.)
- Pending
Links
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F25—REFRIGERATION OR COOLING; COMBINED HEATING AND REFRIGERATION SYSTEMS; HEAT PUMP SYSTEMS; MANUFACTURE OR STORAGE OF ICE; LIQUEFACTION SOLIDIFICATION OF GASES
- F25B—REFRIGERATION MACHINES, PLANTS OR SYSTEMS; COMBINED HEATING AND REFRIGERATION SYSTEMS; HEAT PUMP SYSTEMS
- F25B49/00—Arrangement or mounting of control or safety devices
- F25B49/02—Arrangement or mounting of control or safety devices for compression type machines, plants or systems
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M99/00—Subject matter not provided for in other groups of this subclass
- G01M99/005—Testing of complete machines, e.g. washing-machines or mobile phones
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mechanical Engineering (AREA)
- Thermal Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention discloses a kind of assessment of fault of vapor cycle refrigeration system and diagnostic methods.This method combines fuzzy control with Fault Petri Net, carries out assessment of fault by forward reasoning, determines that failure causes path and fault severity level and confidence level；Fault diagnosis is carried out by backward inference, determines that most probable causes the source of trouble and the preferential diagnostic sequence of failure of failure.Meanwhile assessment of fault and Diagnostic parameters are optimized by mimicry physics, so as to improve positive and backward inference accuracy, strengthening and improving pattern pastes the accident analysis and diagnosis capability of Fault Petri Net.
Description
Technical field
The invention belongs to thermodynamics, fault diagnosis field, in particular to a kind of failure of vapor cycle refrigeration system is commented
Valence and diagnostic method.
Background technique
With the raising that aeroplane performance requires, used electronic equipment is more and more, and heating power is also increasing,
Conventional air refrigeration cycle will increase making for engine bleed and ram-air since performance is lower under conditions of big thermal force
Dosage causes aircraft fuel oil panelty to rise rapidly, and also influences the Stealth Fighter of aircraft.Compared to air refrigeration cycle, steam
Hair refrigeration cycle performance coefficient is higher, so the synthesis environmental control system for introducing vapor cycle refrigeration system is significant.Also because
This, vapor cycle refrigeration system, which breaks down, to have significant impact to aircraft flight.Vapor cycle refrigeration system breaks down original
Because numerous, fault detection larger workload, a large amount of human and material resources need to be spent by checking to be out of order source and solve failure.Therefore right
Vapor cycle refrigeration system establishes the development that fault diagnosis model is conducive to real work.
The failure of vapor cycle refrigeration system has derivative and concurrency, is frequently not single event once failure occurs
Barrier, a failure symptom can derive the failure of correlation, and the failure of a type will lead to the generation of polymorphic type failure symptom.
When therefore carrying out fault diagnosis research, how rapidly to analyze the mechanism of transmission for generation of being out of order and carry out fault location as one
Item urgent problem to be solved.The method of breakdown of refrigeration system diagnosis at present has neural network, expert system, fault tree, obscures and examine
Break.Neural network needs enough learning samples just and can guarantee the reliability of diagnosis；Expert system acquisition knowledge is more difficult,
And the knowledge of expert system has inconsistency, imperfection and uncertainty when knowledge acquisition, and system can not self-perfection；
Fault tree analysis method achievement is cumbersome, is easy mistakes and omissions.
Currently, in fault diagnosis field, Petri network is mainly used for the logical relation of expression system, complete on information table and
Diagnostic reasoning, it lays particular emphasis on the expression of figure, lacks specific description to the dynamic characteristic of net.If Petri network is directly applied
In vapor cycle refrigeration system fault diagnosis, the complexity, dynamic how comprehensive consideration system failure is propagated needed to solve
The problem of fault message of feature and system more complete expression.
Summary of the invention
In order to solve the technical issues of above-mentioned background technique proposes, the present invention is intended to provide a kind of vapor cycle refrigeration system
Assessment of fault and diagnostic method, improve the speed and reliability of diagnostic analysis, express propagation characteristic and spread state well.
In order to achieve the above technical purposes, the technical solution of the present invention is as follows:
A kind of assessment of fault and diagnostic method of vapor cycle refrigeration system, comprising the following steps:
(1) according to the failure effect analysis (FEA) table FMEA of vapor cycle refrigeration system, the improved Fuzzy failure of system is established
Petri net model IFFPN is defined as one ten three-number set IFFPN=(P, T, I, O, Tt, TK, C, M, α, Uμ,λ,w,
F), in which:
P={ p1,p2,…,pnIt is set of library, piFor one of library institute, i=1,2 ..., n indicate system fault event
Set；T={ t1,t2,…,tmIt is transition collection, tjFor one of transition, j=1,2 ..., m indicate the state change of failure,
The replacement in reaction system fault propagation stage changes；I:P × T is input matrix, and expression is transitted towards library institute pi→tjBetween there are
To arc, as transition tjInput arc, and piTo change tjInput matrix；O:T × P is input matrix, and library representation arrives transition
tj→piBetween there are directed arc, as transition tjOutput arc, and piTo change tjOutput matrix；Tt={ tjIt (k) } is point
Fire transition set, k is the duration of ignition；TK={ tk1,tk2,…,tknIt is Tokken collection, tkiFor piTokken, indicate fault message；
C: for willing color of torr set, different colours indicate different confidence levels；M=(m1,m2,…,mn)TBy library identify distribution to
Amount, miRepresent library institute piTokken number, Tokken number indicates failure path number and fault severity level size；α=(α1,
α2,…,αn)TFor event of failure confidence level vector, αiLibrary representation institute piThe confidence level of really degree, subscript T indicate transposition；Uμ=
diag(μ1,μ2,…,μm) it is transition rule reliability matrix, μjIndicate transition tjConfidence level；λ=(λ1,λ2,…,λm)TTo become
Move threshold vector, λjIndicate transition tjThreshold value；W=(w1,w2,…,wn)TFor library institute weight vector, wiLibrary representation institute piPower
Value；F={ f1,f2,…,fnIt is event of failure fuzzy failure rate set, fiLibrary representation institute piFuzzy failure rate；
(2) according to expert system, historical data primarily determine bottom event of failure confidence level, transition confidence level, transition threshold value,
Event of failure weight, and transition threshold value and transition confidence level are optimized using mimicry physics optimization algorithm；
(3) when failure does not occur, initial marking M is determined according to potential failure symptom0Forward reasoning is carried out, intelligence is passed through
Energy reasoning is deduced, and fault propagation path, library institute's fault severity level and confidence level are obtained；
(4) when an error occurs, initial marking is determinedIt carries out backward inference and failure is obtained by intelligent deduction
The priority of diagnosis.
Further, in step (2), transition threshold value and transition confidence level are carried out by mimicry physics optimization algorithm
The step of optimization, is as follows:
(201) according to given improved Fuzzy Fault Petri Net pessimistic concurrency control, the threshold value and confidence level for needing to optimize are determined
Number, so that it is determined that the dimension X of population at individual；
(202) inputting one group of threshold value and confidence packets is optimization aim, as the position vector of particle, according to reasoning letter
Number obtains the real output value of this group of data, calculates the mean square deviation of sampleWith mean square deviationAs the criterion of sample, i.e.,
Fitness function, the performance of evaluation parameter find global optimum's individual:
In above formula, E is sample variance, and N is number of samples, and X is population scale；pijJ-th for i-th of sample is ideal defeated
It is worth out, p'ijFor j-th of real output value of i-th of sample；
(203) population size is set as n, and the quality of i-th of individual is
In above formula, f (xi) be individual i adaptive value, f (xbest) indicate optimum individual target function value, f (xworst) table
Show the target function value of worst individual；
Fictitious force F of the individual i in kth dimension by individual jij,kAnd resultant force of the individual i in kth dimension by other individuals
Fi,kIt is as follows:
In above formula, xi,k、xj,kFor the position of individual i, j in kth dimension, xj,k-xi,kIndicate that individual j to individual i is tieed up in kth
On distance, G is Gravity factor；
(204) the more speed of new individual i and position:
In above formula, vi,kFor the speed of individual i, j in kth dimension, t indicates time, w1For inertia weight, λ1For random value,
w1,λ1∈(0,1)；
(205) updated ideal adaptation angle value is calculated, adaptive optimal control angle value and worst fitness value is updated, judges whether
Meet termination condition, satisfaction then stops calculating, and exports optimal threshold value and confidence level result；If not satisfied, repeating step
(203)-(205)。
Further, specific step is as follows for step (3):
(301) when not breaking down, forward reasoning is carried out, whole library institutes are obtained according to MYCIN confidence level reasoning algorithm
Confidence value, work as αk+1=αkWhen, reasoning terminates, and obtains each event confidence alpha that kth time reasoning obtainsk；
(302) whenWhen establishment, transition are enabled, obtain positive potential enabled transition igniting sequence U (t)
=(U (t1),U(t2),…,U(tm))T, in which:
In above formula, b represents an infinitely great value；
(303) intelligent inference obtains positive transition igniting sequence Uk, work as Uk=(0,0 ..., 0)TWhen stop reasoning, then can obtain
It lights a fire to forward direction and changes set Tt={ tj(k) } and igniting after the identified distribution vector M in libraryk:
In above formula, 1n×mFor the unit matrix of n × m；ITFor the transposed matrix of unit battle array I, " ∧ " is that matrix takes small operator,It is rounded operator for matrix, " ⊙ " is matrix product operator；
(304) it is lighted a fire by forward direction and changes set Tt={ tj(k) } fault propagation path is obtained, is identified by the library after lighting a fire
MkLibrary institute's fault severity level and confidence level are obtained, forward reasoning result is obtained.
Further, in step (301), confidence calculations formula is as follows:
In above formula,For the matrix multiply operator,Big operator is taken for matrix.
Further, specific step is as follows for step (4):
(401) when faulty generation, backward inference is carried out, whole library institutes are obtained according to MYCIN confidence level reasoning algorithm
State value, whenWhen, reasoning terminates, and obtains each event confidence level that kth time reasoning obtains
(402) whenWhen establishment, transition are enabled, obtain reverse potential enabled transition igniting sequence U-
(t)；
(403) intelligent inference obtains inversely changing igniting sequenceWhenWhen stop reasoning, then may be used
Obtain the transition set Tt that inversely lights a fire-={ tj(k) } and the library after igniting is identified
Wherein, subscript "-" represents and reverse value that parameter is corresponded in step (3)；
(404) intelligent inference has inversely been sent out fault estimator A up to process*, asked most further according to Minimal Cut Set
Small cut set G1,G2,…,GS, s be less than library sum n,And the library institute number that each minimal cut set includes is different；
Fault estimator A is sent out*Calculating formula it is as follows:
A=I-O
(405) the easy hair rate for calculating minimal cut set can obtain fault diagnosis priority by minimal cut set easily hair rate height, obtain
To fault converse inference result:
In above formula, d (Gl) indicate minimal cut set easy hair rate, d (pi) indicate minimal cut set GlIn the easy hair that is arrived of each library
Rate.
By adopting the above technical scheme bring the utility model has the advantages that
The present invention combines fuzzy control with Fault Petri Net, introduces mimicry physics optimization algorithm, proposes to be based on changing
Into fuzzy fault Petri network vapor cycle refrigeration system assessment of fault and diagnostic method.In order to improve the standard of fault reasoning
True property enhances the analysis and diagnosis ability of Petri network, carries out parameter optimization to the threshold value of Petri network, weight and confidence level.It will
Mimicry physics optimization algorithm (APO) is combined with Fuzzy Petri Net, and mimicry physics optimization algorithm passes through virtual between individual
Power effect changes speed and the position of individual, mobile towards optimization aim, finally converges on globally optimal solution.With traditional optimization
Algorithm (such as genetic algorithm with particle swarm optimization algorithm) is compared, and APO has preferable global search and avoids falling into local optimum
Ability, and fast convergence rate, stability are good.
Detailed description of the invention
Fig. 1 is overall flow figure of the invention；
Fig. 2 is the flow chart that mimicry physics algorithm optimization threshold value and confidence level are utilized in the present invention.
Specific embodiment
Below with reference to attached drawing, technical solution of the present invention is described in detail.
The present invention devises the assessment of fault and diagnostic method of a kind of vapor cycle refrigeration system, as shown in Figure 1, specific mistake
Journey is as follows.
Step 1: according to the failure effect analysis (FEA) table FMEA of vapor cycle refrigeration system, establishing the improved Fuzzy event of system
Hinder Petri net model IFFPN, is defined as one ten three-number set IFFPN=(P, T, I, O, Tt, TK, C, M, α, Uμ,λ,
W, f), in which:
P={ p1,p2,…,pnIt is set of library, piFor one of library institute, i=1,2 ..., n indicate system fault event
Set；T={ t1,t2,…,tmIt is transition collection, tjFor one of transition, j=1,2 ..., m indicate the state change of failure,
The replacement in reaction system fault propagation stage changes；I:P × T is input matrix, and expression is transitted towards library institute pi→tjBetween there are
To arc, as transition tjInput arc, and piTo change tjInput matrix；O:T × P is input matrix, and library representation arrives transition
tj→piBetween there are directed arc, as transition tjOutput arc, and piTo change tjOutput matrix；Tt={ tjIt (k) } is point
Fire transition set, k is the duration of ignition；TK={ tk1,tk2,…,tknIt is Tokken collection, tkiFor piTokken, indicate fault message；
C: for willing color of torr set, different colours indicate different confidence levels；M=(m1,m2,…,mn)TBy library identify distribution to
Amount, miRepresent library institute piTokken number, Tokken number indicates failure path number and fault severity level size；α=(α1,
α2,…,αn)TFor event of failure confidence level vector, αiLibrary representation institute piThe confidence level of really degree, subscript T indicate transposition；Uμ=
diag(μ1,μ2,…,μm) it is transition rule reliability matrix, μjIndicate transition tjConfidence level；λ=(λ1,λ2,…,λm)TTo become
Move threshold vector, λjIndicate transition tjThreshold value；W=(w1,w2,…,wn)TFor library institute weight vector, wiLibrary representation institute piPower
Value；F={ f1,f2,…,fnIt is event of failure fuzzy failure rate set, fiLibrary representation institute piFuzzy failure rate.
Step 2: bottom event of failure confidence level, transition confidence level, transition threshold are primarily determined according to expert system, historical data
Value, event of failure weight, and transition threshold value and transition confidence level are optimized using mimicry physics optimization algorithm.
Preferably, as shown in Fig. 2, can be optimized using following steps to transition threshold value and transition confidence level:
Step 201: according to given improved Fuzzy Fault Petri Net pessimistic concurrency control, determining the threshold value for needing to optimize and credible
The number of degree, so that it is determined that the dimension X of population at individual.
Step 202: one group of threshold value of input and confidence packets are optimization aim, as the position vector of particle, according to pushing away
Reason function obtains the real output value of this group of data, calculates the mean square deviation of sampleWith mean square deviationJudgement mark as sample
Standard, i.e. fitness function, the performance of evaluation parameter find global optimum's individual:
In above formula, E is sample variance, and N is number of samples, and X is population scale；pijJ-th for i-th of sample is ideal defeated
It is worth out, p'ijFor j-th of real output value of i-th of sample.
Step 203: setting population size as n, the quality of i-th of individual is
In above formula, f (xi) be individual i adaptive value, f (xbest) indicate optimum individual target function value, f (xworst) table
Show the target function value of worst individual.
Fictitious force F of the individual i in kth dimension by individual jij,kAnd resultant force of the individual i in kth dimension by other individuals
Fi,kIt is as follows:
In above formula, xi,k、xj,kFor the position of individual i, j in kth dimension, xj,k-xi,kIndicate that individual j to individual i is tieed up in kth
On distance, G is Gravity factor.
Step 204: the more speed of new individual i and position:
In above formula, vi,kFor the speed of individual i, j in kth dimension, t indicates time, w1For inertia weight, λ1For random value,
w1,λ1∈(0,1)。
Step 205: calculating updated ideal adaptation angle value, update adaptive optimal control angle value and worst fitness value, judgement
Whether termination condition is met, and satisfaction then stops calculating, and exports optimal threshold value and confidence level result；If not satisfied, repeating step
203-205。
Step 3: when failure does not occur, initial marking M being determined according to potential failure symptom0Forward reasoning is carried out, is led to
Intelligent inference deduction is crossed, fault propagation path, library institute's fault severity level and confidence level are obtained.Specific step is as follows:
Step 301: when not breaking down, carrying out forward reasoning, whole libraries are obtained according to MYCIN confidence level reasoning algorithm
Confidence value, work as αk+1=αkWhen, reasoning terminates, and obtains each event confidence alpha that kth time reasoning obtainsk；Confidence level meter
It is as follows to calculate formula:
In above formula:
Multiplication operator A, B, C are respectively m × q, q × n, m * n matrix, then
Take big operator A, B, C are m * n matrix, then cij=max (aij,bij), i=1,2 ..., m, j
=1,2 ..., n.
Step 302: whenWhen establishment, transition are enabled, obtain positive potential enabled transition igniting sequence U
(t)=(U (t1),U(t2),…,U(tm))T, in which:
B represents an infinitely great value
Step 303: intelligent inference obtains positive transition igniting sequence Uk, work as Uk=(0,0 ..., 0)TWhen stop reasoning, then
Positive igniting transition set Tt={ t can be obtainedj(k) } and igniting after the identified distribution vector M in libraryk:
In above formula, 1n×mFor the unit matrix of n × m；ITFor the transposed matrix of unit battle array I.
Taking small operator ∧: C=A ∧ B, A, B, C is m * n matrix, then cij=min (aij,bij), i=1,2 ..., m, j
=1,2 ..., n；
It is rounded operator A, B, C are respectively m × q, q × n, m * n matrix, then
Product Operator ⊙: c=A ⊙ b, A are m * n matrix, and B, C are respectively n, m dimensional vector, thenAnd
aij·bj≠ 0, i=1,2 ..., m, j=1,2 ..., n.
Step 304: being lighted a fire by forward direction and change set Tt={ tj(k) } fault propagation path is obtained, by the library institute after lighting a fire
Identify MkLibrary institute's fault severity level and confidence level are obtained, forward reasoning result is obtained.
Step 4: when an error occurs, determining initial markingBackward inference is carried out to obtain by intelligent deduction
The priority of fault diagnosis.Specific step is as follows:
Step 401: when faulty generation, carrying out backward inference, whole libraries are obtained according to MYCIN confidence level reasoning algorithm
State value, whenWhen, reasoning terminates, and obtains each event confidence level that kth time reasoning obtains
Step 402: whenWhen establishment, transition are enabled, obtain reverse potential enabled transition igniting sequence U-
(t)。
Step 403: intelligent inference obtains inversely changing igniting sequenceWhenWhen stop reasoning, then
Reverse igniting transition set Tt can be obtained-={ tj(k) } and the library after igniting is identified
Wherein, subscript "-" represents the reverse value that parameter is corresponded to step 3.
Step 404: intelligent inference has inversely been sent out fault estimator A up to process*, asked further according to Minimal Cut Set
Minimal cut set G1,G2,…,GS, s be less than library sum n,And the library institute number that each minimal cut set includes is not
Together；Fault estimator A is sent out*Calculating formula it is as follows:
A=I-O
Step 405: calculating the easy hair rate of minimal cut set, it is preferentially suitable to obtain fault diagnosis by minimal cut set easily hair rate height
Sequence obtains fault converse inference result:
In above formula, d (Gl) indicate minimal cut set easy hair rate, d (pi) indicate minimal cut set GlIn the easy hair that is arrived of each library
Rate.
Embodiment is merely illustrative of the invention's technical idea, and this does not limit the scope of protection of the present invention, it is all according to
Technical idea proposed by the present invention, any changes made on the basis of the technical scheme are fallen within the scope of the present invention.
Claims (5)
1. the assessment of fault and diagnostic method of a kind of vapor cycle refrigeration system, which comprises the following steps:
(1) according to the failure effect analysis (FEA) table FMEA of vapor cycle refrigeration system, the improved Fuzzy Fault Petri Net of system is established
Pessimistic concurrency control IFFPN is defined as one ten three-number set IFFPN=(P, T, I, O, Tt, TK, C, M, α, Uμ, λ, w, f),
In:
P={ p1,p2,…,pnIt is set of library, piFor one of library institute, i=1,2 ..., n indicate system fault event set；
T={ t1,t2,…,tmIt is transition collection, tjFor one of transition, j=1,2 ..., m indicate the state change of failure, react
The replacement of system failure propagation stage changes；I:P × T is input matrix, and expression is transitted towards library institute pi→tjBetween there are oriented
Arc, as transition tjInput arc, and piTo change tjInput matrix；O:T × P is input matrix, and library representation arrives transition tj
→piBetween there are directed arc, as transition tjOutput arc, and piTo change tjOutput matrix；Tt={ tjIt (k) } is igniting
Transition set, k is the duration of ignition；TK={ tk1,tk2,…,tknIt is Tokken collection, tkiFor piTokken, indicate fault message；C:
For willing color of torr set, different colours indicate different confidence levels；M=(m1,m2,…,mn)TDistribution vector is identified by library,
miRepresent library institute piTokken number, Tokken number indicates failure path number and fault severity level size；α=(α1,
α2,…,αn)TFor event of failure confidence level vector, αiLibrary representation institute piThe confidence level of really degree, subscript T indicate transposition；Uμ=
diag(μ1,μ2,…,μm) it is transition rule reliability matrix, μjIndicate transition tjConfidence level；λ=(λ1,λ2,…,λm)TTo become
Move threshold vector, λjIndicate transition tjThreshold value；W=(w1,w2,…,wn)TFor library institute weight vector, wiLibrary representation institute piPower
Value；F={ f1,f2,…,fnIt is event of failure fuzzy failure rate set, fiLibrary representation institute piFuzzy failure rate；
(2) bottom event of failure confidence level, transition confidence level, transition threshold value, failure are primarily determined according to expert system, historical data
Event weight, and transition threshold value and transition confidence level are optimized using mimicry physics optimization algorithm；
(3) when failure does not occur, initial marking M is determined according to potential failure symptom0Forward reasoning is carried out, is pushed away by intelligence
Reason is deduced, and fault propagation path, library institute's fault severity level and confidence level are obtained；
(4) when an error occurs, initial marking is determinedIt carries out backward inference and fault diagnosis is obtained by intelligent deduction
Priority.
2. the assessment of fault and diagnostic method of vapor cycle refrigeration system according to claim 1, which is characterized in that in step
(2) as follows to the step of threshold value is optimized with transition confidence level is changed by mimicry physics optimization algorithm in:
(201) according to given improved Fuzzy Fault Petri Net pessimistic concurrency control, of the threshold value and confidence level that need to optimize is determined
Number, so that it is determined that the dimension X of population at individual；
(202) it inputs one group of threshold value and confidence packets is that optimization aim is obtained as the position vector of particle according to inference function
The real output value of this group of data out calculates the mean square deviation of sampleWith mean square deviationAs the criterion of sample, that is, adapt to
Function is spent, the performance of evaluation parameter finds global optimum's individual:
In above formula, E is sample variance, and N is number of samples, and X is population scale；pijFor j-th of ideal output of i-th of sample
Value, p'ijFor j-th of real output value of i-th of sample；
(203) population size is set as n, and the quality of i-th of individual is
In above formula, f (xi) be individual i adaptive value, f (xbest) indicate optimum individual target function value, f (xworst) indicate most
The target function value of poor individual；
Fictitious force F of the individual i in kth dimension by individual jij,kAnd resultant force F of the individual i in kth dimension by other individualsi,kSuch as
Under:
In above formula, xi,k、xj,kFor the position of individual i, j in kth dimension, xj,k-xi,kIndicate individual j to individual i in kth dimension
Distance, G are Gravity factor；
(204) the more speed of new individual i and position:
In above formula, vi,kFor the speed of individual i, j in kth dimension, t indicates time, w1For inertia weight, λ1For random value, w1,λ1
∈(0,1)；
(205) updated ideal adaptation angle value is calculated, adaptive optimal control angle value and worst fitness value is updated, judges whether to meet
Termination condition, satisfaction then stop calculating, and export optimal threshold value and confidence level result；If not satisfied, repeating step (203)-
(205)。
3. the assessment of fault and diagnostic method of vapor cycle refrigeration system according to claim 1, which is characterized in that step
(3) specific step is as follows:
(301) when not breaking down, carry out forward reasoning, according to MYCIN confidence level reasoning algorithm obtain whole libraries set
Certainty value works as αk+1=αkWhen, reasoning terminates, and obtains each event confidence alpha that kth time reasoning obtainsk；
(302) whenWhen establishment, transition are enabled, obtain positive potential enabled transition igniting sequence U (t)=(U
(t1),U(t2),…,U(tm))T, in which:
In above formula, b represents an infinitely great value；
(303) intelligent inference obtains positive transition igniting sequence Uk, work as Uk=(0,0 ..., 0)TWhen stop reasoning, then can be obtained just
Set Tt={ t is changed to ignitingj(k) } and igniting after the identified distribution vector M in libraryk:
In above formula, 1n×mFor the unit matrix of n × m；ITFor the transposed matrix of unit battle array I, " ∧ " is that matrix takes small operator,For square
Battle array is rounded operator, and " ⊙ " is matrix product operator；
(304) it is lighted a fire by forward direction and changes set Tt={ tj(k) } fault propagation path is obtained, by the identified M in library after lighting a firek?
To library institute's fault severity level and confidence level, forward reasoning result is obtained.
4. the assessment of fault and diagnostic method of vapor cycle refrigeration system according to claim 3, which is characterized in that in step
(301) in, confidence calculations formula is as follows:
In above formula,For the matrix multiply operator,Big operator is taken for matrix.
5. the assessment of fault and diagnostic method of vapor cycle refrigeration system according to claim 3, which is characterized in that step
(4) specific step is as follows:
(401) when faulty generation, carry out backward inference, according to MYCIN confidence level reasoning algorithm obtain whole libraries shape
State value, whenWhen, reasoning terminates, and obtains each event confidence level that kth time reasoning obtains
(402) whenWhen establishment, transition are enabled, obtain reverse potential enabled transition igniting sequence U-(t)；
(403) intelligent inference obtains inversely changing igniting sequenceWhenWhen stop reasoning, then can be obtained inverse
Set Tt is changed to igniting-={ tj(k) } and the library after igniting is identified
Wherein, subscript "-" represents and reverse value that parameter is corresponded in step (3)；
(404) intelligent inference has inversely been sent out fault estimator A up to process*, minimal cut set is sought further according to Minimal Cut Set
G1,G2,…,GS, s be less than library sum n,And the library institute number that each minimal cut set includes is different；Event is sent out
Hinder incidence matrix A*Calculating formula it is as follows:
A=I-O
(405) the easy hair rate for calculating minimal cut set can obtain fault diagnosis priority by minimal cut set easily hair rate height, obtain event
Hinder backward inference result:
In above formula, d (Gl) indicate minimal cut set easy hair rate, d (pi) indicate minimal cut set GlIn the easy hair rate that is arrived of each library.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811358466.3A CN109612173A (en) | 2018-11-15 | 2018-11-15 | A kind of assessment of fault and diagnostic method of vapor cycle refrigeration system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811358466.3A CN109612173A (en) | 2018-11-15 | 2018-11-15 | A kind of assessment of fault and diagnostic method of vapor cycle refrigeration system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109612173A true CN109612173A (en) | 2019-04-12 |
Family
ID=66004022
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811358466.3A Pending CN109612173A (en) | 2018-11-15 | 2018-11-15 | A kind of assessment of fault and diagnostic method of vapor cycle refrigeration system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109612173A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110989561A (en) * | 2019-12-26 | 2020-04-10 | 中国航空工业集团公司沈阳飞机设计研究所 | Method for constructing fault propagation model and method for determining fault propagation path |
| CN111076962A (en) * | 2020-01-03 | 2020-04-28 | 中国电建集团华东勘测设计研究院有限公司 | Electromechanical equipment fault diagnosis method for intelligent hydraulic power plant |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS63282070A (en) * | 1987-05-15 | 1988-11-18 | 株式会社日立製作所 | Controller for hydraulic elevator |
| CN1183336A (en) * | 1996-10-18 | 1998-06-03 | 三洋电机株式会社 | Electric appliance quality decision method and refrigerating apparatus inspecting device |
| US5855011A (en) * | 1996-09-13 | 1998-12-29 | Tatsuoka; Curtis M. | Method for classifying test subjects in knowledge and functionality states |
| JP2005241089A (en) * | 2004-02-25 | 2005-09-08 | Mitsubishi Electric Corp | Apparatus diagnosing device, refrigeration cycle device, apparatus diagnosing method, apparatus monitoring system and refrigeration cycle monitoring system |
| CN102175460A (en) * | 2011-01-28 | 2011-09-07 | 中国人民解放军海军工程大学 | Gas path fault diagnostic method for marine gas turbine based on fuzzy Petri net (FPN) |
| CN102918274A (en) * | 2011-05-31 | 2013-02-06 | 日立空调·家用电器株式会社 | Method for detecting abnormality in refrigeration apparatus and apparatus therefor |
| CN107203202A (en) * | 2017-07-03 | 2017-09-26 | 贵州大学 | Systems reliability analysis and method for diagnosing faults based on Fuzzy Petri Net |
-
2018
- 2018-11-15 CN CN201811358466.3A patent/CN109612173A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS63282070A (en) * | 1987-05-15 | 1988-11-18 | 株式会社日立製作所 | Controller for hydraulic elevator |
| US5855011A (en) * | 1996-09-13 | 1998-12-29 | Tatsuoka; Curtis M. | Method for classifying test subjects in knowledge and functionality states |
| CN1183336A (en) * | 1996-10-18 | 1998-06-03 | 三洋电机株式会社 | Electric appliance quality decision method and refrigerating apparatus inspecting device |
| JP2005241089A (en) * | 2004-02-25 | 2005-09-08 | Mitsubishi Electric Corp | Apparatus diagnosing device, refrigeration cycle device, apparatus diagnosing method, apparatus monitoring system and refrigeration cycle monitoring system |
| CN102175460A (en) * | 2011-01-28 | 2011-09-07 | 中国人民解放军海军工程大学 | Gas path fault diagnostic method for marine gas turbine based on fuzzy Petri net (FPN) |
| CN102918274A (en) * | 2011-05-31 | 2013-02-06 | 日立空调·家用电器株式会社 | Method for detecting abnormality in refrigeration apparatus and apparatus therefor |
| CN107203202A (en) * | 2017-07-03 | 2017-09-26 | 贵州大学 | Systems reliability analysis and method for diagnosing faults based on Fuzzy Petri Net |
Non-Patent Citations (1)
| Title |
|---|
| 梁光夏: "基于改进模糊故障Petri网的复杂机电系统故障状态评价与诊断技术研究", 《CNKI自然科学与工程技术类专辑 工程科技Ⅱ辑》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110989561A (en) * | 2019-12-26 | 2020-04-10 | 中国航空工业集团公司沈阳飞机设计研究所 | Method for constructing fault propagation model and method for determining fault propagation path |
| CN111076962A (en) * | 2020-01-03 | 2020-04-28 | 中国电建集团华东勘测设计研究院有限公司 | Electromechanical equipment fault diagnosis method for intelligent hydraulic power plant |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109039739A (en) | A kind of power communication field failure diagnostic method improving fuzzy Petri | |
| CN106096830B (en) | Relay protection state evaluation method and system based on generalized evidence theory | |
| CN111598150B (en) | Transformer fault diagnosis method considering operation state grade | |
| CN110321603A (en) | A kind of depth calculation model for Fault Diagnosis of Aircraft Engine Gas Path | |
| CN109409587A (en) | A kind of airport excavated based on weather data is into traffic flow forecasting method of leaving the theatre | |
| CN109856969A (en) | A kind of failure prediction method and forecasting system based on BP neural network model | |
| CN104765965A (en) | GIS fault diagnosis and reliability analysis method based on fuzzy Petri | |
| CN106060008B (en) | A kind of network intrusions method for detecting abnormality | |
| CN109612173A (en) | A kind of assessment of fault and diagnostic method of vapor cycle refrigeration system | |
| CN105224738B (en) | LSSVM non-gaussian fluctuating wind speed prediction technique | |
| CN109597839B (en) | Data mining method based on avionic combat situation | |
| CN109412155B (en) | Power distribution network power supply capacity evaluation method based on graph calculation | |
| CN109117957A (en) | A kind of Sensitivity Analysis learnt based on Chinese Civil Aviation Maintenance quality evaluation and Manufacturing resource | |
| CN114785573A (en) | Intelligent substation process layer network abnormal flow detection method based on deep learning | |
| CN102968553B (en) | A kind of aircraft lands risk evaluation method | |
| CN113572771B (en) | Power grid CPS network attack identification method and system | |
| CN111083151B (en) | Attack identification method based on deep belief network and wind power management system | |
| CN111127251A (en) | Attack identification method based on LSTM neural network and grid-connected interface device | |
| CN106771847A (en) | A kind of 35kV power distribution networks transmission line lightning stroke Risk Forecast Method | |
| CN114692786B (en) | Civil aircraft fault diagnosis method based on multi-source data fusion | |
| Lyu et al. | Examination on avionics system fault prediction technology based on ashy neural network and fuzzy recognition | |
| Yun et al. | Diagnosis strategy design method based on ant colony algorithm | |
| CN114841063A (en) | Aero-engine residual life prediction method based on deep learning | |
| Li et al. | Application of Improved PSO-BP neural network in fault detection of liquid-propellant rocket engine | |
| Zhang et al. | Runway visual range prediction based on ensemble learning |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190412 |