CN109242093A - A kind of method for evaluating reliability of the motor in electric automobile based on fuzzy neural network - Google Patents
A kind of method for evaluating reliability of the motor in electric automobile based on fuzzy neural network Download PDFInfo
- Publication number
- CN109242093A CN109242093A CN201811178736.2A CN201811178736A CN109242093A CN 109242093 A CN109242093 A CN 109242093A CN 201811178736 A CN201811178736 A CN 201811178736A CN 109242093 A CN109242093 A CN 109242093A
- Authority
- CN
- China
- Prior art keywords
- evaluation
- neural network
- reliability
- fuzzy neural
- matrix
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/043—Architecture, e.g. interconnection topology based on fuzzy logic, fuzzy membership or fuzzy inference, e.g. adaptive neuro-fuzzy inference systems [ANFIS]
-
- 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
- G06Q10/06395—Quality analysis or management
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Development Economics (AREA)
- Educational Administration (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Software Systems (AREA)
- Strategic Management (AREA)
- Computational Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Business, Economics & Management (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Automation & Control Theory (AREA)
- Marketing (AREA)
- Fuzzy Systems (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Game Theory and Decision Science (AREA)
- Health & Medical Sciences (AREA)
- Tourism & Hospitality (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a kind of method for evaluating reliability of motor in electric automobile based on fuzzy neural network, the evaluation index for including the following steps: A1, establishing motor in electric automobile reliability;A2, assessment indicator system is established, formulates evaluation criterion;A3, according to the evaluation index and evaluation criterion, reliability standard is quantified based on fuzzy neural network model;A4, motor in electric automobile motor reliability is judged according to quantized result.It wherein, include that relative defects matrix is determined according to the assessment indicator system in step A3;Calculate the learning sample of fuzzy neural network;Learning sample is brought into fuzzy neural network model and is trained, and precision training is carried out to fuzzy neural network model;By reliability index data to be evaluated, substitutes into the fuzzy neural network model for reaching predetermined accuracy and calculated, obtain evaluation result.It is evaluated with fuzzy neural network model, solves the problems, such as the reliability evaluation under large sample quantity term, obtain reliability evaluation foundation.
Description
Technical field
The present invention relates to motor in electric automobile field more particularly to a kind of motor in electric automobile based on fuzzy neural network
Method for evaluating reliability.
Background technique
Motor is the important component of electric car, and especially for New-energy electric vehicle, driving motor requires work
Make that controllability is high, stable state accuracy is high, dynamic property is good, load requirement, technical performance and in terms of suffer from spy
The height of different requirement, reliability is even more service ability and the service life for directly influencing electric car, and user is fed back
Motor service condition reference can be provided for enterprise, the number such as feedback index such as first-time fault time, failure rate, maintenance time
According to being all the index for evaluating a motor reliability, therefore establish appraisement system by the feedback of existing user, can promote to look forward to
Industry is improved for weak link.
Fuzzy neural network combines the knowledge of neural network algorithm and fuzzy theory, summarizes neural network and fuzzy reason
By the advantages of, integrate study, association, identification, information processing, with powerful structured knowledge ability to express and powerful
Learning ability realizes fuzzy logic inference using neural network structure, is that artificial intelligence and intelligent automation are ground in recent years
Study carefully hot spot, how to be current urgent problem to be solved to motor reliability evaluation using fuzzy neural network.
Summary of the invention
The purpose of the present invention is to solve the evaluation index for how establishing motor reliability and utilize fuzzy neural network
The problem of to motor reliability evaluation, this application provides a kind of reliabilities of motor in electric automobile based on fuzzy neural network
Evaluation method.
In order to solve the above-mentioned technical problem, the application adopts the following technical scheme that
A kind of method for evaluating reliability of the motor in electric automobile based on fuzzy neural network, includes the following steps:
A1, the evaluation index for establishing motor in electric automobile reliability;
A2, assessment indicator system is established, formulates evaluation criterion;
A3, according to the evaluation index and evaluation criterion, reliability standard is quantified based on fuzzy neural network model;
A4, motor in electric automobile motor reliability is judged according to quantized result.
Preferably, in step A1, the evaluation index includes h index, and wherein h is greater than the positive integer equal to 1.
Preferably, in step A2, each index corresponds to m evaluation criterion in the evaluation index, the evaluation index with
The opinion rating standard constitutes reliability evaluation standard value matrix (Y=h*m), wherein the y in matrixijItem indicates i-th
Evaluation criterion value of the evaluation points in j-th of evaluation criterion, wherein m is greater than the positive integer equal to 1.
Preferably, reliability standard is quantified by model in step A3, included the following steps:
A31, relative defects matrix is determined according to the assessment indicator system;
A32, the learning sample that fuzzy neural network is calculated according to relative defects matrix;
A33, it learning sample is brought into fuzzy neural network model is trained, initialize fuzzy neural network, and
Precision training is carried out to fuzzy neural network model;
A34, by reliability index data to be evaluated, substitute into the fuzzy neural network model for reaching predetermined accuracy into
Row calculates, and obtains evaluation result.
Preferably, the 1st level evaluation standard value is set for the relative defects of fuzzy set as pi1=1, m level evaluation
Standard value is p for the relative defects of fuzzy setim=0, p betweenijIt is calculated with linear interpolation, gained member
Element is evaluation criterion index, and all elements constitute evaluation criterion index relative defects matrix.
Preferably, in step A2, if each group of data include h evaluation index in W group data to be tested, then (W* is constituted
H) the detection sample value matrix (X) of a element, wherein the xth in matrixijIndicate i-th of jth group detection data evaluation because
Subvalue, wherein W is greater than the positive integer equal to 1.
Preferably, the evaluation points value is compared with the evaluation criterion value, obtains the phase of opinion rating set
To degree of membership, constitute relative defects matrix (R).
Preferably, according to the difference of the evaluation points, the different evaluation points of setting and the same evaluation criterion value
Comparison result be different relative defects (r), and the relative defects in the case of remaining are calculated using linear interpolation.
Preferably, in the step A3, learning sample and test samples are obtained using interpolation method, Wherein:Indicate that i-th evaluation points of interpolation sample k are evaluated at j-th
Relative defects in grade, TkIndicate the corresponding levels of the standard value of interpolation sample k, 1≤i≤h, 1≤j≤m;Tk=j+q/n,
The value of wherein 1≤j≤n-1, n are adjusted according to the needs of interpolated sample.
Preferably, the learning sample is substituted into program to the initialization for carrying out fuzzy neural network, obtains training sample
Output matrix.
Preferably, the reliability step standard control table for setting each evaluation points, using each classification standard value as precision
The output matrix of training sample is normalized in the test samples of inspection according to the following formula,Its
In, i=1,2...m1, j=1,2...m2, xijIndicate value of i-th evaluation points in j-th of opinion rating,Expression is returned
One change treated value, m1Indicate the number of evaluation points, m2Indicate the number of opinion rating;It is defeated according to the target of test samples
Out with the comparison result of the reality output after neural computing, judge whether this neural network model reaches expected training
Precision.
Preferably, it will be evaluated in the computer program that expected training precision is reached described in data to be evaluated substitution
Obtain result.
Compared with prior art, the invention has the benefit that
A kind of method for evaluating reliability of motor in electric automobile based on fuzzy neural network of the invention, can by establishing
By the assessment indicator system of property, according to the evaluation index and evaluation criterion, reliability standard is quantified based on fuzzy neural network,
Motor in electric automobile motor reliability is judged according to quantized result, realizes evaluation index digitization.
Further, using relative defects matrix, for the reliability standard of description indexes, each reliability index
Data are converted to index for the relative defects of opinion rating set, and abstract evaluation content is changed into data, can
It enough substitutes into mathematical model and is calculated.
Further, the learning sample that fuzzy neural network is calculated by relative defects matrix, is improved with interpolation method
Training precision is trained neural network model using learning sample.
Further, the model that test samples substitute into after study is calculated, target output and warp by test samples
Reality output after crossing neural computing carries out error analysis, judges the neural network model after learning sample training
Whether resultant error meets the requirements.
Further, motor in electric automobile reliability is evaluated with fuzzy neural network model, can solve full-page proof
Reliability evaluation problem under this quantity term, the reliability level for further increasing motor for design department provide foundation, build
Vertical more perfect maintenance.
Detailed description of the invention
Fig. 1 is the flow diagram of method for evaluating reliability of the embodiment of the present invention;
Fig. 2 is the flow diagram that the embodiment of the present invention quantifies reliability standard by model.
Specific embodiment
With reference to embodiment and compares attached drawing the present invention is described in further details.It is emphasized that
Following the description is only exemplary, the range and its application being not intended to be limiting of the invention.
Fuzzy neural network combines the knowledge of neural network algorithm and fuzzy theory, summarizes neural network and fuzzy reason
By the advantages of, integrate study, association, identification, information processing, with powerful structured knowledge ability to express and powerful
Learning ability realizes fuzzy logic inference using neural network structure, is that artificial intelligence and intelligent automation are ground in recent years
Study carefully hot spot.
Fuzzy neural network has following three kinds of forms:
1. fuzzy logic neural network
2. arithmetic fuzzy neural network
3. hybrid fuzzy neural network
Neural network has a variety of feature and advantage:
1. parallel distributed information processing.Neural network has parallel organization, can carry out parallel data processing.It is this simultaneously
Row mechanism can solve extensive real-time computational problem in control system, and the redundancy in parallel computation can make control system
System has very strong fault-tolerance and robustness.
2. neural network is essentially nonlinear system.Theoretically, neural network can be realized any non-linear with arbitrary accuracy
Mapping, network can also be realized compared with the more superior system modelling of other methods.
3. study and adaptive ability.Neural network is trained based on the past data record of institute's research system
's.When the input for being supplied to network is not included in training set, a trained network has inducing ability.Nerve net
Network can also carry out automatic adjusument online.
4. multi-variable system.Neural network can handle many input signals, and have many output quantities, so being easy to
For multi-variable system.
Motor in electric automobile reliability is evaluated with fuzzy neural network model, can solve large sample quantity term
Under reliability evaluation problem, provide foundation for the reliability level that design department further increases motor, establish more perfect
Maintenance.
Specific embodiment one
The method for evaluating reliability of the new energy electric automobile motor based on fuzzy neural network of the present embodiment, including such as
Lower step:
A1, the evaluation index for establishing motor in electric automobile reliability;
A2, assessment indicator system is established, formulates evaluation criterion;
A3, according to the evaluation index and evaluation criterion, reliability standard is quantified based on fuzzy neural network model;
A4, motor in electric automobile motor reliability is judged according to quantized result.
In step A1, being located at the evaluation index includes h index, and each index corresponds to an evaluation points,
Middle h is greater than the positive integer equal to 1.
In step A2, if each index corresponds to m evaluation criterion in the L evaluation index, the evaluation index with
The evaluation criterion constitutes reliability evaluation standard value matrix Y=h*m, wherein the y in matrixijIndicate i-th evaluation because
Evaluation criterion value of the son in j-th of opinion rating, wherein m is greater than the positive integer equal to 1.
In step A3, reliability standard is quantified by model, is included the following steps:
A31, relative defects matrix is determined according to the assessment indicator system;
A32, the learning sample that fuzzy neural network is calculated according to relative defects matrix;
A33, it learning sample is brought into fuzzy neural network model is trained, initialize fuzzy neural network, and
Precision training is carried out to fuzzy neural network model;
A34, by reliability index data to be evaluated, substitute into the fuzzy neural network model for reaching predetermined accuracy into
Row calculates, and obtains evaluation result.
In step A31, the 1st level evaluation standard value is set for the relative defects of fuzzy set as pi1=1, m grades
Other evaluation criterion value is p for the relative defects of fuzzy setim=0, level evaluation standard value is for fuzzy set between
Relative defects pijIt is calculated with linear interpolation:
pij=(yij-yi1)/(yim-yi1) (1)
All elements constitute evaluation criterion index relative defects matrix P.
Equipped with w group data to be tested, each group of data include h evaluation index, then constitute the detection comprising h*w element
Sample value matrix X, wherein the xth in matrixijItem indicates i-th evaluation points value of jth group detection data, and wherein w is greater than
Positive integer equal to 1.
In detection sample value matrix X evaluation points value and reliability evaluation standard value matrix in evaluation criterion value into
Row compares, and obtains the relative defects matrix R of opinion rating set;
It is bigger for numerical value, the higher evaluation points of reliability, the relative defects of opinion rating set are as follows:
Work as xij≤yi1When, rij=0,
Work as xij≥yimWhen, rij=1;
It is bigger for numerical value, the lower evaluation points of reliability, then in contrast, the relative defects of opinion rating set
Are as follows:
Work as xij≥yi1When, rij=0,
Work as xij≤yimWhen, rij=1;
X in the case of remainingijThe relative defects of opinion rating set are solved with linear interpolation:
rij=(xij-yi1)/(yim-yi1) (2)
In step A32, in standard evaluation index relative defects matrix P, more samples are generated using interpolation method,
To improve training precision.
Interpolation sample is calculated according to the element of Y matrix, the y value with No. * is the new sample that interpolation calculation comes out
Data, defining relative defects of i-th evaluation points of interpolation sample k in j-th of opinion rating isDefine interpolation
The corresponding levels of the standard value of sample k is Tk, then:
Wherein 1≤i≤h, 1≤j≤m;
Tk=j+q/n, (4)
The value of wherein 1≤q≤n-1, n can be adjusted according to the needs of interpolated sample.It is obtained more after carrying out interpolation method
More sample datas, chooses a part as the learning sample of neural network in these data, remaining as test samples,
Wherein, learning sample is exactly that model learns how to determine by these samples each opinion rating, test samples be exactly for
Whether the result for examining the model after learning to operate out meets the requirements.
In step A33, learning sample obtained above is updated in computer program and carries out fuzzy neural network
Initialization, obtains the output matrix of learning sample.
After completing initialization procedure, need to carry out accuracy test, we can set the reliability step mark of each evaluation points
The quasi- table of comparisons, the test samples using the standard value of each grade as accuracy test, is normalized it, and normalization is public
Formula are as follows:
Wherein xijIndicate value of i-th evaluation points in j-th of opinion rating,Value after indicating normalized,
m1Indicate the number of evaluation points, m2Indicate the number of opinion rating.The target of known assay sample exports, and passes through nerve net
Reality output after network calculates compares, and acquires error, if error is met the requirements, this neural network model has reached expected instruction
Practice precision, that is, in the new data that interpolation method obtains, accordingly obtained the corresponding evaluation index of evaluation points of each sample
Value, then compares these data and by the data that neural network model runs, sees whether error meets the requirements, with this out
Determine the accuracy of neural network model.
In step A34, through the above steps, neural network model can carry out the certificate authenticity of motor,
Data to be evaluated (motor data evaluated by this model) is substituted into above-mentioned trained program code and is evaluated
Result can be obtained.
Specific embodiment two
The method for evaluating reliability of the new energy electric automobile motor based on fuzzy neural network of the present embodiment,
The evaluation index that motor in electric automobile reliability is described in step A1 includes following index:
1) the average first-time fault time: the first-time fault time of all samples is averaged;
2) it mean up time: is investigating in the time, the total duration of motor operation and the ratio of the number of stoppages;
The present embodiment, which is set, investigated the time as 5 years.
3) it average repair time: for each motor, is investigating in the time, when calculating all maintenances under total maintenance frequency
Between summation, and be averaged;
4) maintenance time rate: maintenance time and the ratio for investigating total run time in the time;
5) normal operation rate: for each motor, it is by the sum of mean up time and average repair time
The total operation time of this motor, normal operation rate are the ratio of mean up time and total operation time.
Above-mentioned evaluation index is set as the assessment indicator system of the present embodiment motor in electric automobile reliability, i.e., 5 are evaluated
The factor.
In step A31, determine that relative defects matrix P method is as follows:
3.1, setting reliability evaluation grade has 4 ranks, is outstanding, good, general, very poor, then 5 in step A1 respectively
Corresponding 4 opinion ratings of evaluation points constitute one 5 × 4 reliability evaluation standard value matrix Y, in matrix
YijItem indicates evaluation criterion value of i-th evaluation points in j-th of opinion rating.
3.2, assume there are n group data to be tested, each group of data include 5 evaluation points values, then this n group data constitutes
The detection sample value matrix X of one 5 × n, the xth in matrixijItem indicates i-th evaluation points value of jth group detection data.
3.3, relative defects of the 1st level evaluation standard value (outstanding) for fuzzy set of i-th evaluation points are set
For pi1The relative defects of=1, the 4th level evaluation standard value (very poor) are pi4=0, p betweenijUse linear interpolation
It is calculated:
pij=(yij-yi1)/(yi4-yi1)。
These relative defects elements constitute evaluation criterion index relative defects matrix P.
3.4, the evaluation points data in matrix X are converted into each factor for the relative defects of opinion rating set,
Form relative defects matrix R.Element in X is compared with evaluation criterion index value, bigger for numerical value, reliability is higher
Evaluation points (such as mean up time), work as xij≤yi1When, rij=0, work as xij≥yi4When, rij=1;Numerical value is got over
Greatly, the lower evaluation points of reliability (such as maintenance time rate), then in contrast, work as xij≥yi1When, rij=0, work as xij≤yi4
When, rij=1;X in the case of remainingijIt is solved with linear interpolation: rij=(xij-yi1)/(yi4-yi1)
In step A32, in standard evaluation index relative defects matrix P, more samples are generated using interpolation method, with
Improve training precision.
Defining relative defects of i-th evaluation points of interpolation sample k in j-th of opinion rating isIn definition
Inserting the corresponding levels of the standard value of sample k is Tk, then:
Wherein 1≤i≤5,1≤j≤4;Tk=j+q/n, wherein the value of 1≤j≤n-1, n can be according to the need of interpolated sample
It is adjusted.More sample datas are obtained after carrying out interpolation method, a part is chosen in these data as neural network
Learning sample, it is remaining as test samples.
In step A33, learning sample obtained above is updated to and carries out the first of fuzzy neural network in computer program
Beginningization obtains the output matrix of training sample.
After completing initialization procedure, need to carry out accuracy test, we can set the reliability step mark of each evaluation points
The quasi- table of comparisons (such as average repair time rate is as outstanding no more than 0.01), using the standard value of each grade as accuracy test
It is normalized in test samples, normalizes formula are as follows:
Wherein xijIndicate value of i-th evaluation points in j-th of opinion rating,Value after indicating normalized,
m1Indicate the number of evaluation points, m2Indicate the number of opinion rating.The target of known assay sample exports, and passes through nerve net
Reality output after network calculates compares, and acquires error, if error is met the requirements, this fuzzy neural network model has reached expection
Training precision.
In step A34, through the above steps, neural network model can carry out the certificate authenticity of motor, will
Data to be evaluated, which substitute into above-mentioned trained computer program evaluate, can be obtained result.
Specific embodiment three
The method for evaluating reliability of the new energy electric automobile motor based on fuzzy neural network of the present embodiment, process
Figure is as shown in Figure 1.Include:
S1, the evaluation index for establishing motor in electric automobile reliability;
S2, assessment indicator system is established, formulates evaluation criterion;
S3, according to the evaluation index and evaluation criterion, reliability standard is quantified based on fuzzy neural network model;
S4, motor in electric automobile motor reliability is judged according to quantized result.
Specific embodiment four
The method for evaluating reliability of the new energy electric automobile motor based on fuzzy neural network of the present embodiment, passes through mould
The flow chart that type quantifies reliability standard is as shown in Figure 2.Include:
S11, initialization program;
S12, subordinated-degree matrix, including reliability evaluation standard value matrix, evaluation criterion index relative defects square are calculated
Battle array, the relative defects matrix for detecting sample value matrix, opinion rating set;
It is trained after S13, fuzzy neural network model initialization;
S14, judge whether fuzzy neural network model reaches expected training precision, it is no, turn S13, be, continues next
Step;
S15, fuzzy neural network model is saved;
S16, data to be tested are substituted into fuzzy neural network model, calls fuzzy neural network model;
S17, evaluation result is obtained.
The above content is a further detailed description of the present invention in conjunction with specific preferred embodiments, and it cannot be said that
Specific implementation of the invention is only limited to these instructions.For those skilled in the art to which the present invention belongs, it is not taking off
Under the premise of from present inventive concept, several equivalent substitute or obvious modifications can also be made, and performance or use is identical, all answered
When being considered as belonging to protection scope of the present invention.
Claims (12)
1. a kind of method for evaluating reliability of the motor in electric automobile based on fuzzy neural network, which is characterized in that including as follows
Step:
A1, the evaluation index for establishing motor in electric automobile reliability;
A2, assessment indicator system is established, formulates evaluation criterion;
A3, according to the evaluation index and evaluation criterion, reliability standard is quantified based on fuzzy neural network model;
A4, motor in electric automobile reliability is judged according to quantized result.
2. method for evaluating reliability according to claim 1, which is characterized in that in step A1, the evaluation index includes h
A index, wherein h is greater than the positive integer equal to 1.
3. method for evaluating reliability according to claim 2, which is characterized in that every in the evaluation index in step A2
One index corresponds to m evaluation criterion, and the evaluation index and the opinion rating standard constitute reliability evaluation standard value matrix
(Y=h*m), wherein m is greater than the positive integer equal to 1, the y in matrixijItem indicates that i-th evaluation points is evaluated at j-th
Evaluation criterion value in standard.
4. method for evaluating reliability according to claim 1, which is characterized in that quantify reliability by model in step A3
Degree includes the following steps:
A31, relative defects matrix is determined according to the assessment indicator system;
A32, the learning sample that fuzzy neural network is calculated according to relative defects matrix;
A33, it learning sample is brought into fuzzy neural network model is trained, initialize fuzzy neural network, and to mould
It pastes neural network model and carries out precision training;
A34, by reliability index data to be evaluated, substitute into the fuzzy neural network model for reaching predetermined accuracy and counted
It calculates, obtains evaluation result.
5. method for evaluating reliability according to claim 4, which is characterized in that setting the 1st level evaluation standard value for
The relative defects of fuzzy set are pi1=1, m level evaluation standard value is p for the relative defects of fuzzy setim=0, it is situated between
In intermediate pijIt is calculated with linear interpolation, gained element is evaluation criterion index, and all elements constitute evaluation criterion and refer to
Mark relative defects matrix.
6. method for evaluating reliability according to claim 5, which is characterized in that in step A2, if in W group data to be tested
Each group of data include h evaluation index, then constitute the detection sample value matrix (X) of (W*h) a element, wherein in matrix
XthijItem indicates i-th evaluation points value of jth group detection data, and W is greater than the positive integer equal to 1.
7. method for evaluating reliability according to claim 6, which is characterized in that by the evaluation points value and the evaluation
Standard value is compared, and obtains the relative defects of opinion rating set, is constituted relative defects matrix (R).
8. method for evaluating reliability according to claim 7, which is characterized in that according to the difference of the evaluation points, if
It is different relative defects (r) that the different evaluation points, which are set, from the comparison result of the same evaluation criterion value, and uses line
Relative defects of the property interpolation calculation in the case of remaining.
9. method for evaluating reliability according to claim 5, which is characterized in that in the step A3, obtained using interpolation method
To learning sample and test samples,
Wherein:Indicate relative defects of i-th evaluation points of interpolation sample k in j-th of opinion rating, TkIt indicates
The corresponding levels of the standard value of interpolation sample k,
1≤i≤h, 1≤j≤m;Tk=j+q/n, wherein the value of 1≤j≤n-1, n are adjusted according to the needs of interpolated sample.
10. method for evaluating reliability according to claim 9, which is characterized in that substitute into the learning sample in program
The initialization for carrying out fuzzy neural network, obtains the output matrix of training sample.
11. method for evaluating reliability according to claim 10, which is characterized in that set the reliability etc. of each evaluation points
The grade standard table of comparisons, the test samples using each classification standard value as accuracy test, to the output matrix of training sample according to
Following formula is normalized,
Wherein, i=1,2 ... m1, j=1,2 ... m2, xijIndicate value of i-th evaluation points in j-th of opinion rating,Table
Value after showing normalized, m1Indicate the number of evaluation points, m2Indicate the number of opinion rating;
The comparison result with the reality output after neural computing is exported according to the target of test samples, judges this nerve
Whether network model reaches expected training precision.
12. method for evaluating reliability according to claim 11, which is characterized in that will reach described in data to be evaluated substitution
It is expected that carrying out evaluation in the computer program of training precision can be obtained result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811178736.2A CN109242093A (en) | 2018-10-10 | 2018-10-10 | A kind of method for evaluating reliability of the motor in electric automobile based on fuzzy neural network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811178736.2A CN109242093A (en) | 2018-10-10 | 2018-10-10 | A kind of method for evaluating reliability of the motor in electric automobile based on fuzzy neural network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109242093A true CN109242093A (en) | 2019-01-18 |
Family
ID=65054453
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811178736.2A Pending CN109242093A (en) | 2018-10-10 | 2018-10-10 | A kind of method for evaluating reliability of the motor in electric automobile based on fuzzy neural network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109242093A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110991849A (en) * | 2019-11-27 | 2020-04-10 | 北京理工大学 | New energy automobile comprehensive index determining method and system |
CN113421643A (en) * | 2021-07-09 | 2021-09-21 | 浙江大学 | AI model reliability judgment method, device, equipment and storage medium |
CN113532628A (en) * | 2020-04-16 | 2021-10-22 | 丰田自动车株式会社 | Abnormal sound evaluation system and abnormal sound evaluation method |
CN113783479A (en) * | 2021-11-11 | 2021-12-10 | 常州市佳博机械制造有限公司 | Brushless direct current motor fuzzy PID control method based on neural network matrix |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101334893A (en) * | 2008-08-01 | 2008-12-31 | 天津大学 | Fused image quality integrated evaluating method based on fuzzy neural network |
CN102400902A (en) * | 2010-09-15 | 2012-04-04 | 中国石油天然气股份有限公司 | Method for evaluating reliability of performance state of reciprocating compressor |
CN102999792A (en) * | 2012-12-20 | 2013-03-27 | 诸暨市供电局 | Method for comprehensive evaluation of power distribution network optimization allocation |
CN107085763A (en) * | 2017-03-31 | 2017-08-22 | 无锡开放大学 | A kind of driving motor for electric automobile system performance evaluation method |
-
2018
- 2018-10-10 CN CN201811178736.2A patent/CN109242093A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101334893A (en) * | 2008-08-01 | 2008-12-31 | 天津大学 | Fused image quality integrated evaluating method based on fuzzy neural network |
CN102400902A (en) * | 2010-09-15 | 2012-04-04 | 中国石油天然气股份有限公司 | Method for evaluating reliability of performance state of reciprocating compressor |
CN102999792A (en) * | 2012-12-20 | 2013-03-27 | 诸暨市供电局 | Method for comprehensive evaluation of power distribution network optimization allocation |
CN107085763A (en) * | 2017-03-31 | 2017-08-22 | 无锡开放大学 | A kind of driving motor for electric automobile system performance evaluation method |
Non-Patent Citations (2)
Title |
---|
刘坤 等: "模糊概率神经网络水质评价模型及其应用模糊概率神经网络水质评价模型及其应用", 《数学的实践与认识》 * |
李佳 等: "基于神经网络的电机可靠性控制技术研究", 《计算机光盘软件与应用》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110991849A (en) * | 2019-11-27 | 2020-04-10 | 北京理工大学 | New energy automobile comprehensive index determining method and system |
CN113532628A (en) * | 2020-04-16 | 2021-10-22 | 丰田自动车株式会社 | Abnormal sound evaluation system and abnormal sound evaluation method |
CN113421643A (en) * | 2021-07-09 | 2021-09-21 | 浙江大学 | AI model reliability judgment method, device, equipment and storage medium |
CN113783479A (en) * | 2021-11-11 | 2021-12-10 | 常州市佳博机械制造有限公司 | Brushless direct current motor fuzzy PID control method based on neural network matrix |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109242093A (en) | A kind of method for evaluating reliability of the motor in electric automobile based on fuzzy neural network | |
CN106355030B (en) | A kind of fault detection method based on analytic hierarchy process (AHP) and Nearest Neighbor with Weighted Voting Decision fusion | |
CN102289590B (en) | Method for estimating operating state of SF6 high-voltage circuit breaker and intelligent system | |
CN106371427B (en) | Industrial process Fault Classification based on analytic hierarchy process (AHP) and fuzzy Fusion | |
CN105117602B (en) | A kind of metering device running status method for early warning | |
CN106845796B (en) | One kind is hydrocracked flow product quality on-line prediction method | |
CN105469196A (en) | Comprehensive evaluation method and comprehensive evaluation system for evaluating mine construction project process | |
CN102496069A (en) | Cable multimode safe operation evaluation method based on fuzzy analytic hierarchy process (FAHP) | |
CN108053148A (en) | A kind of efficient diagnostic method of power information system failure | |
CN109409425B (en) | Fault type identification method based on neighbor component analysis | |
CN106549826A (en) | Intelligent substation switch applied in network performance test appraisal procedure | |
CN106127242A (en) | Year of based on integrated study Extreme Precipitation prognoses system and Forecasting Methodology thereof | |
CN111652271A (en) | Nonlinear feature selection method based on neural network | |
CN114372693B (en) | Transformer fault diagnosis method based on cloud model and improved DS evidence theory | |
CN112084709B (en) | Large-scale generator insulation state evaluation method based on genetic algorithm and radial basis function neural network | |
CN106610584A (en) | Remanufacturing process quality control method based on neural network and expert system | |
CN116109039A (en) | Data-driven anomaly detection and early warning system | |
CN109359742A (en) | A kind of generation method in subway subsystem preventive maintenance period | |
CN109784777B (en) | Power grid equipment state evaluation method based on time sequence information fragment cloud similarity measurement | |
CN116776209A (en) | Method, system, equipment and medium for identifying operation state of gateway metering device | |
CN110533341A (en) | A kind of Livable City evaluation method based on BP neural network | |
CN105931133A (en) | Distribution transformer replacement priority evaluation method and device | |
CN109214089A (en) | A kind of degree of membership appraisal procedure of bridge structural health monitoring index | |
CN117577225A (en) | Water quality evaluation method and device for settling tank | |
CN108053093A (en) | A kind of k- neighbour's method for diagnosing faults based on the conversion of average influence Value Data |
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 |
Application publication date: 20190118 |
|
RJ01 | Rejection of invention patent application after publication |