CN103927582A - Mechanical fault diagnosis method based on collaborative mechanism immune particle swarm network - Google Patents
Mechanical fault diagnosis method based on collaborative mechanism immune particle swarm network Download PDFInfo
- Publication number
- CN103927582A CN103927582A CN201410188049.4A CN201410188049A CN103927582A CN 103927582 A CN103927582 A CN 103927582A CN 201410188049 A CN201410188049 A CN 201410188049A CN 103927582 A CN103927582 A CN 103927582A
- Authority
- CN
- China
- Prior art keywords
- antibody
- group
- sub
- affinity
- network
- 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
Landscapes
- Peptides Or Proteins (AREA)
Abstract
The invention relates to a mechanical fault diagnosis technology, in particular to a mechanical fault diagnosis method based on a collaborative mechanism immune particle swarm network. The problems that according to an existing mechanical fault diagnosis technology, requirements for the number of fault samples are high, and the diagnosis accurate rate is low are solved. The mechanical fault diagnosis method based on the collaborative mechanism immune particle swarm network includes the following steps that (1) mechanical fault samples collected by a sensor are taken as antigens; (2) appetencies between the antigens and antibodies of the immune network are calculated; (3) low-frequency variation and antibody recombination are performed on subgroup bodies A with large appetencies, particle swarm optimization and antibody recombination are performed on subgroup bodies B with small appetencies, and therefore a new antibody group is obtained; (4) the new antibody group is adjusted through the immune network; (5) the step (2), the step (3) and the step (4) are executed circularly; (6) the mechanical fault samples collected by the sensor are input into the new immune network. The mechanical fault diagnosis method is suitable for mechanical fault diagnosis.
Description
Technical field
The present invention relates to technology for mechanical fault diagnosis, specifically a kind of mechanical failure diagnostic method of the immunity particle group network based on synergistic mechanism.
Background technology
Along with commercial production and scientific and technical development, electromechanical equipment is more and more accurate complicated, and automatization level is corresponding improve also.Once electromechanical equipment breaks down, not only can cause equipment itself to damage, and can cause serious work safety accident and personal safety accident.Therefore, mechanical fault diagnosis has very important significance for protection electromechanical equipment, minimizing security incident.Under prior art condition, mechanical fault diagnosis adopts the mode of expert's prophylactic repair mostly.Practice shows, this kind of mode be because self principle is limit, the problem that ubiquity is high to fault sample quantitative requirement, accuracy rate of diagnosis is low.Based on this, be necessary to invent a kind of brand-new mechanical failure diagnostic method, the problems referred to above that exist to solve existing machinery fault diagnosis technology.
Summary of the invention
The present invention, in order to solve the problem that existing machinery fault diagnosis technology is high to fault sample quantitative requirement, accuracy rate of diagnosis is low, provides a kind of mechanical failure diagnostic method of the immunity particle group network based on synergistic mechanism.
The present invention adopts following technical scheme to realize: a kind of mechanical failure diagnostic method of the immunity particle group network based on synergistic mechanism, and the method is to adopt following steps to realize:
1) the mechanical fault sample of sensor being collected is as antigen;
2) calculate the affinity between antigen and the antibody of immunological network;
3), according to the size of affinity, the antibody population of immunological network is equally divided into A sub-group and the little B sub-group of affinity that affinity is large; The large A sub-group of affinity is carried out to low frequency variation and antibody restructuring, and the little B sub-group of affinity is carried out to particle group optimizing and antibody restructuring, obtain thus new antibody population;
4) new antibody population is carried out to immunological network adjusting, obtain thus new immunological network;
5) circulation execution step 2)-step 4); When cycle index reaches 200 times, stop carrying out;
6) the mechanical fault sample of sensor being collected is inputted new immunological network, obtains thus mechanical fault diagnosis result.
Described step 2) in, when calculate between antigen and the antibody of immunological network affinity time, specific formula for calculation is as follows:
D
ij=||A
bi-A
gj|| (2);
In formula (1)-(2): f
ijbe the affinity between j antigen and i antibody of immunological network; A
gjbe j antigen; A
bii the antibody for immunological network; θ is constant, and θ ∈ (0,1); D
ijit is the distance between j antigen and i antibody of immunological network.
Described step 3) in,
When the large A sub-group of affinity being carried out to low frequency variation, the formula that specifically makes a variation is as follows:
In formula (3):
m the variation antibody member for A sub-group; a
mm the original antibody member for A sub-group; T is aberration rate, and T={T ∈ (0,1) };
The large A sub-group of affinity is carried out to antibody restructuring specifically to be referred to: the whole original antibody members that replace A sub-group by whole variation antibody member of A sub-group;
When the little B sub-group of affinity is carried out to particle group optimizing, specifically optimize formula as follows:
x
id(t+1)=v
id(t+1)+x
id(t) (5);
In formula (4)-(5): v
id(t+1) be the t+1 speed of i antibody member of B sub-group constantly; v
id(t) be the t speed of i antibody member of B sub-group constantly; x
id(t+1) be the t+1 position of i antibody member of B sub-group constantly; x
id(t) be the t position of i antibody member of B sub-group constantly; p
id(t) be self optimum position of i antibody member of B sub-group; p
gd(t) be i antibody member of the B sub-group optimum position in field, place; ω is constant and 0 < ω < 1; c
1, c
2, r
1, r
2be constant.
The little B sub-group of affinity is carried out to antibody restructuring specifically to be referred to: the whole original antibody members that replace B sub-group by whole optimization antibody member of B sub-group.
Described step 4), in, the concrete steps of new antibody population being carried out to immunological network adjusting are as follows:
4.1) calculate the degree of being excited of new antibody population, specific formula for calculation is as follows:
In formula (6): A
iaffinity for new antibody population; D
i,jit is the Eucliden distance between j antigen and i B cell; D
i,kit is the Eucliden distance between i B cell and k B cell; α
1, α
2, α
3be constant;
4.2) the low antibody population of the degree of being excited is made a variation, the formula that specifically makes a variation is as follows:
In formula (7):
affinity for the antibody population after variation; b
maffinity for the antibody population before variation; N (0,1) is the normal distribution random number that expectation is 0, variance is 1; σ is aberration rate, and σ=α e
-1/ β;
4.3) antibody population after the high antibody population of the degree of being excited and variation is combined, obtain new immunological network;
4.4) circulation execution step 4.1)-step 4.3); When cycle index reaches 200 times, stop carrying out.
Compare with existing machinery fault diagnosis technology; the mechanical failure diagnostic method of a kind of immunity particle group network based on synergistic mechanism of the present invention has been realized mechanical fault diagnosis based on immunity particle cluster network algorithm; the a small amount of fault sample of its need; just can reach very high accuracy rate of diagnosis; effectively protect thus electromechanical equipment, effectively reduced work safety accident and personal safety accident.
The present invention efficiently solves the problem that existing machinery fault diagnosis technology is high to fault sample quantitative requirement, accuracy rate of diagnosis is low, is applicable to mechanical fault diagnosis.
Embodiment
A mechanical failure diagnostic method for immunity particle group network based on synergistic mechanism, the method is to adopt following steps to realize:
1) the mechanical fault sample of sensor being collected is as antigen;
2) calculate the affinity between antigen and the antibody of immunological network;
3), according to the size of affinity, the antibody population of immunological network is equally divided into A sub-group and the little B sub-group of affinity that affinity is large; The large A sub-group of affinity is carried out to low frequency variation and antibody restructuring, and the little B sub-group of affinity is carried out to particle group optimizing and antibody restructuring, obtain thus new antibody population;
4) new antibody population is carried out to immunological network adjusting, obtain thus new immunological network;
5) circulation execution step 2)-step 4); When cycle index reaches 200 times, stop carrying out;
6) the mechanical fault sample of sensor being collected is inputted new immunological network, obtains thus mechanical fault diagnosis result.
Described step 2) in, when calculate between antigen and the antibody of immunological network affinity time, specific formula for calculation is as follows:
D
ij=||A
bi-A
gj|| (2);
In formula (1)-(2): f
ijbe the affinity between j antigen and i antibody of immunological network; A
gjbe j antigen; A
bii the antibody for immunological network; θ is constant, and θ ∈ (0,1); D
ijit is the distance between j antigen and i antibody of immunological network.
Described step 3) in,
When the large A sub-group of affinity being carried out to low frequency variation, the formula that specifically makes a variation is as follows:
In formula (3):
m the variation antibody member for A sub-group; a
mm the original antibody member for A sub-group; T is aberration rate, and T={T ∈ (0,1) };
The large A sub-group of affinity is carried out to antibody restructuring specifically to be referred to: the whole original antibody members that replace A sub-group by whole variation antibody member of A sub-group;
When the little B sub-group of affinity is carried out to particle group optimizing, specifically optimize formula as follows:
x
id(t+1)=v
id(t+1)+x
id(t) (5);
In formula (4)-(5): v
id(t+1) be the t+1 speed of i antibody member of B sub-group constantly; v
id(t) be the t speed of i antibody member of B sub-group constantly; x
id(t+1) be the t+1 position of i antibody member of B sub-group constantly; x
id(t) be the t position of i antibody member of B sub-group constantly; p
id(t) be self optimum position of i antibody member of B sub-group; p
gd(t) be i antibody member of the B sub-group optimum position in field, place; ω is constant and 0 < ω < 1; c
1, c
2, r
1, r
2be constant.
The little B sub-group of affinity is carried out to antibody restructuring specifically to be referred to: the whole original antibody members that replace B sub-group by whole optimization antibody member of B sub-group.
Described step 4), in, the concrete steps of new antibody population being carried out to immunological network adjusting are as follows:
4.1) calculate the degree of being excited of new antibody population, specific formula for calculation is as follows:
In formula (6): A
iaffinity for new antibody population; D
i,jit is the Eucliden distance between j antigen and i B cell; D
i,kit is the Eucliden distance between i B cell and k B cell; α
1, α
2, α
3be constant;
4.2) the low antibody population of the degree of being excited is made a variation, the formula that specifically makes a variation is as follows:
In formula (7):
affinity for the antibody population after variation; b
maffinity for the antibody population before variation; N (0,1) is the normal distribution random number that expectation is 0, variance is 1; σ is aberration rate, and σ=α e
-1/ β;
4.3) antibody population after the high antibody population of the degree of being excited and variation is combined, obtain new immunological network;
4.4) circulation execution step 4.1)-step 4.3); When cycle index reaches 200 times, stop carrying out.
Claims (4)
1. a mechanical failure diagnostic method for the immunity particle group network based on synergistic mechanism, is characterized in that: the method is to adopt following steps to realize:
1) the mechanical fault sample of sensor being collected is as antigen;
2) calculate the affinity between antigen and the antibody of immunological network;
3), according to the size of affinity, the antibody population of immunological network is equally divided into A sub-group and the little B sub-group of affinity that affinity is large; The large A sub-group of affinity is carried out to low frequency variation and antibody restructuring, and the little B sub-group of affinity is carried out to particle group optimizing and antibody restructuring, obtain thus new antibody population;
4) new antibody population is carried out to immunological network adjusting, obtain thus new immunological network;
5) circulation execution step 2)-step 4); When cycle index reaches 200 times, stop carrying out;
6) the mechanical fault sample of sensor being collected is inputted new immunological network, obtains thus mechanical fault diagnosis result.
2. the mechanical failure diagnostic method of a kind of immunity particle group network based on synergistic mechanism according to claim 1, is characterized in that: described step 2), when calculate between antigen and the antibody of immunological network affinity time, specific formula for calculation is as follows:
D
ij=||A
bi-A
gj|| (2);
In formula (1)-(2): f
ijbe the affinity between j antigen and i antibody of immunological network; A
gjbe j antigen; A
bii the antibody for immunological network; θ is constant, and θ ∈ (0,1); D
ijit is the distance between j antigen and i antibody of immunological network.
3. the mechanical failure diagnostic method of a kind of immunity particle group network based on synergistic mechanism according to claim 1, is characterized in that: described step 3),
When the large A sub-group of affinity being carried out to low frequency variation, the formula that specifically makes a variation is as follows:
In formula (3):
m the variation antibody member for A sub-group; a
mm the original antibody member for A sub-group; T is aberration rate, and T={T ∈ (0,1) };
The large A sub-group of affinity is carried out to antibody restructuring specifically to be referred to: the whole original antibody members that replace A sub-group by whole variation antibody member of A sub-group;
When the little B sub-group of affinity is carried out to particle group optimizing, specifically optimize formula as follows:
x
id(t+1)=v
id(t+1)+x
id(t) (5);
In formula (4)-(5): v
id(t+1) be the t+1 speed of i antibody member of B sub-group constantly; v
id(t) be the t speed of i antibody member of B sub-group constantly; x
id(t+1) be the t+1 position of i antibody member of B sub-group constantly; x
id(t) be the t position of i antibody member of B sub-group constantly; p
id(t) be self optimum position of i antibody member of B sub-group; p
gd(t) be i antibody member of the B sub-group optimum position in field, place; ω is constant and 0 < ω < 1; c
1, c
2, r
1, r
2be constant.
The little B sub-group of affinity is carried out to antibody restructuring specifically to be referred to: the whole original antibody members that replace B sub-group by whole optimization antibody member of B sub-group.
4. the mechanical failure diagnostic method of a kind of immunity particle group network based on synergistic mechanism according to claim 1, is characterized in that: described step 4), the concrete steps of new antibody population being carried out to immunological network adjusting are as follows:
4.1) calculate the degree of being excited of new antibody population, specific formula for calculation is as follows:
In formula (6): A
iaffinity for new antibody population; D
i,jit is the Eucliden distance between j antigen and i B cell; D
i,kit is the Eucliden distance between i B cell and k B cell; α
1, α
2, α
3be constant;
4.2) the low antibody population of the degree of being excited is made a variation, the formula that specifically makes a variation is as follows:
In formula (7):
affinity for the antibody population after variation; b
maffinity for the antibody population before variation; N (0,1) is the normal distribution random number that expectation is 0, variance is 1; σ is aberration rate, and σ=α e
-1/ β;
4.3) antibody population after the high antibody population of the degree of being excited and variation is combined, obtain new immunological network;
4.4) circulation execution step 4.1)-step 4.3); When cycle index reaches 200 times, stop carrying out.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410188049.4A CN103927582A (en) | 2014-05-05 | 2014-05-05 | Mechanical fault diagnosis method based on collaborative mechanism immune particle swarm network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410188049.4A CN103927582A (en) | 2014-05-05 | 2014-05-05 | Mechanical fault diagnosis method based on collaborative mechanism immune particle swarm network |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103927582A true CN103927582A (en) | 2014-07-16 |
Family
ID=51145798
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410188049.4A Pending CN103927582A (en) | 2014-05-05 | 2014-05-05 | Mechanical fault diagnosis method based on collaborative mechanism immune particle swarm network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103927582A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106842920A (en) * | 2017-01-04 | 2017-06-13 | 南京航空航天大学 | For the robust Fault-Tolerant Control method of multiple time delay four-rotor helicopter flight control system |
CN109376652A (en) * | 2018-10-24 | 2019-02-22 | 国网江苏省电力有限公司检修分公司 | Paralleling reactor of extra-high voltage method for diagnosing faults, apparatus and system based on artificial immune particle swarm clustering algorithm |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101056074A (en) * | 2007-05-18 | 2007-10-17 | 吉林大学 | An ultrasonic motor control method based on the immunity particle cluster algorithm |
CN101840635A (en) * | 2010-05-06 | 2010-09-22 | 招商局重庆交通科研设计院有限公司 | Variable speed-limiting control method based on artificial immune particle swarm algorithm |
CN102374936A (en) * | 2010-08-23 | 2012-03-14 | 太原理工大学 | Mechanical failure diagnostic method based on complex immune network algorithm |
-
2014
- 2014-05-05 CN CN201410188049.4A patent/CN103927582A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101056074A (en) * | 2007-05-18 | 2007-10-17 | 吉林大学 | An ultrasonic motor control method based on the immunity particle cluster algorithm |
CN101840635A (en) * | 2010-05-06 | 2010-09-22 | 招商局重庆交通科研设计院有限公司 | Variable speed-limiting control method based on artificial immune particle swarm algorithm |
CN102374936A (en) * | 2010-08-23 | 2012-03-14 | 太原理工大学 | Mechanical failure diagnostic method based on complex immune network algorithm |
Non-Patent Citations (1)
Title |
---|
陈如清等: "《改进的协同粒子群优化算法的研究与应用》", 《振动、测试与诊断》 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106842920A (en) * | 2017-01-04 | 2017-06-13 | 南京航空航天大学 | For the robust Fault-Tolerant Control method of multiple time delay four-rotor helicopter flight control system |
CN106842920B (en) * | 2017-01-04 | 2019-04-30 | 南京航空航天大学 | For the robust Fault-Tolerant Control method of multiple time delay four-rotor helicopter flight control system |
CN109376652A (en) * | 2018-10-24 | 2019-02-22 | 国网江苏省电力有限公司检修分公司 | Paralleling reactor of extra-high voltage method for diagnosing faults, apparatus and system based on artificial immune particle swarm clustering algorithm |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN104268658B (en) | A kind of Forecasting Methodology of bridge structure safe Monitoring Data | |
CN104063586B (en) | Bayesian network failure prediction method based on polymorphic fault tree | |
CN103424654A (en) | Method for assessing voltage sag sensitivity of sensitive equipment | |
CN103064019A (en) | High-voltage circuit interrupter fault diagnosis method based on grey correlation fuzzy clustering | |
CN108121990A (en) | A kind of solar irradiance Forecasting Methodology and device based on all-sky imaging device | |
CN104133994A (en) | Reliability evaluation method fusing multi-source success or failure data | |
CN103927582A (en) | Mechanical fault diagnosis method based on collaborative mechanism immune particle swarm network | |
CN102374936B (en) | Mechanical failure diagnostic method based on complex immune network algorithm | |
CN106056305A (en) | Power generation system reliability rapid assessment method based on state clustering | |
CN104867063A (en) | Power grid multiple major failure recognition method based on a linear weighting method | |
CN104766162A (en) | Opportunistic maintenance method considering different fixed maintenance costs | |
CN105205589A (en) | Evaluation method of boiler efficiency of thermal generator set under different loads | |
CN103473461A (en) | Method for estimating wind power prediction error based on data feature extraction | |
CN103760889B (en) | Fault based on Bayesian network separates fast method | |
BR102021003359A2 (en) | Grain size measurement system, grain collection, movement and measurement system and grain size measurement method | |
CN106021190A (en) | Segment blind signal separating method based on improved particle swarm optimization | |
Min et al. | Highway traffic abnormal state detection based on PCA-GA-SVM algorithm | |
Ruicheng | ODISSEE Test of ABS Anti-lock Braking System | |
Li | Reliability prediction for factory casualty using grey system theory | |
Sun et al. | Quantitative Studies on Traffic Efficiency and Safety Variation Trend of Mixed Traffic Flow with Different Penetration Rates of Automated Vehicles | |
Rezac et al. | Improvement of Diagnostics for Measurement of Multi-MeV Ions Produced in Deuterium Gas-Puff Z-Pinch | |
Bent et al. | Toward Rapid Automated Moment Tensor Solutions for Canada | |
Skidmore | Lepton-number and lepton-flavour violation in B decays | |
Kourkoutis et al. | Measuring Cosmic Rays and their Effect on the Atmosphere | |
Porter et al. | Failure Point Analysis of Membrane Electrode Assemblies with Coating Irregularities |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20140716 |
|
RJ01 | Rejection of invention patent application after publication |