CN107038143A - Belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model - Google Patents
Belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model Download PDFInfo
- Publication number
- CN107038143A CN107038143A CN201710299107.4A CN201710299107A CN107038143A CN 107038143 A CN107038143 A CN 107038143A CN 201710299107 A CN201710299107 A CN 201710299107A CN 107038143 A CN107038143 A CN 107038143A
- Authority
- CN
- China
- Prior art keywords
- antibody
- antigen
- belt conveyer
- affinity
- conveyer scale
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/15—Correlation function computation including computation of convolution operations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Pure & Applied Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Software Systems (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Operations Research (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Control Of Conveyors (AREA)
Abstract
The invention discloses a kind of belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model, monitored in real time by the behavioral parameters to belt conveyer scale, it was found that different from the behavior of normal condition, utilize existing information, analysis judgement is carried out to it, the relation set up between fault type and characteristic utilizes biology immunity principle in modeling process using new algorithm, made improvements on the basis of study aiNet network models, built multilayer immune network model.The inventive method can carry out self clonal vaviation, reduce sample requirement amount so that the model after improvement, which has, distinguishes oneself and oneself non-ability, has significantly heightened the adaptive ability of system.
Description
Technical field
The present invention relates to a kind of belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model, belong to
Manufacturing equipment detection, control, diagnosis and maintenance technical field.
Background technology
With the fast development of modern economy, the trade transportation amount of bulk material steeply rises, and promotes to dynamic weighing
It is required that more and more higher, and the ability of existing belted electronic balance quick diagnosis failure in use is low, and long-time stability are poor,
And there are data fluctuations, it is difficult to meet the requirement of Trade Measures.The development of technology of Internet of things has driven turning for traditional Weighing Apparatus Industry
Type, is supervised with computer and information technology to belt conveyer scale working condition, reduces error caused by human intervention, and then
Ensure the longtime running precision of equipment.
The fault signature that belt conveyer scale is showed in use can change with the change of environment, and some failures are in skin
Belt scale can be just found when running, once equipment is stopped, fault characteristic disappears.Therefore the fault diagnosis of belt conveyer scale is to ensure
The core of accuracy of belt scale, traditional method for diagnosing faults is haveed the shortcomings that to more than sample requirement and response speed is slow, such as BP
Neutral net etc..
The content of the invention
It is an object of the invention to provide a kind of belt conveyer scale fault diagnosis based on improved multilayer artificial immune network model
Method, can quickly, accurately realize the fault diagnosis of belt conveyer scale.
The technical solution for realizing the object of the invention is:Belt conveyer scale based on improved multilayer artificial immune network model
Method for diagnosing faults, comprises the following steps:
Step 1, status data when taking belt conveyer scale real work, determine fault type according to the state of belt conveyer scale, go forward side by side
Row normalized, regard the status data after processing as original antigen;
Step 2, using improved aiNet network models, learning training is carried out to original antigen, antibody is obtained and recognizes skin
The detector of belt scale fault sample, due to antibody and the incomplete matching of antigen so that a kind of detector can recognize difference
Training sample;
Step 3, by obtained antibody add knowledge base in;
Step 4, calculating treat the Euclidean distance between diagnostic sample and existing detector, find out minimum with testing data distance
K detector, fault diagnosis is carried out with k nearest neighbor theory, if the antigen levels that a certain class detector is matched in k detector
At most, then it is judged as such failure;If there is multiclass fault mode, that is, the detector number matched is identical, then takes and antigen
The maximum detector of affinity, and be diagnosed as the fault type;If not detecting fault type yet, jump procedure 5 enters
Adaptive diagnosis layer;
Step 5, using clonal vaviation algorithm unknown failure is learnt, and obtained ripe antibody is added to failure
In detector;
Whether step 6, affinity again between calculating belt conveyer scale fault data and detector, examine detector can be fast
Fast accurately identification antigen, if affine force value, which is more than, suppresses threshold xi, expression can be recognized, then jump to step 3, otherwise send
Warning.
Compared with existing invention, the present invention has following advantage:1) compared with traditional genetic algorithm, the inventive method
There is larger raising in terms of fault recognition rate and adaptivity.2) existing clonal vaviation algorithm is utilized, to artificial
Immune system is improved, and is introduced k nearest neighbor classification and is improved aiNet immune network models, and using the networking model after improving as
Own diagnostic horizon is built on basis, realizes the quick diagnosis to known fault.3) new model also compensate for traditional aiNet nets
Network model can not be recognized to unknown failure or mistaken diagnosis is this blind area of other failures.
Brief description of the drawings
Fig. 1 is flow chart of the method for the present invention.
Fig. 2 is the clonal selection algorithm flow chart for introducing Local Gauss mutation operator.
Embodiment
Below in conjunction with the accompanying drawings and specific experiment the invention will be further described.
Reference picture 1, the belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model includes following step
Suddenly:
Step 1, fault model determined according to state during belt conveyer scale real work, take the status data of belt conveyer scale, go forward side by side
Row normalized, regard the status data after processing as original antigen.
The formula of initialization process is carried out in the step 1 to initial data:
In formula, agijThe element arranged for the i-th row j of state matrix, agij' for normalization after element, N for antigen
Number, in belt conveyer scale fault diagnosis, because fault type is different, therefore should add classification logotype, to reach area into antigen matrix
Divide the purpose of different conditions, be expressed as Ag=[Ag ', T] ∈ RN×(L+1), Ag ' is the antigen matrix after processing, and wherein L is antigen
The element number of vector, corresponding to collection point number, T represents antigen classification identity column, i.e., corresponding to known to belt weigher system
Fault type.N is chosen at random in AntigensmIt is individual to constitute initial memory antibody collection Ab, it is designated as
Step 2, using improved aiNet network models, learning training is carried out to known sample, antibody is obtained and recognizes
The detector of belt conveyer scale fault sample, due to antibody and the incomplete matching of antigen so that a kind of detector can be recognized not
Same training sample.
The step 2 use improvement aiNet network models the step of be trained for:
Step 2.1, the parameter of the learning algorithm of aiNet immune network models, including pruning threshold ξ, suppression threshold value are set
σsDeng taking σ in experiments=0.2 ξ, wherein ξ calculation formula is:
It will remove as the antigen after antibody collection as new antigen collection, and take one of antigen to enter with all antibody
Row affinity computing.Using Euclidean distance calculation formula calculating antibody AbiWith antigen A gjApart from Dij, as D (Abi,Agj):
Defining affinity is:
Step 2.2, the antibody composition antibody subset that affinity exceedes pruning threshold is chosen, the antibody of selection is passed through into clone
Choosing principles are cloned, generation clone collection C, and affinity is bigger, and clone's collection is bigger, each the colony counts of energized antibody
NCCalculated by following formula:
In formula, NAFor the number of antibody in network, round () is the function that rounds up;Each antibody after clone can
Undergone mutation with certain speed, the antibody collection C after variation*Each antibody calculation formula it is as follows:
In formula, AbiFor the antibody of parent, Abi *For the antibody of filial generation,For speed of mutation, affinity size is inversely proportional to;
Step 2.3, antigen and clone's collection C after variation are calculated*Between affinity fi,j', and screened according to affine force value
Antibody, affinity, which is less than, suppresses then being deleted for threshold xi, that is, deletes fi,j' < ξ;The affinity between clone cell is calculated simultaneously
Affinity S i.e. between antibodyi,k, more than suppression threshold value σsExpression antibody between it is much like, delete this antibody-like, that is, delete
Si,k> σs, circulation step 2.2-2.3, until reaching iterations G.
Step 3, the memory antibody after above-mentioned processing is added in knowledge base, the failure that table 1 show experiment is known
Know storehouse description:
The fault knowledge storehouse that table 1 is tested
Step 4, calculating treat the Euclidean distance between diagnostic sample and existing detector, find out minimum with testing data distance
K detector, fault diagnosis is carried out with k nearest neighbor theory, if the antigen levels that a certain class detector is matched in k detector
At most, then it is judged as such failure;If there is multiclass fault mode, that is, the detector number matched is identical, then takes and antigen
The maximum detector of affinity, and be diagnosed as the fault type;If not detecting fault type yet, jump procedure 5 enters
Adaptive diagnosis layer.
Step 5, using clonal vaviation algorithm unknown failure is learnt, and obtained ripe antibody is added to failure
In detector;
In the step 5, learning training, including following step are carried out to the failure for failing to detect using clonal vaviation algorithm
Suddenly:
Step 5.1, take fail in step 4 identification antigen learnt again, i.e., by unknown belt conveyer scale fault data with
A new B cell is generated centered on the average value of its region, the radius for defining B cell is:
In formula,For constant, value is 0.02 in belt conveyer scale fault diagnosis;agimaxI-th for antigen matrix arranges most
Big value;For the average value of the i-th row of antigen matrix;
Step 5.2, antibody is seen as a cell, cell radius is expressed as r=1/ ξ, calculates B cell and failure-free data
The distance between antibody of mark, detects whether to exist overlapping, if existing overlapping, the radius for reducing B cell is detected again,
Until not overlapping;
Step 5.3, antibody collection, antibody number N are generated random in B cell defined in step 5.1A′For:
In formula, K ' is antibody tormation constant, is set as the radius that 50, R is B cell in experiment, Γ is the straight of state space
Footpath;
Step 5.4, the distance between calculation procedure 5.3 is obtained antibody and B cell, according to each antibody and B cell it
Between antibody is arranged apart from size, and clone, colony counts calculation formula:
In formula, β is that value is that 1, i is arrangement sequence number where antibody in multiplication factor, experiment;
Mutation operation is performed to the antibody after clone, introduces a random variation coefficient to realize the variation of cell, is evolved
Equation simplification is:
Abn'=Abn+γ·N(0,1)
Ab in formulan' for variation after antibody, AbnFor the antibody before variation, η is the parameter of control characteristic function decay, and f is
The affinity of antibody and antigen;
The distance between the distance between antibody and B cell after variation, and relatively more female antibody and B cell are recalculated,
Retain the high-quality antibody that affinity is more than threshold value;
Step 5.5, the distance between high-quality antibody is calculated, checks whether there is overlapping, is more than for affinity and suppresses threshold value
σsProgress delete, remaining antibody is updated in tracer.
Step 6:The affinity between belt conveyer scale fault data and detector is calculated again, examines detector whether can be fast
Fast accurately identification antigen, if affine force value, which is more than, suppresses threshold xi, expression can be recognized, then jump to step 3, otherwise send
Warning, the selected genetic algorithm of experiment and original aiNet network models are compared, and comparative result is as shown in table 2:
Table 2
Wherein A1Represent normal condition, A2Represent sensor surface wear-out failure, A3Represent that weighing support gets stuck failure, A4Table
Show weighing support looseness fault, A5Represent sensor wire bad error, A6Represent that idler machining carrying roller (is surveyed by card as unknown failure
Examination).Three kinds of algorithm iteration number of times are selected as 150 times, and the crossover probability of genetic algorithm is 0.6, and mutation probability is 0.001, is fitted
Response function setup is:
Analysis understands that compared with genetic algorithm, multilayer aiNet network models are in fault recognition rate and adaptivity side
There is larger raising in face;New model also compensate for traditional aiNet network models unknown failure can not be recognized or
Mistaken diagnosis is this blind area of other failures;Multilayer aiNet network models are used in the fault detect of belt conveyer scale, by final failure
Diagnostic result arranged, be shown in Table 3:
Table 3
As shown in Table 3, discrimination reaches 100% under normal mode, and reaches 96% for known fault diagnosis rate
More than, for new fault type, mistaken diagnosis is only 1 of normal data, has 92 samples to be correctly diagnosed as new event
Hinder type, only 1 for generally speaking, being misdiagnosed as normal data in 600 test samples, illustrate that the multilayer detects mould
Type is very high for the discrimination of failure, can meet the fault detect requirement of belt weigher system.
In summary, the present invention proposes a kind of belt conveyer scale failure based on improved multilayer artificial immune network model and examined
Disconnected method, mainly for the modified for solving the deficiency of current traditional belt conveyer scale fault diagnosis algorithm and proposing.Belt conveyer scale system
The fault diagnosis of system is mainly monitored in real time by the behavioral parameters to belt conveyer scale, finds the row different from normal condition
Using existing information, analysis judgement to be carried out to it, the relation set up between fault type and characteristic.New algorithm
Using biology immunity principle, made improvements on the basis of study aiNet network models, built multilayer immunological network mould
Type, can carry out self clonal vaviation, reduce sample requirement amount so that model after improvement have distinguish oneself with it is non-oneself
Ability, has significantly heightened the adaptive ability of system.
By presented above, it can be seen that the invention has the advantages that:
(1) compared with traditional genetic algorithm, the method for diagnosing faults is in terms of fault recognition rate and adaptivity
There is larger raising.
(2) existing clonal vaviation algorithm is utilized, artificial immune system is improved, k nearest neighbor classification is introduced perfect
AiNet immune network models, and by improve after networking model based on build own diagnostic horizon, realize to known fault
Quick diagnosis.
(3) new model also compensate for traditional aiNet network models unknown failure can not be recognized or mistaken diagnosis be its
His this blind area of failure.
It should be understood that, although the present specification is described in terms of embodiments, but this narrating mode is only clearly visible,
For the those skilled in the art, under the premise without departing from the principles of the invention, the improvement made also should be regarded as this
The protection domain of invention.
Claims (4)
1. the belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model, it is characterised in that including as follows
Step:
Step 1, status data when taking belt conveyer scale real work, determine fault type, and returned according to the state of belt conveyer scale
One change is handled, and regard the status data after processing as original antigen;
Step 2, using improved aiNet network models, learning training is carried out to original antigen, antibody is obtained and recognizes belt conveyer scale
The detector of fault sample, due to antibody and the incomplete matching of antigen so that a kind of detector can recognize different instructions
Practice sample;
Step 3, by obtained antibody add knowledge base in;
Step 4, calculating treat the Euclidean distance between diagnostic sample and existing detector, find out the k minimum with testing data distance
Individual detector, fault diagnosis is carried out with k nearest neighbor theory, if the antigen levels of a certain class detector matching are most in k detector
It is many, then it is judged as such failure;If there is multiclass fault mode, that is, the detector number matched is identical, then takes and antigen
The maximum detector of affinity, and it is diagnosed as the fault type;If not detecting fault type yet, jump procedure 5 enters suitable
Answering property diagnostic horizon;
Step 5, using clonal vaviation algorithm unknown failure is learnt, and obtained ripe antibody is added to fault detect
In device;
Whether step 6, affinity again between calculating belt conveyer scale fault data and detector, examine detector can be quickly accurate
True identification antigen, if affine force value, which is more than, suppresses threshold xi, expression can be recognized, then jump to step 3, otherwise give a warning.
2. the belt conveyer scale method for diagnosing faults according to claim 1 based on improved multilayer artificial immune network model,
Characterized in that, the formula that the initial data of belt conveyer scale fault signature is normalized the step 1:
In formula, agijThe element arranged for the i-th row j of state matrix, agij' for normalization after element, N be antigen number,
In belt conveyer scale fault diagnosis, because fault type is different, therefore classification logotype should be added into antigen matrix, distinguish different to reach
The purpose of state, is expressed as Ag=[Ag ', T] ∈ RN×(L+1), Ag ' is the antigen matrix after processing, and wherein L is antigen vector
Element number, corresponding to collection point number, T represents antigen classification identity column, i.e., corresponding to failure classes known to belt weigher system
Type.
3. the belt conveyer scale method for diagnosing faults according to claim 1 based on improved multilayer artificial immune network model,
Characterized in that, the step 2 use improvement aiNet network models the step of be trained for:
Step 2.1, N is chosen at random in AntigensmIt is individual to constitute initial antibody collection Ab, it is designated as
New antigen collection is used as using removing as the antigen of antibody collection;
The parameter of step 2.2, the learning algorithm of setting aiNet immune network models, including pruning threshold ξ, suppression threshold value σs, take
The antigen that neoantigen is concentrated carries out affinity computing with all antibody, first using Euclidean distance calculation formula calculating antibody Abi
With antigen A gjApart from Dij, i.e. D (Abi,Agj):
In formula, N is the number of antigen, and L is the element number of antigen vector;
Then affinity is:
Step 2.3, selection affinity exceed the antibody composition antibody subset of pruning threshold, and the antibody of selection is passed through into Immune Clone Selection
Principle is cloned, generation clone collection C, and affinity is bigger, and clone's collection is bigger, each the colony counts N of energized antibodyCBy
Following formula is calculated:
In formula, NAFor the number of antibody in network, round () is the function that rounds up;
Each antibody after clone can be undergone mutation with certain speed, the antibody collection C after variation*Each antibody calculating
Formula is as follows:
In formula, AbiFor the antibody of parent, Abi *For the antibody of filial generation,For speed of mutation, affinity size is inversely proportional to;
Step 2.4, the clone calculated after antigen and variation collect C*Between affinity fi,j', and according to affine force value screening antibodies,
Affinity, which is less than, suppresses then being deleted for threshold xi, that is, deletes fi,j' < ξ;The affinity calculated simultaneously between clone cell is to resist
Affinity S between bodyi,k, more than suppression threshold value σsExpression antibody between it is much like, delete this antibody-like, that is, delete Si,k>
σs, circulation step 2.2-2.3, until reaching iterations G.
4. the belt conveyer scale method for diagnosing faults according to claim 1 based on improved multilayer artificial immune network model,
Characterized in that, the step 5 is the step of being trained using clonal vaviation algorithm to the failure for failing to detect:
Step 5.1, that unknown belt conveyer scale fault data is generated into a new B centered on the average value of its region is thin
Born of the same parents, define B cell radius be:
In formula,For constant, agimaxFor antigen matrix i-th row maximum,For the average value of the i-th row of antigen matrix;
Step 5.2, antibody is seen as to a cell, cell radius is expressed as r=1/ ξ, calculates B cell and is identified with failure-free data
The distance between antibody, detect whether to exist overlapping, if existing overlapping, the radius for reducing B cell is detected again, until
It is not overlapping;
Step 5.3, random generation antibody collection, antibody number N in the B cell that step 5.1 is definedA′For:
In formula, K ' is antibody tormation constant, and R is the radius of B cell, and Γ is the diameter of state space;
The distance between antibody and B cell that step 5.4, calculation procedure 5.3 are obtained, according between each antibody and B cell
Antibody is arranged apart from size, and cloned, colony counts calculation formula:
In formula, β is multiplication factor, and i is the arrangement sequence number where antibody;
Mutation operation is performed to the antibody after clone, introduces a random variation coefficient to realize the variation of cell, evolution equation
It is reduced to:
Abn'=Abn+γ·N(0,1)
In formula, Abn' for variation after antibody, AbnFor the antibody before variation, η is the parameter of control characteristic function decay, and f is anti-
The affinity of body and antigen;
The distance between the distance between antibody and B cell after variation, and relatively more female antibody and B cell are recalculated, is retained
Affinity is more than the high-quality antibody of threshold value;
Step 5.5, the distance between high-quality antibody is calculated, check whether there is overlapping, be more than for affinity and suppress threshold value σsEnter
Row is deleted, and remaining antibody is updated in tracer.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710299107.4A CN107038143A (en) | 2017-04-30 | 2017-04-30 | Belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710299107.4A CN107038143A (en) | 2017-04-30 | 2017-04-30 | Belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107038143A true CN107038143A (en) | 2017-08-11 |
Family
ID=59538343
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710299107.4A Pending CN107038143A (en) | 2017-04-30 | 2017-04-30 | Belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107038143A (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108829755A (en) * | 2018-05-25 | 2018-11-16 | 国网山东省电力公司青岛供电公司 | Transformer fault case based on iteration closest approach matches searching method |
CN112532463A (en) * | 2020-12-17 | 2021-03-19 | 四川长虹电器股份有限公司 | Computer network fault diagnosis method |
CN112699911A (en) * | 2020-06-03 | 2021-04-23 | 武汉市教云慧智信息技术有限公司 | Intelligent marketing model library method based on clonal selection algorithm |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101415256A (en) * | 2008-11-28 | 2009-04-22 | 南京航空航天大学 | Method of diagnosing wireless sensor network fault based on artificial immunity system |
CN102374936A (en) * | 2010-08-23 | 2012-03-14 | 太原理工大学 | Mechanical failure diagnostic method based on complex immune network algorithm |
CN102760208A (en) * | 2012-07-03 | 2012-10-31 | 清华大学 | Dynamic artificial immune fault diagnosis method based on simulation vaccine |
-
2017
- 2017-04-30 CN CN201710299107.4A patent/CN107038143A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101415256A (en) * | 2008-11-28 | 2009-04-22 | 南京航空航天大学 | Method of diagnosing wireless sensor network fault based on artificial immunity system |
CN102374936A (en) * | 2010-08-23 | 2012-03-14 | 太原理工大学 | Mechanical failure diagnostic method based on complex immune network algorithm |
CN102760208A (en) * | 2012-07-03 | 2012-10-31 | 清华大学 | Dynamic artificial immune fault diagnosis method based on simulation vaccine |
Non-Patent Citations (2)
Title |
---|
韩旭明: "若干改进的人工免疫算法及其在大气环境领域中的应用", 《中国博士学位论文全文数据库工程科技Ⅰ辑》 * |
高森祺 等: "基于双层人工免疫网络的皮带秤故障诊断", 《机械设计与制造工程》 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108829755A (en) * | 2018-05-25 | 2018-11-16 | 国网山东省电力公司青岛供电公司 | Transformer fault case based on iteration closest approach matches searching method |
CN108829755B (en) * | 2018-05-25 | 2021-04-13 | 国网山东省电力公司青岛供电公司 | Transformer fault case matching search method based on iteration closest point |
CN112699911A (en) * | 2020-06-03 | 2021-04-23 | 武汉市教云慧智信息技术有限公司 | Intelligent marketing model library method based on clonal selection algorithm |
CN112699911B (en) * | 2020-06-03 | 2022-04-29 | 武汉市教云慧智信息技术有限公司 | Intelligent marketing model library method based on clonal selection algorithm |
CN112532463A (en) * | 2020-12-17 | 2021-03-19 | 四川长虹电器股份有限公司 | Computer network fault diagnosis method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112508105B (en) | Fault detection and retrieval method for oil extraction machine | |
US20190086912A1 (en) | Method and system for generating two dimensional barcode including hidden data | |
CN108173708A (en) | Anomalous traffic detection method, device and storage medium based on incremental learning | |
CN110213244A (en) | A kind of network inbreak detection method based on space-time characteristic fusion | |
JP2021140739A (en) | Program, method for creating learned model, information processing method and information processing device | |
CN108229580A (en) | Sugared net ranking of features device in a kind of eyeground figure based on attention mechanism and Fusion Features | |
CN110348437B (en) | Target detection method based on weak supervised learning and occlusion perception | |
CN106815639A (en) | The abnormal point detecting method and device of flow data | |
US11783471B2 (en) | Method and device for determining whether object includes defect | |
CN107038143A (en) | Belt conveyer scale method for diagnosing faults based on improved multilayer artificial immune network model | |
CN109388944A (en) | A kind of intrusion detection method based on KPCA and ELM | |
CN109146873A (en) | A kind of display screen defect intelligent detecting method and device based on study | |
CN113139500A (en) | Smoke detection method, system, medium and device | |
CN108304567A (en) | High-tension transformer regime mode identifies and data classification method and system | |
US20190034786A1 (en) | System and Method for Making Nuclear Radiation Detection Decisions and/or Radionuclide Identification Classifications | |
CN106250913B (en) | A kind of combining classifiers licence plate recognition method based on local canonical correlation analysis | |
CN114048468A (en) | Intrusion detection method, intrusion detection model training method, device and medium | |
CN112419268A (en) | Method, device, equipment and medium for detecting image defects of power transmission line | |
CN113541985B (en) | Internet of things fault diagnosis method, model training method and related devices | |
Al‐Tahhan et al. | Accurate automatic detection of acute lymphatic leukemia using a refined simple classification | |
CN111786999B (en) | Intrusion behavior detection method, device, equipment and storage medium | |
JP2019158684A (en) | Inspection system, identification system, and discriminator evaluation device | |
CN112836719A (en) | Indicator diagram similarity detection method fusing two classifications and three groups | |
CN115795353B (en) | Underground metal target classification method and system based on unbalanced data set | |
CN106778252A (en) | Intrusion detection method based on rough set theory Yu WAODE algorithms |
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: 20170811 |