CN106650932B - A kind of the intelligent trouble classification method and device of data center's monitoring system - Google Patents
A kind of the intelligent trouble classification method and device of data center's monitoring system Download PDFInfo
- Publication number
- CN106650932B CN106650932B CN201611206596.6A CN201611206596A CN106650932B CN 106650932 B CN106650932 B CN 106650932B CN 201611206596 A CN201611206596 A CN 201611206596A CN 106650932 B CN106650932 B CN 106650932B
- Authority
- CN
- China
- Prior art keywords
- layers
- neuron
- energy
- development network
- input 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000012544 monitoring process Methods 0.000 title claims abstract description 37
- 238000011161 development Methods 0.000 claims abstract description 62
- 239000011159 matrix material Substances 0.000 claims abstract description 43
- 238000012549 training Methods 0.000 claims abstract description 40
- 210000002569 neuron Anatomy 0.000 claims description 129
- 238000010276 construction Methods 0.000 claims description 5
- 210000005036 nerve Anatomy 0.000 claims description 4
- 230000004044 response Effects 0.000 claims description 4
- 238000003745 diagnosis Methods 0.000 abstract description 4
- 239000013598 vector Substances 0.000 description 15
- 230000008569 process Effects 0.000 description 14
- 230000004913 activation Effects 0.000 description 3
- 210000004556 brain Anatomy 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000023886 lateral inhibition Effects 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 210000000225 synapse Anatomy 0.000 description 2
- 230000003213 activating effect Effects 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 210000004907 gland Anatomy 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 230000008555 neuronal activation Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 210000003625 skull Anatomy 0.000 description 1
- 230000000946 synaptic effect Effects 0.000 description 1
- 230000003245 working effect Effects 0.000 description 1
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
- G06N3/08—Learning methods
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
- G05B23/0205—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults
- G05B23/0259—Electric testing or monitoring by means of a monitoring system capable of detecting and responding to faults characterized by the response to fault detection
- G05B23/0275—Fault isolation and identification, e.g. classify fault; estimate cause or root of failure
- G05B23/0281—Quantitative, e.g. mathematical distance; Clustering; Neural networks; Statistical analysis
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Theoretical Computer Science (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Algebra (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Probability & Statistics with Applications (AREA)
- Pure & Applied Mathematics (AREA)
- Automation & Control Theory (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a kind of intelligent trouble classification method of data center's monitoring system and devices, this method comprises: according to the input matrix of the monitor state structure development network of the monitored item of monitoring resource；Learning training is carried out to development network using the input matrix；Classified using the development network after learning training to the failure of monitor state.This method carries out the input matrix of tectonic network according to the specific monitor state of the monitored item of monitoring resource, the intelligent trouble classification to monitoring resource is realized using a kind of intelligent network is development network, after carrying out learning training to development network, the classification to failure can be realized, realize and promote efficiency of fault diagnosis.
Description
Technical field
The present invention relates to data center's monitoring technology fields, more particularly to a kind of intelligence event of data center's monitoring system
Hinder classification method and device.
Background technique
Currently, with the rapid development of Internet, data center carries the function such as the acquisition, storage and analysis of various data
Can, once equipment breaks down, staff needs to go analysis failure according to the alarm details of the monitored item of monitoring, for experience
For old employee abundant, it may be possible to which the cracking failure cause that finds is solved, and general employee is likely to not look for
The basic reason that failure is found to failure cause or take a long time greatly increases data center apparatus operation in this way
Stability and safety, it is impossible to ensure that the normal operation of various businesses, the efficiency of fault diagnosis is lower.
Summary of the invention
The object of the present invention is to provide a kind of intelligent trouble classification method of data center's monitoring system and devices, to realize
Promote efficiency of fault diagnosis.
In order to solve the above technical problems, the present invention provides a kind of intelligent trouble classification method of data center's monitoring system,
This method comprises:
According to the input matrix of the monitor state structure development network of the monitored item of monitoring resource；
Learning training is carried out to development network using the input matrix；
Classified using the development network after learning training to the failure of monitor state.
Preferably, the development network includes X layers, Y layers and Z layers.
Preferably, learning training is carried out to development network using input matrix, comprising:
Using input matrix as X layers, the energy of Y layers of each neuron is calculated；
The maximum neuron j of energy is found out from Y layers of all neurons, and the weight of neuron j is updated；
By Y layers of output as Z layers of input, the neuron k responded is found out from the neuron in Z layers, it is right
The weight of neuron k is updated.
Preferably, the development network using after learning training classifies to the failure of monitor state, comprising:
The input matrix constructed according to X layers calculates the energy of neuron in Y layers, most by energy in Y layers of all neurons
The output of big neuron is set as 1, and the output of other neurons in Y layers other than the maximum neuron of energy is all provided with
It is set to 0；
The output of Y layers of neuron is found out into energy most in Z layers of all neuron as the input of Z layers of neuron
Big neuron i determines that the corresponding fault type of neuron i is the fault type of monitor state.
The present invention also provides a kind of intelligent trouble sorter of data center's monitoring system, for realizing the method,
Include:
Matrix construction module, for according to monitoring resource monitored item monitor state structure development network input square
Battle array；
Training module, for carrying out learning training to development network using the input matrix；
Failure modes module, for being classified using the development network after learning training to the failure of monitor state.
Preferably, the development network includes X layers, Y layers and Z layers.
Preferably, the training module includes:
Computing unit, for calculating the energy of Y layers of each neuron using input matrix as X layers；
First right value update unit will be refreshing for finding out the maximum neuron j of energy from Y layers of all neurons
Weight through first j is updated；
Second right value update unit, for as Z layers of input, finding out Y layers of output from the neuron in Z layers
The neuron k responded, is updated the weight of neuron k.
Preferably, the failure modes module includes:
Setting unit is exported, the input matrix for constructing according to X layers calculates the energy of neuron in Y layers, by Y layers
The output of the maximum neuron of energy is set as 1 in all neurons, by its in Y layers other than the maximum neuron of energy
The output of its neuron is disposed as 0；
Failure determination unit, for by the output of Y layers of neuron, as the input of Z layers of neuron, in Z layers of all mind
The maximum neuron i of energy is found out in member, determines that the corresponding fault type of neuron i is the fault type of monitor state.
The intelligent trouble classification method and device of a kind of data center's monitoring system provided by the present invention are provided according to monitoring
The input matrix of the monitor state structure development network of the monitored item in source；Development network is learnt using the input matrix
Training；Classified using the development network after learning training to the failure of monitor state.As it can be seen that according to the monitoring of monitoring resource
The specific monitor state of item carries out the input matrix of tectonic network, develops network using a kind of intelligent network to realize to monitoring
The intelligent trouble of resource is classified, and after carrying out learning training to development network, the classification to failure can be realized, and in failure
In use process of classifying, if the user thinks that the failure does not meet user's requirement, user can redefine failure, be learned
It practises, can be realized and classify to new failure, to reach the process of an on-line study, and develop network training and be not required to
It iterates, the learning time of network is greatly reduced, and realize a process of an online incremental learning, it can
Meet the customized failure of user, so carrying out failure modes using development network, can quickly analyze failure cause,
And the customized intelligent trouble classification of user may be implemented.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is a kind of flow chart of the intelligent trouble classification method of data center's monitoring system provided by the present invention；
Fig. 2 is development network diagram；
Fig. 3 is failure modes implementation process schematic diagram；
Fig. 4 is a kind of flow chart of the intelligent trouble sorter of data center's monitoring system provided by the present invention.
Specific embodiment
Core of the invention is to provide the intelligent trouble classification method and device of a kind of data center's monitoring system, to realize
Promote efficiency of fault diagnosis.
In order to enable those skilled in the art to better understand the solution of the present invention, below in conjunction in the embodiment of the present invention
Attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is only
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
Every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
Referring to FIG. 1, Fig. 1 is a kind of intelligent trouble classification method of data center's monitoring system provided by the present invention
Flow chart, this method comprises:
S11: according to the input matrix of the monitor state structure development network of the monitored item of monitoring resource；
S12: learning training is carried out to development network using input matrix；
S13: classified using the development network after learning training to the failure of monitor state.
As it can be seen that this method carries out the input matrix of tectonic network according to the specific monitor state of the monitored item of monitoring resource,
The intelligent trouble classification to monitoring resource is realized using a kind of intelligent network is development network, is learnt to development network
After training, the classification to failure can be realized, and in failure modes use process, if the user thinks that the failure is not met
User requires, and user can redefine failure, be learnt, can be realized and classify to new failure, to reach one
The process of a on-line study, and develop network training and do not need to iterate, the learning time of network is greatly reduced, and
A process for realizing an online incremental learning, can satisfy the customized failure of user, thus using development network come
Failure modes are carried out, failure cause can be quickly analyzed, and the customized intelligent trouble classification of user may be implemented.
Based on the above method, specifically, development network includes X layers, Y layers and Z layers.
Further, step S12 the following steps are included:
S21: using input matrix as X layers, the energy of Y layers of each neuron is calculated；
S22: finding out the maximum neuron j of energy from Y layers of all neurons, and the weight of neuron j is carried out more
Newly；
S23: by Y layers of output as Z layers of input, the neuron responded is found out from the neuron in Z layers
K is updated the weight of neuron k.
Further, step S13 the following steps are included:
S31: the input matrix constructed according to X layers calculates the energy of neuron in Y layers, by energy in Y layers of all neurons
The output for measuring maximum neuron is set as 1, by the output of other neurons in Y layers other than the maximum neuron of energy
It is disposed as 0；
S32: the output of Y layers of neuron is found out into energy in Z layers of all neuron as the input of Z layers of neuron
Maximum neuron i is measured, determines that the corresponding fault type of neuron i is the fault type of monitor state.
It is detailed, this method using state university professor Weng Juyang of Michigan, United States simulate human brain law of development and
A kind of intelligent network proposed --- development network, to realize the intelligent trouble classification to monitoring resource.Based on this method, failure
Assorting process specifically includes that (1) construction input matrix, (2) development network training and failure modes.
(1) it constructs input matrix: being that the monitored item of a certain resource of monitoring is classified according to certain mode classification,
"AND" processing, the i.e. monitored item under the same classification are done to the state for the monitored item for belonging to a classification, if there is one
It is abnormal, it is believed that the category is abnormal, if all normal, it is believed that the category is normally, later the monitored item of resource
The state of classification lines up column vector or row vector according to what a certain aligning method was fixed.
(2) develop network training and failure modes: it uses development network.
Wherein, development network is that state university professor Weng Juyang of Michigan, United States simulates the law of development of human brain and mentions
A kind of intelligent network out.
The development network has 3 regions, X, Y and Z, these three regions are similar to the input layer of general neuroid, hidden
Containing layer and output layer, but signal transmission direction and inner workings and general neuroid are completely different, network signal
Figure is as shown in Figure 2.X is contacted usually as sensor with external environment, can be modeled to any sensor type (as regarded
Feel, the sense of hearing or tactile), both can be used as input can also be used as output.The Y layers of brain as development network, it is usually implicit
(being closed in " skull "), cannot directly be contacted with external environment, can only pass through the connection with X, Z region obtain information.Z
Layer is generally as actuator layer, it can is that input is also possible to export, when Z is in extraneous monitor state, Z is network at this time
Input, otherwise, Z provides an output vector to drive actuator (muscle or body of gland) to act on real world.Three regions
Sequence from low to high is X, Y, Z, for example, X layers provide from low to high input to Y layer, Z layers of offer inputing to from high to low
Y layers.Z region is mankind's design or professor in Fig. 2, and the region Y is (naturally or the development) independently generated.The development network
Concrete operating principle be described as follows:
(1) at the t=0 moment, to any region in A={ X, Y, Z }, initialize its adaptive part N=(V, G) and
Vector r is reacted, wherein V is synaptic weight, and G is the age of neuron.
(2) in t=1,2 ... the moment constantly repeats two following steps to any region in A:
1. being calculated as follows using area function f:
(r ', N ')=f (b, t, N) (1)
Wherein b (bottom-up) and t (top-down) is corresponding region respectively from current network response vector r
From bottom to top with top-down input, r ' is its new response vector；
2. being substituted as follows: N ← N ', r ← r '
If X is sensor interface, x ∈ X is constantly in the state supervised by external environment, if Z is actuator interface, only
Have in the case where " teacher " selection, z ∈ Z is just in supervised state, and otherwise Z provides the output of actuator.Only work as X, Y and
After tri- regions Z all at least update once, entire network of developing just is completed once to update.When entirely development network updates two
It is secondary, for specific context (x, z), the prediction of one bout is completed, because development network needs to carry out 2 areas
The parallel update of area update, Y area update and X and Z region, the data of (x, z) carry out the data in the region Y after being transmitted to Y layers
Update, the data in the updated region Y pass to X, Z region completes X, in Z region corresponding data update.
There is weight vector v=(v for any neuron in the A of regionb,vt), corresponding to region input be p=(b,
t).For region Y, there are input b and top-down input t, region X from bottom to top there was only top-down input t, region
Z only has input b from bottom to top.Two normalization of energy definition as follows in region before each neuronal activation
The sum of inner product of vectors afterwards:
Wherein:It is the cynapse vector after normalizationUnit vector,It is the input vector after normalizationUnit vector.Inner product measures two unit vectorsWithMatched degree, because of r (vb,b,vt, t) and=cos
θ, θ are two unit vectorsWithBetween angle.Energy value before activation is between [- 1,1].
To simulate the lateral inhibition (lateral inhibitions) in any region A, only preceding top-k triumph
Neuron can be activated and be updated.Consider that k=1, (being activated) neuron of triumph can be carried out in the following way
Identification:
For k=1, only unique victor can just be activated, response yj=1, other neurons are not swashed
It is living.All connections in development network are all based on what Hebbian learning rules were learnt: activating the pre-synapse of neuron living
It is dynamicIt is activated simultaneously with post-synapse activity y.By taking the region Y as an example (learning method in other regions is similar with Y), if pre-synapse is last
End and post-synapse end are activated together, and the cynapse vector of neuron has a cynapse gainOther nerves not being activated
Member does not change its state.After a neuron j is activated, weight is as follows according to similar Hebbian Policy Updates:
Wherein, ω2(nj) it is learning rate relevant to activation age (activation number), ω1(nj) it is to maintain rate, also, ω1
(nj)+ω2(nj)≡1。ω2(nj) simplest form be ω2(nj)=1/n, this equation gives inputsSample the recurrence of mean value
Calculation method:
Wherein, tiIt is the activationary time of neuron, the age of triumph neuron adds 1, i.e. nj←nj+1。
Based on this method, detailed process is as follows for development network training and failure modes:
(1) develop network training: such as Fig. 2, X layers are the input matrix constructed, the Y layers and Z layers nerve for random initializtion
Member.
The training of the development network of the structure, is divided into two steps: the meter between calculating, Y between X layers and Y layers layer and Z layers
It calculates.
X layers and Y layers of calculating seeks the energy of Y layers of each neuron, then according to formula 3, finds out Y first, in accordance with formula 2
Maximum that neuron j of energy in layer neuron, it is believed that neuron j is activated, the neuron j for the Y layer being followed by activated
Weight, be updated weight according to formula 4, other neurons are done nothing, and the age of last neuron j adds 1.
At this point, it is 1 that the output of Y layers of neuron, which only has the output of neuron j, the output of other neurons is 0, as Z layers of input.
Y layers and Z layers of calculating: at this point, input of Y layers of the output as Z layers, it is assumed that training was specified is k-th of mind of Z layer
It is responded through member, then Z layers of neuron k, is updated weight according to formula 4, other neurons are without updating, last neuron k
Age add 1.
(2) development network failure classification: in the failure modes stage, the weight for developing network is not changing, according to X layers of structure
It is maximum to calculate energy then according to formula 3 according to the energy that formula 2 calculate Y layers of neuron for the input matrix made
The position of that neuron, secondly the output of the Y layers of maximum neuron of energy is set as 1, the output setting of other neurons
It is 0, the output of Y layers of neuron, as the input of Z layers of neuron, then, Z layers of neuron calculate Z layers most according to formula 3
The position i of that big neuron, then it is assumed that final failure modes are the corresponding fault type of Z layers of neuron i.
Based on this method, in specific implementation process, need to construct input matrix, such as Fig. 3, server has multiclass sensor packet
Include the sensors such as temperature, voltage, hard disk, fan, and every a kind of sensor has multiple, and the state of identical sensor is carried out
"AND" calculates, and finally every a kind of sensor is ranked up, is configured to input matrix.It develops in network training process, such as Fig. 3,
If the input matrix of construction is the neuron that X layers, Y layers and Z layers are random initializtion, Y layers are hidden layer, and Z layers are output layer.
Referring to FIG. 4, Fig. 4 is a kind of intelligent trouble sorter of data center's monitoring system provided by the present invention
Structural schematic diagram, for realizing the above method, which includes:
Matrix construction module 101, for according to monitoring resource monitored item monitor state structure development network input
Matrix；
Training module 102, for carrying out learning training to development network using input matrix；
Failure modes module 103, for being classified using the development network after learning training to the failure of monitor state.
As it can be seen that the device carries out the input matrix of tectonic network according to the specific monitor state of the monitored item of monitoring resource,
The intelligent trouble classification to monitoring resource is realized using a kind of intelligent network is development network, is learnt to development network
After training, the classification to failure can be realized, and in failure modes use process, if the user thinks that the failure is not met
User requires, and user can redefine failure, be learnt, can be realized and classify to new failure, to reach one
The process of a on-line study, and develop network training and do not need to iterate, the learning time of network is greatly reduced, and
A process for realizing an online incremental learning, can satisfy the customized failure of user, thus using development network come
Failure modes are carried out, failure cause can be quickly analyzed, and the customized intelligent trouble classification of user may be implemented.
Based on above-mentioned apparatus, specifically, development network includes X layers, Y layers and Z layers.
Further, training module includes:
Computing unit, for calculating the energy of Y layers of each neuron using input matrix as X layers；
First right value update unit will be refreshing for finding out the maximum neuron j of energy from Y layers of all neurons
Weight through first j is updated；
Second right value update unit, for as Z layers of input, finding out Y layers of output from the neuron in Z layers
The neuron k responded, is updated the weight of neuron k.
Further, failure modes module includes:
Setting unit is exported, the input matrix for constructing according to X layers calculates the energy of neuron in Y layers, by Y layers
The output of the maximum neuron of energy is set as 1 in all neurons, by its in Y layers other than the maximum neuron of energy
The output of its neuron is disposed as 0；
Failure determination unit, for by the output of Y layers of neuron, as the input of Z layers of neuron, in Z layers of all mind
The maximum neuron i of energy is found out in member, determines that the corresponding fault type of neuron i is the fault type of monitor state.
A kind of intelligent trouble sorter of data center's monitoring system provided by the present invention has been carried out in detail above
It introduces.Used herein a specific example illustrates the principle and implementation of the invention, the explanation of above embodiments
It is merely used to help understand method and its core concept of the invention.It should be pointed out that for the ordinary skill people of the art
Member for, without departing from the principle of the present invention, can with several improvements and modifications are made to the present invention, these improve and
Modification is also fallen within the protection scope of the claims of the present invention.
Claims (4)
1. a kind of intelligent trouble classification method of data center's monitoring system characterized by comprising
According to the input matrix of the monitor state structure development network of the monitored item of monitoring resource；
Learning training is carried out to development network using the input matrix；
Classified using the development network after learning training to the failure of monitor state；
Wherein, the development network includes X layers, Y layers and Z layers；
Wherein, learning training is carried out to development network using input matrix, comprising:
Using input matrix as X layers, the energy of Y layers of each neuron is calculated；
The maximum neuron j of energy is found out from Y layers of all neurons, and the weight of neuron j is updated；
By Y layers of output as Z layers of input, the neuron k responded is found out from the neuron in Z layers, to nerve
The weight of first k is updated.
2. the method as described in claim 1, which is characterized in that the development network using after learning training is to monitor state
Failure classify, comprising:
According to the input matrix, the energy of neuron in Y layers is calculated, by the maximum nerve of energy in Y layers of all neurons
The output of member is set as 1, and the output of other neurons in Y layers other than the maximum neuron of energy is disposed as 0；
By the output of Y layers of neuron, as the input of Z layers of neuron, it is maximum in Z layers of all neuron to find out energy
Neuron i determines that the corresponding fault type of neuron i is the fault type of monitor state.
3. a kind of intelligent trouble sorter of data center's monitoring system, which is characterized in that for realizing such as claim 1 or
Method described in 2 any one, comprising:
Matrix construction module, for according to monitoring resource monitored item monitor state structure development network input matrix；
Training module, for carrying out learning training to development network using the input matrix；
Failure modes module, for being classified using the development network after learning training to the failure of monitor state；
Wherein, the development network includes X layers, Y layers and Z layers；
Wherein, the training module includes:
Computing unit, for calculating the energy of Y layers of each neuron using input matrix as X layers；
First right value update unit, for finding out the maximum neuron j of energy from Y layers of all neurons, by neuron j
Weight be updated；
Second right value update unit, for Y layers of output as Z layers of input, to be found out progress from the neuron in Z layers
The neuron k of response, is updated the weight of neuron k.
4. device as claimed in claim 3, which is characterized in that the failure modes module includes:
Setting unit is exported, for the energy of neuron in Y layers being calculated, by Y layers of all neurons according to the input matrix
The output of the middle maximum neuron of energy is set as 1, by other neurons in Y layers other than the maximum neuron of energy
Output is disposed as 0；
Failure determination unit, for by the output of Y layers of neuron, as the input of Z layers of neuron, in Z layers of all neuron
In find out the maximum neuron i of energy, determine the corresponding fault type of neuron i be monitor state fault type.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611206596.6A CN106650932B (en) | 2016-12-23 | 2016-12-23 | A kind of the intelligent trouble classification method and device of data center's monitoring system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611206596.6A CN106650932B (en) | 2016-12-23 | 2016-12-23 | A kind of the intelligent trouble classification method and device of data center's monitoring system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106650932A CN106650932A (en) | 2017-05-10 |
| CN106650932B true CN106650932B (en) | 2019-05-28 |
Family
ID=58827281
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611206596.6A Active CN106650932B (en) | 2016-12-23 | 2016-12-23 | A kind of the intelligent trouble classification method and device of data center's monitoring system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106650932B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107247653A (en) * | 2017-06-02 | 2017-10-13 | 郑州云海信息技术有限公司 | A kind of Fault Classification and device of data center's monitoring system |
| CN107396094B (en) * | 2017-08-17 | 2019-02-22 | 上海大学 | Automatic testing method towards camera single in multi-cam monitoring system damage |
| CN109361539A (en) * | 2018-10-16 | 2019-02-19 | 郑州云海信息技术有限公司 | A kind of performance alarm method and apparatus based on development network |
| CN110000781B (en) * | 2019-03-29 | 2021-06-08 | 郑州大学 | Development network-based mobile robot motion direction pre-decision method |
| CN110399238B (en) * | 2019-06-27 | 2023-09-22 | 浪潮电子信息产业股份有限公司 | Disk fault early warning method, device, equipment and readable storage medium |
| CN112988437B (en) * | 2019-12-17 | 2023-12-29 | 深信服科技股份有限公司 | Fault prediction method and device, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103914735A (en) * | 2014-04-17 | 2014-07-09 | 北京泰乐德信息技术有限公司 | Failure recognition method and system based on neural network self-learning |
| CN105630648A (en) * | 2015-12-02 | 2016-06-01 | 武汉物联远科技有限公司 | Multidimensional data deep learning-based data center intelligent control method and system |
| CN105782071A (en) * | 2016-03-04 | 2016-07-20 | 长沙有色冶金设计研究院有限公司 | Water isolation pulp pump fault diagnosis method based on probabilistic neural network |
-
2016
- 2016-12-23 CN CN201611206596.6A patent/CN106650932B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103914735A (en) * | 2014-04-17 | 2014-07-09 | 北京泰乐德信息技术有限公司 | Failure recognition method and system based on neural network self-learning |
| CN105630648A (en) * | 2015-12-02 | 2016-06-01 | 武汉物联远科技有限公司 | Multidimensional data deep learning-based data center intelligent control method and system |
| CN105782071A (en) * | 2016-03-04 | 2016-07-20 | 长沙有色冶金设计研究院有限公司 | Water isolation pulp pump fault diagnosis method based on probabilistic neural network |
Non-Patent Citations (1)
| Title |
|---|
| 基于发育网络的语音图像双模态共发育系统研究;段谊海;《中国优秀硕士学位论文全文数据库信息科技辑》;20160215;I138-1777 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106650932A (en) | 2017-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106650932B (en) | A kind of the intelligent trouble classification method and device of data center's monitoring system | |
| Chui et al. | Predicting at-risk university students in a virtual learning environment via a machine learning algorithm | |
| Thanga Selvi et al. | RETRACTED ARTICLE: An optimal artificial neural network based big data application for heart disease diagnosis and classification model | |
| Khalil et al. | Energy efficiency prediction using artificial neural network | |
| Madhiarasan et al. | Analysis of artificial neural network: architecture, types, and forecasting applications | |
| Marwala | Causality, correlation and artificial intelligence for rational decision making | |
| Buniyamin et al. | Educational data mining for prediction and classification of engineering students achievement | |
| Ahmed et al. | Advancement of deep learning in big data and distributed systems | |
| CN110399238A (en) | A kind of disk failure method for early warning, device, equipment and readable storage medium storing program for executing | |
| Xu et al. | Bipolar fuzzy Petri nets for knowledge representation and acquisition considering non-cooperative behaviors | |
| Chaudhury et al. | An empirical study on attribute selection of student performance prediction model | |
| Sun et al. | GASF–MSNN: A new fault diagnosis model for spatiotemporal information extraction | |
| Chandiok et al. | Machine learning techniques for cognitive decision making | |
| Shahriar et al. | Predicting shellfish farm closures using time series classification for aquaculture decision support | |
| Nayak et al. | Elitist teaching–learning-based optimization (ETLBO) with higher-order Jordan Pi-sigma neural network: a comparative performance analysis | |
| D’Onofrio et al. | Using fuzzy cognitive maps to arouse learning processes in cities | |
| Aishath Murshida et al. | Survey on artificial intelligence | |
| Fen | Efficiency improvement of English online teaching system based on bagging learning flow feature selection | |
| Abd El-Haleem et al. | A generic ai-based technique for assessing student performance in conducting online virtual and remote controlled laboratories | |
| Valsamidis et al. | An approach for LMS assessment | |
| Dake et al. | University Students Behaviour Modelling Using the K‐Prototype Clustering Algorithm | |
| CN109361539A (en) | A kind of performance alarm method and apparatus based on development network | |
| Said et al. | Exploiting computational intelligence paradigms in e-technologies and activities | |
| Devika et al. | Deep Learning in IoT: Introduction, Applications, and Perspective in the Big Data Era | |
| Zhu | [Retracted] Research on Education Teaching Quality Analysis Based on the Neural Network Model |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20191128 Address after: 250001 North Zone, floor 3, building S05, Inspur Science Park, No. 1036, Inspur Road, high tech Zone, Jinan City, Shandong Province Patentee after: Shandong Yingxin Computer Technology Co., Ltd. Address before: 450018 Henan province Zheng Dong New District of Zhengzhou City Xinyi Road No. 278 16 floor room 1601 Patentee before: Zhengzhou Yunhai Information Technology Co. Ltd. |
|
| TR01 | Transfer of patent right |