CN102169555B - Relevance sensing signal multiple sensing element fault location and sensing signal self-recovery method - Google Patents
Relevance sensing signal multiple sensing element fault location and sensing signal self-recovery method Download PDFInfo
- Publication number
- CN102169555B CN102169555B CN 201110077679 CN201110077679A CN102169555B CN 102169555 B CN102169555 B CN 102169555B CN 201110077679 CN201110077679 CN 201110077679 CN 201110077679 A CN201110077679 A CN 201110077679A CN 102169555 B CN102169555 B CN 102169555B
- Authority
- CN
- China
- Prior art keywords
- signal
- relevance
- fault
- sensing element
- transducing signal
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000011084 recovery Methods 0.000 title claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 16
- 230000002463 transducing effect Effects 0.000 claims description 72
- 238000013528 artificial neural network Methods 0.000 claims description 34
- 230000008859 change Effects 0.000 claims description 19
- 230000004807 localization Effects 0.000 claims description 14
- 238000013507 mapping Methods 0.000 claims description 6
- 238000011156 evaluation Methods 0.000 claims description 2
- 230000009467 reduction Effects 0.000 claims description 2
- 238000012546 transfer Methods 0.000 claims description 2
- 238000013178 mathematical model Methods 0.000 claims 1
- 230000007935 neutral effect Effects 0.000 abstract 3
- 238000005314 correlation function Methods 0.000 abstract 1
- LGRFSURHDFAFJT-UHFFFAOYSA-N Phthalic anhydride Natural products C1=CC=C2C(=O)OC(=O)C2=C1 LGRFSURHDFAFJT-UHFFFAOYSA-N 0.000 description 6
- JHIWVOJDXOSYLW-UHFFFAOYSA-N butyl 2,2-difluorocyclopropane-1-carboxylate Chemical compound CCCCOC(=O)C1CC1(F)F JHIWVOJDXOSYLW-UHFFFAOYSA-N 0.000 description 6
- 230000000875 corresponding effect Effects 0.000 description 5
- 238000012360 testing method Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000001537 neural effect Effects 0.000 description 2
- 230000003716 rejuvenation Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 150000003839 salts Chemical class 0.000 description 2
- WYLIRYQDDKDHLT-UHFFFAOYSA-N CC1=CC=CC=C1C.CC1=CC=CC=C1C Chemical group CC1=CC=CC=C1C.CC1=CC=CC=C1C WYLIRYQDDKDHLT-UHFFFAOYSA-N 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000003197 catalytic effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000001771 impaired effect Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003647 oxidation Effects 0.000 description 1
- 238000007254 oxidation reaction Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Abstract
The invention discloses a relevance sensing signal multiple sensing element fault location and sensing signal self-recovery method. The method comprises the steps of: obtaining a training sample; establishing a neutral net corresponding to the signal relevance of a multipath sensing element, and learning the relevance among the sensing signals; by using the learned relevance relation of various signals, inputting a sensing signal containing a fault sensing element into the neutral net, and conducting fault determination of the inputted single according to a difference value between a net output signal and an input signal; positioning the multipath fault sensing element; and recovering the multipath fault sensing signal. On account of the relevance among the sensing signals, the method automatically adjusts node function weight of a hidden layer of the neutral net to learn and utilize the relevance, without solving correlation function. The method is simple, convenient and fast.
Description
Technical field
The present invention relates to many sensing elements localization of fault and transducing signal self-recovery method, relate in particular to a kind of many sensing elements localization of fault and transducing signal self-recovery method for thering is relevance between transducing signal.
Background technology
Many complex processes in objective world need to process the sensing system from the multichannel transducing signal simultaneously, and this promotion can connect the appearance of the Integrated Intelligent Sensors pattern of many sensing elements.These novel integrated sensors connection Mei road sensing elements all can walk abreast and extract the correlated characteristic information of detected object separately, and communicate information in sensor.Typical many sensing elements sensor has the networked smart sensor based on the IEEE1451 standard, and it can connect at most 255 road sensing elements is simultaneously measured, and has standardization, characteristics that integrated level is high.
Connect the multichannel sensing element has brought facility to the detection of many sensor source signals simultaneously, but some problems have also been brought: as, the sensing element (one or more) that connects breaks down simultaneously, detection signal is not followed when measured, is difficult to locate fault element; Under some occasion, each transducing signal of sensor-based system has relevance, and each road transducing signal all can impact other road transducing signal, not yet can utilize the relevance had between each road transducing signal, reconstruct fault sensing element signal value, make transducing signal from recovering.
Summary of the invention
For solving the problem and blemish of above-mentioned middle existence, the invention provides a kind of relevance transducing signal many sensing elements localization of fault and transducing signal self-recovery method.Described technical scheme is as follows:
Relevance transducing signal many sensing elements localization of fault and transducing signal self-recovery method comprise:
A, obtain training sample;
B, foundation meet the neural network of multichannel sensing element signal relevance, and the relevance between the study transducing signal;
Each signal relevance relation of C, utilization study, will, containing the transducing signal input neural network of fault sensing element, according to the difference between network output signal and input signal, carry out the fault judgement to input signal;
D, orient multichannel fault sensing element;
E, recovery multichannel fault transducing signal.
The beneficial effect of technical scheme provided by the invention is:
Broken through and must solve relevance relation between transducing signal and just can carry out the thinking set of localization of fault, utilize to innovation its relevance, by the value of exploratory change multichannel transducing signal within the specific limits, thereby realize that the location of multichannel fault sensing element and signal are from recovering.
The accompanying drawing explanation
Fig. 1 is the process flow diagram of multichannel sensing element localization of fault and transducing signal self-recovery method;
Fig. 2 is the structural representation of the neural network of structure;
Fig. 3 is parallel 9 fork tree fault location algorithm process flow diagrams;
Fig. 4 is expansion 9 fork tree signal reconstruction algorithm flow charts;
Fig. 5 is a kind of phthalic anhydride still testing and control project figure;
Fig. 6 is the neural metwork training procedure chart.
Embodiment
For making the purpose, technical solutions and advantages of the present invention clearer, below in conjunction with accompanying drawing, embodiment of the present invention is described further in detail:
The present embodiment provides a kind of relevance transducing signal many sensing elements localization of fault and transducing signal self-recovery method.
Referring to Fig. 1, the method comprises the following steps:
The above-mentioned training sample obtained comprises abundant training sample.
The neural network of above-mentioned foundation is the neural network that meets multichannel sensing element signal relevance.
Above-mentionedly will, containing the transducing signal input neural network of fault sensing element, can, according to the difference between network output signal and input signal, to input signal, carry out the fault judgement.
Above-mentioned is to change the value of multichannel sensing element signal in fixed range, makes it meet all each road transducing signal x
ibetween the relevance relation, i=1 wherein, 2 ..., n, and orient multichannel fault sensing element according to the difference change conditions of network output signal and input signal.
According to transducing signal relevance relation, continue to change the signal of having located the fault sensing element, make the output signal of neural network and input signal difference be less than setting threshold, thereby automatically recover the fault transducing signal.
Above-mentioned steps 10 specifically comprises with step 20: as shown in Figure 2, for phthalic anhydride Fu 10 tunnel relevance transducing signals, build the neural network of a 10-13-5-13-10 structure.10 road transducing signals under abundant normal operating conditions are formed to the input layer of neural network, relevance relation between 10 road transducing signals under training place normal production conditions.Due to the existence of noise, the condition that network training finishes is not that SSE minimizes, but converges to the threshold value Δ of a setting when it
sSEwith regard to deconditioning, otherwise network will attempt noise is learnt, thereby reduce Generalization Ability of Neural Network.In this network, Δ
sSEbe set as 0.0001, through 1.793 seconds, the network training success, its Fig. 6 is the neural metwork training procedure chart.
Above-mentioned steps 30 specifically comprises: the measuring-signal { x that obtains the multichannel sensing element
i| i=1,2 ..., 10}, by transducing signal X=[x
1, x
2..., x
10]
tin input neural network, observe network output Y=[y
1, y
2..., y
10]
tand the SSE between input signal, judge whether sensing element exists fault, if SSE≤Δ
sSE, the sensing element non-fault; If SSE>Δ
sSE, one or more sensing element non-fault, need to enter step 40 and 50 and carry out localization of fault and signal recovery certainly.
Above-mentioned steps 40 and step 50 specifically comprise: to judging the signal that has the fault sensing element, the Dui Ge road signal tree-shaped fault search that walks abreast within the specific limits, show that SSE reduces the most obvious one group, thereby orient the fault sensing element; After orienting the fault sensing element, proceed the tree-shaped signal reconstruction algorithm of expansion, make SSE≤Δ
sSEthereby, realize that transducing signal is from recovering.
The technological core of the present embodiment is: build a feed-forward type transmission network with symmetrical topological structure, this neural network unique distinction is to have unity gain, that is, and and its input vector x under normal circumstances
i(i=1,2 ..., n) equal output vector y
i(i=1,2 ..., n).Figure 2 shows that the structural drawing of the neural network of structure, comprise an input layer, three hidden layers and an output layer.The hidden layer ground floor is mapping layer, and node dimension p is maximum in whole network, and for extracting input message relevance part, its node transport function can be Sigmoid function (f (x)=1/ (1+e
-x)) or other similar nonlinear functions, the weight of this node layer function is w
c-j(j=1,2 ..., p, p>n); The last one deck of hidden layer is the demapping layer, and the node dimension equates with mapping layer, and for information is carried out to the relevance reduction, its node is nonlinear transfer function, and the weight of this node layer function is w
d-j(j=1,2 ..., p, p>n).
Above-mentioned neural network can be trained with error back propagation (BP) algorithm.Training sample is the measurement data of respectively organizing sensing element under unfaulty conditions, and by enough training samples, neural network can, by automatically adjusting the weight w of each node transport function of hidden layer (comprising mapping layer, bottleneck layer reconciliation mapping layer), make x
1=f
in-1(x
2, x
3..., x
n)=f
out-1(y
2, y
3... y
n)=y
1, x
2=f
in-2(x
1, x
3..., x
n)=f
out-2(y
2, y
3..., y
n)=y
2..., x
n=f
in-n(x
1, x
2..., x
n-1)=f
out-n(y
1, y
2..., y
n-1)=y
2thereby, learn each input signal relevance relation each other.Now, can utilize each signal relevance relation of study, input signal is carried out to the fault judgement.If trouble-free sensing data is input in neural network, network output vector y
i(i=1,2 ..., n) with input vector x
i(i=1,2 ..., n) certainly will equate; Because the weight w of each node transport function is constant, if wherein a road or multichannel sensing element break down, output quantity y so
ito all change, corresponding input quantity x
iwith output vector y
ito there are differences comprehensively.Therefore, definition neural network output vector y
iwith input vector x
ibetween error sum of squares SSE, its formula is as follows:
Index is as sensing element fault evaluation factors.
If connect Ge road sensing element all in the non-fault perfect condition, obviously SSE=0; Because each node transport function weight of neural network w determines in the associate feature of network training stage ,You Ge road transducing signal, if input vector X
iin one or more sensing element break down, output vector Y
ito all change.The thinking of multichannel fault sensing element method for searching is based on the associate feature of transducing signal so, by each road fault element signal of exploratory change within the specific limits (increase, reduce or constant), makes output vector Y at this moment
iwith input vector X
ireach unanimity.Is the difficult point that the method exists: 1. have how many roads sensing element to break down actually? which road specifically again? is the reconstruction value that the sensing element fault-signal occurs on the ②Ge road how many? guarantee output vector Y
iwith input vector X
ireach unanimity.Solve this two problems, the location that will realize the fault sensing element; Another one is to realize the recovery certainly of transducing signal.
IEEE 1451 intelligent sensors that connected the relevant sensing element of 10 road signal of below take are example, and the embodiment how the present embodiment is overcome the above problems is described in further detail: the 10 road transducing signals of establishing intelligent sensor are { x
i| i=1,2 ..., 10}, can utilize each transducing signal X=[x
1, x
2..., x
10]
tfor the input node of neural network, when having one or more sensing element fault, obvious SSE ≠ 0.
While supposing Jin You mono-road sensing element fault, exploratory Gei Ge successively road sensing element signal changes a smaller value δ within the specific limits
s, then the numerical value X ' after changing
1=[x
1+ δ
s, x
2..., x
10]
t, X '
2=[x
1, x
2+ δ
s..., x
10]
t... X '
n=[x
1, x
2..., x
10+ δ
s]
tinput neural network, observe the now variation of SSE.
From formula 4-formula 8, if the input signal after certain group changes can make SSE significantly diminish, illustrate that the transducing signal X ' after changing meets the incidence relation of training signal, thereby judge that this road sensing element breaks down; After the fault sensing element of location, continue to change this road signal value, make SSE be less than the setting threshold Δ
sSE, can recover this road transducing signal.
If exploratory change one road sensing element signal within the specific limits, SSE can't obtain significantly and reduce, a more than road sensing element fault is described, need considers to change the multichannel transducing signal simultaneously, below emphasis the situation that two-way fault sensing element location is arranged in lower 10 road transducing signals simultaneously is described in detail in detail.Now, may exist the sensing element combination of fault to have
kind, table 1 has been listed the combination of all two-way sensing element simultaneous faultss, these be combined into line ordering be respectively 1,2}, 1,3} ..., { 9,10}.
Exploratory successively to the smaller value of sensing element signal change in each combination within the specific limits, if the combination of the sensing element in test is the fault combination, according to the transducing signal incidence relation, change within the specific limits the input numerical value of this two-way sensing element, corresponding SSE will reduce; If instead the test combination is trouble-free, from above-mentioned formula (1), can find out, therefore the value that changes them only can cause SSE further to enlarge, and by changing the input numerical value of every combination, obtains SSE and reduces the most significant combination and can locate impaired sensing element.
Suppose j (j≤10) number and k (k≤10, k ≠ j) number sensing element fault, artificially to them, respectively change a micro-value δ
jand δ
k, now neural network input node will become (referring to formula 2):
X=[x
1,x
2,...,x
j+δ
1,...,x
k+δ
k,...,x
10]
T (2)
The signal of considering a fault sensing element may have three kinds of different {+δ of variation
i,-δ
i, 0} (wherein " 0 " is changed to the trouble-free situation of sensing element), for the signal side-play amount of estimating the test combination needs to carry out 3
2=9 set-up procedures are as follows:
First transducing signal of step 1-3 remains unchanged, and another transducing signal carries out three kinds and changes {+δ
2,-δ
2, 0};
First transducing signal+δ of step 4-6
1, another transducing signal carries out three kinds and changes {+δ
2,-δ
2, 0};
First transducing signal-δ of step 7-9
1, another transducing signal carries out three kinds and changes {+δ
2,-δ
2, 0}.
It is below a kind of parallel fault sensing element method for searching that is similar to tree-shaped path search, 9 steps for each fault combination of estimation, each node need to carry out 9 times and decompose, therefore by this algorithm parallel 9 fork tree algorithms (as Fig. 3) of called after visually.The algorithm ground floor is parallel m node, and each node carries out the calculating of one time 9 fork tree by 9 set-up procedures, and defines a SSE matrix for locating the fault sensing element, referring to formula 3:
The number that wherein m is the fault combination, the row, column value of each data is ordinal number (1~m) and the adjustment mode (1~9) of representing fault combination respectively.
By every group of test sample book input network, by 9 parallel fork tree algorithms, m possible fault combination respectively carried out to the signal offset operation 9 times, result of calculation in input matrix (formula 3), is obtained min (SSE successively
(i, j)), (i≤m, j≤9), be worth corresponding ranks positional value (i, j) according to this and can locate fast the combination of fault sensing element.
If have three road sensing elements to break down in 10 road signals, may exist simultaneously
plant the fault-signal situation, the step at least 3 of estimation side-play amount
3=27; If there is a road sensing element to break down in 10 road signals simultaneously, the fault-signal combined number is
kind, the step at least 3 of estimation side-play amount
aindividual, corresponding calculated amount can be geometric series to be increased, therefore improve algorithm raising localization of fault speed, is very crucial work.
Be generalized to IEEE 1451 intelligent sensors that connect n the relevant property of signal sensing element, if wherein there is fault in the individual element of a (a<n), all possible breakdown sensing element numbers
the step of estimation side-play amount is at least 3
aindividual.
Aforesaid operations has been realized the location of multichannel fault sensing element, and below continuing take the intelligent sensor that has connected 10 road sensing elements (wherein having the two-way fault) is example, and the self-recovery method of fault sensing element signal is described in detail in detail.
The transducing signal internal correlation of grasping according to neural network, after orienting the fault sensing element, 9 fork tree algorithms further expand, and can be used for recovering the fault sensed values.Choose the father node of a node of SSE minimum in the child node layer as lower one deck, change side-play amount step-length δ, proceed 9 fork trees and launch (as Fig. 4).Repeat this step, until SSE is less than the setting threshold Δ
sSEtill, by reading each sensed values X now
r, can realize the reconstruct of transducing signal, self-healing transducing signal is:
X
R=[x
1,x
2,...,x
j+δ
j,...,x
k+δ
k,...,x
10](4)
For the situation that has connected 10 road sensing elements (have three roads or more multichannel sensing element fault) simultaneously, its transducing signal is also the mode that adopts similar expansion 9 fork trees from rejuvenation, the node by choosing SSE minimum in the child node layer is as the father node of lower one deck, change side-play amount step-length δ, proceed tree-shaped expansion.Repeat above-mentioned steps, until SSE is less than Δ
sSEtill, by reading the sensed values X of now reconstruct
r, realize that transducing signal is from recovering.
Be generalized to one and connected the relevant sensing element of n road signal, if wherein there are IEEE 1451 intelligent sensors of fault in a (a<n) circuit component, after the fault sensing element of location, its signal is also by 3 from rejuvenation
aindividual signal set-up procedure, proceed the tree-shaped expansion of expansion.When SSE is less than Δ
sSE, realize the recovery certainly of multichannel fault sensing element signal.
In the present embodiment, obtain SSE matrix of certain group sensing data as following formula:
Min (SSE
(i, j) _ 1)=SSE
(16,4) _ 1=0.95231, its line number i equals 16, means the 16th kind of permutation and combination, looks into sensing fault combination sequence and sees table 1, can locate the corresponding failure sensing element and be combined as { x
2, x
9; Columns j equals 4, means the 4th kind of compensation way, i.e. { x
2+ δ
1, x
9+ δ
2, therefore, can continue to change step-length δ, utilize expansion 9 fork tree algorithms to be reconstructed calculating to it.When expansion depth is 5, SSE
(16,4) _ 5=0.00003<0.0001=Δ
sSE, can think that now signal reconstruction is successful, overall response time is 1.154 seconds.
Table 1
Below with a transducing signal associated close phthalic anhydride (Phthalic Anhydride, C
8h
4o
3) the reactor model is example, embodiment of the present invention is described further in detail:
The main production raw material of phthalic anhydride is o-xylene (Ortho-xylene), and its production technology is in reactor, to use air to be fixed a catalytic oxidation to produce continuously.Obtain stable product quality, must the strict response parameters such as charging rate, temperature and pressure of controlling reactor.Fig. 5 is based on phthalic anhydride still sensing and the control program figure of neural network, in this scheme, relates to the temperature of reaction kettle amount on 3 tunnels, comprising: high-order temperature T
h, the meta temperature T
mand low level temperature T
l; No. 2 reactor amount of pressure, comprise still pressure on top surface P
hwith still bottom pressure P
l; 3 road mass flows, comprise feed rate F
i, salt receiving velocity F
cwith chilled water speed F
l; 1 road cold salt Temperature Quantity T
cand 1 road reacting gas Temperature Quantity T
oetc. quantity of information, these measurers have certain relevance, are applicable to multichannel sensing element localization of fault and signal recovery method that application the present invention proposes.
The above; be only the present invention's embodiment preferably, but protection scope of the present invention is not limited to this, anyly is familiar with in technical scope that those skilled in the art disclose in the present invention; the variation that can expect easily or replacement, within all should being encompassed in protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.
Claims (3)
1. many sensing elements of relevance transducing signal localization of fault and transducing signal self-recovery method, is characterized in that, described method comprises:
A, obtain training sample;
B, foundation meet the neural network of multichannel sensing element signal relevance, and the relevance between the study transducing signal;
Each signal relevance relation of C, utilization study, will, containing the transducing signal input neural network of fault sensing element, according to the difference between network output signal and input signal, carry out the fault judgement to input signal;
D, orient multichannel fault sensing element;
Change the value of multichannel sensing element signal in fixed range, make it meet the relevance relation between all each road transducing signal xi, wherein i=1,2, ..., n, and orient multichannel fault sensing element according to the difference change conditions of network output signal and input signal;
E, recovery multichannel fault transducing signal;
According to transducing signal relevance relation, continue to change the signal of having located the fault sensing element, make the output signal of neural network and input signal difference be less than setting threshold, thereby automatically recover the fault transducing signal;
The neural network algorithm of multichannel sensing element signal relevance is to utilize a kind of symmetrical topological structure neural network algorithm, and learns described relevance by training sample;
The symmetrical topological structure neural network of described foundation has unity gain, its input vector x
iequal output vector y
i, i=1 wherein, 2 ..., n;
Described neural network structure comprises an input layer, three hidden layers and an output layer; The hidden layer ground floor is mapping layer, and node dimension p is maximum in whole network, and for extracting input message relevance part, its node transport function can be Sigmoid function (f (x)=1/ (1+e
-x)) or other similar nonlinear functions, the weight of this node layer function is w
c-jj=1 wherein, 2 ..., p, p>n; The last one deck of hidden layer is the demapping layer, and the node dimension equates with mapping layer, and for information is carried out to the relevance reduction, its node is nonlinear transfer function, and the weight of this node layer function is w
d-j, j=1 wherein, 2 ..., p, p>n.
2. relevance transducing signal many sensing elements localization of fault according to claim 1 and transducing signal self-recovery method, is characterized in that, after setting up symmetrical topological structure neural network, by definition neural network output vector y
iwith input vector x
ibetween error sum of squares
index is as multichannel sensing element fault evaluation factors.
3. according to claim 1 or the described relevance transducing signal of 2 any one many sensing elements localization of fault and transducing signal self-recovery method, it is characterized in that, described method is: Ge road transducing signal x
ibetween there is relevance, and each road transducing signal all can impact other road transducing signal lower applicable, i=1 wherein, 2 ..., n is expressed as by mathematical model: x
1=f
1(x
2, x
3..., x
n), x
2=f
2(x
1, x
3..., x
n) ..., x
n=f
n(x
1, x
2..., x
n-1).
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201110077679 CN102169555B (en) | 2011-03-30 | 2011-03-30 | Relevance sensing signal multiple sensing element fault location and sensing signal self-recovery method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 201110077679 CN102169555B (en) | 2011-03-30 | 2011-03-30 | Relevance sensing signal multiple sensing element fault location and sensing signal self-recovery method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN102169555A CN102169555A (en) | 2011-08-31 |
CN102169555B true CN102169555B (en) | 2013-12-25 |
Family
ID=44490712
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 201110077679 Expired - Fee Related CN102169555B (en) | 2011-03-30 | 2011-03-30 | Relevance sensing signal multiple sensing element fault location and sensing signal self-recovery method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102169555B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107636693B (en) * | 2015-03-20 | 2022-01-11 | 弗劳恩霍夫应用研究促进协会 | Relevance score assignment for artificial neural networks |
CN108416309B (en) * | 2018-03-14 | 2022-02-18 | 揭阳职业技术学院 | Multi-fault sensing signal reconstruction method for intelligent sensor |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6636841B1 (en) * | 1997-04-01 | 2003-10-21 | Cybula Ltd. | System and method for telecommunications system fault diagnostics |
CN100438443C (en) * | 2006-05-08 | 2008-11-26 | 国家数字交换系统工程技术研究中心 | Method and system for fault management |
-
2011
- 2011-03-30 CN CN 201110077679 patent/CN102169555B/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
CN102169555A (en) | 2011-08-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20200150622A1 (en) | Method for detecting abnormity in unsupervised industrial system based on deep transfer learning | |
CN108510153B (en) | Multi-working-condition rotary machine fault diagnosis method | |
CN102169555B (en) | Relevance sensing signal multiple sensing element fault location and sensing signal self-recovery method | |
Morley | Palynology of tertiary and quaternary sediments in Southeast Asia | |
CN104217214A (en) | Configurable convolutional neural network based red green blue-distance (RGB-D) figure behavior identification method | |
CN111737592B (en) | Recommendation method based on heterogeneous propagation collaborative knowledge sensing network | |
CN103810497A (en) | Memristor based image identification system and method | |
CN109491338A (en) | A kind of relevant method for diagnosing faults of multimode procedure quality based on sparse GMM | |
CN110533086A (en) | The semi-automatic mask method of image data | |
CN103322416A (en) | Pipeline weak leakage detecting device and detecting method based on fuzzy hyperbolic chaos model | |
CN112949753A (en) | Satellite telemetry time sequence data anomaly detection method based on binary relation | |
CN112881601A (en) | Moisture detection system based on cloud platform | |
CN112016766A (en) | Oil and gas well drilling overflow and leakage early warning method based on long-term and short-term memory network | |
Sunil et al. | A review on prediction of crop yield using machine learning techniques | |
CN104809205B (en) | A kind of online network of waterways space-time accident detection method | |
CN114359791A (en) | Group macaque appetite detection method based on Yolo v5 network and SlowFast network | |
CN114372538B (en) | Method for convolution classification of scale vortex time series in towed sensor array | |
CN104679844A (en) | Intermittent process batch data synchronizing method based on improved DTW (Dynamic Time Wrapping) algorithm | |
Simon et al. | Prediction and classification of different phases in a fermentation using neural networks | |
Clausius et al. | Effect of climate variability on weaning mass in a declining population of southern elephant seals Mirounga leonina | |
CN115077594B (en) | Swarm unmanned aerial vehicle fault detection method based on LSTM and neighbor trust mechanism | |
CN115800274A (en) | Automatic self-adaption method and device for feeder of 5G power distribution network and storage medium | |
CN114841621A (en) | Method for detecting abnormal operation state of synthetic ammonia process based on online neighbor embedding index | |
CN106893726A (en) | A kind of promoter and restructuring yeast strains | |
CN108416309B (en) | Multi-fault sensing signal reconstruction method for intelligent sensor |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20131225 |
|
CF01 | Termination of patent right due to non-payment of annual fee |