CN103235973A - Transformer fault diagnosis method based on radial basis function neural network - Google Patents
Transformer fault diagnosis method based on radial basis function neural network Download PDFInfo
- Publication number
- CN103235973A CN103235973A CN2013101331562A CN201310133156A CN103235973A CN 103235973 A CN103235973 A CN 103235973A CN 2013101331562 A CN2013101331562 A CN 2013101331562A CN 201310133156 A CN201310133156 A CN 201310133156A CN 103235973 A CN103235973 A CN 103235973A
- Authority
- CN
- China
- Prior art keywords
- neural net
- method based
- base neural
- fault type
- radial base
- 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.)
- Granted
Links
- 238000003745 diagnosis Methods 0.000 title claims abstract description 40
- 238000000034 method Methods 0.000 title claims abstract description 37
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 6
- 230000001537 neural effect Effects 0.000 claims description 41
- 238000012549 training Methods 0.000 claims description 35
- 239000006185 dispersion Substances 0.000 claims description 12
- HSFWRNGVRCDJHI-UHFFFAOYSA-N Acetylene Chemical compound C#C HSFWRNGVRCDJHI-UHFFFAOYSA-N 0.000 claims description 11
- CURLTUGMZLYLDI-UHFFFAOYSA-N Carbon dioxide Chemical compound O=C=O CURLTUGMZLYLDI-UHFFFAOYSA-N 0.000 claims description 11
- 206010020843 Hyperthermia Diseases 0.000 claims description 10
- 230000036031 hyperthermia Effects 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000010606 normalization Methods 0.000 claims description 7
- 238000013021 overheating Methods 0.000 claims description 7
- OTMSDBZUPAUEDD-UHFFFAOYSA-N Ethane Chemical compound CC OTMSDBZUPAUEDD-UHFFFAOYSA-N 0.000 claims description 6
- 125000000219 ethylidene group Chemical group [H]C(=[*])C([H])([H])[H] 0.000 claims description 6
- 238000013461 design Methods 0.000 claims description 4
- 230000006870 function Effects 0.000 description 25
- 239000007789 gas Substances 0.000 description 21
- 238000010586 diagram Methods 0.000 description 9
- 238000012360 testing method Methods 0.000 description 9
- 239000011159 matrix material Substances 0.000 description 6
- VNWKTOKETHGBQD-UHFFFAOYSA-N methane Chemical compound C VNWKTOKETHGBQD-UHFFFAOYSA-N 0.000 description 6
- 229910002092 carbon dioxide Inorganic materials 0.000 description 5
- 230000005611 electricity Effects 0.000 description 3
- 210000002569 neuron Anatomy 0.000 description 3
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 description 2
- 239000001569 carbon dioxide Substances 0.000 description 2
- 229910002091 carbon monoxide Inorganic materials 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 229910052739 hydrogen Inorganic materials 0.000 description 2
- 239000001257 hydrogen Substances 0.000 description 2
- 125000004435 hydrogen atom Chemical class [H]* 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 241001269238 Data Species 0.000 description 1
- VGGSQFUCUMXWEO-UHFFFAOYSA-N Ethene Chemical compound C=C VGGSQFUCUMXWEO-UHFFFAOYSA-N 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 125000002534 ethynyl group Chemical group [H]C#C* 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 229930195733 hydrocarbon Natural products 0.000 description 1
- 150000002430 hydrocarbons Chemical class 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 241000894007 species Species 0.000 description 1
Images
Landscapes
- Other Investigation Or Analysis Of Materials By Electrical Means (AREA)
Abstract
The invention discloses a transformer fault diagnosis method based on a radial basis function neural network. According to the method, the content of characteristic gas in insulating oil can be used as input for the radial basis function neural network, transformer faults are output accurately, and accordingly accuracy in transformer fault diagnosis is improved greatly and safe and reliable transformer operation is ensured.
Description
Technical field
The present invention relates to a kind of transformer fault diagnosis method, relate in particular to a kind of transformer fault diagnosis method based on radial base neural net.
Background technology
Power transformer is one of most important equipment in the system of State Grid, also is electric system one of the maximum equipment that breaks down, and its running status directly influences the security of system of State Grid.Thereby research transformer fault diagnosis technology, the reliability and the security that improve transformer have great importance.
In the research of transformer fault diagnosis, there are very complicated nonlinear mathematical relation in the sign of fault and fault type, and this makes the suitable mathematical model of diagnosis be difficult to find.Wherein, the internal fault of transformer shows as three types of machinery, electricity and heat, and the back is main for two kinds, and mechanical fault also can reveal to come usually with electricity or hotlist.Solid Insulation of Power Transformer material and oil are under the double action of electricity and heat, can produce gases such as hydrogen, hydro carbons, carbon monoxide and carbon dioxide, and these gases will be dissolved in the oil of transformer inside, by these gaseous species and content in the oil are analyzed, just can judge the fault of transformer.Wherein hydrogen, methane, acetylene, ethane, ethene, carbon monoxide, the main gas used when carbon dioxide is failure judgement are called characteristic gas.
Three-ratio method is based upon on the basis of oil dissolved gas, is the diagnostic method of at present the most basic judgement transformer fault.Three-ratio method is by determining C2H2/C2H4, CH4/H2 and these three ratio sizes of C2H4/C2H6, determining according to three ratio coding rules (table 1) and fault type determination methods (table 2) which kind of fault takes place transformer again.But also there is significant limitation in three-ratio method, only when each component concentration of oil dissolved gas all surpasses threshold values, could adopt three-ratio method to carry out the fault diagnosis of transformer.In addition, lack a lot of codings in the three-ratio method, can cause in the coding rule table of three ratios, can not find corresponding ratio combination, can't carry out fault and judge; Simultaneously, if the data that calculate are in the border of three ratios coding, judge that by three-ratio method the transformer fault that draws is inaccurate, be easy to diagnostic error.
The coding rule of table 1 three ratios
Table 2 fault type determination methods
Summary of the invention
The purpose of this invention is to provide a kind of transformer fault diagnosis method based on radial base neural net, can be with the input as radial base neural net of the content of characteristic gas in the insulating oil, output transformer fault accurately, greatly improve the accuracy rate of transformer fault diagnosis, guarantee that transformer moves safely and reliably.
The present invention adopts following technical proposals:
A kind of transformer fault diagnosis method based on radial base neural net may further comprise the steps:
A: collect the training sample data and be used as input vector;
B: fault type is encoded the corresponding tables of establishment training sample and fault type;
C: make up and train RBF Neural Network, till reaching satisfied precision;
D: the diagnosis sample to be tested, sample to be tested is imported radial base neural net, through output vector after the network calculations, obtain diagnostic result.
Training sample data in the described A step are respectively H2, CH4, C2H4, C2H2, C2H6 and CO2 gas content.
Described training sample data are at first passed through after the normalized fan-in network again, and the normalization formula is x
i=(x
i-x
Min)/(x
Max-x
Min), wherein, the numerical value of Xi representation feature gas, X
MinThe numerical value of numerical value minimum in the expression all gas, X
MaxThe numerical value of numerical value maximum in the expression all gas.
The corresponding tables of training sample and fault type is in the described B step: if be encoded to 100000, then fault type is cryogenic overheating; If be encoded to 010000, then fault type is that middle temperature is overheated; If be encoded to 001000, then fault type is high-energy discharge; If be encoded to 000100, then fault type is hyperthermia and superheating; If be encoded to 000010, then fault type is that ground temperature is overheated; If be encoded to 000001, then fault type is shelf depreciation.
Described C step comprises according to the fault type coding schedule of formulating and training sample and makes up radial base neural net, and training network, till reaching satisfied precision; Described radial basis function adopts Gaussian function, and radial basis function is
Wherein, x is n dimension input vector; C is the center of basis function, and with the vector that x has same dimension, δ determines basis function around the width of central point.
Use the Matlab program in the described C step, choice function newrb() carry out network design, method of calling is Net=newrb(P, T, GOAL, spread), wherein, P is input vector, T is the expectation output vector, GOAL is training precision, and spread is the dispersion constant of basic unit radially, and default value is 1.
The dispersion constant spread=10 of the radially basic unit of described radial basis function diagnosis.
Vector through exporting after the network calculations in the described D step got into 0 less than 0.5 o'clock, was greater than or equal at 0.5 o'clock and got into 1.
The present invention can be with the input as radial base neural net of the content of characteristic gas in the insulating oil, base neural net is radially trained, the final radial base neural net that obtains through training is the output transformer fault exactly, greatly improve the accuracy rate of transformer fault diagnosis, guarantee that transformer moves safely and reliably.
Description of drawings
Fig. 1 is process flow diagram of the present invention;
Fig. 2 is transformer fault diagnosis radial base neural net model synoptic diagram of the present invention;
Fig. 3 for when precision be 0.02, when radially the dispersion constant spread of basic unit is 5, radial base neural net training error variation diagram in the MATLAB software;
Fig. 4 is for being 0.02 when precision, and radially the dispersion constant spread of basic unit is 10 o'clock, radial base neural net training error variation diagram in the MATLAB software;
Fig. 5 is for being 0.02 when precision, and radially the dispersion constant spread of basic unit is 15 o'clock, radial base neural net training error variation diagram in the MATLAB software.
Embodiment
As shown in Figure 1, the transformer fault diagnosis method based on radial base neural net of the present invention may further comprise the steps:
A: collect the training sample data and be used as input vector; The training sample data are respectively H2, CH4, C2H4, C2H2, C2H6 and CO2 gas content, and the training sample data at first pass through after the normalized fan-in network again, and the normalization formula is x
i=(x
i-x
Min)/(x
Max-x
Min), wherein, the numerical value of Xi representation feature gas, X
MinThe numerical value of numerical value minimum in the expression all gas, X
MaxThe numerical value of numerical value maximum in the expression all gas;
B: known training sample is encoded the corresponding tables of establishment training sample and fault type with corresponding fault type respectively; The corresponding tables of training sample and fault type is: if be encoded to 100000, then fault type is cryogenic overheating; If be encoded to 010000, then fault type is that middle temperature is overheated; If be encoded to 001000, then fault type is high-energy discharge; If be encoded to 000100, then fault type is hyperthermia and superheating; If be encoded to 000010, then fault type is that ground temperature is overheated; If be encoded to 000001, then fault type is shelf depreciation; Wherein, six codings are respectively six neurons of the output layer of radial base neural net, respectively corresponding six kinds of fault types.
C: make up and train RBF Neural Network; Make up radial base neural net according to the fault type coding schedule of formulating and training sample, and training network, till reaching satisfied precision; Described radial basis function adopts Gaussian function, and radial basis function is
Wherein, x is n dimension input vector; C is the center of basis function, and with the vector that x has same dimension, δ determines basis function around the width of central point; The present invention uses the Matlab program, choice function newrb() carry out network design, method of calling is Net=newrb(P, T, GOAL, spread), wherein, P is input vector, T is the expectation output vector, and GOAL is training precision, and spread is the dispersion constant of basic unit radially, and default value is 1;
D: the diagnosis sample to be tested, sample to be tested is imported radial base neural net, through output vector after the network calculations, obtain diagnostic result; Through the vector of exporting after the network calculations, get into 0 less than 0.5 o'clock, be greater than or equal at 0.5 o'clock and get into 1.
Owing to adopt these 6 kinds of characteristic gas of H2, CH4, C2H4, C2H2, C2H6 and CO2 as the input vector of network among the present invention, so the input layer number of network is confirmed as 6.In the Fault Identification to transformer, adopt 6 kinds of fault types: cryogenic overheating, middle cryogenic overheating, hyperthermia and superheating, low energy discharge, high-energy discharge and shelf depreciation, the node number of output layer also is 6 like this.Fig. 2 is transformer fault diagnosis radial base neural net model synoptic diagram of the present invention, and wherein, C is middle hidden layer.
At the characteristics of transformer fault, adopt Gaussian function as radial basis function (RBF) among the present invention,
Radial basis function is
Wherein, x is n dimension input vector; C is the center of basis function, and with the vector that x has same dimension, δ determines basis function around the width of central point; Among the present invention, use the Matlab program, choice function newrb() carry out network design.During with the radial primary function network approximating function, newrb() function can increase the hidden neuron number of network automatically, till mean square deviation satisfies precision or neuron number and reaches maximum.Method of calling is as follows:
Net=newrb(P, T, GOAL, spread), wherein, P is input vector, and T is the expectation output vector, and GOAL is training precision, and spread is the dispersion constant of basic unit radially, default value is 1.
Radial basis function has the following advantages:
1: the form of performance is simple, even multivariable input can not increase too much complicacy yet;
2: radial symmetry;
3: be convenient to theoretical analysis;
4: the function slickness is good, and the derivative on rank all exists arbitrarily.
For the learning information that guarantees network accurately and network can not occur saturatedly, and the scale of network is only too big, before the data fan-in network, total data is all carried out normalization, and all data are all dropped in [0,1] interval.The normalization formula is x
i=(x
i-x
Min)/(x
Max-x
Min), wherein, the numerical value of Xi representation feature gas, X
MinThe numerical value of numerical value minimum in the expression all gas, X
MaxThe numerical value of numerical value maximum in the expression all gas.
When carrying out simulation training in utilizing MATLAB2008 software, experiment is chosen 269 groups and was carried out the fault type established data as training sample, and 64 groups of fault datas are as test samples.In MATLAB, carry out the language editor, as follows:
p=load('f:\1.txt');
p=p’
t=load('f:\2.txt');
t=t'
p1=load('f:\3.txt');
p1=p1'
pp=load('f:\4.txt');
pp=pp’
net=newrb(p,t,GOAL,spread);
zhenduan=sim(netrbf,p1)
In the above among the language editor, the sample matrix that p representative sample data normalization is later, t represents the fault type matrix of fault sample, p1 represents the matrix after the normalization of sample to be tested, pp represents the right fault type matrix of sample to be tested itself, and the superincumbent content of GOAL and spread was introduced.
As table 3 input layer be connected weights IW and table 4 hidden layer and being connected shown in the weights LW between the output layer between the hidden layer,
Table 3: be connected weights IW between input layer and the hidden layer
0.2092 | 0.8541 | 0.3 | 0.2766 | 0.0144 | 0.0639 |
1 | 0.3833 | 0.037 | 0.0903 | 0.3828 | 0.4998 |
0.3102 | 1 | 0.695 | 1 | 0.0574 | 0.5937 |
0.5041 | 0.4042 | 0.341 | 0.168 | 0.0065 | 0.4344 |
0.2806 | 0.3416 | 0.133 | 0.2258 | 0.0028 | 0.0462 |
0.2112 | 0.2798 | 0.347 | 0.2531 | 0.0225 | 0.0324 |
0.3878 | 0.0065 | 0.0031 | 0.0005 | 0.0006 | 0.056 |
0.3571 | 0.7397 | 0.18 | 0.0508 | 0.0225 | 0.2343 |
0.3878 | 0.0065 | 0.0031 | 0.0005 | 0.0006 | 0.056 |
0.5918 | 0.6725 | 1 | 0.2446 | 0.0038 | 0.1973 |
0.8071 | 0.2623 | 0.169 | 0.1054 | 0.0135 | 0.0979 |
0.1306 | 0.2818 | 0.27 | 0.1155 | 0.0005 | 0.514 |
0.0143 | 0.0303 | 0.034 | 0.0135 | 0 | 0.5739 |
0.4796 | 0.0256 | 0.016 | 0.0096 | 0.0439 | 0.4937 |
0.0898 | 0.1013 | 0.3306 | 0.0087 | 0.007 | 0.0257 |
0.2942 | 0.1509 | 0.031 | 0.0977 | 1 | 0.5208 |
0.4796 | 0.0256 | 0.016 | 0.0096 | 0.0574 | 0.4937 |
0.5429 | 0.2791 | 0.171 | 0.1007 | 0.0091 | 0.1269 |
0.8 | 0.1769 | 0.285 | 0.5344 | 0.0417 | 0.1084 |
0.2735 | 0.3779 | 0.129 | 0.1724 | 0.0176 | 0.0569 |
0.0321 | 0.0276 | 0.0103 | 0.0051 | 0.0123 | 0.5604 |
0.7816 | 0.6678 | 0.116 | 0.1251 | 0.0045 | 0.0176 |
0.4031 | 0.0639 | 0.01 | 0.0337 | 0.0833 | 0.4628 |
0.251 | 0.7128 | 0.216 | 0.2409 | 0.027 | 0.1899 |
0.2173 | 0.2313 | 0.071 | 0.2076 | 0.0225 | 0.0602 |
0.2592 | 0.0159 | 0.0075 | 0.0028 | 0.0372 | 0.469 |
0.0973 | 0.044 | 0.1991 | 0.0481 | 0.005 | 0.0422 |
0.1578 | 0.0342 | 0.0038 | 0.0088 | 0.172 | 0.3759 |
0.0643 | 0.1318 | 0.119 | 0.1852 | 0.0146 | 0.1232 |
0.8265 | 0.1903 | 0.138 | 0.1989 | 0.016 | 0.349 |
0.0306 | 0.035 | 0.2 | 0.0036 | 0.0236 | 0.3023 |
0.0653 | 0.1681 | 0.146 | 0.1592 | 0.0259 | 0.1633 |
0.2085 | 0.008 | 0.0159 | 0.0018 | 0 | 0.0224 |
0.0153 | 0.0619 | 0.166 | 0.1129 | 0.018 | 0.0719 |
0.8265 | 0.1903 | 0.138 | 0.1989 | 0.016 | 0.349 |
0.0214 | 0.0356 | 0.013 | 0.025 | 0 | 0.5857 |
0.0214 | 0.0531 | 0.013 | 0.0788 | 0 | 0.4875 |
0.0214 | 0.0531 | 0.013 | 0.0788 | 0 | 0.4875 |
0.0398 | 0.0114 | 0.138 | 0.0062 | 0 | 0.4109 |
0.0398 | 0.0114 | 0.138 | 0.0062 | 0 | 0.4109 |
0.2173 | 0.1555 | 0.0141 | 0.0623 | 0.6285 | 0.3726 |
0.0286 | 0.2751 | 0.354 | 0.1302 | 0.0006 | 0.0614 |
0.0847 | 0.0296 | 0.075 | 0.0884 | 0.0053 | 0.0012 |
0.2408 | 0.2757 | 0.159 | 0.1537 | 0.0039 | 0.1232 |
0.3347 | 0.0491 | 0.0093 | 0.0131 | 0.0698 | 0.3615 |
0.2408 | 0.0329 | 0.0133 | 0.0089 | 0.0994 | 0.3936 |
0.1051 | 0.0982 | 0.031 | 0.0534 | 0.0045 | 0.2399 |
0.1031 | 0.0861 | 0.013 | 0.0273 | 0.0113 | 0.2627 |
0.299 | 0.026 | 0.0384 | 0.004 | 0.0579 | 0.0468 |
0.3735 | 0.0815 | 0.0478 | 0.026 | 0.3394 | 0.2282 |
0.2668 | 0.0224 | 0.0279 | 0.0037 | 0.052 | 0.0439 |
0.173 | 0.0194 | 0.0034 | 0.0122 | 0.0915 | 0.3838 |
0.2481 | 0.0199 | 0.0284 | 0.0053 | 0.0529 | 0.045 |
0.2599 | 0.1381 | 0.0447 | 0.0133 | 0.0032 | 0.1327 |
0.0573 | 0.2562 | 0.193 | 0.1844 | 0.0134 | 0.1263 |
0.0484 | 0.0236 | 0.0078 | 0.004 | 0.0117 | 0.4776 |
0.3408 | 0.0268 | 0.0478 | 0.001 | 0.2787 | 0.0668 |
0.1071 | 0.0455 | 0.0406 | 0.033 | 0.0017 | 0.4196 |
0.1432 | 0.1302 | 0.0821 | 0.0803 | 0.0304 | 0.5509 |
0.151 | 0.1607 | 0.113 | 0.2322 | 0.0045 | 0.0824 |
0.3837 | 0.0161 | 0.022 | 0.0048 | 0.0757 | 0.1164 |
0.099 | 0.0935 | 0.028 | 0.0943 | 0.0041 | 0.1349 |
0.1798 | 0.0093 | 0.0081 | 0.0025 | 0.031 | 0.0288 |
0.2143 | 0.2824 | 0.016 | 0.0696 | 0.6417 | 0.469 |
0.0296 | 0.0177 | 0.0018 | 0.0051 | 0.0928 | 0.3226 |
0.2337 | 0.0549 | 0.01 | 0.0174 | 0.0851 | 0.4227 |
0.0571 | 0.0273 | 0.0097 | 0.0045 | 0 | 0.7067 |
0.3245 | 0.0767 | 0.052 | 0.0726 | 0.0372 | 0.1639 |
0.3245 | 0.0767 | 0.052 | 0.0726 | 0.0372 | 0.1639 |
0.0987 | 0.2182 | 0.0568 | 0.007 | 0.0305 | 0.1899 |
0.4447 | 0.0471 | 0.0231 | 0.0097 | 0.0652 | 0.9037 |
0.2775 | 0.0236 | 0.0282 | 0.0037 | 0.0619 | 0.056 |
0.1716 | 0.0072 | 0.0049 | 0.0019 | 0.0274 | 0.0332 |
0.0348 | 0.0348 | 0.0207 | 0.0254 | 0 | 0.3731 |
0.0527 | 0.0235 | 0.0075 | 0.0041 | 0.0116 | 0.4863 |
0.4286 | 0.1345 | 0.014 | 0.0433 | 0.2026 | 0.6789 |
0.0005 | 0.0111 | 0.0031 | 0.0027 | 0.0399 | 0.2344 |
0.0224 | 0.0025 | 0.0008 | 0.0016 | 0.0469 | 0.0429 |
0.0816 | 0.0444 | 0.018 | 0.0333 | 0.5741 | 0.1911 |
0.1127 | 0.0096 | 0.0146 | 0 | 0.096 | 0.0256 |
0.1367 | 0.0921 | 0.013 | 0.029 | 0.0124 | 0.3202 |
0.0571 | 0.1923 | 0.096 | 0.1746 | 0.0081 | 0.0367 |
0.0343 | 0.0187 | 0.0109 | 0.0139 | 0.0023 | 0.2263 |
0.0781 | 0.0648 | 0.0235 | 0.0292 | 0 | 0.9788 |
0.0781 | 0.0648 | 0.0235 | 0.0292 | 0 | 0.9788 |
0.2959 | 0.074 | 0.0068 | 0.0088 | 0.1013 | 0.9506 |
0.0235 | 0.041 | 0.014 | 0.0288 | 0.005 | 0.3158 |
0.0559 | 0.0115 | 0.0044 | 0.0015 | 0.0096 | 0.2359 |
0.1194 | 0.1843 | 0.054 | 0.0523 | 0.0033 | 0.0275 |
0.0352 | 0.0147 | 0.0032 | 0.0085 | 0.0221 | 0.0528 |
0.0235 | 0.041 | 0.014 | 0.0288 | 0.005 | 0.3158 |
0.0908 | 0.0067 | 0.003 | 0.0034 | 0 | 0.2572 |
0.299 | 0.0336 | 0.013 | 0.0216 | 0.1351 | 0.1973 |
0.0352 | 0.0224 | 0.0061 | 0.0172 | 0.0012 | 0.1096 |
0.0735 | 0.2972 | 0.221 | 0.0868 | 0.0008 | 0.0738 |
0.0735 | 0.2972 | 0.221 | 0.0868 | 0.0008 | 0.0738 |
0.1327 | 0.1681 | 0.036 | 0.0753 | 0.5629 | 0.0491 |
0.0949 | 0.0222 | 0.011 | 0.0113 | 0.0004 | 0.0411 |
0.1378 | 0.1547 | 0.083 | 0.079 | 0.0016 | 0.2899 |
0.0643 | 0.0108 | 0.003 | 0.0043 | 0.0338 | 0.2813 |
0.451 | 0.0572 | 0.01 | 0.0194 | 0.1959 | 0.6184 |
0.1276 | 0.2625 | 0.2019 | 0.0936 | 0.0008 | 0.4554 |
0.1276 | 0.2625 | 0.2019 | 0.0936 | 0.0008 | 0.4554 |
0.0653 | 0.0484 | 0.103 | 0.0013 | 0 | 0.1189 |
0.0653 | 0.0484 | 0.103 | 0.0013 | 0 | 0.1189 |
0.0051 | 0.0042 | 0.0011 | 0.0022 | 0.0509 | 0.0367 |
0.1224 | 0.0404 | 0.0022 | 0.0014 | 0.4165 | 0.0676 |
0.0143 | 0.0044 | 0.026 | 0.0003 | 0 | 0.0861 |
0.0969 | 0.0135 | 0.0033 | 0.0024 | 0.0036 | 0.6851 |
0.0531 | 0.0028 | 0.001 | 0.0018 | 0.0532 | 0.0621 |
0.1122 | 0.0343 | 0.004 | 0.0062 | 0.0642 | 0.0127 |
0.0459 | 0.0309 | 0.013 | 0.0186 | 0.0007 | 0.2158 |
0.0755 | 0.0464 | 0.026 | 0.025 | 0 | 0.3381 |
0.0755 | 0.0464 | 0.026 | 0.025 | 0 | 0.3381 |
0.1416 | 0.0351 | 0.0068 | 0.0118 | 0.0108 | 0.4264 |
0.0068 | 0.0067 | 0.011 | 0.0134 | 0.0044 | 0.2158 |
0.1796 | 0.0894 | 0.014 | 0.0406 | 0.4458 | 0.2152 |
0.0617 | 0.0684 | 0.0323 | 0.0581 | 0.0015 | 0.2501 |
0.0806 | 0.0182 | 0.0068 | 0.0085 | 0.0026 | 0.2714 |
0.0576 | 0.0439 | 0.1066 | 0.0082 | 0.0493 | 0.1399 |
0.0044 | 0.1298 | 0.118 | 0.0233 | 0 | 0.0367 |
0.0044 | 0.1298 | 0.118 | 0.0237 | 0 | 0.0367 |
0.0408 | 0.0128 | 0.0037 | 0.0034 | 0.0545 | 0.2238 |
0.0949 | 0.0114 | 0.005 | 0.0038 | 0.0349 | 0.1263 |
0.0929 | 0.2421 | 0.04 | 0.1355 | 0.0032 | 0.038 |
0.1263 | 0.1839 | 0.0544 | 0.0152 | 0.0196 | 0.0818 |
0.0533 | 0.018 | 0.1158 | 0.183 | 0.4 | 0 |
0.0449 | 0.043 | 0.0232 | 0.0416 | 0 | 0.1158 |
0.0449 | 0.043 | 0.0232 | 0.0416 | 0 | 0.1158 |
0.1081 | 0.1153 | 0.04 | 0.0627 | 0.0265 | 0.3023 |
0.2245 | 0.0062 | 0.0024 | 0.0004 | 0.0002 | 0.1337 |
0 | 0.0451 | 0.108 | 0.0275 | 0 | 0.0602 |
0.1265 | 0.0773 | 0.04 | 0.0358 | 0.0023 | 0.1788 |
0.0204 | 0.0101 | 0.0032 | 0.0036 | 0 | 0.2405 |
0.0881 | 0.1265 | 0.0714 | 0.0748 | 0.0029 | 0.3436 |
0.0755 | 0.0128 | 0.008 | 0.0055 | 0.0653 | 0.2535 |
0.0316 | 0.0296 | 0.046 | 0.0391 | 0.0028 | 0.1528 |
0.0044 | 0.1298 | 0.118 | 0.0235 | 0 | 0.0367 |
0.0005 | 0.0076 | 0.0028 | 0.0017 | 0.0036 | 0.2279 |
0.3635 | 0.0331 | 0.049 | 0.0355 | 0.0587 | 0.0649 |
0.0567 | 0.1477 | 0.0235 | 0.0718 | 0.0491 | 0 |
0.1633 | 0.15 | 0.045 | 0.0932 | 0.0124 | 0.2399 |
0.0837 | 0.1083 | 0.079 | 0.0764 | 0.01 | 0.6777 |
0.0837 | 0.1083 | 0.079 | 0.0764 | 0.01 | 0.6777 |
0.0195 | 0.0192 | 0.0035 | 0.0076 | 0 | 0.8589 |
0.0571 | 0.0672 | 0.0514 | 0.0468 | 0.0026 | 0.031 |
0.0694 | 0.0706 | 0.025 | 0.0442 | 0.0006 | 0.0398 |
0.0505 | 0.0185 | 0.0062 | 0.0035 | 0.009 | 0.4159 |
0.0561 | 0.0323 | 0.006 | 0.0134 | 0.0146 | 0.1077 |
0.272 | 0.502 | 0.2094 | 0.2019 | 0.0017 | 0.1176 |
0.0486 | 0.0104 | 0.003 | 0.0033 | 0.0391 | 0.3489 |
0.0224 | 0.0022 | 0.0012 | 0.0004 | 0.0002 | 0.2158 |
0.0334 | 0.0098 | 0.0028 | 0.003 | 0 | 0.6678 |
0.1482 | 0.0062 | 0.0126 | 0.001 | 0.0239 | 0.0305 |
0.043 | 0.0188 | 0.01 | 0.0096 | 0.0065 | 0.125 |
0.0547 | 0.0132 | 0.0048 | 0.0017 | 0.0096 | 0.3609 |
0.0476 | 0.0156 | 0.0044 | 0.0038 | 0.0609 | 0.2655 |
0.0166 | 0.0055 | 0.002 | 0.0031 | 0.0133 | 0.053 |
0 | 0.0022 | 0.0016 | 0.0014 | 0.0096 | 0.0164 |
0.0041 | 0.0013 | 0.001 | 0.0045 | 0 | 0.2516 |
0.0157 | 0.0011 | 0.0009 | 0.001 | 0.0243 | 0.0035 |
0.028 | 0.0425 | 0.0327 | 0.0259 | 0.0006 | 0.0343 |
0.0439 | 0.2999 | 0.0649 | 0.0571 | 0.0057 | 0.0952 |
0.1748 | 0.0171 | 0.0284 | 0.0181 | 0.0238 | 0.0304 |
0.0649 | 0.0298 | 0.0102 | 0.0057 | 0.012 | 0.7246 |
0.1592 | 0.0003 | 0.0005 | 0.0102 | 0.1362 | 0.0466 |
0.0184 | 0.002 | 0.001 | 0.0006 | 0.0048 | 0.1541 |
0.0051 | 0.0008 | 0.0009 | 0.0006 | 0.0037 | 0.0713 |
0.3418 | 0.0451 | 0.018 | 0.0269 | 0.1914 | 0.1849 |
0.0152 | 0.009 | 0.0064 | 0.0108 | 0.0025 | 0.1559 |
0.0823 | 0.1141 | 0.0453 | 0.0589 | 0.0032 | 0.2998 |
0 | 0.0451 | 0.108 | 0.0275 | 0 | 0.0602 |
0.0106 | 0.0032 | 0.0011 | 0.0005 | 0.002 | 0.0404 |
0.001 | 0.0007 | 0.0003 | 0.0001 | 0.0001 | 0.0244 |
0.0112 | 0.0054 | 0.0005 | 0.0021 | 0.0079 | 0.0182 |
0.0163 | 0.1594 | 0.092 | 0.0884 | 0 | 0.0966 |
0.0286 | 0.098 | 0.113 | 0.082 | 0 | 0.1423 |
0.0806 | 0.1722 | 0.132 | 0.1058 | 0 | 0.1454 |
0.0806 | 0.1722 | 0.132 | 0.1058 | 0 | 0.1454 |
0.0786 | 0.0807 | 0.044 | 0.0395 | 0.0009 | 0.1911 |
0.0398 | 0.0841 | 0.121 | 0.0775 | 0 | 0.0849 |
0.0551 | 0.1439 | 0.133 | 0.1383 | 0.0146 | 0.1257 |
0.0347 | 0.0052 | 0.0025 | 0.0019 | 0.0011 | 0.08 |
0.0475 | 0.0156 | 0.0044 | 0.0038 | 0.0609 | 0.2655 |
0.0969 | 0.0874 | 0.055 | 0.0414 | 0.0014 | 0.222 |
0.0745 | 0.074 | 0.041 | 0.0358 | 0 | 0.117 |
0.1082 | 0.0161 | 0.004 | 0.0053 | 0.0417 | 0.3418 |
0.0167 | 0.0168 | 0.011 | 0.004 | 0 | 0.9667 |
0.0167 | 0.0168 | 0.011 | 0.004 | 0 | 0.9667 |
0.0433 | 0.0063 | 0.0029 | 0.0054 | 0.0092 | 0.0973 |
0.2041 | 0.0323 | 0.014 | 0.022 | 0.1475 | 0.1541 |
0.0072 | 0.0514 | 0.0171 | 0.0183 | 0 | 0.0849 |
0.0072 | 0.0514 | 0.0171 | 0.0183 | 0 | 0.0849 |
0.0184 | 0.0538 | 0.07 | 0.0348 | 0 | 0.1145 |
0.0173 | 0.0094 | 0.0049 | 0.0024 | 0.0059 | 0.4653 |
0.0534 | 0.0137 | 0.0052 | 0.0018 | 0.0104 | 0.3555 |
0.0197 | 0.004 | 0.0019 | 0.0025 | 0.0097 | 0.0367 |
0.0235 | 0.0027 | 0.002 | 0.0019 | 0.0047 | 0.0065 |
0.0422 | 0.0154 | 0.0042 | 0.0035 | 0.0672 | 0.2762 |
0.0974 | 0.0727 | 0.0292 | 0.0364 | 0 | 0.9751 |
0.0245 | 0.0282 | 0.0036 | 0.0292 | 0.0039 | 0.0256 |
0.0245 | 0.0282 | 0.0036 | 0.0292 | 0.0039 | 0.0256 |
0.0098 | 0.0031 | 0.001 | 0.0004 | 0.0025 | 0.0331 |
0.0005 | 0.0008 | 0.0005 | 0.0004 | 0.002 | 0.0157 |
0.1429 | 0.1614 | 0.085 | 0.0772 | 0.0012 | 0.259 |
0.1429 | 0.1614 | 0.085 | 0.0772 | 0.0012 | 0.259 |
0.1633 | 0.0605 | 0.027 | 0.0032 | 0.0056 | 0.469 |
0.0757 | 0.0692 | 0.0254 | 0.0343 | 0 | 0.9801 |
0.049 | 0.0128 | 0.004 | 0.0045 | 0.0006 | 0.3578 |
0.0305 | 0.0162 | 0.0343 | 0.0285 | 0.0007 | 0.0185 |
0.0827 | 0.0049 | 0.0063 | 0.005 | 0.125 | 0.1788 |
0.0541 | 0.0921 | 0.115 | 0.0984 | 0.0073 | 0.0775 |
0.0824 | 0.0702 | 0.0278 | 0.0354 | 0 | 0.9162 |
0.0356 | 0.0409 | 0.068 | 0.0788 | 0.0016 | 0.0333 |
0.1184 | 0.0135 | 0.007 | 0.0051 | 0.0608 | 0.1392 |
0.1184 | 0.0135 | 0.007 | 0.0051 | 0.0608 | 0.1392 |
0.1133 | 0.0135 | 0.006 | 0.0045 | 0.0484 | 0.1337 |
0.0102 | 0.0027 | 0.002 | 0.0009 | 0.0034 | 0.0448 |
0.0327 | 0.1063 | 0.111 | 0.1186 | 0.0135 | 0.0948 |
0.0327 | 0.1063 | 0.111 | 0.1186 | 0.0135 | 0.0948 |
0 | 0.0025 | 0.0026 | 0.0012 | 0 | 0.0281 |
0.1694 | 0.0438 | 0.0059 | 0.0115 | 0.0656 | 0.9907 |
0.066 | 0.0365 | 0.0143 | 0.0188 | 0 | 0.8533 |
0.2408 | 0.0426 | 0.0077 | 0.0143 | 0.0783 | 1 |
0.1041 | 0.0026 | 0.0008 | 0.0014 | 0.0441 | 0.0429 |
0.2738 | 0.0261 | 0.0217 | 0.0017 | 0.052 | 0.0434 |
0.0296 | 0.0054 | 0.0012 | 0.0007 | 0.0047 | 0.0553 |
0.0494 | 0.0118 | 0.0063 | 0.0081 | 0.0004 | 0.0119 |
0.301 | 0.0247 | 0.0268 | 0.0051 | 0.0575 | 0.0482 |
0.0673 | 0.1681 | 0.145 | 0.1648 | 0.0225 | 0.1516 |
0.0367 | 0.0034 | 0.0018 | 0.0005 | 0.0007 | 0.3393 |
0.0827 | 0.1063 | 0.03 | 0.0228 | 0.0003 | 0.5943 |
0.0316 | 0.0511 | 0.0345 | 0.0259 | 0 | 0.2467 |
0.0143 | 0.0195 | 0.018 | 0.0335 | 0 | 0.046 |
0.0613 | 0.0135 | 0.0055 | 0.0018 | 0.0102 | 0.3584 |
0.1755 | 0.0413 | 0.0109 | 0.0153 | 0.1531 | 0.3412 |
0.0449 | 0.0377 | 0.0096 | 0.0115 | 0.0901 | 0.2467 |
0.0316 | 0.0511 | 0.0345 | 0.0259 | 0 | 0.2467 |
0.1645 | 0.0258 | 0.0044 | 0.0066 | 0.1412 | 0.1837 |
0.1939 | 0.0161 | 0.007 | 0.0032 | 0.0529 | 0.6789 |
0.001 | 0.0075 | 0.0049 | 0.0105 | 0 | 0.0676 |
0 | 0 | 0 | 0.0017 | 0 | 0.0251 |
0.0149 | 0.0073 | 0.0031 | 0.0041 | 0.1216 | 0.0506 |
0.0592 | 0.0968 | 0.0018 | 0.0309 | 0.041 | 0.0929 |
0.0592 | 0.0968 | 0.0018 | 0.0309 | 0.041 | 0.0929 |
0.1439 | 0.0948 | 0.013 | 0.0258 | 0.0124 | 0.3165 |
0.0041 | 0.0009 | 0.0004 | 0.0001 | 0.0005 | 0.0337 |
0.0367 | 0.0114 | 0.004 | 0.0024 | 0.0061 | 0.8722 |
0.0827 | 0.0049 | 0.0063 | 0.005 | 0.125 | 0.1788 |
0.0254 | 0.015 | 0.0075 | 0.0023 | 0.0298 | 0.5869 |
0.1306 | 0.0686 | 0.039 | 0.0591 | 0.0158 | 0.1411 |
0.0346 | 0.0245 | 0.0315 | 0.0074 | 0 | 0.8716 |
0.0092 | 0.0056 | 0.0037 | 0.0056 | 0.0056 | 0.0386 |
0.1183 | 0.051 | 0.0268 | 0.0387 | 0.0165 | 0.0362 |
0.0005 | 0.0011 | 0.0005 | 0.0001 | 0 | 0.0275 |
0.0214 | 0.0067 | 0.001 | 0.0007 | 0.0052 | 0.0831 |
0.0306 | 0.005 | 0.0018 | 0.0016 | 0.0214 | 0.0985 |
0.0582 | 0.1513 | 0.133 | 0.1357 | 0.018 | 0.1269 |
0.152 | 0.2469 | 0.1016 | 0.108 | 0.0018 | 0.1244 |
0.055 | 0.0613 | 0.0282 | 0.0342 | 0 | 0.8701 |
0.0005 | 0.0024 | 0.0007 | 0.0011 | 0.0087 | 0.0028 |
0.0583 | 0.066 | 0.0275 | 0.0321 | 0.0021 | 0.2708 |
0.1653 | 0.0235 | 0.0056 | 0.0056 | 0.0495 | 0.6937 |
0.2041 | 0.0323 | 0.0141 | 0.022 | 0.1475 | 0.1541 |
0.0204 | 0.0081 | 0.0012 | 0.0105 | 0.3085 | 0.0213 |
0.0122 | 0.0027 | 0.002 | 0.0011 | 0.0045 | 0.0448 |
0.0041 | 0.002 | 0.0011 | 0.0003 | 0 | 0.054 |
0.0427 | 0.0652 | 0.157 | 0.1129 | 0.0006 | 0.1306 |
0.0609 | 0.0663 | 0.0246 | 0.0328 | 0 | 0.8556 |
Table 4: be connected weights LW between hidden layer and the output layer
-1.96E+10 | 1.16E+10 | 4.99E+08 | 8.53E+09 | -4.75E+09 | 3.69E+09 |
-5.28E+09 | 4.48E+08 | 5.53E+08 | 3.29E+09 | -3.01E+08 | 1.29E+09 |
6.86E+08 | 2.70E+07 | -1.17E+08 | -7.05E+08 | 9.63E+07 | 1.23E+07 |
1.56E+10 | -1.42E+10 | 1.34E+10 | -1.97E+09 | -5.62E+09 | -7.13E+09 |
5.35E+10 | 1.97E+10 | -2.60E+10 | 4.36E+09 | -3.66E+10 | -1.49E+10 |
-1.38E+10 | -3.28E+10 | -3.41E+09 | 4.75E+10 | 1.54E+10 | -1.29E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
3.22E+09 | -4.10E+09 | -1.67E+09 | 2.18E+09 | 8.43E+08 | -4.71E+08 |
-2.30E+11 | 2.37E+11 | -3.87E+09 | -5.81E+10 | -5.38E+10 | 1.09E+11 |
9.63E+08 | 1.16E+09 | -8.80E+08 | -1.72E+09 | 2.78E+08 | 1.95E+08 |
6.11E+10 | -2.16E+10 | 1.23E+09 | -3.72E+10 | 6.29E+09 | -9.84E+09 |
-6.62E+10 | -3.35E+10 | -1.32E+10 | 1.23E+11 | -2.13E+10 | 1.09E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
-1.61E+11 | 2.85E+10 | 1.25E+10 | 1.23E+11 | -3.41E+10 | 3.10E+10 |
-4.68E+10 | -6.85E+10 | 3.40E+10 | 1.03E+11 | -1.52E+10 | -6.59E+09 |
-1.54E+10 | -2.59E+10 | 1.91E+10 | 1.96E+10 | 6.59E+09 | -4.09E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
-1.94E+11 | 7.11E+10 | -2.41E+09 | 1.13E+11 | -2.36E+10 | 3.59E+10 |
-1.60E+09 | 1.18E+09 | 7.40E+08 | 1.91E+08 | -1.28E+09 | 7.61E+08 |
1.12E+11 | -1.19E+10 | -1.62E+10 | -1.13E+11 | 3.42E+10 | -5.50E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
3.88E+09 | 1.58E+09 | 1.49E+08 | -4.02E+09 | -6.99E+08 | -8.89E+08 |
3.02E+11 | 1.82E+10 | 1.69E+10 | -2.58E+11 | -3.03E+10 | -4.85E+10 |
4.17E+10 | 8.54E+09 | -5.70E+09 | -3.91E+10 | -3.40E+09 | -2.08E+09 |
-2.39E+11 | -9.23E+10 | 7.04E+10 | 2.63E+11 | 6.31E+09 | -8.97E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
8.48E+10 | 2.40E+11 | -8.76E+10 | -2.74E+11 | 5.74E+09 | 3.05E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-7.86E+09 | 1.73E+09 | -5.35E+09 | 7.24E+09 | 4.99E+09 | -7.39E+08 |
8.88E+10 | -5.45E+10 | -4.41E+10 | -4.52E+10 | 9.15E+10 | -3.65E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
3.33E+11 | -2.98E+11 | 2.51E+10 | -1.98E+10 | 7.25E+10 | -1.13E+11 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
2.39E+10 | 6.31E+10 | 2.59E+10 | -1.48E+11 | 3.62E+09 | 3.10E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
-9.27E+10 | 1.20E+11 | 2.34E+10 | 2.46E+10 | -1.17E+11 | 4.24E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
2.09E+11 | 2.68E+11 | -2.13E+11 | -2.18E+11 | -7.22E+10 | 2.66E+10 |
-1.67E+10 | 4.78E+10 | 5.75E+09 | -3.98E+10 | -1.32E+10 | 1.62E+10 |
-1.76E+11 | -5.71E+10 | 4.17E+10 | 2.35E+11 | -2.81E+10 | -1.56E+10 |
2.18E+11 | 1.87E+11 | -1.33E+10 | -4.24E+11 | -4.39E+10 | 7.66E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
1.27E+11 | 7.08E+10 | -7.50E+10 | -9.14E+10 | -1.61E+10 | -1.50E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
1.09E+10 | -8.51E+10 | -7.29E+10 | 1.09E+11 | 6.63E+10 | -2.82E+10 |
-1.39E+11 | 1.10E+10 | 3.17E+10 | 1.31E+11 | -4.36E+10 | 8.03E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
-1.25E+11 | 3.92E+10 | 3.71E+10 | 1.78E+10 | -4.85E+09 | 3.60E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
-2.30E+11 | 2.84E+11 | -1.38E+11 | 5.10E+09 | -2.30E+10 | 1.02E+11 |
1.08E+11 | 3.24E+10 | -5.58E+09 | -1.61E+11 | 2.20E+10 | 4.85E+09 |
2.30E+11 | -2.82E+11 | 2.11E+10 | 1.21E+11 | 5.53E+10 | -1.46E+11 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-3.62E+10 | -5.90E+10 | 4.22E+10 | 4.12E+10 | 1.75E+10 | -5.90E+09 |
1.60E+11 | -5.97E+10 | -3.36E+10 | -8.78E+10 | 3.26E+10 | -1.17E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
9.71E+10 | -7.80E+10 | -1.30E+10 | -4.36E+10 | 4.20E+10 | -4.54E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
-2.93E+10 | 3.51E+11 | -1.33E+11 | -1.29E+11 | -1.11E+11 | 5.19E+10 |
6.59E+10 | -1.62E+09 | -2.29E+10 | -4.51E+10 | 1.57E+10 | -1.20E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-4.79E+10 | -4.61E+10 | 3.85E+10 | 5.22E+10 | 3.51E+09 | -2.05E+08 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-5.14E+10 | 4.22E+10 | -1.22E+09 | -5.94E+09 | -6.34E+09 | 2.27E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
3.26E+10 | 1.28E+11 | -1.14E+11 | -8.72E+09 | -2.31E+10 | -1.41E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-2.71E+11 | -5.10E+10 | 2.80E+11 | 1.35E+11 | -8.99E+10 | -2.92E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-1.53E+11 | 3.31E+11 | -1.23E+11 | -2.45E+10 | -5.56E+10 | 2.62E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
2.33E+11 | -8.99E+10 | -2.28E+10 | -8.92E+10 | 2.23E+10 | -5.35E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
-2.21E+10 | 1.75E+10 | 1.16E+09 | -3.48E+08 | -3.81E+09 | 7.58E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
-1.47E+11 | -1.72E+11 | -1.16E+11 | 5.88E+11 | -4.00E+10 | -1.13E+11 |
0 | 0 | 0 | 0 | 0 | 0 |
-7.59E+10 | -4.42E+09 | 8.46E+09 | 2.39E+10 | 1.90E+10 | 2.89E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
6.00E+10 | 7.92E+10 | 2.54E+10 | -2.16E+11 | 5.28E+10 | -1.07E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
2.09E+11 | -3.47E+10 | -9.85E+10 | -6.35E+10 | 1.09E+10 | -2.35E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-3.56E+11 | -4.34E+11 | 3.64E+11 | 3.77E+11 | 1.01E+11 | -5.13E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-3.97E+10 | -1.59E+11 | 1.45E+11 | 1.09E+11 | -5.55E+10 | 4.02E+08 |
-9.60E+10 | 1.27E+11 | -1.25E+11 | -3.18E+09 | 5.28E+10 | 4.44E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
1.19E+11 | 1.40E+11 | -9.18E+10 | -1.93E+11 | 2.45E+08 | 2.52E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
3.59E+10 | 2.35E+09 | -1.70E+10 | -1.68E+10 | -1.08E+09 | -3.35E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
1.99E+11 | 4.14E+10 | -1.28E+11 | -1.42E+11 | 3.10E+10 | -9.19E+08 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
2.56E+11 | -2.01E+11 | 7.37E+10 | -8.07E+10 | 4.51E+10 | -9.30E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-1.10E+11 | -8.63E+10 | 2.03E+10 | 1.47E+11 | 3.74E+10 | -8.45E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-3.77E+10 | -6.91E+10 | -4.70E+10 | 1.83E+11 | 4.80E+09 | -3.43E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-7.19E+10 | -1.91E+11 | 8.56E+10 | 1.51E+11 | 4.43E+10 | -1.74E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-2.02E+11 | 3.93E+11 | -1.37E+11 | -1.22E+11 | -8.32E+10 | 1.51E+11 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
6.22E+10 | 2.46E+10 | 5.74E+09 | -1.09E+11 | 1.64E+10 | 1.75E+08 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-1.22E+11 | 7.61E+10 | -6.24E+10 | 1.04E+11 | -2.46E+10 | 2.86E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
2.02E+11 | -3.13E+11 | 2.51E+11 | -3.56E+11 | 1.79E+11 | 3.73E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
8.65E+09 | -1.65E+11 | 1.58E+11 | -2.36E+10 | 1.63E+10 | 5.03E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
7.98E+09 | -9.49E+10 | 9.74E+10 | 4.28E+10 | -4.74E+10 | -5.91E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
-1.80E+11 | 1.41E+11 | -1.18E+11 | 2.02E+11 | -4.35E+10 | -1.56E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
2.56E+10 | -1.41E+11 | 8.82E+10 | 2.54E+10 | 2.25E+10 | -2.02E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
3.82E+08 | 1.42E+10 | 2.55E+10 | 9.32E+07 | -9.52E+09 | -3.06E+10 |
0 | 0 | 0 | 0 | 0 | 0 |
0 | 0 | 0 | 0 | 0 | 0 |
2.70E+10 | -1.39E+11 | 4.22E+10 | 5.10E+10 | 2.90E+10 | -9.96E+09 |
0 | 0 | 0 | 0 | 0 | 0 |
Be 0.02 when choosing precision, radially the dispersion constant spread of basic unit is 5 o'clock, and radial base neural net training error variation diagram as shown in Figure 3 in the MATLAB software; Radial base neural net is as follows to the diagnostic result of test samples:
By comparing with the test samples physical fault, there are 53 checks correct, accuracy is 82.81%.
Be 0.02 when choosing precision, radially the dispersion constant spread of basic unit is 10 o'clock, and the radial base neural net training error variation diagram that produces in MATLAB software as shown in Figure 4; Radial base neural net is as follows to the result of test samples fault diagnosis:
By comparing with the test samples physical fault, 55 checks are correct, and accuracy is 85.94%.
Be 0.02 when choosing precision, radially the dispersion constant spread of basic unit is 15 o'clock, the radial base neural net training error variation diagram that in MATLAB software, produces as shown in Figure 5: radial base neural net is as follows to the result of test samples diagnosis:
By comparing with the test samples physical fault, 52 checks are correct, and accuracy is 81.25%, and the mean square deviation of l-G simulation test is 0.0479.
By to the comparative result of spread=5, spread=10, spread=14 as diagnosing shown in the correct counting rate meter:
Diagnose correct counting rate meter
? | The test samples number | Diagnose positive exact figures | Accuracy rate |
SPread=5 | 64 | 53 | 82.81% |
SPread=10 | 64 | 55 | 85.94% |
SPread=15 | 64 | 52 | 81.25% |
Can be drawn by the correct counting rate meter of diagnosis and mean square deviation table, the neural function of RBF when the present invention chooses spread=10 is diagnosed as the neural network model of transformer fault diagnosis, weight matrix LW between weight matrix IW between its input layer and the hidden layer (as table), hidden layer and the output layer sees appendix, the threshold value of hidden layer is equal 0.083255, and the threshold value of output layer is respectively-72129,250000,74263 ,-62261 ,-145000 ,-44636.
In diagnosing the sample to be tested process, sample to be tested is imported radial base neural net, through the vector of exporting after the network calculations, press
Handle, get into 0 when the vector of output less than 0.5 the time, be greater than or equal at 0.5 o'clock and get into 1.
Use the transformer fault diagnosis method based on radial base neural net of the present invention and traditional three-ratio method to compare:
It is as shown in the table to use three-ratio method that 60 groups of sample check data are carried out fault diagnosis result, and can't refer in the table has no idea to determine fault type:
The three-ratio method diagnostic result
1 | l | l | Can't |
l | l | 0 | The low energy discharge |
l | 2 | 2 | Can't |
l | 0 | l | The low energy discharge |
1 | 0 | 2 | High- |
l | |||
0 | 2 | High-energy discharge | |
1 | 0 | 2 | High- |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
|
0 | 0 | Can't |
0 | 0 | l | |
0 | 0 | 2 | Can't |
l | 0 | l | The low energy discharge |
l | 0 | l | The low |
l | |||
0 | 1 | The low energy discharge |
l | 0 | l | The low energy discharge |
l | 0 | l | The low energy discharge |
1 | 0 | l | The low energy discharge |
l | 0 | l | The |
0 | 2 | 2 | Hyperthermia and superheating |
0 | 0 | 2 | Can't |
0 | 2 | 2 | Hyperthermia and superheating |
0 | 2 | 2 | Hyperthermia and superheating |
0 | 2 | 2 | Hyperthermia and superheating |
|
0 | 2 | High- |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
l | 0 | l | The low |
l | |||
0 | 1 | The low | |
l | |||
0 | 0 | Can't | |
|
0 | 1 | The low energy discharge |
1 | 0 | l | The low energy discharge |
l | 0 | l | The low energy discharge |
1 | 0 | 2 | High- |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
0 | 2 | 2 | Hyperthermia and superheating |
0 | 2 | 2 | Hyperthermia and superheating |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
0 | 0 | 2 | Can't |
|
0 | 0 | Can't |
1 | 1 | 0 | High-energy discharge |
1 | 0 | 0 | Can't |
1 | 0 | l | The low energy discharge |
1 | 0 | 0 | Can't |
2 | 0 | 0 | Can't |
1 | 0 | 2 | High- |
0 | 0 | 2 | Can't |
0 | |
0 | The |
0 | 0 | l | Cryogenic overheating |
1 | 0 | l | The low energy discharge |
1 | 0 | l | The low energy discharge |
1 | 0 | 1 | The low energy discharge |
1 | 0 | l | The low energy discharge |
1 | 0 | l | The low energy discharge |
l | 0 | l | The low energy discharge |
2 | 0 | 1 | The low energy discharge |
There are 30 groups of data cannot judge fault type as can be seen from the above table, 10 groups of fault type diagnostic error are arranged, so correctly failure judgement type of 32 groups of data is arranged.
Judge and utilize the transformer fault diagnosis method based on radial base neural net of the present invention to carry out fault type, as mentioned in shown in the correct counting rate meter of diagnosis when spread=10, rate of accuracy reached to 85.94%.
Therefore, utilizing the transformer fault diagnosis method based on radial base neural net of the present invention to carry out fault type judges, can be with the input as radial base neural net of the content of characteristic gas in the insulating oil, output transformer fault exactly, greatly improve the accuracy rate of transformer fault diagnosis, guarantee that transformer moves safely and reliably.
Claims (8)
1. transformer fault diagnosis method based on radial base neural net is characterized in that: may further comprise the steps:
A: collect the training sample data and be used as input vector;
B: fault type is encoded the corresponding tables of establishment training sample and fault type;
C: make up and train RBF Neural Network, till reaching satisfied precision;
D: the diagnosis sample to be tested, sample to be tested is imported radial base neural net, through output vector after the network calculations, obtain diagnostic result.
2. the transformer fault diagnosis method based on radial base neural net according to claim 1, it is characterized in that: the training sample data in the described A step are respectively H2, CH4, C2H4, C2H2, C2H6 and CO2 gas content.
3. the transformer fault diagnosis method based on radial base neural net according to claim 2 is characterized in that: described training sample data are at first passed through after the normalized fan-in network again, and the normalization formula is
, wherein, the numerical value of Xi representation feature gas, Xmin represents the numerical value of numerical value minimum in all gas, Xmax represents the numerical value of numerical value maximum in all gas.
4. the transformer fault diagnosis method based on radial base neural net according to claim 3, it is characterized in that: the corresponding tables of training sample and fault type is in the described B step: if be encoded to 100000, then fault type is cryogenic overheating; If be encoded to 010000, then fault type is that middle temperature is overheated; If be encoded to 001000, then fault type is high-energy discharge; If be encoded to 000100, then fault type is hyperthermia and superheating; If be encoded to 000010, then fault type is that ground temperature is overheated; If be encoded to 000001, then fault type is shelf depreciation.
5. the transformer fault diagnosis method based on radial base neural net according to claim 4, it is characterized in that: described C step comprises according to the fault type coding schedule of formulating and training sample structure radial base neural net, and training network, till reaching satisfied precision; Described radial basis function adopts Gaussian function, and radial basis function is
, wherein, x is n dimension input vector; C is the center of basis function, and with the vector that x has same dimension, δ determines basis function around the width of central point.
6. the transformer fault diagnosis method based on radial base neural net according to claim 5 is characterized in that: use Matlab program, choice function newrb(in the described C step) carry out network design, method of calling is Net=newrb(P, T, GOAL, spread), wherein, P is input vector, and T is the expectation output vector, and GOAL is training precision, spread is the dispersion constant of basic unit radially, and default value is 1.
7. the transformer fault diagnosis method based on radial base neural net according to claim 6 is characterized in that: the dispersion constant spread=10 of the radially basic unit of described radial basis function diagnosis.
8. the transformer fault diagnosis method based on radial base neural net according to claim 7 is characterized in that: the vector through exporting after the network calculations in the described D step, get into 0 less than 0.5 o'clock, and be greater than or equal at 0.5 o'clock and get into 1.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310133156.2A CN103235973B (en) | 2013-04-16 | 2013-04-16 | A kind of Diagnosis Method of Transformer Faults based on radial base neural net |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310133156.2A CN103235973B (en) | 2013-04-16 | 2013-04-16 | A kind of Diagnosis Method of Transformer Faults based on radial base neural net |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103235973A true CN103235973A (en) | 2013-08-07 |
CN103235973B CN103235973B (en) | 2016-04-27 |
Family
ID=48884012
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310133156.2A Active CN103235973B (en) | 2013-04-16 | 2013-04-16 | A kind of Diagnosis Method of Transformer Faults based on radial base neural net |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103235973B (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104063577A (en) * | 2014-05-12 | 2014-09-24 | 吉林省电力科学研究院有限公司 | Method for forecasting characteristic gas development tendency in transformer oil based on generalized recurrent neural network |
CN105335759A (en) * | 2015-11-12 | 2016-02-17 | 南方电网科学研究院有限责任公司 | Generating-probability-model-based transformer fault detection method |
CN105930861A (en) * | 2016-04-13 | 2016-09-07 | 西安西拓电气股份有限公司 | Adaboost algorithm based transformer fault diagnosis method |
CN106067066A (en) * | 2016-05-31 | 2016-11-02 | 西安工程大学 | Method for diagnosing fault of power transformer based on genetic algorithm optimization pack algorithm |
CN103810328B (en) * | 2014-01-16 | 2017-02-22 | 国家电网公司 | Transformer maintenance decision method based on hybrid model |
CN106526370A (en) * | 2016-10-27 | 2017-03-22 | 国家电网公司 | RVM transformer fault diagnosing method based on chemical reaction algorithm |
CN107024267A (en) * | 2017-04-12 | 2017-08-08 | 无锡研测技术有限公司 | Weighing-appliance Transducer-fault Detecting Method based on radial base neural net |
CN110174255A (en) * | 2019-06-03 | 2019-08-27 | 国网上海市电力公司 | A kind of transformer vibration signal separation method based on radial base neural net |
CN110879373A (en) * | 2019-12-12 | 2020-03-13 | 国网电力科学研究院武汉南瑞有限责任公司 | Oil-immersed transformer fault diagnosis method with neural network and decision fusion |
CN111060221A (en) * | 2019-12-31 | 2020-04-24 | 云领电气智能科技(苏州)有限公司 | Transformer overheating fault early warning method based on cyclic neural network |
CN111076927A (en) * | 2019-12-19 | 2020-04-28 | 佛山科学技术学院 | Gear fault diagnosis method, device and system based on deep neural network |
CN111947795A (en) * | 2020-08-18 | 2020-11-17 | 广东电网有限责任公司广州供电局 | Transformer overheating early warning method |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102221651A (en) * | 2011-03-11 | 2011-10-19 | 太原理工大学 | Fault on-line diagnosis and early warning method of flameproof dry-type transformer for mine |
KR20110122238A (en) * | 2010-05-04 | 2011-11-10 | 남창현 | Fault diagnosis method and apparatus using headspace gases for transformer |
-
2013
- 2013-04-16 CN CN201310133156.2A patent/CN103235973B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR20110122238A (en) * | 2010-05-04 | 2011-11-10 | 남창현 | Fault diagnosis method and apparatus using headspace gases for transformer |
CN102221651A (en) * | 2011-03-11 | 2011-10-19 | 太原理工大学 | Fault on-line diagnosis and early warning method of flameproof dry-type transformer for mine |
Non-Patent Citations (4)
Title |
---|
JIANLI YU ETAL: "The fault diagnosis of transformer based on BP neural Network", 《2012 INTERNATIONAL CONFERENCE ON INFORMATION MANAGEMENT,INNOVATION MANAGEMENT AND INDUSTRIAL ENGINEERING》 * |
吴宏岐等: "基于RBF神经网络的变压器故障诊断及matlab仿真研究", 《科学技术与工程》 * |
魏星等: "基于径向基函数网络的大型电力变压器故障诊断", 《江苏电机工程》 * |
魏星等: "基于径向基函数网络的大型电力变压器故障诊断", 《江苏电机工程》, vol. 24, no. 6, 30 November 2005 (2005-11-30), pages 19 - 21 * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103810328B (en) * | 2014-01-16 | 2017-02-22 | 国家电网公司 | Transformer maintenance decision method based on hybrid model |
CN104063577B (en) * | 2014-05-12 | 2017-02-15 | 吉林省电力科学研究院有限公司 | Method for forecasting characteristic gas development tendency in transformer oil based on generalized recurrent neural network |
CN104063577A (en) * | 2014-05-12 | 2014-09-24 | 吉林省电力科学研究院有限公司 | Method for forecasting characteristic gas development tendency in transformer oil based on generalized recurrent neural network |
CN105335759B (en) * | 2015-11-12 | 2019-07-16 | 南方电网科学研究院有限责任公司 | A kind of transformer fault detection method based on generating probability model |
CN105335759A (en) * | 2015-11-12 | 2016-02-17 | 南方电网科学研究院有限责任公司 | Generating-probability-model-based transformer fault detection method |
CN105930861A (en) * | 2016-04-13 | 2016-09-07 | 西安西拓电气股份有限公司 | Adaboost algorithm based transformer fault diagnosis method |
CN105930861B (en) * | 2016-04-13 | 2019-08-06 | 西安西拓电气股份有限公司 | A kind of Diagnosis Method of Transformer Faults based on Adaboost algorithm |
CN106067066A (en) * | 2016-05-31 | 2016-11-02 | 西安工程大学 | Method for diagnosing fault of power transformer based on genetic algorithm optimization pack algorithm |
CN106526370A (en) * | 2016-10-27 | 2017-03-22 | 国家电网公司 | RVM transformer fault diagnosing method based on chemical reaction algorithm |
CN107024267A (en) * | 2017-04-12 | 2017-08-08 | 无锡研测技术有限公司 | Weighing-appliance Transducer-fault Detecting Method based on radial base neural net |
CN110174255A (en) * | 2019-06-03 | 2019-08-27 | 国网上海市电力公司 | A kind of transformer vibration signal separation method based on radial base neural net |
CN110174255B (en) * | 2019-06-03 | 2021-04-27 | 国网上海市电力公司 | Transformer vibration signal separation method based on radial basis function neural network |
CN110879373A (en) * | 2019-12-12 | 2020-03-13 | 国网电力科学研究院武汉南瑞有限责任公司 | Oil-immersed transformer fault diagnosis method with neural network and decision fusion |
CN111076927A (en) * | 2019-12-19 | 2020-04-28 | 佛山科学技术学院 | Gear fault diagnosis method, device and system based on deep neural network |
CN111060221A (en) * | 2019-12-31 | 2020-04-24 | 云领电气智能科技(苏州)有限公司 | Transformer overheating fault early warning method based on cyclic neural network |
CN111947795A (en) * | 2020-08-18 | 2020-11-17 | 广东电网有限责任公司广州供电局 | Transformer overheating early warning method |
CN111947795B (en) * | 2020-08-18 | 2022-07-12 | 广东电网有限责任公司广州供电局 | Transformer overheating early warning method |
Also Published As
Publication number | Publication date |
---|---|
CN103235973B (en) | 2016-04-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103235973B (en) | A kind of Diagnosis Method of Transformer Faults based on radial base neural net | |
CN103062076B (en) | A kind of urban water plant pump station separate unit characteristic curve of pump rating method | |
CN107967485B (en) | Fault analysis method and device for electricity metering equipment | |
CN103064008B (en) | A kind of Nolinear analog circuit soft fault diagnostic method based on Hilbert-Huang transform | |
CN113011481B (en) | Electric energy meter function abnormality assessment method and system based on decision tree algorithm | |
CN105546352A (en) | Natural gas pipeline tiny leakage detection method based on sound signals | |
CN104020401A (en) | Cloud-model-theory-based method for evaluating insulation thermal ageing states of transformer | |
CN105117512A (en) | Transformer early-warning value estimation method and apparatus | |
CN103218534B (en) | Right tail-truncated type lifetime data distribution selection method | |
CN106154163A (en) | Battery life state identification method | |
CN103278715B (en) | Power equipment test method | |
CN105242155A (en) | Transformer fault diagnosis method based on entropy weight method and grey correlation analysis | |
CN104966161A (en) | Electric energy quality recording data calculating analysis method based on Gaussian mixture model | |
CN105510729A (en) | Overheating fault diagnosis method of transformer | |
CN105447646A (en) | Health index assessment method for power distribution system | |
CN109412161A (en) | A kind of Probabilistic Load calculation method and system | |
Farooque et al. | Artificial neural network (ANN) based implementation of Duval pentagon | |
Sun et al. | Data augmentation strategy for power inverter fault diagnosis based on wasserstein distance and auxiliary classification generative adversarial network | |
Liu et al. | A real-time estimation framework of carbon emissions in steel plants based on load identification | |
Chen et al. | Fault diagnosis of transformer based on random forest | |
CN105359003A (en) | Hybrid approach to assisted history matching in large reservoirs | |
CN114595762A (en) | Photovoltaic power station abnormal data sequence extraction method | |
Han et al. | Oil-immersed transformer internal thermoelectric potential fault diagnosis based on decision-tree of KNIME platform | |
Xue et al. | Deep learning based production prediction for an enhanced geothermal system (EGS) | |
CN102254085A (en) | Method for measuring reliability of mechanical/electronic product |
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 | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20191219 Address after: Room 2301-2312, 23 / F, block B, Weiye international, intersection of Zhengkai Avenue and Jinxiu Road, Baisha Town, Zhengdong New District, Zhengzhou City, Henan Province Patentee after: Henan Zhengshu Intelligent Technology Co., Ltd Address before: 450000 Henan Province, Zhengzhou District, No. 27 University Road, No. 2 Patentee before: Zhengzhou Institute of Aeronautical Industry Management |