CN106709149B - A kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based and system - Google Patents
A kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based and system Download PDFInfo
- Publication number
- CN106709149B CN106709149B CN201611055395.0A CN201611055395A CN106709149B CN 106709149 B CN106709149 B CN 106709149B CN 201611055395 A CN201611055395 A CN 201611055395A CN 106709149 B CN106709149 B CN 106709149B
- Authority
- CN
- China
- Prior art keywords
- data
- neural network
- burner hearth
- hearth shape
- aluminium
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/23—Design optimisation, verification or simulation using finite element methods [FEM] or finite difference methods [FDM]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Evolutionary Computation (AREA)
- Geometry (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Control Of Steam Boilers And Waste-Gas Boilers (AREA)
- Manufacture Of Iron (AREA)
Abstract
The invention discloses a kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based and systems, this method comprises: 1) obtaining aluminium cell modeling parameters；2) aluminium cell heat transfer finite element model is constructed according to modeling parameters；Different tests parameter is subjected to permutation and combination, constitutes test group；Each group of test parameters in test group is inputted into aluminium cell heat transfer finite element model respectively, simulation calculation is carried out, obtains the corresponding burner hearth shape of each group of test parameters and shell temperature；3) it is put down with shell temperature, electrolyte level and aluminum water as input variable, using burner hearth shape as output variable, constructs BP neural network；4) based on the simulation result training BP neural network in step 2)；5) real-time measurement and acquisition shell temperature, electrolyte level and aluminum water are flat；Input step 4) in trained BP neural network, the burner hearth shape finally predicted.The present invention can quickly and accurately predict aluminium cell three-dimensional burner hearth shape.
Description
Technical field
The invention belongs to aluminium cell fields, more particularly to a kind of aluminium cell three-dimensional burner hearth shape neural network based
Shape real-time predicting method and system.
Background technique
In aluminium cell, electric current maintains the height of electrolytic cell by there is the fused electrolyte of a constant impedance to generate Joule heat
Temperature operation.And fused electrolyte cooled and solidified in contacting side wall, it forms a circle furnace side and stretches one's legs, energy protective side wall is not by high temperature
Melt attack.Suitable burner hearth shape is maintained, the horizontal current in molten aluminum can be reduced, hydromagnetic stability in slot is improved, has
Conducive to the stable operation of production process.It will lead to slot internal stability without reasonable burner hearth shape to deteriorate, reduce current efficiency, sternly
Even bakie can occur for weight, to directly affect electrolysis safety in production and every technical-economic index.
Melt temperature is also difficult to now during aluminum electrolysis to furnace close to 1000 DEG C and with highly corrosive in slot
Thorax shape carries out directly effective observation.In industrial production, dedicated tool is protruded into when being frequently in change poles, is measured several
Location point, to speculate the burner hearth shape of remaining position.It is close the time required to measuring full tank furnace thorax shape since anode change cycle is longer
One month, when measuring, the burner hearth shape of last measurement position often had occurred and that variation, and this method measuring point has
Limit, the furnace side shape accuracy measured is poor, and site operation personnel can only be assisted to judge situation in slot.
In the prior art, patent application " furnace edge shape of aluminium electrolytic tank on-line monitoring system " (application number:
201110439642.8) in, a kind of furnace edge shape of aluminium electrolytic tank on-line monitoring system is disclosed, on electrolytic cell (1) pot shell
It is electrolysed tank furnace side (2) connection temperature probe (3), temperature probe (3) is electrically connected with upper industrial personal computer (6).The patent application can connect
It is continuous to monitor electrolyzer temperature, but the specific method of analysis had both been calculated without open progress furnace edge shape of aluminium electrolytic tank, more no pair
Calculated result is verified, and can not determine whether slot internal furnace shape is regular, whether have the risk of bakie.
Therefore, it is necessary to design a kind of new aluminium cell three-dimensional burner hearth shape real-time predicting method and system.
Summary of the invention
Technical problem solved by the invention is in view of the deficiencies of the prior art, to provide a kind of aluminium neural network based
Electrolytic cell three-dimensional burner hearth shape real-time predicting method and system, train BP neural network using simulation result data, from
And slot internal furnace shape can be rapidly and accurately predicted according to shell temperature.
Technical solution of the invention is as follows:
A kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based, comprising the following steps:
1) obtain aluminium cell modeling parameters, including each inner lining material attribute data and structural parameters, extraneous heat transfer boundary condition,
Technological parameter；
2) the modeling parameters building aluminium cell heat transfer finite element model obtained according to step 1；By different tests parameter
Different values carry out permutation and combination, constitute test group；Each group of test parameters in test group is inputted aluminium cell respectively to pass
Hot finite element model carries out simulation calculation, obtains the corresponding burner hearth shape of each group of test parameters and shell temperature data, be used for
Subsequent training；
3) BP neural network is constructed；
Choose the flat input variable as the BP neural network of shell temperature, electrolyte level and aluminum water；
Choose output variable of the burner hearth shape as the BP neural network；
4) training BP neural network, obtains the BP neural network debugged；
Vi. shell temperature data are obtained from the simulation result of step 2)；Electrolyte level and aluminum water are obtained from slot control machine
Flat data；And all shell temperature data, electrolyte level and aluminium horizontal data are divided into the first training set and first and surveyed
Examination collection；
Vii. burner hearth shape is obtained from the simulation result of step 2)；And all burner hearth shapes are divided into the second training
Collection and the second test set；
Viii. by shell temperature data, electrolyte level and aluminium horizontal data in first training set and described
Burner hearth shape data in second training set carries out BP neural network training as the inputoutput data of the BP neural network,
Until [mean square error is less than 10 for BP neural network convergence-3It is judged to restraining]；
Ix. shell temperature data, electrolyte level and the aluminium horizontal data in first test set are inputted into the receipts
BP neural network after holding back exports the burner hearth shape data of prediction；
X. by the burner hearth shape data of the prediction compared with the burner hearth shape data in the second test set, prediction effect is examined
Fruit；If predictablity rate is higher than 90%, using current BP neural network as the BP neural network debugged, entrance is next
Step；If predictablity rate is lower than 90%, return step I is repartitioned training set and test set and is trained；
5) temperature sensor, real-time measurement shell temperature data are arranged in aluminum cell side and bottom pot shell；While from
Slot control machine obtains electrolyte level and aluminium horizontal data；The shell temperature data and electrolyte level and aluminum water of real-time measurement are put down
Data input step 4) in the BP neural network debugged, the burner hearth shape data finally predicted.
The aluminium cell heat transfer finite element model and method for solving established in above-mentioned step 2, can select Xu Yujie, Lee
Jie, Yin Chenggang, et al. the model and method for solving that propose in " aluminium cell electric-thermal field strength Coupling method calculation method ".
Preferably, test parameters is one of following parameter or a variety of: electrolyte level, aluminum water in above-mentioned step 2
Flat, pole span height, current strength, mulch thickness, outside air temperature.
Preferably, above-mentioned step 3 constructs BP neural network, comprising:
Determine the structure of BP neural network, including the hidden layer number of plies, each node layer number of input layer, hidden layer, output layer；
Determine neuron connection weight and neuron threshold value.
Preferably, in the step 4) further include:
Respectively to shell temperature data, electrolyte level and the aluminium number of levels in first training set and the first test set
According to being normalized；
The burner hearth shape data in second training set and the second test set is normalized respectively；
BP neural network training is carried out based on the data after normalization.
Preferably, in the step 5), first to real-time measurement shell temperature data, from slot control machine obtain electrolyte level and
Aluminium horizontal data is normalized；Input step 4 again) in the BP neural network debugged, the burner hearth finally predicted
Shape data；Anti-normalization processing finally is carried out to the burner hearth shape data of output finally predicted, the burner hearth after being predicted
Shape data.
Preferably, in above-mentioned method, the temperature sensor is below one or more: thermocouple, temperature-sensitive electricity
Resistance, resistance temperature detector, IC temperature sensor.
Interpolation processing is carried out to the shell temperature data of temperature sensor measurement, to seek missing data and improper number
According to substitution value.
A kind of real-time forecasting system of aluminium cell three-dimensional burner hearth shape neural network based, comprising:
Parameter acquiring unit obtains each inner lining material attribute data of aluminium cell and structural parameters, extraneous heat transfer boundary condition, work
Skill parameter；
Simulation unit, the modeling parameters building aluminium cell heat transfer finite element model for being obtained according to step 1；It will be different
The different values of test parameters carry out permutation and combination, constitute test group；Each group of test parameters in test group is inputted respectively
Aluminium cell heat transfer finite element model, carries out simulation calculation, obtains the corresponding burner hearth shape of each group of test parameters and pot shell temperature
Degree evidence is used for subsequent training；
Neural network construction unit, for constructing BP neural network；
Input/output variable selection unit, it is flat as the BP mind for choosing shell temperature, electrolyte level and aluminum water
Input variable through network；Choose output variable of the burner hearth shape as the BP neural network；
First division unit, for from simulation result in obtain shell temperature data；Electrolyte water is obtained from slot control machine
Gentle aluminium horizontal data；And all shell temperature data, electrolyte level and aluminium horizontal data are divided into the first training set
With the first test set；；
Second division unit, for obtaining burner hearth shape from simulation result；And all burner hearth shapes are divided into
Two training sets and the second test set；
Training unit, by first training set shell temperature data, electrolyte level and aluminium horizontal data and
Burner hearth shape data in second training set carries out BP neural network as the inputoutput data of the BP neural network
Training, until the BP neural network restrains；
Processing unit inputs shell temperature data, electrolyte level and the aluminium horizontal data in first test set
BP neural network after the convergence exports the burner hearth shape data of prediction；
Verification unit, by the burner hearth shape data of the prediction compared with the burner hearth shape data in the second test set, inspection
Test prediction effect；If predictablity rate is higher than 90%, using current BP neural network as the BP neural network debugged；If
Predictablity rate is lower than 90%, then control the first division unit and the second division unit repartition training set and test set again into
Row training；
Temperature detection and processing unit, the temperature sensor including being arranged in aluminum cell side and bottom pot shell, are used for
Real-time measurement shell temperature data；And interpolation processing is carried out to the shell temperature data to temperature sensor measurement, to seek lacking
Lose the substitution value of data and improper data；
Predicting unit: the shell temperature data and electrolyte level of real-time measurement and the input of aluminium horizontal data have been debugged
BP neural network, the burner hearth shape data finally predicted.
Preferably, test parameters is one of following parameter or a variety of: electrolyte water in the simulation unit of above system
It puts down, aluminum water is flat, pole span height, current strength, mulch thickness, outside air temperature.
Preferably, the neural network construction unit of above system, comprising:
First determining module, it is input layer, hidden layer, defeated for determining the structure of BP neural network, including the hidden layer number of plies
Each node layer number of layer out；
Second determining module, for determining neuron connection weight and neuron threshold value.
Preferably, above system further include:
First normalized module, for being carried out respectively to shell temperature data, electrolyte level and aluminium horizontal data
Normalized；
Second normalized module, for burner hearth shape data to be normalized respectively；
Anti-normalization processing module, for carrying out anti-normalization processing to the burner hearth shape data of output finally predicted,
Burner hearth shape data after being predicted.
The utility model has the advantages that
The method of the present invention has following advantage compared with existing burner hearth form measuring method:
(1) burner hearth shape is predicted based on the BP neural network that debugging is completed, it being electrolysed as a result, combining according to test
Actual motion state, the burner hearth shape of each position in aluminium cell is rapidly and accurately obtained, so that it is guaranteed that aluminium cell
Safety in production.
(2) BP neural network is trained using the simulation result of different tests parameter permutation and combination, Neng Gouti
For enough amount of training data, it is ensured that the convergence of neural network.
(3) missing data of temperature sensor and the substitution value of improper data are sought using interpolation method, it is available
More accurate inputoutput data, effectively raises precision of prediction.
Detailed description of the invention
Fig. 1 is the flow diagram of aluminium cell three-dimensional burner hearth shape real-time predicting method provided in an embodiment of the present invention；
Fig. 2 is the temperature of the part aluminium cell heat transfer finite element model determined according to the different values of different tests parameter
Distribution situation；Fig. 2 (a)~(d) be respectively according to 4 groups of parameters determine model profiling temperatures, wherein A, B, C, D, E,
F, G, H, I respectively represent different temperature values (DEG C), respectively equal to 117.362,217.084,316.806,416.528,
516.25,615.972,715.695,815.417,915.139；
Fig. 3 is electrolyzer temperature sensor installation position schematic diagram；
Fig. 4 is the schematic diagram of the real-time forecasting system of aluminium cell three-dimensional burner hearth shape provided in an embodiment of the present invention.
Specific embodiment
The present invention will be further explained below with reference to the attached drawings, but protection scope of the present invention is not limited by embodiment.
Embodiment 1
A kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based, which is characterized in that including with
Lower step:
1) certain 420kA aluminium cell modeling parameters is obtained, including each inner lining material attribute data is changed with structural parameters, the external world
Heat condition, technological parameter, as shown in table 1；
Certain the 420kA electrolytic cell key parameter of table 1
2) the modeling parameters building aluminium cell heat transfer finite element model obtained according to step 1.Meanwhile in order to cover
The most of situation occurred in production needs to calculate the burner hearth shape under the conditions of different tests convenient for being subsequently used for training.Cause
This, carries out permutation and combination for different tests parameter, constitutes test group；Each group of test parameters in test group is inputted into aluminium respectively
Electrolytic cell heat transfer finite element model, carries out simulation calculation, obtains the corresponding burner hearth shape of each group of test parameters and shell temperature
Data；
The aluminium cell heat transfer finite element model and method for solving of above-mentioned foundation, selection Xu Yujie, Lee's Jie, Yin Chenggang,
Et al. the model and method for solving that propose in " aluminium cell electric-thermal field strength Coupling method calculation method ".
The test parameters chosen in the present embodiment is current strength, mulch thickness, by current strength, mulch thickness
Different value input aluminium cells heat transfer finite element models, a part of determining aluminium cell heat transfer finite element model test group
It is as shown in Figure 2:
The permutation and combination of 1 test parameters of table (current strength, mulch thickness) is illustrated
3) BP neural network is constructed；
Include:
Determine the structure of BP neural network, including the hidden layer number of plies, each node layer number of input layer, hidden layer, output layer；
Determine neuron connection weight and neuron threshold value.
Choose the flat input variable as the BP neural network of shell temperature, electrolyte level and aluminum water；
Choose output variable of the burner hearth shape as the BP neural network；
I. the first training set and the first test set and by the shell temperature, electrolyte level and aluminium are divided horizontally into；It will
Burner hearth shape data is divided into the second training set and the second test set；
Ii. the input variable in first training set and the first test set is normalized respectively；It is right respectively
Burner hearth shape data in second training set and the second test set is normalized；
Before neural metwork training, in order to prevent because input only absolute value it is excessive due to make neuron output saturation, and then make
Weighed value adjusting enters error surface flat region, needs the data to training set and test set to be normalized, to each instruction
Practice the data x of collection and test seti, i=1,2 ..., n, using following normalization formula:
Input layer and output layer data are mapped between [0,1], this nondimensionalization side by the data after being normalized
Method solves the problems caused by dimension and numerical value difference in size.
Wherein, xi(k) gained is mapped to the data of original data area after indicating renormalization；xi' (k) indicates normalization
Data afterwards；
Iii. by the input variable after normalized in the first training set and the burner hearth shape in second training set
Shape data carry out BP neural network training as the inputoutput data of the BP neural network, until mean square error function is less than
10-3, the BP neural network convergence；
Iv. the input variable after normalized in the first test set is inputted into the BP neural network after the convergence, it is defeated
The burner hearth shape data predicted out；
V. by the burner hearth shape data in the second test set after the burner hearth shape data of the prediction and normalized
Compare, examines prediction effect；If prediction data accuracy rate is higher than 90%, enters in next step, is lower than 90% return step 4,
Re-start division training；
4) temperature sensor is arranged in aluminum cell side and bottom pot shell, as shown in figure 3, real-time measurement shell temperature number
According to；Electrolyte level and aluminium horizontal data are obtained from slot control machine simultaneously；To the shell temperature data and electrolyte water of real-time measurement
Gentle aluminium horizontal data inputs the good BP neural network of above-mentioned debugging after being normalized, the burner hearth shape finally predicted
Shape.Anti-normalization processing is carried out to the burner hearth shape data of output finally predicted, the burner hearth shape data after being predicted.
It can use following anti-normalization processing formula and model output be mapped to original data area:
xi(k)=x 'i(k)[max(xi)-min(xi)]+min(xi)
Wherein, xi(k) gained is mapped to the data of original data area after indicating renormalization；x′i(k) normalization is indicated
Data afterwards,
In the present embodiment, by the normalization and anti-normalization processing to data, available more accurate prediction data,
Effectively raise precision of prediction.
The temperature sensor selected in the present embodiment is thermocouple.Due to originals such as signal acquisition exception, network communication failures
Cause may cause certain shortage of data or certain data variations exception in temperature sensor data, such as excessive or too small.It is right
In such abnormal data, need to be handled on the basis of qualitative analysis.In the present embodiment, temperature is sought using interpolation method
Spend the substitution value of sensor missing data and improper data.
The substitution value that temperature sensor missing data and improper data are sought using interpolation method can use following
Formula carry out:
x0(i)=x0(k)+[x0(j)-x0(k)]*(i-k)/(j-k)
Wherein: x0(k)、x0It (j) is known normal data, x0It (i) is missing data or the substitution of improper data；
k<i<j；
I indicates the serial number of missing data or abnormal data in data sequence, and j, k indicate given data in data sequence
Serial number, the inside interpolation of normal given data substitutes before and after missing data or abnormal data use.
Certain time point collected shell temperature data such as following table (after interpolation processing):
Shell temperature data after 2 interpolation processing of table
Above data is inputted to the BP neural network model debugged, burner hearth shape data can be obtained, as shown in the table:
Table 3 passes through the calculated each position burner hearth thickness data of BP neural network
3 data of table explanation, the real-time predicting method and system provided through the invention can be quickly and accurately obtained aluminium electricity
Solve the real-time burner hearth thickness of each position of slot.
Claims (10)
1. a kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based, which is characterized in that including following
Step:
1) aluminium cell modeling parameters, including each inner lining material attribute data and structural parameters, extraneous heat transfer boundary condition, technique are obtained
Parameter；
2) the modeling parameters building aluminium cell heat transfer finite element model obtained according to step 1；By the difference of different tests parameter
Value carries out permutation and combination, constitutes test group；Each group of test parameters in test group is inputted aluminium cell heat transfer respectively to be had
Meta-model is limited, simulation calculation is carried out, obtains the corresponding burner hearth shape of each group of test parameters and shell temperature data；
3) BP neural network is constructed；
Choose the flat input variable as the BP neural network of shell temperature, electrolyte level and aluminum water；
Choose output variable of the burner hearth shape as the BP neural network；
4) training BP neural network, obtains the BP neural network debugged；
5) temperature sensor, real-time measurement shell temperature data are arranged in aluminum cell side and bottom pot shell；Simultaneously from slot control
Machine obtains electrolyte level and aluminium horizontal data；By the shell temperature data and electrolyte level and aluminium horizontal data of real-time measurement
Input step 4) in the BP neural network debugged, the burner hearth shape data finally predicted.
2. aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based according to claim 1, special
Sign is, the step 4) specifically includes the following steps:
I. shell temperature data are obtained from the simulation result of step 2)；Electrolyte level and aluminium number of levels are obtained from slot control machine
According to；And all shell temperature data, electrolyte level and aluminium horizontal data are divided into the first training set and the first test set；
Ii. burner hearth shape is obtained from the simulation result of step 2)；And by all burner hearth shapes be divided into the second training set and
Second test set；
Iii. by first training set shell temperature data, electrolyte level and aluminium horizontal data and it is described second instruction
Practice the burner hearth shape data concentrated and carry out BP neural network training as the inputoutput data of the BP neural network, until institute
State BP neural network convergence；
Iv. after shell temperature data, electrolyte level and the aluminium horizontal data in first test set being inputted the convergence
BP neural network, export the burner hearth shape data of prediction；
V. by the burner hearth shape data of the prediction compared with the burner hearth shape data in the second test set, prediction effect is examined；
If predictablity rate is higher than 90%, using current BP neural network as the BP neural network debugged, into next step；If
Predictablity rate is lower than 90%, then return step I, repartitions training set and test set and be trained.
3. aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based according to claim 1, special
Sign is that test parameters is one of following parameter or a variety of in the step 2): electrolyte level, aluminum water are flat, pole span is high
Degree, current strength, mulch thickness and outside air temperature.
4. aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based according to claim 1, special
Sign is that the step 3) constructs BP neural network, comprising:
Determine the structure of BP neural network, including the hidden layer number of plies, each node layer number of input layer, hidden layer, output layer；
Determine neuron connection weight and neuron threshold value.
5. aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based according to claim 2, special
Sign is, in the step 4) further include:
Respectively to shell temperature data, electrolyte level and the aluminium horizontal data in first training set and the first test set into
Row normalized；
The burner hearth shape data in second training set and the second test set is normalized respectively；
BP neural network training is carried out based on the data after normalization.
6. aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based according to claim 1, special
Sign is, in the step 5),
From being first normalized to real-time measurement shell temperature data, from slot control machine acquisition electrolyte level and aluminium horizontal data
Reason；Input step 4 again) in the BP neural network debugged, the burner hearth shape data finally predicted；Finally most to output
The burner hearth shape data predicted eventually carries out anti-normalization processing, the burner hearth shape data after being predicted.
7. aluminium cell three-dimensional burner hearth shape neural network based described according to claim 1~any one of 6 is pre- in real time
Survey method, which is characterized in that the step 5) further includes carrying out interpolation processing to the shell temperature data of temperature sensor measurement,
To seek the substitution value of missing data and improper data.
8. a kind of real-time forecasting system of aluminium cell three-dimensional burner hearth shape neural network based characterized by comprising
Parameter acquiring unit, for obtaining aluminium cell modeling parameters, including each inner lining material attribute data of aluminium cell and knot
Structure parameter, extraneous heat transfer boundary condition, technological parameter；
Simulation unit, the aluminium cell modeling parameters building aluminium cell heat transfer finite element for being obtained according to parameter acquiring unit
Model；The different values of different tests parameter are subjected to permutation and combination, constitute test group；By every battery of tests ginseng in test group
Input aluminium cell heat transfer finite element model, progress simulation calculation obtain the corresponding burner hearth shape of each group of test parameters to number respectively
Shape and shell temperature data；
Neural network construction unit, for constructing BP neural network；
Input/output variable selection unit, for choosing, shell temperature, electrolyte level and aluminum water are flat to be used as the BP nerve net
The input variable of network；Choose output variable of the burner hearth shape as the BP neural network；
First division unit, for obtaining shell temperature data from simulation result；Electrolyte level and aluminium are obtained from slot control machine
Horizontal data；And all shell temperature data, electrolyte level and aluminium horizontal data are divided into the first training set and first
Test set；
Second division unit, for obtaining burner hearth shape from simulation result；And all burner hearth shapes are divided into the second instruction
Practice collection and the second test set；
Training unit, by shell temperature data, electrolyte level and aluminium horizontal data in first training set and described
Burner hearth shape data in second training set carries out BP neural network training as the inputoutput data of the BP neural network,
Until the BP neural network restrains；
Processing unit, will be described in shell temperature data, electrolyte level and the input of aluminium horizontal data in first test set
BP neural network after convergence exports the burner hearth shape data of prediction；
Verification unit is examined pre- by the burner hearth shape data of the prediction compared with the burner hearth shape data in the second test set
Survey effect；If predictablity rate is higher than 90%, using current BP neural network as the BP neural network debugged；If prediction
Accuracy rate is lower than 90%, then controls the first division unit and the second division unit repartitions training set and test set is instructed again
Practice；
Temperature detection and processing unit, the temperature sensor including being arranged in aluminum cell side and bottom pot shell, for real-time
Measure pot shell temperature data；And interpolation processing is carried out to the shell temperature data of temperature sensor measurement, to seek missing data
And the substitution value of improper data；
Predicting unit: the BP mind that the shell temperature data and electrolyte level of real-time measurement and the input of aluminium horizontal data have been debugged
Through network, the burner hearth shape data finally predicted.
9. the real-time forecasting system of aluminium cell three-dimensional burner hearth shape neural network based according to claim 8, special
Sign is, the neural network construction unit, comprising:
First determining module, for determining the structure of BP neural network, including the hidden layer number of plies, input layer, hidden layer, output layer
Each node layer number；
Second determining module, for determining neuron connection weight and neuron threshold value.
10. the real-time forecasting system of aluminium cell three-dimensional burner hearth shape neural network based according to claim 8, special
Sign is,
First normalized module, for carrying out normalizing to shell temperature data, electrolyte level and aluminium horizontal data respectively
Change processing；
Second normalized module, for burner hearth shape data to be normalized respectively；
Anti-normalization processing module is obtained for carrying out anti-normalization processing to the burner hearth shape data of output finally predicted
Burner hearth shape data after prediction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611055395.0A CN106709149B (en) | 2016-11-25 | 2016-11-25 | A kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611055395.0A CN106709149B (en) | 2016-11-25 | 2016-11-25 | A kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106709149A CN106709149A (en) | 2017-05-24 |
| CN106709149B true CN106709149B (en) | 2019-11-19 |
Family
ID=58933935
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611055395.0A Active CN106709149B (en) | 2016-11-25 | 2016-11-25 | A kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106709149B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109283207B (en) * | 2018-11-02 | 2020-09-29 | 中南大学 | Detection device and method for simulating growth process of furnace side of aluminum electrolytic cell |
| CN109732450B (en) * | 2019-02-27 | 2021-02-26 | 重庆理工大学 | Blade polishing and grinding method based on neural network |
| CN116288532A (en) * | 2023-03-13 | 2023-06-23 | 赛富能科技(深圳)有限公司 | Method and equipment for monitoring electrolytic tank |
| CN117253568B (en) * | 2023-11-15 | 2024-01-26 | 沈阳恒泰鑫源精铸耐材有限公司 | Coating process optimization method and system for preparing yttrium oxide crucible |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010011918A2 (en) * | 2008-07-24 | 2010-01-28 | University Of Cincinnati | Methods for prognosing mechanical systems |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102517610A (en) * | 2011-12-26 | 2012-06-27 | 贵阳铝镁设计研究院有限公司 | On-line monitoring system of furnace edge shape of aluminium electrolytic tank |
| US20130304680A1 (en) * | 2012-05-10 | 2013-11-14 | Bp Exploration Operating Company Limited | Predictive corrosion coupons from data mining |
| CN106021760A (en) * | 2016-05-27 | 2016-10-12 | 中南大学 | Aluminum electrolysis cell magnetic field finite element calculation method based on ANSYS |
-
2016
- 2016-11-25 CN CN201611055395.0A patent/CN106709149B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010011918A2 (en) * | 2008-07-24 | 2010-01-28 | University Of Cincinnati | Methods for prognosing mechanical systems |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106709149A (en) | 2017-05-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106709149B (en) | A kind of aluminium cell three-dimensional burner hearth shape real-time predicting method neural network based and system | |
| CN106059496B (en) | A kind of photovoltaic module array performance monitoring and the method and system of Fault Identification | |
| CN102721479B (en) | Online monitoring method for temperature rise of outdoor electrical device | |
| CN102885392B (en) | A kind of technology for making tobacco threds Quality Monitoring Control System and method | |
| CN105137358B (en) | The SOC/SOH Forecasting Methodologies of power battery based on big data self-study mechanism | |
| CN106352542A (en) | Prediction method of remaining bath time of water storage type electric water heater | |
| CN106199305A (en) | Underground coal mine electric power system dry-type transformer insulation health state evaluation method | |
| CN107167497A (en) | A kind of equipment fault detection method and system | |
| CN105320123B (en) | A kind of measuring and controlling equipment health status evaluation system and method based on online monitoring data | |
| CN103954842B (en) | Method for determining grounding resistance value of large-scale grounding system | |
| CN116595327B (en) | Sluice deformation monitoring data preprocessing system and method | |
| CN105400915A (en) | Method and system for quantitatively evaluating distribution of gas flows on blast furnace top | |
| CN110388964A (en) | A kind of methods, devices and systems of tunnel cable data acquisition | |
| CN106119897A (en) | A kind of electrolyzer electric flow distribution and safety monitoring system and electrolytic aluminium equipment | |
| CN109945977B (en) | Thermal fault diagnosis method and system for dry-type transformer | |
| CN109358306A (en) | One kind being based on the intelligent electric energy meter health degree trend forecasting method of GM (1,1) | |
| CN108459269A (en) | A kind of 10kV pvs (pole-mounted vacuum switch)s state on-line evaluation method and apparatus | |
| CN104914854B (en) | Industrial process fault diagnosis method based on KPCA | |
| CN104820166A (en) | Power cable multi-state comprehensive online monitoring system and method for calculating cable core temperature | |
| CN105894706B (en) | A kind of forest fire prediction technique and its system | |
| CN106325258B (en) | Relay protection device state evaluation method based on online monitoring information | |
| CN117421998A (en) | Multi-mode data-based power transmission overhead line health state evaluation system | |
| CN108536968A (en) | A kind of regulation and control method of stress and deformation in welding process | |
| CN105514843B (en) | A kind of 750kV substation secondary device repair methods based on Monitoring Data | |
| CN108286060A (en) | Cathode current abnormity detection method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |