CN106227699B - A kind of blast furnace throat cross temperature measurer center band temperature predicting method and system - Google Patents
A kind of blast furnace throat cross temperature measurer center band temperature predicting method and system Download PDFInfo
- Publication number
- CN106227699B CN106227699B CN201610518985.6A CN201610518985A CN106227699B CN 106227699 B CN106227699 B CN 106227699B CN 201610518985 A CN201610518985 A CN 201610518985A CN 106227699 B CN106227699 B CN 106227699B
- Authority
- CN
- China
- Prior art keywords
- temperature
- center band
- blast furnace
- prediction model
- furnace throat
- 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
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/15—Correlation function computation including computation of convolution operations
-
- C—CHEMISTRY; METALLURGY
- C21—METALLURGY OF IRON
- C21B—MANUFACTURE OF IRON OR STEEL
- C21B7/00—Blast furnaces
- C21B7/24—Test rods or other checking devices
Abstract
A kind of blast furnace throat cross temperature measurer center band temperature predicting method of present invention offer and system, this method include:The process variable of acquisition and the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature;The process variable of acquisition is pre-processed, noise spike saltus step data and high frequency noise data are removed;Blast furnace throat cross temperature measurer center band temperature prediction is carried out using center band temperature prediction model.The system, including:Acquisition module, preprocessing module, prediction module.The present invention predicts that the temperature of following cross temperature measurer center band exports by the historical data of input and output selected by center band temperature prediction model, the relationship between temperature output and control input can accurately be expressed, it can ensure to repair or damage in cross temperature measurer center temperature sensor, so that blast furnace operating personnel accurate judgement is adjusted blast furnace top and the bottom systems in time and foundation is provided, and then realizes the stabilization of blast furnace, efficient, safe direct motion.
Description
Technical field
The invention belongs to blast furnace temperature monitoring technical field, specifically a kind of blast furnace throat cross temperature measurer center band temperature
Spend prediction technique and system.
Background technology
Steel uses most important basic raw material and the most functional material of annual output as human society, extensive
It applies in all trades and professions such as transportation, machinery manufacturing industry, building and military developments on ground.Blast furnace ironmaking is as steel and iron industry
Important procedure, ensure blast furnace ironmaking efficiently, safety and stability direct motion to the sustainable and healthy development of steel and iron industry and reduce the energy
Consumption suffers from important function.It is real since blast furnace process is the process of the physical-chemical reaction of a many kinds of substance complex shape
Now all the time metallurgical to its automation control and unsolved subject problem of automation field, especially in blast furnace exception
Under the working of a furnace, accurately prediction and effective control are carried out to furnace temperature, and realize the intelligent automation control of blast furnace ironmaking process,
Even more current field of metallurgy and the advanced subject for automatically controlling development in science and technology.
Blast furnace temperature includes mainly three aspects:Furnace throat temperature, furnace wall temperature, molten iron temperature in stove.Currently, domestic big portion
The monitoring means for the blast furnace throat temperature divided is from stock gas CO2Sampling analysis is changed into cross temperature measurer temperature curve
Analysis.Since stock gas sample time is long, CO2Constituent analysis time lag is big, and furnace temperature analysis easily error occurs;And cross temperature temperature
Write music line and stock gas CO2Composition profiles have good correspondence:The high ground of furnace throat temperature that cross temperature curve is shown
Square Gas Flow is vigorous, CO2Content is low.And cross temperature measurer has the following advantages:Sample frequency is big, data volume is big, to stove
Condition reacting condition is sensitive, can provide foundation as the adjustment of the timely top and the bottom system of blast furnace operating person.Utilize cross temperature song
Line instructs the adjustment of blast furnace indices, especially when conditions of blast furnace fluctuates abnormal, judges top and the bottom system for blast furnace operating person
The adjustment direction of degree provides foundation, plays positive effect to the fast quick-recovery of conditions of blast furnace, is the effective inspection for avoiding furnace condition disorder
Survey means.
However, blast furnace cross temperature measuring equipment center band temperature, compared with other temperature measuring points height, sensor is easily damaged and replaces week
Phase is long, thus can not monitor gas temperature in stove in time, judges that gas flow distribution affects to blast furnace operating personnel, leads to not
The correctly blast furnaces operating duty such as adjustment cloth, air blast, and then the direct motion of influence blast furnace in time.On the one hand, due to blast furnace complexity
The characteristics of multivariable, multiple time delay, it is very difficult to establish accurate cross temperature point Temperature Mechanism model.On the other hand, base
In the development of the modeling method of data-driven, make it possible accurately to predict cross temperature central temperature.
In view of operation and control amount during blast furnace ironmaking influence of the variation to blast furnace temperature not only have timeliness but also
When having the characteristics that time lag, therefore carrying out the PREDICTIVE CONTROL of blast furnace temperature, also answered under the premise of keeping current furnace condition anterograde
Take into account the benign development of next stove working of a furnace.
Invention content
In view of the deficienciess of the prior art, a kind of blast furnace throat cross temperature measurer center band temperature of present invention offer is pre-
Survey method and system.
The technical scheme is that:
A kind of blast furnace throat cross temperature measurer center band temperature predicting method, including:
The process variable of acquisition and the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature;
The process variable of acquisition is pre-processed, noise spike saltus step data and high frequency noise data are removed;
Blast furnace throat cross temperature measurer center band temperature prediction is carried out using center band temperature prediction model.
It is described to carry out blast furnace throat cross temperature measurer center band temperature prediction, packet using center band temperature prediction model
It includes:
ARMAX is selected to establish center band temperature prediction model;
Determine two polynomial orders of backward shift operator in center band temperature prediction model;
Cross temperature measurer center band temperature prediction model parameter is recognized with recurrent least square method, i.e., moves and calculates after two
Polynomial coefficient in submultinomial matrix, and then obtain final cross temperature measurer center band temperature prediction model;
Blast furnace throat cross temperature measurer center band temperature prediction is carried out using final center band temperature prediction model.
The selection ARMAX establishes center band temperature prediction model, including:
Acquire history blast furnace throat temperature process variable, i.e., each measurement point temperature of blast furnace throat cross temperature measurer and
The top temperature of blast furnace four direction;
Choose the process variable conduct with the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature
The input variable of center band temperature prediction model;
The input variable of center band temperature prediction model is pre-processed, noise spike saltus step data and high-frequency noise are removed
Data;
ARMAX is selected to establish center band temperature prediction model, the rank of two backward shift operator polynomial matrix in the model
It is secondary respectively indicate input variable lag order and output variable lag order, describe input variable and output variable between
Time lag relationship between time lag relationship, the output variable of different moments, two backward shift operator polynomial matrix in the model
Coefficient describes the functional relation between input variable and output variable.
The process variable of the selection and the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature
Centered on the input variable with temperature prediction model, including:
Using factor-analysis approach from the process variable of the history blast furnace throat temperature master of the core out with temperature
The factor;
Pearson correlation analyses are done using main gene and process variable, tentatively choose input variable;
The input variable tentatively chosen and center band temperature are done into Pearson correlation analysis, rejected and center band temperature
Incoherent input variable, the input variable for the center band temperature prediction model finally chosen.
Two polynomial orders of backward shift operator in the determining center band temperature prediction model, including:
According to the polynomial order various combination of two backward shift operators, calculates each order and combine corresponding AIC values;
The codomain of AIC values is divided into N number of node;
Selection is combined with the immediate order of AIC values of each node, and calculating separately cross under the combination of different orders surveys
The goodness of fit of warm device center band temperature prediction model selects the highest order combination of the goodness of fit and is used as two backward shift operators
Polynomial order.
A kind of furnace throat cross temperature measurer center band temperature prediction system, including:
Acquisition module, for acquiring and the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature
Process variable;
Preprocessing module is pre-processed for the process variable to acquisition, removes noise spike saltus step data and high frequency is made an uproar
Sound data;
It is pre- to carry out blast furnace throat cross temperature measurer center band temperature using center band temperature prediction model for prediction module
It surveys.
The prediction module, including:
Model building module, for selecting ARMAX to establish center band temperature prediction model;
Order determining module, for determining two polynomial orders of backward shift operator in center band temperature prediction model;
Parameter identification module, for recognizing cross temperature measurer center band temperature prediction model with recurrent least square method
Parameter, i.e. polynomial coefficient in two backward shift operator polynomial matrix, and then obtain final cross temperature measurer center band
Temperature prediction model;
Temperature prediction module, for carrying out blast furnace throat cross temperature measurer using final center band temperature prediction model
Center band temperature prediction.
The model building module, including:
Process variable acquisition module, the process variable for acquiring history blast furnace throat temperature, i.e. blast furnace throat cross are surveyed
Each measurement point temperature of warm device and the top temperature of blast furnace four direction;
Input variable chooses module, for choosing and the relevant State of Blast Furnace of blast furnace throat cross temperature measurer center band temperature
Input variable with temperature prediction model centered on the process variable of larynx temperature;
Input variable preprocessing module is pre-processed for the input variable to center band temperature prediction model, and removal is made an uproar
Sound spike saltus step data and high frequency noise data;
Prediction model establishes module, for selecting ARMAX to establish center band temperature prediction model, in the model after two
Move the lag order that Operator Polynomial order of matrix time indicates the lag order and output variable of input variable respectively, description output
The time lag relationship between time lag relationship, the output variable of different moments between variable and input variable, two in the model
The coefficient of backward shift operator polynomial matrix describes the functional relation between input variable and output variable.
The input variable chooses module, including:
Main gene chooses module, for using factor-analysis approach from the process variable of the history blast furnace throat temperature
Main gene of the core out with temperature;
First chooses module, and for doing Pearson correlation analyses using main gene and process variable, preliminary selection input becomes
Amount;
Second chooses module, for the input variable tentatively chosen and center band temperature to be done Pearson correlation analysis,
It rejects and the incoherent input variable of center band temperature, the input variable for the center band temperature prediction model finally chosen.
The order determining module, including:
AIC value computing modules, for according to the polynomial order various combination of two backward shift operators, calculating each order group
Close corresponding AIC values;
Node division module, for the codomain of AIC values to be divided into N number of node;
Order combines determining module, for selecting to combine with the immediate order of AIC values of each node, in different orders
The highest order of the goodness of fit is selected in the goodness of fit that cross temperature measurer center band temperature prediction model is calculated separately under combination
Combination is used as two polynomial orders of backward shift operator.
Advantageous effect:
High in order to solve cross temperature measurer central temperature point, sensor is easily damaged and the replacing sensor period is long, with
And when sensor degradation lead to blast furnace operating person can not judge stock gas flow distribution in time, the present invention is based on it is discrete when
Between sequence dynamic modeling thought, the auto regressive moving average modeling algorithm of proposition time series control input under the action of
The center band temperature prediction model for establishing multi output carries out On-line Estimation to five, cross temperature measurer center temperature measuring point temperature.
Ensure when cross temperature central temperature point can not be measured normally, operating personnel can estimate stove according to the temperature prediction value of model
Interior gas fluid distrbution adjusts burden distribution matrix in time, ensures blast furnace stable smooth operation.
The present invention is based on M-ARMAX intelligent modeling methods, blast furnace technology mechanism, it is proposed that a kind of blast furnace throat cross temperature
Device center band temperature predicting method and system.The present invention is gone through by input selected by center band temperature prediction model and output
History data predict the temperature output of following cross temperature measurer center band, have method simple, and built center band temperature
Prediction model can accurately express the relationship between temperature output and control input, and center band temperature prediction model precision is high, energy
It is enough to ensure to repair or damage in cross temperature measurer center temperature sensor, provide accurate furnace throat temperature for blast furnace operating personnel
Data enable blast furnace operating personnel accurate judgement to adjust blast furnace top and the bottom system in time and provide foundation, and then realize blast furnace
Stable, efficient, safe direct motion.
Description of the drawings
Fig. 1 is blast furnace throat cross temperature measurer schematic diagram in the prior art;
Fig. 2 is blast furnace throat cross temperature measurer center band temperature predicting method flow in the specific embodiment of the invention
Figure;
Fig. 3 is step 3 flow chart in the specific embodiment of the invention;
Fig. 4 is step 3-1 flow charts in the specific embodiment of the invention;
Fig. 5 is step 3-1-2 flow charts in the specific embodiment of the invention;
Fig. 6 is step 3-2 flow charts in the specific embodiment of the invention;
Fig. 7 is step 3-2-1 flow charts in the specific embodiment of the invention;
Fig. 8 is step 3-2-3 flow charts in the specific embodiment of the invention;
Fig. 9 is step 3-3 flow charts in the specific embodiment of the invention;
Figure 10 is stove furnace throat cross temperature measurer center band temperature prediction system block diagram in the specific embodiment of the invention;
Figure 11 is prediction module block diagram in the specific embodiment of the invention;
Figure 12 is model building module block diagram in the specific embodiment of the invention;
Figure 13 is that input variable chooses module frame chart in the specific embodiment of the invention;
Figure 14 is order determining module block diagram in the specific embodiment of the invention;
Figure 15 is the cross temperature central temperature prediction effect figure based on MARMAX, wherein (a)~(c) is respectively temperature measuring point
5,6,16 design sketch for being shown as 1000 sampled points, (d)~(e) is respectively the effect that temperature measuring point 15,17 is shown as 300 sampled points
Figure;
In Fig. 1, TT1 is the warm northeast temperature sensor in top, and TT2 is the warm southeast temperature sensor in top, and TT3 is the warm northwest temperature in top
Sensor is spent, TT4 is the southwestern temperature sensor of top temperature, and TT5 is 3 temperature sensor of cross temperature point, and TT6 is cross temperature point 8
Temperature sensor, 10 temperature sensor of TT7 cross temperatures point.
Specific implementation mode
It elaborates below in conjunction with the accompanying drawings to the specific implementation mode of the present invention.
As shown in Figure 1, blast furnace throat cross temperature measurer is mounted in a manner of just cross on blast furnace throat or sealing cover, packet
4 cross temperature rifles are included, are separately mounted to blast furnace throat northwest, southwest, the southeast, 4, northeast direction, thermoelectricity on cross temperature rifle
Even serial number arrangement is as follows:It is 1~No. 10 successively from northwest to the southeast;It is 11~No. 21 successively from southwest to northeast.Cross is surveyed
It is respectively clockwise No. 1, No. 2, No. 3, No. 4 to northeastward that warm rifle, which is numbered from southeastern direction,.Cross temperature rifle draw money on credit with it is short
Point of branch, wherein direction northwest (No. 3) is to draw money on credit, and arranges 6 temperature thermocouples altogether thereon;3 of remaining direction are short branch,
On arrange 5 temperature thermocouples.Site layout project has blast furnace throat cross temperature measurer, blast furnace roof top temperature measurement device, number
According to harvester, computer system.Data collector connects blast furnace temperature-measuring system, and connects department of computer science by communication bus
System.
Temperature element on blast furnace throat cross temperature measurer uses armouring nickel chromium-nickel silicon thermocouple, temperature-measuring range to exist
Between 0-1100 DEG C, the instantaneous highest environment temperature that can be born is less than 1200 DEG C.Involved thermoelectricity in present embodiment
Even model:WRKK-333, Φ 8mm, K Graduation Number, protection pipe material:GH3030 or 1Cr18Ni9Ti.
Blast furnace throat cross temperature measurer has reliability height, continuity for measuring charge level gas temperature in State of Blast Furnace
Well, the features such as digitlization.Blast furnace throat cross temperature measurer is analyzed instead of furnace throat gas sampling, improves the working of a furnace and the type of furnace is sentenced
Disconnected accuracy plays important directive function to rationally adjusting burden distribution system.
Since blast furnace throat cross temperature measurer center band temperature is higher by nearly 500 degree compared with other temperature measuring point temperature, blast furnace
Furnace throat cross temperature measurer center band thermocouple is easily damaged and the replacing sensor period of thermocouple is long, influences the working of a furnace and the type of furnace
Timely judgement.Present embodiment predicts 5, blast furnace throat cross temperature measurer center temperature, and model accuracy is high,
In thermocouple break or the repair and replacement of blast furnace throat cross temperature measurer center band thermometric, it can ensure cross temperature curve
It continuously displays, judges that the working of a furnace adjusts burden distribution system and provides foundation in time for blast furnace operating personnel, to ensure smooth operation of furnace.
For convenience of description, the symbol and terms that use present embodiment are defined as follows:
3 temperature u of temperature measuring point1DEG C (k),;
8 temperature u of temperature measuring point2DEG C (k),;
10 temperature u of temperature measuring point3DEG C (k),;
The warm temperature u in southeast top4DEG C (k),;
The warm temperature u in northwest top5DEG C (k),;
The warm temperature u in northeast top6DEG C (k),;
The warm temperature u in southwest top7DEG C (k),;
5 temperature y of temperature measuring point1DEG C (k),;
6 temperature y of temperature measuring point2DEG C (k),;
16 temperature y of temperature measuring point3DEG C (k),;
15 temperature y of temperature measuring point4DEG C (k),;
17 temperature y of temperature measuring point5DEG C (k),;
Present embodiment provides a kind of blast furnace throat cross temperature measurer center band temperature predicting method, as shown in Fig. 2, packet
It includes:
Step 1, acquisition and the process of the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature become
Amount;
Step 2 pre-processes the process variable of acquisition, removes noise spike saltus step data and high frequency noise data;
Step 3 carries out blast furnace throat cross temperature measurer center band temperature prediction using center band temperature prediction model.
The step 3, as shown in figure 3, including:
Step 3-1, ARMAX is selected to establish center band temperature prediction model;
The step 3-1, as shown in figure 4, including:
Step 3-1-1, the process variable of history blast furnace throat temperature, i.e. each survey of blast furnace throat cross temperature measurer are acquired
The top temperature of amount point temperature and blast furnace four direction;
Step 3-1-2, the mistake of selection and the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature
Input variable with temperature prediction model centered on Cheng Bianliang;
The step 3-1-2, as shown in figure 5, including:
Step 3-1-2-1, using factor-analysis approach from being selected in the process variable of the history blast furnace throat temperature
The main gene of central band temperature;
Step 3-1-2-2, Pearson correlation analyses are done using main gene and process variable, tentatively chooses input variable;
Step 3-1-2-3, the input variable tentatively chosen and center band temperature are done into Pearson correlation analysis, rejected
With the incoherent input variable of center band temperature, the input variable for the center band temperature prediction model finally chosen.
The output of center band temperature prediction model is the survey of five, blast furnace throat cross temperature measurer center in present embodiment
Warm spot temperature includes:y1Temperature measuring point 5, y2Temperature measuring point 6, y2Temperature measuring point 16, y4Temperature measuring point 15, y5Temperature measuring point 17.
Blast furnace is a complicated system, up to 30 kinds of the process variable of blast furnace throat temperature.If all data quilts
It uses, the over-fitting of center band temperature prediction model will necessarily be caused.Therefore, choose 5000 groups of blast furnace real data, using because
Sub-analysis method selects the main gene that can most explain five temperature measuring points in blast furnace throat cross temperature measurer center, this is used in combination
Main gene does Pearson correlation analyses with process variable.Dimension choosing and the output variable phase of incorporation model are reduced for simplified model
Closing property is more than 0.5 variable, selects following 9 input variables:Cross temperature point 3, cross temperature point 4, cross temperature point 8, ten
Word temperature measuring point 10, cross temperature point 20, the warm southeast in top, the warm northwest in top, top temperature southwest, the warm northwest in top.Again because factorial analysis is selected
The main gene gone out does not have practical significance, can not explain all information of five temperature spots in center, therefore the above selected input
It needs to do the rejecting of Pearson correlation analysis and the incoherent variable of central temperature point with five, cross temperature center temperature spot again,
Following 7 variables are finally chosen altogether as input variable:Cross temperature point 3, cross temperature point 8, cross temperature point 10, top temperature east
South, the warm northwest in top, top temperature southwest, the warm northwest in top.
Step 3-1-3, the input variable of center band temperature prediction model is pre-processed, removes noise spike saltus step data
And high frequency noise data;
Due to blast furnace external disturbance influence so that the input variable of center band temperature prediction model there may be with chance error
Difference, for this purpose, rejecting the noise spike saltus step data in input variable using noise spike filtering algorithm;Later, using mobile flat
Equal filtering method rejects the high frequency noise data that temperature value in input variable is less than 10 DEG C.
Step 3-1-4, ARMAX is selected to establish center band temperature prediction model, two backward shift operators are multinomial in the model
Formula order of matrix time indicates the lag order of the lag order and output variable of input variable, description output variable and input respectively
The time lag relationship between time lag relationship, the output variable of different moments between variable, two backward shift operators are more in the model
The coefficient of item formula matrix describes the functional relation between input variable and output variable.
In view of the temperature of five temperature spots in cross temperature center not only has dependence with input variable, and also and its own
Historical data selects ARMAX (Auto Regressive there are the presence that correlativity and blast furnace internal random interfere
Moving Average) centered on band temperature prediction model.Five temperature spots and furnace throat cross centered on output variable simultaneously
It also needs to consider the phase between output variable when also there is correlation between the temperature spot output variable of five, thermometric center, therefore modeling
Pass relationship.In addition modeling data be can real-time data collection, inside blast furnace caused by interference incorporated in blast furnace real data, institute
To define the ARMAX in present embodiment as Multi-outputAuto Regressive MovingAverage, i.e. multi output
ARMAX, establishing center band temperature prediction model is:
Y (k)=A (z-1)y(k)+B(z-1)u(k) (1)
In formula,
U (k)=s [u1(k), u2(k) ..., u7(k)]TFor input vector, y (k)=s [y1(k), y2(k), y3(k), y4(k),
yx(k)]TFor output vector.Backward shift operator polynomial matrix A (z-1) order np be output variable lag order, aij×1,
aij×2..., aij×npIndicate the function of current k moment output temperature point and the output temperature at history k-1, k-2 ..., k-np moment
Relationship.Backward shift operator polynomial matrix B (z-1) order nq be input variable lag order, bij×1, bij×2..., bij×npTable
Show the functional relation of current k moment output temperature point and the input variable temperature at history k-1, k-2 ..., k-nq moment.
Step 3-2, two polynomial orders of backward shift operator in center band temperature prediction model are determined;
The step 3-2, as shown in fig. 6, including:
Step 3-2-1, according to the polynomial order various combination of two backward shift operators, it is corresponding to calculate each order combination
AIC values;
The step 3-2-1, as shown in fig. 7, comprises:
Step 3-2-1-1:The center band temperature prediction model parameter of l groups is recognized with RLS;
Step 3-2-1-2:The order of l groups and the parameter substitution formula of identification (1) are calculated and records center band temperature is pre-
Survey model prediction outputAnd center band temperature prediction model residual epsilonl(k, θN);
Step 3-2-1-3:AIC values are calculated according to formula (2);
Since excessively high model order is easy over-fitting and model complexity, output order np and input order are enabled respectively herein
Nq changes totally 25 groups of order combinations one by one from 1 to 5, remembers l=1, and 2,3 ..., 25 be every group of number;It is calculated separately according to formula (2)
AIC values corresponding to every group of order;
In formula:Centered on d=np+nq band temperature prediction model order and, v is loss function, and N is order number of combinations,
εiFor the center band temperature prediction model residual error combined corresponding to i-th group of order.
Step 3-2-1-4:Judge whether l is less than 25, satisfaction goes to step 3-2-1-1, and otherwise continue to the next step.
Step 3-2-2, the codomain of AIC values is divided into N=10 node;It is selected from 25 AIC codomains according to formula (3)
10 decile nodal value AICnode;
Step 3-2-3, the AIC values AIC of selection and each nodenodeImmediate order combines (totally 10 groups), in difference
The goodness of fit fit for calculating center band temperature prediction model under order combination according to formula (4) respectively, it is highest to select the goodness of fit
Order combination is used as two polynomial orders of backward shift operator.
In formula, n=1,2,3 ..., 10 be the AIC node serial numbers of selection,Centered on band temperature prediction model predict it is defeated
Go out, AICmax, AICmin, AICnodeIndicate that above-mentioned 25 groups of orders combine maximum value in 25 groups of calculated AIC values, most respectively
Small value and the immediate AIC values of nodal value with 10 deciles between the minimum and maximum.
The step 3-2-3, as shown in figure 8, including:
Step 3-2-3-1:Select corresponding A ICnode10 group model orders (np, nq)nodeAnd extract step 3-2-1-2
In corresponding center band temperature prediction model model prediction output
Step 3-2-3-2:According to formula (4) andCalculate the fit value fit of this 10 groups of orders;
Step 3-2-3-3:Find out maximum value fit in 10 groups of fit valuesmax;
Step 3-2-3-4:Find out fitmaxCorresponding model order, as model Optimal order (np, nq)opt;
Step 3-2-3-5:Terminate and preserves Optimal order.
Step 3-3, center band temperature prediction model parameter is recognized with recurrent least square method, i.e. two backward shift operators are more
Item formula matrix, and then obtain final cross temperature measurer center band temperature prediction model;
It is combined based on Optimal order, center band temperature prediction model parameter is recognized with recurrence least square (RLS) technology.
Recurrent least square method is using the training data newly introduced to the center band temperature prediction model estimates of parameters root of previous moment
It is modified to obtain the center band temperature prediction model estimates of parameters at this moment according to recursive algorithm, with training data
It continually introduces, center band temperature prediction model estimates of parameters is continuously available amendment until reaching satisfied precision.It compares
Common common algorithm of least square, RLS calculation amounts and amount of storage are small, and fast convergence rate and capable of realizing is distinguished online
Know.Initial value θ (0)=0, P (0)=α I need to be given with RLS, when α is sufficiently large positive number, the corrective action of RLS is big, convergence
Speed is fast.α=10 are taken herein6。
The step 3-3, as shown in figure 9, including:
Step 3-3-1:Center band temperature prediction model initial parameter values θ (0)=0, inverse matrix initial value P (0)=10 are set6I,
I is unit matrix;
Step 3-3-2:Construct the data vector matrix at k-1 moment
θT=[θ1, θ2, θ3, θ4, θ5] (6)
yi=[-yi(k-1) … -yi(k-np)]
uj=[uj(k) … uj(k-nq)]
θi T=[Ai1, Ai2, Ai3, Ai4, Ai5, Bi1..., Bi7]
Aik=[aik1, aik2..., aiknp],
Bij=[bij0, bij2..., bijnq],
Y (k)=s [y1(k), y2(k), y3(k), y4(k), y5(k)]T
In formula, i=1,2,3,4,5, j=1,2 ..., 7, k=1,2,3,4,5, θTFor parameter matrix.
Step 3-3-3:Gain matrix K (k) is calculated according to formula (7);
Step 3-3-4:According to formula (8) calculating parameter estimate vector θ (k);
Step 3-3-5:Judge whether θ (k) meets formula (11) and shut down criterion, satisfaction goes to step 3-3-8, under otherwise continuing
One step.
Step 3-3-6:Matrix P of matrix (k) is calculated according to formula (9);
Step 3-3-7:Recursion one step k-1 → k, return to step 3-3-2 construct new data vector matrix.
Step 3-3-8:Terminate and preserve the parameter of identification.
By then center band temperature prediction model is in formula (5) and formula (6) substitution formula (1):
The recursive operation of RLS is terminated with the shutdown criterion of formula (11), θ in formulak(i) i-th of element for being parameter vector θ
In the recurrence result of kth time.σ is a certain positive number expression parameter required precision, takes σ=0.05 herein.
Order is determining and parameter identification is after the completion according to M-ARMAX temperature prediction modelsObtain center
Band temperature prediction model predicted value
Relative error δ (%) target function such as formula (12) is introduced to comment the modeling effect of center band temperature prediction model
Estimate, if meeting standard conditions δ≤5% of actual production, terminates the training of this center band temperature prediction model;If be not inconsistent
It closes, then needs to reselect the i.e. collected data of training sample set;
In formula,The predicted value of w center band temperature prediction models, yiTo export actual value, L is training sample capacity.
Step 3-4, blast furnace throat cross temperature measurer center band temperature is carried out using final center band temperature prediction model
Degree prediction.
Present embodiment also provides a kind of stove furnace throat cross temperature measurer center band temperature prediction system, as shown in Figure 10,
Including:
Acquisition module, for acquiring and the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature
Process variable;
Preprocessing module is pre-processed for the process variable to acquisition, removes noise spike saltus step data and high frequency is made an uproar
Sound data;
It is pre- to carry out blast furnace throat cross temperature measurer center band temperature using center band temperature prediction model for prediction module
It surveys.
The prediction module, as shown in figure 11, including:
Model building module, for selecting ARMAX to establish center band temperature prediction model;
Order determining module, for determining two polynomial orders of backward shift operator in center band temperature prediction model;
Parameter identification module, for recognizing cross temperature measurer center band temperature prediction model with recurrent least square method
Parameter, i.e. two backward shift operator polynomial matrix, and then obtain final cross temperature measurer center band temperature prediction model;
Temperature prediction module, for carrying out blast furnace throat cross temperature measurer using final center band temperature prediction model
Center band temperature prediction.
The model building module, as shown in figure 12, including:
Process variable acquisition module, the process variable for acquiring history blast furnace throat temperature, i.e. blast furnace throat cross are surveyed
Each measurement point temperature of warm device;
Input variable chooses module, for choosing and the relevant State of Blast Furnace of blast furnace throat cross temperature measurer center band temperature
Input variable with temperature prediction model centered on the process variable of larynx temperature;
Input variable preprocessing module is pre-processed for the input variable to center band temperature prediction model, and removal is made an uproar
Sound spike saltus step data and high frequency noise data;
Prediction model establishes module, for selecting ARMAX to establish center band temperature prediction model, in the model after two
Move the lag order that Operator Polynomial order of matrix time indicates the lag order and output variable of input variable respectively, description output
The time lag relationship between time lag relationship, the output variable of different moments between variable and input variable, two in the model
The coefficient of backward shift operator polynomial matrix describes the functional relation between input variable and output variable.
The input variable chooses module, as shown in figure 13, including:
Main gene chooses module, the process variable for selecting the history blast furnace throat temperature using factor-analysis approach
In with the relevant main gene of center band temperature;
First chooses module, and for doing Pearson correlation analyses using main gene and process variable, preliminary selection input becomes
Amount;
Second chooses module, for the input variable tentatively chosen and center band temperature to be done Pearson correlation analysis,
It rejects and the incoherent input variable of center band temperature, the input variable for the center band temperature prediction model finally chosen.
The order determining module, as shown in figure 14, including:
AIC value computing modules, for according to the polynomial order various combination of two backward shift operators, calculating each order group
Close corresponding AIC values;
Node division module, for the codomain of AIC values to be divided into N number of node;
Order combines determining module, for selecting to combine with the immediate order of AIC values of each node, in different orders
The highest order combination of the goodness of fit is selected in the goodness of fit of the lower cross temperature measurer center band temperature prediction model respectively of combination
As two polynomial orders of backward shift operator.
Due to system for blast furnace ironmaking complexity dynamic characteristic and easily influenced by raw material, in order to ensure M-ARMAX temperature
The precision of prediction model needs to carry out model using new sample data when the relative error of temperature prediction value is more than 5%
Re -training.
The present invention carries out the software realization of 5 temperature predicting methods in cross temperature measurer center using high-level language C#.It should
Software interface realizes data and shows, inquires and the functions such as prediction result is shown, can easily allow blast furnace operating personnel
Obtain its required blast furnace temperature information.In addition, OPC communication softwares should be equipped with by installing on the computer of the temperature prediction software
It is responsible for carrying out data double-way communication with slave computer and data acquisition device.
Figure 15 (a)~(e) is M-ARMAX temperature models prediction effect figure of the present invention, as can be seen that originally compared with measured value
Five point M-ARMAX temperature predictions of invention cross temperature measurer center are accurate, can closely follow the variation tendency of measured value.In addition, this
Inventive method training speed is fast, model accuracy is high;A kind of ancillary technique as cross temperature measurer, it is ensured that furnace temperature monitors
Continuity and accuracy.Judge that gas fluid distrbution and conditions of blast furnace provide reliable basis in stove for operating personnel.
Claims (8)
1. a kind of blast furnace throat cross temperature measurer center band temperature predicting method, including:
The process variable of acquisition and the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature;
The process variable of acquisition is pre-processed, noise spike saltus step data and high frequency noise data are removed;
Blast furnace throat cross temperature measurer center band temperature prediction is carried out using center band temperature prediction model;
It is characterized in that, described pre- using center band temperature prediction model progress blast furnace throat cross temperature measurer center band temperature
It surveys, including:
ARMAX is selected to establish center band temperature prediction model;
Determine two polynomial orders of backward shift operator in center band temperature prediction model;
Cross temperature measurer center band temperature prediction model parameter is recognized with recurrent least square method, i.e. two backward shift operators are more
Polynomial coefficient in item formula matrix, and then obtain final cross temperature measurer center band temperature prediction model;
Blast furnace throat cross temperature measurer center band temperature prediction is carried out using final center band temperature prediction model.
2. according to the method described in claim 1, it is characterized in that, the selection ARMAX establishes center band temperature prediction model,
Including:
Acquire the process variable of history blast furnace throat temperature, the i.e. each measurement point temperature and blast furnace of blast furnace throat cross temperature measurer
The top temperature of four direction;
Centered on the process variable of selection and the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature
Input variable with temperature prediction model;
The input variable of center band temperature prediction model is pre-processed, noise spike saltus step data and high-frequency noise number are removed
According to;
ARMAX is selected to establish center band temperature prediction model, the order point of two backward shift operator polynomial matrix in the model
Not Biao Shi input variable lag order and output variable lag order, describe input variable and output variable between time lag
Time lag relationship between relationship, the output variable of different moments, the coefficient of two backward shift operator polynomial matrix in the model
Functional relation between input variable and output variable is described.
3. according to the method described in claim 2, it is characterized in that, the selection and blast furnace throat cross temperature measurer center band
Input variable with temperature prediction model centered on the process variable of the relevant blast furnace throat temperature of temperature, including:
Using factor-analysis approach from the process variable of the history blast furnace throat temperature main gene of the core out with temperature;
Pearson correlation analyses are done using main gene and process variable, tentatively choose input variable;
The input variable tentatively chosen and center band temperature are done into Pearson correlation analysis, rejected and center band temperature not phase
The input variable of pass, the input variable for the center band temperature prediction model finally chosen.
4. according to the method described in claim 1, it is characterized in that, being moved after two in the determining center band temperature prediction model
The order of Operator Polynomial, including:
According to the polynomial order various combination of two backward shift operators, calculates each order and combine corresponding AIC values;
The codomain of AIC values is divided into N number of node;
Selection is combined with the immediate order of AIC values of each node, is distinguished under the combination of different orders in cross temperature measurer
It is polynomial as two backward shift operators to select the highest order combination of the goodness of fit for the goodness of fit of central band temperature prediction model
Order.
5. a kind of blast furnace throat cross temperature measurer center band temperature prediction system, including:
Acquisition module, for acquiring and the process of the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature
Variable;
Preprocessing module is pre-processed for the process variable to acquisition, removes noise spike saltus step data and high-frequency noise number
According to;
Prediction module carries out blast furnace throat cross temperature measurer center band temperature prediction using center band temperature prediction model;
It is characterized in that, the prediction module, including:
Model building module, for selecting ARMAX to establish center band temperature prediction model;
Order determining module, for determining two polynomial orders of backward shift operator in center band temperature prediction model;
Parameter identification module, for recurrent least square method identification cross temperature measurer center band temperature prediction model ginseng
It counts, i.e. polynomial coefficient in two backward shift operator polynomial matrix, and then obtains final cross temperature measurer center band temperature
Spend prediction model;
Temperature prediction module, for carrying out blast furnace throat cross temperature measurer center using final center band temperature prediction model
Band temperature prediction.
6. system according to claim 5, which is characterized in that the model building module, including:
Process variable acquisition module, the process variable for acquiring history blast furnace throat temperature, i.e. blast furnace throat cross temperature dress
Each measurement point temperature set;
Input variable chooses module, for choosing and the relevant blast furnace throat temperature of blast furnace throat cross temperature measurer center band temperature
Input variable with temperature prediction model centered on the process variable of degree;
Input variable preprocessing module is pre-processed for the input variable to center band temperature prediction model, removal noise point
Peak saltus step data and high frequency noise data;
Prediction model establishes module, moves and calculates after two for selecting ARMAX to establish center band temperature prediction model, in the model
Submultinomial order of matrix time indicates the lag order of the lag order and output variable of input variable respectively, describes output variable
The time lag relationship between time lag relationship, the output variable of different moments between input variable is moved after two in the model
The coefficient of Operator Polynomial matrix describes the functional relation between input variable and output variable.
7. system according to claim 6, which is characterized in that the input variable chooses module, including:
Main gene chooses module, for using factor-analysis approach select in the process variable of the history blast furnace throat temperature with
The relevant main gene of center band temperature;
First chooses module, for doing Pearson correlation analyses using main gene and process variable, tentatively chooses input variable;
Second chooses module, for the input variable tentatively chosen and center band temperature to be done Pearson correlation analysis, rejects
With the incoherent input variable of center band temperature, the input variable for the center band temperature prediction model finally chosen.
8. system according to claim 5, which is characterized in that the order determining module, including:
AIC value computing modules, for according to the polynomial order various combination of two backward shift operators, calculating each order combination pair
The AIC values answered;
Node division module, for the codomain of AIC values to be divided into N number of node;
Order combines determining module, for selecting to combine with the immediate order of AIC values of each node, is combined in different orders
The highest order combination conduct of the goodness of fit is selected in the goodness of fit of lower cross temperature measurer center band temperature prediction model respectively
Two polynomial orders of backward shift operator.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610518985.6A CN106227699B (en) | 2016-07-04 | 2016-07-04 | A kind of blast furnace throat cross temperature measurer center band temperature predicting method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610518985.6A CN106227699B (en) | 2016-07-04 | 2016-07-04 | A kind of blast furnace throat cross temperature measurer center band temperature predicting method and system |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106227699A CN106227699A (en) | 2016-12-14 |
CN106227699B true CN106227699B (en) | 2018-10-23 |
Family
ID=57519075
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610518985.6A Active CN106227699B (en) | 2016-07-04 | 2016-07-04 | A kind of blast furnace throat cross temperature measurer center band temperature predicting method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106227699B (en) |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108048608B (en) * | 2017-12-12 | 2019-07-23 | 山西太钢不锈钢股份有限公司 | A method of quantization adjusts blast furnace edge airflow |
CN111339636A (en) * | 2020-01-21 | 2020-06-26 | 北京北方华创微电子装备有限公司 | Process environment control method and system |
CN113139275B (en) * | 2021-03-22 | 2022-08-19 | 浙江大学 | Blast furnace throat temperature estimation method based on multilayer ore-coke ratio distribution model |
CN115597715A (en) * | 2022-09-21 | 2023-01-13 | 中冶南方工程技术有限公司(Cn) | Blast furnace soft cross temperature measurement method based on infrared temperature measurement |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101109950A (en) * | 2007-07-23 | 2008-01-23 | 鞍钢股份有限公司 | Blast furnace production process control information intelligence system |
CN102776303A (en) * | 2012-06-27 | 2012-11-14 | 浙江大学 | Method for estimating inner surface temperature of blast furnaces |
CN104498654A (en) * | 2014-12-29 | 2015-04-08 | 燕山大学 | Blast furnace temperature change trend determination method and device |
-
2016
- 2016-07-04 CN CN201610518985.6A patent/CN106227699B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101109950A (en) * | 2007-07-23 | 2008-01-23 | 鞍钢股份有限公司 | Blast furnace production process control information intelligence system |
CN102776303A (en) * | 2012-06-27 | 2012-11-14 | 浙江大学 | Method for estimating inner surface temperature of blast furnaces |
CN104498654A (en) * | 2014-12-29 | 2015-04-08 | 燕山大学 | Blast furnace temperature change trend determination method and device |
Also Published As
Publication number | Publication date |
---|---|
CN106227699A (en) | 2016-12-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106227699B (en) | A kind of blast furnace throat cross temperature measurer center band temperature predicting method and system | |
CN106909705B (en) | Blast furnace molten iron quality forecasting method and system | |
CN105608492B (en) | A kind of polynary molten steel quality flexible measurement method based on robust random weight neutral net | |
CN104651559B (en) | Blast furnace liquid iron quality online forecasting system and method based on multivariable online sequential extreme learning machine | |
CN110066895B (en) | Stacking-based blast furnace molten iron quality interval prediction method | |
CN109935280B (en) | Blast furnace molten iron quality prediction system and method based on ensemble learning | |
CN105886680B (en) | A kind of blast furnace ironmaking process molten iron silicon content dynamic soft measuring system and method | |
CN105821170A (en) | Soft measuring system and method for quality indexes of multielement molten iron of blast furnace | |
KR102103006B1 (en) | Method and Apparatus for Operating Optimal of Equipment based on Machine Learning Model | |
Martín et al. | Hot metal temperature prediction in blast furnace using advanced model based on fuzzy logic tools | |
CN104750902A (en) | Molten iron mass multivariant dynamic soft measurement method based on multi-output support vector regression machine | |
Wu et al. | Integrated soft sensing of coke-oven temperature | |
CN110796305A (en) | Hearth safety early warning method, system, equipment and storage medium | |
Xin et al. | Predicting the alloying element yield in a ladle furnace using principal component analysis and deep neural network | |
Liu et al. | Temporal hypergraph attention network for silicon content prediction in blast furnace | |
CN110457784A (en) | A kind of coal ash fusion temperature prediction technique based on BP network model | |
CN116757078A (en) | Method and system for measuring flow velocity of pulverized coal based on acting force | |
CN115713141A (en) | Parameter adjustment and prediction model acquisition method, device and storage medium | |
CN105385843A (en) | Hot rolled slab heating control method based on section terminal temperature | |
CN111814402B (en) | Heating furnace temperature control method | |
Zheng et al. | Soft measurement modeling based on temperature prediction of LSSVM and ARMA rotary kiln burning zone | |
Zagoskina et al. | Control of the blast furnace thermal state based on the neural network simulation | |
Da-Peng et al. | Intelligent Fault Classification and Identification of Heat Exchange Station Based on Time-Series Analysis | |
Ai-jun et al. | Fault diagnosis expert system using neural networks for roasting process | |
Song et al. | Data‐Driven Approach Using Supervised Learning for Predicting Endpoint Temperature of Molten Steel in the Electric Arc Furnace |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |