CN106435427B - Zinc-plated production air knife is apart from optimal control method - Google Patents
Zinc-plated production air knife is apart from optimal control method Download PDFInfo
- Publication number
- CN106435427B CN106435427B CN201610650208.7A CN201610650208A CN106435427B CN 106435427 B CN106435427 B CN 106435427B CN 201610650208 A CN201610650208 A CN 201610650208A CN 106435427 B CN106435427 B CN 106435427B
- Authority
- CN
- China
- Prior art keywords
- coating
- air knife
- thickness
- apart
- air pressure
- 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
-
- C—CHEMISTRY; METALLURGY
- C23—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; CHEMICAL SURFACE TREATMENT; DIFFUSION TREATMENT OF METALLIC MATERIAL; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL; INHIBITING CORROSION OF METALLIC MATERIAL OR INCRUSTATION IN GENERAL
- C23C—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; SURFACE TREATMENT OF METALLIC MATERIAL BY DIFFUSION INTO THE SURFACE, BY CHEMICAL CONVERSION OR SUBSTITUTION; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL
- C23C2/00—Hot-dipping or immersion processes for applying the coating material in the molten state without affecting the shape; Apparatus therefor
- C23C2/14—Removing excess of molten coatings; Controlling or regulating the coating thickness
- C23C2/16—Removing excess of molten coatings; Controlling or regulating the coating thickness using fluids under pressure, e.g. air knives
- C23C2/18—Removing excess of molten coatings from elongated material
- C23C2/20—Strips; Plates
-
- C—CHEMISTRY; METALLURGY
- C23—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; CHEMICAL SURFACE TREATMENT; DIFFUSION TREATMENT OF METALLIC MATERIAL; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL; INHIBITING CORROSION OF METALLIC MATERIAL OR INCRUSTATION IN GENERAL
- C23C—COATING METALLIC MATERIAL; COATING MATERIAL WITH METALLIC MATERIAL; SURFACE TREATMENT OF METALLIC MATERIAL BY DIFFUSION INTO THE SURFACE, BY CHEMICAL CONVERSION OR SUBSTITUTION; COATING BY VACUUM EVAPORATION, BY SPUTTERING, BY ION IMPLANTATION OR BY CHEMICAL VAPOUR DEPOSITION, IN GENERAL
- C23C2/00—Hot-dipping or immersion processes for applying the coating material in the molten state without affecting the shape; Apparatus therefor
- C23C2/04—Hot-dipping or immersion processes for applying the coating material in the molten state without affecting the shape; Apparatus therefor characterised by the coating material
- C23C2/06—Zinc or cadmium or alloys based thereon
Landscapes
- Chemical & Material Sciences (AREA)
- Chemical Kinetics & Catalysis (AREA)
- Engineering & Computer Science (AREA)
- Materials Engineering (AREA)
- Mechanical Engineering (AREA)
- Metallurgy (AREA)
- Organic Chemistry (AREA)
- Coating With Molten Metal (AREA)
Abstract
The present invention provides a kind of zinc-plated production air knives apart from optimal control method, front and back thickness of coating deviation caused by for overcoming production line operating mode to deviate or other disturbing factors, ensure that front and back coating is uniform, air knife distance when being in saturation critical zone to air pressure simultaneously optimizes, and ensures that the air pressure in feedback control adjusts allowance.It is input with front and back thickness of coating deviation mean value in timing statistics when front and back thickness of coating is inconsistent, by neural network, calculates the correction amount of front and back air knife distance, front and back side air knife distance is adjusted respectively;When air pressure is in saturation region, based on neural network model and current air pressure, when former air knife distance, the reserved allowance of air pressure, producing line speed, calculate rational front and back air knife distance, not only current control accuracy had been met, but also with the air pressure and air knife for adjusting allowance apart from setting value.
Description
Technical field
The present invention relates to zinc-plated production fields, more particularly to a kind of zinc-plated production air knife is apart from optimal control method.
Background technology
Into 21st century, China by big steel country march toward steel power.The industrial structure is adjusted, deep processing is developed
And high value added product, the competitiveness for improving product are the developing direction of China's steel and iron industry.
However, according to statistics, China accounts for the one third of the total metal annual output in the world because of the metal product that corrosion is scrapped every year,
Illustrate that the etching problem for solving steel has very important economic implications to China.Zincincation can effectively prevent steel
Corrosion, prolongs the service life, and the quality of coating is the Automobile Plate user in downstream or the weight that appliance plate is paid close attention to and examined using producer
Point, therefore, steelmaker control zinc coating thickness to control production cost while meeting galvanized product quality and requiring
Level proposes strict requirements.
To promote the automatization level of zinc-plated production control, currently, thickness of coating automatic control system is gradually in some enterprises
It is used, is generally used based on air pressure in industry, the feedback supplemented by air knife distance makes air knife as possible
Distance remains unchanged, and with air pressure for main control variable, calibrator measured value is that feedback quantity realizes closing for thickness of coating
Ring controls.Feedback closed loop control can make strip coating thickness approach setting value as possible, for promoting the control of Zinc Coating Thickness
Precision reduces thickness of coating fluctuation, reduces the excessive consumption of zinc, play certain effect.However, due to zinc-plated production process
Process characteristic, there are two inherent shortcomings for common feedback controling mode.
1), front and back coating offset issue.Above-mentioned feedback control scheme is to be with the thickness of coating summation of coating front and rear surfaces
Controlled variable makes the thickness of coating summation of strip front and rear surfaces track upper setting value.However, Zinc Coating Thickness control targe further includes
An other important quality index, the i.e. uniformity of strip front and rear surfaces coating, if the thickness of coating deviation of front and rear surfaces
It is excessive, cause product not meet quality standard, larger economic loss can be caused.However, in the actual production process, due to band
The factors such as steel plate type, working roller abrasion, machine error influence, even if under stable operating mode, strip front and rear surfaces thickness of coating
Also it will appear and fluctuate by a relatively large margin.And due to the mechanical zero of strip center line and front and back air knife and misaligned (i.e. front and back side gas
The knife of knife is away from measured value and is not equal to its actual range between strip center line), therefore by adjusting front and back unilateral air knife
The method that distance eliminates front and back thickness of coating deviation has prodigious difficulty.
2), pressure saturation problem.Air pressure is adjusted to be limited in the presence of adjusting zone of saturation (i.e. by gas path pipe pressure
Air pressure cannot increase regulating valve, air knife in any case more than gas path pipe pressure when air pressure is close to pipeline pressure
Pressure can not also increase), when air pressure unsaturation, thickness is plated caused by there are producing line velocity variations or other disturbing factors
Degree variation when, control system can by adjust air pressure eliminate above-mentioned disturbed belt come thickness of coating change, guarantee it is zinc-plated
The quality of product conforms to quality requirements;And air pressure, when being in saturation region, when external disturbance occurs, air pressure can not
Increase, lose regulating power, operating personnel have to by reducing line speed to ensure that air pressure there are enough controls
Allowance processed makes thickness of coating meet quality index requirement, the reduction of line speed so that the production capacity and section of upstream and downstream process
It plays all by extreme influence, while a large amount of economic damage can be brought to factory along with certain zinc coating thickness quality fluctuation
It loses.
Invention content
The technical problem to be solved by the present invention is to front and back coating deviations in zinc-plated production feedback control and air pressure to adjust
The problem of saturation.
In order to solve this technical problem, the present invention provides a kind of zinc-plated production air knives apart from optimal control method, packet
Include following steps:
S1:Forward and backward thickness of coating measured value is read, if storage region is less than, institute's Thickness Measurement by Microwave is stored in storage region, into
Enter step S4;If storage region has expired, S2 is entered step;
S2:Measured value earliest in storage region is replaced using current measured value, then, before judging in storage region,
Afterwards coating deviation mean value CW_bias whether be more than front and back drift correction threshold value;If being more than, S3 is entered step, if not exceeded,
Then enter step S4;
S3:It is calculated according to air knife distance, air pressure, line speed and thickness of coating departure real-time optimization optimal
Air knife distance, so that it is determined that forward and backward air knife is apart from knots modification GF_delta and GB_delta;Wherein, preceding thickness of coating needs change
For the departure of change into-CW_bias/2, it is CW_bias/2 that rear thickness of coating, which needs the departure changed,;Then, S5 is entered step;
S4:It apart from knots modification is 0 to enable forward and backward air knife, enters step S5;
S5:Judge whether air pressure enters zone of saturation, if into entering step S6, if not entering, entering step
Rapid S7;
S6:Increase air pressure and adjust allowance, according to air knife after former air knife distance, line speed, increase allowance
Pressure using neural computing thickness of coating predicted value, and then obtains the departure of thickness of coating;
In turn, it is calculated according to the departure real-time optimization of air knife distance, air pressure, line speed and thickness of coating
Optimal air knife distance, determines air knife apart from knots modification, and then the air knife is averagely allocated to forward and backward air knife apart from knots modification and is carried out
Change, obtains forward and backward air knife apart from knots modification, enter step S8;
S7:Forward and backward air knife is remained unchanged apart from knots modification, enters step S8;
S8:New forward and backward air knife distance setting is determined apart from knots modification according to step S3 and the S6 forward and backward air knife confirmed
Value, the air pressure for increasing allowance according to step S6 are air pressure setting value.
Optionally, in the step S6, when using neural computing thickness of coating predicted value, god is further utilized
Thickness of coating predicted value is obtained through neural network forecast module;
The neural network prediction module is input with strip speed, air knife distance, air pressure, and thickness of coating is output
Established, and its historical data based on galvanizing process is learnt, so according to the strip speed of current operation, air knife away from
Thickness of coating is predicted from the sampled value with air pressure.
Optionally, in the step S6, the thickness of coating of required compensation is:
CWm=NN (D (t), P (t), S (t))-NN (D (t), Pini,S(t))
Wherein, D (t) is the air knife distance of t moment, and P (t) is the air pressure of t moment, and S (t) is the production line of t moment
Speed;PiniIt is the air pressure increased after allowance, is also new air pressure setting value.
Optionally, in the step S3 and/or S6, optimization calculate optimal air knife apart from when, utilize real-time optimization mould
Block optimization calculates optimal air knife distance;
The real-time optimization module is minimised as target with the deviation between thickness of coating predicted value and control targe, with gas
Knife is constraint apart from traffic coverage, optimal under air pressure, line speed and thickness of coating setting value that iterative search gives
Air knife distance.
Optionally, real-time optimization algorithm used by the real-time optimization module is specifically described as:
Object function:min|CWt-CWp|
Decision variable:D
Constraints:
CWp=NN (D, Pini,Sini) (6)
CWt=NN (Dini,Pini,Sini)+CWm (7)
D∈[Dmin,Dmax] (8)
Optimizing is iterated to D according to increment type PID algorithm, wherein increasable algorithm describes:
Error=CWt-CWp (9)
P_error=error-error_1 (10)
I_error=error (11)
D_error=error-2*error_1+error_2 (12)
△ u=Kp*P_error+Ki*I_error+Kd*D_error (13)
D=D+ △ u (14)
Wherein, the initial value of D is Dini, error_1 is the previous generation values in error iterative calculation, and error_2 is error
Upper two generations value in iterative calculation.By formula (9)~formula (14) iterative calculation, finds D and make CWp infinite approachs CWt.It is excellent in real time
Change algorithm and finally returns that air knife apart from knots modification △ D, meets formula (15)
△ D=D-Dini (15)
In above formula, △ D are so that thickness of coating is changed needed for CWm in current manufacturing lines speed and air pressure
Total air knife is apart from knots modification, if this knots modification is all applied to former air knife apart from upper, i.e. D_front_delta (t)=
△D;According to formula (5), in order to ensure that thickness of coating is constant, air knife distance should remain unchanged, i.e., forward and backward knife is apart from knots modification size
It is identical, direction on the contrary, so, D_back_delta (t)=- △ D.The △ D are the deviations for needing to change for preceding thickness of coating
Amount-CW_bias/2, and-△ D are to need the departure CW_bias/2 changed to be modified for rear thickness of coating.
The invention discloses the air knives in a kind of Zinc Coating Thickness feedback control apart from optimization method, for overcoming production line work
Front and back thickness of coating deviation caused by condition offset or other disturbing factors ensures that front and back coating is uniform, while being suitable for air knife pressure
Power optimizes air knife distance when being in saturation critical zone, ensures the adjusting allowance of air pressure.When front and back thickness of coating not
When consistent, applying cladding deviation optimization module is input with front and back thickness of coating deviation mean value in timing statistics, passes through nerve net
Network calculates the correction amount of front and back air knife distance, is adjusted respectively to front and back side air knife distance;When air pressure is in saturation region
When, based on neural network model and current air pressure, when former air knife distance, the reserved allowance of air pressure, line speed, meter
Calculate it is rational front and back plays distance, not only met current control accuracy, but with the air pressure and air knife for adjusting allowance away from
From setting value.Method proposed by the invention solves in thickness of coating feedback control since operating mode offset or other disturbances cause
Front and back coating deviation, it is ensured that front and back thickness of coating is uniformly distributed, and prevents unilateral thickness of coating situation not up to standard from occurring.Together
When, it effectively avoids the air pressure of feedback regulation from entering saturation region, contributes to the raising of production line production capacity and putting down for thickness of coating
Surely.
Description of the drawings
Fig. 1 is that the system in an alternate embodiment of the present invention using zinc-plated production air knife apart from optimal control method is illustrated
Figure;
Fig. 2 is the signal of detailed process of the zinc-plated production air knife apart from optimal control method in an alternate embodiment of the present invention
Figure;
Fig. 3 is to control design sketch when thickness of coating setting value changes in an alternate embodiment of the present invention;
Fig. 4 is to control design sketch when line speed changes in an alternate embodiment of the present invention.
Specific implementation mode
Zinc-plated production air knife provided by the invention is carried out apart from optimal control method below with reference to Fig. 1 to Fig. 4 detailed
Description, be optional embodiment of the present invention, it is believed that those skilled in the art are not changing spirit of that invention and content
In the range of, can it be modified and be polished.
First, first related in the prior art problem and scheme are illustrated:
For galvanizing production because its good corrosion resistance is widely used in building, household electrical appliances, automobile and other industries, coating is too thick
The performances such as the resistance to chalking that can not only waste the raw material such as zinc ingot metal, but also the spot weldability of product, adhesion, coating can be influenced, and
Coating is too thin, influences whether that the corrosion resistance of product, product quality cannot be up to standard.Therefore plating uniformity control directly affects
To two aspects of product quality:Thickness of coating and plating uniformity.The zinc-plated main technological parameter of production process includes strip
Speed, air knife height, air knife distance and air pressure.Wherein strip speed is by annealing furnace production capacity, belt steel thickness and heating
The restriction of system, can not be separately adjustable, and air knife height is determined by strip speed, is influenced on zinc coating thickness smaller.Therefore, usually
Air knife distance the most direct and air pressure are influenced as thickness of coating control using on thickness of coating and front and back plating uniformity
The control variable of system processed.
The influence relationship of air pressure and air knife distance to thickness of coating is as follows:
(1) under various air pressures, the increase of zinc coating thickness can be caused by increasing air knife distance, especially under low pressure
Incrementss bigger.
(2) in nominal situation variation range, increasing air pressure can cause to increase the momentum of galvanized sheet, thus cause zinc
The reduction of layer thickness.
Continuous hot galvanizing production system has the characteristics that:
(1) on-line measurement of thickness of coating value needs to use calibrator, in order to ensure the accuracy measured, is surveyed using cold conditions
Measurement of the result of amount as thickness of coating, since cold conditions measures the requirement of technique, the installation site of calibrator is usually apart from gas
Farther out, causing to act on after influence of the controlled quentity controlled variable to thickness of coating at air knife needs longer time can just detect knife, i.e.,
System has larger measurement delay.
(2) galvanizing process is acted on by the physical factor of multiple complexity such as air force, boundary fluid layer, heat transfer, final
Uniformity is mainly coefficient by multiple variables such as air knife distance, air knife air pressure, strip speeds before and after zinc coating thickness and coating
Influence, and these variables are complex to the influence principle of thickness of coating and coating horizontal homogeneity, have it is stronger non-linear,
The use of traditional modelling by mechanism method is difficult to establish accurate system model;
(3) zinc-plated production process disturbing factor is numerous, such as air knife height, die gap, the temperature of strip, plate thickness,
Plate is wide, surface roughness, produces line tension, the factors such as the temperature and chemical composition of zinc pot, belt plate shape and strip shake
Variation will all have an impact thickness of coating and front and back uniformity.
In conclusion zinc-plated is typical time-varying large time delay, non-linear, strong disturbance a production process, control difficulty
Greatly.Most domestic iron and steel enterprise depends on the experience of operative employee and combines bottom circuit using manual hand manipulation at present
The mode of PID is controlled, and control accuracy is low, zinc consumption is big, quality fluctuation is big, or even cannot be guaranteed thickness of coating and coating
The quality requirement of the uniformity.Therefore, thickness of coating automatic control system is also gradually used in some enterprises, for
It improves Zinc Coating Thickness control accuracy and automatization level tool has certain effect.It is analyzed from principle, at present Zinc Coating Thickness system
It generally uses based on air pressure, the feedback supplemented by air knife distance, wherein air knife distance refers to front and rear sides
The distance between air knife (i.e. total air knife distance).However, due to controlling difficulties existing for zinc-plated production process, current feedback
Control method generally faces following two problems:
1), front and back coating offset issue
Current thickness of coating feedback control system with total air knife distance variable in order to control, without consider forward and backward air knife away from
From specific distribution.And under the premise of total air knife distance is certain, although distribution and the coating total thickness of front and back unilateral side air knife distance
Degree association is little, but directly affects the thickness of front and back unilateral coating.The pressure value of usually front and back side air knife is identical, therefore only
Have front and back side air knife apart from it is consistent when, front and back thickness of coating can be made consistent, this requires by strip control in air knife
Near heart line, while being also to ensure safety in production, prevents air knife from scraping the needs of strip.However, due to mechanical reason, in air knife
Heart line can not be completely superposed with strip center line, and there are certain errors, simultaneously as there are measurement error, in actual production
Strip central perpendicularity is there are certain deviation, and cooling tower top roller, sinking roller and stabilizing roller carry out that there are one with production
The abrasion for determining degree causes the departure to change at any time.In addition, the variation with steel plate type also results in it relative to front and back
The air knife distance of both sides shifts.A series of this disturbance factor causes during continuous hot galvanizing, even if total air knife distance
It keeps constant, actual front and back side tool makes strip front and rear surfaces thickness of coating fluctuate away from being still dynamic change.
And the influence principle due to the control variable such as air knife distance, air pressure to coating horizontal homogeneity is complex, has relatively strong
It is non-linear, therefore to the consistency of strip front and rear surfaces thickness of coating control bring difficulty.If the coating of front and rear surfaces
Thickness deviation is excessive, and product is caused not meet quality standard, can cause larger economic loss.Meanwhile zinc-plated production system is deposited
In prodigious measurement delay, if inadequate to the accuracy of front and back zinc coat thickness control, front and back thickness of coating can be caused to adjust back
It is whole, it vibrates.
2), pressure saturation problem
When air pressure is in saturation, air pressure can not improve again, if air knife distance does not change, thickness is plated to meet
Demand for control is spent, then line speed can not be further added by.Therefore air pressure saturation has effect of contraction to production line production capacity.Air pressure
When unsaturated, when the thickness of coating caused by there are line speed variation or other disturbing factors changes, control system can
With by adjust air pressure eliminate above-mentioned disturbed belt come thickness of coating variation, ensure that the quality of galvanizing production meets quality and wants
It asks;And when it is in saturation, air pressure can not increase, lose regulating power, when external disturbance occurs, operator
Member has to by reducing line speed to ensure that air pressure has enough control allowances that thickness of coating is made to meet quality
Index request, the reduction of line speed so that the production capacity and rhythm of upstream and downstream process all receive extreme influence, give factory's band
A large amount of economic loss is carried out.
For coating offset issue, the country also has related patents and document to propose System design based on model method, such as public
The number of opening is the patent of CN103510032.A《The deviation control method of cold rolling hot dip galvanizing plating uniformity》, using knife lip aperture
Influence model to thickness of coating, introducing influences efficiency function matrix, improves plating uniformity control accuracy, is added to simultaneously
This function and filter function item improve coating error function precision, to improve plating uniformity computational accuracy, make zinc-plated
Product surface quality is improved.However, the above method lays particular emphasis on improvement transverse uniformity of plating, it is not related to that operating mode can be overcome
The offset situation inconsistent with front and back coating caused by belt steel thickness.
In order to solve the problems, such as that air pressure is saturated, document《The closed-loop control brief introduction of hot galvanizing line zinc coating thickness》(bang
Dragon, Sichuan is metallurgical, and 2006:28 (6)), the zinc coat thickness control system of VAI companies offer, the system feedback controller are provided
It is made of two functional blocks of control block and saturate block, calls control block first, according to the deviation of thickness of coating, be based on air pressure
For the yield value of thickness of coating, air pressure is calculated and adjusts knots modification, then according to the air pressure knots modification and bottom
The knots modification that the gain of layer air pressure closed-loop control system calculates valve opening is added in existing valve opening value, if its
Value was less than for 80% (regulating power of valve enters zone of saturation), then illustrated that air pressure does not enter saturation, then need not change
Air knife distance value.Conversely, then entering saturate block, set the regulated quantity of air pressure to valve opening is made not enter saturation at this time
The upper limit, and calculate in the case that the air pressure is remaining need to be by the deviation of the thickness of coating of air knife compensated distance, so
Air knife is calculated apart from adjustment amount to the yield value of thickness of coating according to air knife distance afterwards.This regulative mode is based on linear approximation
For Calculation of Sensitivity air knife apart from adjustment amount, precision is relatively low, which is first adjusted to saturation state by air pressure, reuses gas
Knife distance compensates adjusting according to residual deviation, and air pressure will be caused still in saturation critical condition, lose the abundant of adjusting
Amount.
As it can be seen that although it is contemplated that the problem of plating uniformity in existing thickness of coating feedback control system, but lack and solve
The mechanism of front and back thickness of coating dynamic deviation easy tos produce unilateral thickness situation not up to standard;And simultaneously, although existing system is adopted
Air pressure saturated conditions is compensated with air knife distance, but air knife compensated distance gauge is not calculated and carries out optimizing, is caused
Air pressure after compensation still lacks the allowance of adjusting, when external interference is persistently aggravated, needs through constantly regulate gas
Knife distance eliminates the influence of external disturbance, however, knife is by step-length discrete adjustment away from being adjusted, with continuously adjustable
Air pressure is compared, and the precision of adjusting is poor, and it is poor more than overshoot or larger adjusting to be easy to cause.
It revert to from the point of view of technical scheme of the present invention:
In contrast, the purpose of the present invention is be directed to front and back coating deviation existing for existing thickness of coating feedback control system
The problems such as quality of coating fluctuation caused by problem and air pressure saturation and limited production capacity, it is proposed that one kind being based on non-linear god
Air knife in Zinc Coating Thickness feedback control through network model is apart from optimization method.First, using nerual network technique, with strip
Speed, air knife distance, air pressure are input, and thickness of coating is output, establishes zinc-plated production thickness of coating neural network prediction
Model;When thickness of coating is inconsistent before and after appearance, using front and back coating deviation optimization module, with front and back coating in timing statistics
Thickness deviation mean value is input, by neural network, calculates the correction amount of front and back air knife distance, respectively to front and back air knife distance into
Row adjustment;When air pressure is in saturation region, air pressure is set first and adjusts space, then by neural network model,
The front and back side air knife distance of adjustment.The present invention can effectively solve in zinc-plated production process due to operating mode offset, belt steel thickness variation or
The inconsistent situation of front and back thickness of coating caused by other reasons, also can effectively avoid pressure work saturation workspace feelings
Condition ensures to adjust air pressure allowance, significantly improves product quality and front and back coating uniformity, reduces thickness of coating quality fluctuation,
It remains to ensure that product quality meets index request and other effects in the case of promoting production line production capacity.
The present invention proposes the air knife in a kind of galvanization production line thickness of coating feedback control apart from optimization method.It can be fast
Speed reply operating mode offset or pressure work are rapid to optimize air knife distance at saturation region, eliminate operating mode offset and belt steel thickness becomes
Front and back thickness of coating deviation, ensures the adjustment allowance of air pressure caused by changing, and ensures thickness of coating longitudinal uniformity, is carrying
It remains to ensure that product quality meets index request in the case of rising production line production capacity, reduces the generation of substandard product.
Before expansion illustrates this method, under first illustrating, the system for being applicable in this method can be understood as including neural network
Four prediction module, real-time optimization module, front and back coating deviation optimization module, saturated pressure optimization module parts form.
Wherein, neural network prediction module be one with strip speed, air knife distance, air pressure be input, plate thickness
The neural network prediction model that degree is established for output, historical data of the model based on galvanizing process are learnt, can online root
Accurate prediction is made to thickness of coating according to the sampled value of current operation operating mode (strip speed, air knife distance, air pressure), is carried
Influence of the uncontrolled variable to thickness of coating can be surveyed in preceding precognition galvanizing process, the adjustment of variable in order to control provides foundation.Nerve
Network module is the basis of real-time optimization module.
Wherein, the real-time optimization module, the module is between thickness of coating predicted value and zinc coat thickness control target
Deviation be minimised as optimization aim, apart from traffic coverage be constraint with air knife, iterative search gives operating mode (air pressure, life
Producing line speed, thickness of coating setting value) under optimal air knife distance.Specifically, the solution of this module is an optimization problem,
Deviation of the object function between thickness of coating predicted value and control targe, and decision variable is air knife distance, constraints is
Mapping relations between the line speed of neural network prediction module, air knife distance, air pressure and thickness of coating, Yi Jiqi
The technological procedure of knife distance constrains.By, apart from optimizing, being eliminated inclined between thickness of coating predicted value and control targe to air knife
Difference.Real-time optimization module is the basis of front and back coating deviation optimization module and saturated pressure saturation optimization module.
Wherein, the front and back coating deviation optimization module is with the target in order to control of thickness of coating uniformity before and after strip,
Operating mode shift and cause front and back air knife apart from it is inconsistent when, under the premise of ensureing that total thickness of coating is up to standard, before and after optimization
Side air knife distance, with coating deviation before and after elimination.Specifically, before and after appearance when thickness of coating deviation, in order to ensure always to plate
Layer thickness does not change, and to make front and back thickness of coating knots modification in the same size, and direction is on the contrary, correspond, in other works
In the case of condition is immovable, need to ensure that total air knife distance does not change, and front and back side air knife distance change amount is in the same size,
Direction is opposite.Current working (line speed, air knife distance, air pressure) and the thickness of coating input that need to change is real-time
Optimization module, you can obtain the knots modification of front and back air knife distance respectively.
Wherein, the saturated pressure optimization module is to optimize air knife distance when air pressure is operated in saturation region, with
Compensate the adjusting space of air pressure.When air pressure is in saturation, to reduce air pressure and ensureing that coating is constant,
Under the premise of ensureing production safety, then it can reduce air knife distance.Wherein, the knots modification of air knife distance according to line speed,
The thickness of coating that revised air pressure, line speed and needs change, as obtained by real-time optimization module calculating.
Based on four above-mentioned modules, in each control cycle, system operation is shown in steps are as follows:
Without loss of generality, front and back air knife range correction is GF_delta, GB_delta, and the air knife distance of saturated pressure is mended
The amount of repaying is delta_gap, and front and back air knife is GF_SV, GB_SV apart from setting value, this control system deposits comprising a front and back coating
Storage area domain, stores the historical measurements of front and back thickness of coating, and CW_bias is that preceding thickness of coating plates thickness after subtracting in the memory block
The mean value of degree.
S1, front and back thickness of coating measurement value information is read from first-level machine, whether first determine whether front and back coating storage region
It has been expired that, if less than, thickness of coating is preserved and enters the storage region, into S4, otherwise, into S2.It also can be described as:It reads
Institute's Thickness Measurement by Microwave is stored in storage region, enters step S4 by forward and backward thickness of coating measured value if storage region is less than;If storage
Region has been expired, then enters step S2;
S2. use the principle of first in first out that current thickness of coating measured value is saved into front and back coating storage region, and
Override thickness of coating measured value earliest in former storage region.Judge front and back coating deviation mean value CW_ in the storage region
Whether bias is more than the threshold value of front and back drift correction, if then entering S3, otherwise enters S4.It also can be described as:Using current
Measured value replace earliest measured value in storage region and then judge forward and backward coating deviation mean value CW_ in storage region
Bias whether be more than front and back drift correction threshold value;If being more than, S3 is entered step, if not exceeded, then entering step S4;
S3:To ensure that thickness of coating is constant, air knife distance should remain unchanged, i.e., front and back air knife is apart from knots modification size phase
Together, direction is opposite, that is to say, that preceding thickness of coating needs change-CW_bias/2, rear thickness of coating to need to change CW_bias/2.
Air knife distance, air pressure, line speed and the thickness of coating departure that need to change are inputted real-time optimization module respectively, point
Front and back air knife is not calculated apart from knots modification GF_delta and GB_delta, into S4.Wherein, real-time optimization module is according to gas
Knife distance, air pressure and line speed calculate thickness of coating predicted value, then control targe is that thickness of coating predicted value adds
On the thickness of coating amount that need to change, control deviation is the thickness of coating amount that need to change.Air knife distance is carried out according to increment type PID
Optimizing so that the thickness of coating value that revised front and back side air knife distance is predicted with air pressure, line speed meets coating
Thickness control desired value.Then front and back side air knife is that revised air knife distance subtracts air knife apart from original value apart from knots modification.
It can be briefly described for:It is excellent in real time according to air knife distance, air pressure, line speed and thickness of coating departure
Change calculates optimal air knife distance, so that it is determined that forward and backward air knife is apart from knots modification GF_delta and GB_delta;Wherein, preceding plating
It is-CW_bias/2 that layer thickness, which needs the departure changed, and it is CW_bias/2 that rear thickness of coating, which needs the departure changed,;So
Afterwards, S5 is entered step;
S4. GF_delta=0, GB_delta=0 are enabled, without front and back coating drift correction, even forward and backward air knife distance
Knots modification is 0, into S5.
S5:Judge whether air knife enters zone of saturation, if into zone of saturation, enters S6, if not entering into saturation region,
Then enter S7.
S6:Increase air pressure and adjust allowance, is adjusted according to when former air knife distance, air knife height, line speed and increase
The air pressure after allowance is saved, thickness of coating predicted value is calculated by neural network prediction model, calculates predicted value and coating
Deviation between thickness setting value.Increasing, the air pressure after adjusting allowance, air knife distance and line speed input are real
When optimization module, calculate air knife apart from knots modification delta_gap, to ensure that strip center line is overlapped with air knife center line, need
By the air knife at knots modification mean allocation to front and back side air knife, then GF_delta=GF_delta+delta_gap/2, GB_
Delta=GB_delta+delta_gap/2.Into S8.
It, which can be summarized, is described as:
Increase air pressure and adjust allowance, according to air knife pressure after former air knife distance, line speed, increase allowance
Power using neural computing thickness of coating predicted value, and then obtains the departure of thickness of coating;
In turn, it is calculated according to the departure real-time optimization of air knife distance, air pressure, line speed and thickness of coating
Optimal air knife distance, determines air knife apart from knots modification, and then the air knife is averagely allocated to forward and backward air knife apart from knots modification and is carried out
Change, obtains forward and backward air knife apart from knots modification, into S8;
S7:GF_delta, GB_delta are remained unchanged, i.e., forward and backward air knife is remained unchanged apart from knots modification, into S8.
S8:New front and back air knife is calculated apart from setting value GF_SV=GF_SV+GF_delta, GB_SV=GB_SV+GB_
Delta, and under reach PLC.When air pressure is saturated, PiniFor air pressure setting value, and under reach PLC.
The technique effect of above-mentioned technical proposal is analyzed as follows:
(1) present invention uses air knife apart from optimization method in feedback control, effectively overcomes because template changes or it
He disturbs the influence to front and back thickness of coating, ensure that the uniformity of strip front and rear surfaces thickness of coating in feedback control.
(2) present invention uses air knife apart from optimization method in feedback control, effectively overcomes air pressure and is operated in
Saturation region and cause coating surface quality fluctuate and adverse effect is further increased to production capacity, ensure that feedforward control
Middle strip coating quality.
(3) present invention uses real-time optimization method in feedback control, meets process requirements in each operating parameter of guarantee
Meanwhile according to current control parameter and control deviation, air knife is calculated in real time apart from knots modification, while ensureing that thickness of coating meets
Control targe, not only ensure that the surface quality of thickness of coating in feedback control, while optimize operating parameter, front and back to overcome
Coating deviation increases air pressure adjustment allowance, remains to ensure that product quality meets index and wants in the case where promoting production line production capacity
It asks.
In conclusion control system using the present invention, can accurately optimize when deviation occurs for front and back thickness of coating
Air knife distance, fast and effeciently to eliminate front and back thickness of coating deviation, to ensure that the uniformity of thickness of coating, while can
Effectively ensure that air pressure far from saturation region, increases the adjusting space of air pressure, ensure that coating surface quality, carrying
It remains to ensure that product quality meets index request in the case of rising production line production capacity.
Above-mentioned flow further will be unfolded below to be illustrated, detailed design can be regarded as the present invention
Optional or preferred scheme.It is described below to be best understood from, provides following table, to each symbol being set forth below, parameter amount
Carry out a kind of definition, certainly, below table be it is a kind of signal enumerate, respectively meet also can have other to contain above with parameter amount
Justice, still need to be to be hereafter illustrated as standard.
This control system is mainly by neural network prediction module, real-time optimization module, front and back coating deviation optimization module, full
It is formed with four parts such as air pressure optimization modules.
In figure I indicate neural network prediction module, be one with strip speed, air knife distance, air pressure be input, plating
Layer thickness is the neural network prediction model that output is established, and historical data of the model based on galvanizing process learnt, can be
Line is according to the detected value of current operation operating mode (strip speed S (t), air knife distance D (t), air pressure P (t)) to thickness of coating
Accurate prediction is made, the Nonlinear Mapping relationship of the neural network is denoted as NNp (*), expression formula is as follows:
NNp (D (t), S (t), P (t))=CWp (t)
The module is the basis of real-time optimization module operation.
II indicates real-time optimization module in figure, and the module is between thickness of coating predicted value and zinc coat thickness control target
Deviation is minimised as optimization aim, and with air knife apart from technological procedure to constrain, iterative search gives operating mode (air pressure, production
Linear velocity, thickness of coating setting value) under optimal air knife distance.Specifically, the solution of this module is an optimization problem, mesh
Deviation between scalar functions thickness of coating predicted value and control targe, and decision variable is air knife distance, constraints is nerve
Mapping relations between the line speed of neural network forecast module, air knife distance, air pressure and thickness of coating, air knife distance
Technological procedure constrains.By, apart from optimizing, eliminating the deviation between thickness of coating predicted value and control targe to air knife.It is excellent in real time
Change the basis that module is coating deviation optimization module and saturated pressure optimization module.
III indicates front and back coating deviation optimization module in figure, with the target in order to control of thickness of coating uniformity before and after strip,
Operating mode shift and cause front and back air knife apart from it is inconsistent when, under the premise of ensureing that total thickness of coating is up to standard, optimize air knife
Distance, with coating deviation before and after elimination.Specifically, before and after appearance when thickness of coating deviation, in order to ensure total thickness of coating
It does not change, to make front and back thickness of coating knots modification in the same size, direction does not change in other operating modes on the contrary, correspond
In the case of change, need to ensure that total air knife distance does not change, and front and back side air knife distance change amount is in the same size, direction phase
Instead.Operating mode at that time (line speed, air knife distance, air pressure) and the thickness of coating that need to change are inputted real-time optimization mould
Block, you can obtain the knots modification of front and back air knife distance respectively.
IV indicates saturated pressure optimization module in figure, when air pressure is operated in saturation region, is ensureing thickness of coating
Under the premise of, optimize air knife distance, to adjust the adjusting space of air pressure.Specifically, when air pressure is in saturation, give
Determine air pressure and increase adjusting allowance, is then gone out by neural computing and adjust changing for the thickness of coating caused by air pressure
Variable, then air knife distance then to compensate air pressure change caused by thickness of coating, given operating mode (increase adjusting allowance after
Air pressure, air knife distance and line speed) and air knife distance need thickness of coating put into real-time optimization module, calculate
Go out air knife range correction.
Based on above-mentioned feedforward control system framework, in each control cycle, operational flow diagram of the present invention such as Fig. 2 institutes
Show, details are as follows for each step in figure:
S101 (can couple it should be understood that step S1 above):Front and back thickness of coating is read from first-level machine measures value information,
It first determines whether front and back coating storage region has expired, if less than, thickness of coating is preserved and enters the storage region, entered
S104, otherwise, into S102.
S102 (can couple it should be understood that step S2 above):Current thickness of coating is measured using the principle of first in first out
Value is saved into front and back coating storage region, and overrides thickness of coating measured value earliest in former storage region.Judge the storage
In region front and back coating deviation mean value CW_bias whether be more than front and back drift correction threshold value, calculating formula (1):
If CW_bias>θ _ bias then enters S103, otherwise enters S104.
S103 (can couple it should be understood that step S3 above):Front and back coating deviation CW_bias is corrected, and ensures plating thickness
Spend it is constant, then according to the mapping relations formula of neural network (2) and the relational expression (3) of air knife distance and front and back air knife distance
NNp (D (t), P (t), S (t))=CWp (2)
D (t)=D_front (t)+D_back (t) (3)
It is found that when line speed S (t) and air pressure P (t) are constant, it is ensured that CWp is constant, then D (t) should be kept
Constant, i.e., the sum of front and back air knife range correction is zero, meets formula (4)
+ D_back_delta (t)=0 D_front_delta (t) (4)
Formula (5) is converted by formula (4),
D_front_delta (t)=- D_back_delta (t) (5)
I.e. front and back air knife distance change amount, equal in magnitude, direction is opposite.Since thickness of coating deviation is generally little, small
In range, it is believed that front and back air knife distance change is essentially identical to the gain of unilateral thickness of coating, and therefore, front and back thickness of coating is repaiied
Positive quantity is in the same size, direction on the contrary, and so that front and back coating deviation is become 0, it is necessary to make the coating of former air knife distance correction inclined
Residual quantity is-CW_bias/2, and the coating departure of rear air knife distance correction is CW_bias/2.
According to formula (5) it is found that front and back air knife distance change amount, equal in magnitude, therefore only need to calculate the change of former air knife distance
Change amount can be obtained the variable quantity of rear air knife distance.It regard air knife distance D (t) as D respectivelyini, air pressure P (t) be used as Pini、
Line speed S (t) is used as SiniIt is used as CWm with modified preceding thickness of coating departure-CW_bias/2 is intended to, according to real-time optimization
Algorithm calculates D_front_delta (t).
Real-time optimization algorithm describes in this way:
Object function:min|CWt-CWp|
Decision variable:D
Constraints:
CWp=NN (D, Pini,Sini) (6)
CWt=NN (Dini,Pini,Sini)+CWm (7)
D∈[Dmin,Dmax] (8)
Optimizing is iterated to D according to increment type PID algorithm, wherein increasable algorithm describes:
Error=CWt-CWp (9)
P_error=error-error_1 (10)
I_error=error (11)
D_error=error-2*error_1+error_2 (12)
△ u=Kp*P_error+Ki*I_error+Kd*D_error (13)
D=D+ △ u (14)
Wherein, the initial value of D is Dini, error_1 is the previous generation values in error iterative calculation, and error_2 is error
Upper two generations value in iterative calculation.By formula (9)~formula (14) iterative calculation, finds D and make CWp infinite approachs CWt.It is excellent in real time
Change algorithm and finally returns that air knife apart from knots modification △ D, meets formula (15)
△ D=D-Dini (15)
In above formula, △ D are so that thickness of coating is changed needed for CWm in current manufacturing lines speed and air pressure
Total air knife is apart from knots modification.Obviously, it needs exist for adjusting preceding thickness of coating, therefore this knots modification is all applied to former air knife
Apart from upper, i.e. D_front_delta (t)=△ D;According to formula (5), D_back_delta (t)=- △ D.Into S105.In order to
Ensure that thickness of coating is constant, air knife distance should remain unchanged, i.e., front and back knife is identical apart from knots modification size, direction on the contrary, so,
D_back_delta (t)=- △ D.The △ D are the departure-CW_bias/2 for needing to change for preceding thickness of coating, and-△
D is to need the departure CW_bias/2 changed to be modified for rear thickness of coating.
S104 (can couple it should be understood that step S4 above):Enable D_front_delta (t)=0, D_back_delta (t)
=0, into S105.
S105 (can couple it should be understood that step S5 above):Judge whether air pressure enters zone of saturation, if air knife pressure
Power workspace has been enter into zone of saturation, then enters S106, otherwise enters S107.
S106 (can couple it should be understood that step S6 above):Air pressure is enabled to increase Pini=P (t)+△ after adjusting allowance
P then optimizes air knife distance to compensate air pressure and changes caused thickness of coating variable quantity, then the required thickness of coating compensated
CWm is
CWm=NN (D (t), P (t), S (t))-NN (D (t), Pini,S(t)) (16)
Using D (t) as Dini, S (t) is as SiniAnd Pini, CWm put into real-time optimization module, calculate air knife distance
Knots modification D_delta.Then front and back air knife range correction is respectively:
D_front_delta (t)=D_front_delta (t)+D_delta/2 (17)
D_back_delta (t)=D_back_delta (t)+D_delta/2 (18)
Into S108.
S107 (can couple it should be understood that step S7 above):D_front_delta (t), D_back_delta (t) are kept not
Become.Into S108.
S108 (can couple it should be understood that step S8 above):Front and back air knife is calculated apart from setting value DF_SV, DB_SV
DF_SV=DF_SV+D_front_delta (t) (19)
DB_SV=DB_SV+D_back_delta (t) (20)
PLC is reached under DF_SV, DB_SV.When air pressure is saturated, PiniFor air pressure setting value, and under
Reach PLC.
In addition, per at regular intervals, new creation data, sieve caused by system online acquisition interior operation this period
Sample is selected, prediction neural network weight is carried out continuing to train, updates neural network.Specifically, using backpropagation side
Method updates the weight of neural network, and prediction neural network is enable to learn to new work information, can automatic adaptive system spy
The variation of sign.
Below by taking zinc-plated actual production process as an example, illustrate advantageous effect obtained by the present invention:
The design sketch of the front and back thickness of coating drift correction algorithm designed using the method for the present invention is listed in Fig. 3.From figure
In it can be seen that, at 1 hour or so, front and back thickness of coating deviation statistics value reached the threshold value for starting front and back air knife distance optimization
When, air knife distance before and after coating drift correction algorithm optimization so that front and back coating deviation is under control.Illustrate using the present invention
The method can effectively correct front and back thickness of coating deviation, improve product quality.
The design sketch of saturated pressure optimization designed by the invention is listed in Fig. 4.From the figure, it can be seen that for gas
Knife pressure has been operated in the case where saturated pressure edge (67KPa is saturation region), if (time coordinate 30min or so) will be carried
Line speed is risen to improve production capacity, since air pressure can not continue to be turned up air pressure to eliminate speed without allowance is adjusted at this time
Degree rises the influence (coating thickens) to thickness of coating, results in the variation of product quality and a large amount of wastes of zinc liquid.And at this
After invention the method puts into operation (time coordinate is 50min or so), front and back air knife distance is optimized, by ensureing production
Reduce front and back air knife distance in the case of safety so that the adjusting allowance of air pressure is increased, as shown, thickness of coating
It has obtained accurately controlling, has ensured that the product quality after production capacity is promoted.In the above process, under traditional control method,
The increase of galvanization production line speed can cause thickness of coating to deviate Product Process index, and quality of coating is unable to get guarantee, i.e. gas
Pressure long-term work hinders the promotion of production line production capacity in zone of saturation, after the method for the invention, passes through air knife distance
Optimal control, it is ensured that the control allowance of air pressure, remained in the case where promoting production line production capacity ensure product quality
Meet index request.
To sum up, difference with the prior art of the present invention is:
(1), the present invention uses neural network model.Compared with traditional regression model, neural network has very strong non-
Linear fit ability can more accurately reflect the nonlinear characteristic of zinc-plated production process, use linear model to plating with existing
The method that zinc production process is controlled is compared, and neural network model can carry out the full working scope of zinc-plated production high-precision pre-
It surveys, it is ensured that of the invention that all there is good control accuracy and quality under each operating condition of zinc-plated production.
(2), tradition feedback PID needs are repeatedly controlled according to feedback quantity deviation, due to the measurement of zinc-plated production system
Large time delay so that the adjustment process of front and back thickness of coating deviation is very slow, and control effect is poor.The present invention with thickness of coating with
Deviation between control targe is target, is constraint with the technological procedure of air knife distance, using neural network prediction model to gas
Knife distance is iterated optimizing, can disposably eliminate the deviation between thickness of coating and control targe, can be fast and accurately
Ground ensures thickness of coating uniformity.
(3), front and back coating deviation optimization method of the present invention is traditional based on air pressure, total air knife
What the feedback control system supplemented by distance did not had.Specifically, the present invention is considered due to operating mode offset or other interference
Before and after strip caused by factor the problem of coating deviation, using air knife before and after optimal control side air knife distance, it is ensured that
The consistency of strip front and rear surfaces thickness of coating improves Control platform.
(4), saturated pressure optimization method of the present invention is that traditional feedback does not have.Tool
For body, by the optimal control of air knife distance, ensure the control allowance of air pressure, when making air pressure close to saturation region,
It remains to ensure that product quality meets index request in the case where promoting production line production capacity, reduces coating surface quality fluctuation, really
Protect galvanized product quality.
Claims (5)
1. a kind of zinc-plated production air knife is apart from optimal control method, it is characterised in that:Include the following steps:
S1:Forward and backward thickness of coating measured value is read, if storage region is less than, institute's Thickness Measurement by Microwave is stored in storage region, into step
Rapid S4;If storage region has expired, S2 is entered step;
S2:Measured value earliest in storage region, which is replaced, using current measured value then judges forward and backward plating in storage region
Layer deviation mean value CW_bias whether be more than front and back drift correction threshold value;If being more than, enter step S3, if not exceeded, then into
Enter step S4;
S3:Optimal air knife is calculated according to air knife distance, air pressure, line speed and thickness of coating departure real-time optimization
Distance, so that it is determined that forward and backward air knife is apart from knots modification GF_delta and GB_delta;Wherein, preceding thickness of coating needs change
Departure is-CW_bias/2, and it is CW_bias/2 that rear thickness of coating, which needs the departure changed,;Then, S5 is entered step;
S4:It apart from knots modification is 0 to enable forward and backward air knife, enters step S5;
S5:Judge whether air pressure enters zone of saturation, if into entering step S6, if not entering, entering step
S7;
S6:Increase air pressure adjust allowance, according to when former air knife distance, line speed, increase allowance after air pressure,
Using neural computing thickness of coating predicted value, and then obtain the departure of thickness of coating;
In turn, it is calculated according to the departure real-time optimization of air knife distance, air pressure, line speed and thickness of coating optimal
Air knife distance determines air knife apart from knots modification, and then the air knife is averagely allocated to forward and backward air knife apart from knots modification and is changed
Become, obtains forward and backward air knife apart from knots modification, enter step S8;
S7:Forward and backward air knife is remained unchanged apart from knots modification, into S8;
S8:The forward and backward air knife confirmed according to step S3 and S6 determines new forward and backward air knife apart from setting value apart from knots modification, according to
The air pressure for increasing allowance according to step S6 is air pressure setting value.
2. zinc-plated production air knife as described in claim 1 is apart from optimal control method, it is characterised in that:In the step S6
In, when using neural computing thickness of coating predicted value, further thickness of coating is obtained using neural network prediction module
Predicted value;
The neural network prediction module is input with strip speed, air knife distance, air pressure, and thickness of coating is built by output
It is vertical, and its historical data based on galvanizing process is learnt, so according to the strip speed of current operation, air knife distance and
The sampled value of air pressure predicts thickness of coating.
3. zinc-plated production air knife as claimed in claim 2 is apart from optimal control method, it is characterised in that:In the step S6
In, the thickness of coating of required compensation is:
CWm=NN (D (t), P (t), S (t))-NN (D (t), Pini,S(t))
Wherein, D (t) is the air knife distance of t moment, and P (t) is the air pressure of t moment, and S (t) is the line speed of t moment;
PiniIt is the air pressure increased after allowance.
4. zinc-plated production air knife as described in claim 1 is apart from optimal control method, it is characterised in that:In the step S3
And/or in S6, optimization calculate optimal air knife apart from when, utilize real-time optimization module optimization to calculate optimal air knife distance;
The real-time optimization module is minimised as target with the deviation between thickness of coating predicted value and control targe, with air knife away from
It is constraint from traffic coverage, optimal gas under air pressure, line speed and thickness of coating setting value that iterative search gives
Knife distance.
5. zinc-plated production air knife as claimed in claim 4 is apart from optimal control method, it is characterised in that:The real-time optimization mould
Real-time optimization algorithm used by block is specifically described as:
Object function:min|CWt-CWp|
Decision variable:D
Constraints:
CWp=NN (D, Pini,Sini) (6)
CWt=NN (Dini,Pini,Sini)+CWm (7)
D∈[Dmin,Dmax] (8)
Optimizing is iterated to D according to increment type PID algorithm, wherein increasable algorithm describes:
Error=CWt-CWp (9)
P_error=error-error_1 (10)
I_error=error (11)
D_error=error-2*error_1+error_2 (12)
Δ u=Kp*P_error+Ki*I_error+Kd*D_error (13)
D=D+ Δs u (14)
Wherein, the initial value of D is Dini, error_1 is the previous generation values in error iterative calculation, and error_2 is error iteration
Upper two generations value in calculating;By formula (9)~formula (14) iterative calculation, finds D and make CWp infinite approachs CWt;Real-time optimization is calculated
Method finally returns that air knife apart from knots modification Δ D, meets formula (15)
Δ D=D-Dini (15)
In above formula, Δ D is so that thickness of coating is changed total gas needed for CWm in current manufacturing lines speed and air pressure
Knife is apart from knots modification, and Δ D is the departure-CW_bias/2 for needing to change for preceding thickness of coating, and-Δ D is after being directed to
Thickness of coating needs the departure CW_bias/2 changed to be modified;
CWt is zinc coat thickness control target;CWp is neural network prediction value;Dmin is minimum value of the air knife apart from technological procedure;
Dmax is maximum value of the air knife apart from technological procedure;Error is the contemporary deviation in increment type PID iterative process;P_error
Proportional jitter in being calculated for increment type PID;I_error is the integral deviation during increment type PID calculates;D_error is increment
Differential deviation in formula PID calculating;Kp is the proportionality coefficient in increment type PID;Ki is the integral coefficient in increment type PID;Kd
For the differential coefficient in increment type PID;Δ u measures correction amount in order to control.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610650208.7A CN106435427B (en) | 2016-08-08 | 2016-08-08 | Zinc-plated production air knife is apart from optimal control method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610650208.7A CN106435427B (en) | 2016-08-08 | 2016-08-08 | Zinc-plated production air knife is apart from optimal control method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106435427A CN106435427A (en) | 2017-02-22 |
CN106435427B true CN106435427B (en) | 2018-10-02 |
Family
ID=58184345
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610650208.7A Active CN106435427B (en) | 2016-08-08 | 2016-08-08 | Zinc-plated production air knife is apart from optimal control method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106435427B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107794481A (en) * | 2017-10-20 | 2018-03-13 | 甘肃酒钢集团宏兴钢铁股份有限公司 | A kind of stair like control strategy method of galvanized wire air knife position |
CN113667918B (en) * | 2021-07-27 | 2023-08-15 | 首钢京唐钢铁联合有限责任公司 | Control method for thickness switching of hot dip galvanized strip steel coating |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102912275A (en) * | 2012-10-23 | 2013-02-06 | 鞍钢股份有限公司 | Automatic control system for plating thickness of hot galvanizing line |
CN103695830A (en) * | 2013-12-20 | 2014-04-02 | 鞍钢股份有限公司 | Coating thickness control method in hot galvanizing production process |
CN104846306A (en) * | 2015-05-07 | 2015-08-19 | 浙江中控研究院有限公司 | Zincing thickness control system and method |
CN105525248A (en) * | 2015-11-30 | 2016-04-27 | 浙江中控研究院有限公司 | Method for plating thickness feedforward optimal control of galvanization production line |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
KR101185933B1 (en) * | 2010-03-31 | 2012-09-25 | 현대제철 주식회사 | Hot coil for galva annealed steel sheet and method for producing galva annealed steel sheet by using hot coil |
-
2016
- 2016-08-08 CN CN201610650208.7A patent/CN106435427B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102912275A (en) * | 2012-10-23 | 2013-02-06 | 鞍钢股份有限公司 | Automatic control system for plating thickness of hot galvanizing line |
CN103695830A (en) * | 2013-12-20 | 2014-04-02 | 鞍钢股份有限公司 | Coating thickness control method in hot galvanizing production process |
CN104846306A (en) * | 2015-05-07 | 2015-08-19 | 浙江中控研究院有限公司 | Zincing thickness control system and method |
CN105525248A (en) * | 2015-11-30 | 2016-04-27 | 浙江中控研究院有限公司 | Method for plating thickness feedforward optimal control of galvanization production line |
Also Published As
Publication number | Publication date |
---|---|
CN106435427A (en) | 2017-02-22 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105525248B (en) | A kind of galvanization production line thickness of coating feedforward optimal control method | |
CN104846306B (en) | A kind of Zinc Coating Thickness control system and method | |
CN103469137B (en) | A kind of hot dip galvanizing coating thickness dynamic variable specification default control method | |
CN105316613B (en) | A kind of zinc coating thickness control method and system based on time_varying delay offset correction technology | |
CN106637026B (en) | A kind of galvanizing process air pressure method for real-time optimization control and system | |
CN103611734B (en) | A kind of section cooling temperature-controlled process and system | |
CN106167887B (en) | Hot dip galvanizing coating thickness fast switch over method and system away from dynamic compensation based on knife | |
CN101983800B (en) | Secondary cooling water distribution advanced control method for billet continuous casting machine | |
CN103695830A (en) | Coating thickness control method in hot galvanizing production process | |
CN106435427B (en) | Zinc-plated production air knife is apart from optimal control method | |
CN110607435B (en) | Annealing furnace plate temperature control system and method | |
CN106119520A (en) | A kind of cold rolling hot dip galvanizing annealing furnace plate temperature coordinated control system and method | |
CN104289532A (en) | Method for controlling temperature of watermark points of strip steel | |
CN107130199B (en) | A kind of continuous hot galvanizing air knife knife is away from autocontrol method | |
CN109240203B (en) | Multi-model-based continuous casting billet constant weight control method | |
CN109763085A (en) | A kind of temperature controlled method of hot galvanizing line alloying section | |
CN102873106B (en) | Quick and precise elongation control method for temper mill | |
CN110579968A (en) | Prediction control strategy for ultra-supercritical unit depth peak regulation coordination system | |
CN108396275A (en) | A kind of continuous hot galvanizing air knife jetting pressure autocontrol method | |
CN110295338A (en) | A kind of stable strip enters the control method of zinc pot temperature | |
CN103510032A (en) | Deviation value control method for cold rolling hot galvanizing coating uniformity | |
CN105568197A (en) | Method for accurately controlling weight of continuously-hot-galvanized zinc layer | |
CN106521390B (en) | A kind of strip continuous hot galvanizing zinc coat thickness control method based on database | |
JP2939033B2 (en) | Manufacturing method of galvannealed steel sheet | |
CN114130980B (en) | Dynamic secondary cooling control method for continuous casting |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |