CN102366757A - Consecutive mean module and method thereof - Google Patents

Consecutive mean module and method thereof Download PDF

Info

Publication number
CN102366757A
CN102366757A CN2011102749393A CN201110274939A CN102366757A CN 102366757 A CN102366757 A CN 102366757A CN 2011102749393 A CN2011102749393 A CN 2011102749393A CN 201110274939 A CN201110274939 A CN 201110274939A CN 102366757 A CN102366757 A CN 102366757A
Authority
CN
China
Prior art keywords
weights
shift register
value
register
thickness
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN2011102749393A
Other languages
Chinese (zh)
Other versions
CN102366757B (en
Inventor
陈跃华
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wisdri Engineering and Research Incorporation Ltd
Original Assignee
Wisdri Engineering and Research Incorporation Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Wisdri Engineering and Research Incorporation Ltd filed Critical Wisdri Engineering and Research Incorporation Ltd
Priority to CN 201110274939 priority Critical patent/CN102366757B/en
Publication of CN102366757A publication Critical patent/CN102366757A/en
Application granted granted Critical
Publication of CN102366757B publication Critical patent/CN102366757B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Control Of Metal Rolling (AREA)

Abstract

The invention provides a consecutive mean module and a method thereof for solving the mean value of monitoring thickness errors in a monitoring section. The consecutive mean module and method thereof are characterized in that shift registers are used for storing monitoring thickness error values and weights thereof; the thickness error value and weight thereof at the current time are stored in first positions of respective registers; along with the movement of an object corresponding to the error value, the error value and weight thereof move along in the registers; when the corresponding object moves for the whole monitoring section distance, the weight sum of the object on the monitoring section reaches 1.0, and by now, the monitoring thickness error values and weights thereof are multiplied and summarized to obtain the mean value of the monitoring thickness errors on the monitoring section. The consecutive mean module takes possible changes of transmission speed into account, and because the current transmission speed is input at each sampling instant, the mean value of the monitoring thickness errors on the monitoring section can be accurately solved while the speed changes randomly, so that the consecutive mean module can be widely used for solving the mean parameter of speed labile to change.

Description

Consecutive mean module and method thereof
Technical field
The present invention relates to a kind of consecutive mean module of supervision thickness error that be used to ask at supervision section mean value.
Background technology
In cold continuous rolling belt steel rolling process, in order to improve vertical tolerance of band steel, improve the product thickness hit rate, the tandem mill automatic thickness control system has been carried out a large amount of research both at home and abroad.The most frequently used in the past method for controlling thickness is through configuration calibrator after rolling-mill housing the actual (real) thickness of band steel to be measured, and then comes belt steel thickness is carried out FEEDBACK CONTROL through the hydraulic pressure roll gap of regulating milling train.This method for controlling thickness is called keeps watch on AGC (Monitor Automatic Gauge Control).But because the restriction of rolling mill structure, calibrator generally is installed in the place apart from the mill roll-gap certain distance, and the actual like this fluctuation that shuts out thickness must just can obtain through one period lag time, and be totally unfavorable to the control systematic function this lag time.
Might directly accurately measure strip speed because the release of laser velocimeter makes the nineties in last century, therefore not only can accurately obtain the advancing slip value of each frame, and can accurately calculate the deformed area exit thickness through the identical rule of deformed area second flow amount.The meaning of the constant rule of second flow amount is that the mass flow of metal is constant before and after the frame, again because the strip width basically identical before and after the frame then keeps strict proportionate relationship with the speed and the thickness of steel before and after frame, that is:
V en×h en=V ex×h ex
V in the formula En-band steel entrance velocity; V Ex-band steel exports speed; h En-band steel inlet thickness; h Ex-band steel exports thickness.If to band steel section h EnThe actual measurement back is through postponing, as actual measurement h EnBand steel section when getting into the deformed area according to the V of actual measurement this moment EnAnd V ExCan accurately obtain the deformed area exit thickness of this band steel section.This technology has solved the problem of long-term puzzlement cold continuous rolling AGC system design; Because the use of new pattern laser tachymeter can obtain the deformed area exit thickness accurately and can not carry out FEEDBACK CONTROL with lagging behind, thereby successfully thick control precision has been improved an one magnitude.
Second, flow quantity AGC solved problem lag time of keeping watch on AGC, had improved control performance greatly, but kept watch on necessity that AGC still has its existence.Though at first export the calibrator measurement result hysteresis is arranged, its precision is very high, generally can reach 1um; Second flow amount accounting equation is difficult to the precision that reaches such, can utilize outlet calibrator measured value that second flow amount accounting equation is revised, and makes that the precision of second flow quantity AGC is higher; The second, a second flow quantity AGC can't overcome the error that the milling train operating point changes and causes, for example owing to long-term rolling roll wear, thermal expansion of rollers, the advancing slip coefficient of causing changes or the like.The thick difference that these comparatively long-term processes cause still need be kept watch on AGC and overcome.
At present the best supervision AGC method of result of use has been used Smith to estimate to overcome the lag time that the outlet calibrator is measured; If know the exact value of lag time; Smith predictor method " prediction " band steel exports thickness well then, the performance of keeping watch on AGC is also relatively good.If but were with steel exports speed to change, would be a variable lag time that then exports the calibrator measurement, can have a strong impact on the performance of Smith predictor method if change greatly lag time.
The roll gap regulated quantity that original supervision AGC method calculates milling train usually realizes keeping watch on AGC and regulates, and mill roll-gap uses hydraulic press down system control, and the hydraulic press down system precision is high, response is fast, but its formula that calculates the roll gap regulated quantity is:
ΔS = ( 1 + C m K m ) Δh
In the formula, Δ S is for keeping watch on the roll gap regulated quantity of AGC; C mFor-band steel plastic coefficient; K mBe the mill stiffness coefficient; Δ h is band steel exports thickness error monitoring value; Wherein be with steel plastic coefficient C mWith the mill stiffness COEFFICIENT K mTest obtains in the test run stage for milling train, can't obtain exact value at present, and this has influenced the precision of keeping watch on AGC.
Summary of the invention
The technical problem that the present invention will solve is: provide a kind of and be used to ask for the supervision thickness error in the consecutive mean module and the method thereof of keeping watch on section mean value, can accurately ask for the mean value of thickness error on the supervision section under the situation that strip speed changes arbitrarily.
The present invention solves the problems of the technologies described above the technical scheme of being taked to be: the consecutive mean module; It is characterized in that: the consecutive mean module comprises the 4th to the 6th shift register and judging treatmenting module; The monitoring segment length is divided into plurality of sections, and hop count is determined that by strip speed every section length is L S=V * T S, L in the formula SBe every section length, V is band steel transmission speed, T SBe the sampling time;
The storage of the 4th shift register needs the sampled value of the input variable of dynamic calculation mean value, and is every through a sampling time T SAll elements in the register is moved one successively backward, then the input sample value of current time is deposited in first position of this register;
The 5th shift register is used for storing the weights of input sample value, and weights determine that by strip speed the weights of each input sample value are W X=V * T S/ L Mon, in the formula, L MonFor keeping watch on the length of section; Whenever, the weights of all storages in the register are moved one successively backward through a sampling time, then the weights of current time input variable are deposited in first position of this register;
The weights that the 6th shift register is used for storing the input sample value with; All weights additions that are about to corresponding times prior in the 5th shift register obtain; Each sampling instant is with the weights of all storages and move one backward successively; Then the weights of current time input variable are stored in first position of this register, promptly first element of the 6th shift register and the 5th shift register always equates; Three registers are corresponding one by one;
Judging treatmenting module is used for each sampling time and judges that the element more than or equal to 1.0 appears in the 6th shift register at first; If i element more than or equal to 1.0 i-1 element less than 1.0; Then i element of the 5th shift register deducted that i element surpasses 1.0 amount in the 6th shift register, with preceding i element guaranteeing the 5th shift register and be 1.0 just in time; Then preceding i element of the 4th shift register multiply by preceding i element value of corresponding the 5th shift register, and summation obtains keeping watch on thickness error at the mean value of keeping watch on the segment length.
The consecutive mean method is characterized in that: it may further comprise the steps:
The sampled value of the input variable of step 1, input strip speed V, supervision segment length L and need dynamic calculation mean value;
Step 2, the storage of the 4th shift register need the sampled value of the input variable of dynamic calculation mean value, and be every through a sampling time T SAll elements in the register is moved one successively backward, then the input sample value of current time is deposited in first position of this register;
Step 3, the 5th shift register are used for storing the weights of input sample value, and weights determine that by strip speed the weights of each input variable are W X=V * T S/ L Mon, in the formula, L MonFor keeping watch on the length of section; Whenever, the weights of all storages in the register are moved one successively backward through a sampling time, then the weights of current time input variable are deposited in first position of this register;
Step 4, the 6th shift register be used for storing the weights of keeping watch on the thickness error sampled value and; All weights additions that are about to corresponding times prior in the 5th shift register obtain; Each sampling instant is with the weights of all storages and move one backward successively; Then the weights of current time input variable are stored in first position of this register, promptly first element of the 6th shift register and the 5th shift register always equates; Three registers are corresponding one by one;
Step 5, each sampling time are judged the element that occurs at first in the 6th shift register more than or equal to 1.0; If i element more than or equal to 1.0 i-1 element less than 1.0; Then i element of the 5th shift register deducted that i element surpasses 1.0 amount in the 6th shift register, with preceding i element guaranteeing the 5th shift register and be 1.0 just in time; Then preceding i element of the 4th shift register multiply by preceding i element value of corresponding the 5th shift register, and summation obtains keeping watch on thickness error at the mean value of keeping watch on the segment length.
It also comprises step 6, the maximum memory space C of each shift register is set, and the weights W of current time is all judged in each sampling instant XWhether more than or equal to minimum weights W Min=1.0/C, and if less than minimum weights three shift register of this sampling instant all do not upgrade, and only with weights summations, judge weights and whether more than or equal to minimum weights in next sampling instant then; Just carry out the renewal of three shift registers more than or equal to minimum weights up to weights and value.
Operation principle of the present invention is: in the belt steel rolling process; Enable simultaneously under the situation of second flow quantity AGC and supervision AGC; Use exit thickness precomputation device to calculate band steel exports thickness; This one-tenth-value thickness 1/10 is to come out according to the band steel section THICKNESS CALCULATION at current time inlet calibrator place; For the detection thickness that uses the outlet calibrator is revised it, just must compare with the detection thickness at outlet calibrator place then and just obtain second flow amount calculated thickness error with this calculated value synchronous transfer to exporting the calibrator place; The supervision AGC compensating factor that multiply by the calculating of adaptive equalization device has just obtained supervision AGC error amount; Consider that including a moment in this error amount keeps watch on the influence value that the AGC adjusting causes, therefore must this influence value be deducted that the influence value that supervision AGC adjusting causes influences compensation arrangement by supervision AGC and calculates.
Keep watch on the AGC error and ask for it at the mean value of keeping watch on the section, through the ramp generator device, just obtain keeping watch on the average ramp value of AGC error again through the consecutive mean module.
Consider the speed regulating mechanism action delay, the average ramp value of supervision AGC error is compensated action delay through the first first-order lag device then obtain keeping watch on the error correction values of AGC second flow amount equation.
Notice that calculating is to use the thickness data at outlet calibrator place to calculate when keeping watch on the AGC error amount; Therefore when influencing of causing controlled in the supervision of calculating current time; Need be with the second flow amount error correction values synchronous transfer at roll gap place to exporting the calibrator place; Re-use the sampling time-delay that the 21 joint lagging device compensation drops out mouthful calibrator, just obtain current time and keep watch on AGC control next influence value that causes constantly.
Use band steel exports speed adjusting device to calculate band steel exports speed regulation amount according to keeping watch on the average ramp value of AGC error at last; Because rolling mill speed is controlled through transmission device; Therefore the roller of having considered advancing slip and live-roller and working roll directly than situation under, calculate milling train live-roller linear velocity regulated quantity through the live-roller speed adjusting device.This regulated quantity is outputed to the milling train transmission device go just to have realized this high-precision supervision AGC control.
Beneficial effect of the present invention is:
1, because the consecutive mean module all will obtain weights according to the band steel transmission speed of current time in each sampling time; So can accurately ask for the mean value of thickness error on the supervision section under the situation that strip speed changes arbitrarily, the calculating of controlling system for whole AGC provides the foundation.
2, overflow fault can not take place in order to ensure register, the weights W of current time is all judged in each sampling instant XWhether more than or equal to minimum weights W Min=1.0/C, and if less than minimum weights three shift register of this sampling instant all do not upgrade, and only with weights summations, judge weights and whether more than or equal to minimum weights in next sampling instant then; Just carry out the renewal of three shift registers more than or equal to minimum weights up to weights and value.
Description of drawings
Fig. 1 is five frame cold continuous rollings and main detecting element sketch map.
Fig. 2 is the program flow diagram of synchronous transfer model.
Fig. 3 is a program flow diagram of the present invention.
Fig. 4 keeps watch on the schematic diagram of automatic thickness control system for the high accuracy belt steel rolling.
The specific embodiment
Present embodiment is that the consecutive mean module is kept watch on the application in the automatic thickness control system at the high accuracy belt steel rolling.
Fig. 1 is five frame cold continuous rollings and main detecting element sketch map, and five frames of five frame cold continuous rollings are six-high cluster mill all, and the numeral 1~5 above the frame is represented frame respectively 1~No. 5, and rolling direction from left to right.Because the plant equipment of each frame all is the same, form with the equipment that No. 3 frames are example explanation five frame connection rolling machine: No. 3 frames by last backing roll 6, go up intermediate calender rolls 7, top working roll 8, bottom working roll 9, intermediate calender rolls 10, following backing roll 11 are formed down.In addition, jockey pulley 12 is equipped with in the tandem mill porch, coiling machine 13 is equipped with in the exit; Rolling mill transmission system is the intermediate calender rolls transmission; Intermediate calender rolls uses frequency control motor 14,15 to carry out speed governing up and down; Motor is driven by frequency converter 16, and programmable logic controller (PLC) (PLC) 17 is issued frequency converter with the speed regulation amount, and frequency converter is regulated motor speed makes the live-roller linear velocity reach setting value; Milling train uses press down system 18 control roll gap; The control principle is that THICKNESS CONTROL (AGC) is accomplished through regulating the milling train transmission speed, and tension force is then accomplished through regulating roll gap between frame.Instrumentation comprises 4 cover x gamma thickness gages, 3 cover laser velocimeters, some cover pulse coders (pulse coder all is housed) altogether on all bunchers.Wherein 4 cover x gamma thickness gages are respectively frame inlet calibrator 19,1 frame outlet calibrator 20,5 frames inlet calibrator 21,5 frames outlet calibrator 22 No. 1; 3 cover laser velocimeters are respectively 1 frame outlet tachymeter, 23,4 frames outlet tachymeter, 24,5 frames outlet tachymeter 25; Because inlet jockey pulley place does not exist advancing sliply, on the inlet jockey pulley, pulse coder 26 is installed and promptly can be accurately obtained 1 frame strip steel at entry linear velocity in addition, such 1 frame porch need not dispose laser velocimeter.According to such instrument configuration; No. 1 and No. 5 frames can be used second flow quantity AGC and the AGC control of keeping watch on the AGC coupling; But because two frames in tandem mill end have three kinds of control models according to the specification difference of rolled band steel; Be respectively Mode A, B and C, correspond respectively to thicker softer band steel, the band steel of intermediate gauge, thin harder band steel, the implementation method of keeping watch on AGC has some differences.For example following No. 5 frames of C pattern are as smooth frame, and the exit thickness of 5 frames is not controlled, and need on preceding 4 frames, accomplish the adjusting of band steel exports thickness, and the calculating of controlled quentity controlled variable and objective for implementation are with different under other pattern like this; The control method of tension force also has certain difference under A pattern and the B pattern.Because the supervision AGC method after diversity 5 frames of pattern is not suitable for explaining the thought of this method; Therefore present embodiment is that example is done an explanation with the supervision AGC method of No. 1 frame; Supervision AGC control thought after 5 frames is the same, and just implementation method has some differences.
This AGC method may further comprise the steps:
1. calculate the band steel exports thickness h at current time outlet calibrator place Exc, ethg:
In the operation of rolling, enabled simultaneously under the situation of second flow quantity AGC and supervision AGC, precomputation obtains being with the steel exports one-tenth-value thickness 1/10 according to second flow amount equation, should value transfer to outlet calibrator place from the inlet calibrator through two synchronous transfer models, is specially:
(1) according to second flow amount equation precomputation band steel exports one-tenth-value thickness 1/10:
No. 1 the milling train porch is equipped with calibrator can record band steel inlet thickness; Be equipped with pulse coder on the inlet jockey pulley; Advancing slip owing to not existing at the jockey pulley place, therefore can be with the jockey pulley linear velocity as band steel entrance velocity, 1 frame exit is equipped with laser velocimeter can obtain being with steel exports speed; According to the measured value that these instrumentations obtain, use exit thickness precomputation device 27 to calculate band steel exports thickness h with following formula Ex, calc:
h ex , calc = V en × h en V ex
In the formula, h Ex, calcExit thickness for the calculating of second flow amount equation; V EnFor band steel entrance velocity actual value, measure by pulse coder on the inlet jockey pulley; V ExFor band steel exports speed actual value, measure by the outlet laser velocimeter; h EnFor band steel inlet thickness actual value, measure by the inlet calibrator.
(2) use the first synchronous transfer model to obtain being transferred to the exit thickness value h at roll gap place Exc, gap:
Use the first synchronous transfer model, with the h that calculates Ex, calcFrom inlet calibrator synchronous transfer to 1 frame roll gap place, obtain the exit thickness value h at current time roll gap place with band steel entrance velocity Exc, gap, that is:
h exc,gap=TPM1(h ex,calc)
(3) use the second synchronous transfer model to obtain being transferred to the exit thickness value h at outlet calibrator place Exc, ethg:
Use the second synchronous transfer model, with the exit thickness value h at roll gap place Exc, gapWith band steel exports speed from the roll gap synchronous transfer to exporting the calibrator place, obtain the exit thickness value h at current time outlet calibrator place Exc, ethg, that is:
h exc,ethg=TPM2(h exc,gap)。
The present invention adopts unique synchronous transfer model, and the program flow diagram of synchronous transfer model TPM is as shown in Figure 2, and each synchronous transfer model structure is identical.Present embodiment is that example describes with the first synchronous transfer model, and the first synchronous transfer model is made up of first to the 3rd shift register and judge module, and each shift register can store several real variable.At first transmission range is divided into several sections, notices that the segmentation number do not fix, but by the strip speed decision, the length of each section is: L S=V * T S, L in the formula SBe every section length, V is a transmission speed, T SBe the sampling time; T sBe fixed value, look the setting of PLC controller and decide T in the present embodiment s=4ms.
The belt steel thickness sampled value of first shift register storage input, every through a sampling time T sThe data of all storages are moved one, first position of then the belt steel thickness sampled value of current time being deposited this register successively backward;
The length of the second shift register memory segment whenever moves one with the data of all storages through a sampling time successively backward, then the section length of current time is stored in first position of this register;
The 3rd shift register is stored the distance that corresponding belt steel thickness sampled value is passed by, and all section length additions that are about to the corresponding times prior of second shift register obtain, and for example k element of the 3rd shift register is:
B 3 [ k ] = Σ m = 1 k B 2 [ m ] ,
In the formula, B2 and B3 represent the second and the 3rd shift register respectively.
Whenever, through a sampling instant distance value of all storages is moved one successively backward, then the section length of current time is stored in first position of this register, promptly first element of the second and the 3rd shift register is always equal; Three registers are corresponding one by one;
Judge module be used for each sampling time judge one the order three shift registers whether the element more than or equal to transmission range L is arranged; If i element value more than or equal to transmission range i-1 element value less than transmission range; Represent that i element in first shift register transmitted the given transmission distance, then with i element value output in first shift register;
Note need being provided with in advance the maximum memory space N of register here, if strip speed is very slow, the length of each segmentation is all especially little, and the register overflow fault will take place for N section length and also do not reach the transmission range L of setting.Therefore in this mode, memory space N=500 is set, each sampling instant all can be judged section length L SWhether less than minimum section length L Min=L/500, and if less than minimum section length three register of this sampling instant do not upgrade, and only with section length summation, judge section length in next sampling instant then with value whether more than or equal to minimum section length; Just carry out the renewal of three shift registers up to section length more than or equal to minimum section length with value.So just guaranteed that shift register overflow fault can not take place.
2. obtain and keep watch on AGC adaptive equalization factor F Comp:
According to band steel exports speed, export thick poor, whether be in the adaptively selected supervision of this Several Factors of acceleration and deceleration stage AGC compensating factor, be specially:
(1) according to 1 frame muzzle velocity actual value access speed factor F V, more greatly then velocity factor is more little for muzzle velocity, and the velocity factor value is between 0.7~1.0.
(2) choose thickness factor F according to the exit thickness difference h, more greatly then the thickness factor is big more to export thick difference, and thickness factor value is between 0.6~1.0.
(3) whether basis is in acceleration and deceleration stage selection acceleration factor F aIf, be not in the acceleration and deceleration stage, then acceleration factor is taken as 1.0; If be in the acceleration and deceleration stage, then need improve acceleration factor, can be taken as 1.2;
(4) keep watch on AGC adaptive equalization factor F with computes Comp:
F comp=F V×F h×F a
3. obtain and keep watch on the average ramp value of AGC exit thickness error
At first try to achieve the supervision thickness error E of exit thickness Mon, ask for this error then at the mean value of keeping watch on the section, step appears in order to prevent controlled quentity controlled variable, at last also in addition the slope handle and obtain keeping watch on the average ramp value of AGC exit thickness error
Figure BDA0000091731970000072
Be specially:
(1) keeps watch on thickness error E with the computes current time Mon:
E mon(n)=(h ex,thg-h exc,ethg)×F comp-E tpm(n-1) n>1;
E mon(n)=(h ex,thg-h exc,ethg)×F comp n=1;
In the formula, h Ex, thgBe outlet calibrator measured value; E Tpm(n-1) keep watch on the exit thickness influence value that the AGC controlled quentity controlled variable causes current time constantly for last one; N representes current time; Can the exit thickness in this moment be impacted after regulating because last one constantly keeps watch on AGC, therefore moment of the supervision AGC controlled quentity controlled variable of calculating current time must be earlier with on one control the influence value that causes constantly and deduct.If current time is an initial time, the supervision AGC controlled quentity controlled variable that goes up a moment so is 0, so its influence value also is 0.
(2) use the consecutive mean module to ask for E MonAt the mean value of keeping watch on the section
Figure BDA0000091731970000081
Flow chart is as shown in Figure 3:
Keeping watch on AGC is to be used to compensate the trend (thin or thick partially partially) that comparatively long-term process changes the thick difference of outlet that causes, and therefore need ask for E to whole supervisions section (promptly from the frame roll gap to exporting this segment distance of calibrator) MonMean value, the error that interference more at random causes can be cancelled out each other.
Use consecutive mean module MAV to ask for E MonAt the mean value of keeping watch on the section
Figure BDA0000091731970000082
That is:
E mon ‾ = MAV ( E mon ) .
The consecutive mean module is made up of the 4th to the 6th shift register and judging treatmenting module, will the section of monitoring (promptly from the frame roll gap to exporting this segment distance of calibrator) be divided into plurality of sections, segments is determined that by strip speed every section length is L S=V * T S, L in the formula SBe every section length, V is band steel transmission speed, T SBe the sampling time;
The 4th shift register storage input variable E MonSampled value, every through a sampling time T SThe value of all storages in the register is moved one successively backward, then the sampled value of current time input variable is deposited in first position of this register;
The 5th shift register is used for storing the weights of input variable sampled value, and weights determine that by strip speed the weights of each input variable are W X=V * T S/ L Mon, in the formula, L MonFor keeping watch on the length of section; Whenever, the weights of all storages in the register are moved one successively backward through a sampling time, then the weights of current time input variable are deposited in first position of this register;
The weights that the 6th shift register is used for storing the input variable sampled value with, all weights additions that are about to corresponding times prior in the 5th shift register obtain, for example k element in the 6th shift register is:
B 6 [ k ] = Σ m = 1 k B 5 [ m ] ,
In the formula, B5 and B6 represent the 5th and the 6th shift register respectively.
Each sampling instant is with the weights of all storages and move one backward successively, then the weights of current time input variable is stored in first position of this register, and promptly first element of the 6th shift register and the 5th shift register always equates.Three registers are corresponding one by one;
Judging treatmenting module is used for each sampling time and judges that the element more than or equal to 1.0 appears in the 6th shift register at first; If i element more than or equal to 1.0 i-1 element less than 1.0; Then i element of the 5th shift register deducted that i element surpasses 1.0 amount in the 6th shift register, with preceding i element guaranteeing the 5th shift register and be 1.0 just in time; Then preceding i element of the 4th shift register multiply by preceding i element value of corresponding the 5th shift register, i.e. each input variable E constantly MonMultiply by its corresponding weights and sue for peace and just obtained input variable E MonAt the mean value of keeping watch on the segment length That is:
E mon ‾ = Σ j = 1 i ( B 4 [ j ] × B 5 [ j ] ) ,
In the formula, B4 and B5 represent the 4th and the 5th shift register respectively.
Note need being provided with in advance the maximum memory space C of register here, if strip speed is very slow, the weights of each segmentation are all especially little, and the register overflow fault will take place for C weights and also do not reach 1.0.Therefore in this consecutive mean module, memory space C=100 is set, each sampling instant all can be judged current time weights W XWhether more than or equal to minimum weights W Min=1.0/C, and if less than minimum weights three shift register of this sampling instant all do not upgrade, and only with weights summations, judge weights and whether more than or equal to minimum weights in next sampling instant then; Just carry out the renewal of three shift registers more than or equal to minimum weights up to weights and value.
(3) monitor the use of exit ramp generator to generate the average thickness error
Figure BDA0000091731970000093
slope value
Figure BDA0000091731970000094
Calculate the controlled quentity controlled variable of keeping watch on AGC owing to need to use
Figure BDA0000091731970000095
; In general; Avoid AGC controlled quentity controlled variable generation step to change; Can bring bigger impact to the executing agency of controller like this; Therefore used ramp value
Figure BDA0000091731970000096
that ramp generator RGE (Ramp Generator) produces the thick poor mean value of outlet here like this when the variable generation step variation of input; Step can not take place and change in the output valve of ramp generator, but the slope arrives the input variable value.
Ramp generator is calculated as follows ramp value:
Y ramp ( n ) = Y ramp ( n - 1 ) + T S T A ,
In the formula, Y Ramp(n) be current time ramp generator output valve, Y Ramp(n-1) be a last moment ramp generator output valve, T SBe the sampling time, T ABe the slope time, through regulating T AThe speed degree of control ramp, T AMore greatly then ramp must be slow more, T AMore little then ramp must be fast more.
Use ramp generator to produce and keep watch on the average ramp value of exit thickness error
Figure BDA0000091731970000098
promptly:
E R , mon ‾ = RGE ( E mon ‾ )
Like this as
Figure BDA00000917319700000910
when changing; becomes new value at once, reaches new value but each sampling instant changes
Figure BDA00000917319700000912
up to
Figure BDA00000917319700000913
.
4. obtain second flow amount calculated thickness error correction values Δ h MFCKeep watch on the influence value E that AGC controls with current time Tpm(n):
At first with the average ramp value of thickness error
Figure BDA0000091731970000101
Action delay through first first-order lag link compensation executing agency promptly obtains a second flow quantity AGC thickness error correction value Δ h MFCUse the 3rd synchronous transfer model and the second first-order lag link to ask for the influence value E that current time is kept watch on AGC then Tpm(n), be specially:
(1) compensate the action delay that AGC controls executing agency through the first first-order lag link:
Need compensate the action delay of AGC control executing agency in the time of error of calculation correction value, represent AGC control executing agency with the first first-order lag link PT1 here, the PT1 link algorithm of discretization is:
Y pt 1 ( n ) = Y pt 1 ( n - 1 ) + T S T D × ( X n - X n - 1 )
In the formula, Y Pt1(n) be the output valve of PT1 link current time; Y Pt1(n-1) be the output valve in a moment on the PT1 link; T SBe the sampling time of PLC controller; T DBeing PT1 link time constant, being taken as the rise time of AGC control executing agency here, because tandem mill AGC carries out through regulating transmission speed, is the rise time of speed regulating mechanism so here; X nFor the current time input value, here be
Figure BDA0000091731970000103
The value of current time; X N-1For
Figure BDA0000091731970000104
The value in a last moment.
Will
Figure BDA0000091731970000105
Through the rise time of the first first-order lag link compensation speed governor motion, obtain the thickness error correction amount h of second flow quantity AGC MFC, that is:
Δ h MFC = PT 1 SA ( E R , mon ‾ ) ,
Subscript SA representes speed regulating mechanism in the formula.
Calculate and keep watch on the correction thickness error Δ h of AGC flow quantity AGC second MFCAfter, with Δ h MFCDirectly be added to the second flow amount and calculate outlet one-tenth-value thickness 1/10 h Ex, calcGet on, just obtain calculating the outlet one-tenth-value thickness 1/10,, can further improve the control accuracy of second flow quantity AGC like this according to the controlled quentity controlled variable of revised exit thickness value calculating second flow quantity AGC through keeping watch on the revised second flow amount of AGC.
(2) use the 3rd synchronous transfer model with Δ h MFCValue transfers to outlet calibrator place from roll gap, obtains Δ h MFC, ethg, the 3rd synchronous transfer model is identical with the first synchronous transfer model structure.
Notice that calculating is to use the thickness data at outlet calibrator place to calculate when keeping watch on the AGC error amount; Therefore when influencing of causing controlled in the supervision of calculating current time; Need to use the 3rd synchronous transfer model 34 that the second flow amount error correction values at roll gap place is transferred to outlet calibrator place, this model is with the Δ h that calculates MFCWith band steel exports speed from the roll gap synchronous transfer to exporting the calibrator place, that is:
Δh MFC,ethg=TPM3(Δh MFC)
(3) delay time through the detection of second first-order lag link compensation outlet calibrator;
Feed the 3rd first-order lag link with transferring to flow quantity AGC thickness error correction value second that the outlet calibrator goes out, the detection time-delay of compensation outlet calibrator, the time constant of the 3rd first-order lag link detects delay time for the outlet calibrator.Finally obtain current time and keep watch on the AGC controlled quentity controlled variable next influence value E that causes constantly Tpm(n), that is:
E tpm(n)=PT1 ethg(Δh MFC,ethg)
In the formula, subscript ethg representes to export calibrator; The influence value E of current time Tpm(n) supply n+1 to use constantly after calculating.
5. obtain the live-roller linear velocity regulated quantity Δ V that keeps watch on AGC R, mon:
According to keeping watch on the average ramp value of AGC exit thickness error
Figure BDA0000091731970000111
Calculate needed muzzle velocity regulated quantity Δ V Mon, obtain 1 frame live-roller linear velocity regulated quantity Δ V according to the advancing slip factor and roller footpath calculation of parameter again R, mon, be specially:
(1) with the needed 1 frame muzzle velocity regulated quantity Δ V of computes Mon:
Δ V mon = - V ex × E R , mon ‾ h ex , set ,
In the formula, h Ex, setBe band steel exports thickness setting value, by process computer according to rolling procedure making.(2) with the regulated quantity Δ V of computes 1 frame live-roller linear velocity R, mon:
Δ V r , mon = Δ V mon ( 1 + F sl ) × R d
In the formula, F SlBe the advancing slip factor of 1 frame, general value is between 0.03~0.5; R dFor the roller of live-roller and working roll directly than, be the intermediate calender rolls transmission in this example, therefore be exactly the ratio of intermediate calender rolls here with working roll roller footpath.
Through above-mentioned steps, finally obtain second flow amount calculated thickness error correction values Δ h MFCWith the transmission speed regulated quantity Δ V that keeps watch on AGC R, mon, the second flow amount error correction values of trying to achieve is delivered to a second flow quantity AGC program, and the transmission speed regulated quantity is delivered to frequency converter remove to regulate the frame driving motor, just accomplished this supervision AGC and regulated.
It is as shown in Figure 4 that the high accuracy belt steel rolling is kept watch on automatic thickness control system, and programming realizes that it comprises with lower device on PLC:
Exit thickness precomputation device 27 is used for calculating band steel exports thickness in band steel porch according to second flow amount equation:
h ex , calc = V en × h en V ex ,
In the formula, h Ex, calcExit thickness for the calculating of second flow amount equation; V EnFor band steel entrance velocity actual value, measure by pulse coder on the inlet jockey pulley; V ExFor band steel exports speed actual value, measure by the outlet laser velocimeter; h EnFor band steel inlet thickness actual value, measure by the inlet calibrator;
The first synchronous transfer model 28 is used for h Ex, calcDevice with band steel entrance velocity synchronous transfer to mill roll-gap place obtains mill roll-gap place precomputation thickness h Exc, gap
The second synchronous transfer model 29 is used for the precomputation thickness h with the roll gap place Exc, gapTransfer to outlet calibrator place with band steel exports speed synchronization, obtain exporting the precomputation exit thickness h at calibrator place Exc, ethg
Adaptive equalization device 30 is used for, gantry speed poor according to exit thickness, acceleration is asked for and kept watch on AGC adaptive equalization factor F Comp, this device utilizes following formula to calculate:
F comp=F V×F h×F a
In the formula, F VBe the muzzle velocity factor; F hFor exporting the thick poor factor; F aBe acceleration factor;
Be compensated factor F CompAfter, keep watch on thickness error E with the computes current time Mon:
E mon(n)=(h ex,thg-h exc,ethg)×F comp-E tpm(n-1) n>1;
E mon(n)=(h ex,thg-h exc,ethg)×F comp n=1;
In the formula, h Ex, thgBe outlet calibrator measured value; E Tpm(n-1) keep watch on the exit thickness influence value that the AGC controlled quentity controlled variable causes current time constantly for last one; N representes current time; Can the exit thickness in this moment be impacted after regulating because last one constantly keeps watch on AGC, therefore moment of the supervision AGC controlled quentity controlled variable of calculating current time must be earlier with on one control the influence value that causes constantly and deduct.If current time is an initial time, the supervision AGC controlled quentity controlled variable that goes up a moment so is 0, so its influence value also is 0.
Consecutive mean module 31 is used to ask for supervision thickness error E MonAt the mean value of keeping watch on the section
Figure BDA0000091731970000121
Second flow amount calculated thickness error correction device is used to calculate the exit thickness error correction values that the second flow amount is calculated, and comprises the ramp generator 32 and the first first-order lag device 33; Ramp generator 32 is used for calculating
Figure BDA0000091731970000122
Ramp value
Figure BDA0000091731970000123
The rise time that the first first-order lag device 33 is used for the compensation speed governor motion obtains second flow amount calculated thickness round-off error Δ h MFC
The 3rd synchronous transfer model 34 is used for Δ h MFCObtain exporting the thickness round-off error Δ h at calibrator place to outlet calibrator place from the roll gap synchronous transfer MFC, ethg
The second first-order lag device 35 is used to compensate the sampling time-delay that exports calibrator and obtains current time monitoring AGC controlled quentity controlled variable to next exit thickness influence value E that causes constantly Tpm(n);
Keep watch on the AGC adjusting device, be used for calculating and keep watch on the AGC controlled quentity controlled variable, comprise band steel exports speed adjusting device 36 and live-roller speed adjusting device 37; Band steel exports speed adjusting device is used for basis
Figure BDA0000091731970000124
Calculate band steel exports speed regulation amount Δ V MonThe live-roller speed adjusting device is used for according to band steel exports speed regulation amount Δ V MonCalculate live-roller linear velocity regulated quantity Δ V R, mon
The first synchronous transfer model 28, the second synchronous transfer model 29, consecutive mean module 31, the first first-order lag device 33, the 3rd synchronous transfer model 34, the second first-order lag device 35 are corresponding with the first synchronous transfer model, the second synchronous transfer model, consecutive mean module, the first first-order lag device, the 3rd synchronous transfer model, the second first-order lag device that the high accuracy belt steel rolling is kept watch in the automatic thickness control method respectively.
More than combine embodiment that the present invention has been done concrete description, but not as qualification of the present invention, all modifications and variations in claim scope of the present invention all drop on and receive within protection scope of the present invention.

Claims (3)

1. consecutive mean module, it is characterized in that: the consecutive mean module comprises the 4th to the 6th shift register and judging treatmenting module, and the monitoring segment length is divided into plurality of sections, and hop count is determined that by strip speed every section length is L S=V * T S, L in the formula SBe every section length, V is band steel transmission speed, T SBe the sampling time;
The 4th shift register is used to store the sampled value of the input variable that needs dynamic calculation mean value, and is every through a sampling time T SAll elements in the register is moved one successively backward, then the input sample value of current time is deposited in first position of this register;
The 5th shift register is used to store the weights of input sample value, and weights determine that by strip speed the weights of each input sample value are W X=V * T S/ L Mon, in the formula, L MonFor keeping watch on the length of section; Whenever, the weights of all storages in the register are moved one successively backward through a sampling time, then the weights of current time input variable are deposited in first position of this register;
The weights that the 6th shift register is used to store the input sample value with; All weights additions that are about to corresponding times prior in the 5th shift register obtain; Each sampling instant is with the weights of all storages and move one backward successively; Then the weights of current time input variable are stored in first position of this register, promptly first element of the 6th shift register and the 5th shift register always equates; Three shift registers are corresponding one by one;
Judging treatmenting module is used for each sampling time and judges that the element more than or equal to 1.0 appears in the 6th shift register at first; If i element more than or equal to 1.0 i-1 element less than 1.0; Then i element of the 5th shift register deducted that i element surpasses 1.0 amount in the 6th shift register, with preceding i element guaranteeing the 5th shift register and be 1.0 just in time; Then preceding i element of the 4th shift register multiply by preceding i element value of corresponding the 5th shift register, and summation obtains keeping watch on thickness error at the mean value of keeping watch on the segment length.
2. consecutive mean method, it is characterized in that: it may further comprise the steps:
The sampled value of the input variable of step 1, input strip speed V, supervision segment length L and need dynamic calculation mean value;
Step 2, the storage of the 4th shift register need the sampled value of the input variable of dynamic calculation mean value, and be every through a sampling time T SAll elements in the register is moved one successively backward, then the input sample value of current time is deposited in first position of this register;
Step 3, the 5th shift register are used for storing the weights of input sample value, and weights determine that by strip speed the weights of each input variable are W X=V * T S/ L Mon, in the formula, L MonFor keeping watch on the length of section; Whenever, the weights of all storages in the register are moved one successively backward through a sampling time, then the weights of current time input variable are deposited in first position of this register;
Step 4, the 6th shift register be used for storing the weights of keeping watch on the thickness error sampled value and; All weights additions that are about to corresponding times prior in the 5th shift register obtain; Each sampling instant is with the weights of all storages and move one backward successively; Then the weights of current time input variable are stored in first position of this register, promptly first element of the 6th shift register and the 5th shift register always equates; Three registers are corresponding one by one;
Step 5, judging treatmenting module are used for each sampling time and judge that the element more than or equal to 1.0 appears in the 6th shift register at first; If i element more than or equal to 1.0 i-1 element less than 1.0; Then i element of the 5th shift register deducted that i element surpasses 1.0 amount in the 6th shift register, with preceding i element guaranteeing the 5th shift register and be 1.0 just in time; Then preceding i element of the 4th shift register multiply by preceding i element value of corresponding the 5th shift register, and summation obtains keeping watch on thickness error at the mean value of keeping watch on the segment length.
3. consecutive mean method according to claim 2 is characterized in that: it also comprises step 6, the maximum memory space C of each shift register is set, and the weights W of current time is all judged in each sampling instant XWhether more than or equal to minimum weights W Min=1.0/C, and if less than minimum weights three shift register of this sampling instant all do not upgrade, and only with weights summations, judge weights and whether more than or equal to minimum weights in next sampling instant then; Just carry out the renewal of three shift registers more than or equal to minimum weights up to weights and value.
CN 201110274939 2011-09-16 2011-09-16 Dynamic mean module and method thereof Expired - Fee Related CN102366757B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN 201110274939 CN102366757B (en) 2011-09-16 2011-09-16 Dynamic mean module and method thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN 201110274939 CN102366757B (en) 2011-09-16 2011-09-16 Dynamic mean module and method thereof

Publications (2)

Publication Number Publication Date
CN102366757A true CN102366757A (en) 2012-03-07
CN102366757B CN102366757B (en) 2013-07-17

Family

ID=45759370

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 201110274939 Expired - Fee Related CN102366757B (en) 2011-09-16 2011-09-16 Dynamic mean module and method thereof

Country Status (1)

Country Link
CN (1) CN102366757B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111360081A (en) * 2020-03-16 2020-07-03 沈阳工业大学 Five-stand cold continuous rolling mill speed setting method based on global second flow

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH105834A (en) * 1996-06-14 1998-01-13 Kobe Steel Ltd Controller for coarse split cut in rolling equipment for bar steel
JP3569502B2 (en) * 2001-06-14 2004-09-22 東芝三菱電機産業システム株式会社 Thickness control device for continuous rolling mill
CN101147918A (en) * 2006-09-20 2008-03-26 宝山钢铁股份有限公司 Control method for using forward slip adaptive dynamic correction strip steel thickness deviation
CN101623707A (en) * 2009-08-05 2010-01-13 燕山大学 Plate-shape standard curve calculating method of continuous rolling machine
CN101716606A (en) * 2009-12-31 2010-06-02 中冶南方(武汉)自动化有限公司 Rolling mill AGC control system
CN101780480A (en) * 2009-01-16 2010-07-21 宝山钢铁股份有限公司 Thickness control compensation method of strip steel head part when welding line passes through rack

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH105834A (en) * 1996-06-14 1998-01-13 Kobe Steel Ltd Controller for coarse split cut in rolling equipment for bar steel
JP3569502B2 (en) * 2001-06-14 2004-09-22 東芝三菱電機産業システム株式会社 Thickness control device for continuous rolling mill
CN101147918A (en) * 2006-09-20 2008-03-26 宝山钢铁股份有限公司 Control method for using forward slip adaptive dynamic correction strip steel thickness deviation
CN101780480A (en) * 2009-01-16 2010-07-21 宝山钢铁股份有限公司 Thickness control compensation method of strip steel head part when welding line passes through rack
CN101623707A (en) * 2009-08-05 2010-01-13 燕山大学 Plate-shape standard curve calculating method of continuous rolling machine
CN101716606A (en) * 2009-12-31 2010-06-02 中冶南方(武汉)自动化有限公司 Rolling mill AGC control system

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111360081A (en) * 2020-03-16 2020-07-03 沈阳工业大学 Five-stand cold continuous rolling mill speed setting method based on global second flow

Also Published As

Publication number Publication date
CN102366757B (en) 2013-07-17

Similar Documents

Publication Publication Date Title
CN102343365B (en) Method and system for automatic thickness control over high-precision strip steel rolling under monitoring
CN202270720U (en) monitor automatic gauge control equipment for high-precision strip steel rolling
CN102380515B (en) Synchronous transport model and method thereof
CN101934292B (en) Automatic control method for camber and wedge of hot rolled strip roughing mill
CN100369683C (en) Method for automatic controlling thickness in fast high precision plate strip rolling process
CN101618402A (en) Method for controlling planeness of cold-rolling strip steel
CN102641904B (en) Energy consumption forecasting device
CN101600521A (en) Driving is batched or method and the corresponding control device and the devices for taking-up of the devices for taking-up of debatching metal tape
CN102029292A (en) Strip steel thickness feedforward control method based on mechanical property detection
CN101480664B (en) Plate thickness control system, device and method
CN104942019A (en) Automatic control method for width of steel strips during cold rolling
CN100515592C (en) Rolling mill control device and method
KR20110137313A (en) Determining elastic modulus for continuous material web
CN201950071U (en) High precision automatic thickness control device based on thickness meter segment monitoring
CN102553923A (en) Control method for self-adaption compensation of elongation of planisher
CN104338757A (en) Method for controlling rolling force of rolling mill at rolling start state
CN102172637A (en) High-accuracy automatic thickness control method and equipment based on thickness gauge subsection monitoring
KR20160111477A (en) Energy consumption predicting device for rolling line
CN103240279B (en) The control device of continuous hot-rolling mill and the control method of continuous hot-rolling mill
CN102366757B (en) Dynamic mean module and method thereof
CN1050783C (en) Control for constant tension of heat-belt continuous rolling machine and method for counting continuous-rolling thickness and section area
CN111036685A (en) Monitoring thickness control method based on improved estimation compensation
CN105583236B (en) The online acquisition methods of cold-strip steel plastic coefficient
CN102366760B (en) Method for online decoupling through adjustment and control efficacy coefficients of cold-rolled steel strip plate-shape control actuator
CN201632488U (en) Rolling mill AGC control system

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
CF01 Termination of patent right due to non-payment of annual fee
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20130717

Termination date: 20170916