CN104881002B - Optimization system - Google Patents
Optimization system Download PDFInfo
- Publication number
- CN104881002B CN104881002B CN201510089796.7A CN201510089796A CN104881002B CN 104881002 B CN104881002 B CN 104881002B CN 201510089796 A CN201510089796 A CN 201510089796A CN 104881002 B CN104881002 B CN 104881002B
- Authority
- CN
- China
- Prior art keywords
- control
- state
- desired value
- optimization
- opportunity loss
- 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.)
- Expired - Fee Related
Links
- 238000005457 optimization Methods 0.000 title claims abstract description 107
- 238000012360 testing method Methods 0.000 claims abstract description 38
- 238000012937 correction Methods 0.000 claims abstract description 28
- 230000007704 transition Effects 0.000 claims abstract description 10
- 238000000034 method Methods 0.000 description 28
- 238000011156 evaluation Methods 0.000 description 25
- 230000008569 process Effects 0.000 description 19
- 230000008859 change Effects 0.000 description 12
- 230000004044 response Effects 0.000 description 8
- 230000000694 effects Effects 0.000 description 5
- 239000000203 mixture Substances 0.000 description 5
- 230000009467 reduction Effects 0.000 description 5
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 3
- 238000001514 detection method Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 208000037805 labour Diseases 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000002195 synergetic effect Effects 0.000 description 2
- 241000406668 Loxodonta cyclotis Species 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000000178 monomer Substances 0.000 description 1
- 238000011105 stabilization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Manufacturing & Machinery (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Feedback Control In General (AREA)
- General Factory Administration (AREA)
Abstract
The optimization system of the present invention is improved without complicated setting with regard to optimality when can make to be in the transition state before reaching stable state.Have：The multiple control units (41) being controlled to control object (5)；It will be optimized for the variable of the control of control unit and calculate the optimization portion (21) of the desired value of the variable；To whether being to optimize the opportunity loss test section (31) that is detected of opportunity loss state, opportunity loss state representation is optimized under the transition state to before stable state, compared with the desired value calculated based on optimization portion performs the control of control unit, the state of a control of control object can be more improved to the state of optimal state；With desired value correction portion (32), in the case where opportunity loss test section is detected to optimize opportunity loss state, desired value is modified so that state of a control is more nearly optimal state, control unit is based on being controlled control object by the modified desired value of desired value correction portion.
Description
Technical field
The present invention relates to a kind of optimization system.
Background technology
Have been provided with when monitoring control is carried out to the state of production process, by the way that process optimizing is improved production
The system of efficiency (for example, referring to following patent document 1~3).In following patent document 3, disclose to process into
The upper side of multiple local controllers of row control is configured with the optimization of the global optimizer of the optimization of carry out process entirety
System.In the optimization system, the shape of process necessary to global optimizer from local controller, process collection optimization
The information such as state, limitation, are considering with the basis of the cooperation of other local controllers, carrying out for the optimal of process entirety
Change, desired value is sent to each local controller.Each local controller is while consider the scope being responsible for for own controller
The limitation of interior variable, while being controlled such that these variables close to the desired value received from global optimizer.
Prior art literature
Patent document
Patent document 1：Japanese Unexamined Patent Publication 2005-092584 publications
Patent document 2：Japanese Unexamined Patent Application Publication 2002-534729 publications
Patent document 3：No. 612255 specifications of U.S. Patent No.
The content of the invention
Problems to be solved by the invention
Described optimization is for stable state (changing substantially certain state in time) in patent document 3
The optimization that variable carries out.Therefore, can probably be reduced in optimality when reaching the transition state before stable state.Reference
Fig. 6~Fig. 9, is specifically described below.
Fig. 6 shows the optimization system with global optimizer 7 and two local controllers 8,9.Within the system,
The controlled quentity controlled variable CV8 of local controller 8 is input to local controller 9 as control variable CV8, and local controller 9 is based on itself
The operating quantity MV9 and control variable CV8 output control amounts CV9 of controller.In local controller 9, since local can not be operated
The control variable CV8 of controller 8, so control variable CV8 is taken as disturbance variable DV9.
Here, (A), (B) with reference to Fig. 7, how (dry according to operating quantity MV9 or control variable CV8 to controlled quentity controlled variable CV9
Disturb variables D V9) change and change and illustrate.(A) of Fig. 7 is to show to add operating quantity MV9 unit quantities at certain moment
In the case of the figures that how to change in time of controlled quentity controlled variable CV9.(B) of Fig. 7 is to show to make control variable CV8 at certain above-mentioned moment
The figure that how controlled quentity controlled variable CV9 changes in time in the case that (disturbance variable DV9) unit quantity adds.
(A), (B) of Fig. 7 shows the change of the response controlled quentity controlled variable CV9 of to(for) control variable CV8 (disturbance variable DV9)
In the case of the speed ratio of the response speed controlled quentity controlled variable CV9 of change of to(for) operating quantity MV9 is slow.Also, also showing, there are unit quantity
Controlled quentity controlled variable CV9 also increased relation when increasing operating quantity MV9 or control variable CV8 (disturbance variable DV9).
In optimization system in this way, setting optimizes the evaluation function of computing so that controlled quentity controlled variable CV8 and control
In the case of amount CV9 both sides processed are maximized, the result of the optimization based on global optimizer 7 is for example as shown in Figure 8.
(A) of Fig. 8 is shown according to evaluation function sets target value so that the controlled quentity controlled variable CV8 of local controller 8 is i.e. dry
Variables D V9 is disturbed maximumlly to scheme.(B) of Fig. 8 is shown to opposite with the increased directions of controlled quentity controlled variable CV9 of local controller 9
Reduce the figure of the desired value of the operating quantity MV9 of direction setting local controller 9.If increasing operating quantity MV9, due to becoming with interference
The increased synergistic effect of DV9 is measured, controlled quentity controlled variable CV9 exceedes upper limit value, so to the reduction direction setting opposite with increasing direction
The desired value of operating quantity MV9.
(C) of Fig. 8 is shown by the desired value of the setting operation amount MV9 as shown in Fig. 8 (B) so as to by controlled quentity controlled variable
CV9 controls the figure in the situation of the scope no more than the upper limit.
Such as with the optimization system described in patent document 3 come if carrying out such optimized control, such as Fig. 9
(B) shown in, the operating quantity MV9 of local controller 9 declined towards desired value moment.In contrast, the controlled quentity controlled variable of local controller 8
Shown in (A) of CV8, that is, disturbance variable DV9 such as Fig. 9, rise gradually towards desired value.This is because, the response speed of controlled quentity controlled variable CV8
Degree is slower than the response speed of operating quantity MV9, and the controlled quentity controlled variable CV8 of local controller 8 is being input to other local controls
Time etc. is also needed to before device 9.
In this way, due to compared with elevated controlled quentity controlled variable CV8 (disturbance variable DV9) gradually, operating quantity MV9 relatively early undergrounds
Drop, so as shown in Fig. 9 (C), the controlled quentity controlled variable CV9 of local controller 9 is temporarily to the direction change away from desired value, then towards mesh
Scale value is close.Although such state, which becomes, can more improve controlled quentity controlled variable CV9 to optimal state but not improve to optimal
State state, it can be said that be the chance that have lost optimization state (it is following, also referred to as " optimize opportunity loss
State " is only called " opportunity loss state ".).
In (C) of Fig. 9, by according to the contribution a of disturbance variable DV9 in scopes of the controlled quentity controlled variable CV9 not less than upper limit value e
The track d that inside changes operating quantity MV9 and realize, with the contribution b according to the contribution a and operating quantity MV9 of disturbance variable DV9 and
The difference of the track c of obtained controlled quentity controlled variable CV9 reality is opportunity loss f.
Under the situation of (C) of such Fig. 9, for example, by being controlled such that operating quantity MV9 did not had under moment
Drop, but controlled quentity controlled variable CV9 is improved according to the contribution a of disturbance variable DV9, dropped in stages of the PREDICTIVE CONTROL amount CV9 more than upper limit value e
Low operating quantity MV9, then can improve controlled quentity controlled variable CV9 to the state more optimized.
That is, using the optimization system described in patent document 3, it is assumed that can cause to optimize opportunity loss
State, then optimality during transition state before reach stable state may reduce.
It is an object of the invention to provide a kind of optimalizing control system, which sets without complicated
It is fixed can just make especially to reach stable state before transition state when optimality improve.
Means for solving the problems
Optimization system according to the present invention, it is characterised in that have：Multiple control units, the multiple control unit pair
Control object is controlled；Optimization portion, the optimization portion by by the variable of the control of the control unit optimize and based on
Calculate the desired value of the variable；Opportunity loss test section, whether the opportunity loss test section is to being to optimize opportunity loss shape
State is detected, it is described optimization opportunity loss state representation, under the transition state to before stable state, with based on
The control that the desired value that the optimization portion is calculated performs the control unit is compared, can be by the control of the control object
State processed is more improved to the state of optimal state；And desired value correction portion, the desired value correction portion is in the chance
In the case that loss test section is detected as the optimization opportunity loss state, the desired value is modified so that institute
State state of a control and be more nearly optimal state, the control unit is based on the target that have modified by the desired value correction portion
Value is controlled the control object.
The opportunity loss test section can also have the second optimization portion, and the second optimization portion will be used for the control
The variable of the control in portion processed optimizes and calculates the second desired value of the variable, and the opportunity loss test section is to described optimal
Whether the desired value and second desired value that change portion is calculated are compared, to being the optimization opportunity loss state
It is detected.
On the basis of the positive negative direction relative to the value as benchmark, in the control direction based on the desired value and it is based on
In the case that the control direction of second desired value is in opposite direction, the opportunity loss test section can be determined that to be described
Optimize opportunity loss state.
The opportunity loss test section can be in the big situation of the relational threshold value of the desired value and second desired value
Under, it is determined as the optimization opportunity loss state.
The effect of invention
In accordance with the invention it is possible to provide it is a kind of can just make especially to reach stable state without complicated setting before
The optimization system that optimality during transition state improves.
Brief description of the drawings
(A) of Fig. 1 is the figure of the step response for the controlled quentity controlled variable CV for illustrating the operating quantity MV for local controller, and (B) is
Illustrate the figure of the step response of the disturbance variable DV of controlled quentity controlled variable CV to(for) local controller.
(A) of Fig. 2 is the figure for the desired value for illustrating disturbance variable DV, and (B) is the figure of the desired value of exemplified operation amount MV,
(C) it is the global (グ ロ ー バ Le that illustrates controlled quentity controlled variable CV) desired value GV and stable objects value SV be set to opposite direction
Figure.
Fig. 3 is the figure for the composition for illustrating the optimization system in first embodiment.
Fig. 4 is the figure for the composition for illustrating the optimization system in the second embodiment.
Fig. 5 is the figure for the composition for illustrating the optimization system in the 3rd embodiment.
Fig. 6 is the figure for the composition for illustrating optimization system.
(A) of Fig. 7 is the figure of the step response for the controlled quentity controlled variable CV for illustrating the operating quantity MV for local controller, and (B) is
Illustrate the figure of the step response of the disturbance variable DV of controlled quentity controlled variable CV to(for) local controller.
(A) of Fig. 8 is the figure of the desired value for the disturbance variable DV for illustrating optimization system, and (B) is to illustrate optimization system
Operating quantity MV desired value figure, (C) be illustrate optimization system controlled quentity controlled variable CV state of a control figure.
(A) of Fig. 9 is the figure of the state of a control for the disturbance variable DV for illustrating optimization system, and (B) is to illustrate to optimize system
The figure of the state of a control of the operating quantity MV of system, (C) are the figures of the state of a control for the controlled quentity controlled variable CV for illustrating optimization system.
Embodiment
[the principle of the present invention]
Before the embodiment of the explanation present invention, the principle of the present invention is illustrated.The present invention can be comprising complete
The optimization system of office's optimizer and local controller is realized.The principle of the present invention is, in such optimization system, in advance
The situation for reaching above-mentioned optimization opportunity loss state is detected, is detecting above-mentioned optimization opportunity loss state
In the case of, it is controlled to suppress the opportunity loss optimized.
In the optimization system, illustratively, calculated and be used for by both global optimizer and local controller
The desired value of each variable of the control of local controller.Hereinafter, the desired value calculated by global optimizer is known as " global mesh
Scale value ", is known as " stable objects value " by the desired value calculated by local controller.
Here, opportunity loss is to be changed into the meaning with the optimization of global optimizer in stable objects value in an interim state
Scheme caused by opposite value.As one of method for avoiding such opportunity loss, it is contemplated that have and calculated using local controller
Go out the method that stable objects value is lost to avoid chance of occurrence.But using such method, it is necessary to which research in advance may
The pattern of the opportunity loss of generation, in order to avoid chance of occurrence loses and studies intensively the logic of local controller, setting.For so
Study intensively, even the technical staff of processes known is also required to spend substantial amounts of labour in research.Further, due to optimal
The applicable object of change system is mostly extensive and complicated process, so studying intensively itself with regard to extremely difficult as carrying out.This hair
Bright calculate is predicted to be by the actual stable objects value calculated of local controller, to the stable objects value and global object value into
Row is relatively to detect opportunity loss state, so as to suppress to opportunity loss.Thus, it is not necessary to spend the thing of a large amount of labours
Opportunity loss can just be suppressed by first studying.
The detection of opportunity loss state is carried out by comparing global object value and stable objects value.Specifically,
Whether phase on the basis of the positive negative direction of the currency relative to variable is set to according to global object value and stable objects value
Opposite direction, to have determined whether opportunity loss state.
For example, the global object value in the controlled quentity controlled variable for local controller is set bigger than currency, stable objects
In the case that value is set smaller than currency, global object value and stable objects value are just set to opposite direction, thus sentence
Surely opportunity loss state has been arrived.
On the other hand, it is set bigger than currency in the global object value of the controlled quentity controlled variable for local controller, stablizes
In the case that desired value is set also big than currency, then global object value and stable objects value are just set to phase Tongfang
To thus judging not arriving opportunity loss state.
In addition, the detection of opportunity loss state is not limited to the control direction based on global object value and stable objects value
To be judged, as long as global object value and stable objects value can be compared using some benchmark to carry out opportunity loss state
Detection.For example, it can be judged based on the size of global object value and the difference of stable objects value.In such case
Under, when the relational threshold value of global object value and stable objects value is big, then it can be determined that to have arrived opportunity loss state, global object
When the difference of value and stable objects value is below threshold value, then it can be determined that and do not arrive opportunity loss state.
In the case where having detected opportunity loss state, stable objects value is corrected to suppress opportunity loss.Specifically,
The stable objects value for the variable for being judged as opportunity loss state is corrected so that it is close to global object value.With reference to Fig. 1 with
And Fig. 2 is concretely demonstrated.(A) of Fig. 1 is to show to make at certain moment the operating quantity MV unit quantities of local controller to add
In the case of, the controlled quentity controlled variable CV of local controller is increasing the figure of direction instantaneous variation.(B) of Fig. 1 was shown in the above-mentioned some time
Carve make local controller disturbance variable DV unit quantities add in the case of, the controlled quentity controlled variable CV of local controller is in increase direction
On the figure that changes gradually.As disturbance variable DV, the e.g. controlled quentity controlled variable from the output of other local controllers.
In such optimization system, the evaluation function that optimizes computing is being set so that controlled quentity controlled variable CV and dry
Disturb variables D V both of which it is maximized in the case of, the result of the optimization of global optimizer is for example as shown in Figure 2.
(A) of Fig. 2 is shown according to evaluation function sets target value so that disturbance variable DV maximumlly schemes.Fig. 2's
(B) it is that the desired value for the operating quantity MV for showing local controller is set to the increased directions of controlled quentity controlled variable CV with local controller
The figure in opposite reduction direction.If increasing operating quantity MV, due to the increased synergistic effect with disturbance variable DV, controlled quentity controlled variable CV
More than upper limit value, so the desired value to the reduction direction setting operating quantity MV opposite with increase direction.
(C) of Fig. 2 is that the global object value GV and stable objects value SV for showing controlled quentity controlled variable CV are set to opposite direction
Figure.That is, in the present invention, in this case, opportunity loss state is judged as, to the steady of controlled quentity controlled variable CV
Target value SV is modified so that it is close to global object value GV.
It is modified to as by the stable objects value SV for being judged as the controlled quentity controlled variable CV of opportunity loss state close to the overall situation
The means of desired value GV, enumerate for example, the used optimization computing when local controller determines stable objects value
Increase makes controlled quentity controlled variable CV make operating quantity MV close to global mesh close to the evaluation function of global object value or reduction in evaluation function
Weight of the evaluation function of scale value etc..Further, since as long as evaluation function can be corrected to become the variable with more dependency relation
Balance relative to each other, it is possible to the weight of a certain variable among the variable of raising or reduction with dependency relation,
The balance of the weight of two variables can also be changed.Also, the object of opportunity loss is not limited to controlled quentity controlled variable CV or operation
Measure MV.
Next, the embodiment of the present invention is illustrated referring to the drawings.But embodiment described below is
Illustrate, however not excluded that the following various modifications do not expressed or technology are applicable in.That is, the present invention can not depart from the scope of purport
It is interior to carry out various modifications to implement.
[first embodiment]
With reference to Fig. 3, the composition of the optimization system in first embodiment is illustrated.As shown in figure 3, optimize system
System 1 possesses：Global optimizer 2 with optimization portion 21, with opportunity loss test section 31 and desired value correction portion 32
Opportunity loss control device 3 and the local controller 4 with control unit 41.
The variable that optimization portion 21 will be used for the control that control unit 41 carries out control object (process) 5 optimizes, and calculates
Go out the global object value of the variable.The optimization portion 21 of the present invention is characterized in that, using the connection relation of local controller 4,
Carry out the optimization of process entirety can not being realized by the local controller 4 of monomer, being related to stable state.In addition, at this
In embodiment, illustrated using process (especially manufacturing process) as the illustration of control object, but control object
It's not limited to that, such as can be power generator, energy source device, heat resource equipment.
Specifically, optimization portion 21 (is become based on the evaluation function being set and as the value of the benchmark of each variable
The value of the origin of optimization) carry out system entirety optimization.As the value of the benchmark as each variable, such as mistake can be used
The currency of journey, stability forecast value (such as without the stable state in the case of control process predicted value).
As the gimmick of optimization, it is for instance possible to use known linear programming technique, quadratic programming, but as long as being true
Determine the gimmick of global object value, any gimmick can be used.
Optimization portion 21 is selected to pair for 4 sets target value of local controller from the global object value calculated
The variable of elephant is (hereinafter also referred to as " desired value setting object variable ".).Desired value setting object variable for example can be all
The combination or all control of a part for a part and controlled quentity controlled variable CV of operating quantity MV or operating quantity MV
Measure CV.It is particularly preferred to select fully necessary desired value to set object variable so that the close global mesh of all variables
Scale value.This is because if beyond necessarily increase desired value setting object variable, control, which may occur, to be become not
The shortcomings of stablizing.
Whether opportunity loss test section 31 is to being to optimize opportunity loss state to be detected.Optimize opportunity loss state
Refer to, under the transition state to before stable state, with performing control unit 41 according to the desired value calculated by optimization portion 21
Control compare, the state of a control of control object can be further improved to the state of optimal state.
Opportunity loss test section 31 also has the second optimization portion 311, which will be used for control unit 41
The variable of control optimize, calculate the stable objects value of the variable.
The global object value and stable objects value that opportunity loss test section 31 calculates optimization portion 21 are compared, right
Whether it is to optimize opportunity loss state to be detected.
Illustratively, on the basis of the positive negative direction relative to currency, in the control direction based on global object value and
In the case that control direction based on stable objects value is opposite direction, opportunity loss test section 31 is judged to optimizing chance damage
Mistake state.
In addition, opportunity loss test section 31 can be in the big situation of the relational threshold value of global object value and stable objects value
Under, it is judged as optimizing opportunity loss state.
Hereinafter, the function of opportunity loss test section 31 is specifically described.
As long as desired value setting object variable is set to be damaged close to optimization as global object value in chance
The method that mistake test section 31 calculates stable objects value can be any method, it is for instance possible to use by following formula (1) Suo Shi
The difference of such global object value and stable objects value square as evaluation function come the quadratic programming that is minimized.
【Numerical expression 1】
... formula (1)
In formula (1), y is the stable objects value of controlled quentity controlled variable CV, yiIt is stable objects value, the y of i-th of controlled quentity controlled variable CVG iIt is
The global object value of i-th of controlled quentity controlled variable CV, u are the stable objects values of operating quantity MV, ujIt is the stable objects of j-th of operating quantity MV
Value, uG jIt is the global object value of j-th of operating quantity MV, q(y) iIt is the weight for the evaluation function of i-th of controlled quentity controlled variable CV, q(u) j
It is the weight for the evaluation function of j-th of operating quantity MV, Ty、TuIt is the desired value setting pair of controlled quentity controlled variable CV, operating quantity MV respectively
As the index set of variable.Weight can the balance based on the evaluation function in global optimizer 2 by global optimizer 2 come certainly
It is fixed, it can also preset.
At this time, the restrictive condition represented by following formula (2) is endowed between y and u.
Y-y0=G (u-u0) ... formula (2)
In formula (2), G is the matrix (M is that the number of controlled quentity controlled variable CV, N are the numbers of operating quantity MV) of M rows N row, is to represent to control
The change of CV is measured relative to the matrix of the ratio of the variable quantity of operating quantity MV.y0、u0It is the origin optimized, although can assign
Arbitrary operating point, but can also utilize the currency of process, the influence such as interference for considering process stability forecast value (such as
Without the predicted value of the process of the stable state in the case of control) etc..Also, formula (2) is to set up in linear system, but
Comprising in nonlinear system, by around origin carry out linear approximation etc. can formula (2) set up.
Also, for all or a part of controlled quentity controlled variable CV and operating quantity MV, following formula (3), (4) institute are assigned sometimes
Show such upper limit value, lower limit.
【Numerical expression 2】
yLO i≤yi≤yHI i... formula (3)
uLO j≤uj≤uHI j... formula (4)
In formula (3), (4), yHI i、yLO iIt is yiUpper limit value, lower limit, uHI j、uLO jIt is ujUpper limit value, lower limit.
Optimize in computing, evaluation function is minimized so that y, u are no more than these upper limit values, lower limit.It is no more than not finding
In the case of these upper limit values, the solution of lower limit, upper limit value can be relaxed, a part for lower limit solves.
The optimization computing can individually be carried out in units of local controller 4, can also be with system overall time
Ground, which is concentrated, to carry out.It is different from the optimization in optimization portion 21 in the case where integrally being carried out with system, carry out without considering local control
The optimization computing of connection relation between device 4 processed.The result of the optimization computing is to obtain desired value setting object variable
Value under the stable state of all variables comprising stable objects value.For example with the process values of certain variable or stability forecast value
On the basis of, in the case which comes across the opposite direction of global object value, then it is determined as that the variable has arrived opportunity loss state.
Specifically, the value of each variable under stable state is being set to yT=yT 1、yT 2、···、yT M、uT=uT 1、
uT 2、···、uT NWhen, chance damage has been judged by the controlled quentity controlled variable CV of formula (5) establishment and the operating quantity MV of formula (6) establishment
Mistake state.
(yT i- y(0) i)×(yG i- y(0) i) ＜ 0 ... formulas (5)
(uT j- u(0) j)×(uG j- u(0) j) ＜ 0 ... formulas (6)
In formula (5), (6), y(0) i、u(0) jBe i-th of controlled quentity controlled variable CV, j-th operating quantity MV optimization origin
Value.
In addition, in the difference based on global object value and stable objects value, to whether having arrived opportunity loss state and having judged
In the case of, for example, can judge machine by the controlled quentity controlled variable CV and operating quantity MV of formula (7), (8) establishment with given threshold S
State can be lost.
(yT i- yG i)2＞ S ... formulas (7)
(uT j- uG j)2＞ S ... formulas (8)
, can be to one each variable, each local controller 4 or system set overall value on threshold value S.
Threshold value S may not be constant, can also make threshold value according to the weight such as evaluation function, the distance of the origin of self-stabilization desired value
S changes.Also, above-mentioned multiple decision methods can also be combined by forms such as "or", "AND" or most decisions.
Desired value correction portion 32 shown in Fig. 3 is detected in opportunity loss test section 31 to optimize opportunity loss state
In the case of, desired value is modified so that state of a control is closer to optimal state.
Desired value correction portion 32 performs again on the basis of it have modified problem to be used to calculate performed by local controller 4
The optimization computing of stable objects value, thus redefines stable objects value.It is specifically described below.
Initially, to being detected the evaluation function for the variable of opportunity loss state and with the related variable of the variable
It is modified.For example, in the y of corresponding Mr. Yu's controlled quentity controlled variable CVkIn the case of being detected as opportunity loss state, by related yk's
Evaluation function is set to formula (9).
V(y) k(yk)=q(y) k(yk- yG k)2... formula (9)
Formula (9) is meant, in ykEvaluation function is not changed in the case of setting object variable for desired value, in ykIt is not
Again desired value setting object variable is set in the case of desired value setting object variable.
Then, extract and ykRelated operating quantity MV.Specifically, known y is extractedkValue become by operation
The operating quantity MV of change.Such as corresponded to by extraction in the y with gain matrix GkThere is the behaviour of the row of the value beyond 0 in corresponding row
Work amount MV, can extract such operating quantity MV.As such operating quantity MV, such as u is being extractedkIn the case of, will
Its evaluation function is set to formula (10).
V(u) k(uk)=a × q(u) k(uk- uG k)2... formula (10)
In formula (10), q(u) kIt is the weight for optimizing the evaluation function in computing, a is to be for slacken evaluation function
Number (0 less than 1), it is endowed by setting.Can be that for example system is integrally assigned on the adding method of a
One a, for each local controller assign an a, for each variable assign an a the methods of any of.
The optimization computing of desired value correction portion 32 is to perform that modified evaluation function will be not required (to be not detected by chance damage
Lose and the evaluation function of the not related variable of variable with being detected opportunity loss) with correcting the evaluation function of object
With the optimization computing for being set to new evaluation function.
In the optimization computing of desired value correction portion 32, the optimization computing phase with opportunity loss test section 31 is endowed
Same restrictive condition, upper limit value and lower limit.Also, the optimization computing of desired value correction portion 32 can be with local controller 4
Carried out for unit, can also the concentration progress of system overall time.
On having been judged the variable of opportunity loss state by opportunity loss test section 31, pass through desired value correction portion 32
The stable objects value that calculates of optimization computing at least than being approached by the stable objects value that opportunity loss test section 31 calculates
The direction of global object value is calculated, and is more preferably calculated on the basis of the origin of optimization close to the direction of global object value
Go out.
Control unit 41 shown in Fig. 3 is based on by the modified stable objects value of desired value correction portion 32 to control object (mistake
Journey) 5 it is controlled.
As described above, the optimization system 1 in first embodiment, by that with optimization portion 21, will can use
Optimized in the variable of the control of control object (process), and calculate the desired value of variable, by being detected with opportunity loss
Whether portion 31, can be to being to optimize opportunity loss state to be detected, by that with desired value correction portion 32, can be detected
In the case of surveying to optimize opportunity loss state, desired value is modified so that state of a control is closer to optimal shape
State, by with control unit 41, can be controlled based on the desired value being corrected to control object (process).
Therefore, the optimization system 1 in first embodiment, can just improve especially without complicated setting
In optimality when reaching the transition state before stable state.
Also, in the optimization system 1 of first embodiment, can be by using the global optimizer 2 and sheet set
Ground controller 4, and newly set opportunity loss control device 3 to carry out construction systems, therefore can strongly suppress the overall situation for having set
The change of optimizer 2 and local controller 4.
[the second embodiment]
With reference to Fig. 4, the optimization system of the second embodiment is illustrated.It is right in above-mentioned first embodiment
Said using global optimizer 2, opportunity loss control device 3, local controller 4 to form the situation of optimization system 1
It is bright, but in the second embodiment, to using global optimizer 2 and local controller 4 come form the situation of optimization system 1 into
Row explanation.
The optimization system 1 of second embodiment and the difference of the optimization system 1 in first embodiment be, this
Ground controller 4 has opportunity loss test section 31 and mesh possessed by opportunity loss control device 3 in first embodiment
Scale value correction portion 32.
The opportunity loss test section 31 and desired value correction portion 32 of second embodiment are applied to own local control
The first embodiment of optimization computing of the optimization computing this point of device 4 with being applied to opportunity loss control device 3
Function is different, but other functions are identical with the function of first embodiment, so the description thereof will be omitted.
According to the optimization system 1 of the second embodiment, play identical with the optimization system 1 of first embodiment
Effect, and compared with the optimization system 1 of first embodiment, unwanted optimization computing can be omitted by having, can
Such effect is detected opportunity loss based on the newest process status for control.
[the 3rd embodiment]
With reference to Fig. 5, the optimization system of the 3rd embodiment is illustrated.It is right in the second above-mentioned embodiment
Local controller 4 have in first embodiment opportunity loss test section 31 possessed by opportunity loss control device 3 and
The situation of desired value correction portion 32 is illustrated, but in the 3rd embodiment, has to global optimizer 2 and implement first
The situation of opportunity loss test section 31 and desired value correction portion 32 possessed by opportunity loss control device 3 is said in form
It is bright.
The opportunity loss test section 31 and desired value correction portion 32 of 3rd embodiment are passing through opportunity loss test section
31 and desired value correction portion 32 come correct global object value that optimization portion 21 is calculated and export to local controller 4 this
From the opportunity loss test section 31 of first embodiment and the function of desired value correction portion 32 it is different on point, but it is other
Function is identical with the function of first embodiment, so the description thereof will be omitted.
According to the optimization system 1 of the 3rd embodiment, play identical with the optimization system 1 of first embodiment
Effect, and with effect as the local controller 4 set can be utilized same as before.
Symbol description
1 ... optimizes system
2 ... global optimizers
3 ... opportunity loss control devices
4 ... local controllers
5 ... control objects
21 ... optimization portions
31 ... opportunity loss test sections
32 ... desired value correction portions
41 ... control units
311 ... second optimization portions.
Claims (5)
1. a kind of optimization system, it is characterised in that have：
Multiple control units, the multiple control unit are controlled control object；
Optimization portion, the optimization portion will optimize for the variable of the control of the control unit and calculate the mesh of the variable
Scale value；
Whether opportunity loss test section, the opportunity loss test section are described to being to optimize opportunity loss state to be detected
Opportunity loss state representation is optimized, under the transition state to before stable state, and based on the institute of optimization portion
The control that the desired value calculated performs the control unit is compared, and can more improve the state of a control of the control object
To the state of optimal state；And
Desired value correction portion, the desired value correction portion are detected as the optimization chance damage in the opportunity loss test section
In the case of mistake state, the desired value is modified so that the state of a control is more nearly optimal state,
The control unit is controlled the control object based on the desired value that have modified by the desired value correction portion,
The opportunity loss test section also has the second optimization portion, and the second optimization portion is by the control for the control unit
The variable of system optimizes and calculates the second desired value of the variable,
On the basis of the positive negative direction relative to the value as benchmark, in the control direction based on the desired value and based on described
In the case that the control direction of second desired value is in opposite direction, the opportunity loss test section is determined as the optimization machine
State can be lost.
2. system is optimized as claimed in claim 1, it is characterised in that
The opportunity loss test section is determined as in the case where the relational threshold value of the desired value and second desired value is big
The optimization opportunity loss state.
3. system is optimized as claimed in claim 1 or 2, it is characterised in that
Possessing has first control device and multiple second control devices,
The first control device has the optimization portion,
The second control device has the control unit, the opportunity loss test section and desired value correction portion.
4. system is optimized as claimed in claim 1 or 2, it is characterised in that
Possessing has first control device and multiple second control devices,
The first control device has the optimization portion, the opportunity loss test section and desired value correction portion,
The second control device has the control unit.
5. system is optimized as claimed in claim 1 or 2, it is characterised in that
Possessing has first control device, second control device and multiple 3rd control devices,
First control device has the optimization portion,
Second control device has the opportunity loss test section and desired value correction portion,
3rd control device has the control unit.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2014-038437 | 2014-02-28 | ||
| JP2014038437A JP6199203B2 (en) | 2014-02-28 | 2014-02-28 | Optimization system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104881002A CN104881002A (en) | 2015-09-02 |
| CN104881002B true CN104881002B (en) | 2018-04-17 |
Family
ID=53948532
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510089796.7A Expired - Fee Related CN104881002B (en) | 2014-02-28 | 2015-02-27 | Optimization system |
Country Status (2)
| Country | Link |
|---|---|
| JP (1) | JP6199203B2 (en) |
| CN (1) | CN104881002B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7532137B2 (en) * | 2020-08-03 | 2024-08-13 | 株式会社東芝 | OPTIMUM VALUE SEARCH CONTROL DEVICE, OPTIMUM VALUE SEARCH CONTROL METHOD, COMPUTER PROGRAM, AND OPTIMUM VALUE SEARCH CONTROL SYSTEM |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07334207A (en) * | 1994-06-14 | 1995-12-22 | Hitachi Ltd | Method and system for controlling load of plant |
| US7596421B2 (en) * | 2005-06-21 | 2009-09-29 | Kabushik Kaisha Toshiba | Process control system, process control method, and method of manufacturing electronic apparatus |
| JP2007287063A (en) * | 2006-04-20 | 2007-11-01 | Hitachi Ltd | Optimum control method, optimum control system, supervisory control apparatus, and local control apparatus |
| WO2010097891A1 (en) * | 2009-02-24 | 2010-09-02 | 株式会社 東芝 | Plant optimum-operation control system |
| CN101887260A (en) * | 2010-06-30 | 2010-11-17 | 浙江大学 | Industrial process optimal control system and method for adaptive synchronization policy |
| JP5184597B2 (en) * | 2010-09-06 | 2013-04-17 | 株式会社日立産機システム | Communication control apparatus and communication / control system |
| CN102289223B (en) * | 2011-05-16 | 2014-02-19 | 河北省电力公司电力科学研究院 | Method for AVC (Automatic Voltage Control) system to verify power network global optimal control parameters |
-
2014
- 2014-02-28 JP JP2014038437A patent/JP6199203B2/en not_active Expired - Fee Related
-
2015
- 2015-02-27 CN CN201510089796.7A patent/CN104881002B/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| CN104881002A (en) | 2015-09-02 |
| JP6199203B2 (en) | 2017-09-20 |
| JP2015162189A (en) | 2015-09-07 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104935217B (en) | Improved deviation coupling control method suitable for multi-motor system | |
| CN108121215B (en) | Process control loops method of evaluating performance and device based on full loop reconstruct emulation | |
| CN104062901A (en) | Parameter optimization method for control system based on orthogonal optimization and particle swarm optimization method | |
| CN104881002B (en) | Optimization system | |
| CN109325613A (en) | With parallel sequential extreme learning machine to air extract on-line prediction method | |
| CN107402515A (en) | A kind of method of optimized synchronization generator excited system pid parameter | |
| CN106599451B (en) | Multi-objective optimization method for RV reducer main bearing | |
| CN104238624B (en) | A kind of tracking of maximum power point and device | |
| CN105574264A (en) | SVR soft measuring method for kiln tail decomposition rate of cement decomposing furnace | |
| CN104155876B (en) | The separation of a kind of PID controller realizes method | |
| CN104346501A (en) | Integrated optimization method and system for static model of fully-extending boom of crane | |
| CN104008253A (en) | Integrated optimization method of telescopic lifting arm dynamic model | |
| CN104008254A (en) | Integrated optimization method of telescopic lifting arm static model | |
| CN103513648B (en) | The performance test methods of control system and system | |
| CN103853045B (en) | Multivariable control apparatus and method | |
| CN105229265B (en) | System for the valve test compensation that feedovers | |
| CN103105869A (en) | Three-tank single-loop liquid level system | |
| Jia et al. | Research on the flow control strategy of water distributor in Water Injection Well | |
| CN109933937A (en) | A kind of damage measurement method cumulative based on circulation impairment value | |
| CN103970541A (en) | Online self-learning optimizing control method of vacuum salt production circulating water system | |
| CN113886915B (en) | Foundation bearing capacity reliability design method and device based on basic width critical value | |
| CN112366715B (en) | Method for determining transient voltage stability control performance of generator excitation system | |
| CN110609490A (en) | Converter pump simulation control system | |
| Cheng et al. | The design of expert PID parameter optimized by genetic algorithm | |
| CN103199545B (en) | Optimal secondary Gauss controller of dynamic reactive power compensation device and design method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| 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: 20180417 Termination date: 20210227 |