CN106227033A - A kind of pid control algorithm being suitable for multiple constraint target - Google Patents

A kind of pid control algorithm being suitable for multiple constraint target Download PDF

Info

Publication number
CN106227033A
CN106227033A CN201610610309.1A CN201610610309A CN106227033A CN 106227033 A CN106227033 A CN 106227033A CN 201610610309 A CN201610610309 A CN 201610610309A CN 106227033 A CN106227033 A CN 106227033A
Authority
CN
China
Prior art keywords
output valve
knots modification
moment
max
importance degree
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.)
Pending
Application number
CN201610610309.1A
Other languages
Chinese (zh)
Inventor
欧丹林
吴胜
梁逸敏
布莱恩·来恩斯
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.)
Zhejiang Bonyear Technology Co ltd
Original Assignee
Zhejiang Bonyear Technology Co 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 Zhejiang Bonyear Technology Co ltd filed Critical Zhejiang Bonyear Technology Co ltd
Priority to CN201610610309.1A priority Critical patent/CN106227033A/en
Publication of CN106227033A publication Critical patent/CN106227033A/en
Pending legal-status Critical Current

Links

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
    • G05B11/00—Automatic controllers
    • G05B11/01—Automatic controllers electric
    • G05B11/36—Automatic controllers electric with provision for obtaining particular characteristics, e.g. proportional, integral, differential
    • G05B11/42—Automatic controllers electric with provision for obtaining particular characteristics, e.g. proportional, integral, differential for obtaining a characteristic which is both proportional and time-dependent, e.g. P.I., P.I.D.

Abstract

The invention discloses a kind of pid control algorithm being suitable for multiple constraint target, comprise the steps: the first step, if the number of CV is 1, according to pid algorithm, calculate output valve knots modification corresponding for CV.If the number of CV is 0, then make output valve knots modificationIt is zero.Second step, according to pid algorithm, calculates the maximum constrained of output valve knots modification corresponding to each CCVMax, least commitmentmin.3rd step, compares step by step according to importance degree backward, uses following principle: if 1. importance degree rank is lowMore than more higher leveledMax, then useMax is as new；If 2. importance degree rank is lowLess than more higher leveledMin, then useMin is as new；If 3. importance degree rank is low Max andBetween min, then it is continuing withAs new one-level.4th step, by the output valve knots modification generated more afterwards, the final result as controller exports.

Description

A kind of pid control algorithm being suitable for multiple constraint target
Technical field
Patent of the present invention relates to automation field, is specifically related to a kind of New PID control algorithm, goes for processing The process of multiple constraint target.
Background technology
One typical control system, by controlled variable (Controlled Variable, CV), bound variable (Constrained Controlled Variable, CCV), manipulation variable (Manipulated Variable, MV), interference Variable (Disturbance Variable, DV) forms.
Controlled variable (CV) is the output of process variable, need to be controlled near certain desired value with improve technological operation or The variable of product quality performance.
Bound variable (CCV) is the output of process variable, run with guarantee equipment safety in the range of needing to be controlled in certain, Capacity of equipment limits or process optimization.
Handling variable (MV) is process input variable, can be conditioned to ensure controlled variable (CV) near desired value or Person's bound variable (CCV) in the range of.
Disturbance variable (DV) is process input variable, is uncontrollable during the course, but its change can be to controlled change Amount (CV) or bound variable (CCV) impact.Disturbance variable includes measurable disturbances variable and immeasurable disturbance variable, Qi Zhongke Survey disturbance variable and can improve control performance as in feedforward introducing control system, referred to as feed forward variable (Feed Forward Variable, FF).
The control system of application on process industry, based on closed loop feedback control.For realizing feedback control, can use Various control algolithm, such as PID control, fuzzy control, Multimode Control, Model Predictive Control etc..
Wherein, pid algorithm is one of most widely used control algolithm of automation field, with simple in construction, amount of calculation Little, robustness (resistance is transsexual) good and well-known.When realizing pid algorithm by computer, typically based on increment type, its base This computing formula is as follows:
(1)
In formula,
The output valve knots modification of controller.
E control deviation, i.e. desired value (SP) and the deviation of value of feedback (PV), e=SP-PV.
ekThe control deviation of current time (k moment).
ek-1The control deviation in a upper moment (k-1 moment).
ek-2The control deviation in upper upper moment (k-2 moment).
Gain gain.
The Ts sampling time.
Ti time of integration.
Td derivative time.
As can be seen from the above equation, basic pid algorithm is the process of a typical single-input single-output, only comprises one A CV and MV.Its regulation depends on the deviation of setting value and value of feedback, is therefore only applicable to set value calculation.For ease of retouching Stating, in this article, the control process of we this type of single mono-MV of CV is called first kind process.
Fig. 1 show pump outlet flow amount control system, is a simple example.This system only comprises a CV(and pumps out Mouth flow) and a MV(regulation valve), control system continual regulation valve position makes pump discharge flow be maintained at certain setting value.
Although in actual demand for control, first kind process account for great majority (about 60%), but still has some to compare More complicated demand.In addition to setting value, relatively conventional is scope control.It is said that in general, contain a CCV and one MV, the control process of we this kind of mono-MV of single CCV is called Equations of The Second Kind process.
It is illustrated in figure 2 pump reflux protection system, it is simply that a typical scope control.This system comprises a CCV(pump Rate of discharge), a MV(reflux inlet).After pump discharge flow is beyond ceiling value, valve should turn down the electricity saving pump as far as possible Consumption；And after pump discharge flow is less than low limit value, then to open big valve to ensure that pump work is in security interval.
In this example superincumbent, pump discharge flow cannot be below certain value, and this value is referred to as " under constraint by we Limit ".And for scope control, depend on different demands, have and only consider the monolateral of " constraint lower limit " or " the constraint upper limit " Control, also have and need to consider " constraint lower limit " and the polygon control of " the constraint upper limit " simultaneously.These are referred to as Equations of The Second Kind process, this kind of Process cannot process with classical pid algorithm.
Above-mentioned first kind process, Equations of The Second Kind process, be all single-input single-output.And in actual applications, also have more Complicated process, may often be such that multiple constraint target control, and this includes two big classes:
(1) system comprises a controlled variable (CV), one or more bound variable (CCV), handles variable (MV) for one.I Referred to as the 3rd class process.The control of one typical process such as device yield, on the premise of meeting security constraint, maintains In the yield set, and when running counter to security constraint, then must abandon the yield target set, and go preferential satisfied safety the most about Bundle.
(2) system does not comprise controlled variable (CV), only comprises one or more bound variable (CCV), handles variable for one (MV), we term it the 4th class process.One typical process, is the control that contains multiple security constraint of certain variable Journey, any one constraint is run counter to, and is required in the security constraint being adjusted being allowed to return to set.
For 3rd class, the 4th class process compare the first kind, Equations of The Second Kind process, more complicated, therefore cannot be with classics Pid algorithm processes.
Certainly, the most more complicated process, including multiple manipulation variablees (MV).Such complex process, major part is permissible By certain design, it is reduced to multiple containing a process handling variable (MV).But still have significant component of process without Method simplifies, and this part needs to use multivariable Control algorithm, such as Model Predictive Control, for this class process not in discussion herein Within the scope of.
Summary of the invention
As previously described, because the pid algorithm of classics can only process first kind process, and cannot process Equations of The Second Kind, the 3rd class, 4th class process.It is contemplated that propose a kind of new pid algorithm, go for processing multiple constraint target.
It is a feature of the present invention that be applicable to containing 1 handle variable (MV), 0 or 1 controlled variable (CV), 0 or The process of n bound variable (CCV).CV, CCV, according to the difference of its importance degree, distribute corresponding importance degree rank (rank), its The importance degree of middle CV is minimum.
Its calculation procedure is as follows:
The first step, if the number of CV is 1, according to pid algorithm, calculates output valve knots modification corresponding for CV.If CV's is individual Number is 0, then make output valve knots modificationIt is zero.
Second step, according to pid algorithm, calculates the maximum constrained of output valve knots modification corresponding to each CCVMax, Little constraintmin。
3rd step, compares step by step according to importance degree backward, and its basic principle is as follows:
If 1. importance degree rank is lowMore than more higher leveledMax, then useMax is as new；
If 2. importance degree rank is lowLess than more higher leveledMin, then useMin is as new；
If 3. importance degree rank is low?Max andBetween min, then it is continuing withAs new one-level。
4th step, by the output valve knots modification generated more afterwards, the final result as controller exports.
Preferred as one, increase " maximal increment (Max Increment is called for short Inc) ", " maximum decrement (Max Decrement, is called for short Dec) " two parameters, to output valve knots modificationRetrain.
Preferred as one, be the process of 0 for CV number, increase new parameter " cost (Cost) " calculate one excellent The output valve knots modification changedReplace 0 during the above-mentioned first step calculates, make MV move to optimum direction.
Certainly, the pid algorithm of this improvement, its ability still has certain restriction.Compare with Model Predictive Control Algorithm For, on the one hand, cannot be applicable to process the process of multiple MV；On the other hand, even single MV, binding side is being processed During boundary, it controls effect and Model Predictive Control also has a certain distance.
But, compared with the Model Predictive Control Algorithm that amount of calculation is the hugest, above-mentioned multiple constraint target pid algorithm Amount of calculation is the least, implements speed fast, and robustness is more preferable.Therefore, some and uncomplicated " gently application " side are being processed , there is unrivaled advantage in face, is particularly suitable for using in the embedded system of computing capability weakness relatively.
Accompanying drawing explanation
Fig. 1 is the control process example figure of first kind process (single mono-MV of CV)；
Fig. 2 is the control process example figure of Equations of The Second Kind process (single mono-MV of CCV).
Detailed description of the invention
For the pid control algorithm being suitable for multiple constraint target of the present invention, its main purpose is that structure is a kind of general In the first kind described in background technology, Equations of The Second Kind, the 3rd class, the novel pid algorithm of the 4th class process.
1. algorithm is suitable for the definition of process
For general definition, the treatable process of this algorithm, need to meet following condition:
1. variable (MV) is handled for 1
2. 0 or 1 controlled variable (CV)
3. 0 or n bound variable (CCV)
2. the sequence of importance degree
For CV, CCV, introducing sequence index levels, need to sort according to its significance level, wherein CV must be that importance degree is minimum 's.
The implementation method of sequence have a lot of in, a kind of sort method be described below:
1) assume that rank corresponding for CV less than 98, is then set to 99 by CCV.
2) according to demand for control, certain the integer respectively rank of other CCV being set between 1～98.
3) class value is the least, represents it the most important.
4) class value can not repeat, and does not the most allow two or more CCV, has identical rank.
5) class value can be adjacent, it is also possible to is non-conterminous.
If the method using other equivalences realizes similar ranking function, should be regarded as the covering scope of the present invention.
3. controller output valve knots modification corresponding for CVCalculating
If the number of CV is 1, then calculate output valve knots modification corresponding for CV according to pid algorithm；If the number of CV is 0, Then make output valve knots modificationIt is 0.
For the process that number is 1 of CV, output valve knots modification corresponding for CVCan be by formula (1) be launched meter Calculate and obtain,
In formula,
The output valve knots modification calculated by CV,
SPcv,k The desired value of current time (k moment) CV,
SPcv,k-1 The desired value of a upper moment (k-1 moment) CV,
SPcv,k-2 The desired value of a upper moment (k-1 moment) CV,
PVcv,k The value of feedback of current time (k moment) CV,
PVcv,k-1 The value of feedback of a upper moment (k-1 moment) CV,
PVcv,k-2 The value of feedback of upper upper moment (k-2 moment) CV,
GaincvThe gain of CV,
The Ts sampling time,
TicvThe time of integration of CV,
TdcvThe derivative time of CV.
It should be noted that pid algorithm has different ways of realization in order to realize different functions.In this article, as one Plant citing, only list a kind of general PID computing formula.If using other pid algorithms, also should be regarded as containing of the present invention Scope.
4. the maximum constrained of the knots modification of controller output valve corresponding for CCVWithCalculating
For the CCV that importance degree is r, its upper control limit is HLr, its lower control limit is LLr.It is single in view of a lot of CCV Limit retrains, and we define two Boolean quantities UseHLrAnd UseLLrWhether the constraint characterizing correspondence is used.If its value is True, then it represents that corresponding constraint limit is used;If its value is false, then it represents that corresponding constraint limit is not used by.
Then for the CCV that certain importance degree is r, the maximum constrained of the output valve knots modification of its correspondence is, Little it is constrained to, can be by calculated below and obtain.
；
；
IfWithSet up, and；
So,
；
IfWithSet up, and；
So,
；
IfWithSet up, and,
So,
；
IfWithSet up, and,
So,
。
5. numeric ratio is relatively
Calculating corresponding to CV, corresponding to all of CCV、Afterwards, inverse according to importance degree Sequence compares step by step, and its basic principle is as follows:
If 1. importance degree rank is lowMore than more higher leveledMax, then useMax is as new；
If 2. importance degree rank is lowLess than more higher leveledMin, then useMin is as new；
If 3. importance degree rank is low?Max andBetween min, then it is continuing withAs new one-level。
The method that " numeric ratio is relatively " realizes has a variety of, illustrates as one, listed herein a kind of relatively simple Computational methods.If the computational methods using other realize similar function, should be regarded as the scope that the present invention contains.
Carried out numeric ratio relatively after,Meet the optimum controller output under all of constraint premise.It can be straight Connect and directly send controller as result of calculation, it is also possible to carry out next step process again.
6. the constraint of controller output valve knots modification.
Preferred, for the output of MV is carried out regular as one, it is to avoid the fluctuation of CV moment process is stablized cause bright Aobvious impact, is specifically incorporated " maximal increment (Max Increment is called for short Inc) ", " maximum decrement (Max Decrement, letter Claim Dec) " two new parameters.The two parameter is all nonnegative real number, i.e. more than or equal to zero.A upper joint calculates and goes out, through following calculating, after calculating constraint。
If, then；
If, then；
Otherwise。
CalculateAs the result of calculation that controller is final, output is given and is handled executor corresponding to variable MV.
7. the optimization of economic benefit
For having the process of 0 CV, either Equations of The Second Kind process, or the 4th class process, all locate if all of CCV In restriction range, then MV being failure to actuate, i.e.It is zero.Although this can meet demand for security, but in a lot of times, may not It is most economical, there is certain optimization space.
For different processes, MV maximizes sometimes is optimum, and MV minimizes sometimes is optimum.To this end, we draw Enter a new parameter " cost (Cost) ", realize the excitation to MV, be allowed to move to optimum direction.
The definition of Cost and implementation method have a variety of, and a kind of easy to understand implementation is listed below.If adopted Realize the arousal effect being similar to by the method for other equivalences, should be regarded as the covering scope of the present invention.
If Cost is real number, its value scope is between-10.0 to 10.0, then controller exports the initial value of iterative computationCalculated by following program and go out,
If Cost > 0.0, then；
If Cost < 0.0, then；
Otherwise,。
The initial value that calculating is gone out, replace when in above-mentioned calculating Section 3, CV number is 0, it is possible to real The now excitation to MV.I.e. meeting on the premise of Constrained, move to optimum direction, and go beyond the scope at certain bound variable Time, can be adjusted in time again.

Claims (4)

1. the pid control algorithm being suitable for multiple constraint target, it is adaptable to handling variable (MV) containing 1,0 or 1 controlled Variable (CV), the process of 0 or n bound variable (CCV), CV, CCV, according to the difference of its importance degree, distribute the important of correspondence Degree rank, wherein the importance degree of CV is minimum, it is characterised in that: comprise the steps:
The first step, if the number of CV is 1, according to pid algorithm, calculates output valve knots modification corresponding for CV.If CV's is individual Number is 0, then make output valve knots modificationIt is zero.
Second step, according to pid algorithm, calculates the maximum constrained of output valve knots modification corresponding to each CCVMax, minimum are about Bundlemin。
3rd step, compares step by step according to importance degree backward, uses following principle:
If 1. importance degree rank is lowMore than more higher leveledMax, then useMax is as new；
If 2. importance degree rank is lowLess than more higher leveledMin, then useMin is as new；
If 3. importance degree rank is low?Max andBetween min, then it is continuing withAs new one-level。
4th step, by the output valve knots modification generated more afterwards, the final result as controller exports.
2. it is suitable for the pid control algorithm of multiple constraint target as claimed in claim 1, it is characterised in that: increase maximal increment (Max Increment, is called for short Inc), two parameters of maximum decrement (Max Decrement is called for short Dec), to output valve knots modification Retrain.
3. it is suitable for as claimed in claim 1 the pid control algorithm of multiple constraint target, it is characterised in that: it is the mistake of 0 for CV number Journey, increases new parameter " cost (Cost) ", calculates an output valve knots modification optimizedReplace above-mentioned calculating first In step 0, makes MV move to optimum direction.
4. it is suitable for the pid control algorithm of multiple constraint target as claimed in claim 1, it is characterised in that: the output valve that described CV is corresponding Knots modificationCalculated by pid algorithm and obtain:
In formula,
The output valve knots modification calculated by CV,
SPcv,k The desired value of current time (k moment) CV,
SPcv,k-1 The desired value of a upper moment (k-1 moment) CV,
SPcv,k-2 The desired value of a upper moment (k-1 moment) CV,
PVcv,k The value of feedback of current time (k moment) CV,
PVcv,k-1 The value of feedback of a upper moment (k-1 moment) CV,
PVcv,k-2 The value of feedback of upper upper moment (k-2 moment) CV,
GaincvThe gain of CV,
The Ts sampling time,
TicvThe time of integration of CV,
TdcvThe derivative time of CV.
CN201610610309.1A 2016-07-29 2016-07-29 A kind of pid control algorithm being suitable for multiple constraint target Pending CN106227033A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610610309.1A CN106227033A (en) 2016-07-29 2016-07-29 A kind of pid control algorithm being suitable for multiple constraint target

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610610309.1A CN106227033A (en) 2016-07-29 2016-07-29 A kind of pid control algorithm being suitable for multiple constraint target

Publications (1)

Publication Number Publication Date
CN106227033A true CN106227033A (en) 2016-12-14

Family

ID=57534804

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610610309.1A Pending CN106227033A (en) 2016-07-29 2016-07-29 A kind of pid control algorithm being suitable for multiple constraint target

Country Status (1)

Country Link
CN (1) CN106227033A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH09311703A (en) * 1996-05-23 1997-12-02 Rika Kogyo Kk Pid control method by fuzzy inference
CN1449511A (en) * 2000-06-30 2003-10-15 陶氏化学公司 Multi-variable matrix process control
CN103229113A (en) * 2010-09-29 2013-07-31 数学工程公司 Interactive system for controlling multiple input multiple output control (mimo) structures
CN105589448A (en) * 2009-02-02 2016-05-18 费希尔-罗斯蒙特系统公司 Model predictive controller with tunable integral component to compensate for model mismatch

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH09311703A (en) * 1996-05-23 1997-12-02 Rika Kogyo Kk Pid control method by fuzzy inference
CN1449511A (en) * 2000-06-30 2003-10-15 陶氏化学公司 Multi-variable matrix process control
CN105589448A (en) * 2009-02-02 2016-05-18 费希尔-罗斯蒙特系统公司 Model predictive controller with tunable integral component to compensate for model mismatch
CN103229113A (en) * 2010-09-29 2013-07-31 数学工程公司 Interactive system for controlling multiple input multiple output control (mimo) structures

Similar Documents

Publication Publication Date Title
Esfandyari et al. Adaptive fuzzy tuning of PID controllers
CN103645636B (en) A kind of PID controller parameter optimization method
Cui et al. Finite-time stabilization of stochastic low-order nonlinear systems with time-varying orders and FT-SISS inverse dynamics
CN102354107A (en) On-line identification and control method for parameter of alternating current position servo system model
Selamat et al. Comparison of LQR and PID controller tuning using PSO for Coupled Tank System
Prusty et al. Implementation of fuzzy-PID controller to liquid level system using LabVIEW
Bagheri et al. Tuning of dynamic matrix controller for FOPDT models using analysis of variance
Chen et al. Fuzzy differential games for nonlinear stochastic systems: suboptimal approach
del-Muro-Cuéllar et al. On the control of unstable first order linear systems with large time lag: Observer based approach
Ahmadi et al. Optimization of a self-tuning PID type fuzzy controller and a PID controller for an inverted pendulum
CN106227033A (en) A kind of pid control algorithm being suitable for multiple constraint target
Ji et al. Fuzzy dynamic sliding mode controller design for uncertain nonlinear Markovian jump systems
Krishnamoorthy et al. Gas-lift optimization by controlling marginal gas-oil ratio using transient measurements
Yamin et al. SIMPLIFIED ADAPTIVE CONTROL WITH GUARANTEED H8 PERFORMANCE
Aravind et al. Simulation based modeling and implementation of adaptive control technique for Non Linear process tank
Sagdatullin Multidimensional fuzzy control system development of oil transportation and treatment technological processes based on the input/output parameters model in the precise terms set form
Srinivasarao et al. Centralized and decentralized of quadruple tank process
Koshkouei Passivity‐based sliding mode control for nonlinear systems
Midhat Optimal LQR controller design for wing rock motion control in delta wing aircraft
Gopinath Design of fractional order controllers for first order plus time delay systems
Rigatos et al. Nonlinear optimal control for a gas compressor driven by an induction motor
Vhora et al. Design of static and dynamic decoupler for coupled quadruple tank level system
Kudlačák et al. Error behaviour in PID controll systems with dynamic processes
Ramos-Paz et al. Nonlinear optimal tracking control in combination with sliding modes: Application to the pendubot
Wati Interval type-2 fuzzy logic controller for multi input multi output system: A shower system case study

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication

Application publication date: 20161214

RJ01 Rejection of invention patent application after publication