CN106527141A - Glass furnace air-fuel ratio adjustment method based on variable universe fuzzy rule iterative learning - Google Patents

Glass furnace air-fuel ratio adjustment method based on variable universe fuzzy rule iterative learning Download PDF

Info

Publication number
CN106527141A
CN106527141A CN201611104943.4A CN201611104943A CN106527141A CN 106527141 A CN106527141 A CN 106527141A CN 201611104943 A CN201611104943 A CN 201611104943A CN 106527141 A CN106527141 A CN 106527141A
Authority
CN
China
Prior art keywords
harmony
fuel ratio
air
oxygen content
fuzzy
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
CN201611104943.4A
Other languages
Chinese (zh)
Other versions
CN106527141B (en
Inventor
刘民
崔兴华
董明宇
张龙
张亚斌
刘涛
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.)
Qingdao Qingneng Power Technology Co Ltd
Tsinghua University
Original Assignee
Qingdao Qingneng Power Technology Co Ltd
Tsinghua University
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 Qingdao Qingneng Power Technology Co Ltd, Tsinghua University filed Critical Qingdao Qingneng Power Technology Co Ltd
Priority to CN201611104943.4A priority Critical patent/CN106527141B/en
Publication of CN106527141A publication Critical patent/CN106527141A/en
Application granted granted Critical
Publication of CN106527141B publication Critical patent/CN106527141B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention relates to a glass furnace air-fuel ratio adjustment method based on variable universe fuzzy rule iterative learning, which belongs to the field of advanced manufacturing, automation and information. The method is characterized in that: firstly, a data-driven smoke oxygen content index prediction model is built, the input variable is an air-fuel ratio and the output variable is a smoke oxygen content. Meanwhile, on the basis of analyzing a chemical reaction mechanism during a glass furnace combustion process, a fuel thermal value serves as the input, a mechanism model for calculating the theoretical value of the air-fuel ratio is built, and the air-fuel ratio theoretical value acquired by the mechanism model is used for limiting the input value of the above data-driven smoke oxygen content index prediction model. On the basis of smoke oxygen content index prediction, an air-fuel ratio adjustment method based on variable universe fuzzy rule iterative learning is provided and a constraint satisfaction harmony search algorithm is provided for carrying out iterative learning on the variable universe fuzzy rule. When the method of the invention is applied to a glass production process, the furnace combustion condition can be effectively improved.

Description

Air/Fuel Ratio in Glass Furnace method of adjustment based on variable universe fuzzy rule iterative learning
Technical field
The invention belongs to advanced manufacture, automatization and message area, and in particular to a kind of Air/Fuel Ratio in Glass Furnace iteration is adjusted Adjusting method.
Background technology
The Combustion System of glass furnace is produced to improving kiln efficiency of combustion, reducing energy consumption, less waste gas discharge, raising glass Quality has important function, and which is the basis for realizing the efficient control such as glass furnace temperature, pressure, and optimization of air-fuel ratio setting It is the core content of Combustion System.In China's actual glass production process, kiln air-fuel ratio is often set to fixed value, but by Often change in production status such as heating value of natural gas, combustion-supporting air temperatures, carrying out Combustion System based on definite value air-fuel ratio can be to combustion Burn the indexs such as efficiency, energy consumption, discharge amount of exhaust gas, glass product quality to have a negative impact.
The content of the invention
The present invention proposes a kind of Air/Fuel Ratio in Glass Furnace method of adjustment based on variable universe fuzzy rule iterative learning, and which is special Levy and be, methods described is realized on computers successively according to the following steps：
Step (1)：Initialization, sets following basic variable
Setting problem variable：
x1(t)：The gas discharge of t
x2(t)：The combustion-supporting wind flow of t
y(t)：The flue gas oxygen content of t
[ymin,ymax]：Flue gas oxygen content setting value is interval
△C：Air-fuel ratio adjustment amount
C：Actual mixing ratio
CT：Chemically correct fuel
CGas：Heating value of natural gas
Step (2)：Data acquisition
Gather one or more production order of classes or grades at school including above-mentioned gas discharge x1(t), combustion-supporting wind flow x2T (), flue gas contain Oxygen amount y (t) information Store is into model database, and forms training data pair；
Step (3)：Indices prediction data model is modeled
According to the model that sampled data sets up following open loop Narx recurrent neural networks it is：
Y (t)=f2(x1(t),x1(t-1),…,x1(t-9),x1(t-10),x2(t),x2(t-1),…,x2(t-9),x1 (t-10),y(t
-1),y(t-2),…,y(t-10))
Wherein, t is the time, x1The gas discharge of (t) for t, x2T the combustion-supporting wind flow of () for t, y (t) are t The flue gas oxygen content at moment, f2() is closed loop Narx recurrent neural networks, and neutral net hidden layer is 20 neurons.
Step (4)：Mechanism model is modeled
When rule of thumb per the fuel combustion of 4186.8Kj caloric values, 1m is about needed3Air capacity, so as to if calorific value CGas Unit be Kj/m3, then the theoretical model of air-fuel ratio can be modeled as：
CT=CGas/4186.8
And then the present invention is interval for the restriction of air-fuel ratio with [CT*0.95, CT*1.05].
Step (5)：Set up variable universe fuzzy rule
With flue gas oxygen content actual deviation EoAnd its rate of change Δ E with the prediction deviation of indices prediction modeloAs fuzzy The input of rule, outputs of the air-fuel ratio adjustment amount △ C as fuzzy rule.Fuzzy decision table setting is as follows
1 fuzzy decision table of table
Its actual domain is according to EoWith Δ EoSelf-adaptative adjustment, i.e., when error is larger, using big fuzzy domain, error compared with Hour adopts less fuzzy domain.
Fuzzy theory domain classification method is as shown in drawings.Peak dot value a of each fuzzy set after its domain normalization1、a2、a3 And a4The harmonic search algorithm based on constraint satisfaction be given using step (7) carries out off-line search, and search procedure presses step (6) The method for being given sets up object model.
Step (6)：Object model is set up
It is modeled using closed loop Narx recurrent neural networks
Y (t)=f2(x1(t),x1(t-1),…,x1(t-9),x1(t-10),x2(t),x2(t-1),…,x2(t-9),x1 (t-10),y(t
-1),y(t-2),…,y(t-10))
Wherein, t is the time, x1The gas discharge of (t) for t, x2T the combustion-supporting wind flow of () for t, y (t) are t The flue gas oxygen content at moment, f2() is closed loop Narx recurrent neural networks, and neutral net hidden layer is 40 neurons.
Step (7)：Design the harmonic search algorithm based on constraint satisfaction to search for
Step (7.1)：Object function
When being adjusted to air-fuel ratio using variable universe fuzzy rule, in need to ensureing flue gas, oxygen content is in certain interval In the range of, so as to harmony vectorObject function be defined as follows：
f(Xi)=∑ | Eo|
Wherein, EoIt is the deviation of oxygen content in flue gas.
Step (7.2)：Harmony vector
In view of domain Eo、ΔEoIt is both needed to optimize corresponding a with each domain of △ C1、a2、a3And a4.So as to each Harmony vector includes d=4 × 3=12 harmony variate-value.Make each harmony vector be belonging to for Normal Distribution it is interval [0, 1] random number.
If harmony storehouse scale is HMS, then the individual harmony vectors of i-th (i=1,2 ..., HMS) in harmony storehouse can be represented such as Under：
Wherein,The peak dot of fuzzy number in first input domain fuzzy division is represented, needs to meet
The peak dot of fuzzy number in second input domain fuzzy division is represented, needs to meet
The peak dot of fuzzy number in output domain fuzzy division is represented, needs to meet
Step (7.3)：Harmony improves mechanism
Harmony storehouse is improved mechanism using harmony such as memory consideration, syllable adjustment, random selections, a new harmony can be obtained VectorWherein, for each harmony variate-value If the random uniform random number for generating considers probability HMCR less than previously given harmony storehouse,According to the note for being given as the following formula Recall the generation of consideration method：
Wherein, a is belonging to the random integers of interval [1, HMS].
Otherwise, if the uniform random number of above-mentioned random generation is more than or equal to HMCR,By following random selection Method is generated：
Wherein, Randn is the random number for belonging to interval [0,1] or [- 1,0] of Normal Distribution, if harmony variable pair The peak dot of fuzzy number is answered to need less than 0, then in [- 1,0] value, otherwise in [0,1] value.
Further, ifConsider to obtain by remembering, thenAdjusted by following syllable with finely tuning disturbance probability P AR Method is disturbed：
Wherein,It is j-th harmony variate-value in the best harmony vector for obtaining so far.
Step (7.4)：Constraint is processed
It is newly-generated according to said methodIt is unsatisfactory for harmony variable Constraint, the harmony variable of adjacent odd position and even bit need to be exchanged.
Step (7.5)：Harmony storehouse update mechanism
Improve mechanism through harmony, new harmony vector XnewCan be constructed, then, harmony storehouse is carried out more as follows Newly：If XnewWorst harmony X in better than harmony storehouseWorst, then by XnewReplace XWorst。
Step (7.6)：Iterative search mechanism
Iterative step 7.3- steps 7.5 arrive maximum iteration time, obtain the harmony vector of optimization.
According to the above-mentioned Air/Fuel Ratio in Glass Furnace on-line tuning method combined based on indices prediction data and mechanism, this It is bright to have done substantial amounts of emulation experiment, can be seen that from simulation result, reduction of the present invention to glass furnace fume oxygen content has significantly Effect.
Description of the drawings
Fig. 1：The Air/Fuel Ratio in Glass Furnace on-line tuning method hardware system combined based on indices prediction data and mechanism Structural representation.Fig. 2：Shown based on the Air/Fuel Ratio in Glass Furnace on-line tuning method flow that indices prediction data and mechanism combine It is intended to.
Fig. 3：Variable universe fuzzy rule domain divides schematic diagram.
Fig. 4：Flue gas oxygen content object model matched curve.
Fig. 5：Flue gas oxygen content indices prediction models fitting curve.
Fig. 6：Flue gas oxygen content change curve.
Specific embodiment
Dispatching method of the present invention depends on relevant data acquisition system, has modeling client and Modeling Server to realize. Using the software and hardware architecture schematic diagram of the present invention as shown in figure 1, embodiments of the present invention are as follows in actual glass kiln.
Step (1)：Initialization, sets following basic variable
Setting problem variable：
x1(t)：The gas discharge of t
x2(t)：The combustion-supporting wind flow of t
y(t)：The flue gas oxygen content of t
[ymin,ymax]：Flue gas oxygen content setting value is interval
△C：Air-fuel ratio adjustment amount
C：Actual mixing ratio
CT：Chemically correct fuel
CGas：Heating value of natural gas
Step (2)：Data acquisition
Gather one or more production order of classes or grades at school including above-mentioned gas discharge x1(t), combustion-supporting wind flow x2T (), flue gas contain Oxygen amount y (t) information Store is into model database, and forms training data pair；
Step (3)：Indices prediction data model is modeled
According to the model that sampled data sets up following open loop Narx recurrent neural networks it is：
Y (t)=f2(x1(t),x1(t-1),…,x1(t-9),x1(t-10),x2(t),x2(t-1),…,x2(t-9),x1 (t-10),y(t
-1),y(t-2),…,y(t-10))
Wherein, t is the time, x1The gas discharge of (t) for t, x2T the combustion-supporting wind flow of () for t, y (t) are t The flue gas oxygen content at moment, f2() is closed loop Narx recurrent neural networks, and neutral net hidden layer is 20 neurons.
Step (4)：Mechanism model is modeled
When rule of thumb per the fuel combustion of 4186.8Kj caloric values, 1m is about needed3Air capacity, so as to if calorific value CGas Unit be Kj/m3, then the theoretical model of air-fuel ratio can be modeled as：
CT=CGas/4186.8
And then the present invention is interval for the restriction of air-fuel ratio with [CT*0.95, CT*1.05].
Step (5)：Set up variable universe fuzzy rule
With flue gas oxygen content actual deviation EoAnd its rate of change Δ E with the prediction deviation of indices prediction modeloAs fuzzy The input of rule, outputs of the air-fuel ratio adjustment amount △ C as fuzzy rule.Fuzzy decision table setting is as follows
1 fuzzy decision table of table
Its actual domain is according to EoWith Δ EoSelf-adaptative adjustment, i.e., when error is larger, using big fuzzy domain, error compared with Hour adopts less fuzzy domain.
Fuzzy theory domain classification method is as shown in drawings.Peak dot value a of each fuzzy set after its domain normalization1、a2、a3 And a4The harmonic search algorithm based on constraint satisfaction be given using step (7) carries out off-line search, and search procedure presses step (6) The method for being given sets up object model.
Step (6)：Object model is set up
It is modeled using closed loop Narx recurrent neural networks
Y (t)=f2(x1(t),x1(t-1),…,x1(t-9),x1(t-10),x2(t),x2(t-1),…,x2(t-9),x1 (t-10),y(t
-1),y(t-2),…,y(t-10))
Wherein, t is the time, x1The gas discharge of (t) for t, x2T the combustion-supporting wind flow of () for t, y (t) are t The flue gas oxygen content at moment, f2() is closed loop Narx recurrent neural networks, and neutral net hidden layer is 40 neurons.
Step (7)：Design the harmonic search algorithm based on constraint satisfaction to search for
Step (7.1)：Object function
When being adjusted to air-fuel ratio using variable universe fuzzy rule, in need to ensureing flue gas, oxygen content is in certain interval In the range of, so as to harmony vectorObject function be defined as follows：
f(Xi)=∑ | Eo|
Wherein, EoIt is the deviation of oxygen content in flue gas.
Step (7.2)：Harmony vector
In view of domain Eo、ΔEoIt is both needed to optimize corresponding a with each domain of △ C1、a2、a3And a4.So as to each Harmony vector includes d=4 × 3=12 harmony variate-value.Make each harmony vector be belonging to for Normal Distribution it is interval [0, 1] random number.
If harmony storehouse scale is HMS, then the individual harmony vectors of i-th (i=1,2 ..., HMS) in harmony storehouse can be represented such as Under：
Wherein,The peak dot of fuzzy number in first input domain fuzzy division is represented, needs to meet
The peak dot of fuzzy number in second input domain fuzzy division is represented, needs to meet
The peak dot of fuzzy number in output domain fuzzy division is represented, needs to meet
Step (7.3)：Harmony improves mechanism
Harmony storehouse is improved mechanism using harmony such as memory consideration, syllable adjustment, random selections, a new harmony can be obtained VectorWherein, for each harmony variate-value If the random uniform random number for generating considers probability HMCR less than previously given harmony storehouse,According to the note for being given as the following formula Recall the generation of consideration method：
Wherein, a is belonging to the random integers of interval [1, HMS].
Otherwise, if the uniform random number of above-mentioned random generation is more than or equal to HMCR,By following random selection Method is generated：
Wherein, Randn is the random number for belonging to interval [0,1] or [- 1,0] of Normal Distribution, if harmony variable pair The peak dot of fuzzy number is answered to need less than 0, then in [- 1,0] value, otherwise in [0,1] value.
Further, ifConsider to obtain by remembering, thenAdjusted by following syllable with finely tuning disturbance probability P AR Method is disturbed：
Wherein,It is j-th harmony variate-value in the best harmony vector for obtaining so far.
Step (7.4)：Constraint is processed
It is newly-generated according to said methodIt is unsatisfactory for harmony variable Constraint, the harmony variable of adjacent odd position and even bit need to be exchanged.
Step (7.5)：Harmony storehouse update mechanism
Improve mechanism through harmony, new harmony vector XnewCan be constructed, then, harmony storehouse is carried out more as follows Newly：If XnewWorst harmony X in better than harmony storehouseWorst, then by XnewReplace XWorst。
Step (7.6)：Iterative search mechanism
Iterative step 7.3- steps 7.5 arrive maximum iteration time, obtain the harmony vector of optimization.
The parameter of the Air/Fuel Ratio in Glass Furnace on-line tuning method combined based on indices prediction data and mechanism is selected such as Under：
● the sample frequency for the data of indices prediction model modeling was adopted for 3 seconds；
● flue gas oxygen content set point is [3.05,3.15]
● harmony storehouse scale is HMS=50；
● iterationses are 100；
According to above-mentioned the proposed microelectronics production line scheduling method based on index prediction and on-line study, the present invention does Substantial amounts of l-G simulation test, the hardware environment of operation is：P4 2.8GHz CPU, 1024M RAM, operating system be Windows, UNIX。

Claims (3)

1. a kind of Air/Fuel Ratio in Glass Furnace method of adjustment based on variable universe fuzzy rule iterative learning, it is characterised in that include：
According to mechanism model and heating value of natural gas for theoretical air-fuel ratio theoretical value, the theory of glass furnace combustion process is obtained Air-fuel ratio；
According to the flue gas oxygen content indices prediction model and actual mixing ratio and described chemically correct fuel of data-driven, obtain right The flue gas oxygen content answered, and according to variable universe fuzzy rule and the flue gas oxygen content, air-fuel ratio is adjusted.
2. method according to claim 1, it is characterised in that the basis is used for the mechanism mould of theoretical air-fuel ratio theoretical value Type and heating value of natural gas, obtain the chemically correct fuel of glass furnace combustion process, specifically include following steps：
Step 1：Set following basic variable
x1(t)：The gas discharge of t
x2(t)：The combustion-supporting wind flow of t
y(t)：The flue gas oxygen content of t
[ymin,ymax]：Flue gas oxygen content setting value is interval
△C：Air-fuel ratio adjustment amount
C：Actual mixing ratio
CT：Chemically correct fuel
CGas：Heating value of natural gas
Step 2：Indices prediction data model is modeled
Following open loop Narx recurrent neural network models are set up according to sampled data is：
Y (t+1)=f1(x1(t),x1(t-1),…,x1(t-9),x2(t),x2(t-1),…,x2(t-9),y(t),y(t-1),y (t-2),…,y(t-9))
x1T () is t gas discharge, x2T () is t flue gas oxygen content, f for the combustion-supporting wind flow of t, y (t)1() is Closed loop Narx recurrent neural networks；
Step 3：Mechanism model is modeled
When rule of thumb per the fuel combustion of 4186.8Kj caloric values, 1m is about needed3Air capacity, so as to if the unit of calorific value CGas For Kj/m3, then the theoretical model of air-fuel ratio can be modeled as：
CT=CGas/4186.8
Further, it is interval for the restriction of air-fuel ratio with [CT*0.95, CT*1.05].
3. method according to claim 1, it is characterised in that described oxygen-containing according to variable universe fuzzy rule and the flue gas Amount, is adjusted to Air/Fuel Ratio in Glass Furnace, specifically includes following steps：
Step 1：Iteration sets up variable universe fuzzy rule
With flue gas oxygen content actual deviation EoAnd its rate of change Δ E with the prediction deviation of indices prediction modeloAs fuzzy rule Input, outputs of the air-fuel ratio adjustment amount △ C as fuzzy rule；
The actual domain of fuzzy rule is according to EoWith Δ EoSelf-adaptative adjustment, i.e., when error is larger, using big fuzzy domain, error Less fuzzy domain is adopted when less；Fuzzy theory domain classification method is as shown in drawings；Each after its domain normalization is obscured Peak dot value a of collection1、a2、a3And a4；Off-line search is carried out using the harmonic search algorithm based on constraint satisfaction that step 3 is given, Search procedure sets up object model by the method that step 2 is provided；
Step 2：Object model is set up
It is modeled using closed loop Narx recurrent neural networks
Y (t)=f2(x1(t),x1(t-1),…,x1(t-9),x1(t-10),x2(t),x2(t-1),…,x2(t-9),x1(t- 10),y(t-1),y(t-2),…,y(t-10))
Wherein, t is the time, x1The gas discharge of (t) for t, x2T the combustion-supporting wind flow of () for t, y (t) are t Flue gas oxygen content, f2() is closed loop Narx recurrent neural networks；
Step 3：Design the harmonic search algorithm based on constraint satisfaction to search for
Step 3.1：Object function
When being adjusted to air-fuel ratio using variable universe fuzzy rule, in need to ensureing flue gas, oxygen content is in certain interval range It is interior, so as to harmony vectorObject function be defined as follows：
f(Xi)=∑ | Eo|
Wherein, EoIt is the deviation of oxygen content in flue gas；
Step 3.2：Harmony vector
In view of domain Eo、ΔEoIt is both needed to optimize corresponding a with each domain of △ C1、a2、a3And a4；So as to each harmony Vector includes d=4 × 3=12 harmony variate-value；Each harmony vector is made to be that Normal Distribution belongs to interval [0,1] Random number；
If harmony storehouse scale is HMS, then the individual harmony vectors of i-th (i=1,2 ..., HMS) in harmony storehouse can be expressed as follows：
X i = { x 1 i , x 2 i , ... , x j i , ... , x d i } , i = 1 , 2 , ... , H M S
Wherein,The peak dot of fuzzy number in first input domain fuzzy division is represented, needs to meet
- 1 < x 1 i < x 2 i < 0 , 0 < x 3 i < x 4 i < 1
The peak dot of fuzzy number in second input domain fuzzy division is represented, needs to meet
- 1 < x 5 i < x 6 i < 0 , 0 < x 7 i < x 8 i < 1
The peak dot of fuzzy number in output domain fuzzy division is represented, needs to meet
- 1 < x 9 i < x 10 i < 0 , 0 < x 11 i < x 12 i < 1
Step 3.3：Harmony improves mechanism
To harmony storehouse using memory consider, syllable adjustment, random selection etc. harmony improve mechanism, can obtain a new harmony to AmountWherein, for each harmony variate-value If the random uniform random number for generating considers probability HMCR less than previously given harmony storehouse,According to the note for being given as the following formula Recall the generation of consideration method：
x j n e w = x j a
Wherein, a is belonging to the random integers of interval [1, HMS]；
Otherwise, if the uniform random number of above-mentioned random generation is more than or equal to HMCR,By following random selection method Generate：
x j n e w = R a n d n
Wherein, Randn is the random number for belonging to interval [0,1] or [- 1,0] of Normal Distribution, if harmony variable correspondence mould The peak dot of paste number is needed less than 0, then in [- 1,0] value, otherwise in [0,1] value；
Further, ifConsider to obtain by remembering, thenTo finely tune disturbance probability P AR by following syllable method of adjustment Disturbed：
x j n e w = x j B
Wherein,It is j-th harmony variate-value in the best harmony vector for obtaining so far；
Step 3.4：Constraint is processed
It is newly-generated according to said methodIt is unsatisfactory for the pact of harmony variable Beam, need to exchange the harmony variable of adjacent odd position and even bit；
Step 3.5：Harmony storehouse update mechanism
Improve mechanism through harmony, new harmony vector XnewCan be constructed, then, harmony storehouse is updated as follows：If XnewWorst harmony X in better than harmony storehouseWorst, then by XnewReplace XWorst。
Step 3.6：Iterative search mechanism
Iterative step 3.3- steps 3.5 arrive maximum iteration time, obtain the harmony vector of optimization.
CN201611104943.4A 2016-12-05 2016-12-05 Air/Fuel Ratio in Glass Furnace method of adjustment based on variable universe fuzzy rule iterative learning Active CN106527141B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201611104943.4A CN106527141B (en) 2016-12-05 2016-12-05 Air/Fuel Ratio in Glass Furnace method of adjustment based on variable universe fuzzy rule iterative learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201611104943.4A CN106527141B (en) 2016-12-05 2016-12-05 Air/Fuel Ratio in Glass Furnace method of adjustment based on variable universe fuzzy rule iterative learning

Publications (2)

Publication Number Publication Date
CN106527141A true CN106527141A (en) 2017-03-22
CN106527141B CN106527141B (en) 2019-09-20

Family

ID=58354979

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201611104943.4A Active CN106527141B (en) 2016-12-05 2016-12-05 Air/Fuel Ratio in Glass Furnace method of adjustment based on variable universe fuzzy rule iterative learning

Country Status (1)

Country Link
CN (1) CN106527141B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108363302A (en) * 2018-02-13 2018-08-03 集美大学 A kind of dynamic positioning of vessels bottom propeller control method based on harmony search
CN108733921A (en) * 2018-05-18 2018-11-02 山东大学 Coiling hot point of transformer temperature fluctuation range prediction technique based on Fuzzy Information Granulation
CN110471281A (en) * 2019-07-30 2019-11-19 南京航空航天大学 A kind of the Varied scope fuzzy control system and control method of Trajectory Tracking Control

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102654286A (en) * 2012-05-18 2012-09-05 无锡龙山科技有限公司 Intelligent dynamic combustion atmosphere controller
CN103062790A (en) * 2012-12-29 2013-04-24 聚光科技(杭州)股份有限公司 Method for controlling combustion inside heating furnace
CN102354109B (en) * 2011-06-22 2014-06-11 中南大学 Sintering hot air temperature and hot air oxygen content intelligent coordination optimization control method
US8862248B2 (en) * 2010-11-04 2014-10-14 Honda Motor Co., Ltd. Control apparatus
CN105547000A (en) * 2016-02-02 2016-05-04 上海策立工程技术有限公司 Routing inspection type flue gas adjusting system and method of rolled steel based double-heat-accumulation type heating furnace

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8862248B2 (en) * 2010-11-04 2014-10-14 Honda Motor Co., Ltd. Control apparatus
CN102354109B (en) * 2011-06-22 2014-06-11 中南大学 Sintering hot air temperature and hot air oxygen content intelligent coordination optimization control method
CN102654286A (en) * 2012-05-18 2012-09-05 无锡龙山科技有限公司 Intelligent dynamic combustion atmosphere controller
CN103062790A (en) * 2012-12-29 2013-04-24 聚光科技(杭州)股份有限公司 Method for controlling combustion inside heating furnace
CN105547000A (en) * 2016-02-02 2016-05-04 上海策立工程技术有限公司 Routing inspection type flue gas adjusting system and method of rolled steel based double-heat-accumulation type heating furnace

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
花新峰: "模糊控制在热风炉燃烧系统中的应用", 《工业控制计算机》 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108363302A (en) * 2018-02-13 2018-08-03 集美大学 A kind of dynamic positioning of vessels bottom propeller control method based on harmony search
CN108363302B (en) * 2018-02-13 2020-09-11 集美大学 Ship dynamic positioning bottom layer propeller control method based on harmony search
CN108733921A (en) * 2018-05-18 2018-11-02 山东大学 Coiling hot point of transformer temperature fluctuation range prediction technique based on Fuzzy Information Granulation
CN110471281A (en) * 2019-07-30 2019-11-19 南京航空航天大学 A kind of the Varied scope fuzzy control system and control method of Trajectory Tracking Control
CN110471281B (en) * 2019-07-30 2021-09-24 南京航空航天大学 Variable-discourse-domain fuzzy control system and control method for trajectory tracking control

Also Published As

Publication number Publication date
CN106527141B (en) 2019-09-20

Similar Documents

Publication Publication Date Title
Shamshirband et al. A survey of deep learning techniques: application in wind and solar energy resources
CN106527141A (en) Glass furnace air-fuel ratio adjustment method based on variable universe fuzzy rule iterative learning
CN107609667B (en) Heat supply load prediction method and system based on Box _ cox transformation and UFCNN
CN106022521A (en) Hadoop framework-based short-term load prediction method for distributed BP neural network
US20230419073A1 (en) Method for predicting gasification reaction characteristics of biomass char
CN103472866B (en) The pesticide waste liquid incinerator furnace temperature optimization system of intelligent fuzzy system and method
CN116167531A (en) Photovoltaic power generation prediction method based on digital twin
CN112131785A (en) Screening and optimizing method of methanation nickel-based catalyst formula based on ANN-NSGA-II
Li et al. A deep learning framework for non-stationary time series prediction
Liang et al. A wind speed combination forecasting method based on multifaceted feature fusion and transfer learning for centralized control center
Han et al. Deep belief rule based photovoltaic power forecasting method with interpretability
CN104091041B (en) High-order-moment based generated power estimation method and system
CN106529095A (en) Photovoltaic power generation prediction research system based on Matlab
CN115018156A (en) Short-term wind power prediction method
Chen et al. Gold price forecasting based on projection pursuit and neural network
CN115409291A (en) Wind speed corrected wind power prediction method and system
Guo et al. Stabilization temperature prediction in carbon fiber production using empirical mode decomposition and long short-term memory network
Duan et al. Time-series prediction using a regularized self-organizing long short-term memory neural network
CN113051837A (en) Method and device for predicting DLSTM power load based on meteorological parameters
Liao et al. Wind power prediction based on periodic characteristic decomposition and multi-layer attention network
CN113111577A (en) Cement mill operation index decision method based on multi-target cuckoo search
Su et al. Short-term load forecasting of regional integrated energy system based on spatio-temporal convolutional graph neural network
Yan et al. Model Prediction and Optimal Control of Gas Oxygen Content for A Municipal Solid Waste Incineration Process
Tian et al. Wind Speed Prediction based on Long Short-Term Memory network and Policy Suggestions for Wind Power Industry Development
Lv et al. New Teaching Model of Professional Big Data Courses in Universities Based on an Outcome-Oriented Educational Concept

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant