CN108224446A - A kind of automatic combustion Study on Decision-making Method for Optimization of Refuse Incineration Process - Google Patents
A kind of automatic combustion Study on Decision-making Method for Optimization of Refuse Incineration Process Download PDFInfo
- Publication number
- CN108224446A CN108224446A CN201711494362.0A CN201711494362A CN108224446A CN 108224446 A CN108224446 A CN 108224446A CN 201711494362 A CN201711494362 A CN 201711494362A CN 108224446 A CN108224446 A CN 108224446A
- Authority
- CN
- China
- Prior art keywords
- random arrangement
- case
- decision
- network
- real
- 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
Links
Classifications
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F23—COMBUSTION APPARATUS; COMBUSTION PROCESSES
- F23G—CREMATION FURNACES; CONSUMING WASTE PRODUCTS BY COMBUSTION
- F23G5/00—Incineration of waste; Incinerator constructions; Details, accessories or control therefor
- F23G5/50—Control or safety arrangements
Abstract
A kind of automatic combustion real-time optimization decision-making technique of Refuse Incineration Process, it is related to urban solid garbage incinerator Optimized-control Technique field, pass through the Optimal Decision-making of crucial controlled variable (feeder speed and fire grate speed), make automatic combustion control system operation in the ideal range, mainly include the following steps:(1) decision case library is established according to the historical data of burning process;(2) structure training mode pond;(3) random arrangement network is trained so as to obtain random arrangement network retrieval model according to learning-oriented pseudo-metric criterion;(4) target case is input to random arrangement network retrieval model and obtains the solution of K similar cases;(5) average value of K similar cases solution is asked for by case reuse, the decision value of (feeder speed and fire grate speed) is illustrated so as to obtain goal-trail and is output to burning process control system;(6) above-mentioned step (4)~step (5) is repeated, to realize the automatic combustion real-time optimization decision process of burning process.
Description
Technical field
The present invention relates to urban solid garbage incinerator Optimal Decision-making technical fields, and more specifically, the present invention is a kind of
The automatic combustion real-time optimization decision-making technique of Refuse Incineration Process.
Background technology
In recent years, as Urbanization in China is constantly accelerated, domestic waste yield is increased sharply, and harmless treatment energy
Power is but still insufficient.House refuse is affirmed and improved in National Development and Reform Committee in burning city domestic garbage handles opinion on work
The effect of burning disposal and status.Urban solid garbage incineration treatment technology is used widely in developed countries and regions, I
The current incineration treatment of garbage technology of state and equipment reach its maturity, it has also become the important way of urban garbage disposal.Therefore,
The achievement in research of the present invention has broad application prospects.
In Refuse Incineration Process, the burning of stablizing of rubbish is before ensureing to burn economic benefit maximum and environmental impact minimization
It carries.Therefore, under the premise of waste steady burning is ensured, realize the Optimal Decision-making of burning process key controlled variable with important
Realistic meaning.Since Refuse Incineration Process has many characteristics, such as the more, close coupling of interference, non-linear and big inertia, feeder in stove
The optimal setting of speed and fire grate speed is not easy to grasp, and is usually provided by artificial experience, with subjective random, it is difficult to according to
Production status variation is promptly and accurately set, and causes these technic index fluctuation ranges very big, it is difficult to control the model in process stipulation
In enclosing, and make production target up to standard.Therefore, traditional control method is difficult to realize stablize, accurately control.
At present, the Optimization Decision Models of feeder speed and fire grate speed mainly include modelling by mechanism and data in incinerator
Driving modeling.Since Refuse Incineration Process has complex characteristics, mechanism model is caused to be difficult to set up.And building based on data-driven
Mould method is realized by the data acquired with target variable is relevant, easily monitors on-line to target by artificial intelligence technology
Prediction becomes a research hotspot in Optimal Decision-making field.This modeling method mainly has artificial neural network, supporting vector
Machine, multiple linear regression method etc..During using neural net model establishing, lack the effective ways of determining hidden layer and interstitial content;It supports
Vector machine modeling method is for big data set, and training speed is slow, and therefore, the application effect for leading to these methods is bad.It is based on
The modeling method of neural network needs largely have sample representative enough, hidden layers numbers and neuron number determine according to
Rely experience, and convergence rate is slow, is easily trapped into Local Minimum;Model construction of SVM method has well for Small Sample Database
Effect, but it is long for the big-sample data training time, lack the ability of self study, the selection of parameter has uncertainty;It is more
First linear regression method is a problem for the selection of characteristic variable.Therefore, lead to the application effect of these Study on Decision-making Method for Optimization
It is bad.
Reasoning by cases is as a kind of problem solving of artificial intelligence field and machine learning method, it is to solve efficient and increase
The features such as amount formula learning performance is strong is widely applied.Therefore, the present invention utilizes this advantage of reasoning by cases, emphatically from case
Example search method is started with, and modeling is carried out to feeder speed and fire grate speed and realizes real-time optimization decision.
Invention content
In view of the above-mentioned problems, the present invention provides a kind of automatic combustion real-time optimization decision-making technique of Refuse Incineration Process, it can
By the Optimal Decision-making of crucial controlled variable (feeder speed and fire grate speed), automatic combustion control system is made to operate in ideal
In the range of.
In order to achieve the above objectives, the present invention adopts the following technical scheme that:
The automatic combustion real-time optimization decision-making technique of a kind of Refuse Incineration Process, it is characterised in that include the following steps:(1)
Decision case library is established according to the historical data of burning process;(2) structure training mode pond;(3) according to learning-oriented pseudo-metric standard
Random arrangement network is then trained so as to obtain random arrangement network retrieval model;(4) target case is input to random arrangement net
Network retrieval model obtains the solution of K similar cases;(5) average value of K similar cases solution is asked for by case reuse, so as to
The decision value of (feeder speed and fire grate speed) is illustrated to goal-trail and is output to burning process control system;(6) in repetition
The step of stating (4)~step (5), to realize the automatic combustion real-time optimization decision process of burning process.Further specifically include
Following steps:
(1) decision case library is established according to the historical data of burning process;Detailed process is as follows:
By 7 characteristic variable x1~x7(dryer section air quantity x1, burning 1 section of fire grate air mass flow x2, burning 2 sections of fire grate air
Flow x3, boiler export main steam flow x4, primary air flow x5, primary combustion room temperature x6, primary air fan outlet air pressure
x7) historical data and corresponding feeder speed y1With fire grate speed y2It is expressed as feature vector form, forms p source case,
It is stored in decision case library.It is C to remember every source casek, it is represented by following form:
Ck=(Xk;Yk), k=1,2 ..., p (1)
Wherein, p is source case sum;YkIt is kth source case CkIn feeder speed and fire grate velocity amplitude;XkIt is kth
The problem of source case, describes, Xk、YkIt is represented by:
Wherein, xλ,k(λ=1 ..., 7) represent CkIn the λ characteristic variable value;y1,k、y2,kFeeder speed is represented respectively
With fire grate velocity amplitude;(2) structure training mode pond;X is described for the problems in formula (2)k(k=1,2 ..., p), it is therein
There are one the real value data in [0,1] section for each input attribute.Any two problem describes XiAnd XjBetween similarity
It can be weighed by a real number between [0,1], for example, as i ≠ j, XiAnd XjBetween similarity it is high when take a larger reality
Number, otherwise takes a smaller real number.In order to which random arrangement network is used to establish similarity model, defining mode pond D is as follows:
D=(x', x ", δ (x', x ")) | (x', x ") ∈ Xi×Xj, i, j=1 ..., p } (3)
Wherein, × and represent cartesian product, X can be described any two problemiAnd XjBe combined obtain (x', x ");δ
(x', x ") represent Di Li Cray sign functions, when x' and x " completely it is similar when its value be 0, be otherwise 1.According to determining for formula (3)
Justice, the source case that can be stored from case library construct pattern pond D;
(3) random arrangement network is trained so as to obtain random arrangement network retrieval model according to learning-oriented pseudo-metric criterion;
Learning-oriented pseudo-metric is exactly to go to realize the pseudo-metric of x' and x " similarity degrees by learning art, such as random arrangement network.It is right
For the retrieval of similar cases, when being handled using the random arrangement network (there are one hidden layers for tool) of a standard, net
The output of network is exactly equal to 0 or 1 and is nearly impossible, and thus, the property of network is judged using following learning-oriented pseudo-metric criterion
Energy;
(A1)YNN(x', x ") < ε1, when x' and x " is similar;
(A2)YNN(x',x”)≥ε2, when x' and x " is dissimilar;
(A3)|YNN(x',x”)-YNN(x”,x')|≤ε3, to arbitrary x' and x ";
(A4)YNN(x',x”')≤YNN(x',x”)+YNN(x ", x " '), to arbitrary x " ', x' and x " are dissimilar;
Wherein, the problem of x', x " and x " ' expressions are obtained by formula (3) describes;YNN(x', x ") is the defeated of random arrangement network
Go out, represent the similarity degree between x' and x ";ε1,ε2,ε3It is constant, in general ε1=ε3, value 0.2-0.3, ε2Value
For 0.7-0.8;
Using random arrangement network first have to consider network structure choice, i.e., the node number of input layer and output layer and
The selection of hidden layer neuron number.An any given 0 < r < 1 and non-negative sequence of real numbers { μL, makeμL
≤(1-r).For L=1,2 ..., it is denoted as:
Wherein, m is the number of hidden layer neuron;δL,qFor the q in L any given hidden layer neuron range
The value of a neuron;| | | | representing matrix norm;eL-1,qFor q-th of neuron in L-1 hidden layer neuron range
Deviation;If the random basic function g in the selection of the number of hidden layer neuronLMeet with lower inequality:
Wherein, gL(meet for random basic functionOr);Arbitrary biasing(arithmetic number domain);Then the output of random arrangement network is expressed as:
Wherein, arbitrary function f=[f1,f2,...,fm]:(real number field d → m) meets
Output matrix [β1,β2,...,βL] it is Moore Roger Penrose generalized inverse matrix;
The process of random arrangement network, the training sample that will exactly be generated in pattern pond are trained according to learning-oriented pseudo-metric criterion
This input random arrangement network model and the process being trained by step (4)~step (6).The end condition of model can be
It sees whether to meet any of the above-described measurement criterion (A1-A4) with higher proportion on training set and test set, such as α % (α ∈ (0,
It 100]), at this point, can be by YNN(x, y) is for Case Retrieval process;
(4) target case is input to random arrangement network retrieval model and obtains the solution of K similar cases;By target case
The problem of X is describedp+1X is described with the problem of source casek(k=1,2 ..., p) forms p input pair, i.e.,:
Dk:<Xp+1;Xk>, k=1,2 ... p (7)
Recycle p Y of learning-oriented pseudo-metric criterion training outputNN(Xp+1,Xk), according to above-mentioned measurement criterion A1, can count
Go out and Xp+1Similar source case number, it is assumed that be K;
(5) K similar cases solution Y is asked for by case reuse1~YKAverage value, illustrated so as to obtain goal-trail (charging
Device speed Y1,p+1With fire grate speed Y2,p+1) decision value, and be output to burning process control system;
(6) above-mentioned step (4)~step (5) is repeated, to realize the automatic combustion real-time optimization decision mistake of burning process
Journey.
Compared with prior art, the present invention it has the following advantages:1st, the present invention is gone through using what is generated in Refuse Incineration Process
History data establish Optimization Decision Models using reasoning by cases method, and required time is shorter, are conducive to apply in real time;2nd, it avoids
Expertise optimizes the subjectivity of decision;3rd, using the learning-oriented pseudo-metric model carry out case based on random arrangement network
Example retrieval is conducive to avoid weight distribution and the problem apart from trap so that the decision value of feeder speed and fire grate speed meets
The requirement of burning process running optimizatin.
Description of the drawings
Fig. 1 is automatic combustion real-time optimization decision-making technique schematic diagram of the present invention;
Specific embodiment
1000 groups of data that sample data generates in certain incineration treatment of garbage factory combustion process, with reference to Fig. 1 pairs
The specific embodiment of the present invention is described further.
The automatic combustion real-time optimization decision-making technique of a kind of Refuse Incineration Process, it is characterised in that include the following steps:
(1) decision case library is established according to the historical data of burning process;Detailed process is as follows:
By 7 characteristic variable x1~x7(it is dryer section air quantity x respectively1(Nm3/ h), burning 1 section of fire grate air mass flow x2
(Nm3/ h), burning 2 sections of fire grate air mass flow x3(Nm3/ h), boiler export main steam flow x4(t/h), primary air flow x5(Nm3/
H), primary combustion room temperature x6(DEG C), primary air fan outlet air pressure x7(KPa)) historical data and corresponding decision category
Property, i.e. feeder speed y1(cm/min) and fire grate speed y2(cm/min) it is expressed as feature vector form, forms 1000 source cases
Example, is stored in decision case library.It is C to remember every source casek, it is represented by following form:
Ck=(Xk;Yk), k=1,2 ..., 1000 (1)
Wherein, 1000 be source case sum;YkIt is kth source case CkIn feeder speed and fire grate velocity amplitude;XkIt is
The problem of kth source case, describes, Xk、YkIt is represented by:
Wherein, xλ,k(λ=1 ..., 7) represent CkIn the λ characteristic variable value;y1,k、y2,kFeeder speed is represented respectively
With fire grate velocity amplitude;
(2) structure training mode pond;X is described for the problems in formula (2)k(k=1,2 ..., 1000), each of which
All there are one the real value data in [0,1] section for a input attribute.Any two problem describes XiAnd XjBetween similarity can be with
It is weighed by a real number between [0,1], for example, as i ≠ j, XiAnd XjBetween similarity it is high when take a larger real number,
Otherwise a smaller real number is taken.In order to which random arrangement network is used to establish similarity model, defining mode pond D is as follows:
D=(x', x ", δ (x', x ")) | (x', x ") ∈ Xi×Xj, i, j=1 ..., 1000 } (3)
Wherein, × and represent cartesian product, X can be described any two problemiAnd XjBe combined obtain (x', x ");δ
(x', x ") represent Di Li Cray sign functions, when x' and x " completely it is similar when its value be 0, be otherwise 1.According to determining for formula (3)
Justice, the source case that can be stored from case library construct pattern pond D;
(3) random arrangement network is trained so as to obtain random arrangement network retrieval model according to learning-oriented pseudo-metric criterion;
Learning-oriented pseudo-metric is exactly to go to realize the pseudo-metric of x' and x " similarity degrees by learning art, such as random arrangement network.It is right
For the retrieval of similar cases, when being handled using the random arrangement network (there are one hidden layers for tool) of a standard, net
The output of network is exactly equal to 0 or 1 and is nearly impossible, and thus, the property of network is judged using following learning-oriented pseudo-metric criterion
Energy;
(A1)YNN(x', x ") < ε1, when x' and x " is similar;
(A2)YNN(x',x”)≥ε2, when x' and x " is dissimilar;
(A3)|YNN(x',x”)-YNN(x”,x')|≤ε3, to arbitrary x' and x ";
(A4)YNN(x',x”')≤YNN(x',x”)+YNN(x ", x " '), to arbitrary x " ', x' and x " are dissimilar;
Wherein, the problem of x', x " and x " ' expressions are obtained by formula (3) describes;YNN(x', x ") is the defeated of random arrangement network
Go out, represent the similarity degree between x' and x ";ε1,ε2,ε3It is constant, in general ε1=ε3, value 0.2-0.3, ε2Value
For 0.7-0.8;
Using random arrangement network first have to consider network structure choice, i.e., the node number of input layer and output layer and
The selection of hidden layer neuron number.An any given 0 < r < 1 and non-negative sequence of real numbers { μL, makeμL
≤(1-r).For L=1,2 ..., it is denoted as:
Wherein, m is the number of hidden layer neuron;δL,qFor the q in L any given hidden layer neuron range
The value of a neuron;| | | | representing matrix norm;eL-1,qFor q-th of neuron in L-1 hidden layer neuron range
Deviation;If the random basic function g in the selection of the number of hidden layer neuronLMeet with lower inequality:
Wherein, gL(meet for random basic functionOr);Arbitrary biasing(arithmetic number domain);Then the output of random arrangement network is expressed as:
Wherein, arbitrary function f=[f1,f2,...,fm]:(real number field d → m) meets
Output matrix [β1,β2,...,βL] it is Moore Roger Penrose generalized inverse square
Battle array;
The process of random arrangement network, the training sample that will exactly be generated in pattern pond are trained according to learning-oriented pseudo-metric criterion
This input random arrangement network model and the process being trained by step (4)~step (6).The end condition of model can be
It sees whether to meet any of the above-described measurement criterion (A1-A4), such as α %=80% with higher proportion on training set and test set,
It at this point, can be by YNN(x, y) is for Case Retrieval process;
(4) target case is input to random arrangement network retrieval model and obtains the solution of K similar cases;By target case
The problem of X is describedp+1X is described with the problem of source casek(k=1,2 ..., 1000) forms 1000 inputs pair, i.e.,:
Dk:<Xp+1;Xk>, k=1,2 ... 1000 (7) recycle learning-oriented pseudo-metric criterion training output 1000
YNN(Xp+1,Xk), according to above-mentioned measurement criterion A1, can count and Xp+1Similar source case number, it is assumed that be K;
(5) K similar cases solution Y is asked for by case reuse1~YKAverage value, illustrated so as to obtain goal-trail (charging
Device speed Y1,p+1With fire grate speed Y2,p+1) decision value, and be output to burning process control system;
(6) above-mentioned step (4)~step (5) is repeated, to realize the automatic combustion real-time optimization decision mistake of burning process
Journey.
Previous Refuse Incineration Process optimal control is entirely to be carried out by operating personnel with artificial experience, frequent operation, labor
Fatigue resistance is big, once and misoperation can influence production target, cause larger economic loss.Using the automatic of Refuse Incineration Process
Burning real-time optimization decision-making technique, Optimization Decision Models can replace operator to carry out on-line optimization decision to setting value, realize steady
Fixed control.Optimization Decision Models can the decision value of real-time output feed device speed and fire grate speed online, the plan of feeder speed
It is 4.5% to close error, and the error of fitting of fire grate speed is 4.3%, has been achieved the effect that ideal.
Claims (2)
1. the automatic combustion real-time optimization decision-making technique of a kind of Refuse Incineration Process, which is characterized in that include the following steps:(1)
Decision case library is established according to the historical data of burning process;(2) structure training mode pond;(3) according to learning-oriented pseudo-metric standard
Random arrangement network is then trained so as to obtain random arrangement network retrieval model;(4) target case is input to random arrangement net
Network retrieval model obtains the solution of K similar cases;(5) average value of K similar cases solution is asked for by case reuse, so as to
The decision value of (feeder speed and fire grate speed) is illustrated to goal-trail and is output to burning process control system;(6) in repetition
The step of stating (4)~step (5), to realize the automatic combustion real-time optimization decision process of burning process.
2. a kind of automatic combustion real-time optimization decision-making technique of Refuse Incineration Process described in accordance with the claim 1, feature exist
In specifically comprising the following steps:
(1) decision case library is established according to the historical data of burning process;Detailed process is as follows:
By 7 characteristic variable x1~x7(dryer section air quantity x1, burning 1 section of fire grate air mass flow x2, burning 2 sections of fire grate air mass flows
x3, boiler export main steam flow x4, primary air flow x5, primary combustion room temperature x6, primary air fan outlet air pressure x7) go through
History data and corresponding feeder speed y1With fire grate speed y2It is expressed as feature vector form, forms p source case, be stored in
In decision case library.It is C to remember every source casek, it is represented by following form:
Ck=(Xk;Yk), k=1,2 ..., p (1)
Wherein, p is source case sum;YkIt is kth source case CkIn feeder speed and fire grate velocity amplitude;XkIt is kth source
The problem of case, describes, Xk、YkIt is represented by:
Wherein, xλ,k(λ=1 ..., 7) represent CkIn the λ characteristic variable value;y1,k、y2,kFeeder speed and fire grate are represented respectively
Velocity amplitude;
(2) structure training mode pond;X is described for the problems in formula (2)k(k=1,2 ..., p), each input belongs to
Property all there are one the real value data in [0,1] section.Any two problem describes XiAnd XjBetween similarity can be by [0,1]
Between real number weigh, for example, as i ≠ j, XiAnd XjBetween similarity it is high when take a larger real number, otherwise take one
A smaller real number.In order to which random arrangement network is used to establish similarity model, defining mode pond D is as follows:
D=(x', x ", δ (x', x ")) | (x', x ") ∈ Xi×Xj, i, j=1 ..., p } (3)
Wherein, × and represent cartesian product, X can be described any two problemiAnd XjBe combined obtain (x', x ");δ(x',
X ") represent Di Li Cray sign functions, when x' and x " completely it is similar when its value be 0, be otherwise 1.It, can according to the definition of formula (3)
Pattern pond D is constructed with the source case stored from case library;
(3) random arrangement network is trained so as to obtain random arrangement network retrieval model according to learning-oriented pseudo-metric criterion;Study
Type pseudo-metric is exactly to go to realize the pseudo-metric of x' and x " similarity degrees by learning art, such as random arrangement network.For phase
For retrieval like case, when being handled using the random arrangement network (there are one hidden layers for tool) of a standard, network
Output is exactly equal to 0 or 1 and is nearly impossible, thus, the performance of network is judged using following learning-oriented pseudo-metric criterion;
(A1)YNN(x', x ") < ε1, when x' and x " is similar;
(A2)YNN(x',x”)≥ε2, when x' and x " is dissimilar;
(A3)|YNN(x',x”)-YNN(x”,x')|≤ε3, to arbitrary x' and x ";
(A4)YNN(x',x”')≤YNN(x',x”)+YNN(x ", x " '), to arbitrary x " ', x' and x " are dissimilar;
Wherein, the problem of x', x " and x " ' expressions are obtained by formula (3) describes;YNN(x', x ") is the output of random arrangement network, table
Show the similarity degree between x' and x ";ε1,ε2,ε3It is constant, in general ε1=ε3, value 0.2-0.3, ε2Value is 0.7-
0.8;
First have to consider the structure choice of network using random arrangement network, i.e. the node number of input layer and output layer and implicit
The selection of layer neuron number.An any given 0 < r < 1 and non-negative sequence of real numbers { μL, makeμL≤
(1-r).For L=1,2 ..., it is denoted as:
Wherein, m is the number of hidden layer neuron;δL,qFor q-th of god in L any given hidden layer neuron range
Value through member;| | | | representing matrix norm;eL-1, qDeviation for q-th of neuron in L-1 hidden layer neuron range;
If the random basic function g in the selection of the number of hidden layer neuronLMeet with lower inequality:
Wherein, gL(meet for random basic functionOr);Arbitrary biasing(just
Real number field);Then the output of random arrangement network is expressed as:
Wherein, arbitrary function f=[f1,f2,...,fm]:(real number field d → m) meets
Output matrix [β1,β2,...,βL] it is Moore Roger Penrose generalized inverse square
Battle array;
The process of random arrangement network is trained according to learning-oriented pseudo-metric criterion, it is exactly that the training sample generated in pattern pond is defeated
Enter random arrangement network model and the process being trained by step (4)~step (6).The end condition of model can be in training
It sees whether to meet any of the above-described measurement criterion (A1-A4) with higher proportion on collection and test set, such as α % (α ∈ (0,
It 100]), at this point, can be by YNN(x, y) is for Case Retrieval process;
(4) target case is input to random arrangement network retrieval model and obtains the solution of K similar cases;By asking for target case
Topic description Xp+1X is described with the problem of source casek(k=1,2 ..., p) forms p input pair, i.e.,:
Dk:< Xp+1;Xk>, k=1,2 ... p (7)
Recycle p Y of learning-oriented pseudo-metric criterion training outputNN(Xp+1,Xk), according to above-mentioned measurement criterion A1, can count
Go out and Xp+1Similar source case number, it is assumed that be K;
(5) K similar cases solution Y is asked for by case reuse1~YKAverage value, illustrated so as to obtain goal-trail (feeder speed
Spend Y1,p+1With fire grate speed Y2,p+1) decision value, and be output to burning process control system;
(6) above-mentioned step (4)~step (5) is repeated, to realize the automatic combustion real-time optimization decision process of burning process.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711494362.0A CN108224446B (en) | 2017-12-31 | 2017-12-31 | A kind of automatic combustion real-time optimization decision-making technique of Refuse Incineration Process |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711494362.0A CN108224446B (en) | 2017-12-31 | 2017-12-31 | A kind of automatic combustion real-time optimization decision-making technique of Refuse Incineration Process |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108224446A true CN108224446A (en) | 2018-06-29 |
CN108224446B CN108224446B (en) | 2019-06-14 |
Family
ID=62642154
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711494362.0A Expired - Fee Related CN108224446B (en) | 2017-12-31 | 2017-12-31 | A kind of automatic combustion real-time optimization decision-making technique of Refuse Incineration Process |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108224446B (en) |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110608444A (en) * | 2019-08-26 | 2019-12-24 | 光大环保技术研究院(南京)有限公司 | Model-based waste incineration control system |
CN110762540A (en) * | 2019-10-22 | 2020-02-07 | 华南理工大学 | Combustion decision method and automatic combustion control method and system for garbage incinerator |
CN111457392A (en) * | 2020-04-20 | 2020-07-28 | 北京工业大学 | Intelligent setting method for air quantity in urban domestic garbage incineration process |
CN112571424A (en) * | 2020-11-27 | 2021-03-30 | 沈阳工业大学 | Direct constraint control of each axis speed of rehabilitation robot based on SCN walking force estimation |
CN113154404A (en) * | 2021-04-24 | 2021-07-23 | 北京工业大学 | Intelligent setting method for secondary air volume in urban domestic garbage incineration process |
CN113191078A (en) * | 2021-04-26 | 2021-07-30 | 北京工业大学 | Method for forecasting flue gas temperature of primary combustion chamber in municipal solid waste incineration process |
CN113408185A (en) * | 2021-04-24 | 2021-09-17 | 北京工业大学 | Method for setting air flow of grate in incineration process of municipal domestic waste |
WO2022142264A1 (en) * | 2020-12-31 | 2022-07-07 | 深圳市深能环保东部有限公司 | Method for online rapid calculation of garbage incineration calorific value |
CN112571424B (en) * | 2020-11-27 | 2024-05-03 | 沈阳工业大学 | Rehabilitation robot shaft speed direct constraint control based on SCN walking force estimation |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH08221113A (en) * | 1995-02-16 | 1996-08-30 | Hitachi Ltd | Time-series data storage device and process abnormality diagnostic device |
CN102540879A (en) * | 2011-12-28 | 2012-07-04 | 北京工业大学 | Multi-target evaluation optimization method based on group decision making retrieval strategy |
CN104392006A (en) * | 2014-12-17 | 2015-03-04 | 中国农业银行股份有限公司 | Event query processing method and device |
CN105629962A (en) * | 2016-03-03 | 2016-06-01 | 中国铁路总公司 | Failure diagnosis method for high-speed railway train control equipment radio block center (RBC) system |
CN104534507B (en) * | 2014-11-18 | 2017-03-29 | 华北电力大学(保定) | A kind of boiler combustion optimization control method |
CN107168067A (en) * | 2017-06-26 | 2017-09-15 | 北京工业大学 | A kind of waste incinerator Temperature Fuzzy Control method of use reasoning by cases extracting rule |
-
2017
- 2017-12-31 CN CN201711494362.0A patent/CN108224446B/en not_active Expired - Fee Related
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH08221113A (en) * | 1995-02-16 | 1996-08-30 | Hitachi Ltd | Time-series data storage device and process abnormality diagnostic device |
CN102540879A (en) * | 2011-12-28 | 2012-07-04 | 北京工业大学 | Multi-target evaluation optimization method based on group decision making retrieval strategy |
CN104534507B (en) * | 2014-11-18 | 2017-03-29 | 华北电力大学(保定) | A kind of boiler combustion optimization control method |
CN104392006A (en) * | 2014-12-17 | 2015-03-04 | 中国农业银行股份有限公司 | Event query processing method and device |
CN105629962A (en) * | 2016-03-03 | 2016-06-01 | 中国铁路总公司 | Failure diagnosis method for high-speed railway train control equipment radio block center (RBC) system |
CN107168067A (en) * | 2017-06-26 | 2017-09-15 | 北京工业大学 | A kind of waste incinerator Temperature Fuzzy Control method of use reasoning by cases extracting rule |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110608444A (en) * | 2019-08-26 | 2019-12-24 | 光大环保技术研究院(南京)有限公司 | Model-based waste incineration control system |
CN110762540A (en) * | 2019-10-22 | 2020-02-07 | 华南理工大学 | Combustion decision method and automatic combustion control method and system for garbage incinerator |
CN111457392A (en) * | 2020-04-20 | 2020-07-28 | 北京工业大学 | Intelligent setting method for air quantity in urban domestic garbage incineration process |
CN111457392B (en) * | 2020-04-20 | 2021-12-24 | 北京工业大学 | Intelligent setting method for air quantity in urban domestic garbage incineration process |
CN112571424A (en) * | 2020-11-27 | 2021-03-30 | 沈阳工业大学 | Direct constraint control of each axis speed of rehabilitation robot based on SCN walking force estimation |
CN112571424B (en) * | 2020-11-27 | 2024-05-03 | 沈阳工业大学 | Rehabilitation robot shaft speed direct constraint control based on SCN walking force estimation |
WO2022142264A1 (en) * | 2020-12-31 | 2022-07-07 | 深圳市深能环保东部有限公司 | Method for online rapid calculation of garbage incineration calorific value |
CN113154404A (en) * | 2021-04-24 | 2021-07-23 | 北京工业大学 | Intelligent setting method for secondary air volume in urban domestic garbage incineration process |
CN113408185A (en) * | 2021-04-24 | 2021-09-17 | 北京工业大学 | Method for setting air flow of grate in incineration process of municipal domestic waste |
CN113408185B (en) * | 2021-04-24 | 2023-05-23 | 北京工业大学 | Fire grate air flow setting method in urban household garbage incineration process |
CN113191078A (en) * | 2021-04-26 | 2021-07-30 | 北京工业大学 | Method for forecasting flue gas temperature of primary combustion chamber in municipal solid waste incineration process |
Also Published As
Publication number | Publication date |
---|---|
CN108224446B (en) | 2019-06-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108224446B (en) | A kind of automatic combustion real-time optimization decision-making technique of Refuse Incineration Process | |
CN106931453B (en) | The forecasting system and method for circulating fluid bed domestic garbage burning emission of NOx of boiler | |
CN109270842B (en) | Bayesian network-based regional heat supply model prediction control system and method | |
CN107726358B (en) | Boiler Combustion Optimization System based on CFD numerical simulations and intelligent modeling and method | |
CN103676881B (en) | A kind of dynamic bottleneck analytical method of semiconductor production line | |
CN104820707B (en) | One kind being based on the architectonic B/S patterns automatic volume group method of computer realm | |
CN103064289A (en) | Multiple-target operation optimizing and coordinating control method and device of garbage power generator | |
CN105864797B (en) | Real-time prediction system and method for boiler entering heat value of circulating fluidized bed household garbage incineration boiler | |
CN106224939B (en) | Circulating fluid bed domestic garbage burning boiler bed temperature Forecasting Methodology and system | |
US20210365784A1 (en) | Method for deriving fault diagnosis rules of blast furnace based on deep neural network | |
CN105751470B (en) | A kind of injection machine temperature real-time control method | |
CN112766608B (en) | Cement mill system power consumption index prediction method based on extreme learning machine | |
CN107038334A (en) | Circulating fluid bed domestic garbage burning boiler CO emitted smoke system and methods | |
CN106199174A (en) | Extruder energy consumption predicting abnormality method based on transfer learning | |
CN105608295B (en) | The multi-objective genetic algorithm of coking furnace pressure and RBF neural Optimization Modeling method | |
Wang et al. | BP-GA data fusion algorithm studies oriented to smart home | |
CN106444388A (en) | Distributed PID type dynamic matrix control method for furnace pressure of coke furnace | |
CN105975701A (en) | Parallel scheduling disassembly path forming method based on mixing fuzzy model | |
CN106371321A (en) | PID control method for fuzzy network optimization of coking-furnace hearth pressure system | |
Yang et al. | A novel exhaust gas temperature prediction method of hot blast stove | |
CN115526433A (en) | Power plant reheat flue gas baffle operation prediction method based on integrated hybrid model | |
CN102096741A (en) | Thermodynamic system dynamic model establishing method | |
CN106444389A (en) | Method for optimizing PI control by fuzzy RBF neural network based on system of pyrolysis of waste plastic temperature | |
Yongquan et al. | Fuzzy neural PID controller and tuning its weight factors using genetic algorithm based on different location crossover | |
CN114282370B (en) | Disassembly line setting method considering physical and mental loads of operator |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | 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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190614 Termination date: 20211231 |
|
CF01 | Termination of patent right due to non-payment of annual fee |