CN113154404A - Intelligent setting method for secondary air volume in urban domestic garbage incineration process - Google Patents

Intelligent setting method for secondary air volume in urban domestic garbage incineration process Download PDF

Info

Publication number
CN113154404A
CN113154404A CN202110444812.5A CN202110444812A CN113154404A CN 113154404 A CN113154404 A CN 113154404A CN 202110444812 A CN202110444812 A CN 202110444812A CN 113154404 A CN113154404 A CN 113154404A
Authority
CN
China
Prior art keywords
weight
feature
case
weights
secondary air
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
CN202110444812.5A
Other languages
Chinese (zh)
Other versions
CN113154404B (en
Inventor
严爱军
李佳璇
郭益东
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.)
Beijing University of Technology
Original Assignee
Beijing University of Technology
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 Beijing University of Technology filed Critical Beijing University of Technology
Priority to CN202110444812.5A priority Critical patent/CN113154404B/en
Publication of CN113154404A publication Critical patent/CN113154404A/en
Application granted granted Critical
Publication of CN113154404B publication Critical patent/CN113154404B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

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
    • F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F23—COMBUSTION APPARATUS; COMBUSTION PROCESSES
    • F23L—SUPPLYING AIR OR NON-COMBUSTIBLE LIQUIDS OR GASES TO COMBUSTION APPARATUS IN GENERAL ; VALVES OR DAMPERS SPECIALLY ADAPTED FOR CONTROLLING AIR SUPPLY OR DRAUGHT IN COMBUSTION APPARATUS; INDUCING DRAUGHT IN COMBUSTION APPARATUS; TOPS FOR CHIMNEYS OR VENTILATING SHAFTS; TERMINALS FOR FLUES
    • F23L9/00—Passages or apertures for delivering secondary air for completing combustion of fuel 
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]

Landscapes

  • Engineering & Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Mechanical Engineering (AREA)
  • Theoretical Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Biomedical Technology (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Combustion & Propulsion (AREA)
  • Feedback Control In General (AREA)

Abstract

An intelligent setting method for secondary air quantity in an urban domestic garbage incineration process relates to the field of parameter setting of urban solid garbage incinerators, enables the oxygen content of outlet flue gas to be maintained in an ideal range through intelligent setting of the secondary air quantity, and mainly comprises the following steps: (1) establishing a set database according to historical data of the incineration process of the municipal solid waste; (2) initializing parameters; (3) distributing characteristic weight based on a black hole cuckoo algorithm; (4) obtaining solutions of K similar cases through a case retrieval model; (5) and obtaining the average value of the K similar case solutions through case reuse, thereby obtaining the set value of the target case solution. (6) Forming a target case and a set value thereof into a case and storing the case in a set database; (7) and (5) repeating the steps (3) to (6) to realize the intelligent setting process of the secondary air quantity in the incineration process of the municipal domestic waste.

Description

Intelligent setting method for secondary air volume in urban domestic garbage incineration process
Technical Field
The invention relates to the technical field of parameter setting of municipal solid waste incinerators, in particular to an intelligent setting method for secondary air volume in a municipal solid waste incineration process.
Background
China is a large population country, a large energy production country and a large garbage production country at the same time. The population growth, the rapid economic development and the improvement of living standard lead to the massive increase of domestic garbage in China cities, the domestic garbage grows at a speed of 8-10% every year, the pollution is increasingly serious, and great challenges are brought to China. However, the urban domestic garbage yield is high, the garbage treatment capacity is insufficient, and the treatment rate is low in China; the garbage treatment effect is poor, and the standard exceeding phenomenon is common; the composition and the incineration process of the garbage are complex, and how to maintain the stable combustion of the municipal solid waste is a considerable problem, so the research result of the invention has wide application prospect.
In the process of waste incineration, the oxygen content of the outlet flue gas is one of important parameters for controlling the incineration of the urban domestic waste, the oxygen content of the outlet flue gas is very important in the whole waste incineration process and needs to be maintained within 6% -9%, and the heat loss is increased and the heat is taken away due to the overhigh oxygen content of the outlet flue gas; the problems of incomplete waste incineration, low incineration efficiency and the like are caused if the secondary air volume is too low, but the control of the oxygen content of the outlet flue gas is not a single control loop, strong coupling relation exists among variables, the secondary air volume is an important factor influencing the oxygen content of the outlet flue gas, the grate furnace mainly maintains the stability of the oxygen content of the outlet flue gas by adjusting the secondary air volume, further the combustion efficiency is improved, and how to realize intelligent optimization setting of the secondary air volume has important significance.
At present, the research of the intelligent setting method of the secondary air volume in the process of burning the urban domestic garbage mainly comprises a secondary air volume setting method based on mechanism modeling and data driving modeling. For the mechanism modeling-based secondary air volume setting method, as the municipal solid waste incineration process is a complex industrial control process and has the characteristics of multiple input, multiple output, nonlinearity, strong coupling and the like, an accurate mathematical model is difficult to adopt for mechanism modeling; the secondary air volume setting method based on data-driven modeling has the advantages of short research period, low cost, easiness in implementation and the like, wherein the secondary air volume intelligent optimization setting is carried out by using methods such as an artificial neural network, a fuzzy C mean value, fuzzy control and the like, but for a complex nonlinear system, the convergence speed of the artificial neural network is low, and the network weight is adjusted along the direction of local improvement, so that the local minimum is easy to happen, and the network training fails; the fuzzy C mean value has the defects of strong dependence on initialization data, sensitivity to noise and the like; the fuzzy control can not obtain accurate fuzzy rules and membership functions, and is mainly carried out by experience; the above methods cannot obtain an ideal secondary air volume set value.
Case reasoning originates in the eighties of the twentieth century, is an important problem solving method in the field of artificial intelligence, and has a core idea that a new problem is solved according to solving experiences of similar problems in the past, and the case reasoning is widely applied to the fields of medicine, security assessment, image processing and the like.
Disclosure of Invention
Aiming at the problems, the invention provides an intelligent setting method for secondary air quantity in the process of burning urban domestic garbage, which can ensure that the urban domestic garbage incinerator operates stably by optimally setting the secondary air quantity.
In order to achieve the purpose, the invention adopts the following technical scheme:
an intelligent setting method for secondary air quantity in an urban domestic garbage incineration process is characterized by comprising the following steps: (1) establishing a set case library according to historical data of the urban life incineration process; (2) initializing parameters; (3) distributing characteristic weight based on black hole cuckoo search (BH-CS); (4) obtaining solutions of 3 similar cases through a case retrieval model; (5) and (3) calculating the average value of the solutions of the 3 similar cases through case reuse to obtain a set value of the solution of the target case. (6) Forming a case by the target case and the set value thereof and storing the case in a set case library; (7) and (5) repeating the steps (3) to (6) to realize the intelligent setting process of the secondary air quantity in the incineration process of the municipal domestic waste. The method further comprises the following steps:
(1) establishing a set case library according to historical data of the urban life incineration process; the outlet temperature (x) of the secondary air heater is measured according to the characteristic variable1) Mean value of flue gas temperature (x) in a combustion chamber2) Mean value of flue gas temperature (x) of second combustion chamber3) Main steam flow at the boiler outlet (x)4) And the primary air volume (x)5) And the actual value y of the secondary air volume under the current working condition of the evaluation index is expressed into a characteristic vector form after normalization processing, so that n source cases are formed and stored in a set case library. Record each source case as CkExpressed in the following form:
Ck＝(xk,yk),k＝1,2,3,...,n (1)
wherein n is the total number of source cases; ckFor the kth source case, xkInputting normalized characteristic variables for the kth source case; y iskThe actual value of the secondary air volume is correspondingly normalized; x is the number ofkCan be expressed as:
xk＝(x1,k,x2,k,x3,k,x4,k,x5,k) (2)
setting the characteristic variable input of the current working condition normalization as xn+1As follows:
xn+1＝(x1,n+1,x2,n+1,x3,n+1,x4,n+1,x5,n+1) (3)
(2) initializing parameters; let Max iteration number MaxGeneration, weight upper limit wu, weight lower limit wl, and culling probability Pa in the Cuckoo algorithm.
(3) Distributing characteristic weights based on a black hole cuckoo algorithm, wherein the whole algorithm is divided into three stages, namely a first stage, and selecting a fitness function; in the second stage, a CS algorithm is adopted to optimize the characteristic weight, and the characteristic weight is divided into Levy flight and Pa probability elimination; and in the third stage, adding a BH algorithm to continuously optimize the characteristic weight.
The first stage is as follows: selecting a fitness function, and taking Root Mean Square Error (RMSE) of a secondary air volume CBR set model as the fitness function, wherein the Root Mean Square Error (RMSE) is expressed as follows:
Figure BDA0003036423060000031
wherein, ykIs the true value of the secondary air quantity of the kth case;
Figure BDA0003036423060000032
is the setting value of the secondary air volume of the kth case.
And a second stage: levy flight and elimination of feature weights with Pa probability
a) Levy flight is a random walk mechanism, and random update characteristic weights based on Levy flight are expressed as follows:
Figure BDA0003036423060000033
wherein omegai (t)And Ωi (t+1)Respectively representing the characteristic weights of the t generation and the t +1 generation; n is the number of groups of population space characteristic weights;
Figure BDA0003036423060000034
is a step size factor;
Figure BDA0003036423060000035
is a dot product; levy (λ) is a random number obtained by Levy flight;
Figure BDA0003036423060000036
and Levy (λ) as follows:
Figure BDA0003036423060000037
wherein the content of the first and second substances,
Figure BDA0003036423060000038
is a constant with a value of 0.01; omegabestIs the current generation optimal feature weight.
Levy(λ)～μ＝t-λ,(1＜λ≤3) (7)
Figure BDA0003036423060000039
Wherein u and v are random numbers of standard normal distribution, namely the occurrence probability meets the normal distribution, and the probability of the numbers closer to the middle is higher, and the probability of the numbers closer to the two sides is lower; β is a constant over the interval (1,2), typically taken to be 1.5;
Figure BDA00030364230600000310
is represented as follows:
Figure BDA00030364230600000311
where f is the gamma function.
b) Randomly generating feature weights with Pa probability
Recalculating the fitness of the feature weights, and evaluating the optimal feature weights, namely if the fitness of the feature weights generated after Levy flight is smaller, replacing the previous feature weights with the feature weights after Levy flight, detecting a group of worst feature weights in the weight population space according to the probability Pa, namely generating a random number r, if r is greater than Pa, representing the current feature weight difference, eliminating the current feature weight difference and randomly generating a new group of feature weights, otherwise, keeping the feature weights unchanged, wherein the value of Pa is always 0.5, and the feature weight updating mode is represented as follows:
Figure BDA0003036423060000041
wherein omegaj (t)，Ωk (t)Are two random sets of feature weights for the t-th generation; n is the number of groups of population space characteristic weights; r is a random number uniformly distributed in the interval (0, 1).
Figure BDA0003036423060000042
And a third stage: adding BH algorithm to continuously optimize feature weights
Taking the optimal solution of the feature weight as a black hole, and continuously optimizing the feature weight by adopting a BH algorithm updating strategy, wherein the feature weight updating is represented as follows:
Figure BDA0003036423060000043
wherein, rand () is a random number uniformly distributed in the (0,1) interval; omegaBHIs the location of the black hole.
The radius of the black hole is defined as follows:
Figure BDA0003036423060000044
wherein f isBHThe fitness value of the black hole is obtained; f. ofiIs the ith set of feature weight fitness values.
In an iterative process, if the feature weight moves to the black hole radius rBHIn which the distance d between the feature weight and the black hole is smaller than the radius r of the black holeBHThe feature weights are absorbed by the black holes, and after a group of feature weights are absorbed by the black holes and disappear, a new group of feature weights must be randomly generated in the weight population space to ensure that the number of the feature weights is unchanged, wherein the absorption mechanism of the feature weights and the black holes is expressed as follows:
Figure BDA0003036423060000045
wherein omegal (t)Is a set of random feature weights for the t-th generation; omegas (t)Is another set of random feature weights for the t-th generation; rand () is a random number uniformly distributed in the (0,1) interval; d is the distance between the feature weight and the black hole; r isBHIs the radius of the black hole.
And adding the BH algorithm to continuously optimize the feature weight by the BH-CS algorithm on the basis of the CS algorithm, and continuously evaluating the optimal feature weight in the current-generation weight population space until the maximum iteration times are met to obtain the global optimal feature weight.
(4) Obtaining solutions of 3 similar cases through a case retrieval model; calculating input x of characteristic variable after normalization of current working conditionn+1Input x of characteristic variable of historical working conditionkSimilarity of (2)kThe smaller the distance is, the more similar the distance is, the greater the similarity is, the more similar the working conditions are, taking the weighted Euclidean distance formula as an example, the following formula is a calculation target case xn+1And source case xkSimilarity of (2)k：
Figure BDA0003036423060000051
Wherein, ω ispIs the weight, x, of the p-th characteristic variablep,kInputting the p characteristic weight of the kth source case under the historical working condition; x is the number ofp,n+1Is the input of the p characteristic weight of the current working condition;
the following constraints are satisfied:
Figure BDA0003036423060000052
comparison Sk3 similar cases were retrieved.
(5) Obtaining an average value of 3 similar case solutions through case reuse, thereby obtaining a set value of a target case solution;
(6) and forming a case by the target case and the set value, storing the case into a set database, and solving the next suboptimal setting.
(7) And (5) repeating the steps (3) to (6) to realize the setting process of the secondary air volume in the incineration process of the municipal domestic waste.
Compared with the prior art, the invention has the following advantages: 1. the invention adopts historical data in the process of waste incineration, optimizes the set value of the secondary air volume based on case-based reasoning, has short research period, low cost, easy realization and is beneficial to real-time application; 2. the secondary air volume case-based reasoning intelligent setting model is adopted, expert experience is not needed, and limitation and subjectivity of secondary air volume setting are effectively avoided; 3. the characteristic weight is optimized based on the BH-CS algorithm, the problem that the algorithm is easy to fall into local optimization in the later iteration stage is solved, and the prediction precision of the secondary air volume case-based reasoning intelligent setting model can be effectively improved by verifying the characteristic weight distributed by the BH-CS algorithm through a comparison experiment, so that the set value of the secondary air volume meets the requirement of operation optimization in the incineration process.
Drawings
FIG. 1 is a schematic diagram of the method for setting the secondary air volume in the process of incinerating municipal solid waste.
Detailed Description
The sample data is 1000 data generated in the combustion process of a waste incineration plant, and is randomly divided into 800 source cases and 200 test cases, and the specific implementation of the invention is further explained with reference to fig. 1.
A method for setting secondary air volume in the process of burning urban domestic garbage comprises the following steps:
(1) establishing a set case library according to historical data of the incineration process; the detailed process is as follows:
5 characteristic variables of outlet temperature (x) of secondary air heater1) Mean value of flue gas temperature (x) in a combustion chamber2) Mean value of flue gas temperature (x) of second combustion chamber3) Main steam flow at the boiler outlet (x)4) And the primary air volume (x)5) And the actual value y of the secondary air volume under the current working condition of the evaluation index is expressed into a characteristic vector form after normalization processing, 800 source cases are formed and stored in a set case library. Record each source case as CkExpressed in the following form:
Ck＝(xk,yk),k＝1,2,3,...,800 (1)
wherein 800 is the total number of source cases; ckFor the kth source case, xkInputting normalized characteristic variables for the kth source case; y iskThe actual value of the corresponding normalized secondary air volume is obtained; x is the number ofkCan be expressed as:
xk＝(x1,k,x2,k,x3,k,x4,k,x5,k) (2)
setting the characteristic variable input of the current working condition normalization as xn+1As follows:
xn+1＝(x1,n+1,x2,n+1,x3,n+1,x4,n+1,x5,n+1) (3)
(2) initializing parameters; the maximum iteration number is 100, the weight upper limit is 1, the weight lower limit is 0, and the culling probability in the cuckoo algorithm is 0.25.
(3) Distributing characteristic weights based on a black hole cuckoo algorithm, wherein the whole algorithm is divided into three stages, namely a first stage, and selecting a fitness function; in the second stage, a CS algorithm is adopted to optimize the characteristic weight, and the characteristic weight is divided into Levy flight and Pa probability elimination; and in the third stage, adding a BH algorithm to continuously optimize the characteristic weight.
The first stage is as follows: selecting a fitness function, and taking Root Mean Square Error (RMSE) of a secondary air volume CBR set model as the fitness function, wherein the Root Mean Square Error (RMSE) is expressed as follows:
Figure BDA0003036423060000061
wherein, ykIs the true value of the secondary air quantity of the kth case;
Figure BDA0003036423060000062
is the setting value of the secondary air volume of the kth case.
And a second stage: levy flight and elimination of feature weights with Pa probability
a) Levy flight is a random walk mechanism, and random update characteristic weights based on Levy flight are expressed as follows:
Figure BDA0003036423060000063
wherein omegai (t)And Ωi (t+1)Respectively representing the characteristic weights of the t generation and the t +1 generation; n is the number of groups of population space characteristic weights;
Figure BDA0003036423060000064
is a step size factor;
Figure BDA0003036423060000065
is a dot product; levy (λ) is a random number obtained by Levy flight;
Figure BDA0003036423060000066
and Levy (λ) as follows:
Figure BDA0003036423060000067
wherein the content of the first and second substances,
Figure BDA0003036423060000068
is a constant with a value of 0.01; omegabestIs the current generation optimal feature weight.
Levy(λ)～μ＝t-λ,(1＜λ≤3) (7)
Figure BDA0003036423060000071
Wherein u, v are random numbers of a standard normal distribution; β is a constant over the interval (1, 2);
Figure BDA0003036423060000072
is represented as follows:
Figure BDA0003036423060000073
where f is the gamma function.
b) Randomly generating feature weights with Pa probability
Recalculating the fitness of the feature weights, and evaluating the optimal feature weights, namely if the fitness of the feature weights generated after Levy flight is smaller, replacing the previous feature weights with the feature weights after Levy flight, detecting a group of worst feature weights in the weight population space according to the probability Pa, namely generating a random number r, if r > Pa, representing the current feature weight difference, eliminating the current feature weight difference and randomly generating a new group of feature weights, otherwise, keeping the feature weights unchanged, wherein the feature weight updating mode is represented as follows:
Figure BDA0003036423060000074
wherein omegaj (t)，Ωk (t)Are two random sets of feature weights for the t-th generation; r is a random number uniformly distributed in the interval (0, 1).
Figure BDA0003036423060000075
And a third stage: adding BH algorithm to continuously optimize feature weights
Taking the optimal solution of the feature weight as a black hole, and continuously optimizing the feature weight by adopting a BH algorithm updating strategy, wherein the feature weight updating is represented as follows:
Figure BDA0003036423060000076
wherein omegaBHIs the location of the black hole.
In the iterative process, if the feature weight moves to be within the radius of the black hole, namely the distance d between the feature weight and the black hole is smaller than the radius r of the black holeBHThe feature weights are absorbed by the black holes, and after a group of feature weights are absorbed by the black holes and disappear, a new group of feature weights must be randomly generated in the weight population space to ensure that the number of the feature weights is unchanged, wherein the absorption mechanism of the feature weights and the black holes is expressed as follows:
Figure BDA0003036423060000081
wherein omegal (t)，Ωs (t)Are two sets of random feature weights for the t-th generation; r is a random number uniformly distributed in the interval of (0, 1); d is the distance between the feature weight and the black hole; r isBHIs the radius of the black hole.
The radius of the black hole is expressed as follows:
Figure BDA0003036423060000082
wherein f isBHThe fitness value of the black hole is obtained; f. ofiIs the ith set of feature weight fitness values.
And adding the BH algorithm to continuously optimize the feature weight by the BH-CS algorithm on the basis of the CS algorithm, and continuously evaluating the optimal feature weight in the current-generation weight population space until the maximum iteration times are met to obtain the global optimal feature weight.
(4) Obtaining a target case of the current working condition from the test case library, carrying out normalization processing, and calculating the input x of the characteristic variable after the normalization of the current working conditionn+1Input x of characteristic variable of historical working conditionkSimilarity x ofkObtaining solutions of 3 similar cases through a case retrieval model; the following formula is to calculate the target case xn+1And source case xkSimilarity of (2)k：
Figure BDA0003036423060000083
Wherein, ω ispIs the weight of the p characteristic variable, and meets the following constraint conditions:
Figure BDA0003036423060000084
(5) the average value of solutions of 3 similar cases with large similarity values is solved through case reuse, and therefore a set value of secondary air volume is obtained;
(6) forming a case by the target case and the set value, storing the case into a set database, and solving the next suboptimal setting;
(7) repeating the steps (3) to (6) to realize the secondary air volume setting process in the incineration process of the municipal domestic waste;
the invention combines a characteristic weight distribution method based on a BH-CS algorithm with a case reasoning technology, realizes intelligent setting of secondary air volume in the process of incinerating urban domestic garbage, respectively adopts four methods to verify the average fitting error of a set value of the secondary air volume in order to further verify the effectiveness of the set secondary air volume by the method, and uses historical working condition data in a comparison experiment. The case reasoning model based on the characteristic weight distributed by the BH-CS algorithm is recorded as BH-CSCBR, the cuckoo algorithm is recorded as CSCBR, the black hole algorithm is recorded as BHCBR, and the genetic algorithm is recorded as GACBR. The experimental results are as follows: the average fitting errors of the four methods are from high to low that GACBR is 9.21%, CSCBR is 8.79%, BHCBR is 7.94% and BH-CSCBR is 7.47%, and it can be seen that the average fitting error of BH-CSCBR is minimum, the set value of secondary air volume can be better fitted with an actual value, and the method is proved to be easier to obtain global optimal characteristic weight, so that a better set value of secondary air volume is obtained, and the precision of the intelligent setting model of secondary air volume CBR is improved.

Claims (1)

1. An intelligent setting method for secondary air quantity in an urban domestic garbage incineration process is characterized by comprising the following steps: the method specifically comprises the following steps: (1) establishing a set case library according to historical data of the urban life incineration process; the outlet temperature (x) of the secondary air heater is measured according to the characteristic variable1) Mean value of flue gas temperature (x) in a combustion chamber2) Mean value of flue gas temperature (x) of second combustion chamber3) Main steam flow at the boiler outlet (x)4) And the primary air volume (x)5) After normalization processing is carried out on the secondary air volume actual value y under the current working condition of the evaluation index, the secondary air volume actual value y is expressed into a characteristic vector form to form n source cases, and the n source cases are stored in a set case library; record each source case as CkExpressed in the following form:
Ck＝(xk,yk),k＝1,2,3,...,n (1)
wherein n is the total number of source cases; ckFor the kth source case, xkInputting normalized characteristic variables for the kth source case; y iskThe actual value of the secondary air volume is correspondingly normalized; x is the number ofkCan be expressed as:
xk＝(x1,k,x2,k,x3,k,x4,k,x5,k) (2)
setting the characteristic variable input of the current working condition normalization as xn+1As follows:
xn+1＝(x1,n+1,x2,n+1,x3,n+1,x4,n+1,x5,n+1) (3)
(2) initializing parameters; making the maximum iteration number MaxGeneration, the weight upper limit wu, the weight lower limit wl and the elimination probability Pa in the cuckoo algorithm;
(3) distributing characteristic weights based on a black hole cuckoo algorithm, wherein the whole algorithm is divided into three stages, namely a first stage, and selecting a fitness function; in the second stage, a CS algorithm is adopted to optimize the characteristic weight, and the characteristic weight is divided into Levy flight and Pa probability elimination; in the third stage, adding a BH algorithm to continuously optimize the characteristic weight;
the first stage is as follows: selecting a fitness function, and taking Root Mean Square Error (RMSE) of a secondary air volume CBR set model as the fitness function, wherein the Root Mean Square Error (RMSE) is expressed as follows:
Figure FDA0003036423050000011
wherein, ykIs the true value of the secondary air quantity of the kth case;
Figure FDA0003036423050000012
is the set value of the secondary air quantity of the kth case;
and a second stage: levy flight and elimination of feature weights with Pa probability
a) Levy flight is a random walk mechanism, and random update characteristic weights based on Levy flight are expressed as follows:
Figure FDA0003036423050000013
wherein omegai (t)And Ωi (t+1)Respectively representing the characteristic weights of the t generation and the t +1 generation; n is the number of groups of population space characteristic weights;
Figure FDA0003036423050000014
is a step size factor;
Figure FDA0003036423050000021
is a dot product; levy (λ) is a random number obtained by Levy flight;
Figure FDA0003036423050000022
and Levy (λ) as follows:
Figure FDA0003036423050000023
wherein the content of the first and second substances,
Figure FDA0003036423050000024
is a constant with a value of 0.01; omegabestIs the current generation optimal feature weight;
Levy(λ)～μ＝t-λ,(1＜λ≤3) (7)
Figure FDA0003036423050000025
wherein u and v are random numbers of standard normal distribution, namely the occurrence probability meets the normal distribution, and the probability of the numbers closer to the middle is higher, and the probability of the numbers closer to the two sides is lower; β is a constant over the interval (1,2), typically taken to be 1.5;
Figure FDA0003036423050000029
is represented as follows:
Figure FDA0003036423050000026
wherein r is the gamma function;
b) randomly generating feature weights with Pa probability
Recalculating the fitness of the feature weights, and evaluating the optimal feature weights, namely if the fitness of the feature weights generated after Levy flight is smaller, replacing the previous feature weights with the feature weights after Levy flight, detecting a group of worst feature weights in the weight population space according to the probability Pa, namely generating a random number r, if r is greater than Pa, representing the current feature weight difference, eliminating the current feature weight difference and randomly generating a new group of feature weights, otherwise, keeping the feature weights unchanged, wherein the value of Pa is always 0.5, and the feature weight updating mode is represented as follows:
Figure FDA0003036423050000027
wherein omegaj (t)，Ωk (t)Are two random sets of feature weights for the t-th generation; n is the number of groups of population space characteristic weights; r is a random number uniformly distributed in the interval of (0, 1);
Figure FDA0003036423050000028
and a third stage: adding BH algorithm to continuously optimize feature weights
Taking the optimal solution of the feature weight as a black hole, and continuously optimizing the feature weight by adopting a BH algorithm updating strategy, wherein the feature weight updating is represented as follows:
Figure FDA0003036423050000031
wherein, rand () is a random number uniformly distributed in the (0,1) interval; omegaBHIs the location of the black hole;
the radius of the black hole is defined as follows:
Figure FDA0003036423050000032
wherein f isBHThe fitness value of the black hole is obtained; f. ofiIs the ith group of feature weight fitness value;
in an iterative process, if the feature weight moves to the black hole radius rBHIn which the distance d between the feature weight and the black hole is smaller than the radius r of the black holeBHThe feature weights are absorbed by the black holes, and after a group of feature weights are absorbed by the black holes and disappear, a new group of feature weights must be randomly generated in the weight population space to ensure that the number of the feature weights is unchanged, wherein the absorption mechanism of the feature weights and the black holes is expressed as follows:
Figure FDA0003036423050000033
wherein omegal (t)Is a set of random feature weights for the t-th generation; omegas (t)Is another set of random feature weights for the t-th generation; rand () is a random number uniformly distributed in the (0,1) interval; d is the distance between the feature weight and the black hole; r isBHIs the radius of the black hole;
the BH-CS algorithm is added into the BH algorithm on the basis of the CS algorithm to continuously optimize the feature weight, the optimal feature weight in the current generation weight population space is continuously evaluated until the maximum iteration times are met, and the global optimal feature weight is obtained;
(4) obtaining solutions of 3 similar cases through a case retrieval model; calculating input x of characteristic variable after normalization of current working conditionn+1Input x of characteristic variable of historical working conditionkSimilarity of (2)kThe smaller the distance is, the more similar the distance is, the greater the similarity is, the more similar the working condition is, so as to weigh the Euclidean distanceFormula is an example, and the following formula is to calculate the target case xn+1And source case xkSimilarity of (2)k：
Figure FDA0003036423050000034
Wherein, ω ispIs the weight, x, of the p-th characteristic variablep,kInputting the p characteristic weight of the kth source case under the historical working condition; x is the number ofp,n+1Is the input of the p characteristic weight of the current working condition;
the following constraints are satisfied:
Figure FDA0003036423050000041
comparison Sk3 similar cases are retrieved;
(5) obtaining an average value of K similar case solutions through case reuse, thereby obtaining a set value of a target case solution;
(6) forming a case by the target case and the set value, storing the case into a set database, and solving the next suboptimal setting;
(7) and (5) repeating the steps (3) to (6) to realize the setting process of the secondary air volume in the incineration process of the municipal domestic waste.
CN202110444812.5A 2021-04-24 2021-04-24 Intelligent setting method for secondary air quantity in municipal solid waste incineration process Active CN113154404B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110444812.5A CN113154404B (en) 2021-04-24 2021-04-24 Intelligent setting method for secondary air quantity in municipal solid waste incineration process

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110444812.5A CN113154404B (en) 2021-04-24 2021-04-24 Intelligent setting method for secondary air quantity in municipal solid waste incineration process

Publications (2)

Publication Number Publication Date
CN113154404A true CN113154404A (en) 2021-07-23
CN113154404B CN113154404B (en) 2023-01-31

Family

ID=76870146

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110444812.5A Active CN113154404B (en) 2021-04-24 2021-04-24 Intelligent setting method for secondary air quantity in municipal solid waste incineration process

Country Status (1)

Country Link
CN (1) CN113154404B (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113742997A (en) * 2021-08-02 2021-12-03 北京工业大学 Intelligent air quantity optimization setting method for urban solid waste incineration process

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002267134A (en) * 2001-03-13 2002-09-18 Sumitomo Heavy Ind Ltd Combustion control system of refuse incinerator having no boiler facility
CN103900092A (en) * 2014-03-26 2014-07-02 广州环投技术设备有限公司 Automatic combustion control system for municipal solid waste incinerator
CN108224446A (en) * 2017-12-31 2018-06-29 北京工业大学 A kind of automatic combustion Study on Decision-making Method for Optimization of Refuse Incineration Process
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

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2002267134A (en) * 2001-03-13 2002-09-18 Sumitomo Heavy Ind Ltd Combustion control system of refuse incinerator having no boiler facility
CN103900092A (en) * 2014-03-26 2014-07-02 广州环投技术设备有限公司 Automatic combustion control system for municipal solid waste incinerator
CN108224446A (en) * 2017-12-31 2018-06-29 北京工业大学 A kind of automatic combustion Study on Decision-making Method for Optimization of Refuse Incineration Process
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

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
严爱军等: "案例推理分类器的权重分配及案例维护方法", 《计算机应用》 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113742997A (en) * 2021-08-02 2021-12-03 北京工业大学 Intelligent air quantity optimization setting method for urban solid waste incineration process

Also Published As

Publication number Publication date
CN113154404B (en) 2023-01-31

Similar Documents

Publication Publication Date Title
CN103064289B (en) Multiple-target operation optimizing and coordinating control method and device of garbage power generator
Zheng et al. A comparative study of optimization algorithms for low NOx combustion modification at a coal-fired utility boiler
CN105864797B (en) Real-time prediction system and method for boiler entering heat value of circulating fluidized bed household garbage incineration boiler
CN111144609A (en) Boiler exhaust emission prediction model establishing method, prediction method and device
CN104534507B (en) A kind of boiler combustion optimization control method
CN104763999A (en) Power plant pulverized coal boiler combustion performance online optimizing method and system
CN105020705A (en) Method and system for optimizing and controlling combustion performance of circulating fluidized bed boiler in real time
WO2024060488A1 (en) Method based on deep recurrent neural network and evolutionary computation for optimizing combustion of industrial boiler
CN113154404B (en) Intelligent setting method for secondary air quantity in municipal solid waste incineration process
CN111457392B (en) Intelligent setting method for air quantity in urban domestic garbage incineration process
CN112163376A (en) Extreme random tree furnace temperature prediction control method based on longicorn stigma search
CN114564894A (en) Soft measurement method for dioxin emission in grate furnace MSWI process based on residual error fitting mechanism simplified deep forest regression
Xu et al. A novel online combustion optimization method for boiler combining dynamic modeling, multi-objective optimization and improved case-based reasoning
CN105823080B (en) Model-free boiler combustion optimization control method based on numerical optimization extremum search
CN113742997A (en) Intelligent air quantity optimization setting method for urban solid waste incineration process
Xu et al. A new online optimization method for boiler combustion system based on the data-driven technique and the case-based reasoning principle
Teng et al. Prediction of particulate matter concentration in Chengdu based on improved differential evolution algorithm and BP neural network model
CN107977539A (en) Improvement neutral net boiler combustion system modeling method based on object combustion mechanism
CN116720446A (en) Method for monitoring thickness of slag layer of water-cooled wall of boiler in real time
CN113610260B (en) Method for predicting concentration of smoke components in urban household garbage incineration process
CN114943151A (en) MSWI process dioxin emission soft measurement method based on integrated T-S fuzzy regression tree
CN114527646A (en) Multi-loop quasi-diagonal recurrent neural network PID control method for urban solid waste incineration process
CN113408185A (en) Method for setting air flow of grate in incineration process of municipal domestic waste
Hou et al. Multiobjective Operation Optimization for Municipal Solid Waste Incineration Process
Manke A Performance Comparison of Different Back Propagation Neural Networks for Nitrogen Oxides Emission Prediction in Thermal Power Plant

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