CN110751344A - Power plant boiler operation optimization system and method based on intelligent visualization technology - Google Patents

Power plant boiler operation optimization system and method based on intelligent visualization technology Download PDF

Info

Publication number
CN110751344A
CN110751344A CN201911045805.7A CN201911045805A CN110751344A CN 110751344 A CN110751344 A CN 110751344A CN 201911045805 A CN201911045805 A CN 201911045805A CN 110751344 A CN110751344 A CN 110751344A
Authority
CN
China
Prior art keywords
boiler
vector
mapping
optimization
data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201911045805.7A
Other languages
Chinese (zh)
Inventor
鄢烈祥
梁钜亮
周力
刘立柱
彭愿
裴彬
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.)
Hangu Yunzhi Wuhan Technology Co Ltd
Original Assignee
Hangu Yunzhi Wuhan Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hangu Yunzhi Wuhan Technology Co Ltd filed Critical Hangu Yunzhi Wuhan Technology Co Ltd
Priority to CN201911045805.7A priority Critical patent/CN110751344A/en
Publication of CN110751344A publication Critical patent/CN110751344A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/067—Enterprise or organisation modelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
    • G06Q50/06—Electricity, gas or water supply
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
    • Y02E40/70—Smart grids as climate change mitigation technology in the energy generation sector
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Abstract

The invention discloses a power plant boiler operation optimization system and method based on an intelligent visualization technology, wherein the system collects original production data of a boiler; establishing an artificial neural network mapping dimension reduction model, and carrying out dimension reduction mapping on boiler original production data in a multidimensional space to a two-dimensional mapping plane; training and learning through a queue competition algorithm, and solving to obtain model parameter data; drawing the distribution of the boiler multi-dimensional space data points and the contour line of a multi-optimization target on a two-dimensional mapping plane; automatically identifying the optimal point and the stable optimization area of the multi-optimization target contour line of the boiler two-dimensional mapping plane by using an image identification technology; inversely mapping the predicted optimal point to a multidimensional space of original production data of the boiler to obtain operation optimization guidance data; and guiding the production operation of the boiler and checking the optimization effect of the boiler. The invention can realize the visualization of the boiler operation data, solve the multi-objective optimization problem of the boiler and effectively improve the operation efficiency of the boiler.

Description

Power plant boiler operation optimization system and method based on intelligent visualization technology
Technical Field
The invention relates to the technical field of boiler operation optimization, in particular to a power plant boiler operation optimization system and method based on an intelligent visualization technology.
Background
The existing technology has some problems aiming at the problems of multi-objective optimization of boiler operation and optimization with index constraint and determining a stable optimization area. Firstly, the operation optimization of the coal-fired boiler mainly comprises the optimization of improving the combustion efficiency of the boiler and reducing the emission of pollutants such as nitrogen oxides, which is a problem of multi-objective optimization, and the combustion process of the coal-fired boiler is a complex nonlinear process, so that the number of parameters influencing the combustion is very large, and extremely complex coupling relations exist among the parameters. And secondly, the optimization of the boiler operation should search an optimization point and a stable optimization area at the same time so as to ensure the economic safety of the optimization of the boiler operation.
At present, the operation and combustion optimization and adjustment technologies of domestic and foreign boilers mainly comprise three types: the method is characterized in that important parameters of boiler combustion are provided through an online detection technology, combustion optimization and adjustment are realized through modification of devices such as a heating surface and a combustor, and combustion optimization of the boiler is realized through an advanced control idea or an artificial intelligence technology. Because the traditional equipment modification technology has more applications, long construction period and high cost, and the problem of the operating efficiency of the boiler cannot be solved at all, the research on combustion optimization at home and abroad at present mostly focuses on the research of an optimization theory, is mainly based on the aspects of model prediction and multi-objective optimization technology, has a very clear optimization target, and can improve the combustion efficiency and reduce the pollution emission. However, because the boiler efficiency is the balance of the factors of mutual influence of multiple factors, the related theories are complex, the theories are applied less and immature, and the problems of multi-objective optimization and index-constrained optimization of boiler operation cannot be solved and a stable optimization area cannot be determined.
Disclosure of Invention
The invention aims to solve the technical problem of providing a power plant boiler operation optimization system and method based on an intelligent visualization technology aiming at the defects in the prior art.
The technical scheme adopted by the invention for solving the technical problems is as follows:
the invention provides a power plant boiler operation optimization system based on an intelligent visualization technology, which comprises the following units:
the boiler data acquisition unit is used for acquiring original production data of the boiler through an OPC protocol;
the special data management unit of the boiler is used for managing and storing original production data, model parameter data and operation optimization guidance data of the boiler;
the artificial neural network mapping dimension reduction model unit is used for establishing an artificial neural network mapping dimension reduction model and mapping the original boiler production data in a multidimensional space to a two-dimensional mapping plane in a dimension reduction manner;
the training and learning unit for the queue competition algorithm is used for solving and converting model parameters of the artificial neural network mapping model into a nonlinear optimization problem, training and learning are carried out through the queue competition algorithm, and model parameter data are obtained through solving;
the multi-objective function plane contour line unit is used for drawing the distribution of the multi-dimensional space data points of the boiler and the contour line of a multi-optimization target on a two-dimensional mapping plane;
the system comprises an image recognition technology forecast optimization point unit, a data processing unit and a data processing unit, wherein the image recognition technology forecast optimization point unit is used for automatically recognizing an optimal point and a stable optimization area of a multi-optimization target contour line of a boiler two-dimensional mapping plane by using an image recognition technology;
the inverse mapping algorithm unit is used for inversely mapping the predicted optimal point to a multi-dimensional space of original production data of the boiler to obtain operation optimization guide data, and the operation optimization guide data are an optimization point and a stable optimization area represented in the multi-dimensional space;
and the boiler operation optimization guidance unit is used for guiding the production operation of the boiler according to the operation optimization guidance data and checking the optimization effect of the boiler.
Further, the raw production data of the boiler of the present invention includes: coal feeding amount, coal low-grade heating value, water feeding flow, primary air quantity, secondary air quantity, oxygen quantity, boiler combustion efficiency and nitrogen oxide NOx emission;
the model parameter data includes: coal feeding amount, coal low-level heating value, water feeding flow, primary air quantity, secondary air quantity, oxygen quantity, boiler combustion efficiency, nitrogen oxide NOx emission, artificial neural network weight and inverse mapping step length;
the operation optimization guidance data includes: coal feeding amount, primary air quantity, secondary air quantity and oxygen quantity.
Further, the specific method for realizing the artificial neural network mapping dimension reduction model unit comprises the following steps:
raw production data for the boiler multidimensional space includes: coal feeding amount x1Primary air volume x2Secondary air quantity x3Oxygen amount x4Boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX；
Inputting vector X, vector X representing coal feeding quantity X1Primary air volume x2Secondary air quantity x3Oxygen amount x4；
Outputting vector Y which represents the boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX；
Establishing a boiler artificial neural network mapping dimension reduction model:
the formula from the input to the mapping plane is:
Figure BDA0002254107550000031
the formula from the mapping plane to the output is:
Y＝vPT
wherein:
Figure BDA0002254107550000032
Figure BDA0002254107550000034
Figure BDA0002254107550000035
Figure BDA0002254107550000036
wherein Z is1、Z2Is a two-dimensional plane vector, w and v are artificial neural network mapping dimension-reduction model weight vectors, and X is an input vector (coal supply amount X)1Primary air volume x2Secondary air quantity x3Oxygen amount x4)，wi1、wi2(i＝1,2,3,4)Is that the weight vector corresponds to a specific input vector xi(i＝1,2,3,4)Y is the output vector (boiler combustion efficiency η)gtNitrogen oxide NOx emission GNOX)，PTIs the transposed vector of P, vi1、vi2(i＝1,2,3,4,5,6)Is that the weight vector corresponds to a specific non-linear expansion vector pi(i＝1,2,3,4,5,6)The coefficient P is called a nonlinear expansion vector, and the function of the coefficient P is to enhance the nonlinear mapping and approximation capability of the boiler artificial neural network; in the boiler artificial neural network, an input vector X of a boiler 4-dimensional space is firstly mapped to a two-dimensional vector Z, and then the vector Z and an output vector Y are established to form a nonlinear mapping relation through the nonlinear expansion vector P and the effect of the artificial neural network mapping dimension reduction model weight vectors w and v; the purpose of reducing dimension is achieved by means of nonlinear mapping relation, and a distribution rule curve of a vector Y is described by a vector z on a two-dimensional plane.
Further, the specific method for realizing the training and learning unit of the queue competition algorithm comprises the following steps:
determining boiler artificial neural network mapping dimension reduction model weight vectors w and v through a queue competition algorithm; converting the determination problem of the boiler artificial neural network mapping model weight vectors w and v into a nonlinear optimization problem, and training and learning by a queue competition algorithm to solve, namely:
Figure BDA0002254107550000041
wherein n is the total number of sample data, t is the sample pattern, dk(t)，yk(t) given output and network output in t mode respectively, using a queue competition algorithm to obtain w, the formula of v is as follows:
Figure BDA0002254107550000042
Figure BDA0002254107550000043
wherein the training formula is:
Figure BDA0002254107550000044
Figure BDA0002254107550000045
Figure BDA0002254107550000046
the learning formula is:
v(k+1)＝v(k)+Δv+α(v(k)-v(k-1))
w(k+1)＝w(k)+Δw+α(w(k)-w(k-1))
where E is the error between the given output and the network output, wij,vkiIs the weight vector of the artificial neural network model, △ wij,△vkiWeight vector w representing artificial neural network modelij,vkiObey the delta learning rule, t is the sample pattern, dk(t)，yk(t) outputs given in the t-mode and the network outputs respectively,
Figure BDA0002254107550000047
representing partial derivative calculation, η are learning rate and momentum factor respectively, k is iteration number, z1、z2Is a two-dimensional plane vector, xjIs an input vector (coal feed amount x)1Primary air volume x2Secondary air quantity x3Oxygen amount x4)，ykIs the output vector (y)1Boiler combustion efficiency ηgt、y2Emission G of nitrogen oxides NOxNOX)。
Further, the specific method for realizing the multi-target function plane contour line unit comprises the following steps:
the multi-objective function plane contour line is a contour line which draws the distribution of boiler 4-dimensional space data points and a multi-optimization target on a two-dimensional mapping plane; the 4-dimensional spatial data points include: coal feeding amount x1Primary air volume x2Secondary air quantity x3Oxygen amount x4The multiple optimization targets include high combustion efficiency η of boilergtLow emission G of nitrogen oxides NOxNOX。
Further, the specific method for realizing the image recognition technology forecast optimization point unit comprises the following steps:
and automatically identifying the optimal point and the stable optimization area of the isoline of the two-dimensional mapping plane multi-optimization target by using an image identification processing technology comprising image processing software MATLAB and an image processing library in Python language.
Further, the specific method for implementing the inverse mapping algorithm unit of the present invention is as follows:
inversely mapping the forecast optimization points to the original 4-dimensional space of the boiler through an inverse mapping algorithm to obtain optimization points and stable optimization areas represented by original variables of the boiler; the formula of the inverse mapping algorithm is:
xc＝xa+β(xb-xa)
wherein x isa,xbRespectively, a point and a point b on the mapping plane correspond to a point in a two-dimensional plane in a 4-dimensional space of the boiler, and xcIs the corresponding point of any point c on the straight line of the two points a and b in the 4-dimensional space,β, called interpolation or extrapolation step size, whose value is equal to the ratio of the distance between points a, c and a, b, i.e.:
Figure BDA0002254107550000051
when the expression is interpolated, β is less than 1, and when the expression is extrapolated, β is more than 1.
The invention provides a power plant boiler operation optimization method based on an intelligent visualization technology, which comprises the following steps:
s1, configuring an OPC server and developing an OPC client program through an OPC protocol and a data one-way isolation technology, and establishing an OPC data acquisition network to acquire the original boiler production data in the DCS;
s2, managing and storing data required by the power plant boiler operation optimization system based on the intelligent visualization technology, wherein the data comprises original boiler production data, model parameter data and operation optimization guidance data;
s3, establishing a boiler artificial neural network mapping dimension reduction model, and carrying out dimension reduction mapping on boiler original production data in a multidimensional space to a two-dimensional mapping plane;
s4, converting the model parameter solution of the artificial neural network mapping model into a nonlinear optimization problem, training and learning through a queue competition algorithm, and solving to obtain model parameter data;
s5, drawing the distribution of the boiler multi-dimensional space data points and the contour line of the multi-optimization target on a two-dimensional mapping plane;
s6, automatically identifying the optimal point and the stable optimization area of the multi-optimization target contour line of the boiler two-dimensional mapping plane by using an image identification technology;
s7, inversely mapping the predicted optimal point to a multi-dimensional space of original production data of the boiler to obtain operation optimization guide data, wherein the operation optimization guide data are an optimization point and a stable optimization area represented in the multi-dimensional space;
and S8, guiding the production operation of the boiler according to the operation optimization guidance data, and checking the optimization effect.
Further, the specific method for establishing the artificial neural network mapping dimension reduction model in step S3 of the present invention is as follows:
raw production data for the boiler multidimensional space includes: coal feeding amount x1Primary air volume x2Secondary air quantity x3Oxygen amount x4Boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX；
Inputting vector X, vector X representing coal feeding quantity X1Primary air volume x2Secondary air quantity x3Oxygen amount x4；
Outputting vector Y which represents the boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX；
Establishing a boiler artificial neural network mapping dimension reduction model:
the formula from the input to the mapping plane is:
Figure BDA0002254107550000061
the formula from the mapping plane to the output is:
Y＝vPT
wherein:
Figure BDA0002254107550000062
Figure BDA0002254107550000063
Figure BDA0002254107550000064
Figure BDA0002254107550000071
Figure BDA0002254107550000072
wherein Z is1、Z2Is a two-dimensional plane vector, w and v are artificial neural network mapping dimension-reduction model weight vectors, and X is an input vector (coal supply amount X)1Primary air volume x2Secondary air quantity x3Oxygen amount x4)，wi1、wi2(i＝1,2,3,4)Is that the weight vector corresponds to a specific input vector xi(i＝1,2,3,4)Y is the output vector (boiler combustion efficiency η)gtNitrogen oxide NOx emission GNOX)，PTIs the transposed vector of P, vi1、vi2(i＝1,2,3,4,5,6)Is that the weight vector corresponds to a specific non-linear expansion vector pi(i＝1,2,3,4,5,6)The coefficient P is called a nonlinear expansion vector, and the function of the coefficient P is to enhance the nonlinear mapping and approximation capability of the boiler artificial neural network; in the boiler artificial neural network, an input vector X of a boiler 4-dimensional space is firstly mapped to a two-dimensional vector Z, and then the vector Z and an output vector Y are established to form a nonlinear mapping relation through the nonlinear expansion vector P and the effect of the artificial neural network mapping dimension reduction model weight vectors w and v; the purpose of reducing dimension is achieved by means of nonlinear mapping relation, and a distribution rule curve of a vector Y is described by a vector z on a two-dimensional plane.
Further, the specific method for implementing training and learning of the queuing competition algorithm in step S4 of the present invention is as follows:
determining boiler artificial neural network mapping dimension reduction model weight vectors w and v through a queue competition algorithm; converting the determination problem of the boiler artificial neural network mapping model weight vectors w and v into a nonlinear optimization problem, and training and learning by a queue competition algorithm to solve, namely:
Figure BDA0002254107550000073
wherein n is the total number of sample data, t is the sample pattern, dk(t)，yk(t) given output and network output in t mode respectively, using a queue competition algorithm to obtain w, the formula of v is as follows:
Figure BDA0002254107550000074
wherein the training formula is:
Figure BDA0002254107550000077
Figure BDA0002254107550000081
the learning formula is:
v(k+1)＝v(k)+Δv+α(v(k)-v(k-1))
w(k+1)＝w(k)+Δw+α(w(k)-w(k-1))
where E is the error between the given output and the network output, wij,vkiIs the weight vector of the artificial neural network model, △ wij,△vkiWeight vector w representing artificial neural network modelij,vkiObey the delta learning rule, t is the sample pattern, dk(t)，yk(t) outputs given in the t-mode and the network outputs respectively,
Figure BDA0002254107550000082
representing partial derivative calculation, η are learning rate and momentum factor respectively, k is iteration number, z1、z2Is a two-dimensional plane vector, xjIs an input vector (coal feed amount x)1Primary air volume x2Secondary air quantity x3Oxygen amount x4)，ykIs the output vector (y)1Boiler combustion efficiency ηgt、y2Emission G of nitrogen oxides NOxNOX)。
The invention has the following beneficial effects: the invention relates to a power plant boiler operation optimization system and method based on an intelligent visualization technology, which are characterized in that an artificial neural network and a queue competition algorithm are applied to map production data of a multi-dimensional space of a power plant boiler on a plane in a dimensionality reduction manner, a contour line of a multi-objective function is generated, the appearance and the characteristics of the multi-dimensional operation space are displayed in a panoramic manner, and an optimization point and a stable optimization area are found on the plane by an image recognition technology. And inversely mapping the optimization points to a multi-dimensional space through an inverse mapping method for guiding the operation of the power plant boiler.
Drawings
The invention will be further described with reference to the accompanying drawings and examples, in which:
FIG. 1 is a schematic diagram of a system architecture of an embodiment of the present invention;
fig. 2 is a schematic workflow diagram of an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As shown in fig. 1, the power plant boiler operation optimization system based on the intelligent visualization technology of the embodiment of the present invention includes the following units:
the boiler data acquisition unit is used for configuring an OPC server and developing an OPC client program to establish an OPC data acquisition network to acquire original data (coal supply quantity, coal heat value and other physical properties, water supply flow, water supply temperature, primary air quantity, secondary air quantity, oxygen quantity, main steam flow of the boiler, main steam pressure, reheating steam flow of the boiler, reheating steam pressure, exhaust gas temperature, steam pocket pressure, steam pocket water level and nitrogen oxide NOx emission) of a boiler in the DCS through an OPC protocol and a data one-way isolation technology.
And the boiler special database management unit summarizes the data required by the whole system. Namely, boiler raw data (coal supply amount, coal calorific value and other physical properties, water supply flow, water supply temperature, primary air quantity, secondary air quantity, oxygen quantity, boiler main steam flow, main steam pressure, boiler reheat steam flow, reheat steam pressure, exhaust gas temperature, drum pressure, drum water level, nitrogen oxide NOx emission), model parameter data (coal supply amount, coal low calorific value, water supply flow, primary air quantity, secondary air quantity, oxygen quantity, main steam flow, boiler combustion efficiency, nitrogen oxide NOx emission, artificial neural network weight, inverse mapping step length and the like), operation optimization guidance data (coal supply amount, primary air quantity, secondary air quantity, oxygen quantity).
The artificial neural network maps a dimension reduction model unit to convert the production data (coal feeding amount x) of the multidimensional space of the power plant boiler1Primary air volume x2Secondary air quantity x3Oxygen amount x4Boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX) And (4) dimension reduction mapping is carried out on the plane, and the visualized image is formed. Namely, an artificial neural network mapping dimension reduction model is built, and a vector X (coal supply quantity X) is input1Primary air volume x2Secondary air quantity x3Oxygen amount x4) First mapped to a two-dimensional plane, then expanded by non-linearity, superimposed, and finally output. The information transfer process is as follows:
the formula from the input to the mapping plane is:
Figure BDA0002254107550000091
the formula from the mapping plane to the output is:
Y＝vPT(2)
wherein:
Figure BDA0002254107550000092
Figure BDA0002254107550000102
Figure BDA0002254107550000103
Figure BDA0002254107550000104
wherein Z is1、Z2Is a two-dimensional plane vector, w and v are artificial neural network mapping dimension-reduction model weight vectors, and X is an input vector (coal supply amount X)1Primary air volume x2Secondary air quantity x3Oxygen amount x4)，wi1、wi2(i＝1,2,3,4)Is that the weight vector corresponds to a specific input vector xi(i＝1,2,3,4)Y is the output vector (boiler combustion efficiency η)gtNitrogen oxide NOx emission GNOX)，PTIs the transposed vector of P, vi1、vi2(i＝1,2,3,4,5,6)Is that the weight vector corresponds to a specific non-linear expansion vector pi(i＝1,2,3,4,5,6)The function of the coefficient P is to enhance the nonlinear mapping and approximation capability of the boiler artificial neural network, and the coefficient P is an expansion form of a quadratic polynomial. In the boiler artificial neural network, an input vector X (coal feeding quantity X) of a boiler 4-dimensional space1Primary air volume x2Secondary air quantity x3Oxygen amount x4) Firstly mapping to a two-dimensional vector Z, then enabling the Z to pass through a nonlinear expansion vector P, the action of a network weight matrix v and an output vector Y (the boiler combustion efficiency η)gtNitrogen oxide NOx emission GNOX) A non-linear mapping relationship is established. By means of the mapping relation, the purpose of reducing the dimension is achieved, and a distribution rule curve of a vector Y is described by a vector z on a two-dimensional plane.
And (3) training and learning a unit by a queue competition algorithm, and determining boiler artificial neural network mapping dimension reduction model weight vectors w, v. The determination of the boiler artificial neural network mapping model weight vector w, v can be converted into a nonlinear optimization problem (formula 8), and is solved by training and learning of a queue competition algorithm (formulas 9 to 15), namely:
Figure BDA0002254107550000105
wherein n is the total number of sample data, t is the sample pattern, dk(t)，yk(t) given output and network output in t mode respectively, using a queue competition algorithm to obtain w, the formula of v is as follows:
Figure BDA0002254107550000106
Figure BDA0002254107550000111
wherein the training formula is
Figure BDA0002254107550000112
Figure BDA0002254107550000113
Figure BDA0002254107550000114
The learning formula is:
v(k+1)＝v(k)+Δv+α(v(k)-v(k-1)) (14)
w(k+1)＝w(k)+Δw+α(w(k)-w(k-1)) (15)
where E is the error between the given output and the network output, wij,vkiIs the weight vector of the artificial neural network model, △ wij,△vkiWeight vector w representing artificial neural network modelij,vkiObey the delta learning rule, t is the sample pattern, dk(t)，yk(t) outputs given in the t-mode and the network outputs respectively,
Figure BDA0002254107550000115
representing partial derivative calculation, η are learning rate and momentum factor respectively, k is iteration number, z1、z2Is a two-dimensional plane vector, xjIs an input vector (coal feed amount x)1Primary air volume x2Secondary air quantity x3Oxygen amount x4)，ykIs the output vector (y)1Boiler combustion efficiency ηgt、y2Emission G of nitrogen oxides NOxNOX)。
A multi-objective function plane contour line unit draws boiler 4-dimensional space data points (coal feeding amount x) on a two-dimensional mapping plane1Primary air volume x2Secondary air quantity x3Oxygen amount x4) Distribution of (2) and multiple optimization objectives (high boiler combustion efficiency η)gtLow emission G of nitrogen oxides NOxNOX) The contour line of (2).
An image recognition technology forecast optimization point unit automatically recognizes a boiler two-dimensional mapping plane multi-optimization target (high boiler combustion efficiency η) by utilizing an image recognition technology (MATLAB and Python both have related image processing libraries)gtLow emission G of nitrogen oxides NOxNOX) The optimal point and stable optimization region of the contour are then inverse mapped to the original 4-dimensional space of the boiler by the inverse mapping algorithm (equations 16 to 17) described below.
The inverse mapping algorithm unit inversely maps the forecast optimization points to the original 4-dimensional space of the boiler through an inverse mapping algorithm (formulas 16 to 17) to obtain the original variables (coal feeding amount x) of the boiler1Primary air volume x2Secondary air quantity x3Oxygen amount x4) The represented optimization points and the stable optimization region. The inverse mapping can be done using the following inverse mapping algorithm (interpolation, extrapolation equations 16 to 17):
xc＝xa+β(xb-xa) (16)
wherein x isa,xbRespectively, a point and a point b on the mapping plane correspond to a point in a two-dimensional plane in a 4-dimensional space of the boiler, and xcIs the corresponding point of any point c on the straight line of the two points a, b in the 4-dimensional space, β is called the interpolation or extrapolation step size, and its value is equal to the ratio of the distance between the two points a, c and the distance between the two points a, b, i.e.:
Figure BDA0002254107550000121
when interpolated, β < 1, and when extrapolated, β > 1The extrapolation direction is from a → b to c, e.g., from b → a to c, and the two points a, b in the above equation are exchanged. The inverse mapping algorithm (equations 16 to 17) described above establishes any point M on the mapping plane and point x of the original boiler 4-dimensional spaceMOne-to-one correspondence between them.
Boiler operation optimization guiding unit, using original variable (coal feeding quantity x) of boiler1Primary air volume x2Secondary air quantity x3Oxygen amount x4) Optimization point of representation (x)1o、x2o、x3o、x4o) Is used for guiding the production operation of the boiler, namely judging the current values of the coal feeding quantity, the primary air quantity, the secondary air quantity and the oxygen quantity of the boiler, namely the coal feeding quantity x1tPrimary air volume x2tSecondary air quantity x3tOxygen amount x4tThe current value of the boiler is fed to the coal quantity x1tPrimary air volume x2tSecondary air quantity x3tOxygen amount x4tFeeding coal quantity x to optimized value1oPrimary air volume x2oSecondary air quantity x3oOxygen amount x4oThe control is adjusted according to the current value, and the whole system is coordinated and controlled by 10%, 30% and 50% … 100% execution depth from the current value to the optimized value respectively.
The special boiler database management unit acquires original boiler data in the DCS system and adjusts the adjustable boiler space data (coal feeding amount x)1Primary air volume x2Secondary air quantity x3Oxygen amount x4) The data are transmitted to a power plant boiler operation optimization system, and the power plant boiler operation optimization system adjusts the multidimensional space of the boiler (coal feeding amount x) based on the intelligent visualization technology1Primary air volume x2Secondary air quantity x3Oxygen amount x4) By dimension reduction mapping (equations 1 to 7) to a two-dimensional plane Z which displays multiple targets of the boiler (high boiler combustion efficiency η)gtLow emission G of nitrogen oxides NOxNOX) Finding an optimal point c and a stable optimization area a → b by an image recognition technology, mapping the optimal point c to the original 4-dimensional space of the boiler through inverse mapping, and using the original variable (coal feeding amount x) of the boiler1oPrimary air volume x2oSecondary air quantity x3oOxygen amount x4o) To indicate and guide the optimization operation of the boiler.
As shown in fig. 2, in the power plant boiler operation optimization method based on the intelligent visualization technology of the embodiment of the present invention, the power plant boiler operation optimization system extracts important boiler operation data cleaned by data from the boiler dedicated database management unit, automatically establishes the boiler artificial neural network dimension reduction optimization model (formulas 1 to 7) by combining with the boiler design file, the boiler combustion mechanism, the boiler operation requirement, and the like, then determines the mapping dimension reduction model weight vectors w and v by training and learning (formulas 9 to 15) the boiler artificial neural network mapping dimension reduction model through the queue competition algorithm, and reduces the dimension of the boiler multidimensional space production data to the two-dimensional plane Z to draw the boiler multiple targets (high boiler combustion efficiency η)gtLow emission G of nitrogen oxides NOxNOX) Forecasting an optimized point c by using a visualization technology and an image recognition technology through a function contour line, and mapping an optimal point c and a stable optimized area a → b to an original multi-dimensional space (coal feeding amount x) of the boiler through an inverse mapping algorithm (formulas 16 to 17)1oPrimary air volume x2oSecondary air quantity x3oOxygen amount x4o) The method guides the production and operation of the boiler, simultaneously, checks the optimization effect and corrects the parameters of the optimization model, and has a self-learning function.
Compared with the prior art, the technical scheme of the invention has the following advantages:
(1) a boiler artificial neural network mapping dimension reduction model (formulas 1 to 7) and a queue competition algorithm (formulas 9 to 15) are applied, the 4-dimensional space data (coal feeding amount x1, primary air amount x2, secondary air amount x3 and oxygen amount x4) of the boiler are mapped to a plane Z in a dimension reduction mode, a visual image is formed, and an optimal point c and a stable optimization area can be found out through an image recognition technology.
(2) The invention divides the structural model (formula 1 to formula 7) of the boiler artificial neural network into two parts of linear mapping (formula 1) and nonlinear mapping (formula 2), so that the points a, b and c on the mapping plane can be inversely mapped (formula 16 to 17) to the original 4-dimensional space (coal feeding amount x1, primary air amount x2, secondary air amount x3 and oxygen amount x4) by means of two parameters of direction and step length, and the problem of inverse mapping is successfully solved.
(3) Compared with the traditional optimization method, the technical scheme of the invention can solve the problems of multi-target optimization of boiler operation and optimization with index constraint, and can realize the multi-index (high boiler combustion efficiency η gt and low nitrogen oxide NOx emission G) of the boilerNOX) The contour lines are mapped on the same plane Z, and the combustion efficiency η gt of the boiler and the emission G of nitrogen oxides NOx can be determined through the decision of image recognition technology (MATLAB, Python image processing library)NOXThe optimization area with primary and secondary consideration and balance coordination provides an effective means for optimizing the multi-index optimization problem of boiler energy consumption, pollution, cost and the like. In addition, the technical scheme of the invention can determine the boiler optimization operation area meeting the index constraints (formulas 1 to 7) on the mapping plane, thereby effectively solving the constrained boiler operation optimization problem.
(4) An optimized region in which the operation of the boiler is stable can be determined. When certain optimized parameter(s) (coal supply amount x1, primary air amount x2, secondary air amount x3 and oxygen amount x4) change to measurable but uncontrollable parameters, the boiler operating point can be shifted from one point a to another point c in the optimized area by changing other boiler operating parameters (coal supply amount x1, primary air amount x2, secondary air amount x3 and oxygen amount x 4).
(5) Has the function of self-learning. The method can continuously learn and summarize the rule from the data collected by the boiler, automatically correct the optimal operation direction of the boiler, or update the optimal operation point c and the optimal operation area of the boiler.
After the power plant boiler operation optimization system based on the intelligent visualization technology is implemented, when the optimal point c is located inside a boiler sample data space, the optimal point c can be found, when the optimal point c is located outside the boiler sample data space, the optimization direction can be determined, and the optimization target (high boiler combustion efficiency η, low nitrogen oxide NOx emission G) can be generally achieved through optimizationNOX) The improvement is 5-30%.
It will be understood that modifications and variations can be made by persons skilled in the art in light of the above teachings and all such modifications and variations are intended to be included within the scope of the invention as defined in the appended claims.

Claims (10)

1. A power plant boiler operation optimization system based on intelligent visualization technology is characterized by comprising the following units:
the boiler data acquisition unit is used for acquiring original production data of the boiler through an OPC protocol;
the special data management unit of the boiler is used for managing and storing original production data, model parameter data and operation optimization guidance data of the boiler;
the artificial neural network mapping dimension reduction model unit is used for establishing an artificial neural network mapping dimension reduction model and mapping the original boiler production data in a multidimensional space to a two-dimensional mapping plane in a dimension reduction manner;
the training and learning unit for the queue competition algorithm is used for solving and converting model parameters of the artificial neural network mapping model into a nonlinear optimization problem, training and learning are carried out through the queue competition algorithm, and model parameter data are obtained through solving;
the multi-objective function plane contour line unit is used for drawing the distribution of the multi-dimensional space data points of the boiler and the contour line of a multi-optimization target on a two-dimensional mapping plane;
the system comprises an image recognition technology forecast optimization point unit, a data processing unit and a data processing unit, wherein the image recognition technology forecast optimization point unit is used for automatically recognizing an optimal point and a stable optimization area of a multi-optimization target contour line of a boiler two-dimensional mapping plane by using an image recognition technology;
the inverse mapping algorithm unit is used for inversely mapping the predicted optimal point to a multi-dimensional space of original production data of the boiler to obtain operation optimization guide data, and the operation optimization guide data are an optimization point and a stable optimization area represented in the multi-dimensional space;
and the boiler operation optimization guidance unit is used for guiding the production operation of the boiler according to the operation optimization guidance data and checking the optimization effect of the boiler.
2. A power plant boiler operation optimization system based on intelligent visualization technology as claimed in claim 1, characterized in that boiler raw production data comprises: coal feeding amount, coal low-grade heating value, water feeding flow, primary air quantity, secondary air quantity, oxygen quantity, boiler combustion efficiency and nitrogen oxide NOx emission;
the model parameter data includes: coal feeding amount, coal low-level heating value, water feeding flow, primary air quantity, secondary air quantity, oxygen quantity, boiler combustion efficiency, nitrogen oxide NOx emission, artificial neural network weight and inverse mapping step length;
the operation optimization guidance data includes: coal feeding amount, primary air quantity, secondary air quantity and oxygen quantity.
3. A power plant boiler operation optimization system based on an intelligent visualization technology as claimed in claim 2, characterized in that the specific method for realizing the artificial neural network mapping dimension reduction model unit is as follows:
raw production data for the boiler multidimensional space includes: coal feeding amount x1Primary air volume x2Secondary air quantity x3Oxygen amount x4Boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX；
Inputting vector X, vector X representing coal feeding quantity X1Primary air volume x2Secondary air quantity x3Oxygen amount x4；
Outputting vector Y which represents the boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX；
Establishing a boiler artificial neural network mapping dimension reduction model:
the formula from the input to the mapping plane is:
Figure FDA0002254107540000021
the formula from the mapping plane to the output is:
Y＝vPT
wherein:
Figure FDA0002254107540000022
Figure FDA0002254107540000023
Figure FDA0002254107540000025
Figure FDA0002254107540000026
wherein Z is1、Z2Is a two-dimensional plane vector, w and v are artificial neural network mapping dimension-reduction model weight vectors, X is an input vector, and the coal supply amount X1Primary air volume x2Secondary air quantity x3Oxygen amount x4，wi1、wi2Is that the weight vector corresponds to a specific input vector xiY is the output vector, boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX，PTIs the transposed vector of P, vi1、vi2Is that the weight vector corresponds to a specific non-linear expansion vector piThe coefficient P is called a nonlinear expansion vector, and the function of the coefficient P is to enhance the nonlinear mapping and approximation capability of the boiler artificial neural network; in the boiler artificial neural network, an input vector X of a boiler 4-dimensional space is firstly mapped to a two-dimensional vector Z, and then the vector Z and an output vector Y are established to form a nonlinear mapping relation through the nonlinear expansion vector P and the effect of the artificial neural network mapping dimension reduction model weight vectors w and v; the purpose of reducing dimension is achieved by means of nonlinear mapping relation, and a distribution rule curve of a vector Y is described by a vector z on a two-dimensional plane.
4. The power plant boiler operation optimization system based on the intelligent visualization technology as claimed in claim 3, wherein the specific method for realizing the train competition algorithm training and learning unit is as follows:
determining boiler artificial neural network mapping dimension reduction model weight vectors w and v through a queue competition algorithm; converting the determination problem of the boiler artificial neural network mapping model weight vectors w and v into a nonlinear optimization problem, and training and learning by a queue competition algorithm to solve, namely:
Figure FDA0002254107540000031
wherein n is the total number of sample data, t is the sample pattern, dk(t)，yk(t) given output and network output in t mode respectively, using a queue competition algorithm to obtain w, the formula of v is as follows:
Figure FDA0002254107540000032
Figure FDA0002254107540000033
wherein the training formula is:
Figure FDA0002254107540000034
Figure FDA0002254107540000035
Figure FDA0002254107540000036
the learning formula is:
v(k+1)＝v(k)+Δv+α(v(k)-v(k-1))
w(k+1)＝w(k)+Δw+α(w(k)-w(k-1))
training a formula through a queue competition algorithm for continuous iterative convergence, continuously correcting a learning formula, iterating the weight vectors w and v of the artificial neural network mapping model to a certain number of times or approaching a fixed value, and outputting to obtain final results, namely the weight vectors w and v of the artificial neural network mapping model;
where E is between the given output and the network outputError, wij,vkiIs the weight vector of the artificial neural network model, △ wij,△vkiWeight vector w representing artificial neural network modelij,vkiObey the delta learning rule, t is the sample pattern, dk(t)，yk(t) outputs given in the t-mode and the network outputs respectively,representing partial derivative calculation, η are learning rate and momentum factor respectively, k is iteration number, z1、z2Is a two-dimensional plane vector, xjIs an input vector, the coal feed amount x1Primary air volume x2Secondary air quantity x3Oxygen amount x4，ykIs the output vector, y1Boiler combustion efficiency ηgt、y2Emission G of nitrogen oxides NOxNOX。
5. A power plant boiler operation optimization system based on an intelligent visualization technology as claimed in claim 3, wherein the specific method for realizing the multi-objective function plane contour line unit is as follows:
the multi-objective function plane contour line is a contour line which draws the distribution of boiler 4-dimensional space data points and a multi-optimization target on a two-dimensional mapping plane; the 4-dimensional spatial data points include: coal feeding amount x1Primary air volume x2Secondary air quantity x3Oxygen amount x4The multiple optimization targets include high combustion efficiency η of boilergtLow emission G of nitrogen oxides NOxNOX。
6. A power plant boiler operation optimization system based on intelligent visualization technology as claimed in claim 5, characterized in that, the specific method for realizing the image recognition technology forecast optimization point unit is as follows:
and automatically identifying the optimal point and the stable optimization area of the isoline of the two-dimensional mapping plane multi-optimization target by using an image identification processing technology comprising image processing software MATLAB and an image processing library in Python language.
7. A power plant boiler operation optimization system based on intelligent visualization technology as claimed in claim 3, characterized in that, the specific method for realizing the inverse mapping algorithm unit is as follows:
inversely mapping the forecast optimization points to the original 4-dimensional space of the boiler through an inverse mapping algorithm to obtain optimization points and stable optimization areas represented by original variables of the boiler; the formula of the inverse mapping algorithm is:
xc＝xa+β(xb-xa)
wherein x isa,xbRespectively, a point and a point b on the mapping plane correspond to a point in a two-dimensional plane in a 4-dimensional space of the boiler, and xcIs the corresponding point of any point c on the straight line of the two points a, b in the 4-dimensional space, β is called the interpolation or extrapolation step size, and its value is equal to the ratio of the distance between the two points a, c and the distance between the two points a, b, i.e.:
Figure FDA0002254107540000042
when the expression is interpolated, β is less than 1, and when the expression is extrapolated, β is more than 1.
8. A power plant boiler operation optimization method based on an intelligent visualization technology is characterized by comprising the following steps:
s1, configuring an OPC server and developing an OPC client program through an OPC protocol and a data one-way isolation technology, and establishing an OPC data acquisition network to acquire the original boiler production data in the DCS;
s2, managing and storing data required by the power plant boiler operation optimization system based on the intelligent visualization technology, wherein the data comprises original boiler production data, model parameter data and operation optimization guidance data;
s3, establishing a boiler artificial neural network mapping dimension reduction model, and carrying out dimension reduction mapping on boiler original production data in a multidimensional space to a two-dimensional mapping plane;
s4, converting the model parameter solution of the artificial neural network mapping model into a nonlinear optimization problem, training and learning through a queue competition algorithm, and solving to obtain model parameter data;
s5, drawing the distribution of the boiler multi-dimensional space data points and the contour line of the multi-optimization target on a two-dimensional mapping plane;
s6, automatically identifying the optimal point and the stable optimization area of the multi-optimization target contour line of the boiler two-dimensional mapping plane by using an image identification technology;
s7, inversely mapping the predicted optimal point to a multi-dimensional space of original production data of the boiler to obtain operation optimization guide data, wherein the operation optimization guide data are an optimization point and a stable optimization area represented in the multi-dimensional space;
and S8, guiding the production operation of the boiler according to the operation optimization guidance data, and checking the optimization effect.
9. A power plant boiler operation optimization method based on an intelligent visualization technology as claimed in claim 8, wherein the specific method for establishing the artificial neural network mapping dimension reduction model in the step S3 is as follows:
raw production data for the boiler multidimensional space includes: coal feeding amount x1Primary air volume x2Secondary air quantity x3Oxygen amount x4Boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX；
Inputting vector X, vector X representing coal feeding quantity X1Primary air volume x2Secondary air quantity x3Oxygen amount x4；
Outputting vector Y which represents the boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX；
Establishing a boiler artificial neural network mapping dimension reduction model:
the formula from the input to the mapping plane is:
Figure FDA0002254107540000051
the formula from the mapping plane to the output is:
Y＝vPT
wherein:
Figure FDA0002254107540000061
Figure FDA0002254107540000062
Figure FDA0002254107540000063
Figure FDA0002254107540000064
Figure FDA0002254107540000065
wherein Z is1、Z2Is a two-dimensional plane vector, w and v are artificial neural network mapping dimension-reduction model weight vectors, X is an input vector, and the coal supply amount X1Primary air volume x2Secondary air quantity x3Oxygen amount x4，wi1、wi2Is that the weight vector corresponds to a specific input vector xiY is the output vector, boiler combustion efficiency ηgtNitrogen oxide NOx emission GNOX，PTIs the transposed vector of P, vi1、vi2Is that the weight vector corresponds to a specific non-linear expansion vector piThe coefficient P is called a nonlinear expansion vector, and the function of the coefficient P is to enhance the nonlinear mapping and approximation capability of the boiler artificial neural network; in the boiler artificial neural network, an input vector X of a boiler 4-dimensional space is firstly mapped to a two-dimensional vector Z, and then the vector Z and an output vector Y are established to form a nonlinear mapping relation through the nonlinear expansion vector P and the effect of the artificial neural network mapping dimension reduction model weight vectors w and v; the purpose of reducing dimension is achieved by means of nonlinear mapping relation, and a distribution rule curve of a vector Y is described by a vector z on a two-dimensional plane.
10. A power plant boiler operation optimization method based on an intelligent visualization technology as claimed in claim 9, wherein the concrete method for realizing the training and learning of the queue competition algorithm in the step S4 is as follows:
determining boiler artificial neural network mapping dimension reduction model weight vectors w and v through a queue competition algorithm; converting the determination problem of the boiler artificial neural network mapping model weight vectors w and v into a nonlinear optimization problem, and training and learning by a queue competition algorithm to solve, namely:
Figure FDA0002254107540000071
wherein n is the total number of sample data, t is the sample pattern, dk(t)，yk(t) given output and network output in t mode respectively, using a queue competition algorithm to obtain w, the formula of v is as follows:
Figure FDA0002254107540000072
Figure FDA0002254107540000073
wherein the training formula is:
Figure FDA0002254107540000074
the learning formula is:
v(k+1)＝v(k)+Δv+α(v(k)-v(k-1))
w(k+1)＝w(k)+Δw+α(w(k)-w(k-1))
where E is the error between the given output and the network output, wij,vkiIs the weight vector of the artificial neural network model, △ wij,△vkiWeight vector w representing artificial neural network modelij,vkiObey the delta learning rule, t is the sample pattern, dk(t)，yk(t) outputs given in the t-mode and the network outputs respectively,representing partial derivative calculation, η are learning rate and momentum factor respectively, k is iteration number, z1、z2Is a two-dimensional plane vector, xjIs an input vector, the coal feed amount x1Primary air volume x2Secondary air quantity x3Oxygen amount x4，ykIs the output vector, y1Boiler combustion efficiency ηgt、y2Emission G of nitrogen oxides NOxNOX。
CN201911045805.7A 2019-10-30 2019-10-30 Power plant boiler operation optimization system and method based on intelligent visualization technology Pending CN110751344A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911045805.7A CN110751344A (en) 2019-10-30 2019-10-30 Power plant boiler operation optimization system and method based on intelligent visualization technology

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911045805.7A CN110751344A (en) 2019-10-30 2019-10-30 Power plant boiler operation optimization system and method based on intelligent visualization technology

Publications (1)

Publication Number Publication Date
CN110751344A true CN110751344A (en) 2020-02-04

Family

ID=69281215

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911045805.7A Pending CN110751344A (en) 2019-10-30 2019-10-30 Power plant boiler operation optimization system and method based on intelligent visualization technology

Country Status (1)

Country Link
CN (1) CN110751344A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105631151A (en) * 2016-01-06 2016-06-01 陈威宇 Modeling method for pulverized coal fired boiler combustion optimization
CN106327021A (en) * 2016-08-31 2017-01-11 西安艾贝尔科技发展有限公司 Boiler combustion optimization air distribution method based on online model prediction
CN108644805A (en) * 2018-05-08 2018-10-12 南京归图科技发展有限公司 Boiler intelligent combustion optimal control method based on big data
CN110017470A (en) * 2019-02-27 2019-07-16 中国大唐集团科学技术研究院有限公司火力发电技术研究院 Boiler combustion multiobjective optimization control method based on differential evolution algorithm

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105631151A (en) * 2016-01-06 2016-06-01 陈威宇 Modeling method for pulverized coal fired boiler combustion optimization
CN106327021A (en) * 2016-08-31 2017-01-11 西安艾贝尔科技发展有限公司 Boiler combustion optimization air distribution method based on online model prediction
CN108644805A (en) * 2018-05-08 2018-10-12 南京归图科技发展有限公司 Boiler intelligent combustion optimal control method based on big data
CN110017470A (en) * 2019-02-27 2019-07-16 中国大唐集团科学技术研究院有限公司火力发电技术研究院 Boiler combustion multiobjective optimization control method based on differential evolution algorithm

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
鄢烈祥 等: "基于人工神经网络降维映射的统计优化方法" *
鄢烈祥 等: "降维映射分析法及其应用" *

Similar Documents

Publication Publication Date Title
CN103576655B (en) A kind of power boiler burning subspace modeling and Multipurpose Optimal Method and system
US8185216B2 (en) Plant controlling device and method, thermal power plant, and its control method
CN107726358A (en) Boiler Combustion Optimization System and method based on CFD numerical simulations and intelligent modeling
CN105910169A (en) Urban heating system heating network regulating method and system based on mechanism model prediction control
CN106642067A (en) Liquid level control system and method of boiler
CN105546508B (en) Thermal power plant&#39;s Control on Main-steam Temperature and method based on event trigger mechanism
CN105974793A (en) Power plant boiler combustion intelligent control method
CN103744294B (en) Based on the multiple goal soot blowing and optimal method of fuzzy control, server and system
Liu et al. Attention mechanism-aided data-and knowledge-driven soft sensors for predicting blast furnace gas generation
CN109695892B (en) Method and device for protecting wall temperature of heating surface of boiler
CN107016176A (en) A kind of hybrid intelligent overall boiler burning optimization method
Hou et al. Coordinated control system modeling of ultra-supercritical unit based on a new fuzzy neural network
WO2019237316A1 (en) Knowledge-transfer-based modeling method for blast furnace coal gas scheduling system
Niu et al. Case-based reasoning based on grey-relational theory for the optimization of boiler combustion systems
CN104807039A (en) Variable dimensionality reduction modeling method for boiler combustion optimization
Zhou et al. Modeling error PDF optimization based wavelet neural network modeling of dynamic system and its application in blast furnace ironmaking
CN117211973A (en) Ship liquid ammonia fuel supply control system and method
CN105808945B (en) A kind of hybrid intelligent boiler efficiency burning optimization method
CN108762086B (en) Secondary reheat steam temperature control device and control system based on model predictive control
CN111881120B (en) Intelligent operation optimization method for boiler
Vieira et al. Methodology for ranking controllable parameters to enhance operation of a steam generator with a combined Artificial Neural Network and Design of Experiments approach
CN110751344A (en) Power plant boiler operation optimization system and method based on intelligent visualization technology
WO2020052413A1 (en) Combustion control optimization method and apparatus for thermal generator sets and readable storage medium
CN111413864A (en) 600MW supercritical thermal power generating unit modeling and control method
CN116432515A (en) Boiler efficiency operation instruction optimization method and system based on artificial intelligence

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