CN114186472B - Design method of multi-input multi-output urban solid waste incineration process model - Google Patents
Design method of multi-input multi-output urban solid waste incineration process model Download PDFInfo
- Publication number
- CN114186472B CN114186472B CN202110998327.2A CN202110998327A CN114186472B CN 114186472 B CN114186472 B CN 114186472B CN 202110998327 A CN202110998327 A CN 202110998327A CN 114186472 B CN114186472 B CN 114186472B
- Authority
- CN
- China
- Prior art keywords
- layer
- input
- output
- solid waste
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 76
- 230000008569 process Effects 0.000 title claims abstract description 61
- 238000004056 waste incineration Methods 0.000 title claims abstract description 55
- 239000010813 municipal solid waste Substances 0.000 title claims abstract description 49
- 238000013461 design Methods 0.000 title claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 47
- 238000007781 pre-processing Methods 0.000 claims abstract description 8
- 230000006870 function Effects 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 21
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 20
- 239000001301 oxygen Substances 0.000 claims description 20
- 229910052760 oxygen Inorganic materials 0.000 claims description 20
- UGFAIRIUMAVXCW-UHFFFAOYSA-N Carbon monoxide Chemical compound [O+]#[C-] UGFAIRIUMAVXCW-UHFFFAOYSA-N 0.000 claims description 18
- 239000003546 flue gas Substances 0.000 claims description 18
- 238000013528 artificial neural network Methods 0.000 claims description 17
- 238000002485 combustion reaction Methods 0.000 claims description 12
- 210000002569 neuron Anatomy 0.000 claims description 12
- 230000002159 abnormal effect Effects 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 7
- 238000009826 distribution Methods 0.000 claims description 3
- 238000001035 drying Methods 0.000 claims 2
- 238000012360 testing method Methods 0.000 abstract description 35
- 230000009467 reduction Effects 0.000 abstract description 6
- 238000005457 optimization Methods 0.000 abstract description 5
- 238000002474 experimental method Methods 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 21
- 230000007246 mechanism Effects 0.000 description 15
- 239000002910 solid waste Substances 0.000 description 11
- 238000004458 analytical method Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000004519 manufacturing process Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 4
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 2
- 239000000779 smoke Substances 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000005065 mining Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 239000002994 raw material Substances 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computer Hardware Design (AREA)
- Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Biomedical Technology (AREA)
- Geometry (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Incineration Of Waste (AREA)
Abstract
Aiming at the problem that a controlled object model is difficult to build in the urban solid waste incineration process, the invention designs a multi-input multi-output urban solid waste incineration process model design method; firstly, describing a core process flow of an urban solid waste incineration process and analyzing influencing factors of a model; then, a modeling strategy of a controlled object model is designed according to the control characteristics of the urban solid waste incineration process, and the strategy consists of a working condition identification module, a data preprocessing module, a characteristic reduction module, a controlled object model training module and a controlled object model testing module; finally, experiments show that the effectiveness of the controlled object model in the urban solid waste incineration process lays a foundation for researching an optimization control algorithm of the urban solid waste incineration process.
Description
Technical Field
Aiming at the problems that the internal mechanism of the urban solid waste incineration process is complex and a controlled object model is difficult to build, the invention designs a multi-input multi-output controlled object model based on a Takagi-Sugeno fuzzy neural network, which consists of a working condition identification module, a data preprocessing module, a characteristic reduction module, a controlled object model training module and a controlled object model test module, solves the problems that the internal mechanism of the urban solid waste incineration process is difficult to analyze, the multivariable coupling is strong and the internal rule is difficult to dig, and lays a model foundation for researching the optimal control of the urban solid waste incineration process.
Background
The urban solid waste incineration technology has the outstanding advantages of reduction, reclamation, harmlessness and the like, and becomes one of the main technical means for treating urban solid waste in the world at present. By 2016, 303 solid waste incineration power plants operated in China and China have been provided, and the total treatment capacity reaches 3.04 multiplied by 10 5 t/day, wherein the solid waste incineration power plant using the mechanical grate furnace has 220 seats, the ratio of which exceeds 72 percent, and the grate furnace has become the main incinerator type adopted by the solid waste incineration of China city. Urban solid waste incineration based on grate furnace is a process with various uncertainty characteristics such as strong nonlinearity, strong coupling, large time variation and the like, the complex process inevitably involves the problems of various control fields, the solid waste incineration state can be stabilized by means of advanced control technology, the operation efficiency is improved, and the establishment of an accurate controlled object model is the basis and necessary preparation for implementing the control technology. Therefore, the invention has important significance for the design and research of the object model in the urban solid waste incineration process。
Conventional controlled object models are typically built based on a mechanism analysis, referred to as a mechanism model, also known as a "white-box model". The mechanism model is constructed according to the principles of material balance equation, energy balance equation, biological law, chemical dynamics and the like, and a relatively accurate mathematical model is built by deducing the functional relationship among the operation variable, the state variable and the controlled variable. The mechanism model has the capability of intuitively reflecting the relation between the internal rule and the structure of the system, however, unlike the traditional complex industrial process, the raw material used in the urban solid waste incineration process is solid waste, and the mechanism model has the characteristics of complexity and variability in nature. Factors influencing the components of the solid waste are various, including season climate, classification degree of the solid waste, living standard, living habit, environmental protection consciousness and the like of people in the area. For strong nonlinear industrial processes such as urban solid waste incineration, the model constructed based on mechanism analysis is difficult to analyze the properties and internal mechanism of a strong nonlinear system, and is difficult to be applied to the solid waste incineration process under multiple conditions. In recent years, with the rise of artificial intelligence, a machine learning method based on data driving provides a solution for modeling a control object in a municipal solid waste incineration process.
The data driving model is constructed by mining the mapping relation of the input and output data of the system, and is also called a black box model. The artificial neural network is widely applied to process analysis of a complex industrial system due to good learning capability, computing capability and nonlinear approximation capability, and has important application value when being used in controlled object model design with unknown internal mechanism. Artificial neural networks have found increasing application in the modeling of controlled objects in industrial processes, and have become a popular research point.
The urban solid waste incineration process is a typical multi-input multi-output industrial process, the solid waste heat value is difficult to determine, the internal mechanism reaction is complex, a plurality of operation amounts are seriously coupled with the controlled amount, the system rule is difficult to mine, and the urban solid waste incineration process has typical fuzzy characteristics. Aiming at complex industrial processes such as urban solid waste incineration, the fuzzy neural network provides a good solution. The fuzzy neural network is used as a fuzzy self-adaptive scheme, has nonlinear processing and analysis capability of a fuzzy system and parameter learning and dynamic optimization capability of an artificial neural network, has been widely studied in recent years, becomes an important branch in intelligent computing and neuroscience, and is a technology superior to that of the artificial neural network and the fuzzy system which are independently used.
According to the analysis, the invention designs a multi-input multi-output controlled object model based on a Takagi-Sugeno type fuzzy neural network aiming at the characteristics of the urban solid waste incineration process. Firstly, identifying urban solid waste incineration operation conditions according to an incineration mechanism and expert experience and preprocessing data; then, extracting key operation quantity and controlled quantity capable of reflecting the state of the system; then, constructing a plurality of back-piece sub-networks, and optimizing local parameters and overall parameters of the network by adopting a gradient descent algorithm, so that convergence accuracy and output synchronism of the model are ensured; and finally, verifying the validity of the controlled object model through the process data of a certain solid waste incineration plant in Beijing city.
Disclosure of Invention
The invention obtains a multi-input multi-output controlled object model based on a Takagi-Sugeno fuzzy neural network, which consists of a working condition identification module, a data preprocessing module, a characteristic reduction module, a controlled object model training module and a controlled object model testing module, so that the accurate prediction of key controlled variables is realized, the problem that a controlled object model is difficult to build in the urban solid waste incineration process is solved, and a foundation is laid for researching urban solid waste incineration optimization control;
the invention adopts the following technical scheme and implementation steps:
a design method of a multi-input multi-output urban solid waste incineration process model comprises the following steps:
1. a design method of a multi-input multi-output urban solid waste incineration process model is characterized by comprising the following steps:
(1) The working condition identification module: the module builds a working condition identification expert judgment mechanism based on primary wind pressure, divides working conditions according to primary wind pressure set values, and builds corresponding controlled object models according to different working conditions;
(2) And a data preprocessing module: the module preprocesses acquired data through abnormal data rejection and data normalization, and the calculation steps are as follows:
(1) abnormal data rejection: firstly, normal distribution of data is detected by drawing a quantile graph, abnormal data is removed by 3 sigma criterion, and key controlled variables at 1-T moment are collected: the main steam flow, furnace temperature and flue gas oxygen content are defined as Y s (T), wherein s=1, 2,..q, q is 3, t=1, 2,.. s (t) corresponding residual error ε s (t) is:
calculating standard deviation sigma of data set s The method comprises the following steps:
when Y is s (t) corresponding residual error ε s (t) when the following conditions are met:
|ε s (t)|＞3σ s (3)
then for this Y s (T) performing a culling operation while letting t=t-1;
(2) and (3) data normalization processing: extracting key operation variables in the urban solid waste incineration process: the dry stage grate air flow (left 1, right 1, left 2, right 2), the combustion 1 stage grate air flow (left 1, right 1, left 2, right 2), the combustion 2 stage furnace exhaust air flow (left 1, right 1, left 2, right 2), the burn-out stage grate air flow (left, right), the secondary air flow, the dry stage grate velocity (left inner, right inner, left outer, right outer), the combustion 1 stage grate velocity (left inner, right inner, left outer, right outer), the combustion 2 stage grate velocity (left inner, right inner, left outer, right outer), and the burn-out stage grate velocity (left inner, right inner) are defined as X i (T), wherein i=1, 2,.. N is 29, t=1, 2, where, T, data X will be acquired i (t) and Y s (t) performing normalization processing, wherein the calculation formula is as follows:
wherein x is i (t) represents data X i (t) normalized value, y s (t) represents data Y s (t) normalized value, X i All data representing the ith parameter during the acquisition period, Y s All data representing the s-th parameter during the acquisition period;
(3) Characteristic about Jian Mokuai: calculating the key operation variable x i (t) and key controlled variable y s Pearson correlation coefficient between (t), defined as ρ ds The calculation method comprises the following steps:
according to the calculation result, according to ρ ds Is ranked by the absolute value of (3), and the operation variable ranked as the first 3 is selected and is marked as x i (t), wherein i=1, 2,..n, n is 3;
(4) Multiple-input multiple-output Takagi-Sugeno type fuzzy neural network training module: the model structure of the module design is composed of a front part network and a back part network, wherein the front part network comprises 5 layers of an input layer, a membership function layer, a rule layer, a back part layer and an output layer, the back part network comprises 3 layers of the input layer, the rule layer and the back part layer, and the mathematical description is as follows:
(1) input layer: the layer has n neurons in total, n is 3, the function of the layer transmits input values, and when a t sample enters, the output of the input layer is as follows:
x i (t),i＝1,2,...,n (7)
(2) membership function layer: the layer has n multiplied by m neurons, m is 12, the output of each node represents the membership value of the corresponding input quantity, and the membership function is as follows:
wherein, c ij (t) and delta ij (t) the center and width of membership functions, respectively, whose initial values are generated by rand random functions in the range of [0,2]Random real numbers uniformly distributed among the two;
(3) rule layer: the layer is provided with m neurons, a fuzzy continuous multiplication operator is adopted as a fuzzy logic rule, and the output of the rule layer is as follows:
(4) back part layer: the layer has m times q neurons, q is 3, each node performs linear summation of T-S type model rules, and the function of the layer is to calculate the output back-part parameters corresponding to each rule
The back-part parameters are calculated by the back-part network, n+1 variables are input into the input layer of the back-part network, wherein the input of the 0 th node is a constant, namely x 0 (t) =1, the back-part parameters are transmitted back to the back-part layer of the front-part network, and the calculation process is as follows:
in the method, in the process of the invention,
is a parameter of the fuzzy system, the initial value is set to 0.3, x 0 (t),x 1 (t),…, x n (t) is an input variable;
(5) output layer: the layer is provided with q output nodes, each node performs weighted summation on input parameters, and the calculation formula is as follows:
(6) model parameter learning: the gradient descent algorithm is used to adjust the network parameters, first, the error calculation method is defined as follows:
wherein y is s (t) is the s-th actual output corresponding to the t-th input sample,
is the s-th calculated output corresponding to the t-th input sample, e s And (t) defining the error between the two, and defining the center, the width and the fuzzy system parameter updating algorithm of the network according to the error as follows:
wherein, eta is the online learning rate, and the value range of eta is [0.01,0,05 ]]，c ij (t-1)、δ ij (t-1) and
respectively inputting parameters of the center, the width and the fuzzy system of the network membership function layer when the t-1 th sample is input, and inputting training sample data x after finishing the parameter updating i (t+1), repeating the steps (1) - (6) until all training samples are input, wherein the number of the training samples is 80% of the total number of samples T, and then performing iterative training on the modelUntil the iteration number reaches the maximum iteration value It max ，It max 500 a;
(5) After model training is completed, the building of the multi-input multi-output urban solid waste incineration process model is completed, and at the moment, primary air flow, secondary air flow and fire grate speed are input into the model, and then the model outputs main steam flow, hearth temperature and smoke oxygen content.
The invention mainly comprises the following steps:
(1) The invention solves the problems that the internal mechanism of the urban solid waste incineration process is difficult to analyze, the multivariable coupling is strong, and the internal rules are difficult to excavate, and lays a model foundation for researching the optimization control of the urban solid waste incineration process;
(2) Aiming at the technical characteristics of domestic urban solid waste incineration, a modeling strategy with multiple-station identification and feature reduction is designed, and the model has good robustness and applicability;
(3) The controlled object model established by the invention has multi-output learning capability, and utilizes complementary information among multiple tasks to accurately fit multiple controlled quantities at the same time, and updates network parameters online;
(4) The modeling method for the controlled object in the urban solid waste incineration process based on the Takagi-Sugeno fuzzy neural network has strong learning capacity, high modeling precision and wide application value;
drawings
FIG. 1 is a process flow diagram of the urban solid waste incineration process of the invention
FIG. 2 is a control flow chart of the urban solid waste incineration process of the invention
FIG. 3 is a diagram of a method for constructing a controlled object model in a municipal solid waste incineration process based on data driving according to the invention
FIG. 4 is a schematic diagram of a Takagi-Sugeno fuzzy neural network based multiple-input multiple-output model of the present invention
FIG. 5 is a model training process RMSE variation of the invention
FIG. 6 is a model training sample fitting effect of the present invention
FIG. 7 is a graph showing the effect of fitting the model test sample of the present invention
FIG. 8 is a graph of the effect of fitting the training samples to the temperature of the furnace according to the invention
FIG. 9 is a graph showing the fitting effect of the oxygen content training sample for flue gas according to the present invention
FIG. 10 is a graph of the effect of the main steam flow test sample fit of the present invention
FIG. 11 is a graph showing the effect of fitting the temperature test sample to the furnace chamber according to the present invention
FIG. 12 is a graph showing the fitting effect of the oxygen content test sample of the flue gas according to the present invention
FIG. 13 is a graph of the main steam flow test sample test error of the present invention
FIG. 14 is a graph of the test error of the furnace temperature test sample of the present invention
FIG. 15 is a graph of the test error of the oxygen content of flue gas test sample according to the present invention
Detailed Description
The invention obtains a multi-input multi-output controlled object model based on a Takagi-Sugeno fuzzy neural network, which consists of a working condition identification module, a data preprocessing module, a characteristic reduction module, a controlled object model training module and a controlled object model testing module, so that the accurate prediction of key controlled variables is realized, the problem that a controlled object model is difficult to build in the urban solid waste incineration process is solved, and a foundation is laid for researching urban solid waste incineration optimization control;
the experiment collects the process data of the solid waste incineration power plant in a certain city, the sampling frequency is 1 s/time, and the total collection is 2 multiplied by 10 5 Group data were used as experimental samples;
a design method of a multi-input multi-output urban solid waste incineration process model comprises the following steps:
(1) The working condition identification module: the module builds a working condition identification expert judgment mechanism based on primary wind pressure, divides working conditions according to primary wind pressure set values, and builds corresponding controlled object models according to different working conditions;
(2) And a data preprocessing module: the module preprocesses acquired data through abnormal data rejection and data normalization, and the calculation steps are as follows:
(1) abnormal data rejection: firstly, normal distribution of data is detected by drawing a quantile graph, abnormal data is removed by 3 sigma criterion, and key controlled variables at 1-T moment are collected: the main steam flow, furnace temperature and flue gas oxygen content are defined as Y s (T), wherein s=1, 2,..q, q is 3, t=1, 2,.. s (t) corresponding residual error ε s (t) is:
calculating standard deviation sigma of data set s The method comprises the following steps:
when Y is s (t) corresponding residual error ε s (t) when the following conditions are met:
|ε s (t)|＞3σ s (3)
then for this Y s (T) performing a culling operation while letting t=t-1;
(2) and (3) data normalization processing: extracting key operation variables in the urban solid waste incineration process: the dry stage grate air flow (left 1, right 1, left 2, right 2), the combustion 1 stage grate air flow (left 1, right 1, left 2, right 2), the combustion 2 stage furnace exhaust air flow (left 1, right 1, left 2, right 2), the burn-out stage grate air flow (left, right), the secondary air flow, the dry stage grate velocity (left inner, right inner, left outer, right outer), the combustion 1 stage grate velocity (left inner, right inner, left outer, right outer), the combustion 2 stage grate velocity (left inner, right inner, left outer, right outer), and the burn-out stage grate velocity (left inner, right inner) are defined as X i (T), wherein i=1, 2,.. N is 29, t=1, 2, where, T, data X will be acquired i (t) and Y s (t) performing normalization processing, wherein the calculation formula is as follows:
wherein x is i (t) represents data X i (t) normalized value, y s (t) represents data Y s (t) normalized value, X i All data representing the ith parameter during the acquisition period, Y s All data representing the s-th parameter during the acquisition period;
(3) Characteristic about Jian Mokuai: calculating the key operation variable x i (t) and key controlled variable y s Pearson correlation coefficient between (t), defined as ρ ds The calculation method comprises the following steps:
according to the calculation result, according to ρ ds Is ranked by the absolute value of (3), and the operation variable ranked as the first 3 is selected and is marked as x i (t), wherein i=1, 2,..n, n is 3;
(4) Multiple-input multiple-output Takagi-Sugeno type fuzzy neural network training module: the model structure of the module design is composed of a front part network and a back part network, wherein the front part network comprises 5 layers of an input layer, a membership function layer, a rule layer, a back part layer and an output layer, the back part network comprises 3 layers of the input layer, the rule layer and the back part layer, and the mathematical description is as follows:
(1) input layer: the layer has n neurons in total, n is 3, the function of the layer transmits input values, and when a t sample enters, the output of the input layer is as follows:
x i (t),i＝1,2,...,n (7)
(2) membership function layer: the layer has n multiplied by m neurons, m is 12, the output of each node represents the membership value of the corresponding input quantity, and the membership function is as follows:
wherein, c ij (t) and delta ij (t) the center and width of membership functions, respectively, whose initial values are generated by rand random functions in the range of [0,2]Random real numbers uniformly distributed among the two;
(3) rule layer: the layer is provided with m neurons, a fuzzy continuous multiplication operator is adopted as a fuzzy logic rule, and the output of the rule layer is as follows:
(4) back part layer: the layer has m times q neurons, q is 3, each node performs linear summation of T-S type model rules, and the function of the layer is to calculate the output back-part parameters corresponding to each rule
The back-part parameters are calculated by the back-part network, n+1 variables are input into the input layer of the back-part network, wherein the input of the 0 th node is a constant, namely x 0 (t) =1, the back-part parameters are transmitted back to the back-part layer of the front-part network, and the calculation process is as follows:
in the method, in the process of the invention,
is a parameter of the fuzzy system, the initial value is set to 0.3, x 0 (t),x 1 (t),…, x n (t) is an input variable;
(5) output layer: the layer is provided with q output nodes, each node performs weighted summation on input parameters, and the calculation formula is as follows:
(6) model parameter learning: the gradient descent algorithm is used to adjust the network parameters, first, the error calculation method is defined as follows:
wherein y is s (t) is the s-th actual output corresponding to the t-th input sample,
is the s-th calculated output corresponding to the t-th input sample, e s And (t) defining the error between the two, and defining the center, the width and the fuzzy system parameter updating algorithm of the network according to the error as follows:
wherein, eta is the online learning rate, and the value range of eta is [0.01,0,05 ]]，c ij (t-1)、δ ij (t-1) and
respectively inputting parameters of the center, the width and the fuzzy system of the network membership function layer when the t-1 th sample is input, and inputting training sample data x after finishing the parameter updating i (t+1), repeating the steps (1) - (6) until all training samples are input, wherein the number of the training samples is 80% of the total number of samples T, and then performing iterative training on the model until the number of iterations reaches a maximum iteration value It max ，It max 500 a;
(5) After model training is completed, the building of the multi-input multi-output urban solid waste incineration process model is completed, and at the moment, primary air flow, secondary air flow and fire grate speed are input into the model, and then the model outputs main steam flow, hearth temperature and smoke oxygen content.
FIG. 1 is a process flow diagram of the urban solid waste incineration process of the invention
FIG. 2 is a control flow chart of the urban solid waste incineration process of the invention
FIG. 3 is a schematic diagram of a multiple-input multiple-output model based on Takagi-Sugeno fuzzy neural network according to the present invention
FIG. 4 is a graph of the RMSE variation during the main steam flow training process of the invention
FIG. 5 is a graph of the variation of RMSE in the furnace temperature training process of the invention
FIG. 6 is a graph showing the variation of RMSE in the flue gas oxygen content training process according to the invention
FIG. 7 is a graph of the effect of the main steam flow training sample fit of the present invention
FIG. 8 is a graph of the effect of fitting the training samples to the temperature of the furnace according to the invention
FIG. 9 is a graph showing the fitting effect of the oxygen content training sample for flue gas according to the present invention
FIG. 10 is a graph of the effect of the main steam flow test sample fit of the present invention
FIG. 11 is a graph showing the effect of fitting the temperature test sample to the furnace chamber according to the present invention
FIG. 12 is a graph showing the fitting effect of the oxygen content test sample of the flue gas according to the present invention
FIG. 13 is a graph of the main steam flow test sample test error of the present invention
FIG. 14 is a graph of the test error of the furnace temperature test sample of the present invention
FIG. 15 is a graph of the test error of the oxygen content of flue gas test sample according to the present invention
The technical flow of the urban solid waste incineration process is shown in figure 1; the control flow of the urban solid waste incineration process is shown in figure 2; the structure of a multiple-input multiple-output model based on a Takagi-Sugeno fuzzy neural network is shown in FIG. 3; taking training sample data as the input of a model, wherein the change of the RMSE in the model training process is shown in figures 4-6; FIG. 4 illustrates the variation of RMSE for the main steam flow training process, X axis: training steps, Y axis: training a RMSE value; FIG. 5 shows the RMSE variation during furnace temperature training, X axis: training steps, Y axis: training a RMSE value; FIG. 6 shows the RMSE variation during the flue gas oxygen content training, X axis: training steps, Y axis: training a RMSE value; the fitting effect of the model training sample is shown in fig. 7-9; FIG. 7 is a main steam flow training sample fitting effect, X-axis: sample number, Y axis: main steam flow, the unit is t/h, the black line is predicted output, and the gray line is actual output; FIG. 8 is a graph of the effect of the furnace temperature training sample fit, X-axis: sample number, Y axis: the hearth temperature is in the unit of DEG C, the black line is predicted output, and the gray line is actual output; FIG. 9 shows the fitting effect of the oxygen content training sample of the flue gas, X axis: sample number, Y axis: the oxygen content of the flue gas is shown in units of prediction output and actual output of black lines and gray lines; inputting test sample data as a trained model, wherein the fitting effect of the model test sample is shown in fig. 10-12; FIG. 10 is a plot of the main steam flow test sample fitting effect, X axis: sample number, Y axis: the main steam flow rate is t/h, the black line is predicted output, and the gray line is actual output; FIG. 11 shows the effect of the furnace temperature test sample, X axis: sample number, Y axis: the hearth temperature is in the unit of DEG C, the black line is predicted output, and the gray line is actual output; FIG. 12 is a graph showing the fitting effect of the oxygen content test sample of the flue gas, X axis: sample number, Y axis: the oxygen content of the flue gas, the unit is the predicted output, and the black line is the actual output; the test errors of the model test samples are shown in fig. 13 to 15; FIG. 13 is a main steam flow test sample test error, X axis: sample number, Y axis: main steam flow, unit is t/h; FIG. 14 shows the test error of the furnace temperature test sample, X axis: sample number, Y axis: furnace temperature in degrees celsius; FIG. 15 shows the measurement error of the oxygen content test sample of the flue gas, X axis: sample number, Y axis: oxygen content of flue gas, the unit is; the result shows the effectiveness of the model in modeling the controlled object in the urban solid waste incineration process.
Claims (1)
1. A design method of a multi-input multi-output urban solid waste incineration process model is characterized by comprising the following steps:
(1) The working condition identification module: dividing working conditions according to primary wind pressure set values, and constructing corresponding controlled object models according to different working conditions;
(2) And a data preprocessing module: the module preprocesses acquired data through abnormal data rejection and data normalization, and the calculation steps are as follows:
(1) abnormal data rejection: firstly, normal distribution of data is detected by drawing a quantile graph, abnormal data is removed by 3 sigma criterion, and key controlled variables at 1-T moment are collected: the main steam flow, the furnace temperature and the oxygen content of the flue gas are defined as Y s (T), wherein s=1, 2,..q, q is 3, t=1, 2,.. s (t) corresponding residual error ε s (t) is:
calculating standard deviation sigma of data set s The method comprises the following steps:
when Y is s (t) corresponding residual error ε s (t) when the following conditions are met:
|ε s (t)|＞3σ s (3)
then for this Y s (T) performing a culling operation while letting t=t-1;
(2) and (3) data normalization processing: extracting key operation variables in the urban solid waste incineration process: the drying section grate air flow comprises: left 1, right 1, left 2, right 2; the combustion 1-stage fire grate air flow comprises: left 1, right 1, left 2, right 2; the flow rate of the fire grate of the 2 sections of combustion comprises: left 1, right 1, left 2, right 2; the air flow of the fire grate of the burn-out section comprises left and right; secondary air flow; the drying section grate speed includes: left inner, right inner, left outer, right outer; the combustion section 1 fire grate speed comprises a left inner part, a right inner part, a left outer part and a right outer part; the combustion 2-stage fire grate speed comprises a left inner part, a right inner part,Left outer, right outer; and the speed of the fire grate of the burn-out section comprises left inner part and right inner part; define it as X i (T), wherein i=1, 2,.. N is 29, t=1, 2, where, T, data X will be acquired i (t) and Y s And (t) carrying out normalization processing, wherein the calculation formula is as follows:
wherein x is i (t) represents data X i (t) normalized value, y s (t) represents data Y s (t) normalized value, X i All data representing the ith parameter during the acquisition period, Y s All data representing the s-th parameter during the acquisition period;
(3) Characteristic about Jian Mokuai: calculating the key operation variable x i (t) and key controlled variable y s Pearson correlation coefficient between (t), defined as ρ ds The calculation method comprises the following steps:
according to the calculation result, according to ρ ds Is ranked by the absolute value of (3), and the operation variable ranked as the first 3 is selected and is marked as x i (t), wherein i=1, 2,..n, n is 3;
(4) Multiple-input multiple-output Takagi-Sugeno type fuzzy neural network training module: the model structure of the module design is composed of a front part network and a back part network, wherein the front part network comprises 5 layers of an input layer, a membership function layer, a rule layer, a back part layer and an output layer, the back part network comprises 3 layers of the input layer, the rule layer and the back part layer, and the mathematical description is as follows:
(1) input layer: the layer has n neurons in total, n is 3, the function of the layer transmits input values, and when a t sample enters, the output of the input layer is as follows:
x i (t),i＝1,2,...,n (7)
(2) membership function layer: the layer has n multiplied by m neurons, m is 12, the output of each node represents the membership value of the corresponding input quantity, and the membership function is as follows:
wherein, c ij (t) and delta ij (t) the center and width of membership functions, respectively, whose initial values are generated by rand random functions in the range of [0,2]Random real numbers uniformly distributed among the two;
(3) rule layer: the layer is provided with m neurons, a fuzzy continuous multiplication operator is adopted as a fuzzy logic rule, and the output of the rule layer is as follows:
(4) back part layer: the layer has m times q neurons, q is 3, each node performs linear summation of T-S type model rules, and the function of the layer is to calculate the output back-part parameters corresponding to each rule
The back-part parameters are calculated by the back-part network, n+1 variables are input into the input layer of the back-part network, wherein the input of the 0 th node is a constant, namely x 0 (t) =1, the back-part parameters are transmitted back to the back-part layer of the front-part network, and the calculation process is as follows:
in the method, in the process of the invention,
is a parameter of the fuzzy system, the initial value is set to 0.3, x 0 (t),x 1 (t),…,x n (t) is an input variable;
(5) output layer: the layer is provided with q output nodes, each node performs weighted summation on input parameters, and the calculation formula is as follows:
(6) model parameter learning: the gradient descent algorithm is used to adjust the network parameters, first, the error calculation method is defined as follows:
wherein y is s (t) is the representation data Y s (t) the normalized value of (c),
is the s-th calculated output corresponding to the t-th input sample, e s And (t) defining the error between the two, and defining the center, the width and the fuzzy system parameter updating algorithm of the network according to the error as follows:
wherein eta is the online learning rate, and eta is takenThe value range is [0.01,0,05 ]]，c ij (t-1)、δ ij (t-1) and
respectively inputting parameters of the center, the width and the fuzzy system of the network membership function layer when the t-1 th sample is input, and inputting training sample data x after finishing the parameter updating i (t+1), repeating the steps (1) - (6) until all training samples are input, wherein the number of the training samples is 80% of the total number of samples T, and then performing iterative training on the model until the number of iterations reaches a maximum iteration value It max ，It max 500 a;
(5) After model training is completed, the building of the multi-input multi-output urban solid waste incineration process model is completed, and at the moment, primary air flow, secondary air flow and fire grate speed are input into the model, and then the model outputs main steam flow, hearth temperature and flue gas oxygen content.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110998327.2A CN114186472B (en) | 2021-08-27 | 2021-08-27 | Design method of multi-input multi-output urban solid waste incineration process model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110998327.2A CN114186472B (en) | 2021-08-27 | 2021-08-27 | Design method of multi-input multi-output urban solid waste incineration process model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114186472A CN114186472A (en) | 2022-03-15 |
| CN114186472B true CN114186472B (en) | 2023-05-02 |
Family
ID=80601030
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110998327.2A Active CN114186472B (en) | 2021-08-27 | 2021-08-27 | Design method of multi-input multi-output urban solid waste incineration process model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114186472B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115755779B (en) * | 2022-08-27 | 2024-08-30 | 北京工业大学 | Control-oriented data-driven urban solid waste incineration full-flow modeling method |
| CN115542740B (en) * | 2022-10-07 | 2024-07-16 | 北京工业大学 | Multivariable active disturbance rejection control method based on data driving for urban solid waste incineration process |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101476731A (en) * | 2008-12-31 | 2009-07-08 | 华南理工大学 | Refuse incineration control method based on refuse thermal value soft measurement |
| CN105864797A (en) * | 2016-04-01 | 2016-08-17 | 浙江大学 | Real-time prediction system and method for boiler entering heat value of circulating fluidized bed household garbage incineration boiler |
| CN112733876A (en) * | 2020-10-28 | 2021-04-30 | 北京工业大学 | Soft measurement method for nitrogen oxides in urban solid waste incineration process based on modular neural network |
| CN113191078A (en) * | 2021-04-26 | 2021-07-30 | 北京工业大学 | Method for forecasting flue gas temperature of primary combustion chamber in municipal solid waste incineration process |
-
2021
- 2021-08-27 CN CN202110998327.2A patent/CN114186472B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101476731A (en) * | 2008-12-31 | 2009-07-08 | 华南理工大学 | Refuse incineration control method based on refuse thermal value soft measurement |
| CN105864797A (en) * | 2016-04-01 | 2016-08-17 | 浙江大学 | Real-time prediction system and method for boiler entering heat value of circulating fluidized bed household garbage incineration boiler |
| CN112733876A (en) * | 2020-10-28 | 2021-04-30 | 北京工业大学 | Soft measurement method for nitrogen oxides in urban solid waste incineration process based on modular neural network |
| CN113191078A (en) * | 2021-04-26 | 2021-07-30 | 北京工业大学 | Method for forecasting flue gas temperature of primary combustion chamber in municipal solid waste incineration process |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114186472A (en) | 2022-03-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114186472B (en) | Design method of multi-input multi-output urban solid waste incineration process model | |
| CN111045326B (en) | Tobacco shred drying process moisture prediction control method and system based on recurrent neural network | |
| CN108224446A (en) | A kind of automatic combustion Study on Decision-making Method for Optimization of Refuse Incineration Process | |
| CN114282443B (en) | Residual service life prediction method based on MLP-LSTM supervised joint model | |
| CN106529818A (en) | Water quality evaluation prediction method based on fuzzy wavelet neural network | |
| CN111695666A (en) | Wind power ultra-short term conditional probability prediction method based on deep learning | |
| Ding et al. | MIMO modeling and multi-loop control based on neural network for municipal solid waste incineration | |
| Tang et al. | A deep learning model for measuring oxygen content of boiler flue gas | |
| CN114527646A (en) | Multi-loop quasi-diagonal recurrent neural network PID control method for urban solid waste incineration process | |
| Yang et al. | Application of fuzzy neural network PID algorithm in oil pump control | |
| Jin et al. | Simplification of ANFIS based on importance-confidence-similarity measures | |
| Yang et al. | A novel exhaust gas temperature prediction method of hot blast stove | |
| CN114740713B (en) | Multi-objective optimization control method for wet flue gas desulfurization process | |
| CN110684547A (en) | Optimized control method for biomass pyrolysis carbonization kiln | |
| CN113191078B (en) | Flue gas temperature forecasting method for primary combustion chamber in urban household garbage incineration process | |
| CN115542740B (en) | Multivariable active disturbance rejection control method based on data driving for urban solid waste incineration process | |
| CN113779858B (en) | Combustion optimization method and system, storage medium and electronic equipment | |
| CN114819063A (en) | Semi-supervised coke oven flame path temperature soft measurement method based on manifold regularization | |
| Li et al. | Long short-term memory based on random forest-recursive feature eliminated for hot metal silcion content prediction of blast furnace | |
| CN113869359A (en) | Modular neural network-based prediction method for nitrogen oxides in urban solid waste incineration process | |
| CN118009322B (en) | Method and system for controlling hearth temperature in urban solid waste incineration process through end-to-side cloud cooperation | |
| Wang et al. | Predictive Modeling of Henry’s Law Constant in Chemical Structures Using LSSVM and ANFIS Algorithms | |
| CN113627469B (en) | Hot rolled strip steel plate convexity prediction method based on fuzzy reasoning algorithm | |
| Ding et al. | Multi-input and multi-output modeling method based on TS fuzzy neural network and its application | |
| He et al. | Application of Takagi-Sugeon fuzzy system to regional economy modeling |
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 |