An intelligent control method, device, medium and product for a municipal solid waste incineration process

Abstract

The application discloses a kind of urban solid waste incineration process intelligent control method, equipment, medium and product, it is related to urban solid waste incineration technical field, the method includes obtaining the current time of urban solid waste incineration process's pusher speed, grate speed, primary air flow, secondary air flow, combustion time, main steam flow and furnace oxygen content;According to the current time of pusher speed, grate speed, combustion time and main steam flow, utilize main steam flow prediction model, determine the main steam flow of next time;According to the current time of primary air flow, secondary air flow and furnace oxygen content, utilize furnace oxygen content prediction model, determine the furnace oxygen content of next time;Combination is based on the switching rule of expert knowledge, utilize gradient descent algorithm, to main steam flow controller and furnace oxygen content controller are solved, determine optimal control rate, to control urban solid waste incineration process.The application improves solid waste treatment efficiency, realizes the fast stable control of main steam flow and furnace oxygen content.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of municipal solid waste incineration, and in particular to a municipal solid waste incineration process intelligent control method, device, medium and product. BACKGROUND

[0002] With the rapid development of China's economy and the continuous acceleration of urbanization, the production of municipal solid waste has increased dramatically, and municipal solid waste treatment has become one of the focus problems of environmental protection. Municipal solid waste incineration technology has the characteristics of obvious volume and weight reduction, resource recycling, and full harmless treatment, and has become the main way of municipal solid waste treatment in China. However, since garbage classification in China is still in its infancy, the calorific value of solid waste fluctuates greatly, and there are problems such as unstable incineration process, low waste heat utilization efficiency, and high pollutant emission concentration. Therefore, it is of great theoretical significance and application value to realize the control of the municipal solid waste incineration process, improve the efficiency of solid waste treatment and the stability of the incineration process. SUMMARY

[0003] The purpose of the present application is to provide a municipal solid waste incineration process intelligent control method, device, medium and product to improve the efficiency of solid waste treatment and the stability of the incineration process.

[0004] To achieve the above-mentioned purpose, the present application provides the following solutions:

[0005] In a first aspect, the present application provides a municipal solid waste incineration process intelligent control method, comprising:

[0006] obtaining the current time pusher speed, the current time grate speed, the current time primary air flow, the current time secondary air flow, the current time combustion time, the current time main steam flow and the current time furnace oxygen content of the municipal solid waste incineration process;

[0007] determining the next time main steam flow according to the current time pusher speed, the current time grate speed, the current time combustion time and the current time main steam flow, using a main steam flow prediction model; wherein the main steam flow prediction model is obtained by training a fuzzy neural network using a first training data set;

[0008] determining the next time furnace oxygen content according to the current time primary air flow, the current time secondary air flow and the current time furnace oxygen content, using a furnace oxygen content prediction model; wherein the furnace oxygen content prediction model is obtained by training a fuzzy neural network using a second training data set;

[0009] Based on the main steam flow at the next moment and the furnace oxygen at the next moment, combined with the switching rule based on expert knowledge, the main steam flow controller and the furnace oxygen controller are solved by using the gradient descent algorithm to determine the optimal control rate to control the municipal solid waste incineration process; the switching rule based on expert knowledge includes a first rule, a second rule and a third rule; the first rule is that when the back pressure of the primary air prefilter at the current moment is greater than a first set pressure and less than a second set pressure, the learning rate of the main steam flow controller is increased by a set value, and the burning time at the current moment is reduced by a set burning time; the second rule is that when the change rate of the back pressure of the primary air prefilter at the current moment is greater than a third set pressure, it is determined that the current is screening, the solving of the main steam flow controller and the furnace oxygen controller is performed, and the first rule is not executed, and when the change rate of the back pressure of the primary air prefilter at the current moment is less than the third set pressure, it is determined that the current is incinerating, the solving of the main steam flow controller and the furnace oxygen controller and the first rule are executed; the third rule is that when the difference between the main steam flow set value at the next moment and the main steam flow at the next moment is greater than a set difference, the material pushing is delayed, and the burning time at the current moment is increased by a set burning time; the optimal control rate includes the pushing speed increment, the grate speed increment, the burning time increment, the primary air flow increment and the secondary air flow increment.

[0010] Optionally, the fuzzy neural network is trained by using the first training data set, specifically including:

[0011] The first training data set is obtained; the first training data set includes the historical pushing speed, the historical grate speed, the historical burning time, the historical main steam flow and the actual value of the main steam flow at the corresponding moment of the training municipal solid waste incineration process;

[0012] The fuzzy neural network is trained by taking the historical pushing speed, the historical grate speed, the historical burning time and the historical main steam flow of the training municipal solid waste incineration process as inputs and taking the actual value of the main steam flow at the corresponding moment as output, the model parameters of the fuzzy neural network are updated by using the LM algorithm, and a main steam flow prediction model is obtained.

[0013] Optionally, the fuzzy neural network is trained by taking the historical pushing speed, the historical grate speed, the historical burning time and the historical main steam flow of the training municipal solid waste incineration process as inputs and taking the actual value of the main steam flow at the corresponding moment as output, the model parameters of the fuzzy neural network are updated by using the LM algorithm, and a main steam flow prediction model is obtained, specifically including:

[0014] The historical pushing speed, the historical grate speed, the historical burning time and the historical main steam flow are input into the current fuzzy neural network to obtain the predicted value of the main steam flow at the corresponding moment;

[0015] The loss function value is determined based on the predicted main steam flow rate at the corresponding time and the actual main steam flow rate at the corresponding time.

[0016] Determine whether the loss function value is less than a set threshold;

[0017] If so, the current fuzzy neural network will be used as the main steam flow prediction model;

[0018] If not, the LM algorithm is used to update the model parameters of the current fuzzy neural network, and the following is returned: "Input the feeder speed, grate speed, combustion time and main steam flow rate at the historical time into the current fuzzy neural network to obtain the predicted value of the main steam flow rate at the corresponding time."

[0019] Optionally, the main steam flow controller is:

[0020] ;

[0021] in, This represents the increment of the main steam flow control rate at time t+1. , The feeder speed increment at time t+1; This represents the increment of the grate speed at time t+1; t+1 is the combustion time increment; T is the transpose; Let t be the learning rate of the main steam flow controller; The objective function of the main steam flow controller; Let be the feeder speed at time t; Let be the grate speed at time t; Let t be the burning time at time t.

[0022] Optionally, the objective function of the main steam flow controller is:

[0023] ;

[0024] in, for The main steam flow rate setpoint at any given time; for Main steam flow rate at any given time; Let t be the main steam flow control rate. .

[0025] Optionally, the furnace oxygen controller is:

[0026] ;

[0027] in, This represents the increment of the furnace oxygen control rate at time t+1. , This represents the incremental airflow at time t+1. The increment of secondary airflow at time t+1; T is the transpose; Let t be the learning rate of the furnace oxygen controller. The objective function for the furnace oxygen quantity controller; Let be the airflow rate at time t; Let be the secondary airflow rate at time t.

[0028] Optionally, the objective function of the furnace oxygen controller is:

[0029] ;

[0030] in, for The setpoint for furnace oxygen level at any given time; for The amount of oxygen in the furnace at any given time; Let be the furnace oxygen control rate at time t. .

[0031] In a second aspect, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the intelligent control method for the urban solid waste incineration process described in any one of the above-mentioned methods.

[0032] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent control method for the urban solid waste incineration process described above.

[0033] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the intelligent control method for the urban solid waste incineration process described above.

[0034] According to the specific embodiments provided in this application, this application has the following technical effects:

[0035] The application provides a municipal solid waste incineration process intelligent control method, device, medium and product. The current moment pusher speed, current moment grate speed, current moment primary air flow, current moment secondary air flow, current moment combustion time, current moment main steam flow and current moment furnace oxygen content of the municipal solid waste incineration process are acquired. According to the current moment pusher speed, current moment grate speed, current moment combustion time and current moment main steam flow, a main steam flow prediction model is used to determine the next moment main steam flow. The main steam flow prediction model is obtained by training a fuzzy neural network by using a first training data set. According to the current moment primary air flow, current moment secondary air flow and current moment furnace oxygen content, a furnace oxygen content prediction model is used to determine the next moment furnace oxygen content. The furnace oxygen content prediction model is obtained by training a fuzzy neural network by using a second training data set. Based on the next moment main steam flow and next moment furnace oxygen content, a switching rule based on expert knowledge is combined, and a gradient descent algorithm is used to solve a main steam flow controller and a furnace oxygen content controller to determine an optimal control rate to control the municipal solid waste incineration process. The application establishes an accurate and effective main steam flow prediction model and furnace oxygen content prediction model based on a fuzzy neural network, combines a switching rule based on expert knowledge, and uses a gradient descent algorithm to solve the optimal control rate of the pusher speed, grate speed, combustion time, primary air flow, secondary air flow and other operating quantities in real time, thereby improving the solid waste treatment efficiency and realizing rapid and stable control of the main steam flow and furnace oxygen content. BRIEF DESCRIPTION OF DRAWINGS

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0037] Figure 1 A flowchart of a municipal solid waste incineration process intelligent control method provided by an embodiment of the present application;

[0038] Figure 2 A flowchart of the municipal solid waste incineration process intelligent control method of the present application in actual application;

[0039] Figure 3 A municipal solid waste incineration process main steam flow prediction result curve;

[0040] Figure 4 A municipal solid waste incineration process furnace oxygen content prediction result curve;

[0041] Figure 5 A municipal solid waste incineration process main steam flow control effect diagram;

[0042] Figure 6 The control effect diagram of the furnace oxygen quantity of the municipal solid waste incineration process;

[0043] Figure 7 The control result curve diagram of the pusher speed, the grate speed and the combustion time of the municipal solid waste incineration process;

[0044] Figure 8 The control result curve diagram of the primary air flow and the secondary air flow of the municipal solid waste incineration process;

[0045] Figure 9 The structural schematic diagram of a computer device provided by an embodiment of the present application is shown. DETAILED DESCRIPTION

[0046] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0047] The above purposes, features and advantages of the present application can be more obvious and easy to understand. The present application will be further described in detail below with reference to the drawings and specific embodiments.

[0048] The present application relates to an intelligent autonomous control method of municipal solid waste incineration process data and knowledge double driving (i.e. an intelligent control method of municipal solid waste incineration process); a main steam flow prediction model and a furnace oxygen quantity prediction model of the municipal solid waste incineration process are established, a model prediction control strategy is designed, the primary air, the secondary air, the pusher speed, the grate speed and the combustion time are automatically adjusted, a plurality of switching rules based on expert knowledge are integrated, the control performance of the controller in the unstable state is improved, and the tracking control of the furnace oxygen quantity and the main steam flow is realized. It belongs to the field of municipal solid waste treatment and the field of intelligent control.

[0049] In an exemplary embodiment, as shown in Figure 1 An intelligent control method of municipal solid waste incineration process is provided, comprising the following steps:

[0050] S1: obtaining the current time pusher speed, the current time grate speed, the current time primary air flow, the current time secondary air flow, the current time combustion time, the current time main steam flow and the current time furnace oxygen quantity of the municipal solid waste incineration process.

[0051] S2: determining the main steam flow at the next time point according to the pusher speed at the current time point, the grate speed at the current time point, the combustion time at the current time point, and the main steam flow at the current time point, by using a main steam flow prediction model; wherein the main steam flow prediction model is obtained by training a fuzzy neural network by using a first training data set.

[0052] As an optional implementation, the fuzzy neural network is trained by using the first training data set, and specifically includes:

[0053] S21: obtaining a first training data set; the first training data set includes the historical pusher speed, the historical grate speed, the historical combustion time, the historical main steam flow, and the actual value of the main steam flow at the corresponding time point of the training municipal solid waste incineration process.

[0054] S22: training the fuzzy neural network by taking the historical pusher speed, the historical grate speed, the historical combustion time, and the historical main steam flow of the training municipal solid waste incineration process as inputs and taking the actual value of the main steam flow at the corresponding time point as output, updating the model parameters of the fuzzy neural network by using the LM algorithm, and obtaining the main steam flow prediction model.

[0055] As an optional implementation, S22 specifically includes:

[0056] S221: inputting the historical pusher speed, the historical grate speed, the historical combustion time, and the historical main steam flow into the current fuzzy neural network to obtain the predicted value of the main steam flow at the corresponding time point.

[0057] S222: determining the loss function value according to the predicted value of the main steam flow at the corresponding time point and the actual value of the main steam flow at the corresponding time point.

[0058] S223: determining whether the loss function value is less than a set threshold.

[0059] S224: if yes, taking the current fuzzy neural network as the main steam flow prediction model.

[0060] S225: if no, updating the model parameters of the current fuzzy neural network by using the LM algorithm, and returning to “inputting the historical pusher speed, the historical grate speed, the historical combustion time, and the historical main steam flow into the current fuzzy neural network to obtain the predicted value of the main steam flow at the corresponding time point”.

[0061] S3: determining the furnace oxygen content at the next time point according to the primary air flow at the current time point, the secondary air flow at the current time point, and the furnace oxygen content at the current time point, by using a furnace oxygen content prediction model; wherein the furnace oxygen content prediction model is obtained by training a fuzzy neural network by using a second training data set.

[0062] In practical applications, the construction process of the main steam flow prediction model and the furnace oxygen amount prediction model is as follows:

[0063] 1. Data collection. The historical city solid waste incineration process (training city solid waste incineration process) historical time pusher speed, historical time grate speed, historical time burning time, historical time main steam flow, corresponding time main steam flow actual value, historical time primary air flow, historical time secondary air flow and corresponding time furnace oxygen amount actual value are collected.

[0064] 2. Determine the input and output variables of the operation index model (main steam flow prediction model and furnace oxygen amount prediction model), and establish the corresponding prediction model: determine the input variables and corresponding output variables of the main steam flow prediction model and the furnace oxygen amount prediction model, and use fuzzy neural network (Fuzzy Neural Network, FNN) to establish the operation index model.

[0065] (1) Determine the input variables of each operation index model:

[0066] Since the established city solid waste incineration process main steam flow prediction model and furnace oxygen amount prediction model contain the main operating variables. The main operating variables of the city solid waste incineration plant using the inverse push grate furnace are the pusher speed, the grate speed, the primary air flow, the secondary air flow and the burning time, among which the pusher speed and the grate speed mainly affect the feeding speed, the burning time mainly affects the feeding frequency, and the primary air flow and the secondary air flow mainly affect the oxygen supply during the combustion process. Based on the above rules, the input and output variables of the main steam flow prediction model and the furnace oxygen amount prediction model are:

[0067] The input variables of the main steam flow prediction model are , the input dimension , wherein is the pusher speed, is the grate speed, is the burning time, is the main steam flow. The output variable is the predicted next time main steam flow.

[0068] The input variables of the furnace oxygen amount prediction model are , the input dimension , wherein is the primary air flow, is the secondary air flow, is the furnace oxygen amount. The output variable is the predicted next time furnace oxygen amount.

[0069] (2) A fuzzy neural network (FNN) is used to establish a main steam flow prediction model and a furnace oxygen prediction model. The method is as follows:

[0070] The FNN-based model can be represented as:

[0071] (1)

[0072] in, For the number of prediction models, For time t+1, the first... The prediction output of a prediction model. for Time of the first The input to a prediction model, for Time of the first The first prediction model One input, For the first The number of input variables for a prediction model. It is the number of fuzzy rules in the FNN. for Time of the first The prediction model is the first one. The input and the first The center of each membership function; for Time of the first The prediction model is the first one. The input and the first The width of each membership function. The sampling time for the urban solid waste incineration process.

[0073] For the first A first-order Takagi-Sugeno-Kang (TSK) fuzzy rule is defined as follows:

[0074] (2)

[0075] in, and for Time of the first The first prediction model The 0th coefficient and the 1st coefficient of the first-order TSK fuzzy rule Each coefficient.

[0076] For the The center of each prediction model ,width and fuzzy rule coefficients , which is to be determined by the online learning algorithm given by equation (4) - equation (7) for a given.

[0077] The model accuracy is detected by periodic evaluation. The absolute percentage error (MAPE) can accurately measure the accuracy of the model, the moment the prediction output of the th prediction model,

[0078] (3)

[0079] wherein, the prediction output of the moment the prediction output of the th prediction model, the expected output of the moment the expected output of the th prediction model, the data volume of model training.

[0080] The model accuracy threshold is set. If , the model accuracy is high, and the model is kept unchanged. If , the model accuracy is low, and the model needs to be updated by online learning.

[0081] The loss function of the model online learning can be defined as:

[0082] (4)

[0083] wherein, the loss function of the moment the loss function of the th prediction model, the prediction output error of the moment the prediction output error of the th prediction model.

[0084] An adaptive Levenberg-Marquardt (LM) algorithm is proposed to update the model parameters. The parameter updating method is:

[0085] (5)

[0086] wherein, the iteration number of the adaptive LM algorithm, the maximum iteration number of the adaptive LM algorithm, the unit matrix, the prediction output error of the moment the prediction output error of the a parameter matrix of the prediction model; is the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; a center of the kth prediction model input and membership function; is the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; a width of the kth prediction model input and membership function; is the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; a fuzzy rule coefficient of the kth prediction model. is an adaptive learning rate of the kth prediction model at the tth iteration; and are the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; a Hessian matrix and a gradient vector of the kth prediction model at the tth iteration, defined as:

[0087] (6)

[0088] wherein, is the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; a loss function of the kth prediction model at the tth iteration, is the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; a prediction output error of the kth prediction model at the tth iteration. T is transpose; is the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; a Jacobian matrix of the kth prediction model at the tth iteration, defined as:

[0089] (7)

[0090] wherein, is the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; a center of the kth prediction model at the tth iteration; a width of the kth prediction model at the tth iteration; is the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; a width of the kth prediction model at the tth iteration; is the kth prediction model at the tth iteration; the kth prediction model at the tth iteration; is the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1 the first coefficient of the nth input and the rth membership function of the kth prediction model at the tth iteration at time t+1

[0091] (8)

[0092] wherein, the prediction output of the kth prediction model at the tth iteration at time t+1 the prediction output of the kth prediction model at the tth iteration at time t+1 the prediction output of the kth prediction model at the tth iteration at time t+1 the kth input of the kth prediction model at time t, the kth input of the kth prediction model at time t, the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1 the center of the nth input and the rth membership function of the kth prediction model at time t+1

[0093] (9)

[0094] (10)

[0095] the adaptive learning rate of the kth prediction model at the tth iteration of the adaptive LM algorithm is:

[0096] (11)

[0097] wherein, is a two-norm, is a constant, ​.

[0098] S4: based on the main steam flow at the next moment and the furnace oxygen at the next moment, combining the switching rule based on expert knowledge, using the gradient descent algorithm to solve the main steam flow controller and the furnace oxygen controller, determine the optimal control rate to control the municipal solid waste incineration process; the switching rule based on expert knowledge includes the first rule, the second rule and the third rule; the first rule is that when the current time once air preheater back pressure is greater than the first set pressure and less than the second set pressure, the learning rate of the main steam flow controller increases by a set value, and the current combustion time decreases by a set combustion time; the second rule is that when the current time once air preheater back pressure rate of change is greater than the third set pressure, it is determined that the current is screening, the solving of the main steam flow controller and the furnace oxygen controller is executed, and the first rule is not executed, when the current time once air preheater back pressure rate of change is less than the third set pressure, it is determined that the current is incineration, the solving of the main steam flow controller and the furnace oxygen controller and the first rule are executed; the third rule is that when the difference between the next time main steam flow set value and the next time main steam flow is greater than the set difference, the material pushing is delayed, and the current combustion time increases by a set combustion time; the optimal control rate includes the pushing speed increment, the grate speed increment, the combustion time increment, the primary air flow increment and the secondary air flow increment.

[0099] As shown in Figure 2 In actual application, first, DCS (Distributed Control System) data is collected and DCS data is stored. Then, it is judged whether the main steam flow prediction model and the furnace oxygen prediction model are adaptive, if adaptive, the main steam flow is predicted by using the main steam flow prediction model, the deviation of the main steam flow from the main steam flow set value is calculated, and the control rate of the pushing speed, the grate speed and the combustion time is calculated, based on the current control rate, the pushing speed, the grate speed and the combustion time are given, the field operation result is output, while predicting the main steam flow, the furnace oxygen is also predicted by using the furnace oxygen prediction model, the deviation of the furnace oxygen from the furnace oxygen set value is calculated, and the control rate of the primary air flow and the secondary air flow is calculated, based on the current control rate, the primary air valve opening and the secondary air valve opening are given, and the field operation result is output. If not adaptive, the main steam flow prediction model parameters and the furnace oxygen prediction model parameters are updated, and the subsequent steps of predicting the main steam flow and the furnace oxygen are performed.

[0100] In actual application, the main steam flow controller and the furnace oxygen controller are established, the switching rule based on expert knowledge is established based on expert knowledge, the optimal control rate is solved, and the stable and rapid tracking control of the main steam flow and the furnace oxygen is realized.

[0101] (1) The target function of the main steam flow controller for tracking control is:

[0102] (12)

[0103] wherein, is the target function of the main steam flow controller; is the main steam flow set value at time t; is the main steam flow at time t; is the main steam flow control rate at time t, ; is the pusher speed at time t; is the grate speed at time t; is the combustion time at time t. The control rate is updated as:

[0104] (13)

[0105] wherein, is the main steam flow control rate increment at time t+1;

[0106] , is the pusher speed increment at time t+1; is the grate speed increment at time t+1; is the combustion time increment at time t+1. The main steam flow control rate increment is solved as: (14)

[0107] wherein, is the learning rate of the main steam flow controller at time t.

[0108] The main steam flow control rate increment is specifically: (15)

[0109] (16)

[0110] (17)

[0111] wherein, is the predicted output of the main steam flow prediction model at time t+1.

[0112] Through the above continuous iteration, the main steam flow tracking control is realized.

[0113]

[0114] (2) The target function of the furnace oxygen quantity controller for tracking control is:

[0115] ​​ (18)

[0116] in, The objective function for the furnace oxygen quantity controller; for The setpoint for furnace oxygen level at any given time; for The amount of oxygen in the furnace at any given time; Let be the furnace oxygen control rate at time t. ; Let be the airflow rate at time t; Let be the secondary airflow rate at time t.

[0117] Control rate updated to:

[0118] (19)

[0119] in, This represents the increment of the furnace oxygen control rate at time t+1. , This represents the incremental airflow at time t+1. Let be the increment of secondary air flow at time t+1. The increment of furnace oxygen control rate is calculated as follows:

[0120] (20)

[0121] in, Let be the learning rate of the furnace oxygen controller at time t. The specific increment of the furnace oxygen control rate is as follows:

[0122] (twenty one)

[0123] (twenty two)

[0124] in, This is the predicted output of the furnace oxygen quantity prediction model at time t+1.

[0125] Through continuous iteration, the goal is to achieve furnace oxygen quantity tracking and control.

[0126] (3) Switching rules based on expert knowledge.

[0127] Due to the significant fluctuations in the urban solid waste incineration process, the controller may not respond in a timely manner. To further improve the controller's response speed and stability, a switching rule based on expert knowledge combined with model predictive control was developed based on the experience of on-site operators to improve control effectiveness.

[0128] 1) Pressure after the primary air preheater is an important reference to show the amount of solid waste and the amount of reduction with incineration each time the solid waste is pushed in, after each time the solid waste is pushed in, increases, and the solid waste is pushed in once will be increased by about 70 Pa to 120 Pa as the incineration proceeds, gradually decreases and finally returns to the initial state. Based on the above expert experience, the first rule is established: when the pressure after the air preheater of the primary air at the current time is greater than the first set pressure and less than the second set pressure, the learning rate of the main steam flow controller is increased by the set value, and the current combustion time is reduced by the set combustion time.

[0129] In this embodiment, the first rule is as follows:

[0130] If , the learning rate of the main steam flow controller is increased and the combustion time is reduced .

[0131] 2) There is a screening process during the operation process, about every half an hour, the accumulated ash in the air chamber is blown out of the air chamber horizontally, lasting about 30 seconds. During the screening process, there will be a sharp fluctuation, at the beginning of the screening, rapidly decreases by about 300 Pa to 400 Pa, at the end of the screening, rapidly rises to the state before screening. Screening will interfere with the judgment of the change, which may lead to misjudgment that the solid waste in the furnace has been completely incinerated. Based on the above expert knowledge, the second rule is established: when the change rate of the pressure after the air preheater of the primary air at the current time is greater than the third set pressure, it is determined that the current is screening, and the solving of the main steam flow controller and the furnace oxygen amount controller is executed, and the first rule is not executed. When the change rate of the pressure after the air preheater of the primary air at the current time is less than the third set pressure, it is determined that the current is incinerating, and the solving of the main steam flow controller and the furnace oxygen amount controller and the first rule are executed.

[0132] In this embodiment, the second rule is as follows:

[0133] If , the current is screening, at this time only the controller operation is executed, and the first rule is not executed, if , it is a normal incineration condition, the controller operation and the first rule are executed.

[0134] 3) There is a long time lag between the solid waste incineration and the main steam flow, and the controller has a certain lag in tracking the main steam flow. When the solid waste in the furnace is incinerated, there is a lag in the increase of the main steam flow, and only relying on the controller will cause the incineration time Given lower, the controller response speed is low when the main steam flow increases, which leads to the situation of pushing material in advance. Based on the above expert knowledge, the third rule is established. When the difference between the next time main steam flow set value and the next time main steam flow is greater than the set difference, the pushing material is delayed, and the current time combustion time is increased by the set combustion time.

[0135] In this embodiment, the third rule is as follows:

[0136] If , at this time the main steam flow is in a higher state, the pushing material should be delayed, .

[0137] The above expert rules ensure the control response speed in the boundary condition, avoiding the situation of continuously thickening or thinning the material layer. The model prediction control method is used for regulation and control when not in the boundary condition, realizing stable tracking control.

[0138] The application utilizes the collected historical data to establish the main steam flow prediction model and the furnace oxygen amount prediction model of the municipal solid waste incineration process. As an embodiment, the actual data from a municipal solid waste incineration plant in Beijing is used to verify the effectiveness of the method. The original data sampling interval is 30 seconds, and there are 1500 groups of experimental data. The variables required for establishing the main steam flow prediction model and the furnace oxygen amount prediction model are extracted. The 1500 groups of data are used for pre-training of the prediction model, and the control effect of the main steam flow and the furnace oxygen amount of the municipal solid waste incineration process is tested under any initial value.

[0139] (1) The main steam flow prediction model and the furnace oxygen amount prediction model based on fuzzy neural network are established. Based on 1000 groups of historical data, the adaptive LM algorithm is used to train the main steam flow prediction model and the furnace oxygen amount prediction model. The prediction result of the main steam flow is as shown in Figure 3 , X axis: time, unit is 30 seconds, Y axis: main steam flow, unit is t / h. The prediction result of the furnace oxygen amount is as shown in Figure 4 , X axis: time, unit is 30 seconds, Y axis: furnace oxygen amount, unit is %.

[0140] (2) Based on the obtained main steam flow prediction model and the furnace oxygen amount prediction model of the municipal solid waste incineration process, the switching rule based on expert knowledge is used to solve the optimal control rate. The set value of the main steam flow is given in the following way: every 5 hours, 75 t / h, 78 t / h, 76 t / h, 74 t / h, and 72 t / h are given in turn. The set value of the furnace oxygen amount is given in the following way: every 5 hours, 5.5%, 7%, 5%, 6%, and 4.5% are given in turn. The control effect of the main steam flow of the municipal solid waste incineration process is as shown in Figure 5 , X axis: time, unit is 30 seconds, Y axis: main steam flow, unit is t / h. The control effect of the furnace oxygen amount of the municipal solid waste incineration process is as shown inFigure 6 Figure 6 shows the control results of the pusher speed, the grate speed and the combustion time in the municipal solid waste incineration process, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the furnace oxygen content, the unit is %. The control results of the primary air flow and the secondary air flow in the municipal solid waste incineration process are shown in Figure 7, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the primary air flow, the unit is m Figure 7 Figure 6 shows the control results of the pusher speed, the grate speed and the combustion time in the municipal solid waste incineration process, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the furnace oxygen content, the unit is %. The control results of the primary air flow and the secondary air flow in the municipal solid waste incineration process are shown in Figure 7, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the primary air flow, the unit is m Figure 8 Figure 6 shows the control results of the pusher speed, the grate speed and the combustion time in the municipal solid waste incineration process, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the furnace oxygen content, the unit is %. The control results of the primary air flow and the secondary air flow in the municipal solid waste incineration process are shown in Figure 7, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the primary air flow, the unit is m 3 Figure 6 shows the control results of the pusher speed, the grate speed and the combustion time in the municipal solid waste incineration process, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the furnace oxygen content, the unit is %. The control results of the primary air flow and the secondary air flow in the municipal solid waste incineration process are shown in Figure 7, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the primary air flow, the unit is m 3 Figure 6 shows the control results of the pusher speed, the grate speed and the combustion time in the municipal solid waste incineration process, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the furnace oxygen content, the unit is %. The control results of the primary air flow and the secondary air flow in the municipal solid waste incineration process are shown in Figure 7, wherein the X axis represents time, the unit is 30 seconds, and the Y axis represents the primary air flow, the unit is m

[0141] (3) The MAPE is used to quantitatively evaluate the accuracy of the prediction model, and the calculation result is that the MAPE result of the main steam flow prediction model is 1.45%, the prediction accuracy is 98.55%, and the MAPE result of the furnace oxygen content prediction model is 4.09%, the prediction accuracy is 95.91%. The integral absolute error (IAE) is used to quantitatively evaluate the control effect, and the calculation result is that the IAE result of the main steam flow control is 0.0328, the control accuracy is 96.72%, and the IAE result of the furnace oxygen content control is 0.041, the control accuracy is 95.9%.

[0142] The application establishes an accurate and effective main steam flow prediction model and a furnace oxygen content prediction model, realizes the rapid and stable control of the main steam flow and the furnace oxygen content by solving the optimal control rate of the pusher speed, the grate speed, the combustion time, the primary air flow and the secondary air flow in real time, and has important significance for the stable operation and sustainable development of the municipal solid waste incineration process.

[0143] In an exemplary embodiment, a computer device is provided, comprising a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned intelligent control method of the municipal solid waste incineration process when executing the computer program.

[0144] In an exemplary embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program implements the above-mentioned intelligent control method of the municipal solid waste incineration process when executed by a processor.

[0145] In an exemplary embodiment, a computer program product is provided, comprising a computer program, and the computer program implements the above-mentioned intelligent control method of the municipal solid waste incineration process when executed by a processor.

[0146] In an exemplary embodiment, a computer device is provided, which can be a server or a terminal, and the internal structure diagram thereof can be as shown in Figure 9As shown in the figure. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through the system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capability. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and the external device. The communication interface of the computer device is used to communicate with the terminal outside through the network connection. The computer program is executed by the processor to realize a kind of urban solid waste incineration process intelligent control method.

[0147] Those skilled in the art can understand that, Figure 9 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or less components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0148] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant regulations.

[0149] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to a memory, a database or other medium used in the embodiments provided in the present application can include at least one of a non-volatile and a volatile memory. The non-volatile memory can include a read-only memory (ROM), a magnetic tape, a floppy disk, a flash memory, an optical storage, a high-density embedded non-volatile memory, a resistive random access memory (ReRAM), a magnetoresistive random access memory (MRAM), a ferroelectric random access memory (FRAM), a phase change memory (PCM), a graphene memory, etc. The volatile memory can include a random access memory (RAM) or an external cache memory, etc. As an illustration but not limitation, the RAM can be in various forms, such as a static random access memory (SRAM) or a dynamic random access memory (DRAM), etc.

[0150] The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a blockchain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0151] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combinations of the technical features do not exist contradictory, they should be considered as the scope of the present application.

[0152] The principles and implementation modes of the present application are described by applying specific examples in the present application. The above-mentioned embodiments are only used to help understand the method and its core idea of the present application; meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range can be changed. In conclusion, the content of the present application should not be understood as a limitation.

Claims

1. An intelligent control method for municipal solid waste incineration process, characterized in that, The application relates to a method for controlling a municipal solid waste incineration process. The method comprises the following steps: acquiring a current moment pusher speed, a current moment grate speed, a current moment primary air flow, a current moment secondary air flow, a current moment combustion time, a current moment main steam flow and a current moment furnace oxygen content of the municipal solid waste incineration process; determining a next moment main steam flow according to the current moment pusher speed, the current moment grate speed, the current moment combustion time and the current moment main steam flow by using a main steam flow prediction model, wherein the main steam flow prediction model is obtained by training a fuzzy neural network by using a first training data set; determining a next moment furnace oxygen content according to the current moment primary air flow, the current moment secondary air flow and the current moment furnace oxygen content by using a furnace oxygen content prediction model, wherein the furnace oxygen content prediction model is obtained by training a fuzzy neural network by using a second training data set; 2. The intelligent control method for municipal solid waste incineration process according to claim 1, characterized in that, solving a main steam flow controller and a furnace oxygen content controller by using a gradient descent algorithm to determine an optimal control rate for controlling the municipal solid waste incineration process based on the next moment main steam flow and the next moment furnace oxygen content and in combination with a switching rule based on expert knowledge, wherein the switching rule based on expert knowledge comprises a first rule, a second rule and a third rule; the first rule is that when a current moment primary air pre-accumulator back pressure is greater than a first set pressure and less than a second set pressure, a learning rate of the main steam flow controller is increased by a set value, and the current moment combustion time is reduced by a set combustion time; the second rule is that when a change rate of the current moment primary air pre-accumulator back pressure is greater than a third set pressure, it is determined that the current is in a screening process, the solving of the main steam flow controller and the furnace oxygen content controller is performed, and the first rule is not performed; when the change rate of the current moment primary air pre-accumulator back pressure is less than the third set pressure, it is determined that the current is in an incineration process, the solving of the main steam flow controller and the furnace oxygen content controller and the first rule are performed; the third rule is that when a difference between a next moment main steam flow set value and the next moment main steam flow is greater than a set difference value, the pushing is delayed, and the current moment combustion time is increased by a set combustion time; and the optimal control rate comprises a pusher speed increment, a grate speed increment, a combustion time increment, a primary air flow increment and a secondary air flow increment. The method for training the fuzzy neural network comprises the following steps: acquiring a first training data set, wherein the first training data set comprises historical moment pusher speeds, historical moment grate speeds, historical moment combustion times, historical moment main steam flows and actual values of main steam flows at corresponding moments of a training municipal solid waste incineration process; training the fuzzy neural network by taking the historical moment pusher speeds, the historical moment grate speeds, the historical moment combustion times and the historical moment main steam flows of the training municipal solid waste incineration process as inputs and taking the actual values of the main steam flows at the corresponding moments as outputs, updating model parameters of the fuzzy neural network by using an LM algorithm, and obtaining the main steam flow prediction model.

3. The intelligent control method for municipal solid waste incineration process according to claim 2, characterized in that, The fuzzy neural network is trained by taking the historical moment pusher speed, the historical moment grate speed, the historical moment combustion time and the historical moment main steam flow of the training urban solid waste incineration process as inputs, and taking the actual value of the main steam flow at the corresponding moment as output, the model parameters of the fuzzy neural network are updated by using the LM algorithm to obtain a main steam flow prediction model, specifically including: The historical moment pusher speed, the historical moment grate speed, the historical moment combustion time and the historical moment main steam flow are input into the current fuzzy neural network to obtain the predicted value of the main steam flow at the corresponding moment; According to the predicted value of the main steam flow at the corresponding moment and the actual value of the main steam flow at the corresponding moment at the corresponding moment, a loss function value is determined; Determine whether the loss function value is less than a set threshold value; If yes, the current fuzzy neural network is used as the main steam flow prediction model; If not, the model parameters of the current fuzzy neural network are updated by using the LM algorithm, and the step of inputting the historical moment pusher speed, the historical moment grate speed, the historical moment combustion time and the historical moment main steam flow into the current fuzzy neural network to obtain the predicted value of the main steam flow at the corresponding moment is returned.

4. The intelligent control method for municipal solid waste incineration process according to claim 1, wherein, The main steam flow controller is: ； wherein, is the primary steam flow control rate increment for time t+1; , is the pusher speed increment for time t+1; is the grate speed increment for time t+1; is the burn time increment for time t+1; T is the transpose; is the learning rate of the primary steam flow controller at time t; is the objective function of the primary steam flow controller; is the pusher speed at time t; is the grate speed at time t; is the burn time at time t.

5. The intelligent control method for municipal solid waste incineration process as claimed in claim 4 wherein, The objective function of the main steam flow controller is: ； wherein, is the main steam flow rate set value at the time t; is the main steam flow rate at the time t; is the main steam flow rate control rate at the time t, .

6. The intelligent control method for municipal solid waste incineration process as claimed in claim 1 wherein, The furnace oxygen amount controller is: ； wherein, is the furnace oxygen control rate increment at time t+1; , is the primary air flow increment at time t+1; is the secondary air flow increment at time t+1; T is transpose; is the learning rate of the furnace oxygen controller at time t; is the objective function of the furnace oxygen controller; is the primary air flow at time t; is the secondary air flow at time t.

7. The intelligent control method for municipal solid waste incineration process according to claim 6, characterized in that, The objective function of the furnace oxygen amount controller is: ； wherein, is the furnace oxygen amount set value at time t; is the furnace oxygen amount at time t; is the furnace oxygen amount control rate at time t, .

8. A computer device comprising: A memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor executes the computer program to implement the urban solid waste incineration process intelligent control method of any one of claims 1-7.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the urban solid waste incineration process intelligent control method of any one of claims 1-7.

10. A computer program product comprising a computer program, characterized in that, The computer program is executed by the processor to implement the urban solid waste incineration process intelligent control method of any one of claims 1-7.

Images