Coordinated system increment calculation and coal feed stratification control method, equipment and medium

Abstract

The present application discloses a method, device and medium for calculating the incremental amount of a coordinated system and controlling the coal feeding amount in layers. The method may include: establishing an online identification model of a neural network for the calorific value of coal; performing secondary division of sub-models based on the online identification model of the neural network for the calorific value of coal to form a gridded sub-model object set; establishing a dynamic mathematical model of a controlled object of a multivariable coordinated system with three inputs and three outputs to determine the internal parameters of the gridded sub-model object set; judging the current load point and the quality of the coal, and calculating the optimal control incremental vector at the current moment; distributing the total incremental amount of coal feeding according to the principle of layered control based on the number of units currently in operation of the pulverizing system and the actual amount of coal of each unit. The present invention solves the problem of controlling the boiler-machine coordination system when a thermal power unit is operating under large amplitude, high speed and unstable coal quality conditions, and improves the safety, stability and flexible adjustment capabilities of the unit.

Description

Technical Field

[0001] The present invention relates to the field of thermal automatic control of thermal power plants, and more particularly to a method, device and medium for coordinating system increment calculation and coal feed rate layered control. Background Art

[0002] In recent years, the large-scale integration of clean energy into the grid has placed higher demands on the auxiliary services provided by thermal power units in peak load regulation and frequency regulation. When grid load demands fluctuate frequently and significantly, there is an urgent need to improve the coordinated control flexibility and unit safety of thermal power units.

[0003] When the units are operating under large-scale and high-speed variable operating conditions, the coordinated control of thermal power units has the following problems:

[0004] (1) Currently, the coal sources for most units are complex, and the quality of the coal varies greatly and is difficult to measure accurately online in real time. Traditional coal quality correction methods have slow response speeds and limited accuracy. Subsequent coal quality modeling research represented by intelligent algorithms is difficult to carry out large-scale engineering applications due to factors such as the high dimensionality of the required data and the complex algorithm structure. The uncertainty of coal quality poses a great challenge to the precise input of boiler heat under coordinated control, seriously affecting the regulation accuracy and adjustment speed of the coordinated control system.

[0005] (2) The PID+proportional feedforward control strategy commonly used in thermal power units can only adapt to a small range of variable operating conditions. When the unit frequently operates under a large range of variable operating conditions, the nonlinear dynamic characteristics of the unit will become prominent, and the original control strategy, parameters, and feedforward signals will no longer be adapted, and the coordinated control quality under wide loads will deteriorate.

[0006] (3) Most of the current model-based predictive control algorithms only use the difference between the set value and the actual value of the controlled variable and the control quantity increment as optimization indicators, without considering whether the control quantity is close to the boundary value within the adjustment range. If the control quantity is adjusted near the critical value for a long time, it may bring hidden dangers to the safe operation of the unit.

[0007] (4) The boiler master control signals output by the current coordinated control system are all sent to each pulverizing system in the form of a comprehensive (single) instruction, without considering the differences in coal quality in each pulverizing bin. When the coal feed amounts of each pulverizing system are different and the calorific value of the coal used is significantly different, the comprehensive (single) instruction will change the average calorific value of the coal entering the furnace, and the parameters such as the furnace combustion conditions, steam pressure, and steam temperature will fluctuate greatly. The increase in internal disturbances will bring greater hidden dangers to the control performance and safety performance of the unit.

[0008] Therefore, it is necessary to develop a method, equipment and medium for coordinating system increment calculation and coal feed stratification control.

[0009] The information disclosed in the background technology section of the present invention is only intended to deepen the understanding of the general background technology of the present invention, and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art already known to those skilled in the art. Summary of the Invention

[0010] The present invention proposes a method, device and medium for coordinated system increment calculation and coal feed stratification control, which solves the problem of boiler-turbine coordinated system control when thermal power units operate under large amplitude, high speed and unstable coal quality conditions, and improves the safety, stability and flexible adjustment capabilities of the units.

[0011] In a first aspect, an embodiment of the present disclosure provides a method for coordinating system increment calculation and coal feed rate stratified control, comprising:

[0012] Establish a neural network online identification model for coal calorific value;

[0013] The start-stop load point of the pulverizing system and the load point of the turbine valve opening change are used as the sub-model demarcation points of the local linearization of the nonlinear system. According to the coal calorific value neural network online identification model, the sub-model is secondary divided according to the different calorific values ​​of the coal in each sub-model to form a gridded sub-model object set.

[0014] Based on the field data, a dynamic mathematical model of the controlled object of the three-input and three-output multivariable coordinated system is established, and the internal parameters of the gridded sub-model object set are determined;

[0015] Determine the current load point and coal quality, traverse the grid to obtain the dynamic mathematical model of the coordinated system controlled object that is adapted to the current working conditions, and calculate the optimal control increment vector Δu at the current moment: (Δu1, Δu2, Δu3) T , where Δu1 represents the increment of the turbine valve opening, Δu2 represents the total increment of coal feed, and Δu3 represents the increment of feed water;

[0016] According to the number of running pulverizing systems and the actual amount of coal in each system, the total increment of coal feed Δu2 is distributed according to the principle of stratified control.

[0017] Preferably, the coal calorific value neural network online identification model is established by training a long short-term memory network.

[0018] Preferably, the input data of the coal calorific value neural network online identification model include: the steam-water separator outlet pressure P1, the steam-water separator outlet temperature T1, the economizer inlet feed water temperature T2, the economizer inlet feed water pressure P2, the economizer inlet feed water flow S1, the superheater outlet steam flow S2, T0 and the discrete sequence of the instantaneous coal quantity of each coal feeder within 3 minutes {C1, C2, ..., C f}, where f is the number of discrete data under the current sampling accuracy;

[0019] The output data of the coal calorific value neural network online identification model is the industrial analysis actual measured value of the coal calorific value at time T0.

[0020] Preferably, establishing the gridded sub-model object set includes:

[0021] The minimum load of the unit under deep peak regulation is F d %Pe, maximum load is F u %Pe, there are m operating modes of the pulverizing system on the boiler side, and the corresponding load points are: F 11 %Pe,F 12 %Pe,...,F 1m %Pe;

[0022] When the steam turbine adopts the partial steam inlet mode, there are n ways to control the four main steam regulating doors of the steam turbine, and the corresponding load points are: F 21 %Pe,F 22 %Pe,...,F 2n %Pe;

[0023] Arrange the load points under the two groups of division in ascending order to obtain the preliminary sub-model load cutoff points: F1%Pe, F2%Pe, ..., F n+m %Pe, Pe is the rated power of the system;

[0024] In each load range, the sub-model is divided into two parts according to the calorific value of coal. The step disturbance data of the input and output variables of the coal calorific value neural network online identification model in the load range is used as the modeling data source to obtain a grid database covering all typical coal blending methods under full load conditions.

[0025] Preferably, the input variables of the dynamic mathematical model of the controlled object of the multivariable coordination system include the turbine throttle opening U1, the total fuel quantity U2, and the feed water flow U3;

[0026] The output variables of the dynamic mathematical model of the controlled object of the multivariable coordination system include unit power Y1, main steam pressure Y2, and separator outlet temperature Y3;

[0027] The elements in the grid database S are sequentially brought into the model to establish the grid sub-model transfer function of the coordinated control object covering all typical coal blending methods under full load conditions, which is recorded as

[0028] Preferably, the optimal control increment vector Δu at the current moment is calculated: (Δu1, Δu2, Δu3)T include:

[0029] Based on the model, a predictive controller is designed. The difference between the expected output and the actual output in the next P steps, the control increment in the next C steps, and the quadratic function of the criticality of the control quantity are used as performance indicators to obtain the optimal control increment vector Δu at the current moment: (Δu1, Δu2, Δu3) T .

[0030] Preferably, the current load point and coal quality are determined, the dynamic mathematical model of the coordinated system controlled object adapted to the current working conditions is obtained by traversing the grid, and the optimal control increment vector Δu at the current moment is calculated: (Δu1, Δu2, Δu3) T include:

[0031] According to the current load point and coal calorific value, the mathematical model of the coordinated system controlled object that matches the model set is automatically selected and recorded as T q , based on model T q A predictive controller is designed, using the difference between the expected output and the actual output in the next P steps, the control increment in the next C steps, and the quadratic function of the critical degree of the control quantity as performance indicators. The performance indicator J(k) of the algorithm at time k is:

[0032] J(k)=[YY S ] T Q[YY S ]+△u T R△u-(U+△uU min ) T W[U max -(U+△u)]

[0033] Where Y is the actual value of the output, which includes the predicted information of the output at future moments; S is the output set value; Q is the output weight coefficient matrix; △u is the control quantity increment at time k; R is the control quantity increment weight coefficient matrix; U is the control quantity at time k-1; U min and U max are the lower limit and upper limit of the control quantity interval respectively; W is the weight coefficient matrix of the critical degree of the control quantity, where:

[0034] Q=diag(Q1,Q2,...Q i ,...,Q P ), Q i =diag(q i )

[0035] R=diag(R1,R2,...R i ,...,R C ), R i=diag(r i )

[0036] W=diag(W1,W2,...W i ,...,W C ), W i =diag(w i )

[0037] Where q i is the weight coefficient corresponding to the i-th controlled quantity, Q i q i The diagonal matrix of r i is the weight coefficient corresponding to the i-th control quantity increment, R i For r i The diagonal matrix of w i is the weight coefficient corresponding to the critical degree of the i-th control quantity, W i w i The diagonal matrix of ;

[0038] The optimal control increment vector Δu at the current moment is obtained by minimizing the performance index: (Δu1, Δu2, Δu3) T .

[0039] Preferably, according to the number of pulverizing systems currently in operation and the actual coal quantity of each system, the total increment of coal feed Δu2 is distributed according to the principle of stratified control, including:

[0040] Calculate the current coal feeding ratio of each coal feeder:

[0041] P j =M j / (M1+M2+...+M c ), j=1,2,...,c

[0042] Where M j is the instantaneous coal feeding amount of the jth coal feeder at time k, c is the number of coal feeders in operation, P j is the proportion of coal supply of the j-th coal feeder to the total coal supply;

[0043] P j It represents the proportion of coal feeding rate of a single coal feeder, and also the contribution rate of the coal calorific value of a single coal feeder to the total combustion calorific value. In order to ensure the stability of the total combustion calorific value after the control amount changes, the increment of the coal feeding amount of a single coal feeder is:

[0044] △u 2j =△u2×P j ,j=1,2,...,c

[0045] And satisfy: U cmin ≤U 2j +△u2j ≤U cmax

[0046] Where: △u2 is the total increment of coal feeding; U 2j is the actual coal quantity of the j-th coal feeder at time k-1; △u 2j is the coal feeding increment of the j-th coal feeder at time k; U cmin and U cmax They are the lower and upper limits of coal quantity for a single coal feeder respectively.

[0047] As a specific implementation of the embodiment of the present disclosure,

[0048] In a second aspect, an embodiment of the present disclosure further provides an electronic device, the electronic device comprising:

[0049] a memory storing executable instructions;

[0050] A processor runs the executable instructions in the memory to implement the method for coordinating system increment calculation and coal feed rate stratification control.

[0051] In a third aspect, an embodiment of the present disclosure further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for coordinating system incremental calculation and coal feed quantity stratified control.

[0052] The methods and apparatus of the present invention have other features and advantages that will be apparent from or will be described in detail in the accompanying drawings and subsequent detailed descriptions incorporated herein, which together serve to explain the specific principles of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] The above and other objects, features and advantages of the present invention will become more apparent through a more detailed description of exemplary embodiments of the present invention with reference to the accompanying drawings, wherein like reference numerals generally represent like components throughout the exemplary embodiments of the present invention.

[0054] Figure 1 A flow chart showing the steps of a method for coordinating system increment calculation and coal feed rate stratified control according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0055] The preferred embodiments of the present invention will be described in more detail below. Although the preferred embodiments of the present invention are described below, it should be understood that the present invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0056] To facilitate understanding of the solutions and effects of the embodiments of the present invention, three specific application examples are given below. Those skilled in the art should understand that these examples are only for facilitating understanding of the present invention, and any specific details thereof are not intended to limit the present invention in any way.

[0057] Example 1

[0058] Figure 1 A flow chart showing the steps of a method for coordinating system increment calculation and coal feed rate stratified control according to an embodiment of the present invention is shown.

[0059] like Figure 1 As shown, the method for incremental calculation and stratified control of coal feeding amount of the coordinated system includes: step 101, establishing a neural network online identification model of coal calorific value; step 102, using the start-stop load point of the pulverizing system and the load point of the turbine valve opening change as the sub-model demarcation point of the local linearization of the nonlinear system, and performing secondary sub-model division according to the different calorific values ​​of the coal in each sub-model according to the neural network online identification model of the coal calorific value, to form a gridded sub-model object set; step 103, based on field data, establishing a dynamic mathematical model of the controlled object of the three-input and three-output multivariable coordinated system, and determining the internal parameters of the gridded sub-model object set; step 104, judging the current load point and coal quality, traversing the grid to obtain the dynamic mathematical model of the controlled object of the coordinated system adapted to the current working conditions, and calculating the optimal control increment vector Δu at the current moment: (Δu1, Δu2, Δu3) T , where Δu1 represents the increment of the turbine valve opening, Δu2 represents the total increment of coal feed, and Δu3 represents the increment of water feed; step 105, according to the number of running units of the current pulverizing system and the actual amount of coal of each unit, the increment of coal feed Δu2 is distributed according to the principle of stratified control.

[0060] In one example, a long short-term memory network is trained to establish an online recognition model of coal calorific value neural network.

[0061] In one example, the input data of the coal calorific value neural network online identification model include: the steam-water separator outlet pressure P1, the steam-water separator outlet temperature T1, the economizer inlet feed water temperature T2, the economizer inlet feed water pressure P2, the economizer inlet feed water flow S1, the superheater outlet steam flow S2, T0 and the discrete sequence of the instantaneous coal quantity of each coal feeder within 3 minutes {C1, C2, ..., C f}, where f is the number of discrete data under the current sampling accuracy;

[0062] The output data of the coal calorific value neural network online identification model is the industrial analysis measured value of the coal calorific value at time T0.

[0063] In one example, establishing a meshed sub-model object set includes:

[0064] The minimum load of the unit under deep peak regulation is F d %Pe, maximum load is F u %Pe, there are m operating modes of the pulverizing system on the boiler side, and the corresponding load points are: F 11 %Pe,F 12 %Pe,...,F 1m %Pe;

[0065] When the steam turbine adopts the partial steam inlet mode, there are n ways to control the four main steam regulating doors of the steam turbine, and the corresponding load points are: F 21 %Pe,F 22 %Pe,...,F 2n %Pe;

[0066] Arrange the load points under the two groups of division in ascending order to obtain the preliminary sub-model load cutoff points: F1%Pe, F2%Pe, ..., F n+m %Pe, Pe is the rated power of the system;

[0067] In each load range, the sub-model is divided into two parts according to the calorific value of coal. The step disturbance data of the input and output variables of the coal calorific value neural network online identification model in the load range is used as the modeling data source to obtain a grid database covering all typical coal blending methods under full load conditions.

[0068] In one example, the input variables of the dynamic mathematical model of the controlled object of the multivariable coordination system include the turbine throttle opening U1, the total fuel quantity U2, and the feed water flow rate U3;

[0069] The output variables of the dynamic mathematical model of the controlled object of the multivariable coordination system include unit power Y1, main steam pressure Y2, and separator outlet temperature Y3;

[0070] The elements in the grid database S are brought into the model one by one, and the grid sub-model transfer function covering the coordinated control object of all typical coal blending methods under full load conditions is established, which is recorded as

[0071]

[0072] In one example, the optimal control increment vector Δu at the current moment is calculated as: (Δu1, Δu2, Δu3) T include:

[0073] Based on the model, a predictive controller is designed. The difference between the expected output and the actual output in the next P steps, the control increment in the next C steps, and the quadratic function of the criticality of the control quantity are used as performance indicators to obtain the optimal control increment vector Δu at the current moment: (Δu1, Δu2, Δu3) T .

[0074] In one example, the current load point and coal quality are determined, the grid is traversed to obtain the dynamic mathematical model of the coordinated system controlled object that is adapted to the current operating conditions, and the optimal control increment vector Δu at the current moment is calculated: (Δu1, Δu2, Δu3) T include:

[0075] According to the current load point and coal calorific value, the mathematical model of the coordinated system controlled object that matches the model set is automatically selected and recorded as T q , based on model T q A predictive controller is designed, using the difference between the expected output and the actual output in the next P steps, the control increment in the next C steps, and the quadratic function of the critical degree of the control quantity as performance indicators. The performance indicator J(k) of the algorithm at time k is:

[0076] J(k)=[YY S ] T Q[YY S ]+△u T R△u-(U+△uU min ) T W[U max -(U+△u)]

[0077] Where Y is the actual value of the output, which includes the predicted information of the output at future moments; S is the output set value; Q is the output weight coefficient matrix; △u is the control quantity increment at time k; R is the control quantity increment weight coefficient matrix; U is the control quantity at time k-1; U min and U max are the lower limit and upper limit of the control quantity interval respectively; W is the weight coefficient matrix of the critical degree of the control quantity, where:

[0078] Q=diag(Q1,Q2,...Q i ,...,Q P ), Q i =diag(q i )

[0079] R=diag(R1,R2,...R i ,...,R C ), R i =diag(r i )

[0080] W=diag(W1,W2,...W i ,...,W C ), W i =diag(w i )

[0081] Where q i is the weight coefficient corresponding to the i-th controlled quantity, Q i q i The diagonal matrix of r i is the weight coefficient corresponding to the i-th control quantity increment, R i For r i The diagonal matrix of i is the weight coefficient corresponding to the critical degree of the i-th control quantity, W i w i The diagonal matrix of ;

[0082] The optimal control increment vector Δu at the current moment is obtained by minimizing the performance index: (Δu1, Δu2, Δu3) T .

[0083] In one example, based on the number of pulverizing systems currently in operation and the actual coal quantity of each system, the total increment of coal feed Δu2 is distributed according to the principle of stratified control, including:

[0084] Calculate the current coal feeding ratio of each coal feeder:

[0085] P j =M j / (M1+M2+...+M c ), j=1,2,...,c

[0086] Where M j is the instantaneous coal feeding amount of the jth coal feeder at time k, c is the number of coal feeders in operation, P j is the proportion of coal supply of the j-th coal feeder to the total coal supply;

[0087] P j It represents the proportion of coal feeding rate of a single coal feeder, and also the contribution rate of the coal calorific value of a single coal feeder to the total combustion calorific value. In order to ensure the stability of the total combustion calorific value after the control amount changes, the increment of the coal feeding amount of a single coal feeder is:

[0088] △u 2j =△u2×P j ,j=1,2,...,c

[0089] And satisfy: U cmin ≤U 2j +△u 2j ≤U cmax

[0090] Where: △u2 is the total increment of coal feeding; U 2j is the actual coal quantity of the j-th coal feeder at time k-1; △u 2j is the coal feeding increment of the j-th coal feeder at time k; U cmin and U cmax They are the lower and upper limits of coal quantity for a single coal feeder respectively.

[0091] Specifically, a long short-term memory (LSTM) network is used for training to establish a coal calorific value neural network online identification model with a streamlined structure, high fitting accuracy, and low dependence on data volume. The advantages are:

[0092] (1) LSTM neural network is suitable for processing and predicting events with long intervals and delays in time series, which is consistent with the large inertia and large delay characteristics of the controlled objects in the coordination system. In terms of principle and structure, it ensures the rationality and reliability of the model output.

[0093] (2) Due to the long period of coal quality testing and insufficient accumulation of historical data, the calorific value estimation of coal based on LSTM neural network adapted to small sample data has an accurate and stable estimation effect.

[0094] The input data of the neural network online identification model of coal calorific value are: the steam-water separator outlet pressure P1, steam-water separator outlet temperature T1, economizer inlet feed water temperature T2, economizer inlet feed water pressure P2, economizer inlet feed water flow S1, superheater outlet steam flow S2, T0 and the discrete sequence of instantaneous coal quantity of each coal feeder within 3 minutes {C1, C2, ..., C f}, where f is the number of discrete data points at the current sampling accuracy. The output data of the coal calorific value neural network online identification model is: the industrial analysis measured value of coal calorific value at time T0.

[0095] The start-stop load point of the pulverizing system and the load point of the turbine valve opening change are used as the sub-model demarcation points for the local linearization of the nonlinear system. Within each sub-model, secondary sub-model division is performed according to the different calorific values ​​of the coal to form a gridded sub-model object set, which specifically includes:

[0096] Under deep peak regulation, if the minimum load of the unit is F d %Pe, maximum load is F u %Pe, depending on the load range, the operating state of the boiler side pulverizing system has m modes, and the corresponding load points are: F 11 %Pe,F 12 %Pe,...,F 1m %Pe;

[0097] When the steam turbine adopts the partial steam inlet mode, there are n ways to control the four main steam regulating doors of the steam turbine, and the corresponding load points are: F 21 %Pe,F 22 %Pe,...,F 2n %Pe;

[0098] Arrange the load points under the two groups of division in ascending order to obtain the preliminary sub-model load cutoff points: F1%Pe, F2%Pe, ..., F n+m %Pe, Pe is the rated power of the system;

[0099] In each load range, the sub-model is divided into two parts according to the calorific value of coal. Taking the [F1%Pe, F2%Pe] range as an example, there are a typical ways of coal distribution in actual operation, corresponding to the calorific value of coal of type a {R1, R2, ..., Ra}. When R1 coal is selected, the step disturbance data of the input and output variables of the coal calorific value neural network online identification model in the [F1%Pe, F2%Pe] load range is used as one of the modeling data sources, denoted as S 1,2,1 , and so on, we get a grid database covering all typical coal blending methods under full load conditions

[0100] Based on field data, a dynamic mathematical model of the controlled object of the three-input and three-output multivariable coordinated system is established, and the internal parameters of the gridded sub-model object set are determined. The input variables of the multivariable model are the turbine valve opening U1, the total fuel quantity U2, and the feed water flow rate U3; the output variables are the unit power Y1, the main steam pressure Y2, and the separator outlet temperature Y3. The elements in the grid data S containing the step disturbance characteristics are sequentially brought into the model to establish the gridded sub-model transfer function of the coordinated control object covering all typical coal blending methods under full load conditions, which is recorded as

[0101] Determine the current load point and coal quality, traverse the grid to obtain a dynamic mathematical model of the coordinated system controlled object that is compatible with the current operating conditions, design a predictive controller based on the model, and use the difference between the expected output and the actual output in the next P steps, the control increment in the next C steps, and the quadratic function of the criticality of the control quantity as performance indicators to obtain the optimal control increment vector Δu at the current moment: (Δu1, Δu2, Δu3) T , where Δu1 represents the increment of the turbine valve opening, Δu2 represents the increment of the coal supply, and Δu3 represents the increment of the water supply. Based on the current load point and the calorific value of the coal, the mathematical model of the controlled object of the coordinated system that is suitable for the current load point is automatically selected in the model set and is denoted as T q , based on model T qA predictive controller is designed, using the difference between the expected output and the actual output in the next P steps, the control increment in the next C steps, and the quadratic function of the critical degree of the control quantity as performance indicators. The performance indicator J(k) of the algorithm at time k is:

[0102] J(k)=[YY S ] T Q[YY S ]+△u T R△u-(U+△uU min ) T W[U max -(U+△u)]

[0103] Where Y is the actual value of the output, which includes the predicted information of the output at future moments; S is the output set value; Q is the output weight coefficient matrix; △u is the control quantity increment at time k; R is the control quantity increment weight coefficient matrix; U is the control quantity at time k-1; U min and U max are the lower limit and upper limit of the control quantity interval respectively; W is the weight coefficient matrix of the critical degree of the control quantity. Among them:

[0104] Q=diag(Q1,Q2,...Q i ,...,Q P ), Q i =diag(q i )

[0105] R=diag(R1,R2,...R i ,...,R C ), R i =diag(r i )

[0106] W=diag(W1,W2,...W i ,...,W C ), W i =diag(w i )

[0107] Where q i is the weight coefficient corresponding to the i-th controlled quantity, Q i q i The diagonal matrix of r i is the weight coefficient corresponding to the i-th control quantity increment, R i For r i The diagonal matrix of i is the weight coefficient corresponding to the critical degree of the i-th control quantity, W i w i The diagonal matrix of .

[0108] The optimal control increment vector Δu at the current moment is obtained by minimizing the performance index: (Δu1, Δu2, Δu3) T , where Δu1 represents the increase in the opening of the turbine regulating valve, Δu2 represents the total increase in the coal feed rate, and Δu3 represents the increase in the water feed rate.

[0109] According to the number of pulverizing systems currently in operation and the actual amount of coal in each system, the total increment of coal feed Δu2 is distributed according to the principle of stratified control to ensure that the dynamic adjustment will not lead to deviation of the combustion calorific value in the furnace, while reducing the large fluctuations in steam temperature and pressure caused by internal disturbances on the boiler side.

[0110] Calculate the current coal feeding ratio of each coal feeder:

[0111] P j =M j / (M1+M2+...+M c ), j=1,2,...,c

[0112] Where M j is the instantaneous coal feeding amount of the jth coal feeder at time k, c is the number of coal feeders in operation, P j is the proportion of coal supply of the j-th coal feeder to the total coal supply;

[0113] P j It represents the proportion of coal feeding rate of a single coal feeder, and also the contribution rate of the coal calorific value of a single coal feeder to the total combustion calorific value. In order to ensure the stability of the total combustion calorific value after the control amount changes, the increment of the coal feeding amount of a single coal feeder is:

[0114] △u 2j =△u2×P j ,j=1,2,...,c

[0115] And satisfy: U cmin ≤U 2j +△u 2j ≤U cmax

[0116] Where: △u2 is the total increment of coal feeding; U 2j is the actual coal quantity of the j-th coal feeder at time k-1; △u 2j is the coal feeding increment of the j-th coal feeder at time k; U cmin and U cmax They are the lower and upper limits of coal quantity for a single coal feeder respectively.

[0117] The present invention significantly improves the load tracking capability and operating stability of the unit under large-scale and high-speed variable operating conditions and enhances the adaptability of the control system through technical means such as coal quality identification under small sample data, global modeling of nonlinear controlled objects and automatic model screening, indicator optimization of model predictive control, and stratified control of coal feed rate.

[0118] Taking the coordinated system of a 660MW supercritical unit at a power plant as an example, this paper describes the present invention in detail using system modeling and optimization control methods. 198MW (30% Pe), 440MW (66.7% Pe), 528MW (80% Pe), and 660MW (100% Pe) were selected as the initial sub-model load cutoff points. Three typical coal blending methods were used for the units, corresponding to three types of coal heating value. Grid database data covering all typical coal blending methods under full-load conditions was used for modeling and predictive control optimization.

[0119] Practical application shows that when the unit load change is 99MW (15%Pe) and the load change rate is 7.92MW / min (1.2%Pe / min), the maximum dynamic load deviation is reduced from the original 9.2MW to 4.5MW, and the maximum dynamic deviation of the main steam pressure is reduced from the original 1.2MPa to 0.75MPa. The main steam temperature deviation and reheat steam temperature deviation are controlled within ±5℃ and ±5.5℃ respectively. The speed and accuracy of the coordinated control system are significantly improved.

[0120] Example 2

[0121] The present disclosure provides an electronic device comprising: a memory storing executable instructions; and a processor running the executable instructions in the memory to implement the above-mentioned method for coordinating system increment calculation and coal feed amount stratified control.

[0122] An electronic device according to an embodiment of the present disclosure includes a memory and a processor.

[0123] The memory is used to store non-transitory computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), a hard disk, flash memory, etc.

[0124] The processor may be a central processing unit (CPU) or other form of processing unit having data processing capability and / or instruction execution capability, and may control other components in the electronic device to perform desired functions. In one embodiment of the present disclosure, the processor is used to execute the computer-readable instructions stored in the memory.

[0125] Those skilled in the art should understand that in order to solve the technical problem of how to obtain a good user experience, this embodiment may also include well-known structures such as a communication bus and an interface, and these well-known structures should also be included in the scope of protection of this disclosure.

[0126] For detailed description of this embodiment, please refer to the corresponding description in the aforementioned embodiments, which will not be repeated here.

[0127] Example 3

[0128] An embodiment of the present disclosure provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the method for coordinating system increment calculation and coal feed rate stratified control is implemented.

[0129] According to an embodiment of the present disclosure, a computer-readable storage medium stores non-transitory computer-readable instructions, which, when executed by a processor, execute all or part of the steps of the aforementioned methods of the embodiments of the present disclosure.

[0130] The above-mentioned computer-readable storage media include, but are not limited to, optical storage media (e.g., CD-ROMs and DVDs), magneto-optical storage media (e.g., MOs), magnetic storage media (e.g., magnetic tapes or mobile hard disks), media with built-in rewritable non-volatile memory (e.g., memory cards), and media with built-in ROM (e.g., ROM cartridges).

[0131] Those skilled in the art should understand that the above description of the embodiments of the present invention is only for the purpose of illustrative purposes only to illustrate the beneficial effects of the embodiments of the present invention, and is not intended to limit the embodiments of the present invention to any given examples.

[0132] While various embodiments of the present invention have been described above, the above description is intended to be illustrative, not exhaustive, and not limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

Claims

1. A method for coordinating system increment calculation and coal feed rate stratification control, characterized in that: include: Establish a neural network online identification model for coal calorific value; The start-stop load point of the pulverizing system and the load point of the turbine valve opening change are used as the sub-model demarcation points of the local linearization of the nonlinear system. According to the coal calorific value neural network online identification model, the sub-model is secondary divided according to the different calorific values ​​of the coal in each sub-model to form a gridded sub-model object set. Based on the field data, a dynamic mathematical model of the controlled object of the three-input and three-output multivariable coordinated system is established, and the internal parameters of the gridded sub-model object set are determined; Determine the current load point and coal quality, traverse the grid to obtain the dynamic mathematical model of the coordinated system controlled object that is adapted to the current working conditions, and calculate the optimal control increment vector Δu at the current moment: (Δu1, Δu2, Δu3) T , where Δu1 represents the increment of the turbine valve opening, Δu2 represents the total increment of coal feed, and Δu3 represents the increment of feed water; According to the number of running pulverizing systems and the actual amount of coal in each system, the coal feed increment Δu2 is distributed according to the principle of stratified control.

2. The method for coordinated system increment calculation and coal feed rate layered control according to claim 1, wherein: The coal calorific value neural network online identification model is established by training the long short-term memory network.

3. The method for coordinated system increment calculation and coal feed rate layered control according to claim 1, wherein: The input data of the coal calorific value neural network online identification model include: the steam-water separator outlet pressure P1, the steam-water separator outlet temperature T1, the economizer inlet feed water temperature T2, the economizer inlet feed water pressure P2, the economizer inlet feed water flow S1, the superheater outlet steam flow S2, T0 and the discrete sequence of the instantaneous coal quantity of each coal feeder within 3 minutes {C1, C2, ..., C f }, where f is the number of discrete data under the current sampling accuracy; The output data of the coal calorific value neural network online identification model is the industrial analysis actual measured value of the coal calorific value at time T0.

4. The method for coordinated system increment calculation and coal feed rate layered control according to claim 1, wherein: Establishing the gridded sub-model object set includes: The minimum load of the unit under deep peak regulation is F d %Pe, maximum load is F u %Pe, there are m operating modes of the pulverizing system on the boiler side, and the corresponding load points are: F 11 %Pe,F 12 %Pe,...,F 1m %Pe; When the steam turbine adopts the partial steam inlet mode, there are n ways to control the four main steam regulating doors of the steam turbine, and the corresponding load points are: F 21 %Pe,F 22 %Pe,...,F 2n %Pe; Arrange the load points under the two groups of division in ascending order to obtain the preliminary sub-model load cutoff points: F1%Pe, F2%Pe, ..., F n+m %Pe, Pe is the rated power of the system; In each load range, the sub-model is divided into two parts according to the calorific value of coal. The step disturbance data of the input and output variables of the coal calorific value neural network online identification model in the load range is used as the modeling data source to obtain a grid database covering all typical coal blending methods under full load conditions.

5. The method for coordinated system increment calculation and coal feed rate layered control according to claim 4, wherein: The input variables of the dynamic mathematical model of the controlled object of the multivariable coordination system include the turbine valve opening U1, the total fuel quantity U2, and the feed water flow rate U3; The output variables of the dynamic mathematical model of the controlled object of the multivariable coordination system include unit power Y1, main steam pressure Y2, and separator outlet temperature Y3; The elements in the grid database S are sequentially brought into the model to establish the grid sub-model transfer function of the coordinated control object covering all typical coal blending methods under full load conditions, which is recorded as 6. The method for coordinated system increment calculation and coal feeding amount layered control according to claim 1, wherein: Calculate the optimal control increment vector Δu at the current moment: (Δu1, Δu2, Δu3) T include: Based on the model, a predictive controller is designed. The difference between the expected output and the actual output in the next P steps, the control increment in the next C steps, and the quadratic function of the criticality of the control quantity are used as performance indicators to obtain the optimal control increment vector Δu at the current moment: (Δu1, Δu2, Δu3) T .

7. The method for coordinated system increment calculation and coal feed rate layered control according to claim 6, wherein: Determine the current load point and coal quality, traverse the grid to obtain the dynamic mathematical model of the coordinated system controlled object that is adapted to the current working conditions, and calculate the optimal control increment vector Δu at the current moment: (Δu1, Δu2, Δu3) T include: According to the current load point and coal calorific value, the mathematical model of the coordinated system controlled object that matches the model set is automatically selected and recorded as T q , based on model T q A predictive controller is designed, using the difference between the expected output and the actual output in the next P steps, the control increment in the next C steps, and the quadratic function of the critical degree of the control quantity as performance indicators. The performance indicator J(k) of the algorithm at time k is: J(k)=[Y-Y S ] T Q[Y-Y S ]+△u T R△u-(U+△u-U min ) T W[U max -(U+△u)] Where Y is the actual value of the output, which includes the predicted information of the output at future moments; S is the output set value; Q is the output weight coefficient matrix; △u is the control quantity increment at time k; R is the control quantity increment weight coefficient matrix; U is the control quantity at time k-1; U min and U max are the lower limit and upper limit of the control quantity interval respectively; W is the weight coefficient matrix of the critical degree of the control quantity, where: Q=diag(Q1,Q2,...Q i ,...,Q P ),Q i =diag(q i ) R=diag(R1,R2,...R i ,...,R C ),R i =diag(r i ) W=diag(W1,W2,...W i ,...,IN C ),IN i =diag(w i ) Where q i is the weight coefficient corresponding to the i-th controlled quantity, Q i q i The diagonal matrix of r i is the weight coefficient corresponding to the i-th control quantity increment, R i For r i The diagonal matrix of i is the weight coefficient corresponding to the critical degree of the i-th control quantity, W i w i The diagonal matrix of ; The optimal control increment vector Δu at the current moment is obtained by minimizing the performance index: (Δu1, Δu2, Δu3) T .

8. The method for coordinated system increment calculation and coal feed rate layered control according to claim 1, wherein: According to the number of pulverizing systems currently in operation and the actual amount of coal in each system, the total increment of coal feed Δu2 is distributed according to the principle of stratified control, including: Calculate the current coal feeding ratio of each coal feeder: P j =M j / (M1+M2+...+M c ),j=1,2,...,c Where M j is the instantaneous coal feeding amount of the jth coal feeder at time k, c is the number of coal feeders in operation, P j is the proportion of coal supply of the j-th coal feeder to the total coal supply; P j It represents the proportion of coal feeding rate of a single coal feeder, and also the contribution rate of the coal calorific value of a single coal feeder to the total combustion calorific value. In order to ensure the stability of the total combustion calorific value after the control amount changes, the increment of the coal feeding amount of a single coal feeder is: △u 2j =△u2×P j ,j=1,2,...,c And satisfy: U cmin ≤U 2j +△u 2j ≤U cmax Where: △u2 is the total increment of coal feeding; U 2j is the actual coal quantity of the j-th coal feeder at time k-1; △u 2j is the coal feeding increment of the j-th coal feeder at time k; U cmin and U cmax They are the lower and upper limits of coal quantity for a single coal feeder respectively.

9. An electronic device, characterized in that: The electronic device comprises: a memory storing executable instructions; A processor, wherein the processor runs the executable instructions in the memory to implement the method for coordinated system increment calculation and coal feed rate stratification control according to any one of claims 1 to 8.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method for coordinated system increment calculation and coal feed rate layered control according to any one of claims 1 to 8.

Images