CN101709863A

CN101709863A - Hybrid control method for furnace pressure system of coal-fired boiler

Info

Publication number: CN101709863A
Application number: CN200910155792A
Authority: CN
Inventors: 张日东; 薛安克; 葛铭; 王俊宏; 李春富
Original assignee: Hangzhou Electronic Science and Technology University
Current assignee: Hangzhou Electronic Science and Technology University
Priority date: 2009-12-18
Filing date: 2009-12-18
Publication date: 2010-05-19
Anticipated expiration: 2029-12-18
Also published as: CN101709863B

Abstract

The invention relates to a hybrid control method for a furnace pressure system of a coal-fired boiler, which is characterized by firstly establishing a process model based on the real-time process data of the furnace pressure of the coal-fired boiler and digging out the basic process characteristics; then establishing a proportional-integral (PI) control circuit based on the process model; and finally implementing predictive PI control on PI differentiation control and the furnace pressure object of the coal-fired boiler wholly by computing the parameters of a predictive PI controller. The method of the invention makes up for the deficiency of traditional control, effectively facilitates the design of the controller, ensures the control performance to be elevated and simultaneously meets the given production performance index. The control technology provided by the invention can effectively reduce the error between the technological parameters of the ideal furnace pressure and the actual furnace pressure, further make up for the deficiency of the traditional controller and simultaneously ensure the control device to operate in the optimum state so as to ensure the technological parameter of the furnace pressure in the production process to be strictly controlled.

Description

Hybrid control method for furnace pressure system of coal-fired boiler

技术领域technical field

本发明属于自动化技术领域，涉及一种燃煤锅炉炉膛压力系统的预测比例积分控制(预测PI)与比例积分微分控制(PID)的混合控制方法。The invention belongs to the technical field of automation, and relates to a hybrid control method of predictive proportional integral control (predictive PI) and proportional integral differential control (PID) of a furnace pressure system of a coal-fired boiler.

背景技术Background technique

燃煤锅炉是电力生产部门的重要动力设备，其要求是供给合格的蒸汽，使燃煤锅炉发汽量适应负荷的需要。为此，生产过程的各个主要工艺参数必须严格控制。然而燃煤锅炉设备是一个复杂的被控对象，输入量与输出量之间相互关联。对于锅炉炉膛压力系统来说：蒸汽负荷发生变化引起蒸汽压力和过热蒸汽温度变化的同时，也引起炉膛压力发生变化；燃料量的变化直接影响蒸汽压力，过热蒸汽温度、过剩空气和炉膛负压的变化；减温水的变化会导致过热蒸汽温度、蒸汽压力发生变化，进一步导致炉膛压力等的变化。这些不利因素导致传统的控制手段精度不高，又进一步导致后续生产控制参数不稳定，产品合格率低，锅炉效率低下。目前实际工业中燃煤锅炉的炉膛压力控制基本上采用传统的简单的控制手段，甚至手动操作，控制参数完全依赖技术人员经验，使生产成本增加，控制效果很不理想。我国燃煤锅炉控制与优化技术比较落后，能耗居高不下，控制性能差，自动化程度低，很难适应节能减排以及间接环境保护的需求，这其中直接的影响因素之一便是燃煤锅炉系统的控制方案问题。Coal-fired boiler is an important power equipment in the power production department. Its requirement is to supply qualified steam so that the steam output of coal-fired boiler can meet the needs of the load. For this reason, each main process parameter of the production process must be strictly controlled. However, coal-fired boiler equipment is a complex controlled object, and the input and output are interrelated. For the boiler furnace pressure system: the change of steam load causes the change of steam pressure and superheated steam temperature, and also causes the change of furnace pressure; the change of fuel quantity directly affects the steam pressure, superheated steam temperature, excess air and furnace negative pressure. Changes; changes in desuperheating water will lead to changes in superheated steam temperature and steam pressure, which will further lead to changes in furnace pressure. These unfavorable factors lead to low precision of traditional control methods, which further lead to instability of subsequent production control parameters, low product qualification rate, and low boiler efficiency. At present, the furnace pressure control of coal-fired boilers in the actual industry basically adopts traditional simple control means, even manual operation, and the control parameters are completely dependent on the experience of technicians, which increases the production cost and the control effect is not ideal. my country's coal-fired boiler control and optimization technology is relatively backward, with high energy consumption, poor control performance, and low degree of automation. It is difficult to meet the needs of energy conservation, emission reduction, and indirect environmental protection. One of the direct influencing factors is coal combustion. Control scheme issues for boiler systems.

发明内容Contents of the invention

本发明的目标是针对现有的燃煤锅炉炉膛压力系统控制技术的不足之处，提供一种燃煤锅炉炉膛压力系统混合控制方法，具体是基于预测比例积分与比例积分微分控制的混合控制方法。该方法弥补了传统控制方式的不足，保证控制具有较高的精度和稳定性的同时，也保证形式简单并满足实际工业过程的需要。The object of the present invention is to provide a hybrid control method for the furnace pressure system of a coal-fired boiler for the shortcomings of the existing coal-fired boiler furnace pressure system control technology, specifically a hybrid control method based on predictive proportional-integral and proportional-integral-derivative control . This method makes up for the deficiency of the traditional control method, ensures high precision and stability of the control, and at the same time ensures the simplicity of the form and meets the needs of the actual industrial process.

本发明方法首先基于燃煤锅炉炉膛压力实时过程数据建立过程模型，挖掘出基本的过程特性；然后基于该过程模型建立比例积分微分控制回路；最后通过计算预测PI控制器的参数，将比例积分微分控制与燃煤锅炉炉膛压力对象整体实施预测PI控制。The method of the present invention first establishes a process model based on the real-time process data of the furnace pressure of a coal-fired boiler, and digs out the basic process characteristics; then establishes a proportional-integral-differential control loop based on the process model; Control and coal-fired boiler furnace pressure object integral implementation of predictive PI control.

本发明的技术方案是通过数据采集、过程辨识、预测机理、数据驱动、优化等手段，确立了一种燃煤锅炉炉膛压力系统的预测PI与比例积分微分控制的混合控制方法，利用该方法可有效提高控制的精度，，同时满足给定的生产性能指标。The technical solution of the present invention is to establish a hybrid control method of predicting PI and proportional-integral-derivative control of the furnace pressure system of a coal-fired boiler through means such as data collection, process identification, prediction mechanism, data drive, and optimization. Effectively improve the accuracy of control, while meeting the given production performance indicators.

本发明方法的步骤包括：The steps of the inventive method comprise:

(1)利用燃煤锅炉炉膛压力实时过程数据建立过程模型，具体方法是：(1) Use the real-time process data of coal-fired boiler furnace pressure to establish a process model, the specific method is:

首先建立燃煤锅炉炉膛压力实时运行数据库，通过数据采集装置采集N组实时过程运行数据，将采集的实时过程运行数据作为数据驱动的样本集合，表示为{X_i，y(i)}_i＝1 ^N，i＝1，2，…，N，其中X_i表示第i组工艺参数的输入数据，y(i)表示第i组工艺参数的输出值。First, a real-time operation database of coal-fired boiler furnace pressure is established, and N sets of real-time process operation data are collected through the data acquisition device, and the collected real-time process operation data is used as a data-driven sample set, expressed as {X _i , y(i)} _{i= 1} ^N , i=1, 2, ..., N, where Xi _i represents the input data of the i-th group of process parameters, and y(i) represents the output value of the i-th group of process parameters.

然后以该炉膛压力实时过程运行数据集合为基础建立基于最小二乘法的离散差分方程形式的局部受控自回归滑动平均模型：Then, based on the real-time process operation data set of the furnace pressure, a locally controlled autoregressive moving average model in the form of discrete difference equations based on the least squares method is established:

y_L(k)＝Φ^TX，Φ＝[a′₁，a′₂，…，a′_n，b′₀，b′₁，…，b′_m-1]^T y _L (k)=Φ ^T X, Φ=[a′ ₁ , a′ ₂ ,…, a′ _n , b′ ₀ , b′ ₁ ,…, b′ _m-1 ] ^T

X＝[y(k-1)，…，y(k-n)，u(k-d-1)，…，u(k-d-m)]^T X=[y(k-1),...,y(kn), u(kd-1),...,u(kdm)] ^T

其中，y_L(k)表示当前时刻过程模型的工艺参数的输出值，X表示过程模型的工艺参数的过去时刻的输入和输出数据的集合，u(k)表示当前过程模型工艺参数对应的控制变量，k为当前的递推步数，Φ表示通过辨识得到的模型参数的集合，T表示矩阵的转置，n，m，d+1分别为对应实际过程的输出变量阶次、输入变量阶次、实际过程的时滞。Among them, y _L (k) represents the output value of the process parameters of the process model at the current moment, X represents the set of input and output data of the process parameters of the process model at the past time, u(k) represents the control corresponding to the process parameters of the current process model variable, k is the current number of recursion steps, Φ represents the set of model parameters obtained through identification, T represents the transposition of the matrix, n, m, d+1 are the output variable order and input variable order corresponding to the actual process, respectively. Second, the time delay of the actual process.

采用的辨识手段为：The identification methods used are:

${Φ Φ}_{k k} = = {Φ Φ}_{k k - - 11} + + \overset{&OverBar; &OverBar;}{K K} ((k k)) [[y the y ((k k)) - - {Φ Φ}_{k k}^{T T} {X x}_{k k}]]$

$\overset{&OverBar; &OverBar;}{K K} ((k k)) = = P P ((k k - - 11)) {X x}_{k k} [[{X x}_{k k}^{T T} P P ((k k - - 11)) {X x}_{k k} + + γ γ {]]}^{- - 11}$

其中，K和P为辨识中的两个矩阵，

γ为遗忘因子，

为单位矩阵。Among them, K and P are two matrices in identification,

γ is the forgetting factor,

is the identity matrix.

(2)采用典型的响应曲线法设计炉膛压力过程模型的比例积分微分控制器，具体方法是：(2) Design the proportional integral differential controller of the furnace pressure process model by using the typical response curve method, the specific method is:

a.将过程模型的比例积分微分控制器停留在手动操作状态，操作拨盘使其输出有阶跃变化，由记录仪表记录过程模型的输出值，将过程模型输出值yL(k)的响应曲线转换成无量纲形式y_L ^*(k)，具体是：

a. Keep the proportional integral differential controller of the process model in the manual operation state, operate the dial to make the output have a step change, record the output value of the process model by the recording instrument, and record the response curve of the process model output value yL(k) Converted to the dimensionless form y _L ^* (k), specifically:

其中，y_L(∞)是过程模型的比例积分微分控制器的输出有阶跃变化时的过程模型输出y_L(k)的稳态值。Among them, y _L (∞) is the steady-state value of the process model output y _L (k) when the output of the proportional integral differential controller of the process model has a step change.

b.选取满足

的两个计算点k₁和k₂，，依据下式计算比例积分微分控制器所需要的参数K、T和τ：b. Choose to satisfy

The two calculation points k ₁ and k ₂ , calculate the parameters K, T and τ required by the proportional integral differential controller according to the following formula:

K＝y_L(∞)/qK=y _L (∞)/q

T＝2(k₁-k₂)T=2(k ₁ -k ₂ )

τ＝2k₁-k₂ τ=2k ₁ -k ₂

其中，q为过程模型的比例积分微分控制器输出的阶跃变化幅度。Among them, q is the step change magnitude of the proportional-integral-differential controller output of the process model.

c.计算过程模型的比例积分微分控制器的参数，具体是：c. Calculate the parameters of the proportional integral differential controller of the process model, specifically:

K_c＝1.2T/Kτ _Kc = 1.2T/Kτ

T_i＝2τT _i =2τ

T_d＝0.5τT _d =0.5τ

其中K_c为比例积分微分控制器的比例参数，T_i为比例积分微分控制器的积分参数，T_d分别为比例积分微分控制器的微分参数。Among them, K _c is the proportional parameter of the proportional integral differential controller, T _i is the integral parameter of the proportional integral differential controller, and T _d is the differential parameter of the proportional integral differential controller.

(3)设计预测比例积分比例积分微分控制器，具体步骤是：(3) Designing a predictive proportional-integral proportional-integral-derivative controller, the specific steps are:

d.将过程模型的比例积分微分控制器停留在自动操作状态，操作拨盘使其输入有阶跃变化，由记录仪表记录实时过程的输出，将过程输出值y(k)的响应曲线转换成无量纲形式y^*(k)，具体是：y^*(k)＝y(k)/y(∞)d. Keep the proportional integral differential controller of the process model in the automatic operation state, operate the dial to make the input have a step change, record the output of the real-time process by the recording instrument, and convert the response curve of the process output value y(k) into Dimensionless form y ^* (k), specifically: y ^* (k) = y(k)/y(∞)

其中，y(∞)是过程模型的比例积分微分控制器的输入有阶跃变化时的过程模型输出y(k)的稳态值。Among them, y(∞) is the steady-state value of the process model output y(k) when the input of the proportional integral differential controller of the process model has a step change.

e.选取满足y(k₃)＝0.39，y(k₄)＝0.63的另两个计算点k₃和k₄，依据下式计算预测比例积分比例积分微分控制器所需要的参数K₁，T₁和τ₁：e. Select the other two calculation points k ₃ and k ₄ satisfying y(k ₃ )=0.39, y(k ₄ )=0.63, and calculate the parameter K ₁ required for predicting the proportional integral proportional integral differential controller according to the following formula, T ₁ and τ ₁ :

K₁＝y(∞)/q₁ K ₁ =y(∞)/q ₁

T₁＝2(k₃-k₄)T ₁ =2(k ₃ -k ₄ )

τ₁＝2k₃-k₄ τ ₁ =2k ₃ -k ₄

其中，q₁为过程模型的比例积分微分控制器输入的阶跃变化幅度。Among them, q ₁ is the magnitude of the step change of the proportional integral differential controller input of the process model.

f.将步骤e得到的参数转化为拉普拉斯形式的局部受控传递函数模型：f. Convert the parameters obtained in step e into a locally controlled transfer function model of the Laplace form:

$\frac{y the y ((s the s))}{{q q}_{11} ((s the s))} = = \frac{11}{{λ λ}_{11} s the s + + 11} {e e}^{- - {L L}_{11} s the s}$

其中，s为拉普拉斯变换算子，λ₁为局部受控传递函数模型的时间常数，L₁为局部受控传递函数模型的时滞，y(s)表示当前时刻过程模型的输出值的拉普拉斯变换，q₁(s)表示过程模型的比例积分微分控制器输入的拉普拉斯变换。Among them, s is the Laplace transform operator, λ ₁ is the time constant of the locally controlled transfer function model, L ₁ is the time lag of the locally controlled transfer function model, and y(s) represents the output value of the process model at the current moment The Laplace transform of q ₁ (s) represents the Laplace transform of the input of the proportional-integral-differential controller of the process model.

λ₁＝T₁ λ ₁ = T ₁

L₁＝τ₁ L ₁ =τ ₁

g.依据步骤f计算出的模型参数整定预测比例积分比例积分微分控制器的参数，具体方法是：g. according to the model parameter calculation that step f calculates the parameter of forecasting proportional-integral proportional-integral-derivative controller, concrete method is:

①对该对象设计预测比例积分控制器。选期望的闭环传递函数模型为G_q2(s)① Design a predictive proportional-integral controller for the object. Select the desired closed-loop transfer function model as G _q2 (s)

${G G}_{q q 22} ((s the s)) = = \frac{11}{{λ λ}_{22} s the s + + 11} {e e}^{- - {L L}_{22} s the s}$

λ₂为期望的闭环传递函数模型的时间常数，L₂为期望的闭环传递函数模型的时滞，L₂＝L₁；λ ₂ is the time constant of the desired closed-loop transfer function model, L ₂ is the time lag of the desired closed-loop transfer function model, L ₂ =L ₁ ;

②预测比例积分比例积分微分控制器的传递函数G_c1(s)可由下式来表示②The transfer function G _c1 (s) of the predicted proportional-integral proportional-integral-derivative controller can be expressed by the following formula

${G G}_{c c 11} ((s the s)) = = \frac{{λ λ}_{11} s the s + + 11}{(({λ λ}_{22} s the s + + 11 - - {e e}^{- - {L L}_{22} s the s}))}$

③依据步骤②得到当前的预测比例积分比例积分微分控制器参数值u(s)③According to step ②, get the current predicted proportional-integral-proportional-integral-derivative controller parameter value u(s)

$u u ((s the s)) = = \frac{(({λ λ}_{22} s the s + + 11 - - {e e}^{- - {L L}_{22} s the s}))}{{λ λ}_{11} s the s + + 11} y the y ((s the s))$

本发明提出的一种基于数据驱动的模型选取和预测PI-PID混合控制方法弥补了传统控制的不足，并有效地方便了控制器的设计，保证控制性能的提升，同时满足给定的生产性能指标。A data-driven model selection and prediction PI-PID hybrid control method proposed by the present invention makes up for the shortcomings of traditional control, and effectively facilitates the design of the controller, ensures the improvement of control performance, and meets the given production performance at the same time index.

本发明提出的控制技术可以有效减少理想炉膛压力工艺参数与实际炉膛压力工艺参数之间的误差，进一步弥补了传统控制器的不足，同时保证控制装置操作在最佳状态，使生产过程的炉膛压力工艺参数达到严格控制。The control technology proposed in the present invention can effectively reduce the error between the ideal furnace pressure process parameters and the actual furnace pressure process parameters, further make up for the shortcomings of the traditional controller, and at the same time ensure that the control device operates in the best state, so that the furnace pressure in the production process Process parameters are strictly controlled.

具体实施方式Detailed ways

以循环流化床锅炉系统炉膛压力过程控制为例：Take the furnace pressure process control of circulating fluidized bed boiler system as an example:

这里以该系统炉膛压力回路的控制作为例子加以描述。炉膛压力不仅受到空气流量的影响，同时也受燃料流量，减温水流量和蒸汽流量的影响。调节手段采用进空气量，其余的影响作为不确定因素。Here, the control of the furnace pressure loop of the system is described as an example. Furnace pressure is not only affected by air flow, but also by fuel flow, desuperheating water flow and steam flow. The adjustment method adopts the air intake volume, and the rest of the effects are regarded as uncertain factors.

(1)建立该循环流化床锅炉系统的炉膛压力过程模型。(1) Establish the furnace pressure process model of the circulating fluidized bed boiler system.

通过数据采集装置采集实时过程炉膛压力运行数据，将采集的实时过程炉膛压力运行数据作为数据驱动的样本集合采用最小二乘法推理，建立基于最小二乘法的离散差分方程形式的炉膛压力过程模型。The real-time process furnace pressure operation data is collected by the data acquisition device, and the collected real-time process furnace pressure operation data is used as a data-driven sample set, and the least square method is used to reason, and the furnace pressure process model based on the discrete difference equation of the least square method is established.

其中，系统调用推理机采用最小二乘法进行炉膛压力过程模型参数的辨识，这些参数包括元素Φ中变量的个数和具体数值。Among them, the system calls the inference engine to identify the parameters of the furnace pressure process model by using the least square method, and these parameters include the number and specific values of variables in the element Φ.

其中y(k)是实际炉膛压力测量值，Φ_k ^TX_k是炉膛压力过程模型的输出值。Where y(k) is the measured value of the actual furnace pressure, and Φ _k ^T X _k is the output value of the furnace pressure process model.

这个过程是第一步推理过程。这个第一步推理是初步挖掘实际炉膛压力回路的基本特性。This process is the first step reasoning process. This first step of reasoning is to initially dig out the basic characteristics of the actual furnace pressure circuit.

(2)设计炉膛压力过程模型的比例积分微分控制器，具体方法是典型的响应曲线法。(2) Design the proportional integral differential controller of the furnace pressure process model, the specific method is the typical response curve method.

第一步：将炉膛压力比例积分微分控制器停留在“手动操作”状态，操作进空气量的拨盘使进空气量控制器输出有个阶跃变化，由记录仪表记录炉膛压力过程模型的输出值，将炉膛压力过程模型输出值y_L(k)的响应曲线转换成无量纲形式y_L ^*(k)：Step 1: Keep the furnace pressure proportional integral differential controller in the "manual operation" state, operate the dial of the intake air volume to make the output of the intake air volume controller have a step change, and record the output of the furnace pressure process model by the recording instrument value, transform the response curve of the furnace pressure process model output value y _L (k) into the dimensionless form y _L ^* (k):

${y the y}_{L L}^{* *} ((k k)) = = {y the y}_{L L} ((k k)) / / {y the y}_{L L} ((\infty \infty))$

其中，y_L(∞)是炉膛压力过程模型输出y_L(k)的稳态值。Among them, y _L (∞) is the steady-state value of the furnace pressure process model output y _L (k).

第二步：选取2个计算点，

依据以下计算公式计算炉膛压力比例积分微分控制器所需要的参数T和τ：Step 2: Select 2 calculation points,

Calculate the parameters T and τ required by the furnace pressure proportional integral differential controller according to the following calculation formula:

K＝y_L(∞)/qK=y _L (∞)/q

T＝2(k₁-k₂)T=2(k ₁ -k ₂ )

τ＝2k₁-k₂ τ=2k ₁ -k ₂

其中，q为炉膛压力比例积分微分控制器输出的阶跃变化幅度。Among them, q is the step change amplitude of the furnace pressure proportional integral differential controller output.

第三步：依据第二步计算出的K，T和τ整定炉膛压力比例积分微分控制器的参数：The third step: according to the K, T and τ calculated in the second step, the parameters of the furnace pressure proportional integral differential controller are adjusted:

K_c＝1.2T/Kτ _Kc = 1.2T/Kτ

T_i＝2τT _i =2τ

T_d＝0.5τT _d =0.5τ

其中K_c，T_i，T_d分别为比例积分微分控制器的比例参数，积分参数，微分参数。Among them, K _c , T _i , and T _d are proportional parameters, integral parameters, and differential parameters of the proportional-integral-derivative controller, respectively.

(3)设计炉膛压力过程的预测PI-PID控制器，具体方法是：(3) Design the predictive PI-PID controller of the furnace pressure process, the specific method is:

针对设计的炉膛压力比例积分微分控制器和过程模型组成的基本控制回路建立该锅炉炉膛压力实时运行过程数据库，通过数据采集装置采集炉膛压力实时过程运行数据，依据炉膛压力实时过程运行数据建立预测PI-PID控制所需的预测模型，基于该预测模型设计相应的炉膛压力实时过程预测PI-PID控制器，具体步骤是：Aiming at the basic control loop composed of the designed furnace pressure proportional-integral-derivative controller and the process model, the real-time operation process database of the boiler furnace pressure is established, and the real-time process operation data of the furnace pressure is collected through the data acquisition device, and the prediction PI is established according to the real-time process operation data of the furnace pressure - The prediction model required for PID control, based on the prediction model, the corresponding furnace pressure real-time process prediction PI-PID controller is designed, and the specific steps are:

第一步：将炉膛压力比例积分微分控制器停留在“自动操作”状态，操作炉膛压力比例积分微分控制器的输入使炉膛压力比例积分微分控制器的输入有个阶跃变化，由记录仪表记录炉膛压力实时过程的输出，将炉膛压力实时过程输出值y(k)的响应曲线转换成无量纲形式y^*(k)：Step 1: Keep the furnace pressure proportional integral differential controller in the "automatic operation" state, operate the input of the furnace pressure proportional integral differential controller to make the input of the furnace pressure proportional integral differential controller have a step change, which is recorded by the recording instrument The output of the furnace pressure real-time process, the response curve of the furnace pressure real-time process output value y(k) is converted into a dimensionless form y ^* (k):

y^*(k)＝y(k)/y(∞)y ^* (k) = y(k)/y(∞)

其中，y(∞)是炉膛压力实时过程输出y(k)的稳态值。Among them, y(∞) is the steady-state value of the furnace pressure real-time process output y(k).

第二步：选取2个计算点，y(k₃)＝0.39，y(k₄)＝0.63，依据以下计算公式计算炉膛压力预测PI-PID控制器所需要的参数K₁，T₁和τ₁：Step 2: Select two calculation points, y(k ₃ )=0.39, y(k ₄ )=0.63, calculate the parameters K ₁ , T ₁ and τ required by the furnace pressure prediction PI-PID controller according to the following calculation formula ₁ :

K₁＝y(∞)/q₁ K ₁ =y(∞)/q ₁

T₁＝2(k₃-k₄)T ₁ =2(k ₃ -k ₄ )

τ₁＝2k₃-k₄ τ ₁ =2k ₃ -k ₄

其中，q₁为炉膛压力比例积分微分控制器输入的阶跃变化幅度。Among them, q ₁ is the step change amplitude of the furnace pressure proportional integral differential controller input.

第三步：将第二步得到的参数转化为拉普拉斯形式的局部受控传递函数模型：Step 3: Transform the parameters obtained in the second step into a locally controlled transfer function model in Laplace form:

其中，y(s)表示当前时刻炉膛压力过程模型输出值的拉普拉斯变换，q₁(s)表示炉膛压力过程模型的比例积分微分控制器输入的拉普拉斯变换。Among them, y(s) represents the Laplace transform of the output value of the furnace pressure process model at the current moment, and q ₁ (s) represents the Laplace transform of the proportional-integral-derivative controller input of the furnace pressure process model.

λ₁＝T₁ λ ₁ =T ₁

L₁＝τ₁ L ₁ =τ ₁

第四步：依据第三步计算出的模型参数整定炉膛压力预测PI-PID控制器的参数，具体方法是：The fourth step: according to the model parameters calculated in the third step, set the parameters of the furnace pressure prediction PI-PID controller, the specific method is:

Claims

1. A hybrid control method for a coal-fired boiler furnace pressure system, characterized in that the method comprises the following steps:

(1) Use the real-time process data of coal-fired boiler furnace pressure to establish a process model, the specific method is:

Firstly, establish the real-time operation database of coal-fired boiler furnace pressure, collect N groups of real-time process operation data through the data acquisition device, and use the collected real-time process operation data as a data-driven sample set, expressed as {x _i , y(i)} _{i= 1} ^N , i=1, 2, ..., N, where x _i represents the input data of the i-th group of process parameters, and y(i) represents the output value of the i-th group of process parameters;

Then, based on the real-time process operation data set of the furnace pressure, a locally controlled autoregressive moving average model in the form of a discrete difference equation based on the least square method is established:

{y the y}_{L L} ((k k)) = = {Φ Φ}^{T T} X x,, Φ Φ = = {[[{a a}_{11}^{' '},, {a a}_{22}^{' '},, . . . . . .,, {a a}_{n no}^{' '},, {b b}_{00}^{' '},, {b b}_{11}^{' '},, . . . . . .,, {b b}_{m m - - 11}^{' '}]]}^{T T}

X=[y(k-1),...,y(kn), u(kd-1),...,u(kdm)] ^T

Among them, y _L (k) represents the output value of the process parameters of the process model at the current moment, x represents the set of input and output data of the process parameters of the process model at the past time, and u(k) represents the control variable corresponding to the process parameters of the current process model , k is the current recursion steps, Φ represents the set of model parameters obtained through identification, T represents the transposition of the matrix, n, m, d+1 are the output variable order and input variable order corresponding to the actual process , the time lag of the actual process;

The identification methods used are:

{Φ Φ}_{k k} = = {Φ Φ}_{k k - - 11} + + \overset{&OverBar; &OverBar;}{K K} ((k k)) [[y the y ((k k)) - - {Φ Φ}_{k k}^{T T} {X x}_{k k}]]

\overset{&OverBar; &OverBar;}{K K} ((k k)) = = P P ((k k - - 11)) {X x}_{k k} {[[{X x}_{k k}^{T T} P P ((k k - - 11)) {X x}_{k k} + + γ γ]]}^{- - 11}

where k and P are two matrices in identification, γ is the forgetting factor, is the identity matrix;

(2) Design the proportional integral differential controller of the furnace pressure process model by using the typical response curve method, the specific method is:

a. Keep the proportional integral differential controller of the process model in the manual operation state, operate the dial to make the output have a step change, record the output value of the process model by the recording instrument, and record the response of the process model output value y _L (k) The curve is transformed into a dimensionless form y _L ^* (k), specifically:

Among them, y _L (∞) is the steady-state value of the process model output y _L (k) when the output of the proportional integral differential controller of the process model has a step change;

b. Choose to satisfy

K=y _L (∞)/q

T=2(k ₁ -k ₂ )

τ=2k ₁ -k ₂

Where q is the step change amplitude of the proportional integral differential controller output of the process model;

c. Calculate the parameters of the proportional integral differential controller of the process model, specifically:

_Kc = 1.2T/Kτ

T _i =2τ

T _d =0.5τ

Where K _c is the proportional parameter of the proportional integral differential controller, T _i is the integral parameter of the proportional integral differential controller, T _d is the differential parameter of the proportional integral differential controller respectively;

(3) Designing a predictive proportional-integral proportional-integral-derivative controller, the specific steps are:

d. Keep the proportional integral differential controller of the process model in the automatic operation state, operate the dial to make the input have a step change, record the output of the real-time process by the recording instrument, and convert the response curve of the process output value y(k) into Dimensionless form y ^* (k), specifically: y ^* (k) = y(k)/y(∞)

Among them, y(∞) is the steady-state value of the process model output y(k) when the input of the proportional integral differential controller of the process model has a step change;

e. Select the other two calculation points k ₃ and k ₄ satisfying y(k ₃ )=0.39, y(k ₄ )=0.63, and calculate the parameter K ₁ required for predicting the proportional integral proportional integral differential controller according to the following formula, T ₁ and τ ₁ :

K ₁ =y(∞)/q ₁

T ₁ =2(k ₃ -k ₄ )

τ ₁ =2k ₃ -k ₄

Where _q1 is the step change amplitude of the proportional integral differential controller input of the process model;

f. Convert the parameters obtained in step e into a locally controlled transfer function model of the Laplace form:

\frac{y the y ((s the s))}{{q q}_{11} ((s the s))} = = \frac{11}{{λ λ}_{11} s the s + + 11} {e e}^{- - {L L}_{11} s the s}

Where s is the Laplace transform operator, λ ₁ is the time constant of the locally controlled transfer function model, L ₁ is the time lag of the locally controlled transfer function model, y(s) represents the output value of the process model at the current moment Laplace transform, q ₁ (s) represents the Laplace transform of the proportional integral differential controller input of the process model;

λ ₁ =T ₁

L ₁ =τ ₁

g. according to the model parameter calculation that step f calculates the parameter of forecasting proportional-integral proportional-integral-derivative controller, concrete method is:

① Design a predictive proportional-integral controller for the object; select the desired closed-loop transfer function model as G _q2 (s)

{G G}_{q q 22} ((s the s)) = = \frac{11}{{λ λ}_{22} s the s + + 11} {e e}^{- - {L L}_{22} s the s}

λ ₂ is the time constant of the desired closed-loop transfer function model, L ₂ is the time lag of the desired closed-loop transfer function model, L ₂ =L ₁ ;

②The transfer function G _c1 (s) of the predicted proportional-integral proportional-integral-derivative controller can be expressed by the following formula

{G G}_{c c 11} ((s the s)) \frac{{λ λ}_{11} s the s + + 11}{(({λ λ}_{22} s the s + + 11 - - {e e}^{- - {L L}_{22} s the s}))}

③According to step ②, the current predictive proportional-integral-proportional-integral-derivative controller parameter value u(s) is obtained.

u u ((s the s)) = = \frac{{λ λ}_{22} s the s + + 11 - - {e e}^{- - {L L}_{22} s the s}}{{λ λ}_{11} s the s + + 11} y the y ((s the s))