CN105240846B - The Process of Circulating Fluidized Bed Boiler control method of multivariable GPC optimization - Google Patents

Abstract

A kind of Process of Circulating Fluidized Bed Boiler control method of multivariable GPC optimization of the present invention, including step：1st, loop initialization fluidized-bed combustion boiler combustion controller design parameter；2nd, the process variable data of the Process of Circulating Fluidized Bed Boiler according to collection, combustion process model is set up by multivariable gradual reducing memory recursion least squares identification；3rd, according to multivariable GPC characteristic to PID₁、PID₂、PID₃Controller parameter is optimized to be adjusted, and the result decoupling that will adjust；4th, the controller parameter according to obtained by, controls coal-supplying amount, primary and secondary air quantity, and then control bed temperature, main steam pressure and furnace outlet flue gas oxygen content respectively；5th, into next moment, step 2 to step 5 is repeated.Control accuracy of the present invention is higher, and tracking velocity is very fast, and non-overshoot, steady-state error is small, and anti-coupling ability is strong, and control process is smoothed, and is that a kind of form is simple, realizes convenient Process of Circulating Fluidized Bed Boiler multivariable control techniques.

Description

Multivariable generalized predictive control optimized circulating fluidized bed boiler combustion process control method

Technical Field

The invention belongs to the technical field of boiler combustion process control, and relates to a multivariable proportional-integral-derivative (PID) control method for a circulating fluidized bed boiler combustion process optimized through multivariable generalized predictive control, in particular to a circulating fluidized bed boiler combustion process control method optimized through multivariable generalized predictive control.

Background

Circulating Fluidized Bed Boilers (CFBB) have the advantages of wide fuel adaptability, low-pollution combustion, strong load regulation capacity and the like, and are increasingly widely applied to industries such as power generation, heat supply, steel making, chemical industry and the like in recent years. The combustion process is a main system of the circulating fluidized bed boiler, and the control quality of the combustion process directly influences the safety and the economical efficiency of the operation of the circulating fluidized bed boiler.

The existing control method for the combustion process of the circulating fluidized bed boiler basically adopts the traditional multivariable proportional-integral-derivative (PID) control, the PID control has the advantages of convenience in implementation, low model dependency, simple principle and the like, but the combustion process has the complex characteristics of strong coupling, multivariable, slow time variation, large hysteresis and the like, the control quality of the traditional multivariable PID is not high, the problems of low boiler combustion efficiency, serious abrasion, high auxiliary power consumption and the like are caused, and the requirements of energy conservation, emission reduction and environmental protection are difficult to adapt.

Generalized Predictive Control (GPC) is an advanced computer control technique that is suitable for industrial processes where it is difficult to build accurate mathematical models and where the dynamic process is complex. The circulating fluidized bed boiler combustion process is a typical multivariable, strongly coupled, nonlinear, dynamic process complex control system, and an accurate mathematical model is difficult to establish. Chinese patent 'a multivariable control method for combustion process of circulating fluidized bed boiler', application No. CN201110422623.4, discloses a multivariable generalized predictive control method for combustion process of circulating fluidized bed boiler, which can theoretically improve the control quality of combustion process, but because the control method is established on the basis of description and solution of constrained optimization problem, the cost of environment and training and maintenance of computer is higher, and the multivariable generalized predictive control method is not as simple in form, easy to understand and master, difficult to implement and difficult to be practically applied to the control of combustion process of circulating fluidized bed boiler.

Disclosure of Invention

The invention aims to provide a circulating fluidized bed boiler combustion process control method optimized by multivariable generalized predictive control aiming at the application defects of the multivariable generalized predictive control method.

In order to solve the technical problems in the prior art, the invention adopts the following technical scheme:

a multivariable PID control method for a combustion process of a circulating fluidized bed boiler optimized by multivariable generalized predictive control comprises the following steps of 1, initializing design parameters of a combustion process controller of the circulating fluidized bed boiler: bed temperature, main steam pressure and furnace outlet flue gas oxygen content prediction step number P₁、P₂、P₃Controlling the weighting coefficient r according to the bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth₁、r₂、r₃Bed temperature, main steam pressure and furnace outlet flue gas oxygen content reference trajectory coefficient β₁、β₂、β₃Identification of initial values by multivariate fading memory recursive least squaresP (0), forgetting factor μ;

characterized in that the method further comprises the following steps:

step 2, identifying and establishing a combustion process model through a multivariate fading memory recursive least square method according to the collected process variable data of the combustion process of the circulating fluidized bed boiler;

step 3, PID is controlled according to the characteristics of multivariable generalized predictive control₁、PID₂、PID₃Optimizing and setting the parameters of the controller, and decoupling the setting result;

step 4, obtaining PID according to step 3₁、PID₂、PID₃Controller parameters, respectivelyThe control quantity is formed to control the coal feeding quantity, the primary air quantity and the secondary air quantity, and further control the bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth;

and 5, entering the next moment, returning to the step 2, and repeating the process from the step 2 to the step 5.

The specific process of the step 3 is as follows:

3-a, calculating a backward shift operator q according to the equation set of the loss map^-1Polynomial ofi＝1,2,3；

3-b, establishing a predicted value y of the bed temperature at the moment k + j of the combustion process of the circulating fluidized bed boiler_1pPredicted values y of main steam pressure at moments (k + j) and (k + j)_2pPredicted value y of oxygen content of flue gas at furnace outlet at (k + j) and k + j moments_3p(k+j)；

3-c, establishing a multi-step predicted value Y of the bed temperature₁Main steam pressure multi-step prediction value Y₂And the multi-step predicted value Y of the oxygen content of the flue gas at the outlet of the hearth₃；

3-d, establishing an objective function J of bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth₁(k)、J₂(k)、J₃(k) Control step number M of bed temperature, main steam pressure and oxygen content of hearth outlet flue gas₁、M₂、M₃Taking the value as 1;

3-e, increasing the coal feeding amount by delta u₁(k) Primary air volume increment delta u₂(k) Secondary air quantity increment delta u₃(k) Form conversion, substitution into the objective function J₁(k)、J₂(k)、J₃(k) The optimization is carried out, and the decoupling processing is carried out on the optimization result to obtain a formula:

wherein,

f₁＝F¹¹y₁(k)+H¹¹Δu₁(k-1)+H¹²Δu₂(k-1)+H¹³Δu₃(k-1)

f₂＝F²²y₂(k)+H²¹Δu₁(k-1)+H²²Δu₂(k-1)+H²³Δu₃(k-1)

f₃＝F³³y₃(k)+H³¹Δu₁(k-1)+H³²Δu₂(k-1)+H³³Δu₃(k-1)

Y_1r＝[y_1r(k+1),y_1r(k+2),…,y_1r(k+P₁)]^T

y_1r(k+j)＝β₁ ^jy_1r(k+j-1)+(1-β₁ ^j)y_1s(k),j＝1,…,P₁

Y_2r＝[y_2r(k+1),y_2r(k+2),…,y_2r(k+P₂)]^T

y_2r(k+j)＝β₂ ^jy_2r(k+j-1)+(1-β₂ ^j)y_2s(k),j＝1,…,P₂

Y_3r＝[y_3r(k+1),y_3r(k+2),…,y_3r(k+P₃)]^T

y_3r(k+j)＝β₃ ^jy_3r(k+j-1)+(1-β₃ ^j)y_3s(k),j＝1,…,P₃

e₁(k)、e₁(k-1)、e₁(k-2) are errors between the bed temperature reference trajectory value and the actual value at the time k, the time k-1 and the time k-2 respectively;

e₂(k)、e₂(k-1)、e₂(k-2) errors between the reference trajectory values and actual values of the main steam pressure at the time k, the time k-1 and the time k-2, respectively;

e₃(k)、e₃(k-1)、e₃(k-2) errors between the reference track value and the actual value of the oxygen content of the flue gas at the outlet of the hearth at the time k, the time k-1 and the time k-2 respectively; r is₁,r₂,r₃To control the weighting coefficients;is the prediction error of the bed temperature at the moment k,is the prediction error of the main steam pressure at time k,predicting error of oxygen content of the flue gas at the outlet of the hearth at the moment k; h is₁、h₂、h₃Predicting error correction matrixes for bed temperature, main steam pressure and oxygen content of flue gas at a hearth outlet; f. of₁、f₂、f₃Is at bed temperature,Main steam pressure and hearth outlet flue gas oxygen content free movement term; p₁、P₂、P₃Predicting the number of steps for bed temperature, main steam pressure and oxygen content in the flue gas at the outlet of the furnace, P₁、P₂、P₃The values are different; m₁、M₂、M₃Controlling the number of steps for bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; r is₁、r₂、r₃The weighting coefficients are used for controlling bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; y is_1r、Y_2r、Y_3rThe matrix is composed of reference track values at different moments of bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; y is_1r(k+j)、y_2r(k+j)、y_3r(k + j) is the value of bed temperature, main steam pressure and reference track of oxygen content of the flue gas at the outlet of the furnace at the moment of k + j β₁、β₂、β₃Reference trajectory coefficients are provided for bed temperature, main steam pressure and oxygen content of flue gas at a hearth outlet; y is_1s(k)、y_2s(k)、y_3s(k) Respectively setting values of bed temperature, main steam pressure and oxygen content of flue gas at the outlet of the hearth at the moment k; y is₁(k) Represents the actual output of the bed temperature at time k, y₂(k) Representing the actual output of the main steam pressure at time k, y₃(k) Representing the actual output of the oxygen content of the flue gas at the outlet of the hearth at the moment k; i is₁，I₂，I₃Is an identity matrix; Δ u₁(k-1)、Δu₂(k-1)、Δu₃And (k-1) respectively represents the coal feeding quantity increment, the primary air quantity increment and the secondary air quantity increment at the k-1 moment.

3-f, calculating the PID of the k time by using the formula of the step 3-e₁、PID₂、PID₃Parameters of the controller:

wherein, V₁(k)＝[ν₁₀(k),ν₁₁(k),ν₁₂(k)]^T，V₂(k)＝[ν₂₀(k),ν₂₁(k),ν₂₂(k)]^T，V₃(k)＝[ν₃₀(k),ν₃₁(k),ν₃₂(k)]^T；k_p1(k)、k_i1(k)、k_d1(k) PID of coal supply at time k₁Proportional, integral, differential coefficients of the controller; k is a radical of_p2(k)、k_i2(k)、k_d2(k) PID of primary air volume at time k₂Proportional, integral, differential coefficients of the controller; k is a radical of_p3(k)、k_i3(k)、k_d3(k) PID of secondary air volume at time k₃Proportional, integral, and differential coefficients of the controller.

In step 4, the coal supply amount, the primary air volume and the secondary air volume at time k are:

u₁(k)＝u₁(k-1)+k_p1(k)[e₁(k)-e₁(k-1)]+k_i1(k)e₁(k)+k_d1(k)[e₁(k)-2e₁(k-1)+e₁(k-2)]

u₂(k)＝u₂(k-1)+k_p2(k)[e₂(k)-e₂(k-1)]+k_i2(k)e₂(k)+k_d2(k)[e₂(k)-2e₂(k-1)+e₂(k-2)]

u₃(k)＝u₃(k-1)+k_p3(k)[e₃(k)-e₃(k-1)]+k_i3(k)e₃(k)+k_d3(k)[e₃(k)-2e₃(k-1)+e₃(k-2)]

wherein u is₁(k)、u₂(k)、u₃(k) Respectively shows the coal feeding quantity, the primary air quantity and the secondary air quantity at the moment k, u₁(k-1)、u₂(k-1)、u₃(k-1) represents the coal supply amount, the primary air amount and the secondary air amount at the time of k-1, respectively.

In the step 3, the control weighting coefficients r of the bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth₁、r₂、r₃The value range of (A) is 0.001-0.1.

Compared with the prior art, the invention has the following advantages and beneficial effects:

1. the method combines multivariable generalized predictive control and multivariable PID control technology, so that the multivariable PID control has the excellent control quality and the predictive function of the multivariable generalized predictive control, the control quality of the traditional multivariable PID control is effectively improved, and the defect of the actual application of the multivariable generalized predictive control is overcome.

2. The simulation control test of the combustion process of the circulating fluidized bed boiler shows that the method has the advantages of high control precision, high tracking speed, no overshoot, small steady-state error, strong coupling resistance and smooth control process, and is a multivariable control technology with high control quality, simple form and convenient realization of the combustion process of the circulating fluidized bed boiler.

Drawings

FIG. 1 is a schematic view of a circulating fluidized bed boiler combustion process control system for carrying out the method of the present invention;

FIG. 2 is a flow chart of the method of the present invention;

FIG. 3 is a simulation result fitting curve of the controlled bed temperature in the combustion process of the circulating fluidized bed boiler controlled by the method of the present invention;

FIG. 4 is a controlled quantity bed temperature tracking error fitting curve of the combustion process of the circulating fluidized bed boiler under the control of the method of the present invention;

FIG. 5 is a fitting curve of simulation results of the amount of coal supplied by the combustion process of the circulating fluidized bed boiler controlled by the method of the present invention;

FIG. 6 is a fitting curve of simulation results of the main steam pressure of the controlled quantity in the combustion process of the circulating fluidized bed boiler under the control of the method of the present invention;

FIG. 7 is a controlled main steam pressure tracking error fitting curve of the circulating fluidized bed boiler in the combustion process under the control of the method of the present invention;

FIG. 8 is a fitting curve of simulation results of the primary air quantity of the combustion process control quantity of the circulating fluidized bed boiler under the control of the method of the present invention;

FIG. 9 is a simulation result fitting curve of the oxygen content of the flue gas at the outlet of the controlled-quantity furnace of the circulating fluidized bed boiler in the combustion process under the control of the method of the present invention;

FIG. 10 is a graph of error-fitted tracking of oxygen content in flue gas at the outlet of a controlled quantity furnace of a circulating fluidized bed boiler during combustion under the control of the method of the present invention;

FIG. 11 is a fitting curve of simulation results of secondary air quantity of the combustion process control quantity of the circulating fluidized bed boiler under the control of the method of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings.

The combustion process of the circulating fluidized bed boiler has the complex characteristics of strong coupling, multivariable, slow time change, large hysteresis and the like, and the controlled quantity is as follows: bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth, and the control quantity is as follows: coal feeding amount, primary air quantity and secondary air quantity.

FIG. 1 is a schematic view of a circulating fluidized bed boiler combustion process control system for carrying out the method of the present invention. The system consists of a microcomputer, an A/D converter, a D/A converter, a sensor, an actuator and the like, wherein control software written according to the method is solidified in the microcomputer. The control system carries out timing sampling on analog parameters of bed temperature, main steam pressure and hearth outlet flue gas oxygen content in the combustion process of the circulating fluidized bed boiler through a sensor group, the analog parameters are converted into digital signals through an A/D converter, a microcomputer carries out operation according to a multivariable generalized predictive control optimized control method of the circulating fluidized bed boiler combustion system, the coal feeding quantity, the primary air quantity and the secondary air quantity which need to be adjusted are calculated, the analog signals are converted into analog signals through the D/A converter, and an actuator group is directly controlled to control the bed temperature, the main steam pressure and the hearth outlet flue gas oxygen content of the combustion system, so that the whole circulating fluidized bed boiler combustion process control system is formed.

FIG. 2 is a flow chart of the method of the present invention, and as shown in FIG. 2, the multivariable PID control method of the combustion process of the circulating fluidized bed boiler optimized by multivariable generalized predictive control of the present invention comprises the following steps:

step 1, initializing design parameters of a combustion process controller of the circulating fluidized bed boiler: bed temperature, main steam pressure and furnace outlet flue gas oxygen content prediction step number P₁、P₂、P₃Controlling the weighting coefficient r according to the bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth₁、r₂、r₃Bed temperature, main steam pressure and furnace outlet flue gas oxygen content reference trajectory coefficient β₁、β₂、β₃Identification of initial values by multivariate fading memory recursive least squaresP (0), forgetting factor μ;

step 2, according to the collected process variable data of the combustion process of the circulating fluidized bed boiler, identifying and establishing a model of the combustion process through a multivariate fading memory recursive least square method, and specifically comprising the following steps of:

2-a, collecting process variables of the combustion process of the circulating fluidized bed boiler at each moment in a control period, wherein the process variables specifically comprise control quantity: coal feeding amount, primary air volume and secondary air volume, controlled amount: bed temperature, main steam pressure and hearth outlet flue gas oxygen content;

2-b, establishing a matrix H (k) according to the process variable data obtained in the step 2-a, wherein the matrix H (k) is in the form of:

Δu₂(k-1),Δu₂(k-2),…,Δu₂(k-d₁₂),Δu₃(k-1),Δu₃(k-2),…,Δu₃(k-d₁₃)]

Δu₂(k-1),Δu₂(k-2),…,Δu₂(k-d₂₂),Δu₃(k-1),Δu₃(k-2),…,Δu₃(k-d₂₃)]

Δu₂(k-1),Δu₂(k-2),…,Δu₂(k-d₃₂),Δu₃(k-1),Δu₃(k-2),…,Δu₃(k-d₃₃)]

wherein T is a matrix transposition symbol; c. C₁₁、、c₂₂、c₃₃Is the output order, d₁₁、d₁₂、d₁₃、d₂₁、d₂₂、d₂₃、d₃₁、d₃₂、d₃₃The method comprises the steps of inputting order, and determining according to collected process variable data of the combustion process of the circulating fluidized bed boiler; y is₁(k-1) is the value at the time of bed temperature k-1, y₁(k-2) is the value at the time of bed temperature k-2, y₁(k-c₁₁) Is the bed temperature k-c₁₁A value of a time of day; y is₂(k-1) is the value at the moment of the main steam pressure k-1, y₂(k-2) is the value at the moment of the main steam pressure k-2, y₂(k-c₂₂) Is the main steam pressure k-c₂₂A value of a time of day; y is₃(k-1) is the value at the moment when the oxygen content of the flue gas at the outlet of the furnace is k-1, y₃(k-2) is the value at the moment when the oxygen content of the flue gas at the outlet of the furnace is k-2, y₃(k-c₃₃) Is the oxygen content of the flue gas at the outlet of the hearthk-c₃₃A value of a time of day; Δ u₁(k-1) is the increment of the coal supply at the time k-1, Δ u₁(k-2) is the increment of the coal feed amount at the time of k-2, Δ u₁(k-d₁₁) The coal feeding amount is in k-d₁₁An increment of time; Δ u₂(k-1) is the increment of the primary air volume at the time k-1, Δ u₂(k-2) is the increment of the primary air volume at the time k-2, Δ u₂(k-d₂₂) The primary air quantity is in k-d₂₂An increment of time; Δ u₃(k-1) is the increment of the secondary air volume at the time k-1, Δ u₃(k-2) is the increment of the secondary air volume at the time k-2, Δ u₃(k-d₃₃) The secondary air quantity is in k-d₃₃The increment of the time of day.

2-c, establishing a model identification matrix by using a multivariate fading memory recursive least square methodThe specific method comprises the following steps:

whereinMu is forgetting factor, K (k) is weight factor, P (k) is positive definite covariance matrix, and initial value P (0) of P (k) is α²I, α is a positive number, I is an identity matrix;initial valueThe value is 0; y (k) ═ y₁(k),y₂(k),y₃(k)]^T，y₁(k)、y₂(k) And y₃(k) The bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth at the moment k are respectively.

2-d, using the model identification matrix obtained in step 2-cEstablishing a combustion process system matrix A (q)^-1) And B (q)^-1) The form is as follows:

wherein,

q^-1is the postshift operator, A, B is the postshift operator q^-1The polynomial of (a) is determined,a, B respectively^-1C, d are the order of the system, determined from the collected process variable data of the combustion process of the circulating fluidized bed boiler, A (q)^-1) And B (q)^-1) Is a system matrix.

2-e, establishing a model of the combustion process, wherein the form is as follows:

A(q^-1)y(k)＝B(q^-1)u(k-1)+ξ(k)/Δ

wherein u (k-1) ═ u₁(k-1),u₂(k-1),u₃(k-1)]^T，u₁(k-1)、u₂(k-1)、u₃(k-1) represents the coal supply amount, the primary air flow rate and the secondary air flow rate at the time of k-1, and ξ (k) [ ξ ]₁(k),ξ₂(k),ξ₃(k)]^T，ξ₁(k)、ξ₂(k)、ξ₃(k) White noise at time k; Δ is a difference operator, Δ ═ 1-q^-1。

Step 3, PID is controlled according to the characteristics of multivariable generalized predictive control₁、PID₂、PID₃The method comprises the following steps of optimizing and setting parameters of a controller, and decoupling a setting result, and specifically comprises the following steps:

3-a, calculating a backward shift operator q according to the equation set of the loss map^-1Polynomial of(i ═ 1,2,3), the specific method is:

wherein N is_AIs matrix A (q)^-1) Order of (1), N_BIs the matrix B (q)^-1) Order of (1), P₁,P₂,P₃To predict the number of steps.

3-b, establishing a predicted value y of the bed temperature at the moment k + j in the CFBB combustion process_1pPredicted values y of main steam pressure at moments (k + j) and (k + j)_2pPredicted value y of oxygen content of flue gas at furnace outlet at (k + j) and k + j moments_3p(k+j)：

Wherein,

is the initial predicted value of the bed temperature at the moment k + j in the combustion process,is the initial predicted value of the main steam pressure at the moment k + j,is an initial predicted value of the oxygen content of the flue gas at the outlet of the hearth at the moment k + j,is the prediction error of the bed temperature at the moment k,is the prediction error of the main steam pressure at time k,is the prediction error of the oxygen content of the flue gas at the outlet of the hearth at the moment k, y₁(k) Represents the actual output of the bed temperature at time k, y₂(k) Representing the actual output of the main steam pressure at time k, y₃(k) And the actual output of the oxygen content of the flue gas at the outlet of the hearth at the moment k is shown.

3-c, establishing a multi-step predicted value Y of the bed temperature₁Main steam pressure multi-step prediction value Y₂And the multi-step predicted value Y of the oxygen content of the flue gas at the outlet of the hearth₃：

Wherein,

f₁＝F¹¹y₁(k)+H¹¹Δu₁(k-1)+H¹²Δu₂(k-1)+H¹³Δu₃(k-1)

f₂＝F²²y₂(k)+H²¹Δu₁(k-1)+H²²Δu₂(k-1)+H²³Δu₃(k-1)

f₃＝F³³y₃(k)+H³¹Δu₁(k-1)+H³²Δu₂(k-1)+H³³Δu₃(k-1)

Y₁＝[y_1p(k+1),y_1p(k+2),…,y_1p(k+P₁)]^T

Y₂＝[y_2p(k+1),y_2p(k+2),…,y_2p(k+P₂)]^T

Y₃＝[y_3p(k+1),y_3p(k+2),…,y_3p(k+P₃)]^T

h₁、h₂、h₃predicting error correction matrixes for bed temperature, main steam pressure and oxygen content of flue gas at a hearth outlet; f. of₁、f₂、f₃Is the free movement term of bed temperature, main steam pressure and the oxygen content of the smoke at the outlet of the hearth.

3-d, establishing an objective function J of bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth₁(k)、J₂(k)、J₃(k) Control step number M of bed temperature, main steam pressure and oxygen content of hearth outlet flue gas₁、M₂、M₃All take 1:

wherein,

Y_1r＝[y_1r(k+1),y_1r(k+2),…,y_1r(k+P₁)]^T

Y_2r＝[y_2r(k+1),y_2r(k+2),…,y_2r(k+P₂)]^T

Y_3r＝[y_3r(k+1),y_3r(k+2),…,y_3r(k+P₃)]^T

y_1r(k+j)＝β₁ ^jy_1r(k+j-1)+(1-β₁ ^j)y_1s(k)

y_2r(k+j)＝β₂ ^jy_2r(k+j-1)+(1-β₂ ^j)y_2s(k)

y_3r(k+j)＝β₃ ^jy_3r(k+j-1)+(1-β₃ ^j)y_3s(k)

β₁、β₂、β₃reference trajectory coefficients are provided for bed temperature, main steam pressure and oxygen content of flue gas at a hearth outlet; p₁、P₂、P₃Predicting the number of steps for bed temperature, main steam pressure and oxygen content in the flue gas at the outlet of the furnace, P₁、P₂、P₃The values are different; m₁、M₂、M₃Controlling the number of steps for bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; r is₁、r₂、r₃The weighting coefficients are used for controlling bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; the set value of the bed temperature at the moment k is y_1s(k) And the main steam pressure set value at the moment k is y_2s(k) And the set value of the oxygen content of the flue gas at the outlet of the hearth at the moment k is y_3s(k)；Y_1r、Y_2r、Y_3rThe reference track of bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; y is_1r(k+j)、y_2r(k+j)、y_3rAnd (k + j) is the value of the bed temperature, the main steam pressure and the reference track of the oxygen content of the flue gas at the outlet of the hearth at the moment of k + j.

3-e, increasing the coal feeding amount by delta u₁(k) Primary air volume increment delta u₂(k) Secondary air quantity increment delta u₃(k) Form conversion, substitution into the objective function J₁(k)、J₂(k)、J₃(k) The optimization is carried out, and the decoupling processing is carried out on the optimization result to obtain a formula：

Wherein,

first, the coal supply quantity increment delta u₁(k) Primary air volume increment delta u₂(k) Secondary air quantity increment delta u₃(k) PID form of (1):

Δu₁(k)＝k_p1(k)[e₁(k)-e₁(k-1)]+k_i1(k)e₁(k)+k_d1(k)[e₁(k)-2e₁(k-1)+e₁(k-2)]

Δu₂(k)＝k_p2(k)[e₂(k)-e₂(k-1)]+k_i2(k)e₂(k)+k_d2(k)[e₂(k)-2e₂(k-1)+e₂(k-2)]

Δu₃(k)＝k_p3(k)[e₃(k)-e₃(k-1)]+k_i3(k)e₃(k)+k_d3(k)[e₃(k)-2e₃(k-1)+e₃(k-2)]

wherein e is₁(k)、e₁(k-1)、e₁(k-2) errors between reference trajectory values and actual values of bed temperature at time k, time k-1 and time k-2, respectively, e₂(k)、e₂(k-1)、e₂When (k-2) is each kThe error between the reference trajectory value and the actual value of the main steam pressure at the moment k-1 and the moment k-2, e₃(k)、e₃(k-1)、e₃(k-2) errors between reference trajectory values and actual values of oxygen content of flue gas at the outlet of the hearth at the time k, the time k-1 and the time k-2, respectively, e₁(k)＝y_1r(k)-y₁(k)，e₂(k)＝y_2r(k)-y₂(k)，e₃(k)＝y_3r(k)-y₃(k)，k_p1(k)、k_i1(k)、k_d1(k) PID of coal supply at time k₁Proportional, integral, differential coefficient, k, of the controller_p2(k)、k_i2(k)、k_d2(k) PID of primary air volume at time k₂Proportional, integral, differential coefficient, k, of the controller_p3(k)、k_i3(k)、k_d3(k) PID of secondary air volume at time k₃Proportional, integral, and differential coefficients of the controller.

Then the coal feeding quantity is increased by delta u₁(k) Primary air volume increment delta u₂(k) Secondary air quantity increment delta u₃(k) Rewriting to matrix form:

wherein,

V₁(k)＝[ν₁₀(k),ν₁₁(k),ν₁₂(k)]^T，V₂(k)＝[ν₂₀(k),ν₂₁(k),ν₂₂(k)]^T

ν₁₀(k)＝k_p1(k)+k_i1(k)+k_d1(k)，ν₁₁(k)＝-k_p1(k)-2k_d1(k)，ν₁₂(k)＝k_d1(k)

ν₂₀(k)＝k_p2(k)+k_i2(k)+k_d2(k)，ν₂₁(k)＝-k_p2(k)-2k_d2(k)，ν₂₂(k)＝k_d2(k)

ν₃₀(k)＝k_p3(k)+k_i3(k)+k_d3(k)，ν₃₁(k)＝-k_p3(k)-2k_d3(k)，ν₃₂(k)＝k_d3(k)

then rewriting Δ u in matrix form₁(k)、Δu₂(k)、Δu₃(k) Substituting the objective function J₁(k)、J₂(k)、J₃(k) Let us order

Further convert V into₁、V₂、V₃Are combined to formA set of ternary primary matrix equations:

wherein,

finally solving a matrix equation set to obtain a formula:

3-f, calculating the PID of the k time by using the formula of the step 3-e₁、PID₂、PID₃Parameters of the controller:

step 4, obtaining PID according to step 3₁、PID₂、PID₃The controller parameters respectively form control quantities for controlling the coal feeding quantity, the primary air quantity and the secondary air quantity, and further controlling the bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth:

u₁(k)＝u₁(k-1)+k_p1(k)[e₁(k)-e₁(k-1)]+k_i1(k)e₁(k)+k_d1(k)[e₁(k)-2e₁(k-1)+e₁(k-2)]

u₂(k)＝u₂(k-1)+k_p2(k)[e₂(k)-e₂(k-1)]+k_i2(k)e₂(k)+k_d2(k)[e₂(k)-2e₂(k-1)+e₂(k-2)]

u₃(k)＝u₃(k-1)+k_p3(k)[e₃(k)-e₃(k-1)]+k_i3(k)e₃(k)+k_d3(k)[e₃(k)-2e₃(k-1)+e₃(k-2)]

wherein u is₁(k)、u₂(k)、u₃(k) Respectively shows the coal feeding quantity, the primary air quantity and the secondary air quantity at the moment k, u₁(k-1)、u₂(k-1)、u₃(k-1) represents the coal supply amount, the primary air amount and the secondary air amount at the time of k-1, respectively.

And 5, entering the next moment and returning to the step 2.

In order to verify the effectiveness of the method, in a Matlab2010 simulation environment, the method is programmed by using an M language, and a simulation test is carried out:

wherein, the bed temperature, the main steam pressure and the prediction step number P of the oxygen content of the flue gas at the outlet of the hearth₁＝11、P₂8 and P₃13, controlling the weighting coefficient r by the bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth₁＝0.001、r₂0.001 and r₃0.001, bed temperature, main steam pressure and furnace outlet flue gas oxygen content reference trajectory coefficient β₁＝0.75、β₂0.75 and β₃＝0.75。

The following model of the combustion process of the circulating fluidized bed boiler is adopted:

fig. 3-11 are fitting curves of the simulation results of the dynamic process of 3 control loops of the CFBB combustion process under the control of the method of the present invention, which are respectively: FIG. 3 is a simulation result fitting curve of the controlled bed temperature in the combustion process of the circulating fluidized bed boiler controlled by the method of the present invention, FIG. 4 is a tracking error fitting curve of the controlled bed temperature in the combustion process of the circulating fluidized bed boiler controlled by the method of the present invention, FIG. 5 is a simulation result fitting curve of the controlled quantity of coal fed in the combustion process of the circulating fluidized bed boiler controlled by the method of the present invention, FIG. 6 is a simulation result fitting curve of the controlled quantity of main steam pressure in the combustion process of the circulating fluidized bed boiler controlled by the method of the present invention, FIG. 7 is a tracking error fitting curve of the controlled quantity of main steam pressure in the combustion process of the circulating fluidized bed boiler controlled by the method of the present invention, FIG. 8 is a simulation result fitting curve of the controlled quantity of primary air volume in the combustion process of the circulating fluidized bed boiler controlled by the method of the present invention, FIG. 9 is, FIG. 10 is a fitting curve of tracking error of oxygen content of flue gas at the outlet of a controlled quantity hearth of a combustion process of a circulating fluidized bed boiler under the control of the method of the invention, and FIG. 11 is a fitting curve of simulation result of secondary air quantity of the combustion process control quantity of the circulating fluidized bed boiler under the control of the method of the invention.

When t is 0s, the bed temperature is set to 900 ℃, the main steam pressure is 3Mpa, and the oxygen content of the flue gas at the outlet of the hearth is 3.5%. When t is 300s, the main steam pressure is increased to 4Mpa, and the bed temperature and the oxygen content of the furnace outlet flue gas are kept unchanged. As can be seen from the observation of FIG. 3, FIG. 6 and FIG. 9, the method of the present invention can reach the set values of bed temperature, main steam pressure and oxygen content of the flue gas at the furnace outlet at a higher speed, and has excellent dynamic performance and no overshoot. When t is 300s, when the main steam pressure is increased to 4Mpa, the coupling influence on the bed temperature and the oxygen content of the flue gas is small, and the decoupling performance is strong. As can be seen from the observation of the graphs in FIG. 5, FIG. 8 and FIG. 11, the steady-state errors of the method for tracking the bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth are smaller, and the tracking controlled quantity precision is higher. As can be seen from fig. 4, 7 and 10, the coal supply amount, the primary air volume and the secondary air volume change relatively smoothly, which is beneficial to the protection of the on-site actuator.

In conclusion, the multivariable generalized predictive control optimized multivariable PID control method for the combustion process of the circulating fluidized bed boiler combines multivariable generalized predictive control and multivariable PID control technology, so that the multivariable PID control has excellent control quality and predictive function of multivariable generalized predictive control, the control quality of the traditional multivariable PID control is effectively improved, and the defects of actual application of the multivariable generalized predictive control are overcome. The simulation control test of the combustion process of the circulating fluidized bed boiler shows that the method has the advantages of high control precision, high tracking speed, no overshoot, small steady-state error, strong coupling resistance and smooth control process, and is a multivariable control technology with high control quality, simple form and convenient realization of the combustion process of the circulating fluidized bed boiler.

Claims (3)

1. A multivariable generalized predictive control optimized circulating fluidized bed boiler combustion process control method comprises the following steps of 1, initializing design parameters of a circulating fluidized bed boiler combustion process controller: bed temperature, main steam pressure and furnace outlet flue gas oxygen content prediction step number P₁,P₂,P₃Controlling the weighting coefficient r according to the bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth₁,r₂,r₃Bed temperature, main steam pressure and furnace outlet flue gas oxygen content reference trajectory coefficient β₁,β₂,β₃Memory recursion with multiple variablesIdentification of initial value by least square methodP (0), forgetting factor μ;

the method is characterized in that: the method further comprises the following steps:

step 2, identifying and establishing a combustion process model through a multivariate fading memory recursive least square method according to the collected process variable data of the combustion process of the circulating fluidized bed boiler;

step 3, PID is controlled according to the characteristics of multivariable generalized predictive control₁、PID₂、PID₃The parameters of the controller are optimized and set, and the setting result is decoupled: the specific process is as follows:

3-a, calculating a backward shift operator q according to the equation set of the loss map^-1Polynomial of

3-b, establishing a predicted value y of the bed temperature at the moment k + j of the combustion process of the circulating fluidized bed boiler_1pPredicted values y of main steam pressure at moments (k + j) and (k + j)_2pPredicted value y of oxygen content of flue gas at furnace outlet at (k + j) and k + j moments_3p(k+j)；

3-c, establishing a multi-step predicted value Y of the bed temperature₁Main steam pressure multi-step prediction value Y₂And the multi-step predicted value Y of the oxygen content of the flue gas at the outlet of the hearth₃；

3-d, establishing an objective function J of bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth₁(k)、J₂(k)、J₃(k) Control step number M of bed temperature, main steam pressure and oxygen content of hearth outlet flue gas₁、M₂、M₃Taking the value as 1;

3-e, increasing the coal feeding amount by delta u₁(k) Primary air volume increment delta u₂(k) Secondary air quantity increment delta u₃(k) Form conversion, substitution into the objective function J₁(k)、J₂(k)、J₃(k) The optimization is carried out, and the decoupling processing is carried out on the optimization result to obtain a formula:

$\left[\begin{array}{c}{V}_1\\ V_2\\ V_3\end{array}\right]={\left[\begin{array}{ccc}{I}_1& -L_1{G}^{12}{\hat{e}}_2& -L_1{G}^{13}{\hat{e}}_3\\ -L_2{G}^{21}{\hat{e}}_1& I_2& -L_2{G}^{23}{\hat{e}}_3\\ -L_3{G}^{31}{\hat{e}}_1& -L_3{G}^{32}{\hat{e}}_2& I_3\end{array}\right]}^{-1}\left[\begin{array}{c}{L}_1(h_1{\tilde{e}}_1+f_1-Y_{1r})\\ L_2(h_2{\tilde{e}}_2+f_2-Y_{2r})\\ L_3(h_3{\tilde{e}}_3+f_3-Y_{3r})\end{array}\right]$

wherein,

$L_1={\[(r_1+G^{11T}{G}^{11}){{\hat{e}}_1}^T{\hat{e}}_1\]}^{-1}{{\hat{e}}_1}^T{G}^{11T}$

$L_2={\[(r_2+G^{22T}{G}^{22}){{\hat{e}}_2}^T{\hat{e}}_2\]}^{-1}{{\hat{e}}_2}^T{G}^{22T}$

$L_3={\[(r_3+G^{33T}{G}^{33}){{\hat{e}}_3}^T{\hat{e}}_3\]}^{-1}{{\hat{e}}_3}^T{G}^{33T}$

f₁＝F¹¹y₁(k)+H¹¹Δu₁(k-1)+H¹²Δu₂(k-1)+H¹³Δu₃(k-1)

f₂＝F²²y₂(k)+H²¹Δu₁(k-1)+H²²Δu₂(k-1)+H²³Δu₃(k-1)

f₃＝F³³y₃(k)+H³¹Δu₁(k-1)+H³²Δu₂(k-1)+H³³Δu₃(k-1)

Y_1r＝[y_1r(k+1),y_1r(k+2),…,y_1r(k+P₁)]^T

$y_{1r}(k+j)={{\mathrm{\β}}_1}^j{y}_{1r}(k+j-1)+(1-{{\mathrm{\β}}_1}^j)y_{1s}\left(k\right),j=1,...,P_1$

Y_2r＝[y_2r(k+1),y_2r(k+2),…,y_2r(k+P₂)]^T

$y_{2r}(k+j)={{\mathrm{\β}}_2}^j{y}_{2r}(k+j-1)+(1-{{\mathrm{\β}}_2}^j)y_{2s}\left(k\right),j=1,...,P_2$

Y_3r＝[y_3r(k+1),y_3r(k+2),…,y_3r(k+P₃)]^T

$y_{3r}(k+j)={{\mathrm{\β}}_3}^j{y}_{3r}(k+j-1)+(1-{{\mathrm{\β}}_3}^j)y_{3s}\left(k\right),j=1,...,P_3$

$h_1={\[1,1,...,1\]}_{P_1\×1}^T,h_2={\[1,1,...,1\]}_{P_2\×1}^T,h_3={\[1,1,...,1\]}_{P_3\×1}^T$

${\hat{e}}_1=\[e_1\left(k\right),e_1(k-1),e_1(k-2)\],{\hat{e}}_2=\[e_2\left(k\right),e_2(k-1),e_2(k-2)\],{\hat{e}}_3=\[e_3\left(k\right),e_3(k-1),e_3(k-2)\]$

e₁(k)、e₁(k-1)、e₁(k-2) are errors between the bed temperature reference trajectory value and the actual value at the time k, the time k-1 and the time k-2 respectively; e.g. of the type₂(k)、e₂(k-1)、e₂(k-2) errors between the reference trajectory values and actual values of the main steam pressure at the time k, the time k-1 and the time k-2, respectively; e.g. of the type₃(k)、e₃(k-1)、e₃(k-2) errors between the reference track value and the actual value of the oxygen content of the flue gas at the outlet of the hearth at the time k, the time k-1 and the time k-2 respectively; r is₁,r₂,r₃To control the weighting coefficients;is the prediction error of the bed temperature at the moment k,is the prediction error of the main steam pressure at time k,predicting error of oxygen content of the flue gas at the outlet of the hearth at the moment k; h is₁、h₂、h₃Predicting error correction matrixes for bed temperature, main steam pressure and oxygen content of flue gas at a hearth outlet; f. of₁、f₂、f₃Is a free movement term of bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; p₁、P₂、P₃Predicting the number of steps for bed temperature, main steam pressure and oxygen content in the flue gas at the outlet of the furnace, P₁、P₂、P₃The values are different; m₁、M₂、M₃Controlling the number of steps for bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; r is₁、r₂、r₃The weighting coefficients are used for controlling bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; y is_1r、Y_2r、Y_3rThe matrix is composed of reference track values at different moments of bed temperature, main steam pressure and oxygen content of the flue gas at the outlet of the hearth; y is_1r(k+j)、y_2r(k+j)、y_3r(k + j) is the value of bed temperature, main steam pressure and reference track of oxygen content of the flue gas at the outlet of the furnace at the moment of k + j β₁、β₂、β₃Reference trajectory coefficients are provided for bed temperature, main steam pressure and oxygen content of flue gas at a hearth outlet; y is_1s(k)、y_2s(k)、y_3s(k) Respectively setting values of bed temperature, main steam pressure and oxygen content of flue gas at the outlet of the hearth at the moment k; y is₁(k) Represents the actual output of the bed temperature at time k, y₂(k) Representing the actual output of the main steam pressure at time k, y₃(k) Representing the actual output of the oxygen content of the flue gas at the outlet of the hearth at the moment k; i is₁，I₂，I₃Is a unit ofA matrix; Δ u₁(k-1)、Δu₂(k-1)、Δu₃(k-1) respectively representing coal feeding quantity increment, primary air quantity increment and secondary air quantity increment at the k-1 moment;

3-f, calculating the PID of the k time by using the formula of the step 3-e₁、PID₂、PID₃Parameters of the controller:

$\begin{array}{ccc}\left\{\begin{array}{c}{k}_{d1}\left(k\right)=v_{12}\left(k\right)\\ k_{p1}\left(k\right)=-v_{11}\left(k\right)-2v_{12}\left(k\right)\\ k_{i1}\left(k\right)=v_{10}\left(k\right)+v_{11}\left(k\right)+v_{12}\left(k\right)\end{array}\right.& \left\{\begin{array}{c}{k}_{d2}\left(k\right)=v_{22}\left(k\right)\\ k_{p2}\left(k\right)=-v_{21}\left(k\right)-2v_{22}\left(k\right)\\ k_{i2}\left(k\right)=v_{20}\left(k\right)+v_{21}\left(k\right)+v_{22}\left(k\right)\end{array}\right.& \left\{\begin{array}{c}{k}_{d3}\left(k\right)=v_{32}\left(k\right)\\ k_{p3}\left(k\right)=-v_{31}\left(k\right)-2v_{32}\left(k\right)\\ k_{i3}\left(k\right)=v_{30}\left(k\right)+v_{31}\left(k\right)+v_{32}\left(k\right)\end{array}\right.\end{array}$

wherein, V₁(k)＝[ν₁₀(k),ν₁₁(k),ν₁₂(k)]^T，V₂(k)＝[ν₂₀(k),ν₂₁(k),ν₂₂(k)]^T，V₃(k)＝[ν₃₀(k),ν₃₁(k),ν₃₂(k)]^T；k_p1(k)、k_i1(k)、k_d1(k) PID of coal supply at time k₁Proportional, integral, differential coefficients of the controller; k is a radical of_p2(k)、k_i2(k)、k_d2(k) PID of primary air volume at time k₂Proportional, integral, differential coefficients of the controller; k is a radical of_p3(k)、k_i3(k)、k_d3(k) PID of secondary air volume at time k₃Proportional, integral, differential coefficients of the controller;

step 4, obtaining PID according to step 3₁、PID₂、PID₃The controller parameters respectively form control quantities to control the coal feeding quantity, the primary air quantity and the secondary air quantity, and further control the bed temperature, the main steam pressure and the oxygen content of the flue gas at the outlet of the hearth;

and 5, entering the next moment, returning to the step 2, and repeating the process from the step 2 to the step 5.

2. The method as claimed in claim 1, wherein the method comprises the steps of: in the step 4, the coal feeding amount, the primary air volume and the secondary air volume at the time k are as follows:

u₁(k)＝u₁(k-1)+k_p1(k)[e₁(k)-e₁(k-1)]+k_i1(k)e₁(k)+k_d1(k)[e₁(k)-2e₁(k-1)+e₁(k-2)]

u₂(k)＝u₂(k-1)+k_p2(k)[e₂(k)-e₂(k-1)]+k_i2(k)e₂(k)+k_d2(k)[e₂(k)-2e₂(k-1)+e₂(k-2)]

u₃(k)＝u₃(k-1)+k_p3(k)[e₃(k)-e₃(k-1)]+k_i3(k)e₃(k)+k_d3(k)[e₃(k)-2e₃(k-1)+e₃(k-2)]

wherein u is₁(k)、u₂(k)、u₃(k) Respectively shows the coal feeding quantity, the primary air quantity and the secondary air quantity at the moment k, u₁(k-1)、u₂(k-1)、u₃(k-1) represents the coal supply amount, the primary air amount and the secondary air amount at the time of k-1, respectively.

3. The method as claimed in claim 1, wherein the step 3 is performed by using the control weighting coefficients r of bed temperature, main steam pressure and furnace exit flue gas oxygen content₁、r₂、r₃The value range of (A) is 0.001-0.1.