CN105276561A - Self-adaption predictive control method for main steam pressure of boiler - Google Patents

Abstract

The invention discloses a self-adaption predictive control method for the main steam pressure of a boiler. The method specifically comprises the steps that (a) an empirical model is established, that is, the coupling relationship between the main steam pressure and the bed temperature is established; (b) static decoupling is conducted; (c) parameter identification is conducted, that is, standards are introduced into the decoupled model so as to participate in the identification of model parameters; (d) data sampling is conducted, that is, unit step data of the decoupled main steam pressure and bed temperature object are sampled; and (e) optimizing calculation is conducted, and the identified model is calculated through a dynamic matrix algorithm. According to the method, by decoupling the coupling relationship between the main steam pressure and the bed temperature, identifying the object properties of the circulating fluidized bed boiler and finally automatically adjusting the multi-parameter non-linear and time-varying complex system with multiple variables being coupled closely through the self-adaption predictive control method, the control effect is within the optimum range, and the modeling process and the control method have universality and popularization significance is achieved for relative control process.

Description

A kind of adaptive prediction control method of boiler main steam pressure

[technical field]

The present invention relates to the technical field of boiler implosion, particularly the automatic control technology field of CFBB main steam pressure.

[background technology]

Combustion technology of circulating fluidized is efficient, the oligosaprobic New combustion technique of one grown up on the basis of bubbling fluidized bed combustion technology.CFBB (CirculatingFluidizedBedBoiler is called for short CFBB) is a kind of device of burn fossil matter fuel production steam.Under the burner hearth of boiler operates in a kind of special flow dynamic characteristic, fine grained is conducted through burner hearth by with the gas speed exceeding average grain diameter particle terminal velocity, there are again enough solid back-mixings to ensure the uniformity of temperature profile in burner hearth simultaneously, fuel and desulfurizing agent repeatedly circulate in burner hearth, repeatedly carry out low-temperature burning and desulfurization, thus reach the efficient burning of fuel and the low emission of pollutant.

CFBB is made up of burner hearth, gas-solid separator, solid material recirculation device, external heat exchangers etc. usually, and compared with other boilers, its main feature is: 1. fuel preparation system is simple; 2. the ignition temperature of boiler is lower; 3. material burning in fluidisation state in burner hearth, enhanced burning intensity and efficiency of combustion greatly; 4. the material circulating system etc. of uniqueness.

Exactly because CFBB time become, the characteristic such as non-linear, multi-state, so it is very natural to introduce Self Adaptive Control.Self Adaptive Control, by the state of continuous measuring system, performance and parameter, obtains the operating index of current system and makes comparisons with expectation index, thus the structure and parameter of adjustment controller, to ensure the optimum of system cloud gray model under certain meaning or suboptimum state.When this characteristic of adaptive controller just ensure that change or variable working condition when CFBB object occurs, controller can adapt to this change by adjustment parameter, thus ensures control effects.

[summary of the invention]

Object of the present invention solves the problems of the prior art exactly, a kind of adaptive prediction control method of boiler main steam pressure is proposed, because there is strong coupled relation in main steam pressure and bed temperature, the Mathematical Modeling to simplifying is needed first to carry out decoupling zero, then carry out identification, Self Adaptive Control is carried out to main steam pressure.

For achieving the above object, the present invention proposes a kind of adaptive prediction control method of boiler main steam pressure, concrete steps comprise:

A () sets up empirical model: according to the practical experience of execute-in-place, sets up the coupled relation between main steam pressure and bed temperature;

B () static decoupling: carry out static decoupling to the model of (a) step, eliminates association between the two;

(c) parameter identification: criterion is introduced to the model after decoupling zero, carries out the identification of model parameter;

(d) data sampling: determine sampling time T, the main steam pressure after sampling decoupling zero and the unit step data of bed temperature object;

E () optimizes calculating: utilize the model of dynamic matrix control algorithm to identification to calculate, on-line correction model error.

As preferably, described coupling model accumulates gradually in practical experience, the scope larger to parameter identification process and selection, leaves certain allowance to different operating mode, boiler etc.

As preferably, the time static decoupling mode that described Uncoupled procedure adopts, be ripe application and the effect of considering the method, different decoupling zero modes finally all can have influence on the quality of control effects.

As preferably, it is utilize dynamic matrix control algorithm to realize that described adaptive prediction controls, and needs the unit step data obtaining main steam pressure and bed temperature object in optimizing process.

Beneficial effect of the present invention: the present invention is by the coupled relation between decoupling zero main steam pressure and bed temperature, the plant characteristic of identification CFBB, finally by self-adaptation control method automatically regulate this multi-parameter, non-linear, time become and the closely-coupled complication system of multivariable, make control effects in optimized scope, be conducive to real-time control and the on-line correction of computer, improve reaction speed and the control position precision of procedures system, modeling process and control method have universality, have dissemination to relevant control process.

Feature of the present invention and advantage will be described in detail by reference to the accompanying drawings by embodiment.

[accompanying drawing explanation]

Fig. 1 is the control block diagram of the static decoupling of the adaptive prediction control method of a kind of boiler main steam pressure of the present invention;

Fig. 2 is the simulation architecture figure of the adaptive prediction control method of a kind of boiler main steam pressure of the present invention.

[detailed description of the invention]

Consult Fig. 1, Fig. 2, the adaptive prediction control method of a kind of boiler main steam pressure of the present invention, concrete steps comprise:

Step one, practical experience according to execute-in-place, set up the coupled relation between main steam pressure and bed temperature;

Step 2, static decoupling is carried out to the model of step one, eliminate association between the two;

Step 3, to after decoupling zero model introduce criterion, carry out the identification of model parameter;

Step 4, determine sampling time T, the main steam pressure after sampling decoupling zero and the unit step data of bed temperature object;

Step 5, the model of dynamic matrix control algorithm to identification is utilized to calculate, on-line correction model error.

Described coupling model accumulates gradually in practical experience, the scope larger to parameter identification process and selection, leaves certain allowance to different operating mode, boiler etc.; The time static decoupling mode that described Uncoupled procedure adopts, be ripe application and the effect of considering the method, different decoupling zero modes finally all can have influence on the quality of control effects; It is utilize dynamic matrix control algorithm to realize that described adaptive prediction controls, and needs the unit step data obtaining main steam pressure and bed temperature object in optimizing process.

In step one, when combustion rate disturbance and steam turbine adopt hydraulic speed regulation system, the dynamic characteristic of main steam vapour pressure is:

$G_{P1}\left(s\right)=\frac{K_{P1}}{{(1+T_{P1}s)}^2}{e}^{-{\mathrm{\τ}}_{P1}s}$

When bed temperature automatic control system drops into operation with closed ring, air output does step disturbance, and the broad sense dynamic characteristic of the main steam pressure obtained is:

$G_{P2}\left(s\right)=\frac{K_{P1}}{{(1+T_{P2}s)}^3}{e}^{-{\mathrm{\τ}}_{P2}s}$

Under coal-supplying amount step disturbance, the dynamic characteristic of bed temperature is:

$G_{\mathrm{\θ}1}\left(s\right)=\frac{K_{\mathrm{\θ}1}}{{(1+T_{\mathrm{\θ}1}s)}^2}{e}^{-{\mathrm{\τ}}_{\mathrm{\θ}1}s}$

Under air output disturbance, the dynamic characteristic of bed temperature is:

$G_{\mathrm{\θ}1}\left(s\right)=\frac{K_{\mathrm{\θ}2}}{1+T_{\mathrm{\θ}2}s}$

To sum up, main steam pressure can be obtained and pass letter with being coupled of bed temperature:

$\left[\begin{array}{c}P\\ T\end{array}\right]=\left[\begin{array}{c}\frac{K_{P2}}{{(1+T_{P2}s)}^3}{e}^{-{\mathrm{\τ}}_{p2}s}\frac{K_{P1}}{{(1+T_{P1}s)}^2}{e}^{-{\mathrm{\τ}}_{P1}s}\\ \frac{K_{\mathrm{\θ}2}}{1+T_{\mathrm{\θ}2}s}\frac{1+T_{\mathrm{\θ}2}s}{(1+T_{\mathrm{\θ}2}s)2}{e}^{-{\mathrm{\τ}}_{\mathrm{\θ}1}s}\end{array}\right]\left[\begin{array}{c}V\\ M\end{array}\right]$

Wherein, M is fuel quantity, and V is air output, and P is main steam pressure, and T is bed temperature.

In step 5, because CFBB has larger lag characteristic, and in step 4, the parameter of identification has certain fluctuation, so have selected the dynamic matrix control algorithm having and overcome and postpone with compared with strong robustness, detailed process is as follows:

The sampled value a of determination object step response _i=a (iT), wherein T is the sampling period.T _ntend towards stability after=NT, i.e. a _n≈ a (∞), then object multidate information can be described as finite aggregate a={a ₁, a ₂... a _n, a ^tfor model vector, N is modeling time domain.

${\tilde{y}}_M(k+i|k)={\tilde{y}}_O(k+i|k)+\underset{j=1}{\overset{\min (M,i)}{\Σ}}{a}_{i-j+1}\mathrm{\Δ}u(k+i-1),i=1,2,\mathrm{3....}N$

Wherein, for the k moment exports the prediction in k+i moment, for the initial prediction in k moment.

Vector form: ${\tilde{y}}_{PM}\left(k\right)={\tilde{y}}_{PO}\left(k\right)+{\mathrm{A\Δu}}_M\left(k\right)$

Wherein ${\tilde{y}}_{PM}\left(k\right)=\[\begin{array}{c}{\tilde{y}}_M(k+1|k)\\ \mathrm{...}\\ {\tilde{y}}_M(k+P|k)\end{array}\],{\tilde{y}}_{PO}\left(k\right)=\[\begin{array}{c}{\tilde{y}}_O(k+1|k)\\ \mathrm{...}\\ {\tilde{y}}_O(k+P|k)\end{array}\],$

Optimality criterion:

$\min J\left(k\right)=\underset{i=1}{\overset{P}{\Σ}}{q}_i{\[w(k+i)-{\tilde{y}}_M(k+i|k)\]}^2+\underset{j=1}{\overset{M}{\Σ}}{r}_j{\mathrm{\Δu}}^2(k+j-1)$

Wherein q _i, r _jfor weights coefficient.

After vector form, local derviation obtain:

${\mathrm{\Δu}}_M\left(k\right)={(A^{T}QA+R)}^{-1}{A}^{T}Q\[w_P\left(k\right)-{\tilde{y}}_{PO}\left(k\right)\],$ Wherein Q=diag (q ₁... q _p), R=diag (r ₁... r _m).

According to the requirement of rolling optimization, then only get first value of Δ u (k), that is:

$\mathrm{\Δ}u\left(k\right)=C^T{\mathrm{\Δu}}_M\left(k\right)=d^T\left(w_P\right(k)-\widetilde{y_{PO}\left(k\right)}),$ Wherein C ^t=[10...0].

Optimum by obtaining output valve at feedback compensation:

actual to the prediction output in k+1 moment and the system y (k+1) that exports is compared, forms output error: thus the prediction revising future time instance exports initial value ${\tilde{y}}_{PO}(k+1)={\tilde{y}}_{cor}(k+1)={\tilde{y}}_{N1}\left(k\right)+He(k+1),$ Wherein for the prediction output valve in k moment.

The course of work of the present invention:

The present invention is by the coupled relation between decoupling zero main steam pressure and bed temperature, the plant characteristic of identification CFBB, finally by self-adaptation control method automatically regulate this multi-parameter, non-linear, time become and the closely-coupled complication system of multivariable, make control effects in optimized scope, be conducive to real-time control and the on-line correction of computer, improve reaction speed and the control position precision of procedures system, modeling process and control method have universality, have dissemination to relevant control process.

Above-described embodiment is to explanation of the present invention, is not limitation of the invention, anyly all belongs to protection scope of the present invention to the scheme after simple transformation of the present invention.

Claims (4)

1. the adaptive prediction control method of a boiler main steam pressure, it is characterized in that: described control method there is strong coupled relation according to main steam pressure and bed temperature, first carries out decoupling zero, then carry out identification, carry out Self Adaptive Control to main steam pressure again, concrete steps comprise:

A () sets up empirical model: according to the practical experience of execute-in-place, sets up the coupled relation between main steam pressure and bed temperature;

B () static decoupling: carry out static decoupling to the model of (a) step, eliminates association between the two;

(c) parameter identification: criterion is introduced to the model after decoupling zero, carries out the identification of model parameter;

(d) data sampling: determine sampling time T, the main steam pressure after sampling decoupling zero and the unit step data of bed temperature object;

E () optimizes calculating: utilize the model of dynamic matrix control algorithm to identification to calculate, on-line correction model error.

2. the adaptive prediction control method of a kind of boiler main steam pressure as claimed in claim 1, it is characterized in that: in described (a) step, coupling model accumulates gradually in practical experience, the scope larger to parameter identification process and selection, leave certain allowance to different operating mode, boiler etc.

3. the adaptive prediction control method of a kind of boiler main steam pressure as claimed in claim 1, it is characterized in that: in described (b) step, the time static decoupling mode that Uncoupled procedure adopts, be ripe application and the effect of considering the method, different decoupling zero modes finally all can have influence on the quality of control effects.

4. the adaptive prediction control method of a kind of boiler main steam pressure as claimed in claim 1, it is characterized in that: in described (d), (e) step, it is utilize dynamic matrix control algorithm to realize that adaptive prediction controls, and needs the unit step data obtaining main steam pressure and bed temperature object in optimizing process.