CN1055317C - On-line Control Method of Continuous Annealing Furnace - Google Patents

Abstract

The invention discloses an online control method of a continuous annealing furnace, which is characterized in that a dynamic mathematical model of the whole furnace temperature of the annealing furnace is established based on a heat conduction theory to form a hybrid intelligent control system based on mathematical model quantitative calculation and artificial experience qualitative reasoning. The method comprises the steps of carrying out hybrid intelligent coordination on optimization control based on mathematical model quantitative calculation, presetting based on fast and slow variable separation and closed loop feedback intelligent correction by introducing artificial experience on the basis of classical PID to obtain total control quantity of furnace temperature of each furnace section, calculating various control signals through furnace temperature distribution and intelligent split-range, and realizing on-line control of a large annealing furnace by specific implementation of basic automation facilities.

Description

On-line Control Method of Continuous Annealing Furnace

The invention relates to an on-line control method of a continuous annealing furnace, in particular to a method for on-line control operation of a large-scale continuous annealing furnace by using a computer through a mathematical model.

In the prior art, large-scale continuous annealing furnaces generally adopt small loop control and computer monitoring systems equipped with various furnace temperature and belt temperature instruments, and the belt temperature is indirectly controlled by manually setting the furnace temperature. Temperature-based cascade control system. However, the entire annealing furnace has a huge structure, both heating and cooling in the furnace, the heat transfer mode is very different, and the mechanism is complicated. In addition, the strip steel type, specification, annealing curve and unit speed often change, and the thermal inertia of the furnace And large, making the working condition difficult to stabilize. Since the above-mentioned control system does not directly control the target zone temperature in the furnace, critical oscillation occurs in some furnace areas when the range adjustment is implemented according to the process requirements. Serious faults such as crooked or even broken belts, and high energy consumption. With the further development of computer technology, a method of using computer to control large annealing furnace through mathematical model has emerged. Most of the mathematical models used are static mathematical models, empirical simple dynamic models, etc., which cannot dynamically describe the furnace temperature and belt temperature distribution of the whole furnace, and cannot realize online operation control.

The purpose of the present invention is to provide an online control method for a continuous annealing furnace. By establishing a dynamic mathematical model of the temperature of the entire annealing furnace, a hybrid intelligent control system based on quantitative calculation of the model and qualitative reasoning based on artificial experience is formed, and the online control of the large-scale annealing furnace is carried out. Computer controlled operation.

The purpose of the present invention is achieved like this:

An on-line control method for a continuous annealing furnace. Based on the heat conduction theory, a dynamic mathematical model of the temperature of the entire annealing furnace is established to form a hybrid intelligent control system based on quantitative calculation of the mathematical model and qualitative reasoning based on manual experience. The specific steps are as follows:

1. The system enters the process parameters such as strip steel type, specification and annealing curve through the management computer, and sends them to the process computer.

2. The process computer collects parameters such as furnace temperature, belt temperature, unit operating speed and current operating conditions at each measuring point through basic automation facilities such as measuring instruments.

3. The host computer reads the above parameters from the process computer and is controlled by the hybrid intelligent control system:

a. Based on the optimal control of the quantitative calculation of the mathematical model, the furnace temperature control increment of each furnace section is obtained,

b. Based on the preset separation of fast and slow variables, the fast variables are separated from the system mathematical model, and the furnace temperature compensation control increments of each furnace section caused by the fast variable changes are preset,

c. On the basis of classic PID (proportional-integral-derivative control), artificial experience is introduced to carry out closed-loop feedback intelligent correction, and qualitative reasoning obtains the furnace temperature control increment of each furnace section.

d. Mix and intelligently coordinate the above three kinds of control increments based on quantitative calculation based on mathematical models and qualitative reasoning based on experience to determine the total control amount of furnace temperature in each furnace section,

e. The total control amount of furnace temperature in each furnace section determined by the hybrid intelligent coordination is distributed according to the quantity and logical relationship, and various control signals are obtained through multi-channel splitting, etc.

4. The process computer reads the above-mentioned control signals from the upper computer, and applies them to the production process through various regulators of the on-site basic automation facilities to realize online closed-loop control.

The whole furnace strip temperature dynamic mathematical model established by the present invention is a strip steel temperature dynamic distribution mathematical model represented by a partial differential equation, the annealing furnace is expanded according to the strip steel moving direction, and a three-dimensional space coordinate system of x, y, z is established, wherein The x, y, and z directions represent the strip thickness, furnace length, and furnace width respectively. Assuming that the strip temperature gradient in the Z direction of the furnace width is zero, the specific heat, density, and thermal conductivity of the strip are constant. According to Fourier’s thermal conductivity law, The two-dimensional unstable heat conduction equation describing the dynamic distribution of temperature in the whole furnace is obtained: $\frac{\∂T(x,\mathrm{the\; y},t)}{\∂t}=\frac{K_{\mathrm{the\; s}}}{\mathrm{C\ρ}}[\frac{{\∂}^{2}T(x,\mathrm{the\; y},t)}{{\∂x}^2}+\frac{{\∂}^{2}T(x,\mathrm{yt})}{{\∂\mathrm{the\; y}}^2}]-v\left(t\right)\frac{\∂T(x,\mathrm{the\; y},t)}{\∂\mathrm{the\; y}}---\left(1\right)$ Where: T(x ₁ y ₁ t) is the band temperature

t is time, t≥0

C is the specific heat of the strip

ρ is the density of strip steel

K _s is the thermal conductivity of the strip

v(t) is the speed of the unit to determine the upper and lower surface boundary conditions of the steel strip required to solve equation (1), and then use the time-space discretization technology to process, and transform the formula (1) into the furnace temperature as the control vector and the belt temperature as the state State-space mathematical models of vectors.

The whole furnace strip temperature dynamic mathematical model established by the present invention can also be the strip temperature tracking mathematical model represented by partial differential equations, the annealing furnace is expanded according to the moving direction of the strip, and the x, y moving coordinate system is established, wherein x, y are the positions of the tracking units in the direction of strip thickness and furnace length, respectively, and neglecting the transverse heat conduction of the strip along the furnace length, the movement of any small unit of the strip can be regarded as the movement of the boundary field. Assuming that the specific heat of the strip , density, and thermal conductivity are all constants. According to Fourier's thermal conductivity law, the one-dimensional unstable thermal conductivity equation describing the strip temperature tracking is obtained: $\frac{\∂T{S}_f(x,t)}{\∂t}=\frac{K_{\mathrm{the\; s}}}{\mathrm{C\ρ}}\frac{{\∂}^{2}T{S}_f(x,t)}{{\∂x}^2}---\left(1a\right)$ Where: TS _f is the temperature of the tracking unit

t is time, t≥0

C is the specific heat of the strip

ρ is the density of strip steel

K _s is the thermal conductivity of the strip to determine the upper and lower surface boundary conditions of the strip required to solve the equation (1a), and then use the time-space discretization technology to process the formula (1a) into a state variable with the strip temperature and the furnace temperature as A state-space mathematical model of the control variables. Under relatively stable working conditions, the strip temperature tracking mathematical model is equivalent to the strip temperature dynamic distribution mathematical model, and the strip temperature tracking mathematical model with a simple structure and a small amount of calculation can reflect the strip temperature distribution law of the whole furnace.

The optimal control based on the quantitative calculation of the mathematical model first needs to conduct online identification of the parameters in the mathematical model that are affected by the steel type, specification, unit speed, furnace condition, etc., and then predict the furnace temperature adjustment required to compensate the temperature deviation according to the mathematical model. Finally, optimization is carried out according to the objective function, and the furnace temperature control increment of each furnace section is obtained.

The preset based on the separation of fast and slow variables is to separate the speed V of the unit from the mathematical model as a fast variable, and calculate the dynamic compensation amount of the furnace temperature required when V changes ΔV according to the mathematical model, and calculate the dynamic compensation amount of the furnace temperature caused by the speed change. The furnace temperature compensation control increment is preset.

The artificial experience introduced by the closed-loop feedback intelligent correction mainly includes: increase the proportional action when the temperature deviation is large, add the integral action in the medium proportion when the deviation is medium, reduce the proportional action when the deviation is small, and add the integral action when the overshoot is over. When the integral function is canceled, the large, medium and small temperature deviation is also determined by manual actual experience.

Hybrid intelligent coordination is to record three kinds of control increments as μ ₁ , μ ₂ , μ ₃ respectively, and the total control quantity is μ. According to the process mechanism and operation experience, the knowledge base of the mapping relationship between the three control quantities and the total control quantity is established {μ ₁ , μ ₂ , μ ₃ }→μ, and obtain the total furnace temperature control quantity μ of each furnace section through knowledge reasoning.

Description of drawings:

Fig. 1 is a multi-level hierarchical control structure diagram of the method of the present invention.

Fig. 2 is a flow chart of the hybrid intelligent control process of the method of the present invention.

Figure 3 is a schematic diagram of the structure and process flow of a continuous hot-dip galvanizing annealing furnace.

The present invention will be described in detail below in conjunction with the accompanying drawings and by using the method of the present invention for on-line control of a continuous hot-dip galvanizing annealing furnace as an example.

The continuous hot-dip galvanizing annealing furnace shown in Figure 3 is a vertical structure and consists of four sections: preheating F ₁ , reduction F ₂ , controlled cooling F ₃ and spray cooling C ₁ , and these four sections have 16 furnace zones in total. The entrance of the preheating section _F1 and the exit of each furnace section are respectively installed with a belt temperature measuring instrument 1RT-5RT, which can obtain 5 belt temperature measurement values. A furnace temperature measuring instrument 1TC-16TC is installed in each furnace area, and 16 belt temperature measurement values can be obtained. The preheating section _F1 heats the steel strip by direct combustion and cleans the surface of the steel strip. Excessive heating may cause belt breakage or hot scoop in the furnace. The reduction section _F2 is heated by many radiant tubes that are symmetrical on both sides to continue to increase the band temperature. The band temperature 3RT at the exit of this section determines the mechanical properties of the product and therefore has strict requirements. The cooling section F ₃ is cooled by radiant tubes, and the spray cooling section C ₁ uses cooling gas to directly cool the steel strip. The strip temperature 5RT directly affects the quality of galvanizing, which can neither be too high nor too low. The whole annealing process must meet a certain annealing curve.

Based on the heat conduction theory, the present invention establishes a dynamic mathematical model of the annealing furnace temperature in the whole furnace, and constitutes a hybrid intelligent control system based on quantitative calculation of the model and qualitative reasoning based on manual experience, so as to meet the requirements of the annealing curve. The system adopts a multi-level hierarchical control structure, as shown in Figure 1, the process computer is collected by the basic automation facilities and read in various parameters by the management computer, and the upper computer reads the above parameters from the process computer to perform hybrid intelligent control and obtain the control The signal is applied to the production process by the process computer through the basic automation facilities, and the closed-loop online control of the continuous hot-dip galvanizing annealing furnace is carried out.

The mathematical model used to quantitatively calculate the control amount of the furnace temperature in the embodiment is a mathematical model of the dynamic distribution of the steel strip temperature expressed by a partial differential equation. Expand the hot-dip galvanizing annealing furnace according to the moving direction of the strip, and establish a three-dimensional coordinate system of x, y, and z, where the x, y, and z directions represent the strip thickness, furnace length, and furnace width respectively. The temperature gradient of the steel is zero, and the specific heat, density, and thermal conductivity of the strip steel are all constant. According to Fourier's heat conduction law, a two-dimensional unstable heat conduction equation describing the dynamic distribution of the entire furnace strip temperature is obtained: $\frac{\∂T(x,\mathrm{the\; y},t)}{\∂t}=\frac{K_{\mathrm{the\; s}}}{\mathrm{C\ρ}}[\frac{{\∂}^{2}T(x,\mathrm{the\; y},t)}{{\∂x}^2}+\frac{{\∂}^{2}T(x,\mathrm{yt})}{{\∂\mathrm{the\; y}}^2}]-v\left(t\right)\frac{\∂T(x,\mathrm{the\; y},t)}{\∂\mathrm{the\; y}}---\left(1\right)$ Among them: T(x, y, t) is the band temperature

t is time, t≥0

C is the specific heat of the strip

ρ is the density of strip steel

K _s is the thermal conductivity of the strip

v(t) is the speed of the unit to determine the boundary conditions on the upper and lower surfaces of the strip required to solve equation (1): the energy transfer between the surface of the strip in the furnace and the annealing furnace is mainly radiation and convection, assuming that the heat transfer on the upper and lower surfaces of the strip If it is symmetrical, only the boundary conditions on the lower surface of the strip are considered, and the relational formula of the heat flux on the lower surface of the strip is obtained:

q(y,t)=ε(y)F _a σ[TZ ⁴ (y,t)-T ⁴ (0,y,t)]+h _c F _a [TZ(y,t)

-T(0, y, t)]...(2) where: q(y, t) is the heat flux on the lower surface of the strip

F _a is the effective heat transfer area

σ is the Stefan-Boltzmann constant (Stefan-Boltzmann) constant

TZ(y, t) is the furnace temperature distribution

h _c is the convective heat transfer coefficient of the furnace gas to the lower surface of the strip

ε(y) is the total effective blackness coefficient of the entire furnace body to the lower surface of the strip,

ε(y)＝φ _sw ε _s +[ε _w +ε _g (y)]/2 where: ε _s is the blackness coefficient of strip steel

ε _w is the total effective blackness coefficient of the furnace wall to the lower surface of the strip

ε _g is the furnace gas blackness coefficient

φ _sw is the radiation angle coefficient between the furnace wall and the lower surface of the steel strip

In order to simplify the model for practical engineering application, the comprehensive equivalent heat transfer coefficient h(y ₂ t) is introduced:

h(y,t)=ε(y)σ[TZ ² (y,t)+T ² (0,y,t)][TZ(y,t)-T(0,y,t)]+h _c

...(3) Equation (2) can be simplified into a linearized boundary condition:

q(y, t)=h(y, t) F _a [TZ ² (y, t)-T(0, y, t) ... (4) So the boundary condition (4) can be expressed as: ${\frac{\∂T(x,\mathrm{the\; y},t)}{\∂x}|}_{x=0}=-\frac{h(\mathrm{the\; y},t)}{K_{\mathrm{the\; s}}}[Z(\mathrm{the\; y},t)-T(0,\mathrm{the\; y},t)]---\left(6\right)$ In the actual on-line control process, h(y, t) is determined by real-time recursive estimation.

Using space-time discretization technology to engineer the above mathematical model. Divide the steel strip in the furnace into N _x ×N _y networks, and the time step is Δt. For convenience, iΔx, jΔy and kΔt are abbreviated as i, j and k, respectively. Appropriately using the forward difference and backward difference approximation, the band temperature distribution can be modeled as:

T(i,j,k+1)=aT(i+1,j,k)+(1-2a-2b-c)T(i,j,k)+aT(i-1,j,k) +

bT(i, j+1, k)+(b+c)T(i, j-1, k)...(7) where:

Take the average temperature T(y, t) in the thickness direction of the strip at the coordinate y as the approximate value of the strip temperature TS(y, t) at y, $\stackrel{-}{T}(j,k+1)=\frac{1}{N_x}\underset{i=1}{\overset{N_x}{\mathrm{\Σ}}}T(i,j,k+1)$ =(1-2b-c)T(j,k)+bT(j+1,k)+(b+c)T(j-1,k)+ $\frac{\mathrm{\σ}}{N_x}[T(N_x+1,j,k)-T(N_x,j,k)+T(0,j,k)-T(i,j,k)]---\left(8\right)$ ∴TS(j,k+1)=(1-2b-cd _jk )TS(j,k)+(b+c)TS(j-1,k)

类似地，可以导出：+bTS(j+1, k)+d _jk TZ(j, k) ... (9) where:

Similarly, it is possible to export:

TS(0,k+1)=(1-bd _0k )TS(0,k)+bTS(j,k)+d _0k TZ(0,k)...(10) $\mathrm{TS}(N_{\mathrm{the\; y}},k+1)=(b+c)\mathrm{TS}(N_{\mathrm{the\; y}}-1,k)+(1-b-c-d_{N_{\mathrm{the\; y}}K})\mathrm{TS}(N_{\mathrm{the\; y}},k)$ $+d_{N_{\mathrm{the\; y}}K}Z(0,k)----\left(11\right)$ If remember:

X(k)=[TS(0,k),TS(j,k),...,TS(N _y ,k)] ^T

U(k)=[TZ(0,k), TZ(j,k),..., TZ(N _y ,k)] ^T $B\left(k\right)=\mathrm{diag}[d_{0k},d_{1k},\&Center\; Dot;\&Center\; Dot;\&Center\; Dot;\·\·\·,d_{N_{\mathrm{the\; y}}K}]$

Then finally derive a state-space mathematical model with the belt temperature as the state variable and the furnace temperature as the control variable

X(k+1)=A(k)X(k)+B(k)U(k) ... (12) Various characteristic parameters such as steel type, specification, unit speed and furnace condition are included in A(k) ) and B(k).

The specific control steps are as follows (see Figure 1 and Figure 2):

1. The system inputs process parameters such as strip steel type, specification, annealing curve, etc. through the management computer, and sends them to the process computer.

2. The process computer collects parameters such as furnace temperature, belt temperature, unit operating speed and current operating conditions at each measuring point through basic automation facilities such as measuring instruments.

3. The host computer reads the above parameters from the process computer and is controlled by the hybrid intelligent control system:

a. Carry out optimal control based on mathematical model quantitative calculation. Firstly, the online identification of A(K) and B(K) in the state space mathematical model formula (12) including steel types, specifications, unit operating speed, furnace conditions and other parameters is carried out, and then the band temperature deviation ΔX( k+1) the required furnace temperature adjustment ΔU(k), and finally optimize it with the goal of saving energy, and obtain the furnace temperature control increment of each furnace section.

b. Presetting based on the separation of fast and slow variables. In order to quickly overcome the disturbance of fast variables, the operating speed V of the unit is separated from the mathematical model shown in formula (12) as a fast variable, and the dynamic compensation amount of the furnace temperature required when V changes ΔV is calculated according to the mathematical model. The resulting furnace temperature compensation control increment of each furnace section is preset.

c. Introduce artificial experience on the basis of classic PID for closed-loop feedback intelligent correction. The manual experience introduced mainly includes: increase the proportional action when the temperature is large, add the integral action in the medium proportion when the deviation is medium, reduce the proportional action when the deviation is small, add the integral action when overshooting, and cancel the integral action when it is called back. The large, medium and small of the band temperature deviation are also determined by manual experience. Based on this qualitative reasoning, the furnace temperature control increment of each furnace section is obtained to overcome a large number of uncertainties in the continuous hot-dip galvanizing annealing process.

d. Carry out mixed intelligent coordination on the above three kinds of control increments obtained from quantitative calculation based on mathematical model and qualitative reasoning based on experience. The working conditions are different, and the proportions and functions of the three control quantities are different. The three control quantities are respectively recorded as μ ₁ , μ ₂ , and μ ₃ , and the total control quantity is μ, which is established according to the process mechanism and operating experience. The knowledge base {μ ₁ , μ ₂ , μ ₃ )→μ of the mapping relationship between the three control quantities and the total control quantity is used to obtain the total control quantity of the furnace temperature in each furnace section through knowledge reasoning.

e. The total furnace temperature control amount of each furnace section determined by the hybrid intelligent coordination is distributed according to the quantity and logical relationship, and converted into various control signals through multi-channel division. The conversion process is expressed as a series of production rules according to the process mechanism and human experience, which can be used for real-time reasoning of the control program.

4. The lower computer reads the above-mentioned control signals from the upper computer, and applies them to the production process through various regulators of the on-site basic automation facilities to realize online closed-loop operation control.

Compared with the prior art, the present invention has the advantages that: the present invention establishes a dynamic mathematical model of the temperature of the annealing furnace to form a hybrid intelligent control system based on quantitative calculation of the mathematical model and qualitative reasoning based on artificial experience, and truly realizes the control of On-line control of large annealing furnace. It has been proved by actual use on large-scale continuous hot-dip galvanizing annealing furnaces that the mathematical model established by the method of the present invention has high precision, and the hybrid intelligent control solves the problems caused by the inability to control the strip temperature and the incoordination between the furnace temperature and the strip speed in the past. Troubleshooting and quality problems such as tape and hot scoops ensure the normal production, stable product quality, increased production and reduced energy consumption.

The method of the invention can be used for online control of all vertical and horizontal continuous annealing furnaces.

Claims (5)

1. online controlling method for continuously annealing furnace, it is characterized in that: by being foundation with the heat transfer theory, set up the whole furnace zone of annealing furnace temperature dynamic mathematical models, formation is based on the mathematical model quantitative Analysis with based on the hybrid intelligent Controlling System of artificial experience qualitative reasoning, and concrete steps are as follows:

(1) system is by supervisory computer input tape steel steel grade, specification, and processing parameters such as annealing curve are sent into process computer,

(2) the process computer parameters such as furnace temperature, band temperature, unit operation speed and current operational condition of gathering each point position by basic automatization facilities such as metrical instruments,

(3) host computer reads above-mentioned parameter from process computer, is controlled by the hybrid intelligent Controlling System:

A. based on the optimal control of mathematical model quantitative Analysis, draw each stove section Control for Kiln Temperature increment,

B. preset based on fast slow variable is isolating, fast variable separated from system mathematic model, fast variable is changed each the stove section Furnace temperature compensation control increment that causes preset,

C. introduce artificial experience and carry out the correction of closed loop feedback intelligence on the classical PID basis, qualitative reasoning draws each stove section Control for Kiln Temperature increment,

D. carry out the hybrid intelligent coordination to above-mentioned based on the mathematical model quantitative Analysis with based on three kinds of control increment of experience qualitative reasoning gained, determine each stove section furnace temperature overhead control amount,

E. will carry out furnace temperature according to quantity and logical relation through each stove section furnace temperature overhead control amount that hybrid intelligent coordinate to determine and distribute, obtain various control signals through hyperchannel branch journey etc.,

(4) process computer reads above-mentioned control signal from host computer, puts on production process through the various setters of on-the-spot basic automatization facility, realizes online closed-loop control;

Described whole furnace zone temperature dynamic mathematical models are belt steel temperature DYNAMIC DISTRIBUTION mathematical models of representing with partial differential equation, annealing furnace is launched by band steel travel direction, and set up x, y, z three dimensional space coordinate system, wherein x, y and z direction are represented tape thickness, furnace superintendent, stove cross direction respectively, suppose that the wide Z direction of stove belt steel temperature gradient is zero, specific heat, density, the thermal conductivity of band steel are constant, according to sad heat conduction law in the richness, obtain describing the two dimensions and unstable heat conduction equation of full furnace zone temperature DYNAMIC DISTRIBUTION: $\frac{\∂T(x,y,t)}{\∂t}=\frac{K_s}{\mathrm{C\ρ}}[\frac{{\∂}^{2}T(x,y,t)}{{\∂x}^2}+\frac{{\∂}^{2}T(x,\mathrm{yt})}{{\∂y}^2}]-v\left(t\right)\frac{\∂T(x,y,t)}{\∂y}---\left(1\right)$ Wherein: T (x ₁y ₁T) be the band temperature

T is the time, t 〉=0

C is the specific heat of band steel

ρ is the density of band steel

K _sThermal conductivity for the band steel

V (t) determines the band steel upper and lower surface final condition that solving equation (1) is required for unit speed, adopts the space-time discretization technique to handle again, and it is that control vector, band temperature are the space mathematical model of state vector that formula (1) is turned to the furnace temperature;

Described whole furnace zone temperature dynamic mathematical models also can be to follow the tracks of mathematical model with the belt steel temperature that partial differential equation are represented, annealing furnace is launched by band steel travel direction, and set up x, y moving coordinate system, wherein x, y are respectively the position of tracking cell on tape thickness and furnace superintendent direction, ignore the transverse heat transfer of band steel along the furnace superintendent direction, then just can be considered moving of boundary field with arbitrarily small unitary the moving of steel, specific heat, density, the thermal conductivity of supposing the band steel are constant, according to sad heat conduction law in the richness, obtain describing the one dimension unsteady heat conduction equation that belt steel temperature is followed the tracks of: $\frac{\∂T{S}_f(x,t)}{\∂t}=\frac{K_s}{\mathrm{C\ρ}}\frac{{\∂}^{2}T{S}_f(x,t)}{{\∂x}^2}---\left(1a\right)$ Wherein: TS _fTemperature for tracking cell

T is the time, t 〉=0

C is the specific heat of band steel

ρ is the density of band steel

K _sDetermine the band steel upper and lower surface final condition that solving equation (1a) is required for the thermal conductivity of band steel, adopt the space-time discretization technique to handle again, formula (1a) is turned to being that state variables, furnace temperature are the space mathematical model of controlled variable with temperature.

2. method according to claim 1, it is characterized in that described optimal control based on the mathematical model quantitative Analysis, at first to carry out on-line identification to the parameter that influenced by steel grade, specification, unit speed, the working of a furnace etc., then according to the required furnace temperature regulated quantity of mathematical model prediction compensating band temperature deviation, be optimized according to getting objective function at last, obtain each stove section Control for Kiln Temperature increment.

3. method according to claim 1, it is characterized in that describedly presetting based on fast slow variable is isolating, be that unit speed V is separated from mathematical model as fast variable, required furnace temperature dynamic compensation amount when going out V changes delta V according to calculated with mathematical model presets caused each the stove section Furnace temperature compensation control increment of velocity variations.

4. method according to claim 1, it is characterized in that described closed loop feedback intelligence proofreaies and correct the artificial experience of being introduced and mainly comprise: strengthen proportional action when being with warm large deviation, in during deviation moderate proportions add the integration effect, reduce proportional action during little deviation, and when overshoot, add integral action, cancellation integral action during readjustment is also determined by artificial practical experience with the large, medium and small of warm deviation.

5. method according to claim 1 is characterized in that it is three kinds of control increment to be remembered respectively make μ that described hybrid intelligent is coordinated ₁, μ ₂, μ ₃, the overhead control amount is μ, sets up three kinds of manipulated variables and overhead control amount mapping relations knowledge base { μ according to process mechanism and service experience ₁, μ ₂, μ ₃} → μ obtains each stove section furnace temperature overhead control amount μ by knowledge reasoning.

Images