JPH066735B2 - Combustion control method for continuous heating furnace - Google Patents

Description

TECHNICAL FIELD The present invention relates to combustion control of a continuous heating furnace having a plurality of furnace zones.

[Conventional technology]

In the fuel control of the continuous multi-zone heating furnace, the method to optimize the material heating curve (heating pattern) online is
In recent years, several proposals have been made as a powerful and flexible method for optimizing the operation such as ensuring the burning quality and energy saving for various charging of different materials.

As a typical example thereof, the system described in Japanese Patent Laid-Open No. 61-199016 divides the heating furnace into n meshes in the furnace length direction, and the k-th mesh exists for each material in the furnace. Change Δw _k from the current value of the fuel flow rate w _k when
And material average temperature _s during extraction, soaking degree (temperature difference between surface and inside) And the relationship between the k-th mesh furnace temperature _gi and the variation from the current value are approximated by a linear expression, _s , Fuel flow rate integrated value for each material under the constraint conditions regarding The optimum flow rate w _KOPT (k = 1, 2, ...,
n) is obtained using linear programming. Each zone setting furnace temperature _T'gA (i) (i is a material index) for each material is obtained from this w _KOPT , and each zone setting furnace temperature is determined as a weighted average value for each material.

On the other hand, in Japanese Patent Laid-Open No. 58-210120, the furnace temperature is optimized instead of the fuel flow rate, but the linear programming method is used in the same manner.

[Problems to be Solved by the Invention]

The problem will be described below with reference to Japanese Patent Laid-Open No. 61-199016, which optimizes the fuel flow rate, but the problem is the same when the furnace temperature is optimized as in Japanese Patent Laid-Open No. 58-210120.

Now, in the conventional continuous heating furnace control method as described above,
The relationship between the amount of change in the fuel flow rate from the current value and the amount of change in each of the material average temperature, soaking degree, and furnace temperature at the time of extraction is approximated by a linear equation, but these relationships are essentially Since it is nonlinear, the error increases as the amount of change increases. Therefore, when the size of the material, the planned in-reactor time, and the target extraction temperature are different, the amount of change also becomes large, which causes an increase in the error of the linear approximation. Since it is necessary to make the fuel flow rate finer by increasing the value to a certain degree, the number of parameters to be optimized (fuel flow rate change amount for each material category) for all materials in the reactor increases, and if linear approximation is not performed, nonlinear When trying to perform optimization up to now, the amount of calculation becomes enormous and it becomes difficult to execute online.

On the other hand, in the operation of the heating furnace, the operation schedule (remaining time of the material in the furnace, etc.) is often changed depending on the work situation of the process before and after, but such a direct operation amount (fuel flow rate) is set as the optimization target parameter. When is selected, the schedule change has a large influence on the optimum parameter value, and optimization calculation needs to be performed every time the schedule change occurs, resulting in a large calculation load. Further, since the optimization calculation requires a certain amount of time, the action amount change action with respect to the schedule change is delayed by this amount of time, and the effect of optimization is reduced.

[Means for Solving the Problem] The present invention solves the problems of the conventional control as described above, and a target curve of temperature rise from charging to extraction for each material (target temperature rise curve) Is set for each material, and each future material temperature does not fall below the target temperature rise curve for a specified time within the same zone, or each material temperature is averaged or given priority to this target temperature rise curve for a specific material. Such a fuel flow rate is output from time to time in a predetermined control cycle, and each material is divided into several groups by grouping similar items that are close to each other, and the target temperature rise curve has a common value in each group. The value of the common parameter (for example, the material heating rate in each zone) is determined by a non-linear optimization method such as the steepest descent method so as to minimize a predetermined evaluation function value.

The value of the evaluation function is obtained by simulating the heating furnace operation until a predetermined material is extracted by a process model (heating furnace simulator) that includes a nonlinear heat balance equation and a heat conduction equation and can be calculated at high speed. Since linear approximation of the equation is not performed here, an error due to linearization does not occur unlike the conventional technique. Further, the function of the common parameter of the target temperature rising curve which is the optimization target parameter may be significantly smaller than the number of future fuel flow rate time series values in the conventional technique, and the optimization calculation becomes efficient accordingly.

On the other hand, regarding the change of the operating schedule of the heating furnace, if the target temperature rising curve is expressed as a function of the remaining in-reactor time, for example, when the remaining in-reactor time of a material is changed, the temperature of the material after a predetermined time Since the target value is also automatically changed on the target temperature rising curve, it is not necessary to change the common parameter value defining the target temperature rising curve so much. The same applies when the target extraction temperature of the material is changed, and in the present invention, the change of the operation schedule has little influence on the optimum value of the optimal change parameter, and therefore, the optimization calculation is performed every time the schedule is changed. There is no need to do it, the calculation load is reduced,
The operation amount change at the time of changing the schedule can be performed immediately without waiting for the optimization calculation, and the effect of the optimization is maintained.

[Action]

 The present invention will be described in detail below.

The target plant of the present invention is a multi-zone continuous heating furnace,
One example is shown in FIG. In the heating furnace, various materials with different material data (thickness, width, length, material, charging temperature, extraction target temperature, in-furnace time or extraction pitch prediction value with preceding material) are loaded from the charging side to the extraction side. The temperature is raised to the extraction target temperature while being conveyed to.

The present invention is a method of controlling such a heating furnace, and a functional block diagram is shown in FIG.

Of these, the first performance processing function (first function) is activated at a predetermined cycle, and the plant performance φ [reactor temperature (gas temperature) in each zone of the heating furnace, the observed values of the fuel flow rate, and The material data including the material position of the material to be charged into the furnace] is input to calculate and change the temperature distribution value (current value) in the thickness direction of each zone wall and each material that cannot be directly observed. In addition to φ, φ ′ and an estimated value of an unknown parameter in the heating furnace simulator, which is an important process model in the present invention, in order to cope with the change in the characteristics of the heating furnace. Is calculated based on the plant performance φ ′, and the fuel flow rate set value calculation function (second function) 2 is started. Here, the sampling cycle is set equal to the simulation cycle Δt of the heating furnace simulator. Φ ′ updated by the first function, Is used for other functions.

Next, the second function is started from the first function, and the fuel flow rate set value F _SV of each combustion zone (zone in which fuel is burned) of the heating furnace is set to each zone furnace temperature, current value of material temperature, and target temperature rise of each material. Calculate from the curve and output to the plant.

Finally, the target temperature increase curve optimization function (third function) 3 is to divide the material into several groups in the extraction order, and to supply the target temperature increase curve of each material (take the same value in the same group). The value is calculated so that the value of the predetermined evaluation function becomes the minimum. The obtained value A ^* is used in the second function.

Here, the evaluation function is set according to the purpose of operation such as quality assurance and energy saving during material baking. The value of the evaluation function is obtained by simulating the operation of the heating furnace from the present time until the predetermined material is extracted using a heating furnace simulator that is a process model. This third
The function does not have to be activated in synchronization with the first and second functions, and may be activated independently in a predetermined cycle.

Hereinafter, a heating furnace simulator which is a basic process model used in the present invention, a furnace temperature prediction model used for controlling the upper and lower limits of the furnace temperature, a material temperature rising model used for calculating a fuel flow rate set value, a target temperature rising curve of the material , And the evaluation function used for optimization of the target temperature rise curve are described below.
The third function will be described in detail.

(1) Heating furnace simulator The heating furnace simulator is for simulating the operation of the heating furnace at a predetermined time interval Δt.
From the wall temperature, the fuel flow rate, the initial value (value at the initial time) regarding the material data, and the value of each zone fuel flow rate for each Δt after the initial time, the furnace temperature, the wall temperature, and the material for each Δt after the initial time Calculate the temperature.

If this initial time is taken as the current time, future operations can be simulated. Although the heating furnace simulator is a known technique, it plays an important role in the present invention and is also related to the successive parameter estimation described later, and therefore the processing means at time t will be described in 1) to 3). Also, in the following description,
Let Δt be the unit of time. (That is, Δt = 1) 1) Simulate the material conveyance that has occurred during the time [t-1, t]. Any material extracted during this period will be removed from the in-core material data. When there is a space above a predetermined interval on the charging side, a material to be charged (or a dummy material having a constant size and material) is charged.

2) For each zone furnace temperature at time t, the heat balance in the zone shown below is calculated from the fuel flow rate of the zone and the zone on the extraction side of the zone at time (t-1), the wall temperature of the zone, the material surface temperature Calculate using an equation.

Q _Ai + Q _Mi + Q _Gi-1 + Q _Gi + Q _Si −Q _Wi −Q _Li = 0 (1) Q _Wi = σ ・ φ _GW・ S _Wi・ (T _Fi ⁴ + T _Wi ⁴ ) …… (3) Q _Li ＝ λ _Ai・ T _Fi ＋ λ _Bi …… (4) where i: band index (i = 1, 1 2, ...; the strip at the extraction end is i
= 1); j: index of i-th zone material; Q _Ai : preheated air sensible heat; Q _Hi : input fuel calorific value; Q _Gi-1 : inflowing waste gas sensible heat (other than soaking zone); Q _Gi : Outflow waste gas sensible heat; Q _Si : Heat input to material; Q _Wi : Heat input to furnace wall; Q _Li : Heat loss and correction term; σ: Stefan Boltzmann constant; φ _GS : Overall heat transfer coefficient (gas ~ materials); φ _GW: overall heat transfer coefficient (gas-furnace wall); _{S j:} material surface seat; _{S w:} furnace wall area; T _Fi: i-th band oven temperature (gas temperature); T _Sij: the j Surface temperature of material; T _Wi : furnace wall temperature; λ _Ai , λ _Bi : heat loss coefficient, values are sequentially estimated using actual operation data (described later).

If the number of bands included in the heating furnace and n _z, (1) Equation heat balance equation because established in each band becomes total n _Z number, whereas unknowns also n _Z number present in each band furnace temperature To do. Therefore, the furnace temperature T _Fi can be obtained by using simultaneous heat balance equations, but since this equation is a quartic equation regarding T _{Fi, the} solution is obtained by the known Newton-Raphson method. Figure 3 illustrates the concept of heat balance in the belt.

3) The thickness direction distribution of each zone wall temperature at time t and the thickness direction distribution of each zone material temperature are calculated from the thickness direction temperature distribution at time (t-1) and the furnace temperature at time t as follows. The known heat conduction equation is obtained by using an equation discretized in the time and thickness directions.

The boundary condition at this time is Given in. Here, T: temperature (wall temperature or material temperature); t: time: x: coordinate in the thickness direction; α: thermal conductivity; K: boundary condition. Since a considerable amount of calculation is required to obtain the temperature distribution in the thickness direction from the above equation for all the materials in the furnace, the known weighted residual method is used to improve the calculation efficiency.

A simulation is performed by repeating the processes 1) to 3) described above by updating the time t one by one.

(2)) Furnace temperature prediction model In the present invention, the operation amount of control is the fuel flow rate, but in actual equipment, the upper and lower limits of each zone furnace temperature are determined from the viewpoint of equipment or material quality. Therefore, it is necessary to set the fuel flow rate set value so that the furnace temperature does not exceed the upper and lower limits, and for this reason, a model that represents the dynamic relationship between the fuel flow rate change and the furnace temperature change is required. Therefore, it was newly developed as a furnace temperature prediction model. The model formula is shown below.

However, ΔT _Fi (t) = T _Fi (t) −T _Fi (t−1) ΔF _i (t) = F _i (t) −F _i (t−1) where i: band index; T _Fi : furnace Temperature; F _i : Fuel flow rate; a _Fi , C _{Fk, i} (k = 0, 1, ..., m _F ): Appropriate constant (value is determined for each band from actual data); m _F : Appropriate positive integer (When Δt = 2 min, m _F = 2 is sufficient).

(3) Material temperature increase model In the present invention, it is necessary to obtain a fuel flow rate for each material such that the temperature of each material matches the target temperature increase curve of the material after a predetermined time. Therefore, a simple model that represents the dynamic relationship between the fuel flow rate change and the material temperature change is required, and was newly developed as a material temperature rise model. The model formula is shown below.

However, ΔT _si (t) = T _Sj (t) −T _Si (t−1) ΔF _n (t) = F _n (t) −F _n (t−1) i: band index; j: material index (first zone i); n: zone index; T _Sj : material temperature (thickness direction average value); S _Hj = material specific heat; t _Hi = material thickness; b _nj : to material j of n-th zone furnace temperature T _Fn A constant that is determined by the position of the material. 0 ≦ b _nj ≦ 1, And B _nj (n = i-1, i, i +) in the i-th band
An example of 1) is shown in FIG.

a _Si , c _{Sk, n} : constant (value is determined for each band from actual data) m _S : positive integer (m Δ = 2 min, approximately m _S = 2).

In the second term on the right side of the equation (6), when i = 1 (band at the extraction end) or i = n _Z (band at the charging end), n = 0 or n = n _Z +1 may be satisfied. However, in these cases, b _nj = c _{SK, n} = 0.

(4) Target temperature rising curve of material Set the target temperature rising curve for each material for the materials that have been charged into the furnace and the predetermined number of materials to be charged.

This curve represents the target value of the temperature rise of the material temperature (average temperature in the material) during the period from the charging of the material to the extraction, and is a function of time. The value of the target temperature rise curve at the time of material charging is equal to the actual value of the material charging temperature (average temperature in the material at the time of charging) (predicted value for the material to be charged), or in the vicinity of it. Value (If it is not necessary to raise the temperature immediately after charging, it is effective in saving energy to set the target value slightly lower than the charging temperature).

On the other hand, if the remaining in-reactor time until material extraction is less than or equal to the predetermined aging heat time, the target temperature rise curve is set in order to reduce the uneven heat distribution in the material (temperature difference in material, surface temperature-center temperature) during extraction. Make it equal to the extraction target temperature of the material. The value of this aging heat time is determined in consideration of the charging temperature of the material, rolling specifications and the like.

A plurality of straight lines or curves are continuously connected from the charging to the start of the heat-up to complete the target temperature rising curve. Some of the various parameters (straight line or curve gradient, heat-up time, etc.) that determine the target temperature rise curve have common values within the material group described below, and are set in the target temperature rise curve optimization process. Value that minimizes the evaluation function (optimum value)
Is required.

An example of the target temperature rising curve is shown in FIG. In this example, the heating furnace is from the charging side to the pre-tropical zone (PHz), the # 1 heating zone (# 1H
z), # 2 heating zone (# 2 Hz), and soaking zone (SZ). The zones other than PHz are combustion zones (fuel combustion zones). The target temperature rise curve is a quadratic curve for PHz and # 1Hz, a cubic curve for # 2Hz, and a straight line from SZ entry to the start of heat treatment. Connected by a factor).

(5) Evaluation function In the operation of the heating furnace, it is necessary to comprehensively optimize some objectives such as material baking quality assurance and energy saving. Therefore, the evaluation function J _RF is set by the following formula.

Here, i: band index; j: material index (a material existing between the extraction end and a predetermined position on the charging side is a target, and is referred to as an evaluation target material); All combustion zones (zones that burn fuel); Sum of all materials to be evaluated; From the present, the sum of the time until all the evaluation target materials are extracted; ΔT _SEj ; Extraction material temperature deviation (difference of the j-th material average temperature at the time of extraction from the extraction target temperature (average-target)); Material within temperature deviation (extraction time of the j material within temperature deviation (surface - _{center)); f 1, f 2} : appropriate function; _{F i:} fuel flow rate of the i _{band; w 1, w 2, W} 3: Common weighting factors for extraction temperature deviation, in-material heat excursion, and fuel flow rate; w _1Sj , w _2Sj ; material-specific weighting factor for j-th material The first to third terms of the evaluation function are the extraction temperatures of the materials, respectively. It shows the evaluation of deviation, heat deviation in material, and fuel flow rate.

The common weighting coefficient is set to an appropriate value according to the operation purpose (material baking quality or energy saving is important), and the material-specific weighting coefficient is, for example, whether the j-th material needs to have a smaller internal bias than other materials. The value is determined for each material, considering whether or not it is necessary.

Functions f ₁ and f ₂ are | ΔT _SEj |, In order to prevent the value of p from increasing, the function value increases as these values increase.

This evaluation function is used for target temperature increase curve optimization (third function) described later.

(6) Performance processing function (first function) Simulation cycle of heating furnace simulator Δt (=
It is activated in the same cycle as 1) and performs the following processes 1) to 3).

1) Update the temperature distribution in the thickness direction of each strip wall temperature and each material temperature.
From the previous value of the temperature distribution and the actual value of each zone temperature, the temperature distribution is updated using Equation (5) of the heating furnace simulator to obtain the current value.

2) Sequential parameter estimation Heat loss coefficient λ _Ai , λ _Bi (i
The band index) is sequentially estimated using the performance data each time performance data is obtained in a predetermined cycle. The formula for updating the estimated value at time t is shown below.

Where Q _Li (t) = Q _Ai (t) = Q _Hi (t) + Q _Gi-1 (t) + Q _Gi (t) -Q _Si (t) -Q _Wi (t) ...... (10) x _i (t) = [F _i (t), 1] ^T However, the unit of t is the estimated value change period, Each of the estimated values of λ _Ai and λ _Bi at time t, F _i
Is the i-th zone fuel flow rate, Q _Li (t) is the actual loss heat value at time t, and the value on the right side of equation (10) obtained by converting equation (1) is used as the actual data at time t. It can be calculated.

On the other hand, k (t) is a parameter correction gain vector (two-dimensional) with respect to the estimation error of Q _Li (value in {} of the equation (9)), and the time t is determined by a known method such as the recursive least squares method.
Find the value at. With this parameter sequential estimation function, the heat loss coefficient reflecting the actual data is always obtained, and the accuracy of the heating furnace simulator is maintained even if the characteristics of the furnace change. After this processing ends, the second function is activated.

3) Activate function 2.

(7) Fuel flow rate set value calculation / output function (second function) A schematic flow chart of this function is shown in FIG. 6, and the processing procedure according to the figure is described in 1) to 5) below.

1) Initialize the simulation time t to the current time t ₀ .

2) In each fuel zone in the furnace, the fuel flow rate set value F _i (t) of the zone is calculated in order from the extraction side by the procedure of a) to d) below.

a) The material temperature (average temperature) T after the predetermined time N ₃ for the material existing in the extraction side predetermined range of the band (the i-th band) and the charging side band (the (i + 1) band).
_Si (t + N ₃ ) is the target temperature rise curve value T _Sj ^* (t
The i-th zone fuel flow rate set value F _{Si, j} (t) for the material, which is required to substantially match + N ₃ , is obtained (j is a material index). Here, the (i + 1) th zone extraction side slab is also included because the i-th zone fuel flow rate also affects the temperature rise of the material at this position.

The values soaking zone is N _3, since it is highly necessary to match as soon as possible material temperature at the target, and a small value, in other band, optimization and or scheduled extracted pitch operation on the target Atsushi Nobori curve will be described later In order to avoid large fluctuations in the fuel flow rate, which is the manipulated variable, when the target value fluctuates due to fluctuations, etc., it is set larger than the soaking zone. Also, from T _Sj (t) to T _Sj ^*
The target trajectory of the material temperature change during (t + N ₃ ) is represented by a straight line connecting the two as shown in FIG. 7, thereby suppressing a large change in the fuel flow rate. Although various methods are conceivable for calculating F _{Si, j} (t), it is generally effective to satisfy the slab temperature tracking performance to the target orbit and the control stability to prevent the excessive fluctuation of the fuel flow rate. Prediction control (Genera
lized Predictive Control). Using this method, the manipulated variable u (t) (= F _{Si, j} (t)) that minimizes the following function J _FL is calculated by using the material heating model (7).
It can be obtained by using the formula).

However, y (t + k) = T _Sj (t + k) (k = N _1, N ₁ +1, ..., N ₂ ) Δu (t + k) = u (t + k) −u (t + k -1) (k = 0, 1, ..., N _U-1 ) Here, F _i (t-1): i-th heat amount setting value at time (t-1) λ: operation amount fluctuation weighting factor (λ ≧ 0) N _1, N _2, N _U : positive integer (N ₁ ≧ N ₂ ) y _r : Target trajectory of material temperature (y) The first term of the evaluation function J _FL represents the deviation of the material temperature from the target trajectory y _r , and the second term represents the fluctuation of the fuel flow rate. Therefore, control is performed so as to reduce both of them. In order to ensure the stability of control, N ₂ is set to a value of 2 or more when the control cycle is 2 minutes. The value of N _U is 1
Then the calculation becomes easier. A specific calculation procedure for _{obtaining the} manipulated variable F _{Si, j} (t) is publicly known and will be omitted.

b) The i-th zone fuel flow rate set value F _i (t) is the predetermined time (N ₃ )
However, each future material temperature is determined so as not to fall below the target temperature rising curve, or each material temperature is determined to be the value of the target temperature rising curve on an average or preferentially with respect to a specific material. The calculation formula of F _i (t) in the former case is represented by the following formula.

Here, w _{Li, j} represents the weight for the i-th band of the j-th material, and the value is determined as shown in FIG. 8 depending on the position of the j-th material in the furnace. In the figure, d _{k, i} (k = 1, ..., 5) is an appropriate constant determined for each band.

c) The i-th zone fuel flow rate set value F _i (t) is modified so that the i-th zone furnace temperature does not exceed a predetermined upper or lower limit value determined in terms of equipment or material quality.

For this purpose, first, the furnace temperature prediction model (6) is used to predict the future furnace temperature at the time point d from the furnace temperature up to the present, the actual fuel flow rate value and F _{Si, j} (t). When the value exceeds the upper or lower limit of the furnace temperature, the prediction model again
Correct the fuel flow rate set value so that the furnace temperature matches the upper limit value or the lower limit value.

d) If the fuel flow rate set value after the correction in the above c) exceeds the predetermined upper and lower limit values of the equipment, the correction is performed so that it does not exceed the upper and lower limit values.

3) If the simulation time t is a predetermined time (t + N ₄ ) for completing the simulation, proceed to 5). In other cases, increment t by 1 and proceed to step 4).

4) The furnace temperature at time t of each zone from the extraction side to the charging side, the temperature distribution in the thickness direction of each zone wall temperature and each material temperature, and these previous values at time (t-1) and 2) From each zone fuel flow rate obtained in step 1, the value is obtained using a heating furnace simulator, and the process returns to step 2).

5) Fuel flow rate set value F actually output in each fuel zone
_{From SVi} (t ₀ ) and the set value F _i (t + k) (k = 0, 1, ..., N ₄ ) obtained by the actual fuel flow rate actual value and 2), a large fluctuation of the set value is suppressed. Calculated by appropriate filtering (smoothing) and output to the fuel flow controller.

(8) Target heating curve optimization function (third function) A schematic flow chart of this function is shown in FIG. 9, and the processing procedure according to the drawing will be described in detail in 1) to 2).

1) Optimization pretreatment a) Material group division: In a predetermined cycle, the material in the furnace and the material to be charged into the furnace are searched in order from the extraction side in the furnace length direction, and similar to the following division criteria. The above materials are divided into the same group as much as possible, and the whole is divided into several material groups.

Group division criteria: i) to iv) below between adjacent materials
When is established, the group division is performed with that as the boundary of the material group.

i) The time spent in each zone differs greatly.

ii) The charging temperatures are very different.

iii) The extraction target temperatures are significantly different. iv) The group length exceeds the prescribed length.

b) Initialization of common parameters The values of supply parameters (see the above description of the target temperature rising curve) that determine the target temperature rising curves of the respective materials are initialized for each material group. Here, when the control is started up, a predetermined initial value is set equally for all groups, but in other cases, it is determined by the averaging process described below.

Here, i is a group index, j is a material index in the i-th group, a _xi is a common parameter a _x , an initial setting value (a desired value) in the i-th group, a ′ _xij
Is the current value of the common parameter for the j-th slab in the i-th group (that is, the value obtained in the previous optimization process), Indicates relaxation related to materials in the i-th group (this time i-th group)
When the material belonging to the group was divided into two or more groups last time, the value of a ′ _xij varies depending on j).

2) Search for optimum value of common parameter Here, the value of the evaluation function of Eq. (8) is the minimum in the common parameter space (that is, the optimum value of common parameter).
Is searched using the steepest descent method, which is a known method for nonlinear optimization. To obtain the value of the evaluation function, use a heating furnace simulator until the future time when the predetermined material of the charging completion or the planned charging material (for example, the material on the charging side from the fuel zone inlet by a predetermined distance) is extracted. To simulate the operation. This simulation is the same as the simulation for calculating the flow rate set value shown in FIG. The search consists of the following direction search and line search.

a) Direction search In the common parameter space, the direction (steepest descent method) in which the gradient in which the value of the evaluation function decreases decreases is obtained. For this purpose, it is necessary to obtain the gradient of the evaluation relation value change in each direction in the space, which is to move the common parameter value in that direction by a predetermined minute amount to obtain the evaluation function value at that point,
This is done by calculating the amount of change in the value from the original position.

When the steepest descent direction is found as a result of the direction search, the following b) is performed. If not obtained, the current point is considered to be the optimum value (that is, the optimum value A ^* ) that minimizes the evaluation function value, and the processing is ended.

b) Straight line search If the steepest descent direction is obtained as a result of the direction search, the point at which the evaluation function value becomes the minimum in that direction is obtained. For this search, a known method such as Armijo method is used,
Detailed explanation is omitted. After the minimum point of the evaluation function is obtained, the direction search and the straight line search are repeated until the optimum point is obtained.

〔Example〕

Hereinafter, a simulation result using actual operation data will be described as an example of the present invention.

The target plant is a continuous heating furnace for billet slabs at a steel mill continuous hot rolling plant. As for the heating furnace, the furnace bottom zone, the first heating zone (# 1Hz), the second heating zone (# 2H)
z), the third heating zone (# 3 Hz), and the soaking zone (Sz). The extraction side is the combustion zone from # 2 Hz. 10 and 11 show the results of actual operation of this heating furnace by the conventional method and the results of the simulation of the operation of the same material as the actual operation using the present invention.

First, Fig. 10 shows the material temperature during extraction (average temperature in the material).
The comparison of the temporal transition of is shown together with the extraction target temperature. The material targeted here is ordinary steel having a thickness of about 250 mm, a width of 760 to 1460 mm, and a charging temperature of 50 to 900 ° C. The target extraction temperature is from 1100 ℃ to 12 ℃ as shown in the figure.
It changes greatly up to 30 ℃. It is desirable that the extraction temperature of each material does not fall below this target temperature.

However, as can be seen from the figure, in the actual operation result by the conventional method, there is insufficient baking of the material after the extraction target temperature rises, where the extraction temperature is below the target temperature.
On the other hand, in the simulation result according to the present invention,
Insufficient baking does not occur, and when the extraction target temperature falls, the extraction temperature falls more quickly than in the conventional method, preventing the material from being burnt.

Next, FIG. 11 shows a comparison of the transition of the uneven heat distribution in the material during extraction. It is not preferable that the unbalanced value has a large positive value, and several peaks of unbalanced temperature appear in the conventional method, but in the present invention, the value of each peak is small.

On the other hand, regarding the fuel flow rate in this embodiment, the flow rate of the present invention is reduced by about 10% as compared with the conventional method. In the present invention, the flow rate for each zone was greatly reduced in the second heating zone on the charging side, decreased slightly in the soaking zone, and increased in the third heating zone, which reduces waste gas loss and during extraction. It also serves the purpose of suppressing uneven heat distribution in the material.

As described above, in the present embodiment, the present invention has shown preferable results over the conventional operation in all of the extraction temperature of the material, the in-material heat deviation during the extraction, and the fuel flow rate.

〔The invention's effect〕

As described above, according to the present invention, the future operation of the heating furnace is simulated using the heating furnace simulator including the non-linear equation, so that the value of the predetermined evaluation function can be minimized for each material. The target temperature rise curve is optimized online, and the unknown parameter values in the furnace simulator are successively estimated online based on the actual data, which enables highly accurate operation optimization. It is possible to obtain effects such as securing a material extraction temperature, suppressing uneven heat distribution in the material at the time of extraction, and reducing the fuel flow rate.

[Brief description of drawings]

FIG. 1 is a functional block diagram of the present invention, FIG. 2 is a diagram showing an example of a target plant, FIG. 3 is a diagram showing heat balance in a zone, and FIG. 4 is a diagram showing the degree of influence of furnace temperature on materials. FIG. 5 is a diagram showing a target temperature rise curve of the material, FIG. 6 is a flow chart of fuel flow rate setting value calculation, FIG. 7 is a diagram showing a target trajectory of the material temperature, and FIG. 8 is depending on the position of the material. FIG. 9 is a diagram showing weighting factors, FIG. 9 is a flow chart of target temperature rising curve optimization processing, FIG. 10 is a transition diagram of material extraction temperature in the embodiment, and FIG. 11 is transition of in-material unbalanced heat during extraction in the embodiment. Each figure is shown.

Claims (6)

[Claims] 1. A combustion control method for a continuous heating furnace having a plurality of furnace zones, wherein a plant performance (φ) is input from the heating furnace at a predetermined cycle,
Calculated and updated the temperature distribution actual values of each zone furnace wall and each material, and added these actual values to the above-mentioned plant actual, and the corrected plant actual (φ '), and the heating obtained based on the corrected actual plant. The first function, which is a performance processing function that outputs the estimated value (θ) of the unknown parameter in the reactor simulator to the second and third functions, and the i-th band fuel at time t that is activated following the first function The flow rate setting value (Fi (t)) is set so that each material temperature in the zone in the future and for a predetermined time (N ₃ ) does not fall below each target temperature rising curve, or each material temperature is averaged. Alternatively, the second function, which is a function for calculating and outputting the fuel flow rate set value, which is set so that the value of the target temperature rising curve is preferentially set for the specific material, and is output to the heating furnace, and either continuously or independently of the second function. The heating condition is similar because it is started in a predetermined cycle The values of the common parameters of the target temperature rising curve of each material, which have the same value in the same group, are collected into a group, and the value of the evaluation function having the evaluation terms for the extraction temperature deviation, the material internal heat deviation, and the fuel flow rate is calculated. A fuel for a continuous heating furnace, which is provided with a third function that is a target heating curve optimization function that obtains the minimum value and outputs the obtained optimum value (A ^* ) to the second function. Control method. 2. In the calculation of each zone fuel flow rate set value in the second function according to claim 1, it is necessary to set the temperature of each material in the furnace to a value on the target temperature rise curve for a predetermined time in the future. The value of the fuel flow rate for each material
Obtained based on a material temperature rise model that represents the dynamic relationship between changes in fuel flow rate and changes in material temperature, and then calculate the maximum or weighted average value of the fuel flow rate for each material within the same zone. Characterized by setting the flow rate,
Fuel control method for continuous heating furnace. 3. In the calculation of each zone fuel flow rate set value in the second function according to claim 1, a change in each zone furnace temperature due to a change in each zone fuel flow rate set value from present to future is calculated as a fuel flow rate change and a furnace Prediction is performed using a furnace temperature prediction model that represents dynamic changes with temperature changes, and when it is predicted that the furnace temperature will exceed the upper and lower limits set for equipment or material quality, the furnace temperature is set to the upper limit or A combustion control method for a continuous heating furnace, characterized in that the fuel flow rate set value is corrected using the furnace temperature prediction model so as to be equal to the lower limit value. 4. The target temperature rising curve of each material according to claim 1 is set as a function of the remaining in-reactor time, and at the time of charging the material, the value is set to a charging temperature or a value in the vicinity of the charging temperature. When the remaining in-reactor time is less than a predetermined aging heat time, it is made equal to the extraction target temperature of the material, and each zone heating rate is set as one of the parameters from the charging to the start of aging heat. A combustion control method for a continuous heating furnace, characterized in that a plurality of straight lines or curves are continuously connected, and the target temperature rising curve is corrected by correcting the values of the above parameters. 5. The third function according to claim 1, wherein in the heating furnace, several similar and continuous judgments are made based on a division criterion including each zone time, charging temperature, and extraction target temperature. A method for controlling combustion in a continuous heating furnace, characterized in that the materials are grouped in the same group, and some of the various parameters that define the target temperature rise curve for each material are set to common values within the same group. 6. The heat loss actual value according to the first function of claim 1, wherein an unknown parameter value that is a heat loss coefficient in a heating furnace simulator that is a process model used for simulating a heating furnace operation is obtained. Was calculated from the heat balance formula in the belt, and using the difference between this actual value and the estimated loss of heat calculated from the previously estimated loss of heat coefficient and the fuel flow rate,
Characterized in that the unknown parameter is sequentially estimated,
Combustion control method for continuous heating furnace.