JPH02294419A - Method for controlling furnace heat in blast furnace - Google Patents

Abstract

PURPOSE:To stabilize blast furnace operation and to control furnace heat to the desired temp. by predicting variations of the furnace heat for a short period of time and for a long period of time in the blast furnace operation, obtaining the overall furnace heat level based on these, selecting operational condition for stabilizing the furnace heat and setting the operational condition according to this. CONSTITUTION:Plural thermometers are set in height direction and circumferential direction in the blast furnace and at first, as hot blast in the circumferential part in the furnace is caused mush to flow for the short time, the variation of the furnace heat for the short period of time caused by reduction endothermic reaction of FeO at the circumferential part in the furnace is predicted. Successively, based on lowering of raw ore level caused by increase of level lowering velocity of the raw material in the furnace, the variation of the furnace heat for the long period of time is predicted. Based on these first and second predicted results of the furnace heats, each furnace heat level is obtd. with member-ship function and the overall furnace heat level is obtd. with fuzzy inference from the present first rule. These furnace heat levels and the operational data showing the present operation condition are collated with the present second rule and inference is preformed, and by this method, the operation action for stabilizing the furnace condition is selected, and by operating according to this, the furnace condition in the blast furnace is stabilized and the furnace heat is controlled to the desired temp.

Description

DETAILED DESCRIPTION OF THE INVENTION (Industrial Application Field) The present invention relates to a blast furnace heat control method for stable operation of a blast furnace.

(Prior art) As a blast furnace furnace heat control system that estimates the temperature of hot metal in a blast furnace and manages and controls it, for example, Japanese Patent Application Laid-Open No. 62-27
There is a system disclosed in Publication No. 0708.

This blast furnace furnace heat control system includes a data input means that takes in data from various sensors installed in the blast furnace at a predetermined timing, and an unloading speed that is determined based on the data from the sensors. Means for creating various data indicating the status of the blast furnace such as pressure loss, shaft pressure, shaft temperature, fixed sonde temperature, gas utilization rate, furnace rotor sonde temperature, etc., and comparing the various data with reference data to determine whether it is true or false. a storage means for temporarily storing authenticity data; a knowledge base means storing various knowledge bases based on experience and achievements regarding blast furnaces; and authenticity data of the storage means and the knowledge. The inference means infers the furnace heat level and the furnace heat transition based on the knowledge base of the base means and determines the amount of action for the blast furnace.

According to this system, truth data is created from the data of various sensors installed in the blast furnace, and predetermined inferences as artificial intelligence are made using this data and a knowledge base based on experience, etc. stored in the knowledge base means. Therefore, the existing experience can be fully utilized, and even if the system is implemented on a computer, its capacity will be extremely small. Furthermore, even when the blast furnace changes over time, it is only necessary to change the memory contents of the knowledge base means, and the change is extremely easy.

(Problems to be Solved by the Invention) However, the furnace heat control method using the blast furnace furnace heat control system described above has the following problems.

■ Since the truth data used for inference is obtained by comparing the standard data and sensor data, if the sensor data fluctuates near the standard data, the truth data values will become completely opposite, and the furnace heat will change. This leads to a decrease in the accuracy of change inference.

- Also, since the reference data itself is a fixed value, it cannot respond to changes in the average value of sensor data (level changes) due to changes in the properties of raw materials in the furnace. This also leads to a decrease in the accuracy of inferring changes in furnace heat.

■ Since there is no means to select the operational action to be executed, the operational action performed by this blast furnace furnace heat control system is fixed (blow humidity), and it cannot be said that the optimal operational axis has been performed.

■ Since the amount of operational action is determined by referring to the action table in the knowledge base, it is difficult to determine the amount of action at a detailed level.

This invention was made in order to solve the above-mentioned problems ① to ②. It predicts furnace heat changes with high precision and quantitatively, and adjusts the optimum operating acillin to the optimum control amount in response to the furnace heat changes. The purpose of this study is to obtain a method for controlling blast furnace heat that can be carried out in the following manner.

(Means for Solving the Problems) A blast furnace furnace heat control method according to the present invention incorporates at least one piece of blast furnace data indicating the internal condition of the blast furnace, and compares the blast furnace data with a reference value to control the furnace heat. a first step of outputting a first prediction result that predicts a short-term change; a second step of importing one blast furnace data and hot metal temperature and outputting a second prediction result predicting a long-term change in furnace heat by adding the influence of an operation action to these data; The furnace heat level of each of the first and second prediction results is determined by the membership function, and the total furnace heat level is determined by fuzzy inference from these furnace heat levels and the predetermined first rule.
By comparing the operation data indicating the overall furnace heat level and the current operating status with a predetermined second rule and making inferences, it is possible to determine the operational actions that should be taken to stabilize the furnace heat. a fourth step of selecting a certain execution operation action; and a fifth step of determining an operation action amount, which is a control amount of the execution operation action, using a function specific to the execution operation action using the overall furnace heat level as an argument. , a sixth step of executing the execution operation action according to the operation action amount.

(Operation) In the first step of the present invention, since the reference value is changed from moment to moment according to the blast furnace data for a predetermined period in the past, it is possible to respond to changes in the level of the blast furnace data.

In addition, in the third step, the furnace heat levels for each of the first and second prediction results are obtained using a membership function, and based on the predetermined first rule and each furnace heat level, fuzzy inference is performed. Since the overall furnace heat level is being determined, a quantitative overall furnace heat level is obtained by adding the second prediction result, which predicts long-term furnace heat changes, to the first prediction result, which predicts short-term furnace heat changes. be able to.

(Example) The following may be considered as a cause of the decrease in furnace heat of a blast furnace.

High-temperature air (gas flow) blown up from the blast furnace lining to adjust the temperature and amount of hot metal is usually blown into the center of the furnace. However, due to reasons such as raw material charging conditions and charge distribution, a large amount of gas may suddenly flow to the periphery of the furnace.

As a result, Fe　O+C　−　Fe　+CO The endothermic reaction of is promoted and the furnace heat decreases.

By the way, when a large amount of gas flows around the furnace, Na
, K, Pb, etc., and the stagnation layer peel off and fall off the wall, causing the temperature of the furnace wall in that area to rise rapidly. If this rapid temperature rise is detected, a decrease in furnace heat can be predicted.

(A-2) Second Reason for Decrease in Furnace Heat The following may be considered as a cause of the decrease in furnace heat in the blast furnace.

When the unloading speed in the blast furnace increases for the same reason as (A-1), the level of the cohesive zone in the blast furnace decreases due to so-called raw ore unloading, resulting in a decrease in furnace heat.

By the way, when the cohesive zone level decreases, the furnace wall temperature at the corresponding portion also decreases rapidly. If this rapid temperature drop is detected, a decrease in furnace heat can be predicted.

(A-3) Third Reason for Decrease in Furnace Heat Furthermore, the following may be considered as a contributing factor to the decrease in furnace heat in the blast furnace.

High-temperature air (gas flow) blown up from the blast furnace lining to adjust the temperature and amount of hot metal is usually blown into the center of the furnace. However, (A-1). For the same reason as (A-2), part of the gas flow may flow to the periphery of the furnace. If this state continues for a long time, heat dissipation from the gas flow from the blast furnace wall becomes greater than during normal operation, resulting in a decrease in furnace heat.

By the way, when a part of the gas flow steadily flows into the periphery of the furnace, the temperature of the furnace wall gradually increases. If such a slow temperature rise over a relatively long period of time is detected, a decrease in furnace heat can be predicted.
I can.

(A-4) 1st to 3rd Preparation [1 Means Figure 1 (a)
．． (b) is a side sectional view and a plan sectional view, respectively, showing the arrangement of an inner wall thermometer used in an embodiment of the present invention.

There are seven inner wall thermometers 3 in the height direction of the blast furnace 1 (three on the back, two on the abdomen, and two on the morning glory), as shown in FIG. Install 4 pieces in the circumferential direction. In other words, a total of 28 inner wall thermometers 3 are installed in four directions and seven levels.

The inner wall thermometer is, for example, disclosed in Japanese Utility Model Application No. 57-81 by the present applicant.
531, Utility Model Publication No. 59-16816 may be used, and FIG. 2 is a conceptual diagram showing an inner wall thermometer (hereinafter referred to as "rFM sensor") disclosed in the latter.

In the figure, reference numeral 4 denotes a sheath type thermometer having two conductive wires 5 buried in parallel insulatively and having a temperature sensing part 6 on the front end side. The temperature-sensing parts 6 are arranged in parallel so that they are arranged at different lengthwise positions, and a sheath-type dummy rod 7 is connected to the tips of the temperature-sensing parts 6 to align the leading ends. The sheath type dummy rod 7 has two conductive wires 5 buried in parallel insulating manner, and the sheath type 4-1
It has substantially the same thermal conductivity as the hot body 4. The FM sensor 3 is formed by keeping the sheath type n1 hot body 4 in a non-contact manner with an insulating material 8 and housing it in a sheath tube 9.

FIG. 3 is an explanatory diagram of the installation of the FM sensor 3. In the figure, 10 to 13 are the walls of the blast furnace, 10 is a brick, and 1
1 is a stave, 12 is a stamp, and 13 is an iron skin. F
As shown in the figure, the M sensor 3 is installed inside the furnace wall by welding to a packing 14 and a welded portion 15. Note that 16 is a filler, and 17 is a milk injection port into which the filler 16 is poured.

Note that due to the installation and structure of the FM sensor 3 described here, the FM sensor 3 itself is eroded along with the erosion of the furnace wall, and the sheath type temperature sensing element 4 may be exposed inside the furnace near the furnace wall. In addition to "furnace wall temperature", "furnace temperature near the furnace wall"
This means that 1111 are set. Hereinafter, a concept including both will be described as "furnace wall temperature." As mentioned above, compared to conventional thermometers such as sheathed thermocouples, the FM sensor 3 has a large number of 1113 fixed points, satisfies rapid temperature measurement response, is capable of long-term continuous temperature measurement, and is highly reliable. Efforts are being made to improve performance, durability, and workability.

Each FM sensor 3 measures the inner wall temperature of the blast furnace 1 at every predetermined sampling time Δt, as shown in FIG.

Here, let T be the inner wall temperature of the i-th FM sensor 3 at time j, and let 1 sample j. l If the inner wall temperature before the ring time Δt is ".+-t.t," then the inner wall temperature difference (difference value) ΔT between T and T J.l j-1.1
is, J. 1 ΔT −T −T. J, i J. IJ-1.1-(1)
becomes. This state is shown in FIG.

This difference value ΔT j,t is multiplied by a weight wi in consideration of the height, circumferential direction, etc. of each FM sensor 3. Furthermore, the difference value ΔT. For negative values, Vtj・l is 10, and for other values, the positive/negative parameter Vt indicating v1-1 is also multiplied to obtain the corrected difference value (positive difference value) CT at time j. .

j. ! C T j, 1 strict wt ” %' t −Δ”j.
t - (2) Next, the total FM of the corrected difference value CT
The sum of j and l for the sensor 3 is taken, and this is set as S T 1 i.

STI, -mo CTJ. 1-(3) and according to the following equation (4), if the value of the total difference value sT1j becomes larger than the predetermined threshold value ε1, it is assumed that there has been a rapid temperature rise, and an evaluation point of 1 is given for the predetermined period.

S T 1 j. ≧81...
(4) The above is the first prediction means based on the reason ('A-1).

The second prediction means based on the reason (A-2) is as shown below.

(2) In formula (2), the positive and negative parameters Vi are the difference values ΔT
is positive, v t −Q−, that j. For values other than 1, it is set to V 1 - 1, and then the corrected difference value C T j. The total sum of the absolute values of s for all FM sensors 3 is taken and this is set as ST2j.

ST2 − Σ IcT l
...(3)゜j ・、．． j・l Then, according to the following equation (4), if the value of the sum of the difference values ST2j based on the equation (3) is larger than the predetermined threshold ε2, it is assumed that there has been a sudden temperature drop due to the descent of the raw ore, An evaluation point of 1 will be given for a predetermined period of time.

ST2・≧ε ...(4)゜J
2 C. The third preliminary 1N-1 means based on the reason is as shown below.

As in the first prediction means, the positive/negative parameter v1 in equation (2) is vl-0,
For other cases, it is set to v,-1. Also time j
Let cTj-k,1 be the corrected difference value k sampling times before (that is, Δt
For M sensor 3...(3)'' and then (4)'', this moving average sum ST3
If the value of j becomes larger than a predetermined threshold value ε3, it is assumed that there has been a gradual temperature rise for a long period of time, and an evaluation point of 1 is given for the predetermined period.

ST3j≧ε3 ...(4)゜Since the first to third prediction means described above are each performed using the furnace wall temperature difference (difference value), they are accurate regardless of whether the absolute value of the furnace wall temperature is above or below. predictions can be made. Shigamo, F
The M sensor 3 can be arranged to cover the entire circumference of the blast furnace due to its good workability and good temperature measurement response, and by being able to grasp continuous inner wall temperature differences, more accurate predictions can be made.

(A-5) Threshold learning Each threshold value ε (any of εl.ε2.ε3, hereinafter collectively referred to as "ε") of the first to third pre-71IJ means* which are furnace wall temperature prediction means ) is the optimum furnace heat drop prediction 1
1? This is determined in advance so that a prediction of J or furnace heat rise can be obtained. However, the initially optimal threshold value ε1 may become non-optimal due to changes in various conditions such as the production plan and raw material conditions during blast furnace operation. Therefore, it is necessary to change the threshold value ε1 to an optimal value during blast furnace operation, and learning of the threshold value ε8 is performed as shown below.

When measuring the furnace wall temperature of a blast furnace, FM, which is a furnace wall thermometer,
As described in (A-4), the temperature point of the sensor 3 changes with the erosion of the furnace wall. For this reason, not only the value of the measured furnace wall temperature itself but also the peak value (maximum value) of the furnace wall temperature difference, that is, the sensor sensitivity, may vary depending on the erosion of the furnace wall. Since the first to third prediction means make predictions based on the furnace wall temperature difference, in such a case, the prediction δ
Ill accuracy will also vary.

In order to solve this problem, the following processing is performed. First, going backwards from the current time t to the past S over a time period h, the average values STI to ST3 and the standard j deviations σ to σ3 of each sum STS 1 to ST3j in this interval h are calculated, and the following (5) to (7) ), determine threshold values ε to ε3.

l 11 ●ST1+bl ・σl ...(5)2
2φST2+b2 Hayabusa σ2 ... (6) s s
S T 3 + b a φσ3 (7) Note that a - a and b - b3 are coefficients.

tat In this way, by automatically determining the threshold value ε1 based on changes in each sampling data in the first to third pre-AI1 means at the appropriate time, various factors such as production plans and raw material conditions can be adjusted during blast furnace operation. Predictions can always be made with high accuracy regardless of fluctuations in sensor sensitivity due to changes in conditions.

(A-6) Comprehensive prediction If an evaluation point of 1 is given by any of the above-mentioned first to third prediction means, 1 is given, otherwise 0 is given as the prediction result of the furnace wall temperature prediction means IIl. Output as M.

Note that the following abnormal value correction processing may be performed. That is, when predicting by the first to third prediction means, T
．． When (T) satisfies the following formulas (8) to (lJ.i j-1.1 0), T,, -T,■,, > AT -(8
)T. ＞T...(
9) J. l lIaX T . <T,,n
... (lO)3.1 (However, ΔT, T and T are predetermined and I
Iax a+In is the threshold value. )T. T. Replaced with j.
t J-1.1 Correct abnormal values. Thereby, prediction accuracy can be further improved.

B. Calculate Sol loss C (solution inlet carbon) ffi (kg/t-p) at each sampling time Δt based on furnace top gas component analysis using gas chromatography, ventilation conditions, raw material loading conditions, etc. do.

Here, the amount of Sol loss C at time j is expressed as X. and J is k sampling time before time j (i.e., ΔtXk
time), the amount of Sol loss C is Then, J−k The predetermined time width nΔt at the current time j, ilΔ[
The moving averages x, x i' M of (J!>n) can be calculated using the equation.

(lla). Moving average x based on the (flb) formula. x
i' MH is calculated for each sampling time Δt, and the following (
12》 formula, 1 -X)4 -X IM・=(12)X The prediction result l by the fourth prediction means is obtained. An increase in the amount of SolX loss C indicates that the Solloss reaction, which is an endothermic reaction, is promoted. Therefore, if I>oX, the furnace heat will decrease, and if I<Q, the furnace heat will increase. It can be said that

Note that when calculating the moving average of the Sol loss C amount, since the instantaneous value of the Sol loss C amount may generate an abnormal value due to noise or the like, the following abnormal value correction processing may be performed. That is, if the Sol loss C amount at time j is x, and the Sol loss C amount one sampling time Δt before is X, then the absolute value of the Sol loss C difference J-1 minute value ΔX. is (lx −x l) J
J J-1 Ru. An abnormal value is found by comparing this Δxj with a preset threshold value ε2, and when an abnormal value is detected, the previous moving average value is replaced with the current moving average value. By applying this method, it is possible to make a more accurate prediction.

C. Gas chromatography N2 prediction means In addition, the amount of nitrogen (%) in the top gas detected by gas chromatography (hereinafter referred to as "gas chromatography N2 amount") has a strong negative correlation with the amount of Sol loss C.
Instead of increasing the amount, by increasing or decreasing the gas chromatin N2 amount,
Blast furnace heat changes can be predicted.

As a result, if the gas chromatography N2 amount at the current time j is yj, and the gas chromatography Nu at k sampling times before time j (that is, ΔtXk hours before) is y, then 2 j−k predetermined time width at the current time j Each moving average y.nΔt, itΔt (1)>n). yIM can be calculated with H.

(13a). Based on formula (13b) < Y , Y J2
Calculate each ring time Δt, and use formula (l4) below to calculate 1 ” Y H Y I M
...(l4)y Obtain the prediction result I by the fifth prediction means. If y, ■, > 0, a rise in furnace heat is predicted, and if I, <Q, a decrease in furnace heat is predicted.

Note that the abnormal value correction process used in the Sol loss C prediction means is
It is of course possible to apply the present invention to gas chromatography N2 prediction means.

B. C. Since the two prediction n1 means described above are based on a moving average for each sampling time Δt, predictions can be obtained quickly and the prediction accuracy can be said to be sufficiently reliable.

D. Charging Pitch Prediction Means It is empirically known that the number of times raw material is charged from the top of the furnace per predetermined period of time (hereinafter referred to as "charging bit") affects the furnace heat. In other words, it is known that when the charging pitch increases, the furnace heat decreases, and when the charging pitch decreases, the furnace heat increases.

This is presumed to be mainly due to the following reasons.

■ When the charging pitch increases (decreases), the Solloss reaction, which is an endothermic reaction, is promoted (suppressed), and as a result, the furnace heat decreases (increases).

■ As the charging pitch increases (decreases), the amount of pig iron tapped increases (decreases) and the time from material charging to tapping becomes shorter (longer).

As a result, the time for the hot metal to exchange heat in the blast furnace decreases (increases), so the furnace heat decreases (increases).

Therefore, the prediction result I by the charging pitch prediction means is obtained from the following equation (l5).

p However, ? ITCHN1: Number of times of raw material charging in the past 11 minutes PI TCHil2: Number of times of raw material charging in the past 12 minutes (Evening, 9■) In other words, if Ip>0, is it predicted that the furnace heat will decrease? IL,,
■, If it is <0, it means that a rise in furnace heat is predicted.

In addition, it can be said that the pre-n1 means described in B-Y-D basically performs pre-11P1 by comparing one piece of blast furnace data indicating the inside situation of the blast furnace with reference data.

Standard data ε * x J! M, yρM, P
Since ITCHN/ρ2 is set sequentially according to changes in blast furnace data, the above A, to D. The pren1 means described above can maintain a constant furnace heat estimation accuracy even if there is a change in the heat balance due to a change in sensor sensitivity or a change in raw material properties.

E. Long-term outlook prediction method A. ~D. All of the prediction means described above were means for predicting changes in furnace heat from a relatively short-term outlook based on the current blast furnace situation, but the prediction means described below are only for a long-term outlook.

Long-term outlook prediction is performed based on the following equation (16).

(t+Δt') −a ● (XN (t)−XM (t))+b● (
2 (1+Δt') -2 (1+Δt'))+y(t) ... (I6) However, Δt'> Δt, and XN(t): Sol loss C in the past N hours from time t
Quantity average value XM(t): M hours past from time t (M >> N)
Sol loss cffi average value Z (t + Δt'): operation action (heating action,
The predicted value of melt temperature at time (t + Δt') due to the influence of heat reduction action (hereinafter referred to as "actual action prediction 111 +
"value". ) Z (t + Δt'): Average value of actual action predicted value Z (t) over the past N hours from time (t + Δt') V (t): Melt temperature y (t) over the past N hours from time t It is an average value.

Also, coefficients a, b and XN Ct), Z (t+Δt'
), y (t), the set time N changes from time to time so as to improve the prediction accuracy of hot metal temperature abnormality, which will be described later.

Below, regarding the actual action prediction 1111 value Z (t),
This will be explained in detail with reference to FIG. Actual action prediction value 2
(1) is the predicted value of the hot metal temperature at time t when operational actions such as wind temperature and humidity control are performed.

As shown in Figure 6, when a heating action occurs at time 0, the response (force that affects the hot metal temperature y (t))
gradually appears, and the hot metal temperature difference T, (Z(t)−Z (
t)) increases. In this way, the predicted value of the hot metal temperature based on the hot metal temperature change after performing the operation action is the actual action prediction 71?1 value Z (t). This predicted actual action value Z (t) is calculated by determining the response of hot metal temperature to actions such as wind temperature and humidity control using a mathematical model or data analysis. As the actual action prediction value 2 (1) using a mathematical model, for example, there is one disclosed in "Development of a Blast Furnace Unsteady Simulation Program"("Tetsu to Hagane" k 12. vo 173 s 825 (1987) p. 89).

Next, a method for determining the coefficients a, b and the set time N will be described. These values a, b, and N are determined so that the evaluation function J shown by the following equation (l7) is minimized.

J-pH ”l +p2 ”2 ”3 ”3 +p4
E4...(17) However, p1 to p4 are coefficients E1: Missed rate of abnormal drop in hot metal temperature E2: False alarm rate of abnormal drop in molten iron temperature E3: Missed rate of abnormal rise in hot metal temperature E4: False alarm rate of abnormal rise in hot metal temperature be. In addition, abnormal drop in hot metal temperature (hereinafter simply ``drop'')
That's what it means. ) means that the temperature of hot metal falls below the lower control limit temperature "CL", and abnormal rise in melt temperature (hereinafter simply referred to as "rise").

) means that the hot metal temperature is the control upper limit temperature T. It is assumed that the number exceeds 11. It is also assumed that conditions other than those described above are stable.

E1 to E4 are determined by the following formulas (l8) to (2l).

utJ+uh1 UE2 Table 1 However, DG: Decrease number of passes (decrease prediction was correct) DEI:
Number of missed declines (failed to predict decline) DE2: Number of false alarms of decline (incorrect prediction of decline) U G: Number of passes for rise (incorrect prediction of decline) UEI:
Number of missed climbs (failed to predict climbs) UE2: Number of false climb warnings (wrong predictions of climbs made).

The above six judgments are performed when the operation action is performed based on the prediction (t+Δt') of this example (in other words, when a decrease (increase) is predicted, a heat increase (heat decrease) action is performed, and a stable prediction is made. (No action is taken if

In addition, the operation action is the prediction (t+Δt') of this example.
Table 2 shows the determination in the case where no operational action is taken in response to the prediction of decrease or increase in this embodiment.

Table 2 also shows that the operation action is predetermined A1'g(t+Δ
t'), and the operational action is contrary to the prediction yCt+Δt') of this example, in other words, even though the reduction (increase) prediction Alj has been made, the heat decrease (increase) The judgment when the heat) action is performed is as shown in Table 3.

Table 3 further shows that the operational action is predicted in this example (t+Δt'
), and the determination in the case where an operation action is performed at the time of stable state prediction in this embodiment is as shown in Table 4.

A fourth scheduled time N is determined. Note that the determinations in Tables 1 to 4 described above are performed every predetermined time interval p' using the predicted value V (t) and the actual hot metal temperature value y (t) within the time width g'.

As a result, the coefficients a, b and setting time N of equation (lB) are:
Abnormal decrease in furnace heat during the evening period. In order to improve the accuracy of predicting the presence or absence of an abnormal rise, the blast furnace heat preheat l1-1 according to this embodiment is changed every time 1', so that an abnormal rise or fall in the furnace heat occurs in the time width 9'. It is expected that it will be possible to accurately predict 11I1 whether or not.

Based on the judgments in Tables 1 to 4 above, (l8) to (
21) From formula 1N-1 failure rate E1 to E for the past predetermined period
4 is calculated. Then, the coefficients a and b are set so that the evaluation function J obtained from equation (17) is minimized. ~E.
Prediction results obtained by each prediction means IPl4 (prediction means A
．． ), Ix (prediction means B,), I (prediction means C
, ), I (prediction means D.). y
-p
-(t+Δt') (prediction means E.) By comparing each of them with the membership functions M-M5 shown in FIGS. 7 to 11, furnace heat evaluation points P1 to P5 are obtained. For example, by each prediction means, IPM-1
, I-4, I--0.35, I,
-0.55. When 9xy (t+Δt') - 1477.5 is obtained, P (summer P value point) is P B - 1.0{ P,, ( Ix evaluation score) is P S - 0.5P M
- 0.5 P3 (evaluation point of 1) is NB - 0.25M10,75 P4 (evaluation point of 1) is P B - 0.75M
−0.25 P5 (evaluation point of (t+Δt')) is evaluated as N S −0.75 M −0.25.

In addition, in FIGS. 7 to 11, NB, NM, NS,
M, PS, PM, and PB are furnace heat evaluation parameters, and each parameter is: NB...fairly low NM...low NS...slightly low M...stable PS...slightly high PM...・High PB...means quite high. These parameters have a goodness of fit between 0 and 1. Moreover, the shape of the membership function M-M5 shown in FIGS. 7 to 11 can be changed as appropriate depending on (1) differences in blast furnace, controlled temperature, etc.

(P-2) Determining the intermediate conclusion The furnace heat evaluation points P1 to P5 obtained in this way are compared with the furnace heat evaluation rule in the IF-TI{EN format created in advance, and if successful, the intermediate conclusion CI get. The verification is successful at the furnace heat evaluation point p − shown in the IF section (condition section).
Furnace heat evaluation parameters of p5 (NB, NM, NS, M, PS, PM,
PB) are all positive values. If each furnace heat evaluation point P - P5 exemplified in (P-1) is reached, the verification of Rules 1 to 10 shown in Table 5 is successful.

Table 5 DS: Small furnace heat decrease DB: Large furnace heat decrease N: Furnace thermal stability THEN part (conclusion part) DS, DB, N & 1 parameter of furnace heat change indicating the degree of change in the furnace heat level, There are also UB (large furnace heat rise) and US (furnace heat rise/JX) forces. In addition, the suitability of each furnace heat change parameter is determined by the value of the furnace heat evaluation parameter in section
When a plurality of furnace heat evaluation parameters exist in the IF section as in 8 to 8, the minimum furnace heat evaluation parameter is used to determine.

In this example, the intermediate conclusion C derived by the THEN part
There are three furnace heat change parameters for I: DS, DB, and N. As the evaluation value of these furnace heat change parameters, the maximum degree of conformity among the same furnace heat change parameters is adopted.

That is, DS-0.75 (Rule 9), DB-0.
5 (Rule 7.8), N-0.5 (Rule 1)
. An intermediate conclusion CI of UB-US-0 (no adoption rule) is obtained.

(1'-3) Inference of overall reactor heat level Once the intermediate conclusion CI is obtained, the membership function BDB(y) of each reactor heat change parameter is determined as shown in FIG.
BDs(y). BN (y), Bun (V) Bus
Evaluation values for (y) are 0.5 and 0.75, respectively.
0.5, 0. A composite area BB(y) (indicated by diagonal lines) is created, which is an area whose upper part is cut at 0.

The y-coordinate y of the position of the center of gravity G of the region surrounded by the composite region BB (y) and the y-axis is determined by the following equation (22).

This y* is the overall reactor heat level obtained by fuzzy reasoning, and in the example of FIG. 12, y*-o. It becomes 5. The closer the overall furnace heat level y* is to 1.5, the stronger the tendency for the furnace heat level to decrease, and the closer y is to -1.5, the stronger the tendency for the furnace heat level to increase. If is close to 0, it indicates that the furnace heat level tends to be stable.

In this way, the total furnace heat level y* is determined by fuzzy estimation? Since it is obtained by , it is a quantitative value. Also, A. ~C. The short-term outlook prediction result 'PM''x, 1,, I described in section 1. The reactor heat level and the long-term outlook p described in section D.
(2) Since the total furnace heat level y'' is determined based on the predicted result (t+Δt') of the furnace heat level, it can be said that the total furnace heat level y is a highly accurate prediction result.

G. Selection of execution operation action F. The total furnace heat level y determined by y and the current wind temperature. By comparing operating conditions such as humidity control, pulverized coal flow rate, and 0/C against the knowledge base, the next operating action to be executed is selected. The knowledge base is
The rules are written in IF-THEN format, and these rules have been extracted based on the operator's know-how, and are based on constraints on blast furnace equipment, responses to changes in actual melt temperature, and past considerations. The impact of operational actions, etc. are taken into account. An example of the rule is shown below.

■ I F y > C I A N D [There is no operation action in the past d hours] AND Hot metal temperature Y (t) < Cz AND Wind temperature < C3 T II E N Wind temperature lowering action ■ I F
y > C r A N D [There is no operation action in the past d hours] AND Hot metal temperature y (t) < C Z A N D
Air temperature > C3 AND Humidity control <04 T II E N Humidity control lowering action In this way, taking into account the overall furnace heat level and the current operating conditions, operational actions are determined based on the knowledge base accumulated by the operator's know-how. Since the selection is made, the optimal operational action can be automatically selected.

H. Calculation of operational action amount G. When the execution operation action is selected, the operation action amount A, which is the control amount of the execution operation action, becomes the next (2
3) Calculated from the formula.

A-f(y)...(23)
The function f(y) differs depending on the operational action taken, and can be estimated from the overall furnace heat level y and the amount of actual action taken by the operator. FIG. 13 shows the function f when the executed operation action is the wind temperature lowering action.

The function f(y) can be determined by the least squares method based on past data, and the function f in Figure 13 is f (y) -0.19-0.61y +15
．． 17(y”) 2*3 +87.32 (Y) -18.5(V*)
'- 57.83(y "") 5 ・
...(24).

In this way, since the operational action amount is determined by a function using the overall furnace heat level y as an argument, it is possible to finely adjust the operational action amount.

1. Execution of operational actionsG. The execution operation action determined in H. The blast furnace furnace heat is automatically controlled in order to stabilize the furnace heat by executing it according to the operation action amount determined in . Also, A. ~H
By displaying the results described in , on a CRT or the like, they can be used as an action guide for operators.

J. Flow of Blast Furnace Furnace Heat Control FIG. 14 is a flowchart showing the process flow when a blast furnace heat control method according to an embodiment of the present invention is carried out using a computer. Further, FIG. 15 is a conceptual diagram showing a furnace heat control system to which a blast furnace furnace heat control method according to an embodiment of the present invention is applied. Hereinafter, the processing procedure in this embodiment will be explained with reference to FIG. 14.

First, in step S1, A. ~D. As mentioned above, the short-term outlook prediction method can be used to predict the short-term outlook. i-1 to obtain various short-term outlook prediction results.

Next, in step S2, E. As mentioned above, the long-term outlook based on equation (l6) a? A long-term outlook prediction is performed by one means, and a long-term outlook prediction n1 result is obtained.

Then, in step S3, F. As mentioned above, the furnace heat evaluation points P-P5 of the short-term forecast 811 results and the long-term forecast results obtained in steps Sl and S2 are expressed as membership function M.
-M5, and from these furnace heat l evaluation points P-P5 and the furnace heat evaluation rule, the overall furnace heat level y is determined by Falzi inference.

After that, in step S4, G. As mentioned above, using the knowledge base that takes into account the overall furnace heat level y and the operating conditions, the execution operation action that is the optimal operation action that should stabilize the furnace heat is selected.

Then, in step S5, H. As described above, the control amount of the execution operation action selected in step S4 is determined by a function using the overall furnace heat level as an argument.

Then, in step S6, l. As mentioned above, the execution operation action selected in step S4 is transferred to step S5.
Execute according to the amount of operational action determined in . Thereafter, steps 81 to S6 are repeated to continue furnace heat control.

K. Supplementary note: A. The furnace wall temperature prediction described in 1-
In one method, an FM sensor is used as the inner wall thermometer, but a normal aPI temperature sensor (for example, a sheathed thermocouple) can be used as a substitute, although there are problems with the service life. In addition, the reliability of using a stave thermometer or a brick-embedded thermometer. Ha1
Although the prediction accuracy will be slightly lower due to the low temperature responsiveness, it can be used as a substitute.

(Effects of the Invention) As explained above, according to the present invention, in the first step, the reference value is changed from moment to moment according to the blast furnace data for a predetermined period in the past, so that it can respond to level changes in the blast furnace data. By doing so, high prediction accuracy can always be maintained.

In addition, in the third step, the furnace heat level of each of the first and second prediction results is determined by the membership function, and is calculated by fuzzy inference from the predetermined first rule and each of the furnace heat levels. , to determine the overall furnace heat level and predict short-term furnace heat changes? IFI L 1st
Based on the prediction results of 11) I L. It is possible to obtain a quantitative comprehensive furnace heat level with high accuracy by taking into account the second prediction result.

By comparing the overall furnace heat level obtained in this way and the current operating status with the second rule,
In the fourth step, to select the execution operation action,
The optimal operational action can be selected.

Then, in the fifth step, the operation action amount of the executed operation action is determined by a function using the overall furnace heat level as an argument, so that the action amount can be determined at a detailed level.

[Brief explanation of the drawing]

Figure 1(a). (b) is a side sectional view, a plan sectional view, and FIG. Figure 3 is a conceptual diagram and installation explanatory diagram of the FM sensor, Figure 4 is a graph showing the change over time in the furnace wall temperature measured by the FM sensor, and Figure 5 is the change over time in the difference value of the furnace wall temperature measured by the FM sensor. A graph showing,
Figure 6 is an explanatory diagram showing the situation after the heating action, Figures 7 to 11 are graphs each showing the membership function M - M5, and Figure 12 is a graph explaining the method for determining the overall furnace heat level y. , FIG. 13 is a graph showing a function that determines the operational action amount, FIG. 14 is a flowchart showing the processing procedure of a blast furnace furnace heat control method which is an embodiment of the present invention, and FIG. 15 is an embodiment of the present invention. It is a conceptual diagram of a furnace heat control system to which a blast furnace furnace heat control method is applied. 1... Blast furnace, 3... FM sensor

Claims (1)

[Claims] (1) A first system that outputs a first prediction result predicting a short-term change in furnace heat by capturing at least one piece of blast furnace data indicating the inside condition of the blast furnace and comparing the blast furnace data with a reference value. The reference value changes from moment to moment according to the blast furnace data for a predetermined period in the past, and incorporates at least one blast furnace data indicating the inside condition of the blast furnace and the hot metal temperature, and incorporates the hot metal temperature into these data. The second model predicts long-term changes in furnace heat by taking into account the effects of actions.
a second step of outputting a prediction result of the first and second prediction results, calculating the furnace heat level of each of the first and second prediction results using a membership function, and using these furnace heat levels and a predetermined first rule, The third step is to calculate the overall furnace heat level by fuzzy inference, and the operation data indicating the overall furnace heat level and the current operating status are compared with a predetermined second rule to infer the furnace heat level. a fourth step of selecting an execution operation action that is an operation action that should be performed to stabilize the execution operation; and a fourth step of selecting an execution operation action that is an operation action that should be performed to stabilize the execution operation; A method for controlling heat in a blast furnace, further comprising: a fifth step of obtaining an action-specific function; and a sixth step of executing the execution operation action according to the operation action amount.