JPH06279827A - Method for predicting slopping and device therefor - Google Patents

Abstract

PURPOSE:To improve the predicting precise of slopping. CONSTITUTION:This device is constituted with a sensor treating part 1 for editing measured values from plural and different kinds of slopping sensors 1a, 1b..., a model arithmetic part 2 for estimating slag property based on molten iron information 2a and operational information 2b and a slopping inference part 3 for computing the developing likelihood of the slopping in each measured item from the results of the sensor treating part 1 and the model arithmetic part 2 and also for predicting the development of the slopping based on a slopping developing index value computed from these developing likehoods.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention is applied to converter operation, etc.
The present invention relates to a method for predicting a so-called sloping phenomenon, in which slag overflows from a furnace port or a tap hole during blowing, and an apparatus therefor.

[0002]

2. Description of the Related Art In the operation of a converter, as the blowing process progresses, the components in the molten steel are oxidized and the slag volume increases. CO gas bubbles generated by the decarburization reaction are trapped in the slag, so that the slag overflows from the furnace opening.
A so-called sloping phenomenon occurs.

The above-mentioned sloping which occurs during the blowing of the converter causes a decrease in the yield of steel output due to the scattering of granular iron to the outside of the furnace, and also becomes a disturbance of the blowing control to lower the appropriate blow-stopping ratio and the furnace pressure. It becomes a disturbance of control and causes a decrease in gas recovery rate. Therefore, control technology that predicts the occurrence of sloping and introduces a sloping inhibitor such as carbonaceous material in order to suppress it is indispensable not only for improving the iron yield but also for improving the accuracy of blowing control. Has become the technology of.

Conventionally, various methods have been proposed as a method of predicting the sloping in advance. For example, the method disclosed in JP-A-57-140812 for measuring the absolute level of the slag surface by applying microwaves to the slag surface, and the method disclosed in JP-A-54-33790 installed around the furnace opening Method for measuring acoustic changes with a microphone, and Japanese Patent Laid-Open No. 54-11
Various methods have been proposed, such as the method of continuously measuring the pressure in the furnace disclosed in Japanese Patent No. 9316, and the method of predicting sloping by image-processing the photographed image of the molten slag surface.

[0005]

However, each of the sloping prediction methods described above is based on a single measuring end (sensor).
None of them have satisfactory predictive accuracy, and each has the following problems.

First, in the case of using the microwave, an antenna serving as a sensor main body is placed on the upper part of the furnace opening of the converter, so that the sensor is seriously damaged by dust or a frame, and there is a problem in maintainability. It is difficult to maintain stable operation. Further, when utilizing the acoustic change, there is a problem that the acoustic receiving microphone is difficult to maintain and stable to operate, like the microwave method. Furthermore, in the case of the prediction method based on the furnace port pressure, it is easy to be affected by disturbances such as opening and closing of skirts and changes in the amount of exhaust gas generated, so if the time from the start of pressure change to the occurrence of sloping is extremely short, There was a problem that the countermeasures could not be taken when the correspondence with slopping was not clear. In the case of the method of predicting the outflow slag from the furnace port and tap hole by image processing, there was a problem that sloping detection delay occurred and countermeasures such as introduction of inhibitor could not be done in time.

Therefore, a main object of the present invention is to provide a slopping prediction method and a device therefor, which have higher prediction accuracy and no time delay than the conventional sloping prediction methods, and correspondingly suppress the inhibitor. It is possible to take appropriate countermeasures such as determining the appropriate input amount.

[0008]

[Means for Solving the Problems] The above-mentioned problem is to provide a single sloping sensor, or a plurality of different types of sloping sensors during the converter or hot metal pretreatment process, and to measure signals from these sloping sensors. Calculate the sloping occurrence probability for each measurement item based on this, estimate the slag physical properties from the hot metal information and operation information, calculate the sloping occurrence accuracy based on this slag physical property, and obtain from the sloping occurrence accuracy. This can be solved by predicting the occurrence of sloping based on the obtained sloping occurrence index value.

The apparatus therefor is provided with a plurality of single or plural different types of sloping sensors provided in the converter or the hot metal pretreatment process, and a measurement from the single or plural sloping sensors during blowing. Sensor processing unit for receiving and editing signals, model calculation unit for estimating slag physical properties from hot metal information and operation information, and sloping for each measurement item based on information from the sensor processing unit and model calculation unit It comprises a sloping inference unit that calculates the accuracy, calculates a sloping occurrence index value from these sloping occurrence accuracy, and predicts the occurrence of sloping based on the sloping occurrence index value.

[0010]

Since the conventional sloping prediction methods are all based on one measurement end, it is not possible to independently perform high-precision prediction due to the influence of disturbances and the like due to the characteristics of each measurement method. could not.

Therefore, in the present invention, a single sensor,
Or, in order to predict sloping based on a combination of multiple and different types of measurement end data, and a model operation unit that estimates slag physical properties, complement each other's problems,
As a result, it is possible to predict sloping with high accuracy.

Further, in the present invention, as described above, in addition to the sloping occurrence accuracy based on the actual measurement information based on a single or a plurality of measurement ends, the sloping occurrence based on the physical property estimated value of the slag based on the hot metal information and the operation information. The sloping occurrence index value is calculated in consideration of the accuracy.

Therefore, in the case of the prediction based on the data of only the actual measurement information, since the prediction is performed based on the current information, the prediction of the sloping may be delayed. By using a method that predicts sloping occurrence based on slag physical property information, that is, a method that can estimate future slag physical property value changes, there is no time delay in prediction, and sloping occurs quickly. Will be able to predict.

As a result of adopting the sloping prediction method as described above, in actual converter operation, highly accurate sloping prediction can be performed in real time on the basis of the blowing information that changes from moment to moment during blowing. Since it will be realized, it will be possible to quickly and appropriately take countermeasures such as addition of an inhibitor, which will greatly contribute to stable operation of the converter operation.

[0015]

EXAMPLES The present invention will be described below with reference to the case of combining a plurality of sloping sensors and a slag physical property estimation model. In the case of the sloping prediction method according to the present invention,
As shown in the block diagram of FIG. 1, each sensor 1a,
1b, 1c, ... Receives measurement signals from the sensor processing unit 1 that edits data necessary for inferring occurrence of sloping based on these measurement results, and slag information based on information from the hot metal information 2a and operation information 2b. Model calculation unit 2 for estimating physical properties, the sensor processing unit 1 and model calculation unit 2
And a sloping inference unit 3 that estimates the sloping occurrence probability P (i) based on the data from the above data, and controls the amount of sloping inhibitor such as graphite based on the information from the sloping inference unit 3. And a sloping control unit 4 for switching.

In the sensor processing unit 1, the measurement results obtained by the sensors 1a, 1b, ... Are edited,
They are sorted and the result is sent to the sloping inference unit 3. The sloping inference unit 3 calculates the sloping occurrence probability P (i) for each measurement item, for example, according to the following sloping occurrence determination rule.

The determination rules for predicting sloping by utilizing acoustic changes, predicting based on the furnace opening pressure, and predicting based on the exhaust gas flow rate will be described below.

(1) Predicting sloping by utilizing acoustic change When the sound accompanying the blowing of top blowing oxygen is reduced by fdB or more from the start of blowing for 3 to 5 minutes from the start of blowing Then, for example, the sloping occurrence probability P is converted according to the following determination rule. P (acoustic change) = min (1.0, e × (acoustic change amount-f)) where e, f; constant, acoustic change amount; blowing start sound-current sound (dB).

(2) Forecasting based on the furnace port pressure When the furnace port pressure rapidly increases When the furnace port pressure increases by hmmH20 or more in k seconds, for example, the sloping occurrence probability P is converted according to the following judgment rule. . P (furnace pressure) = min (1.0, g x (furnace pressure change-
h)) where g, h; constant, amount of change in furnace pressure; furnace pressure before k seconds −
The current furnace pressure is (mmH20)) / k.

When the furnace port pressure becomes negative When the furnace port pressure becomes negative, the sloping occurrence probability P is converted by, for example, the following determination rule. P (furnace port pressure) = min (1.0, −u × furnace port pressure)) where u is a constant.

(3) Prediction based on the flow rate of exhaust gas For example, although one minute or more has elapsed from the start of blowing and the top-blown oxygen amount has not decreased by p% or more as compared with the data k seconds before, When the exhaust gas flow rate decreases by n% or more, the sloping accuracy P is converted according to the following determination rule. P (exhaust gas flow rate) = min (1.0, r x (flow rate change rate-
s)) where p, n, r, s; constant, flow rate change rate;
Current flow rate) / flow rate before k seconds x 100.

On the other hand, the model computing section 3 creates data for predicting future sloping occurrence by calculating the slag basicity based on the following equation, for example. Basicity = WCaO / WSiO ₂ −−−−−−−−− (1) WCaO = Σ ((CaO content in auxiliary raw material) × (amount of dissolved auxiliary raw material))
× 1/100 × S where S; CaO slagging ratio WSiO ₂ = ((HMSi × HMwt + CMSi × CMwt + SCSi × SCwt) − (C
alSi x HMwt + CMSi x undissolved CMwt + SCSi x undissolved SCwt))
× 1/100 × 60/28 + Σ ((CaO content in the auxiliary material) × (amount of dissolved auxiliary material)) Note that the dissolution of hot metal Si, cold pig iron, scrap and each auxiliary material is It can be estimated that it progresses with a delay.

The basicity obtained by the above-mentioned formula, and other information such as the Si content in the hot metal, the amount of quicklime added, the top-blown oxygen velocity, and the top-blown oxygen amount are the sloping inference unit 3
Then, the slag height H is obtained by the following equation, and the slag occurrence probability P is obtained from this slag height H.

For example, when 15 minutes or less after the start of blowing, H
(Slag height) is represented by the following formula. H (slag height) = α × Si content in hot metal × f (basicity) × oxygen integrated amount by top blowing × top blowing oxygen velocity ---
−−− (2) where α is a unit conversion gain, f (basicity) is a function having a basicity x as a variable, and for example, f (x) = a + bx
(A and b are constants).

The conversion to the sloping accuracy P is performed, for example, according to the following determination rule. When H (slag height) ≥ c P (slag height) = min (1.0, d x (H (slag height)
-C)) where c and d are constants.

As described above, the sloping accuracy P (i) based on the actual measurement is obtained for each measurement item, and
The estimated sloping occurrence probability P (i) based on the basicity calculated by the model calculation unit is obtained, and the sloping occurrence index value Q is obtained from the following equation from the total sloping occurrence probability P (i).

Q = (Max (P (i)) + α · F (P (j))) × β −−−−−−−− (3) where F is the average value of the data of j ≠ i. , Α are constants for adjusting to the actual sloping level. Further, β is a coefficient determined according to the basicity obtained from the equation (1), and can be set as shown in FIG. 3 as an example. The first term in the previous equation (3) is a term for capturing the occurrence of sloping detected by the characteristics of each sensor, and if one of the multiple sloping prediction methods is used to predict the occurrence of sloping, It is possible to determine that sloping is present from this item. Also,
The second term is to make up for the shortcomings of the individual prediction methods, and even if the occurrence of sloping cannot be predicted by a single prediction method, by adding the average value of the sloping occurrence probabilities by other prediction methods, Even if all the measurement items do not reach the prediction level, it can be determined that there is sloping if the total value exceeds a certain value. Further, by multiplying by the coefficient β, it becomes possible to avoid unnecessary sloping prediction in an atmosphere in which sloping is unlikely to occur physically, and consequently improve prediction accuracy.

The constant α can be obtained as α = (1-1 / (m-1)) × m ---------- (4), where m is the number of all data. it can.

Then, if the sloping index value Q is larger than a certain threshold value, it is predicted that sloping has occurred, and if it is smaller than the threshold value, it is judged that sloping has not occurred.

[Example] Hereinafter, a sloping prediction based on a furnace end pressure and an exhaust gas flow rate measurement end and a sloping prediction method based on an estimated value of a slag height calculated from hot metal information and operation information will be used. The sloping prediction method according to the invention will be described in detail. The furnace port pressure may be measured directly using a furnace port pressure gauge, or may be indirectly obtained by using an acoustic sensor or a vibration sensor as an index of the furnace port pressure. Good.

FIG. 2 (A) is a graph showing the results of furnace pressure measurement, exhaust gas flow rate measurement and estimated slag height, with the horizontal axis representing the elapsed time from the start of blowing.
Sloping occurs at t ₁ , t ₂ , and t ₃ on the horizontal axis.

FIG. 2B shows a result of calculating the sloping occurrence probability P (i) for each sensor according to the above-mentioned sloping occurrence determination rule for each measurement item based on each measurement result described above. Shown in the upper row. In the case of sloping prediction based on the furnace port pressure, it is possible to predict the sloping that occurred at times t ₁ and t ₃ among the three sloping occurrences, and in the case of sloping prediction based on the exhaust gas flow rate. Was able to predict the sloping that occurred at times t ₁ and t ₂ . Moreover, in the case of sloping prediction based on the estimated slag height, the sloping that occurred at times t ₁ and t ₃ could be predicted.

As described above, the estimated rate of sloping occurrence based on a single sensor or the like (the number of occurrences of sloping is the denominator and the number of times that can be predicted is the numerator) is shown on the right side of FIG. 2 (B). As described above, it is 2/3 to 1.5 / 3 (the sloping at the time t ₃ of the estimation based on the slag height is evaluated as 0.5), and is estimated in any sensor-based determination. There is an occurrence of sloping that cannot be obtained.

However, as shown in Table 1, the sloping occurrence accuracy P based on each sensor is integrated to
When the presence / absence of sloping is predicted and determined from the result of the sloping occurrence index value Q obtained by the equation (3), the threshold value γ = 2 indicating the occurrence of sloping is exceeded for all three times, and all the slopping occurrences are exceeded. While the occurrence of roping is predicted, the estimated sloping level (Q) and the actual sloping level indicating the magnitude of the actually occurring sloping are in good agreement.

The actual sloping level is
1: No sloping, 2: Small sloping occurred, 3:
It is a numerical value defined by occurrence of large sloping.

[0036]

[Table 1]

As described above in detail, in the prediction of sloping occurrence based on any of the method of the furnace port pressure, the exhaust gas flow rate, and the slag level, it is not possible to predict all the sloping that has occurred, whereas the slopping according to the present invention can be predicted. In the case of the roping prediction method, it becomes possible to predict the occurrence of all sloping. Further, since the determination level corresponds to the magnitude of the sloping that occurs, the amount of the sloping inhibitor added is proportional to the magnitude of the sloping determination level. As compared with the conventional method, it is possible to take fine-grained and appropriate countermeasures and avoid excessive countermeasures.

The information for estimating the occurrence of sloping is not limited to the case where there are three pieces of information as in the above embodiment. For example, when there are n pieces of information, the constant α is obtained by the equation (4) and the sloping The same determination can be made and the same result can be obtained by setting the threshold of occurrence; γ to γ = n × 2/3 −−−−−−−−−−− (5).

Here, 2 in the equation (5); a threshold value when there are three pieces of estimated information, and 3; means the number of original pieces of estimated information.

[0040]

As described above in detail, according to the present invention, the accuracy of prediction is significantly higher than that of the conventional prediction of occurrence of sloping, and there is no time delay. Moreover, since the index value also corresponds to the magnitude of sloping, the corresponding measure can be an appropriate amount.

[Brief description of drawings]

FIG. 1 is a block diagram of a sloping prediction method according to the present invention.

FIG. 2 is a diagram illustrating sloping prediction for each procedure.

FIG. 3 is a diagram for obtaining a constant β in a formula for calculating a sloping occurrence index value.

[Explanation of symbols]

1 ... Sensor processing unit, 1a, 1b, 1c ... Sensor, 2
... Model calculation unit, 2a ... Hot metal information, 2b ... Operation information, 3
… Slopping inference unit, 4… Slopping suppression unit

[Procedure amendment]

[Submission date] March 30, 1993

[Procedure Amendment 1]

[Document name to be corrected] Drawing

[Name of item to be corrected] Figure 1

[Correction method] Change

[Correction content]

[Figure 1]

Claims (2)

[Claims] 1. A single sloping sensor or a plurality of different types of sloping sensors are provided during a converter or hot metal pretreatment process, and a sloping sensor is provided for each measurement item based on measurement signals from these sloping sensors. In addition to calculating the sloping occurrence probability, estimating the slag physical properties from the hot metal information and operation information, calculating the sloping occurrence accuracy based on this slag physical property, and using the sloping occurrence index value calculated from these slag occurrence accuracy values. A method for predicting sloping, comprising predicting the occurrence of sloping based on the above. 2. Single or multiple and dissimilar sloping sensors provided during a converter or hot metal pretreatment process and for receiving and editing measurement signals from the single or multiple sloping sensors during blowing. Of the sensor processing unit, a model calculation unit for estimating slag physical properties from hot metal information and operation information, and a sloping occurrence probability for each measurement item based on information from the sensor processing unit and the model calculation unit, A sloping prediction device comprising: a sloping occurrence index value calculated from these sloping occurrence accuracies; and a sloping inference unit that predicts sloping occurrence based on the sloping occurrence index value.