JPH08269518A - Method for guiding to charging condition of treating agent in pretreatment operation of molten iron - Google Patents

Abstract

PURPOSE: To provide a method for guiding to charging conditions of treating agents in pretreatment operation of molten iron capable of adequately guiding to the charging conditions of the treating agents in the pretreatment operation of the molten iron with good accuracy while taking the variations in practicable use into consideration. CONSTITUTION: The charging conditions of the treating agents are determined by changing the operation conditions including the charging conditions of the treating agent and repetitively inputting these conditions from a work station 5 into a neural network 7 disposed in a host computer 3. The charging conditions are evaluated by adding the variations in the practicable use in an evaluation section 39 and the adequate guidance to the charging conditions of the treating agents is provided. As the result, high estimation accuracy is easily realized regardless of whether the corresponding characteristics of the charging conditions of the treating agent and the component concn. after the treatment exhibit linearity or nonlinearity. The adequate guidance to the charging conditions of the treating agents is thus made possible.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method of guiding the conditions for adding a treating agent for removing a predetermined component in hot metal in a hot metal pretreatment operation.

[0002]

2. Description of the Related Art When a large amount of sulfur (S), phosphorus (P) and silicon (Si) is contained in hot metal, the production efficiency at the time of steel making decreases. Therefore, desulfurization treatment, dephosphorization treatment, and desiliconization treatment are performed as the hot metal pretreatment. For example, in desulfurization treatment, lime (CaO ), And a sulfur component is removed using a desulfurization agent such as carbide (CaC ₂ ). It is important to determine the input conditions such as the input amount of the desulfurizing agent and the input time in consideration of the balance between reducing the S concentration in the hot metal below the target concentration and not increasing the operating cost. Therefore, conventionally, the desulfurization operation is modeled by a statistical method such as multiple regression, and the S concentration is estimated by using the multiple regression equation to determine the conditions for adding the desulfurizing agent.

[0003]

However, as is clear from the example of operation results (from "Iron and Steel" 1978A21 P64) shown in FIG. 8, for example, the correspondence between the amount of desulfurization agent input and the desulfurization rate after operation. The characteristics are non-linear. Therefore, the target value of the desulfurization rate is divided into multiple sections, and multiple multiple regression equations that are different for each section are used for approximation, but there are various obstacles in improving accuracy. Therefore, the accuracy of the input amount that was guided was not always satisfactory. The amount of guidance given using the above multiple regression equation is an average amount that does not consider variations in operation, and even if the desulfurizing agent is introduced according to the guidance, about half of the charges are stochastically stochastic. The target S concentration exceeds the target S concentration. Therefore, during actual operation,
In order to keep the target S concentration below the target, operators use the rule of thumb or subjectivity to add more desulfurizing agent than the guidance amount, but it is based on the rule of thumb or subjectivity. There is a drawback that the deviation easily occurs and the accuracy of the guidance amount itself is poor, so that the desulfurization agent is insufficiently supplied or excessively supplied, resulting in insufficient desulfurization or uneconomical. Incidentally, the desiliconizing agent used in the desiliconizing treatment, the dephosphorizing agent used in the dephosphorizing treatment, the blowing oxygen, the temperature adjusting additive, and the like also have the same problems as the above desulfurizing agent. The same applies to the time required to add the agent, and an excellent technique for solving this problem is desired. The present invention has been conceived in view of the above circumstances, and an object thereof is to accurately guide the conditions for introducing a treating agent in a hot metal pretreatment operation and to appropriately provide guidance in consideration of variations in practical use. It is an object of the present invention to provide a method of guiding the conditions for introducing a treating agent in a possible hot metal pretreatment operation.

[0004]

In order to achieve the above object, a first aspect of the present invention is to provide operating conditions at the time of hot metal pretreatment including conditions for adding a predetermined treatment agent for removing a predetermined component in the hot metal. Is a input, and the processing agent charging condition guidance method in the hot metal pretreatment operation is used to determine the processing agent charging condition using a neural network that outputs the above-mentioned predetermined component concentration in the processed hot metal. A second invention is a hot metal pretreatment operation in which the treatment agent is one of a S-removing agent, a Si-removing agent and a P-removing agent for removing any one of S, Si and P components. It is a method of introducing the conditions for the treatment agent guidance in the above. In the third invention, the above-mentioned input condition is
This is a method for guiding the conditions for charging the treatment agent in the hot metal pretreatment operation including the amount and time of the charge. The fourth invention is
Processing in the hot metal pretreatment operation in which the processing agent injection conditions determined by repeatedly inputting the operating conditions into the above neural network are evaluated in consideration of variations in practical use, and appropriate processing agent injection conditions are determined. This is a guidance method of agent injection conditions.

[0005]

According to the first aspect of the present invention, the operating conditions during the hot metal pretreatment including the conditions for adding the predetermined treatment agent for removing the predetermined components in the hot metal are input, and the predetermined component concentration in the hot metal is input. By using a neural network that outputs the above, the concentration of the above-mentioned predetermined component in the hot metal after the treatment when the conditions for feeding the above-mentioned treatment agent under various operating conditions are varied, Conditions are decided. Therefore, regardless of whether the corresponding characteristic between the operating condition and the predetermined component concentration exhibits linearity or non-linearity, a high estimation accuracy can be easily realized, so that the treatment agent dosing condition can be properly adjusted. It is possible to give guidance. In the second aspect of the invention, the conditions for introducing any one of the S-removing agent, the Si-removing agent, and the P-removing agent are varied, and the S, Si, and P corresponding to the treating agent are changed.
The concentration of the component in the hot metal of any one of the above is estimated by using a neural network, and a suitable feeding condition of the above treatment agent is determined. Therefore, regardless of whether the corresponding characteristics of the conditions for introducing the S-removing agent, the Si-removing agent, and the P-removing agent and the concentrations of S, Si, and P show linearity or non-linearity, high estimation accuracy can be easily achieved. Therefore, it is possible to properly guide the conditions for adding the S-removing agent, the Si-removing agent, and the P-removing agent. In the third aspect of the invention, the treatment agent is estimated by estimating the concentration of the predetermined component in the hot metal after treatment when the amount and time of the treatment agent are varied under various operating conditions. The suitable dose and time of the are determined. Therefore, it is possible to easily realize high estimation accuracy regardless of whether the corresponding characteristics of the amount and time of the treatment agent charged and the predetermined component concentration exhibit linearity or non-linearity. It is possible to properly give guidance on the charging conditions including the charging amount and the charging time. In the fourth aspect of the present invention, the processing conditions of the processing agent determined by repeatedly inputting the processing conditions into the neural network while changing the operating conditions are evaluated in consideration of practical variations, and the appropriate processing agent is input. Determine the conditions. Therefore,
It is possible to provide guidance on suitable processing agent charging conditions according to the above evaluation conditions.

[0006]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments embodying the present invention will be described below with reference to the accompanying drawings for the understanding of the present invention. In addition,
The following examples are examples embodying the present invention and are not of the nature to limit the technical scope of the present invention.

FIG. 1 is a block diagram showing an apparatus configuration in hot metal pretreatment. A process control host computer 3 is provided to control the process of a hot metal pretreatment facility (hereinafter simply referred to as “equipment”) 1, and an operator workstation 5 operated by an operator is connected to the host computer 3. . Inside the host computer 3, there is provided a neural network 7 which is used to determine a charging condition including a charging amount and a charging time of the processing agent at the time of operation. The structure of the neural network 7 will be described later.

In the facility 1, a desulfurization section 11 for removing the sulfur (S) component in the hot metal, a desiliconization section 13 for removing the silicon (Si) component, and a phosphorus (P) component are removed. And a dephosphorization unit 15 for the purpose. Each of the desulfurization unit 11, the desiliconization unit 13, and the dephosphorization unit 15 is a desulfurization agent,
A desiliconizing agent and a dephosphorizing agent (including blown oxygen and a temperature adjusting additive) are added to reduce the concentration of the components to be removed in the hot metal. Inside the host computer 3,
In addition to the neural network 7 described above, a control unit 31 that controls the processes of the desulfurization unit 11, the desiliconization unit 13, and the dephosphorization unit 15 of the melting equipment 1, and data used by the control unit 31 for process control. The process control data storage unit 33, the operation result data storage unit 35 for storing the operation result data in the facility 1, and the processing agent input amount and input time determined using the neural network 7. The evaluation part 39 for evaluating is provided.

The workstation 5 is an operating means of the host computer 3 and includes a display section 51, a printing section 53 and an input section 55. The display unit 51 is configured by a display such as a CRT or LCD for displaying data, the printing unit 53 is configured by a printer for printing out data, and the input unit 55 is configured by a keyboard, a mouse and various switches. There is.

Each part 7, 3 inside the host computer 3
1, 33, 35 and 39 are connected to each other or to the equipment 1 and the workstation 5 via a signal path indicated by a solid arrow. The signal path of this solid arrow is
This means that the devices are actually connected via signal lines or that data (signals) are transferred between different functions. The dashed arrows between the "operator" and the equipment 1 and the workstation 5 represent the relationship in which operations or observations without signal exchange, work, or the like intervene. The operator is
The host computer 3 controls the process of the equipment 1 by inputting commands and data as necessary from the input unit 55 while monitoring various data and guide messages displayed on the display unit 51. In the host computer 3, the control unit 31 has the process control data storage unit 33.
The control data is read from and the operation command signal is sent to the equipment 1 according to the procedure according to this data. In the equipment 1, the desulfurization section 11, the desiliconization section 13, and the dephosphorization section 15 are activated in response to the sent command signal. The operator performs a predetermined work while observing the processing situation in the facility 1.

During operation, various data relating to the progress status of processing and the internal state of the equipment 1 are sent from the equipment 1 to the control section 31 at predetermined timing. The control unit 31
The contents of the sent data are analyzed, a command signal corresponding to the data is sent to the equipment 1, and a guide message is displayed on the display unit 51 of the workstation 5. In addition, the control unit 3
1 stores a predetermined one of the data sent from the equipment 1 in the operation result data storage unit 35. Among the data to be stored are the amount of hot metal, the concentrations of sulfur (S), silicon (Si), manganese (Mn), and phosphorus (P) contained in the hot metal, the hot metal temperature, and the desulfurization unit 11 And a treating time which is an amount of treating agent and a treating agent used in each of the decalcification section 13 and the dephosphorization section 15, and
It includes the lance immersion depth, the number of times the slag was topped, the slag thickness, and the number of times the lance was used. In the following description, the amount of the processing agent added and the amount of time the agent is added are referred to as the injection conditions.

Of the data stored in the process control data storage unit 33, the processing agent charging conditions are estimated by using the neural network 7 to estimate the component concentrations after processing under various operating conditions. It is determined in advance.
The operator operates the workstation 5 to select an appropriate performance data from the performance performance data storage unit 35 of the past multiple performance data, or input an appropriate performance data from the input unit 55. Then, the neural network 7 is made to learn this in advance as the teacher data. The neural network 7 includes the amount of hot metal, the concentration of each component of S, Si, Mn, and P in the hot metal before the treatment, the hot metal temperature,
Desulfurizing agent, desiliconizing agent, and dephosphorizing agent used in facility 1, equivalent amount of input amount, processing time, lance immersion depth,
Input the operating conditions including the number of times to use slag after topping, the thickness of slag, and the number of times to use lance. When the desulfurization treatment is supported, the S concentration after treatment is treated, and when the desulfurization treatment is supported, the Si concentration after treatment is treated. When the dephosphorization process is supported, the P concentration after the process is output.

When deciding the treating agent charging condition, the operator inputs the planned operating condition (including the treating agent charging condition) in the learned neural network 7 by using the workstation 5. The neural network 7 outputs the processed component concentration corresponding to it in response to the input of the operating condition and sends it to the workstation 5. In the workstation 5, the output result of the neural network 7 is displayed on the display unit 51. If the output result displayed on the display unit 51 is a desirable value within the target concentration, the operator inputs into the evaluation unit 39 the conditions for adding the treating agent included in the operating conditions. The evaluation unit 39 sets an appropriate variation width with respect to the input processing agent input condition,
Furthermore, by comparing the processing agent charging condition determined this time with the processing agent charging condition in the past operation results, the processing agent charging condition of this time is evaluated, and the evaluation result is sent to the workstation 5.

In the workstation 5, the evaluation result sent from the evaluation unit 39 is displayed on the display unit 51, and the operator looks at the evaluation result, changes the treatment agent charging condition as necessary, and again performs the neural network. 7 estimates the S concentration, determines the processing agent charging condition, causes the evaluation unit 39 to evaluate the value, and if a desired evaluation result is obtained from the evaluation unit 39, the processing agent charging condition is actually operated. It is stored in the process control data storage unit 33 for use in The stored data is stored in the control unit 31 at the next operation.
And the desulfurization unit 11, the desiliconization unit 13, or the dephosphorization unit 15 is instructed to add the treatment agent under the determined conditions.

Next, with reference to FIGS. 2 to 7, the details of the method of guiding the conditions for introducing the treatment agent by the neural network 7 and the evaluation section 39 will be described. In the following description, the amount of the desulfurization treatment agent added in the desulfurization unit 11 will be described as an example.

FIG. 2 shows the structure of the neural network 7.
It is a schematic diagram which shows. Neural network 7 is
It is composed of three layers, a power layer A, an intermediate layer B, and an output layer C.
It Input layer A is m neurons A₁~ A_mThe middle layer
B is n neurons B₁~ B _n, The output layer C is
Neuron C₁Including each. Each neuron of input layer A
A (hereinafter referred to as “input neuron”) A₁~ A_mIs each
Each of the neurons in the middle layer B (hereinafter "intermediate neuron")
Say B₁~ B_nFor the weighting factor W_ij(I = 1 to 1
m, j = 1 to n). For example, input new
Ron A₁Is an intermediate neuron B₁~ B_nAgainst each
Coupling coefficient W₁₁~ W_1mInput neuron A_mIs
Intermediate neuron B₁~ B_nFor each of the weighting factors W
_m1~ W_mnAre combined in. Input neuron A₂~
A_m-1Intermediate neuron B₁~ B_nTo the connection relation to
Although illustration is omitted, it is similar to the above. Middle New
Ron B₁~ B_nAre output layer C neurons
(Hereinafter referred to as "output neuron") C₁The weighting factor W_k
They are connected at (k = 1 to n).

Input neuron A₁~ A_mTo each
Corresponding data X_iWhen (i = 1 to m) is set,
This data X₁~ X_mIs the coupling relation between the input layer A and the intermediate layer B.
A few W _ijThe value W multiplied by_ij× X_iIs each intermediate neuron B
₁~ B_nIs set to. For example, input neuron A₁
Data X₁Is set, the intermediate neuron B₁~
B_nIs the corresponding coupling coefficient W₁₁~ W_1mX₁Multiplied by
Value W₁₁× X₁~ W_1m× X₁Are input respectively.

[0018]

[Equation 1]

Intermediate neuron B ₁ to which data is input
B _n is an intermediate neuron B calculated from the sum of input values W _ij × X _i.
A value U _j is calculated by subtracting a predetermined threshold θ _j for each of _{1 to} B _n (see Expression (1)), and the value U _j is further converted by a sigmoid function f to calculate a value V _j (Expression (1 See 2)). The sigmoid function is f (x) = 1 / (1 + exp (-x))
It is represented by the formula (see Formula (3)). Each of the intermediate neurons B _{1 to} B _n outputs the calculated value V _{1 to} V _n . A value W ₁ × V obtained by multiplying the output neuron C ₁ by the coupling coefficients W _{1 to} W _n corresponding to the values V _{1 to} V _n , respectively.
₁ ~W _n × V _n are input. The output neuron C ₁ is
A value U is calculated by subtracting a predetermined threshold value θ from the total of input values (see equation (4)), and the value U is converted by a sigmoid function f (see equation (3)) similar to that in the intermediate layer. The value Y is output (see formula (5)).

To the neural network 7 configured as described above, teacher data is input from the workstation 5 and learned prior to estimating the amount of the processing agent input. That is, the operator uses the workstation 5 to select an appropriate number of data from the past multiple times of operation record data stored in the operation record data storage unit 35 (see FIG. 1), or Appropriate teacher data is prepared and set in the input layer A and the output layer C of the neural network 7 depending on whether the data is theoretically considered to be appropriate or whether new virtual data is created.

When the neural network 7 is used to determine the input amount of the desulfurization treatment agent, among the input neurons A _{1 to} A _m , A ₁ is the amount of hot metal X ₁ and A ₂ is the S concentration before treatment. X ₂ is the amount of desulfurizing agent added to A ₃ , X ₃ is A
₄ , the hot metal temperature X ₄ , A ₅ the treatment time X ₅ , A ₆ the Si concentration X ₆ before treatment, A ₇ the P concentration X ₇ before treatment,
The Mn concentration X ₈ of pretreatment A _8, a lance immersion depth X ₉ to A _9, the Haikasu after Topido use count X ₁₀ to A _10, A ₁₁
For slag thickness X ₁₁ , for A _{12 the} number of lances used X
The _12, A ₁₃ to to A _m inputs other operating conditions X ₁₃ to X _m, respectively (hereinafter, these as "teacher input values"). Then, from the output neuron C ₁ , S after desulfurization treatment
The Y value that matches the actual density value or the theoretical virtual value (both are referred to as "teacher output value" below) is output.
The coupling coefficient W _ij between the input layer A and the intermediate layer B and the coupling coefficient W _k between the intermediate layer B and the output layer C are corrected. Coupling coefficient W
_ij, the initial value of W _k, for example, be determined by using a predetermined random number table. Then, in correcting the coupling coefficient, for example, a back propagation (error propagation) method is used.
In this back-propagation method, forward calculation and error back-propagation calculation in the opposite direction are repeated to reduce the square value of the error between the output Y and the teacher output value, and the coupling coefficient W _ij , W _k is modified.

By the above learning, the operator finds that the neural network 7 is suitable for practical use, for example, as shown in FIG. 3, the amount of desulfurizing agent input (equivalent basic unit of carbide) in the past operation results, and after desulfurization treatment. If it becomes possible to accurately represent the corresponding characteristics with the S concentration of, the amount of desulfurization agent input for the operation scheduled in the future is determined using this neural network 7. As is clear from the past results of FIG. 3, generally, in the correspondence characteristics of the desulfurizing agent input amount and the post-treatment S concentration, when the input amount is small, the change rate of the S concentration with respect to the change of the input amount is large, and It is known that the change rate of the S concentration with respect to the change of the input amount is small when the input amount is large, but the output example of the neural network 7 shown in FIG. 3 has continuity in which this corresponding characteristic is easy to handle. The relationship is expressed accurately.

Data X ₁ representing the planned operating conditions by the operator
~ X _m are set to the corresponding input neurons A _{1 to} A _m from the workstation 5, the neural network 7 causes the estimated S concentration by forward calculation from the input layer A to the intermediate layer B to the output layer C. Y is output.
If the estimated S concentration is a desirable value within the target value or less, the operator adopts the amount of the desulfurizing agent added at that time, and performs the evaluation process described later on the amount.

FIG. 4 is a characteristic diagram showing the contents of the input amount correction processing in the evaluation section 39 (see FIG. 1). In the figure, when the target S concentration under a certain operating condition is s ₁ ,
It is assumed that the neural network 7 represents the corresponding characteristic T ₁ between the amount of desulfurizing agent added and the S concentration under the operating conditions. In this case, input amount corresponding to the target S concentration s ₁ on the corresponding characteristic T ₁ is t ₁ showing the point P _1. This t ₁ is an average input amount, and in actual operation, even if this guidance amount is input, the target S concentration estimated by the neural network 7 is not necessarily obtained due to various disturbances. In other words, stochastically, when the guidance amount of desulfurizing agent is added, about half of the charge is the target S.
It will be higher than the concentration (that is, not achieved).

Therefore, in this embodiment, the amount of processing agent charged is controlled.
The variation width ε taking into consideration disturbances etc.
Set. Guidance was given to achieve the target S concentration
Input amount is t₁, The S concentration after treatment is s₁From
The largest difference s deviated by ε ₂When there is a case
Think This concentration s₂Is the characteristic line T₁Upper point P₂Corresponding to
The input amount is t₂Is. S concentration after treatment s₂The eyes
Standard S concentration s₁In order to reduce to P₂Corresponding to
Input t₂Point P₁Input t corresponding to₁Matched with
Characteristic line T so that₁Parallel to the horizontal axis (input amount) direction
Characteristic line T₂To set. neural network
T represented by 7₁Goal achievement point P above ₁Is the characteristic line
T₂Point P above₃And the input amount is "t₁To "t
₁−t₂"2t"₁-T₂It will be. Follow
In this case, the maximum is "2t₁-T₂The amount of
The target concentration can be sufficiently achieved without excess or deficiency.

Here, whether or not the value of ε is appropriate in consideration of the operating cost becomes a problem. Next, the economic evaluation of the input amount will be described. FIG. 5 is a characteristic diagram illustrating the relationship between the S concentration estimated by the neural network 7 and the S concentration in the operation record. For example, the variation width D as shown in Table 1 is set based on this characteristic diagram.

[0027]

[Table 1]

The relation between ε and the variation width D is ε = α
It is represented by the formula of × D. α is a coefficient that changes between 0 and 1, and the target achievement rate of the S concentration changes depending on the value of α. FIG. 6 shows the change in the target underachievement rate when α is changed under a certain operating condition. In consideration of quality and certainty of operation, it is preferable that the target underachievement rate is as close to 0 as possible, but it is also important to keep the operation cost below a predetermined level in practical use. For example, in actual operation, the target underachievement rate is 1
If it is known that the most desirable balance between quality and certainty of operation and cost can be achieved when set to 2%, 0.7 of α value corresponding to 12% is selected as shown in the figure. . Once α is determined, multiply it by the variation width D and ε
Is calculated, and the input amount “2t ₁ −t ₂ ” corresponding to this ε is calculated.
Is required. This “2t ₁ −t ₂ ” is the guidance amount in consideration of variations in practical use.

Next, a specific quantitative evaluation of the guidance amount will be described. FIG. 7 shows the input amount when the non-reaching rate is set to 12%, that is, the α value is set to 0.7, and the variation width D is determined,
It is a comparison diagram with the input actual amount when the unachieved rate is 12% in the past operational results. In the figure, the target S concentration is 10p
In pm, it can be seen that the amount of input was insufficient in the past operation and the amount of S other than that was excessive except for 30 ppm. Then, when the evaluation is carried out over all the target S concentrations, if the amount is input according to the guidance amount, 26
% Desulfurizing agent can be saved. Also, the operator
Depending on the result of the evaluation, in order to obtain a more desirable evaluation, the input amount is re-determined by using the neural network 7 again, or the guidance amount in which the values of α and D in the evaluation unit 39 are adjusted to reconsider the variation is used. To ask.

In this way, the operator evaluates the guidance amount, repeats the process of determining the input amount using the neural network 7 as necessary, and finally selects the desired guidance amount using the workstation 5. Then, the value is stored in the process control data storage unit 33 (see FIG. 1), and the process control host computer 3 (see FIG.
Control) to control the desulfurization operation. As a result, it is possible to accurately guide the amount of desulfurizing agent to be added in the hot metal pretreatment operation, and to take into account variations in practical use.

As the teacher input value at the time of learning of the neural network 7 and the input data at the time of estimating the S concentration, among the above X _{1 to} X _m , the amount of hot metal X ₁ , the S concentration before processing X ₂ , the desulfurization, The amount of agent input X ₃ , the hot metal temperature X ₄ , and the processing time X ₅ are essential data.
Concentrations of other Si, P, Mn components before treatment X _{6 to} X ₈ , lance immersion depth X ₉ , number of times to use topped after sludge X ₁₀ , slag thickness X ₁₁ , number of times to use lance X ₁₂ , other operations Condition X
Add _13- X _m . More preferably, the other operating conditions X _{13 to} X _m include the respective amounts of the silicidation agent for removing the Si component and the dephosphorization agent for removing the P component. Let

On the other hand, the neural network 7 is made to correspond to the de-causing process of the de-causing unit 13 and the estimated Si after the process is performed.
In the case of the configuration for outputting the concentration, in place of the S concentration before the treatment and the amount of the desulfurizing agent added, the Si concentration of the treating agent and the amount of the desiliconizing agent are essential data. In addition, the dephosphorization unit 15
When the estimated P concentration after the treatment is output corresponding to the dephosphorization treatment of the above, the P concentration before the treatment and the addition of the dephosphorization agent in place of the S concentration before the treatment and the amount of the desulfurization agent added. Use quantity as required data. Further, the neural network 7 has a plurality of output neurons in the output layer C,
All of the estimated S concentration, estimated Si concentration, and estimated P concentration, or any two of them are output at the same time, so that the amounts of the three types of treatment agents to be desulfurized, decalcified, and dephosphorized are collectively determined. You may Furthermore, the neural network 7 may be used to determine the processing time which is the processing agent charging time, the combination of the processing agent charging amount and the processing time, or other charging conditions.

[0033]

As described above, the method for guiding the conditions for introducing the treating agent in the hot metal pretreatment operation according to the present invention is configured as described above, so that the correspondence characteristic between the operating conditions and the predetermined component concentration is linear. Since high estimation accuracy can be easily achieved regardless of whether or not
It is possible to properly guide the conditions for adding the treatment agent. In particular, the predetermined component is any one of S, Si, and P, and the processing agent is a S-removing agent, a Si-removing agent, and a P-removing agent.
If any of these agents is used, de-S agent, Si-free
Conditions for adding chemicals and de-P agent and S, Si, P after treatment in hot metal
It is possible to easily achieve a high estimation accuracy regardless of whether the corresponding characteristic with the concentration of S shows linearity or non-linearity.
It is possible to properly provide guidance on the conditions for introducing the Si-eliminating agent and the P-eliminating agent. In addition, by including the input amount and the input time in the input condition, it is possible to easily obtain a high estimation accuracy regardless of whether the corresponding characteristic between the input condition and the predetermined component concentration exhibits linearity or non-linearity. As a result, it is possible to properly guide the charging conditions including the charging amount and the charging time of the processing agent. Furthermore, the processing conditions are determined by repeatedly inputting the processing conditions into the neural network, and the evaluation conditions are evaluated by adding a suitable variation range, so that the appropriate processing conditions are determined. It is possible to provide guidance on proper conditions for adding the treatment agent according to the evaluation conditions. As described above, it is possible to provide a treatment agent introduction condition guidance method in the hot metal pretreatment operation capable of accurately providing treatment agent introduction conditions and appropriately considering variations in practical use.

[Brief description of drawings]

FIG. 1 is a block diagram showing an apparatus configuration in a hot metal pretreatment according to an embodiment of the present invention.

FIG. 2 is a schematic diagram showing the configuration of a neural network.

FIG. 3 is a characteristic diagram showing the corresponding characteristics of the amount of desulfurization agent input and the S concentration after desulfurization treatment, which is represented by a neural network.

FIG. 4 is a characteristic diagram showing the details of the input amount correction processing in the evaluation unit.

FIG. 5 is a characteristic diagram showing an example of the relationship between the S concentration estimated by the neural network and the S concentration in the past operation record.

FIG. 6 is a characteristic diagram showing a change in the target underachievement rate when the coefficient of the variation width is changed under a certain operating condition.

FIG. 7 is a comparison diagram of the input amount when the coefficient of variation width is set to 0.7 and the past input actual amount.

FIG. 8 is a characteristic diagram showing the correspondence characteristics between the amount of desulfurization agent input and the desulfurization rate in past operational performance examples.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Hot metal pretreatment facility 3 ... Process control host computer 5 ... Operator workstation 7 ... Neural network 11 ... Desulfurization part 13 ... Desalination part 15 ... Dephosphorization part 31 ... Control part 33 ... Process control data storage part 35 ... operation result data storage unit 39 ... evaluation unit A ... input layer B ... intermediate layer C ... output layer A ₁ to A _m ... input neuron B ₁ .about.B _n ... intermediate neuron C ₁ ... output neuron

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Kiyoshi Ebina 1 Kanazawa-machi, Kakogawa City, Hyogo Prefecture Kamido Steel Works, Ltd. Kakogawa Steel Works (72) Takashi Fujita Kanazawa-cho, Kakogawa City, Hyogo Prefecture Kamido Co., Ltd. Inside the Kakogawa Works

Claims (4)

[Claims] 1. The input of operating conditions at the time of hot metal pretreatment, including the conditions for introducing a predetermined treatment agent for removing a predetermined component in the hot metal, and the output of the predetermined component concentration in the hot metal after the treatment A method for guiding the conditions for introducing the treatment agent in the hot metal pretreatment operation for determining the conditions for introducing the treatment agent by using a neural network. 2. The conditions for adding the treatment agent are S, Si, P.
S-removing agent for removing any of the components
The method for guiding the conditions for introducing the treating agent in the hot metal pretreatment operation according to claim 1, wherein the condition for introducing the treating agent is either the i-agent or the P-eliminating agent. 3. The method of introducing a treatment agent condition in a hot metal pretreatment operation according to claim 2, wherein the condition of the supply includes a supply amount and a supply time. 4. The processing agent charging conditions determined by repeatedly inputting operating conditions into the neural network are repeatedly evaluated in consideration of practical variations,
The method for guiding the conditions for introducing the treatment agent in the hot metal pretreatment operation according to claim 1, wherein an appropriate condition for introducing the treatment agent is determined.