JP6911808B2 - Control device for sinter manufacturing equipment, sinter manufacturing equipment and sinter manufacturing method - Google Patents

Description

A control device for a sinter manufacturing facility that can determine operating conditions that satisfy the control values in multiple control items when manufacturing sinter with a sinter, and sinter that manufactures sinter under the operating conditions. Regarding the method of producing ore.

When producing sinter with a Dwightroid sinter, various operating techniques have been developed and implemented to improve the strength of the sinter, the production rate of the sinter, and the like. In Patent Document 1, the relationship between the position where the exhaust gas temperature discharged from the sintering machine in the machine length direction of the Dwightroid type sintering machine becomes a predetermined temperature and the position where the maximum temperature becomes the maximum temperature, using the past operation record data. Is disclosed, and a method of predicting the maximum temperature position from a predetermined temperature position in the current operation and controlling the pallet speed so that the maximum temperature position becomes a predetermined position is disclosed.

Japanese Unexamined Patent Publication No. 2006-307259

In Patent Document 1, the fluctuation of the sinter end point can be reduced by adjusting the pallet speed so that the maximum temperature position becomes a predetermined position, whereby the quality of the sinter, the yield of the product, and the production can be reduced. It is disclosed that the sex can be improved. However, the control items when manufacturing sinter are not only the quality, product yield, and productivity of sinter, but also an electrostatic precipitator that removes exhaust gas emitted from the sinter (in the following description, electricity). The pre-EP temperature of the exhaust gas and the pre-EP negative pressure of the exhaust gas for preventing damage to the dust collector) need to be controlled in the same manner. The method disclosed in Patent Document 1 only adjusts the pallet speed, and controls not only the quality, product yield, and productivity of the sinter, but also the pre-EP temperature of the exhaust gas and the negative pre-EP pressure of the exhaust gas. The adjustment allowance was small to determine the operating conditions of the sintering machine that also satisfied the values, and it was difficult to determine the operating conditions that satisfied these control values.

The present invention has been made in view of such problems of the prior art, and an object of the present invention is to widen the adjustment allowance of operating conditions and to produce a product sinter, which is a control item when producing a product sinter. Manufacture of sinter that can increase the possibility that the operating conditions of the sinter that can satisfy not only the strength and the production rate of the product sintered ore but also the control value of the pre-EP temperature of the exhaust gas and the negative pressure before the EP of the exhaust gas can be determined. The purpose is to provide a control device for the equipment.

上記（１）式において、Ｓ_ｅ追加前は、回帰式モデルに入力変数を追加する前の残差平方和であり、Ｓ_ｅ追加後は、回帰式に入力変数を追加した後の残差平方和であり、Φ_ｅ追加前は、入力変数を追加する前のΦ_ｅであり、Φ_ｅ追加後は、入力変数を追加した後のΦ_ｅであり、Φ_ｅは下記（２）式で算出される値である。
Φ_ｅ＝ｎ−２・・・（２）
上記（２）式において、ｎは、入力変数の数である。
［４］鉄含有原料、ＣａＯ含有原料および凝結材を配合して焼結原料とする原料配合装置と、前記焼結原料に水を添加して造粒する造粒機と、前記造粒された焼結原料を焼結機のパレット台車に装入して装入層を形成させ、前記装入層を焼結する焼結機と、破砕機と、冷却機と、篩分け装置と、［１］から［３］の何れか１つに記載の焼結鉱製造設備の制御装置と、を有する、焼結鉱製造設備。
［５］鉄含有原料、ＣａＯ含有原料および凝結材を含む焼結原料に水を添加して造粒し、造粒された焼結原料を焼結機のパレット台車に装入して装入層を形成させ、前記装入層を焼結して焼結鉱を製造する焼結鉱の製造方法であって、少なくとも前記焼結原料のコークス比、造粒水分量、前記装入層の層厚およびパレット台車の進行速度を含むガイダンス項目の設定値と操業諸元とからなる要求条件を設定する要求条件設定ステップと、出力変数としての焼結鉱強度、焼結鉱生産率、排ガスのＥＰ前温度および排ガスのＥＰ前負圧と、入力変数としての前記ガイダンス項目の設定値および操業諸元とが過去の操業条件ごとに格納されたデータベースを用いて、前記要求条件と前記データベースの入力変数との類似度を過去の操業条件ごとに算出する類似度算出ステップと、前記類似度と前記入力変数とを用いて、焼結鉱強度、焼結鉱生産率、排ガスのＥＰ前温度および排ガスのＥＰ前負圧のそれぞれを出力変数とした局所回帰式を作成し、前記局所回帰式と前記要求条件とを用いて焼結鉱強度、焼結鉱生産率、排ガスのＥＰ前温度および排ガスのＥＰ前負圧の予測結果を算出する予測結果算出ステップと、前記予測結果算出ステップで算出された焼結鉱強度、焼結鉱生産率、排ガスのＥＰ前温度および排ガスのＥＰ前負圧の予測結果の全てが予め定められた管理値を満足する場合に、前記要求条件を焼結鉱製造条件として設定する製造条件設定ステップと、前記製造条件設定ステップで設定された焼結鉱製造条件で、焼結鉱を製造する焼結鉱製造ステップと、を有し、前記予測結果算出ステップでは、前記入力変数のうち、トレランスが０．１より大きい入力変数を用いて前記局所回帰式を作成し、前記製造条件設定ステップでは、算出された焼結鉱強度、焼結鉱生産率、排ガスのＥＰ前温度および排ガスのＥＰ前負圧の予測結果の少なくとも１つが前記管理値を満足しない場合には、前記要求条件設定ステップで用いた前記ガイダンス項目の設定値を変更して、再び、前記類似度算出ステップ、前記予測結果算出ステップおよび前記製造条件設定ステップを繰り返し実施する、焼結鉱の製造方法。
［６］前記ガイダンス項目の設定値のそれぞれを変化させる１単位を予め定めておき、前記要求条件設定ステップでは、前記要求条件に代えて前記ガイダンス項目のそれぞれの設定値に対して１単位を加算した値および１単位を減算した値をそれぞれ組み合わせた８１行のガイダンス項目と操業諸元とからなる要求条件行列を作成し、前記類似度算出ステップでは、前記要求条件設定ステップで定められた前記要求条件行列ごとに類似度を算出し、前記予測結果算出ステップでは、前記類似度と、前記入力変数と用いて、焼結鉱強度、焼結鉱生産率、排ガスのＥＰ前温度および排ガスのＥＰ前負圧の予測結果からなる８１行の予測結果行列を算出し、前記製造条件設定ステップでは、前記予測結果行列に前記管理値を全て満足するものがある場合には、前記予測結果を算出するのに用いた要求条件を焼結鉱製造条件として設定し、前記予測結果行列に前記管理値を全て満足するものがない場合には、前記管理値を満足しない項目において、最も管理値に近い予測結果を算出するのに用いた要求条件をガイダンス項目の設定値とし、このガイダンス項目の設定値に対して１単位を加算および１単位を減算した値を組み合わせた要求条件行列を作成して、再び、前記類似度算出ステップ、前記予測結果算出ステップおよび前記製造条件設定ステップを繰り返し実施する、［５］に記載の焼結鉱の製造方法。
［７］前記予測結果算出ステップでは、前記データベースに格納された入力変数のうち、下記（１）で算出される分散比が２以上となる入力変数を用いて前記局所回帰式を作成する、請求項［５］または請求項［６］に記載の焼結鉱の製造方法。

上記（１）式において、Ｓ_ｅ追加前は、回帰式に入力変数を追加する前の残差平方和であり、Ｓ_ｅ追加後は、回帰式に入力変数を追加した後の残差平方和であり、Φ_ｅ追加前は、入力変数を追加する前のΦ_ｅであり、Φ_ｅ追加後は、入力変数を追加した後のΦ_ｅであり、Φ_ｅは下記（２）式で算出される値である。
Φ_ｅ＝ｎ−２・・・（２）
上記（２）式において、ｎは、入力変数の数である。 The features of the present invention that solve such a problem are as follows.
[1] Water is added to a sintered raw material containing an iron-containing raw material, a CaO-containing raw material, and a coagulant to granulate, and the granulated sintered raw material is charged into a pallet carriage of a sintering machine to charge a layer. A control device that sets the manufacturing conditions of a sinter manufacturing facility that sinters the charge layer to produce sinter, and at least the coke ratio of the sinter raw material, the amount of granulated water, and so on. A requirement setting means for setting a requirement condition consisting of a setting value of a guidance item including a layer thickness of the charging layer and a traveling speed of a pallet carriage and an operation specification, a sinter strength as an output variable, and a sinter ore. The requirement condition using a database in which the production rate, the pre-EP temperature of the exhaust gas, the pre-EP negative pressure of the exhaust gas, the set value of the guidance item as an input variable, and the operation specifications are stored for each past operation condition. Using the similarity calculation means for calculating the similarity between the sinter and the input variable of the database for each past operating condition, and the sinter strength, the sinter production rate, and the exhaust gas using the similarity and the input variable. A local regression equation was created with each of the pre-EP temperature and the pre-EP negative pressure of the exhaust gas as output variables, and the sinter strength, sinter production rate, and EP of the exhaust gas were used using the local regression equation and the required conditions. Prediction result calculation means for calculating the prediction result of pre-temperature and pre-EP negative pressure of exhaust gas, and the calculated sinter strength, sinter production rate, pre-EP temperature of exhaust gas and pre-EP negative pressure of exhaust gas The prediction result calculation means includes the manufacturing condition setting means for setting the required condition as the sinter manufacturing condition when all of the above satisfy the predetermined control values, and the prediction result calculating means is among the input variables, the tolerance. The local regression equation was created using an input variable with a value greater than 0.1, and the production condition setting means used the calculated sinter strength, sinter production rate, pre-EP temperature of exhaust gas, and pre-EP of exhaust gas. If at least one of the negative pressure prediction results does not satisfy the control value, the setting value of the guidance item set by the requirement condition setting means is changed, and the similarity is calculated again by the similarity calculation means. , A control device for sinter manufacturing equipment that allows the sinter strength, sinter production rate, pre-EP temperature of exhaust gas, and pre-EP negative pressure of exhaust gas to be calculated by the prediction result calculation means.
[2] One unit for changing each of the set values of the guidance item is predetermined, and the required condition setting means sets one unit for each set value of the guidance item instead of the required condition. A requirement condition matrix consisting of 81 rows of guidance items and operation specifications, each of which is a combination of the added value and the subtracted value of 1 unit, is created, and the similarity calculation means is defined by the requirement condition setting means. The similarity is calculated for each requirement matrix, and the prediction result calculation means uses the similarity and the input variable to obtain the sinter strength, the sinter production rate, the pre-EP temperature of the exhaust gas, and the EP of the exhaust gas. The prediction result matrix of 81 rows consisting of the prediction results of the pre-negative pressure is calculated, and the manufacturing condition setting means calculates the prediction result when the prediction result matrix satisfies all the control values. If the requirements used in the above are set as the sinter production conditions and there is no one in the prediction result matrix that satisfies all the control values, the prediction that is closest to the control value in the items that do not satisfy the control values. The requirement condition used to calculate the result is used as the setting value of the guidance item, and a requirement condition matrix is created by combining the value obtained by adding 1 unit and subtracting 1 unit from the setting value of this guidance item, and again. , The similarity calculation means is used to calculate the similarity, and the prediction result calculation means is used to calculate the sinter strength, the sinter production rate, the pre-EP temperature of the exhaust gas, and the pre-EP negative pressure of the exhaust gas. 1] The control device for the sinter manufacturing equipment according to the above.
[3] The prediction result calculation means creates the local regression equation using the input variables stored in the database and having a variance ratio of 2 or more calculated in (1) below. The control device for the sintered ore manufacturing facility according to 1] or [2].

In the above (1), S e _{added before,} a residual sum of squares before adding the input variables to regression models, S _after e _added, residual square after adding the input variable in the regression equation the sum, _{before adding} [Phi _e, a previous [Phi _e to add an input variable, _{after adding} [Phi _e is [Phi _e after adding the input variable, [Phi _e is calculated by the following equation (2) Is the value to be.
Φ _e = n-2 ... (2)
In the above equation (2), n is the number of input variables.
[4] A raw material blending device for blending an iron-containing raw material, a CaO-containing raw material, and a coagulant to be a sintered raw material, a granulator for adding water to the sintered raw material to granulate, and the granulated product. A sintering machine that sinters the charging layer by charging the sintering raw material into the pallet carriage of the sintering machine to form the charging layer, a crusher, a cooler, a sieving device, and [1 ] To [3], the sinter manufacturing facility comprising the control device for the sinter manufacturing facility according to any one of [3].
[5] Water is added to a sintered raw material containing an iron-containing raw material, a CaO-containing raw material, and a coagulant to granulate, and the granulated sintered raw material is charged into a pallet carriage of a sintering machine to charge a layer. Is a method for producing a sinter, which comprises forming a sinter and sintering the sinter layer to produce a sinter. And the requirement setting step to set the requirement condition consisting of the setting value of the guidance item including the traveling speed of the pallet trolley and the operation specifications, and the sinter strength, sinter production rate, and exhaust gas before EP as output variables. Using a database in which the negative pressure before EP of temperature and exhaust gas, the set value of the guidance item as an input variable, and the operation specifications are stored for each past operation condition, the requirement condition and the input variable of the database are used. Sinter strength, sinter production rate, pre-EP temperature of exhaust gas and EP of exhaust gas using the similarity calculation step for calculating the similarity of the above for each past operating condition, and the similarity and the input variable. A local regression equation was created with each of the pre-negative pressure as an output variable, and the sinter strength, sinter production rate, pre-EP temperature of the exhaust gas, and pre-EP of the exhaust gas were used using the local regression equation and the required conditions. The prediction result calculation step for calculating the negative pressure prediction result, the sinter strength, the sinter production rate, the pre-EP temperature of the exhaust gas, and the pre-EP negative pressure prediction result of the exhaust gas calculated in the prediction result calculation step. When all of them satisfy the predetermined control values, the sinter is performed by the sinter production condition set in the production condition setting step for setting the required condition as the sinter production condition and the sinter production condition set in the production condition setting step. It has a sinter manufacturing step for manufacturing ore, and in the prediction result calculation step, the local regression equation is created using the input variable having a tolerance greater than 0.1 among the input variables, and the manufacturing is performed. In the condition setting step, if at least one of the calculated sinter strength, sinter production rate, pre-EP temperature of exhaust gas and negative pre-EP pressure of exhaust gas does not satisfy the control value, the requirement A method for producing sinter, wherein the set value of the guidance item used in the condition setting step is changed, and the similarity calculation step, the prediction result calculation step, and the production condition setting step are repeatedly carried out again.
[6] One unit for changing each of the set values of the guidance item is predetermined, and in the requirement condition setting step, one unit is added to each set value of the guidance item instead of the required condition. A requirement condition matrix consisting of 81 rows of guidance items and operation specifications, each of which is a combination of the value obtained and the value obtained by subtracting one unit, is created. Similarity is calculated for each condition matrix, and in the prediction result calculation step, using the similarity and the input variable, the sintered ore strength, the sintered ore production rate, the pre-EP temperature of the exhaust gas, and the pre-EP of the exhaust gas are used. An 81-row prediction result matrix composed of negative pressure prediction results is calculated, and in the manufacturing condition setting step, if the prediction result matrix satisfies all the control values, the prediction result is calculated. If the requirements used in the above are set as the sintered ore production conditions and there is no one in the prediction result matrix that satisfies all the control values, the prediction result closest to the control value in the items that do not satisfy the control values. The requirement condition used to calculate is used as the setting value of the guidance item, and a requirement condition matrix is created by combining the value obtained by adding 1 unit and subtracting 1 unit from the setting value of this guidance item. The method for producing a sintered ore according to [5], wherein the similarity calculation step, the prediction result calculation step, and the production condition setting step are repeatedly carried out.
[7] In the prediction result calculation step, the local regression equation is created using the input variables stored in the database and having a variance ratio of 2 or more calculated in (1) below. The method for producing a sintered ore according to item [5] or claim [6].

In the above (1), _{before adding} S e, a residual sum of squares before adding the input variable to the regression equation, _{after adding} S e is the residual sum of squares after adding the input variable in the regression equation and a, _{before adding} [Phi _e, a previous [Phi _e to add an input variable, _{after adding} [Phi _e is [Phi _e after adding the input variable, [Phi _e is calculated by the following equation (2) Value.
Φ _e = n-2 ... (2)
In the above equation (2), n is the number of input variables.

In the control device of the sinter manufacturing facility of the present invention, the sinter raw material that can satisfy each of the control values of the sinter strength of the product, the production rate of the sinter of the product, the pre-EP temperature of the exhaust gas, and the negative pressure before the EP of the exhaust gas. Determines the coke ratio, the amount of granulated water, the thickness of the charging layer, and the traveling speed of the pallet trolley. As described above, the control device of the sinter manufacturing facility according to the present invention adjusts the coke ratio of the sinter raw material, the amount of granulated water, the layer thickness of the charging layer, and the traveling speed of the pallet trolley. The adjustment allowance can be expanded compared to the conventional example, and as a result, the operating conditions can satisfy the control values of the strength of the sinter of the product, the production rate of the sinter of the product, the pre-EP temperature of the exhaust gas, and the negative pressure before the EP of the exhaust gas. It can increase the possibility of being determined.

It is a schematic diagram which shows an example of the sinter manufacturing equipment 10 which can carry out the sinter manufacturing method which concerns on this embodiment. It is a functional block diagram of the control device 90. It is a figure which shows an example of the operation record database 110. It is a flow chart which shows an example of the manufacturing process of a sinter using a requirement condition matrix. It is a flow chart which shows an example of the manufacturing process of a sinter using a requirement condition matrix. An example of the requirement condition matrix is shown. An example of the prediction result matrix is shown. It is a graph which shows the strength distribution of a sinter. It is a graph which shows the production amount distribution of a sinter.

Hereinafter, the present invention will be described through embodiments of the invention. FIG. 1 is a schematic view showing an example of a sinter manufacturing facility 10 including a control device according to the present embodiment. The iron-containing raw material 12 is conveyed from the raw material piles piled up in the yard 11 to the compounding tank 22 by the transfer conveyor 14. The iron-containing raw material 12 contains various brands of iron ore and dust generated in the steelworks.

The raw material blending device 20 includes a plurality of blending tanks 22, 24, 25, 26, 28. The iron-containing raw material 12 is stored in the compounding tank 22. The compounding tank 24 stores the CaO-containing raw material 16 containing limestone, quicklime, and the like, and the compounding tank 25 stores the MgO-containing raw material 17 containing dolomite, refined nickel slag, and the like. The compounding tank 26 stores a coagulant 18 containing coke breeze and anthracite crushed to a particle size of 1 mm or less using a rod mill. Further, in the compounding tank 28, the return ore (sintered ore sieving powder) having a particle size of 5 mm or less, which is under the sieve of the sinter, is stored. A predetermined amount of each raw material is cut out from the mixing tanks 22 to 28 of the raw material mixing device 20, and these are mixed to form a sintered raw material. The sintered raw material is conveyed to the drum mixer 36 by the conveyor 30. The MgO-containing raw material 17 is an optional blended raw material and may or may not be blended with the sintered raw material.

The sintered raw material conveyed to the drum mixer 36 is charged into the drum mixer 36, and an appropriate amount of water 34 is added to granulate the pseudo particles having an average particle size of, for example, 3.0 to 6.0 mm. The granulated sintering raw material is conveyed to the sintering raw material charging device 42 of the sintering machine 40 by the conveying conveyor 38. The drum mixer 36 is an example of a granulator that granulates a sintered raw material, and a pelletizer granulator may be used instead of the drum mixer 36. In the present embodiment, the average particle size of the pseudo particles is the arithmetic mean particle size, and is Σ (Vi × di) (where Vi is the abundance ratio of particles in the i-th particle size range, and di is i. It is a particle size defined by), which is a representative particle size of the second particle size range.

The sintering machine 40 is, for example, a downward suction type dwightroid sintering machine. The sintering machine 40 includes a sintering raw material charging device 42, an endlessly movable pallet carriage 44, an ignition furnace 46, a city gas supply device 47, and a windbox 48. The sintered raw material granulated from the sintered raw material charging device 42 is charged into the pallet carriage 44, and a charged layer of the sintered raw material is formed.

The charge layer is ignited in the ignition furnace 46. Then, by sucking air through the windbox 48, the charging layer takes in the city gas and oxygen supplied from the city gas supply device 47 provided above into the charging layer, and the city gas and the coagulant are taken into the charging layer. A combustion / melting zone is formed by burning 18. The combustion and melting zones move below the charging layer as the pallet carriage 44 moves due to the suction of air by the wind box 48. In this way, the charging layer is sintered to form a sintered cake. In the present embodiment, the city gas is an example of a gaseous fuel, and instead of the city gas, a blast furnace gas, a coke oven gas, a blast furnace / coke oven mixed gas, a converter gas, a city gas, a natural gas, a methane gas, an ethane gas, and a propane gas. , A flammable gas selected from shale gas and mixed gases thereof may be used.

The EP50 is connected to the exhaust gas flow path 49 of the windbox 48. The air sucked through the windbox 48 is discharged as exhaust gas from the exhaust gas flow path 49 to the EP50. A thermometer and a pressure gauge are provided in the exhaust gas flow path 49, and the pre-EP temperature of the exhaust gas and the pre-EP negative pressure of the exhaust gas are measured by these. EP50 removes sintered dust contained in exhaust gas. The exhaust gas from which the sintered dust has been removed is released to the atmosphere through a chimney (not shown) or the like.

The sintered cake is crushed by a crusher 60 into a sintered ore. The sinter crushed by the crusher 60 is cooled by the cooler 62. The sinter cooled by the cooler 62 is sieved by a sieving device 70 having a plurality of sieves, and is sieved into a product sinter 72 having a particle size of more than 5 mm and a return ore 74 having a particle size of 5 mm or less. Will be done.

The product sintered ore 72 is conveyed to the blast furnace 80 by the transfer conveyor 76, and is charged into the blast furnace 80 as a raw material for the blast furnace. On the other hand, the return ore 74 is conveyed to the compounding tank 28 of the raw material compounding apparatus 20 by the transfer conveyor 78. In the present embodiment, the particle size of the product sintered ore 72 and the particle size of the returned ore 74 mean the particle size that can be sieved by a sieve. For example, if the particle size exceeds 5 mm, a sieve having a mesh size of 5 mm is used. The particle size to be sieved on the sieve, and the particle size of 5 mm or less is the particle size to be sieved under the sieve using a sieve having a mesh size of 5 mm. In the following description, the product sintered ore 72 may be simply referred to as "sintered ore".

FIG. 2 is a functional block diagram of the control device 90. The control device 90 includes a storage unit 92, a processing unit 94, an input unit 96, and an output unit 98. The control device 90 is, for example, a general-purpose computer such as a workstation or a personal computer.

The storage unit 92 is composed of a flash memory capable of updating and recording, a hard disk connected by a built-in or data communication terminal, an information recording medium such as a memory card, a reading / writing device thereof, and the like. The storage unit 92 stores in advance a program for realizing various functions of the sinter manufacturing facility 10, data used for executing the program, and the like.

The operation record database 110 is further stored in the storage unit 92. FIG. 3 is a diagram showing an example of the operation record database 110. The operation record database 110 contains the sinter strength, the sinter production rate, the pre-EP temperature of the exhaust gas, the negative pre-EP pressure of the exhaust gas, the Fe concentration, the SiO ₂ concentration, and the Al ₂ O ₃ concentration for each past operating condition. , CaCO ₃ concentration, CaO concentration, particle size, city gas intensity, coke ratio, granulated water content, layer thickness of charging layer, traveling speed of pallet trolley 44, etc. There is. Here, the sinter strength, the sinter production rate, the pre-EP temperature of the exhaust gas and the pre-EP negative pressure of the exhaust gas are stored as output variables, and the Fe concentration, the SiO ₂ concentration, the Al ₂ O ₃ concentration, and the CaCO ₃ concentration are stored. , CaO concentration, particle size, lime compounding ratio, city gas intensity, coke ratio, granulated water content, layer thickness of charging layer and traveling speed of pallet carriage 44 are stored as input variables. In the following description, the pre-EP temperature of the exhaust gas is simply referred to as “pre-EP temperature”, and the pre-EP negative pressure of the exhaust gas is simply referred to as “pre-EP negative pressure”. Further, the granulated water content represents the water content contained in the pseudo particles obtained by granulating the sintering raw material.

For example, No. 1 of the operation record database 110. 1 is No. Has been operating conditions described in the first row (operational specifications and guidance items) at carrying out the production of sintered ore, the strength of the sintered ore produced is Y ₁ ^1, its sinter production rate Y ₂ is ^1, EP before temperature is _Y ^{3 1,} EP before negative pressure indicates a past of preparation was _Y ^{4 1.} In the present embodiment, among the input variables, the Fe concentration, SiO ₂ concentration, Al ₂ O ₃ concentration, CaCO ₃ concentration, Ca O concentration, particle size, and city gas basic unit cannot be adjusted by the operator, or are adjusted by the operator. Items that are not included, and these are collectively referred to as operation specifications. On the other hand, among the input variables, the coke ratio, the amount of granulated water, the layer thickness of the charge layer, and the traveling speed of the pallet carriage 44 are management items for which the operator has conventionally set conditions, and these are collectively guidance items. And. In this embodiment, the guidance items are the coke ratio, the granulated water content, the layer thickness of the charging layer, and the traveling speed of the pallet carriage 44. However, the guidance items are not limited to this, and may include at least the coke ratio, the amount of granulated water, the layer thickness of the charging layer, and the traveling speed of the pallet carriage 44, and may include other conditions. Further, Fe concentration, SiO ₂ concentration, Al ₂ O ₃ concentration, CaCO ₃ concentration, CaO concentration, particle size, and city gas basic unit are also examples of operating specifications, and the operating specifications are other components of the sintered raw material. It may include the component concentration and other conditions of the equipment of the sinter production facility 10.

See FIG. 2 again. The processing unit 94 is a CPU or the like, and controls the operation of the sinter manufacturing facility 10 by using an input signal output from the input unit 96 and a program or data stored in the storage unit 92. The processing unit 94 includes a requirement condition setting means 100, a similarity calculation means 102, a prediction result calculation means 104, and a manufacturing condition setting means 106. Each of these means implements a requirement condition setting step, a similarity calculation step, a prediction result calculation step, a production condition setting step, and a sinter production step using the program and data stored in the storage unit 92.

The input unit 96 is, for example, an input device such as a keyboard, a mouse, a touch panel, and various switches. The input unit 96 outputs an input signal corresponding to an input operation from the operator to the processing unit 94.

The output unit 98 outputs a control signal of the device output from the processing unit 94 to each device in the sinter manufacturing facility 10. As a result, the control device 90 controls the operation of each device of the sinter manufacturing facility 10. Further, the output unit 98 may be a display device such as an LCD or a CRT display.

When the control device 90 receives input from the operator of the control values of the product sinter strength, the product sinter production rate, the exhaust gas pre-EP temperature, and the exhaust gas pre-EP negative pressure, the requirement condition setting step is performed. , The similarity calculation step, the prediction result calculation step, and the manufacturing condition setting step are carried out, and the coke ratio of the sintered raw material, the granulated water content, and the load, which are the control items for which the operator has set the conditions in the past, are carried out. Calculate the set value of the management item that can satisfy the control value with respect to the layer thickness of the incoming layer and the traveling speed of the pallet trolley. Next, the operation of each facility in the sinter manufacturing facility 10 is controlled based on the set value of the control item. For example, the control device 90 controls the amount of each raw material cut out from the blending tanks 22, 24, 25, 26, 28 of the raw material blending device 20 so that the coke ratio of the sintered raw material becomes the calculated set value. Then, the coke ratio of the sintered raw material is adjusted. Further, the control device 90 adjusts the amount of water added by the drum mixer 36 so that the amount of granulated water content of the pseudo particles becomes the calculated set value. Further, the control device 90 controls the sintering raw material charging device 42 and the pallet carriage 44 of the sintering machine 40 so that the layer thickness of the charging layer and the traveling speed of the pallet carriage become the calculated set values. , Adjust the layer thickness of the charging layer or the traveling speed of the pallet carriage 44.

Next, the requirement condition setting step, the similarity calculation step, the prediction result calculation step, and the manufacturing condition performed by the requirement condition setting means 100, the similarity calculation means 102, the prediction result calculation means 104, and the manufacturing condition setting means 106. The processing of the setting step and the sinter manufacturing step will be described. When the input unit 96 receives the management value of the management item, the operation specifications, and the set value of the guidance item from the operator, the input unit 96 outputs these information to the processing unit 94. The setting value of the guidance item does not have to be input by the operator every time, and the optimum value of the guidance item derived last time may be used. In addition, among the operation specifications, the set value regarding the component concentration of the sintered raw material is measured at the timing when the raw material pile stored in the yard 11 is switched, and the set value is used while the raw material pile is used. May be. Upon acquiring these information from the input unit 96, the requirement condition setting means 100 executes the requirement condition setting step and sets the requirement condition composed of the set value of the guidance item and the operation specifications.

The similarity calculation means 102 carries out the similarity calculation step on condition that the required conditions are set. In the similarity calculation step, the similarity calculation means 102 calculates the similarity for each of the past operation conditions stored in the operation record database 110. Here, the similarity is a value determined from the similarity function including the function of the distance indicating the closeness between the required condition and the input variable for each past operating condition stored in the operation record database 110.

In the similarity calculation step, the similarity calculation means 102 first converts the required condition into the input vector shown in the following equation (3).

_{^{X r = [X 1 r,}} X 2 r, ····, X M r] T ··· (3)
The similarity calculation means 102 creates a regression equation for calculating each of the output variables stored in the operation record database 110 shown in FIG. This regression equation is expressed by the following equation (4).

Y = b + a ₁ x X ₁ + a ₂ x X ₂ + ... + a _M x X _M ... (4)
However, in the above equation (4), Y is an output variable, and b, a ₁ , a ₂ , ... a _M are parameters of the regression equation.

The partial regression coefficient vector of the following equation (5), which is obtained by extracting only the coefficient from this parameter by removing the constant b, is used as the influence coefficient used in the distance calculation.

α = [a ₁ , a ₂ , ... a _M ] ^T ... (5)
The similarity calculation means 102 calculates the distance between the required condition and the input variable stored in the operation record database 110 for each past operation condition. For this purpose, first, for calculating the distance L from the input variable of the above equation (3) with respect to _{a certain point x = [x 1} , x ₂ , ... x _M ] ^T in the input space (requirement condition space). Determine the distance function. The distance function for calculating the distance L is expressed by the following equation (6) in consideration of the influence coefficient of the above equation (5).

(6) is a requirement, the absolute value of the difference between the input variables stored in the operational result database 110, a multiplied by the absolute value of the influence coefficients a _m, by performing a process of adding up all of the conditions There is. (5) influence coefficient given by equation (partial regression coefficients) a _m, since also considered contribution of each input variable X _M with respect to the change amount of the sintered ore strength as an output variable, (6) a function of the distance Represents a weighted distance that takes into account its contribution.

Using the distance function shown by the equation (6), the distance from the required condition is calculated for each past operating condition stored in the operating record database 110. That is, the distance from ^Xr indicating the required condition is obtained for each past operating condition stored in the operation record database 110. Specifically, n-th (n = 1, 2, · · ·, N) and operating conditions, the distance ^{L n} between ^{X r} indicating the input operation specifications and guidance items, the following equation (7) Can be calculated with.

L ⁿ = L (X ⁿ , X ^r , α) ... (7)
However, in the equation (5), X ⁿ = [X ₁ ⁿ , X ₂ ⁿ , ... X _M ⁿ ] ^T (n = 1, 2, ..., N).

Further, the input variables in 1~N th operating conditions, summarized the distance l between the X ^r representing the requirements, it can be expressed by the following equation (8).

l = [L ¹ , L ² , ..., L ^N ] ^T ... (8)
The similarity calculation means 102 calculates the similarity for each of the past operating conditions after calculating the distance from the required conditions. The similarity function W that expresses the proximity from the requirement condition for that purpose is defined as the following equation (9).

W (L, p, l) = exp (-[L / {p · σ (l)}] ² ) ... (9)
However, in the above equation (9), σ (l) represents the standard deviation of l used for normalization, p is a parameter for converting the distance into the similarity, and in the present embodiment, it is the bandwidth. do. The similarity calculation means 102 uses the similarity function W of the equation (9) to obtain the similarity between the required condition and the operating condition stored as an input variable for each of the past operating conditions stored in the operating record database 110. calculate. And requirements, similarity ^{W n} of an input variable of the n-th stored in operating results database 110 (n = 1, 2, · · ·, N) can be calculated using the following equation (10).

W ⁿ = W (L ⁿ , p, l) ... (10)
However, in the above equation (10), n = 1, 2, ..., N.

The degree of similarity between the required condition and the input variable in the 1st to Nth operating conditions can be summarized by the following equation (11).

w = [W ¹ , W ² , ..., W ^N ] ^T ... (11)
The prediction result calculation means 104 executes the output variable prediction step on the condition that the above equation (11), which is the similarity, is calculated in the similarity calculation step. The prediction result calculation means 104 creates a regression equation for calculating an output variable using the input variable of the operation record database 110. This regression equation can be expressed by the following equation (12).

Y = b + a ₁ x X ₁ + a ₂ x X ₂ + ... + a _M x X _M ... (12)
In the above equation (12), Y is the sinter strength, the sinter production rate, the pre-EP temperature and the pre-EP negative pressure. This formula is a formula that can calculate the prediction results of sinter strength, sinter production rate, pre-EP temperature, and pre-exhaust EP negative pressure, which are output variables. The prediction result calculation means 104 uses the input variables of the operation record database 110 and the similarity w to set the parameters (b, a ₁ , a ₂ , ... a _M ) of the regression equation of equation (12) to the similarity. Calculated using the weighted least squares method with w as the weight. As described above, the regression equation represented by the equation (12) calculated by using the weighted least squares method with the similarity w as the weight is a local regression equation.

Further, in calculating the parameters of the regression equation of the above equation (12), the prediction result calculation means 104 uses an input variable having a tolerance greater than 0.1 among the input variables of the operation record database 110. The tolerance will be described below. Table 1 shows n data in which the output variable and the input variable are associated with each other. Tolerance will be explained using this data.

In the data shown in Table 1, the multiple regression calculation formula of the output variable using the input variables x1 and x2 is the following formula (13).

In such a case, the residual e _i is the difference between the data i-th actual and predicted values is expressed by the following equation (14), the residual sum of squares S _e is expressed by the following equation (15), the deviation The sum of squares S _yy is represented by the following equation (16) and the following equation (17).

The multiple correlation coefficient R is expressed by the following equation (18) using the residual sum of squares Se and the deviation sum of squares Syy.

Tolerance is a value represented by the following equation (19) using the multiple correlation coefficient R calculated by the above equation (18). An input variable with a tolerance greater than 0.1 is determined to have no multicollinearity. If an input variable having multicollinearity is used, the accuracy of the regression equation deteriorates. Therefore, the prediction result calculation means 104 uses an input variable having a tolerance greater than 0.1 when calculating the parameters of the regression equation of the above equation (12).

T = 1-R ² ... (19)
In equation (19), T is tolerance.

Further, in calculating the parameters of the regression equation of the above equation (12), it is preferable to use the input variables having a variance ratio F of 2 or more among the input variables of the operation record database 110. Here, the dispersion ratio F is a value calculated by the following equation (1).

In the above (11), S e _{added before,} a residual sum of squares before adding the input variables to regression models, S _after e _added, residual square after adding the input variable in the regression equation It is a sum. For example, when confirming the variance ratio of the input variable x1, the sum of squared residuals calculated using data other than x1 is _{before the addition of} Se, and the residual sum of squares calculated using the data including x1. The sum is _{after the addition} of Se. Further, Φ _e is a value calculated by the following equation (2), _{before adding Φ e,} it is Φ e before adding an input variable _{, and after adding} Φ e, Φ after adding an input variable. _e .

Φ _e = n-2 ... (2)
In the above equation (2), n is the number of input variables. Also in [Phi _e, for example, in the case of confirming the dispersion ratio of the input variables x1, next _before other _additional [Phi _e is [Phi _e calculated by the data except x1, calculated in data including x1 [Phi _e Is _{after adding} Φ e.

The prediction result calculation means 104 repeatedly carries out the above processing to create local regression equations capable of calculating the prediction results of the sinter strength, the sinter production rate, the pre-EP temperature and the pre-EP negative pressure, respectively. The prediction result calculation means 104 calculates the prediction results of the sinter strength, the sinter production rate, the pre-EP temperature and the pre-EP negative pressure by using the local regression equation and the required conditions.

The manufacturing condition setting means 106 carries out the manufacturing condition setting step on the condition that the predicted results of the sinter strength, the sinter production rate, the pre-EP temperature and the pre-EP negative pressure are calculated in the prediction result calculation step. In the manufacturing condition setting step, does the manufacturing condition setting means 106 satisfy all of the calculated sinter strength, sinter production rate, pre-EP temperature and pre-EP negative pressure prediction results satisfy these control values? Judge whether or not. The sinter strength, the sinter production rate, the pre-EP temperature, and the pre-EP negative pressure control values are predetermined and stored in the storage unit 92.

In the present embodiment, the lower limit of the sinter strength is stored in the storage unit 92 as the control value of the sinter strength. The manufacturing condition setting means 106 reads the lower limit value of the sinter strength from the storage unit 92, and compares the calculated prediction result of the sinter strength with the lower limit value. When the predicted result of the sinter strength is larger than the lower limit value, the manufacturing condition setting means 106 determines that the predicted result of the sinter strength satisfies the control value. On the other hand, when the prediction result of the sinter strength is equal to or less than the lower limit value, the manufacturing condition setting means 106 determines that the prediction result of the sinter strength does not satisfy the control value.

Similarly, the storage unit 92 has a lower limit of the sinter production rate as a control value of the sinter production rate, a lower limit of the pre-EP temperature as a control value of the pre-EP temperature, and a control of the negative pressure before EP. The negative pressure before EP as a value is stored. The manufacturing condition setting means 106 reads out these control values from the storage unit 92, compares the calculated prediction result with the control value, and determines whether or not the control value is satisfied.

The manufacturing condition setting means 106 provides guidance on the optimum conditions from the required conditions in which all of the prediction results of the pre-EP temperature, pre-EP negative pressure, sinter strength and sinter production rate satisfy predetermined control values. Set the item to sinter production conditions. Here, the guidance item of the optimum condition means that the prediction result having the highest pre-EP temperature is obtained. When multiple prediction results with the highest pre-EP temperature are obtained, the one with the lowest pre-EP negative pressure may be set as the optimum condition, and priority is given to sinter strength and sinter production rate in that order. The degree goes down.

On the other hand, when at least one of the predicted results of the sinter strength, the sinter production rate, the pre-EP temperature and the pre-EP negative pressure does not satisfy the predetermined control values, the manufacturing condition setting means 106 is required. In the setting step, the setting value of the guidance item used for setting the requirement condition is changed, and the similarity calculation step 102 is performed again by using the changed setting value of the guidance item, and the prediction result is predicted. The calculation means 104 is made to carry out the prediction result calculation step.

The similarity calculation step, the prediction result calculation step, and the manufacturing condition setting step may be repeated a predetermined number of times. Further, the setting range of the guidance item and one unit to be changed may be predetermined for each guidance item, and all combinations of the guidance items within the setting range may be implemented. Sinter strength, sinter production rate, before EP even if the predetermined end conditions are satisfied, or even if all combinations of guidance items within the set range are set as required conditions and the above-mentioned treatment is performed. When all the prediction results of the temperature and the negative pressure before EP of the exhaust gas do not satisfy the predetermined control values, the manufacturing condition setting means 106 ends the process.

The processing unit 94 carries out the sinter production step on the condition that the guidance item is set in the sinter production condition in the production condition setting step. The processing unit 94 controls each device of the sinter manufacturing facility 10 under the set sinter manufacturing conditions to manufacture the sinter. As a result, not only the sinter strength and the sinter production rate but also the pre-EP temperature and the pre-EP negative pressure can be produced while satisfying predetermined control values.

As described above, the control device 90 of the sinter manufacturing facility according to the present embodiment satisfies each of the control values of the sinter strength, the sinter production rate, the pre-EP temperature and the pre-EP negative pressure. , Coke ratio, granulated water content, layer thickness of charging layer and traveling speed of pallet trolley, which are guidance items. In the control device 90 of the sinter manufacturing equipment according to the present embodiment, since these guidance items are adjusted, the adjustment allowance of the operating conditions is wider than in the conventional example in which only the traveling speed of the pallet carriage is adjusted, and as a result, the product is produced. It is possible to increase the possibility that operating conditions that can satisfy all of the sinter strength, the production rate of the product sinter, the pre-EP temperature and the pre-EP negative pressure control values can be determined.

Further, the control device 90 of the sinter manufacturing facility according to the present embodiment uses a local regression equation using an operation record database 110 accumulating past operation conditions to determine the coke ratio of the sinter raw material, the granulated water content, and the like. Determine the layer thickness of the charging layer and the traveling speed of the pallet trolley. As described above, in the method for producing sinter according to the present embodiment, since the operator does not set these conditions through trial and error, the fluctuation of the strength and composition of the sinter due to individual differences of the operator is reduced, and the sinter is baked. Improves the quality of sinter.

In the above example, an example of setting one requirement condition in the requirement condition setting step is shown, but the present invention is not limited to this. When the requirement condition setting means 100 acquires information indicating a guidance item from the input unit 96 in the requirement condition setting step, the acquired guidance item is set as a setting value of the guidance item, and one unit is set for each setting value of the guidance item. A requirement matrix may be calculated by combining three values having different values, and the requirement matrix may be used as a requirement condition. In this case, one unit for changing each of the guidance items is predetermined and stored in the storage unit 92.

Next, the processing of the similarity calculation step using the requirement condition matrix, the output variable prediction step, and the manufacturing condition setting step will be described with reference to FIGS. 4 and 5. 4 and 5 are examples of flow charts for setting sinter production conditions. The flow shown in FIGS. 4 and 5 is started on the condition that the processing unit 94 has acquired the input signal indicating the guidance item.

The requirement condition setting means 100 first records the flag 1 in the input signal (S101). This flag is a flag that identifies the processing status in the sinter manufacturing process using the requirement matrix. The input unit 96 adds and subtracts a predetermined unit for each set value of the guidance item, calculates three values for each of the guidance items, and combines these requirements for 81 lines. Create a matrix (S102). FIG. 6 shows an example of a requirement condition matrix. As shown in FIG. 6, the operation specifications are constant values in the requirement condition matrix, and only the guidance items have different values in each row.

The similarity calculation means 102 carries out the similarity calculation step using the requirement condition matrix, and the prediction result calculation means 104 carries out the prediction result calculation slap using the similarity. As a result, the prediction result calculation means 104 calculates a prediction result matrix composed of 81 rows of prediction results corresponding to the requirement condition matrix (S103). FIG. 7 shows an example of the prediction result matrix. FIG. 7 shows an example of a prediction result matrix associated with a guidance item in the requirement condition matrix, but the present invention is not limited to this, and the guidance item may not be associated with the guidance item.

The manufacturing condition setting means 106 determines whether or not the flag of the input signal is 1 (S104). When the manufacturing condition setting means 106 determines that the flag of the input signal is 1, the process proceeds to S105 (S104: Yes). On the other hand, when the manufacturing condition setting means 106 determines that the flag of the input signal is not 1, the process proceeds to S112 (S104: No).

The manufacturing condition setting means 106 determines whether or not there is a prediction result including the pre-EP temperature exceeding the lower limit value in the prediction result matrix (S105). The lower limit of the pre-EP temperature is a control value of the exhaust gas temperature before the EP50 determined for the purpose of preventing a failure of the EP50, and the control value is preset and stored in the storage unit 92. ..

The manufacturing condition setting means 106 reads the lower limit value from the storage unit 92 and compares the pre-EP temperature in the prediction result matrix with the lower limit value. When the manufacturing condition setting means 106 has a prediction result including the pre-EP temperature exceeding the lower limit in the prediction result matrix (S105: Yes), the manufacturing condition setting means 106 predicts the pre-EP temperature exceeds the lower limit. Are all extracted (S106).

On the other hand, when there is no pre-EP temperature exceeding the lower limit in the prediction result matrix (S105: No), the manufacturing condition setting means 106 specifies the maximum value of the pre-EP temperature in the prediction result matrix, and predicts including the maximum value. The requirement condition used for calculating the result is set as the setting value of the guidance item (S107), and the process is returned to S102. In S102, the requirement condition setting means 100 again adds 1 unit and subtracts 1 unit to the set value of the guidance item, calculates three values for each, and recreates the requirement condition matrix by combining these. Then, the processes of S103 to S105 are repeatedly carried out.

In S106, after extracting the prediction result in which the pre-EP temperature exceeds the lower limit value, the manufacturing condition setting means 106 adds 1 to the flag of the input signal and sets the flag of the input signal to 2 (S108).

The manufacturing condition setting means 106 determines whether or not the prediction result extracted in S106 includes a prediction result including the pre-EP negative pressure that is less than the upper limit of the pre-EP negative pressure (S109). The upper limit value of the negative pressure before EP is a predetermined control value of exhaust gas negative pressure for the purpose of preventing failure of EP50, and the control value is preset and stored in the storage unit 92. ..

When the prediction result extracted in S106 includes the prediction result including the pre-EP negative pressure less than the upper limit value (S109: Yes), the manufacturing condition setting means 106 has the upper limit of the pre-EP negative pressure. All the prediction results that are less than the value are extracted (S110). On the other hand, when the prediction result extracted in S106 does not have an EP pre-negative pressure less than the upper limit value (S109: No), the manufacturing condition setting means 106 specifies the minimum value of the EP pre-negative pressure in the prediction result extracted in S106. Then, the requirement condition used for calculating the prediction result including the minimum value is set as the setting value of the guidance item (S111), and the process is returned to S102. In S102, the requirement condition setting means 100 again adds 1 unit and subtracts 1 unit to the set value of the guidance item, calculates three values for each, and recreates the requirement condition matrix by combining these. do. When the signal is returned to S102 by the processing of S111, the flag of the input signal is set to 2 in the processing of S108. Therefore, in the subsequent flow, the processing of S105 to S108 is not executed by the processing of S104. The manufacturing condition setting means 106 extracts the prediction result in which the pre-EP temperature exceeds the lower limit value from the calculated prediction result matrix (S113), and proceeds to the process in S109.

In S110, after extracting the prediction result that the negative pressure before EP is less than the upper limit value, the manufacturing condition setting means 106 adds 1 to the flag of the input signal and sets the flag of the input signal to 3 (S114).

The manufacturing condition setting means 106 determines whether or not the prediction result extracted in S110 includes a prediction result including a sinter strength exceeding the lower limit of the sinter strength (S115). The lower limit of the sinter strength is, for example, a control value of the sinter strength required when the product sintered ore 72 is charged into the blast furnace, and the control value is preset and stored in the storage unit 92. It is stored in.

When the production condition setting means 106 has a prediction result including a sinter strength exceeding the lower limit value in the prediction result extracted in S110 (S115: Yes), the production condition setting means 106 has a sinter strength lower limit value. All the prediction results exceeding the above are extracted (S116). On the other hand, when there is no prediction result including the sinter strength exceeding the lower limit value in the prediction result extracted in S110 (S115: No), the manufacturing condition setting means 106 determines the sinter strength in the prediction result extracted in S110. The requirement condition used for specifying the maximum value and calculating the prediction result including the maximum value is set as the setting value of the guidance item (S117), and the process is returned to S102. In S102, the requirement condition setting means 100 again adds 1 unit and subtracts 1 unit to the set value of the guidance item, calculates three values for each, and recreates the requirement condition matrix by combining these. do. When the signal is returned to S102 by the processing of S117, the flag of the input signal is set to 3 in the processing of S114, so that the processing of S105 to S114 is not executed by the processing of S104 and S112 in the subsequent flow. The manufacturing condition setting means 106 extracts a prediction result in which the pre-EP temperature exceeds the lower limit value and the pre-EP negative pressure is less than the upper limit value from the calculated prediction result matrix (S119), and proceeds to the process in S115.

After extracting the prediction result in which the sinter strength exceeds the lower limit value in S116, the manufacturing condition setting means 106 adds 1 to the flag of the input signal and sets the flag of the input signal to 4 (S120).

The production condition setting means 106 determines whether or not the prediction result extracted in S116 includes a prediction result that exceeds the lower limit of the productivity of the sinter (S121). The lower limit of the sinter production rate is a control value of the sinter production rate required in the production of the sinter, and the control value is predetermined and stored in the storage unit 92. There is.

When the production condition setting means 106 has a prediction result including the production rate of sinter exceeding the lower limit value in the prediction result extracted in S116 (S121: Yes), the production condition setting means 106 has a lower limit value of the production rate. All the prediction results exceeding the above are extracted (S122). On the other hand, when there is no prediction result including the sinter strength exceeding the lower limit value in the prediction result extracted in S116 (S121: No), the manufacturing condition setting means 106 maximizes the production rate in the prediction result extracted in S116. The requirement condition used for specifying the value and calculating the prediction result including the maximum value is set as the setting value of the guidance item (S123), and the process is returned to S102. In S102, the requirement condition setting means 100 again adds 1 unit and subtracts 1 unit to the set value of the guidance item, calculates three values for each, and recreates the requirement condition matrix by combining these. do. When the signal is returned to S102 by the processing of S123, the flag of the input signal is set to 4 in the processing of S120. Therefore, in the subsequent flow, the processing of S105 to S120 is not executed by the processing of S104, S112, and S118.

After extracting the prediction result in which the production rate exceeds the lower limit in S122, the production condition setting means 106 sets the requirement condition used for calculating the prediction result in the sinter production condition (S125). The processes shown in FIGS. 4 and 5 are completed. Even in the process using the requirement condition matrix shown in FIGS. 4 and 5, the process of returning to S102 by S107, S111, S117 and S123 may be performed a predetermined number of times, and may be performed a predetermined number of times. However, if there is no prediction result matrix that satisfies the control value, the error display corresponding to each step may be output. For example, when the output unit 98 is a display device such as an LCD or a CRT display, and the number of times S107 is executed is a predetermined number of times, a prediction result matrix including the pre-EP temperature exceeding the lower limit value is displayed. The output unit 98 may output the fact that there is no such thing. As a result, the operator can immediately grasp which control value should be reviewed when reviewing the control value. In addition, the setting range of the guidance item is set in advance for each guidance item, all the combinations of the guidance items within the set range are implemented, and even if all the combinations are implemented, the pre-EP temperature exceeding the lower limit value. An error display may be output when there is no prediction result matrix containing. Further, if there is no prediction result matrix that satisfies the control value even if the process of returning to S102 by S107, S117, and S123 is performed a predetermined number of times, the process may proceed to S108, S114, S120, and S123. In this case, in S113, S119, and S124, the value closest to the control value is extracted for the item in the skipped process.

FIG. 8 is a graph showing the strength distribution of the sinter. FIG. 8A shows the strength distribution of the sinter before producing the sinter under the production conditions determined by the treatment using the requirement condition matrix shown in FIGS. 4 and 5. FIG. 8B shows the strength distribution of the sinter after the sinter is produced under the production conditions determined by the treatment using the requirement condition matrix shown in FIGS. 4 and 5. In both FIGS. 8A and 8B, the horizontal axis is the sinter strength (%) and the vertical axis is the frequency. Under the manufacturing conditions determined by the process using the requirement condition matrix, the pre-EP temperature exceeded the predetermined lower limit value, and the pre-EP negative pressure also became less than the predetermined upper limit value.

As shown in FIG. 8, by producing the sinter under the production conditions determined by the treatment using the requirement matrix, the average value of the sinter strength can be increased and the sinter strength fluctuates. Has also become smaller. In this way, by determining the production conditions and producing the sinter by the treatment using the requirement condition matrix shown in FIGS. 4 and 5, the production of the sinter having a weak strength is suppressed, and the sinter is performed. It was confirmed that the quality of the ore was improved.

FIG. 9 is a graph showing the production amount distribution of sinter. FIG. 9A shows the production amount distribution of the sinter before producing the sinter under the production conditions determined by the treatment using the requirement condition matrix shown in FIGS. 4 and 5. FIG. 9B shows the production amount distribution of the sinter after the sinter is produced under the production conditions determined by the treatment using the requirement condition matrix shown in FIGS. 4 and 5. In both FIGS. 8A and 8B, the horizontal axis represents the amount of sinter produced (t / h), and the vertical axis represents the frequency. The value obtained by dividing the amount of sinter produced by the area of the pallet carriage of the sinter (m ² ) is the sinter production rate.

As shown in FIG. 9, by producing the sinter under the production conditions determined by the treatment using the requirement matrix, the average value of the production amount of the sinter can be increased, and the sinter can be produced. Fluctuations in production have also become smaller. As described above, it was confirmed that the productivity of the sinter was increased by producing the sinter by determining the production conditions by the treatment using the requirement condition matrix shown in FIGS. 4 and 5.

In the examples shown in FIGS. 4 and 5, it is an example of determining whether or not each control value is satisfied in the order of pre-EP temperature, pre-EP negative pressure, sinter strength, and sinter production rate. Indicated. However, the order of pre-EP temperature, pre-EP negative pressure, sinter strength and sinter production rate is not limited, and the order of checking these items is not particularly limited.

Further, in the present embodiment, an example is shown in which the control device 90 carries out the processing of each step in the method for producing sinter. However, the present invention is not limited to this, and a part or all of the processing of each step may be performed by the operator.

Although the present invention has been described above using the embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments. It will be apparent to those skilled in the art that various changes or improvements can be made to the above embodiments. It is clear from the description of the claims that such modified or improved forms may also be included in the technical scope of the present invention.

In addition, the execution order of the operations, procedures, and steps in the devices, systems, and methods shown in the claims, the specification, and the drawings is specified as "before", "prior to", and the like. It can be realized in any order as long as the output of the previous process is not used in the subsequent process. Even if the scope of claims, the specification, and the flow in the drawings are explained using "first", "next", etc. for convenience, it does not mean that it is essential to carry out in this order. do not have.

10 Sintered ore production equipment 11 yards 12 Iron-containing raw material 14 Conveyor conveyor 16 CaO-containing raw material 17 MgO-containing raw material 18 Coagulant 20 Raw material compounding device 22 Mixing tank 24 Mixing tank 25 Mixing tank 26 Mixing tank 28 Mixing tank 30 Conveyor 34 Water 36 Drum mixer 38 Conveyor conveyor 40 Sinter machine 42 Sintered raw material charging device 44 Pallet trolley 46 Ignition furnace 47 City gas supply device 48 Windbox 49 Exhaust flow path 50 EP
60 Crusher 62 Cooler 70 Sieving device 72 Product Sintered ore 74 Return ore 76 Conveyor conveyor 78 Conveyor conveyor 80 Blast furnace 90 Control device 92 Storage unit 94 Processing unit 96 Input unit 98 Output unit 100 Requirement condition setting means 102 Similarity calculation Means 104 Prediction result calculation means 106 Manufacturing condition setting means 110 Operation record database

Claims (7)

Iron-containing raw materials, a pseudo-particles were granulated by adding water to the sintering raw material containing a CaO-containing material and flocculation materials, to form a charged to the sintering bed on the pallet truck sintering machine the pseudo particles, A control device that sets the manufacturing conditions of a sinter manufacturing facility that sinters the charging layer to produce sinter while collecting the exhaust gas discharged from the sinter with an electrostatic collector.
At least the coke ratio of the sintered raw material, the amount of water contained in the pseudo-particles, the layer thickness of the charging layer, the setting value of the guidance item including the traveling speed of the pallet carriage, and the required conditions including the operation specifications are set. Requirements setting means and
The sinter strength as output variables, the sinter production rate, the temperature before the electrostatic precipitator of the exhaust gas, the negative pressure before the electrostatic precipitator of the exhaust gas, and the set values and operation specifications of the guidance items as input variables are the past operations. A similarity calculation means for calculating the similarity between the required condition and the input variable of the database for each past operating condition using the database stored for each condition.
Using said similarity and said input variables to create sinter strength, sintered ore production rate, the local regression equation, respectively as the output variable of the exhaust gas electrostatic precipitator before temperature and the exhaust gas of an electrostatic precipitator prior to a negative pressure , sintered ore strength using the above requirements and the local regression equation, sintered ore production rate, and the prediction result calculation means for calculating a prediction result of the exhaust gas electrostatic precipitator before temperature and the exhaust gas of an electrostatic precipitator prior to negative pressure,
When all of the calculated sinter strength, sinter production rate, exhaust gas pre-dust collector temperature, and exhaust gas pre- dust collector negative pressure prediction results satisfy predetermined control values, the above requirements are met. Including a manufacturing condition setting means for setting as a sinter manufacturing condition,
The prediction result calculation means creates the local regression equation using the input variables having a tolerance greater than 0.1 among the input variables.
The manufacturing condition setting means is used when at least one of the calculated sinter strength, sinter production rate, exhaust gas pre-electrostatic collector temperature, and exhaust gas pre-electrostatic collector negative pressure does not satisfy the control value. Changes the set value of the guidance item set by the requirement condition setting means, causes the similarity calculation means to calculate the similarity again, and uses the prediction result calculation means to calculate the sinter strength, the sinter production rate, and the like. A control device for sinter manufacturing equipment that calculates the prediction results of the temperature before the electrostatic collector of exhaust gas and the negative pressure before the electric dust collector of exhaust gas. One unit for changing each of the set values of the guidance items is predetermined.
The requirement condition setting means has 81 lines of guidance items and operation specifications in which a value obtained by adding 1 unit to each setting value of the guidance item and a value obtained by subtracting 1 unit are combined in place of the requirement condition. Create a requirement matrix consisting of
The similarity calculation means calculates the similarity for each requirement matrix defined by the requirement condition setting means.
The prediction result calculating unit, said similarity, comprising the use as input variable, sinter strength, sintered ore production rate, the exhaust gas electrostatic precipitator before temperature and the exhaust gas of an electrostatic precipitator prior to negative pressure prediction result 81 Calculate the row prediction result matrix and
When the prediction result matrix satisfies all the control values, the production condition setting means sets the requirement conditions used for calculating the prediction result as the sinter production condition.
When the prediction result matrix does not satisfy all of the control values, the requirement condition used for calculating the prediction result closest to the control value in the items that do not satisfy the control values is the set value of the guidance item. Then, a requirement condition matrix in which 1 unit is added and 1 unit is subtracted from the set value of this guidance item is created, and the similarity is calculated again by the similarity calculating means, and the prediction result is obtained. The control device for a sinter manufacturing facility according to claim 1, wherein the calculation means calculates the prediction results of the sinter strength, the sinter production rate, the temperature before the electrostatic collector of the exhaust gas, and the negative pressure before the electrostatic collector of the exhaust gas. The prediction result calculation means creates the local regression equation using the input variables stored in the database and having a variance ratio of 2 or more calculated in (1) below. The control device for the sintered ore manufacturing facility according to claim 2.

In the above (1), S e _{added before,} a residual sum of squares before adding the input variables to regression models, S _after e _added, residual square after adding the input variable in the regression equation the sum, _{before adding} [Phi _e, a previous [Phi _e to add an input variable, _{after adding} [Phi _e is [Phi _e after adding the input variable, [Phi _e is calculated by the following equation (2) Is the value to be.
Φ _e = n-2 ... (2)
In the above equation (2), n is the number of input variables. A raw material blending device that blends an iron-containing raw material, a CaO-containing raw material, and a coagulant to form a sintered raw material,
A granulator that adds water to the sintering raw material to granulate pseudo-particles,
A sintering machine in which the pseudo-particles are charged into a pallet carriage of a sintering machine to form a charging layer, and the charging layer is sintered.
A sinter manufacturing facility having a crusher, a cooler, a sieving device, and a control device for the sinter manufacturing facility according to any one of claims 1 to 3. Iron-containing raw materials, a pseudo-particles were granulated by adding water to the sintering raw material containing a CaO-containing material and flocculation materials, to form a charged to the sintering bed on the pallet truck sintering machine the pseudo particles, A method for producing sinter, which produces sinter by sintering the charge layer while collecting the exhaust gas discharged from the sinter with an electrostatic collector.
At least the coke ratio of the sintered raw material, the amount of water contained in the pseudo-particles, the layer thickness of the charging layer, the setting value of the guidance item including the traveling speed of the pallet carriage, and the required conditions including the operation specifications are set. Requirements setting steps and
The sinter strength as output variables, the sinter production rate, the temperature before the electrostatic precipitator of the exhaust gas, the negative pressure before the electrostatic precipitator of the exhaust gas, and the set values and operation specifications of the guidance items as input variables are the past operations. Using the database stored for each condition, the similarity calculation step for calculating the similarity between the required condition and the input variable of the database for each past operating condition, and
Using said similarity and said input variables to create sinter strength, sintered ore production rate, the local regression equation, respectively as the output variable of the exhaust gas electrostatic precipitator before temperature and the exhaust gas of an electrostatic precipitator prior to a negative pressure , sintered ore strength using the above requirements and the local regression equation, sintered ore production rate, and the prediction result calculating step of calculating a prediction result of the exhaust gas electrostatic precipitator before temperature and the exhaust gas of an electrostatic precipitator prior to negative pressure,
When all of the predicted results of sinter strength, sinter production rate, exhaust gas pre-dust collector temperature and exhaust gas pre- dust collector negative pressure calculated in the prediction result calculation step satisfy predetermined control values. In addition, a production condition setting step for setting the above-mentioned requirements as a sinter production condition, and
It has a sinter manufacturing step of manufacturing sinter under the sinter manufacturing conditions set in the manufacturing condition setting step.
In the prediction result calculation step, the local regression equation is created using the input variables having a tolerance greater than 0.1 among the input variables.
In the manufacturing condition setting step, when at least one of the calculated sinter strength, sinter production rate, exhaust gas pre-dust collector temperature, and exhaust gas pre- dust collector negative pressure does not satisfy the control value. Changes the set value of the guidance item used in the requirement condition setting step, and repeats the similarity calculation step, the prediction result calculation step, and the production condition setting step again to manufacture the sinter. Method. One unit for changing each of the set values of the guidance items is set in advance.
In the requirement condition setting step, 81 lines of guidance items and operation specifications are combined with a value obtained by adding 1 unit to each setting value of the guidance item and a value obtained by subtracting 1 unit in place of the requirement condition. Create a requirement matrix consisting of
In the similarity calculation step, the similarity is calculated for each requirement matrix defined in the requirement condition setting step.
Wherein in the prediction result calculating step, said degree of similarity, comprising the use as input variable, sinter strength, sintered ore production rate, the exhaust gas electrostatic precipitator before temperature and the exhaust gas of an electrostatic precipitator prior to negative pressure prediction result 81 Calculate the row prediction result matrix and
In the production condition setting step, when the prediction result matrix satisfies all the control values, the requirement conditions used for calculating the prediction result are set as the sinter production conditions.
When the prediction result matrix does not satisfy all of the control values, the requirement condition used for calculating the prediction result closest to the control value in the items that do not satisfy the control values is set as the setting value of the guidance item. Then, a requirement condition matrix is created by combining the values obtained by adding 1 unit and subtracting 1 unit from the set value of this guidance item, and again, the similarity calculation step, the prediction result calculation step, and the manufacturing condition. The method for producing a sintered ore according to claim 5, wherein the setting step is repeatedly carried out. In the prediction result calculation step, the local regression equation is created using the input variables stored in the database and having a variance ratio of 2 or more calculated in (1) below. The method for producing a sintered ore according to claim 6.

In the above (1), _{before adding} S e, a residual sum of squares before adding the input variable to the regression equation, _{after adding} S e is the residual sum of squares after adding the input variable in the regression equation and a, _{before adding} [Phi _e, a previous [Phi _e to add an input variable, _{after adding} [Phi _e is [Phi _e after adding the input variable, [Phi _e is calculated by the following equation (2) Value.
Φ _e = n-2 ... (2)
In the above equation (2), n is the number of input variables.

Images