JP2017128805A - Operation method of blast furnace - Google Patents
Operation method of blast furnace Download PDFInfo
- Publication number
- JP2017128805A JP2017128805A JP2017006316A JP2017006316A JP2017128805A JP 2017128805 A JP2017128805 A JP 2017128805A JP 2017006316 A JP2017006316 A JP 2017006316A JP 2017006316 A JP2017006316 A JP 2017006316A JP 2017128805 A JP2017128805 A JP 2017128805A
- Authority
- JP
- Japan
- Prior art keywords
- blast furnace
- statistic
- index
- calculated
- furnace
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
Description
本発明は、高炉の操業方法に関し、特に、炉内での通気不良に伴う棚吊りや吹き抜けなどの異常現象を未然に防止しながら操業する高炉の操業方法に関する。 The present invention relates to a method for operating a blast furnace, and more particularly, to a method for operating a blast furnace that operates while preventing abnormal phenomena such as shelf hanging and blow-out due to poor ventilation in the furnace.
銑鉄を生産する高炉（「溶鉱炉」ともいう）では、通常、炉頂から原料である鉄鉱石（「鉱石」とも記す）と鉄鉱石の還元材であるコークスとをそれぞれ交互に装入し、鉱石層とコークス層とを層状に形成させている。そして、炉内における鉱石層とコークス層との堆積後の分布を調整することにより、炉内でのガスの流れを制御している。高炉の安定操業を維持するためには、高炉内に良好な通気性を維持し、炉下部の羽口と呼ばれる孔から高炉内部に供給される高温空気の流れを安定化させることが重要である。高炉内の通気性は、装入される鉱石及びコークスの性状や粒度の影響を受けるが、これ以外に炉頂からの装入物の装入方法及び装入物の炉内における分布状況によっても大きく影響を受ける。 In a blast furnace that produces pig iron (also referred to as a “blast furnace”), iron ore (also referred to as “ore”) that is a raw material and coke that is a reduction material of iron ore are normally charged alternately from the top of the furnace. The layer and the coke layer are formed in layers. And the flow of the gas in a furnace is controlled by adjusting the distribution after the deposition of the ore layer and the coke layer in a furnace. In order to maintain stable operation of the blast furnace, it is important to maintain good air permeability in the blast furnace and stabilize the flow of high-temperature air supplied into the blast furnace from a hole called tuyere at the bottom of the furnace. . The air permeability in the blast furnace is affected by the properties and particle size of the ore and coke to be charged, but besides this, it depends on the charging method of the charge from the top of the furnace and the distribution of the charge in the furnace. It is greatly affected.
この高炉の通気性が何らかの原因によって悪化し、炉内におけるガスの円滑な流れが阻害された場合には、棚吊り、スリップ、吹き抜けと呼ばれる異常現象、つまり炉況異常が起こることは良く知られている。ここで、「棚吊り」とは、正常な状態であれば炉上部から順次降下する（以下、「荷下がり」という）はずの鉱石及びコークスが、停止する状態を指す。「スリップ」とは、荷下がりが停止している鉱石及びコークスが、突如、荷下がりする現象である。また、「吹き抜け」とは、炉下部から供給された高温のガスが、何らかの理由により急激に炉上部へと噴出する現象である。 It is well known that when the air permeability of this blast furnace deteriorates for some reason and the smooth flow of gas in the furnace is hindered, an abnormal phenomenon called shelf suspending, slipping, or blow-through, that is, abnormal furnace conditions occurs. ing. Here, “shelf hanging” refers to a state in which ores and coke that should descend sequentially from the upper part of the furnace (hereinafter referred to as “unloading”) in a normal state are stopped. “Slip” is a phenomenon in which ore and coke that have stopped unloading suddenly unload. “Blow-through” is a phenomenon in which high-temperature gas supplied from the lower part of the furnace is suddenly ejected to the upper part of the furnace for some reason.
これら高炉の異常現象は、（１）装入物の正常な荷下がりを乱し、炉内に層状に充填された装入物の構造を破壊し炉内のガス流れを乱す、（２）炉下部への装入物の急激な荷下がりにより鉱石の直接還元の増加による炉温の低下をもたらす、（３）炉上部に吹き抜けた高温のガスが炉頂部の付帯設備を損傷させるなどの問題を引き起こし、高炉の円滑な操業を阻害して莫大な損害をもたらす。そのため、高炉の操業において、これらの異常現象の予兆検知は重要な課題である。 The abnormal phenomena of these blast furnaces are: (1) disturb the normal unloading of the charge, destroy the charge structure layered in the furnace and disturb the gas flow in the furnace, (2) the furnace Due to the sudden unloading of the charge in the lower part, the furnace temperature decreases due to an increase in direct reduction of ore. (3) The high temperature gas blown through the upper part of the furnace damages incidental equipment at the top of the furnace. Cause, and hinders the smooth operation of the blast furnace, causing enormous damage. Therefore, in the operation of the blast furnace, the detection of these abnormal phenomena is an important issue.
従来、高炉の異常現象を回避或いは予知するために種々の技術が報告されている。例えば、特許文献１には、高炉シャフト部の炉壁に複数個の貫通孔を設け、吹き抜けの発生要因となる炉内のガスを抜き、吹き抜けを防止する方法が提案されている。また、特許文献２には、高炉炉壁の高さ方向に高炉内の圧力を測定するための圧力測定装置を複数段設けて測定装置群とし、これら測定装置群ごとに各段の差圧を測定し、測定した差圧の時間経過に伴う変化を３次元表示し、３次元表示した差圧の変化を監視することにより、操業異常を予知する方法を提案している。 Conventionally, various techniques have been reported for avoiding or predicting abnormal phenomena in a blast furnace. For example, Patent Document 1 proposes a method in which a plurality of through holes are provided in a furnace wall of a blast furnace shaft portion, and gas in the furnace that causes blow-out is extracted to prevent blow-through. In Patent Document 2, a plurality of pressure measuring devices for measuring the pressure in the blast furnace are provided in the height direction of the blast furnace wall to form a measuring device group, and the differential pressure at each stage is set for each measuring device group. It proposes a method for predicting abnormal operation by measuring, measuring the three-dimensional change of the measured differential pressure over time, and monitoring the change of the three-dimensionally displayed differential pressure.
しかしながら、これらの技術は、棚吊り、スリップ、吹き抜けなどの高炉の異常現象を回避或いは予知するための技術としては、十分とはいえない。例えば、特許文献１に記載の技術の場合には、高炉に新たにガス抜き用の孔を設置する、または、ガス抜き用の管を設置するなど、新たな設備の設置を必要とする。また、特許文献２に記載の技術の場合には、差圧を３次元表示することで監視すべきデータ量が増えるために、異常予兆の検知はかえって複雑で困難となる。 However, these techniques are not sufficient as techniques for avoiding or predicting abnormal phenomena in the blast furnace such as shelf hanging, slipping, and blow-through. For example, in the case of the technique described in Patent Document 1, it is necessary to install new equipment such as newly installing a degassing hole in the blast furnace or installing a degassing pipe. In the case of the technique described in Patent Document 2, since the amount of data to be monitored increases by displaying the differential pressure in three dimensions, detection of an abnormality sign is rather complicated and difficult.
本発明は上記事情に鑑みてなされたもので、その目的とするところは、高炉にガス抜き用の孔を設置するなどの新たな設備を設けることなく、高炉に設置されている複数個の圧力センサーで測定されるシャフト圧に対して統計処理を施すことで監視すべきデータ量を少なくし、この少ないデータ量で、通気不良に伴う棚吊り、スリップ、吹き抜けなどの高炉の異常現象を正確に予知し、異常現象を未然に防止することのできる高炉の操業方法を提供することである。 The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a plurality of pressures installed in the blast furnace without providing new equipment such as installing a vent hole in the blast furnace. By applying statistical processing to the shaft pressure measured by the sensor, the amount of data to be monitored is reduced, and with this small amount of data, abnormal phenomena in the blast furnace such as shelves, slips, and blowouts caused by poor ventilation can be accurately performed. It is to provide a method of operating a blast furnace that can predict and prevent an abnormal phenomenon.
本発明者らは、上記課題を解決すべく、鋭意検討した。その結果、本発明者らは、高炉に数多く設置された圧力センサーで測定される複数個のシャフト圧の測定データに対して主成分分析という統計的な処理を施して、これら複数個の測定データを一元的な指標に表現することで、高炉内の通気性の判定及び高炉内での異常現象の予知、つまり高炉の炉況異常の予知を精度良く行うことが可能になるとの知見を得た。 The present inventors diligently studied to solve the above problems. As a result, the present inventors performed statistical processing called principal component analysis on a plurality of shaft pressure measurement data measured by a number of pressure sensors installed in the blast furnace, and the plurality of measurement data We have obtained the knowledge that it is possible to accurately determine the air permeability in the blast furnace and the prediction of abnormal phenomena in the blast furnace, that is, the prediction of abnormal conditions in the blast furnace by expressing the above as a unified index. .
本発明は上記知見に基づきなされたものであり、その要旨は以下のとおりである。
［１］高炉に設置された複数の圧力センサーで測定される複数点のシャフト圧データから高炉の操業異常を予知しながら操業する高炉の操業方法であって、
操業中に前記圧力センサーで測定された複数点のシャフト圧データに対して主成分分析を行い、主成分分析によってＴ２統計量及び／またはＱ統計量を算出し、算出したＴ２統計量を正常な操業範囲のシャフト圧測定データに基づいて算出したＴ２統計量の最大値で除した指数（Ｔ２統計量指数）、及び／または、算出したＱ統計量を正常な操業範囲のシャフト圧測定データに基づいて算出したＱ統計量の最大値で除した指数（Ｑ統計量指数）を用いて、高炉操業の異常を予知することを特徴とする、高炉の操業方法。
［２］高炉に設置された複数の圧力センサーで測定される複数点のシャフト圧データから高炉の炉況異常を予知しながら操業する高炉の操業方法であって、
操業中に前記圧力センサーで測定された複数点のシャフト圧データに対して主成分分析を行い、主成分分析によってＴ２統計量及び／またはＱ統計量を算出し、算出したＴ２統計量を正常な操業範囲のシャフト圧測定データに基づいて算出したＴ２統計量の最大値で除した指数（Ｔ２統計量指数）、及び／または、算出したＱ統計量を正常な操業範囲のシャフト圧測定データに基づいて算出したＱ統計量の最大値で除した指数（Ｑ統計量指数）を用いて高炉の炉況異常を予知し、高炉操業条件を調整することを特徴とする、高炉の操業方法。
［３］任意の８時間以内における前記Ｑ統計量指数の１．０超えの回数が３回を超えると前記炉況異常が発生する可能性があると判定し、任意の８時間以内における前記Ｑ統計量指数の１．０超えの３回目のピークの発生後に、高炉操業条件を調整することを特徴とする、上記［２］に記載の高炉の操業方法。
The present invention has been made based on the above findings, and the gist thereof is as follows.
[1] A method of operating a blast furnace that operates while predicting an abnormal operation of the blast furnace from a plurality of shaft pressure data measured by a plurality of pressure sensors installed in the blast furnace,
A principal component analysis is performed on the shaft pressure data at a plurality of points measured by the pressure sensor during operation, T 2 statistics and / or Q statistics are calculated by principal component analysis, and the calculated T 2 statistics are calculated. An index (T 2 statistic index) divided by the maximum value of T 2 statistic calculated based on shaft pressure measurement data in the normal operating range and / or the calculated Q statistic of shaft pressure in the normal operating range A method of operating a blast furnace, characterized by predicting abnormalities in blast furnace operation using an index (Q statistic index) divided by the maximum value of Q statistics calculated based on measurement data.
[2] A method of operating a blast furnace that operates while predicting abnormal blast furnace conditions from a plurality of shaft pressure data measured by a plurality of pressure sensors installed in the blast furnace,
A principal component analysis is performed on the shaft pressure data at a plurality of points measured by the pressure sensor during operation, T 2 statistics and / or Q statistics are calculated by principal component analysis, and the calculated T 2 statistics are calculated. An index (T 2 statistic index) divided by the maximum value of T 2 statistic calculated based on shaft pressure measurement data in the normal operating range and / or the calculated Q statistic of shaft pressure in the normal operating range Blast furnace operation characterized by predicting abnormal blast furnace conditions and adjusting blast furnace operating conditions using an index (Q statistic index) divided by the maximum Q statistic calculated based on measurement data Method.
[3] When the number of times the Q statistic index exceeds 1.0 in any 8 hours exceeds 3, it is determined that the furnace condition abnormality may occur, and the Q in any 8 hours is determined. The method for operating a blast furnace according to the above [2], wherein the blast furnace operating conditions are adjusted after the occurrence of the third peak exceeding 1.0 of the statistical index.
本発明によれば、少ないデータ量で、高炉の通気不良に起因する炉内の異常現象を精度良く予知することができ、その結果、通気不良に起因する棚吊りや吹き抜けなどの異常現象の発生前に、適切な操業アクションを適用することが可能となり、高炉の安定操業が実現される。 According to the present invention, it is possible to accurately predict an abnormal phenomenon in the furnace caused by poor ventilation of the blast furnace with a small amount of data, and as a result, occurrence of abnormal phenomena such as shelf hanging and blow-by caused by defective ventilation. Before, it becomes possible to apply an appropriate operation action, and stable operation of the blast furnace is realized.
以下、本発明を詳細に説明する。はじめに、高炉の複数箇所に設置されている圧力センサーによって測定されるシャフト圧の測定データに対して行う主成分分析について説明する。 Hereinafter, the present invention will be described in detail. First, a principal component analysis performed on shaft pressure measurement data measured by pressure sensors installed at a plurality of locations in a blast furnace will be described.
主成分分析とは、同期する複数個（複数次元）のデータ群について、元のデータ群の持つ情報量の損失をできる限り小さくしつつ、元のデータの持つ特徴が良く反映された少数の変数へと置換（低次元化）する数学的処理を指す。これは、例えば高炉のシャフト圧のデータの場合であれば、高炉１基に対してシャフト圧を測定する圧力センサーは約３０箇所（３０点）設置されているが、これに主成分分析を適用することで、３０点のデータ群の特徴を良好に反映する数個の変数に仮に置き換えられたとすれば、これら３０点のデータ群全てを観察することなく、主成分分析により生成された少数の変数を監視することで、炉内の状態をより簡便に推定可能であることを表している。尚、同期とは、図１に示すように、プロセスにおける時間推移または操業アクションに対して、操業上の変数の挙動に協調性があることを指す。 Principal component analysis is a small number of variables that reflect the characteristics of the original data well while minimizing the loss of information in the original data group as much as possible for multiple synchronized data groups. Refers to mathematical processing that replaces (decreases). For example, in the case of blast furnace shaft pressure data, about 30 (30 points) pressure sensors for measuring shaft pressure for one blast furnace are installed, and principal component analysis is applied to this. Thus, if it is replaced by several variables that well reflect the characteristics of the 30-point data group, a small number of data generated by principal component analysis can be obtained without observing all of the 30-point data group. It shows that the state in the furnace can be estimated more easily by monitoring the variables. As shown in FIG. 1, the synchronization means that the behavior of the operation variable is cooperative with respect to the time transition or the operation action in the process.
以下、刊行物１（刊行物１；http://manabukano.brilliant-future.net/document/text-PCA.pdf）に記載された内容に沿って、主成分分析の具体的な手法を説明する。 In the following, a specific method for principal component analysis will be described according to the contents described in Publication 1 (Publication 1; http://manabukano.brilliant-future.net/document/text-PCA.pdf). .
主成分分析では、下記の（１）式に示されるＰ個の変数｛ｘｐ｝（ｐ＝１，２，…，Ｐ）の持つ情報を、情報の損失を最小限に抑えながら、変数｛ｘｐ｝の一次結合として与えられる互いに独立なＭ個（Ｍ≦Ｐ）の主成分｛ｚｍ｝（ｍ＝１，２，…，Ｍ）を用いて表現する。 In the principal component analysis, the information of the P variables {x p } (p = 1, 2,..., P) represented by the following equation (1) is reduced to the variable { x p } is expressed using M (M ≦ P) principal components {z m } (m = 1, 2,..., M) which are given as linear combinations.
（１）式において、ｗｐｍは結合係数を表し、ｘｐは前述のようにシャフト圧などのセンサーによるデータ群に相当する。尚、（１）式に示すｗｐｍは、下記の（２）式に示す条件を満足する必要がある。 In the equation (1), w pm represents a coupling coefficient, and x p corresponds to a data group by a sensor such as a shaft pressure as described above. Incidentally, w pm shown in the equation (1) needs to satisfy the condition shown in the following equation (2).
また、（１）式において、Ｍ＝１の場合ならば、複数個（Ｐ個）のシャフト圧データは一つの主成分データ｛ｚ１｝に変換されたことになる。また、Ｍ＝１の場合、主成分データ｛ｚ１｝を第１主成分という。第１主成分｛ｚ１｝は（１）式で与えられるため、（１）式の結合係数ｗｐｍを下記の（３）式のようなベクトル表記とする。 In the equation (1), if M = 1, a plurality (P) of shaft pressure data is converted into one principal component data {z 1 }. When M = 1, the principal component data {z 1 } is referred to as a first principal component. Since the first principal component {z 1 } is given by the equation (1), the coupling coefficient w pm of the equation (1) is represented by a vector notation as the following equation (3).
また、操業中の或る時刻におけるセンサーデータｘを、下記の（４）式で示すベクトル表記で表す。 Further, sensor data x at a certain time during operation is represented by a vector notation represented by the following equation (4).
この時、センサーデータｘに対応する第１主成分｛ｚ１｝は、下記の（５）式で表される。 At this time, the first principal component {z 1 } corresponding to the sensor data x is expressed by the following equation (5).
第１主成分｛ｚ１｝の分散σ２ z1は、下記の（６）式で表される。 The variance σ 2 z1 of the first principal component {z 1 } is expressed by the following equation (6).
（６）式において、Ｎはデータのサンプル数を表す。（６）式のＴは転置行列を表す。また、（６）式において、Ｖは共分散行列であり、共分散行列Ｖは下記の（７）式で表される。 In the equation (6), N represents the number of data samples. T in Equation (6) represents a transposed matrix. In the equation (6), V is a covariance matrix, and the covariance matrix V is expressed by the following equation (7).
第１主成分｛ｚ１｝は、（２）式の条件を満たす条件下で、（６）式に示す分散σ２ z1が最大となるように決定される必要がある。これはLagrange未定乗数法を用いて解くことが可能であり、乗数λを用いて下記の（８）式に示す変数Ｊ１を最大にする結合係数ｗ１を求めればよい。 The first principal component {z 1 } needs to be determined so that the variance σ 2 z1 shown in the equation (6) is maximized under the condition satisfying the equation (2). This can be solved using the Lagrange undetermined multiplier method, and the coupling coefficient w 1 that maximizes the variable J 1 shown in the following equation (8) may be obtained using the multiplier λ.
（８）式の最大値を与える結合係数ｗ１を求めるには、変数Ｊ１の結合係数ｗ１による偏微分値が０となる結合係数ｗ１を求めればよく、結局、（８）式の偏微分から下記の（９）式に示す条件式が得られる。（９）式のＩは、対角項が１、それ以外の成分が０である単位行列を表す。 To determine the coupling coefficients w 1 that gives the maximum value of the equation (8) may be determined the coupling coefficient w 1 of partial derivatives by coupling coefficient w 1 of the variable J 1 becomes 0, after all, (8) of The conditional expression shown in the following expression (9) is obtained from the partial differentiation. In equation (9), I represents a unit matrix in which the diagonal term is 1 and the other components are 0.
（９）式の条件式は固有値問題であり、乗数λが満たす条件は下記の（１０）式の固有方程式を用いて表される。 The conditional expression (9) is an eigenvalue problem, and the condition satisfied by the multiplier λ is expressed using the following eigen equation (10).
したがって、乗数λ及び第１主成分｛ｚ１｝は共分散行列Ｖの最大固有値及び固有ベクトルとして求めることができる。 Therefore, the multiplier λ and the first principal component {z 1 } can be obtained as the maximum eigenvalue and eigenvector of the covariance matrix V.
次元圧縮後のデータを元のＰ次元空間上の座標で表現すると、下記の（１１）式のようになる。 When the dimension-compressed data is expressed by coordinates in the original P-dimensional space, the following equation (11) is obtained.
Ｐは主成分の結合係数からなる行列であり、下記の（１２）式のようになる。 P is a matrix composed of the coupling coefficients of the principal components, and is represented by the following equation (12).
操業中のシャフト圧の変動範囲を定義するには、原点からの距離に対応する指標を用いればよい。そこで、下記の（１３）式に示すＴ２統計量を用いる。 In order to define the fluctuation range of the shaft pressure during operation, an index corresponding to the distance from the origin may be used. Therefore, the T 2 statistic shown in the following equation (13) is used.
（１３）式に示すＴ２統計量を用いれば、操業上におけるシャフト圧の変動範囲を見積もることが可能となり、シャフト圧の大小から操業の正常・異常を判定することができる。一方、操業中のシャフト圧の同期関係が乱れるような、操業の本質的な変動とは異なる、センサー異常や高炉内の何らかの異常を検知するためには、（１３）式に示すＴ２統計量に直交する指標である、下記の（１４）式に示すＱ統計量を用いればよい。 If the T 2 statistic shown in the equation (13) is used, it is possible to estimate the fluctuation range of the shaft pressure in the operation, and the normality / abnormality of the operation can be determined from the magnitude of the shaft pressure. On the other hand, in order to detect a sensor abnormality or any abnormality in the blast furnace, which is different from the essential fluctuation of the operation, such that the synchronous relationship of the shaft pressure during operation is disturbed, the T 2 statistic shown in the equation (13) Q statistic shown in the following equation (14), which is an index orthogonal to
図２に、Ｔ２統計量及びＱ統計量の数学的なイメージを示す。今、簡便のためシャフト圧のデータは２点と仮定する。これらシャフト圧のデータが同期するものであるならば、操業中に測定されたシャフト圧のデータを２次元的にプロットすると図２のようになる。この時、データの有する特徴を良く表現する指標として、先に説明した新たな情報量として定義されたＴ２統計量は、図２のｚ１軸の値にほぼ相当し、操業中のデータはｚ１軸方向を移動する。一方、測定されたシャフト圧から何らかの理由で同期性が失われ、シャフト圧のプロットがｚ１軸上から外れることがある。この時、ｚ１軸に対して直交する新たな情報量としてＱ統計量が定義され、プロットがｚ１軸上から逸脱した量、即ちプロットされた点からｚ１軸上に下ろされた垂線の長さが、シャフト圧の逸脱量（異常量）として検出されることになる。 FIG. 2 shows a mathematical image of T 2 statistics and Q statistics. For simplicity, it is assumed that the shaft pressure data is 2 points. If the shaft pressure data are synchronized, the shaft pressure data measured during operation is plotted two-dimensionally as shown in FIG. At this time, the T 2 statistic defined as the new information amount described above as an index that well expresses the characteristics of the data substantially corresponds to the value of the z 1 axis in FIG. 2, and the data in operation is z Move in one axis direction. On the other hand, for some reason in synchrony is lost from the measured shaft pressure, there is the plot of the shaft pressure deviates from the z 1 axis. In this case, Q statistic is defined as a new amount of information that is orthogonal to the z 1 axis plots the amount deviating from the z 1 axis, that is, from the plotted points z 1 axis to downed perpendicular of The length is detected as a deviation amount (abnormal amount) of the shaft pressure.
そのため、上記の手法による異常予知を行なうためには、図２の楕円領域に相当する正常な操業範囲におけるデータベースを事前に構築する必要がある。これは、まず炉内に異常が発生していない正常な操業区間について、センサーデータの時系列データに対して上記手法を適用し、Ｔ２統計量及びＱ統計量の時系列データを作成する。また、この正常な時間区間におけるＴ２統計量及びＱ統計量の最大値を求める。正常な時間区間についてこれらＴ２統計量及びＱ統計量の最大値を求めることは、正常な操業を行なっている場合のセンサーデータの変動幅及び正常な操業範囲からの逸脱量の最大値を求めていることになる。 Therefore, in order to perform abnormality prediction by the above method, it is necessary to construct in advance a database in a normal operation range corresponding to the elliptical region in FIG. First, the above method is applied to the time series data of the sensor data for a normal operation section in which no abnormality has occurred in the furnace, and time series data of T 2 statistics and Q statistics are created. In addition, the maximum values of the T 2 statistic and the Q statistic in the normal time interval are obtained. Determining the maximum value of T 2 statistic and Q statistic for the normal time interval, the maximum value of the deviation amount from the fluctuation range and normal operating range of the sensor data when doing the normal operation Will be.
高炉内の操業異常を予知する場合には、異常が起こりつつある操業範囲のセンサーデータから算出したＴ２統計量を、正常な操業範囲のセンサーデータから算出したＴ２統計量の最大値で除した指数（以下、「Ｔ２統計量指数」と記す）を用いて異常を予知する、及び／または、異常が起こりつつある操業範囲のセンサーデータから算出したＱ統計量を、正常な操業範囲のセンサーデータから算出したＱ統計量の最大値で除した指数（以下、「Ｑ統計量指数」と記す）を用いて異常を予知する。 When predicting an abnormal operation in the blast furnace, the T 2 statistic calculated from the sensor data in the operating range where the abnormality is occurring is divided by the maximum value of the T 2 statistic calculated from the sensor data in the normal operating range. exponent (hereinafter referred to as "T 2 statistic index") to predict the abnormality with, and / or abnormalities of the Q statistic calculated from sensor data operating range is occurring, normal operation range An anomaly is predicted using an index (hereinafter referred to as “Q statistic index”) divided by the maximum value of the Q statistic calculated from the sensor data.
例として、表１に示す各時刻のシャフト圧のＴ２統計量及びＱ統計量を求める。 As an example, the T 2 statistic and the Q statistic of the shaft pressure at each time shown in Table 1 are obtained.
表１のデータからデータ行列を構築すると、下記の（１５）式が得られる。 When a data matrix is constructed from the data in Table 1, the following equation (15) is obtained.
（１５）式の行列Ｘから共分散行列を求めて主成分を算出し、（１３）式からＴ２統計量を算出し、また、（１４）式からＱ統計量を算出すると、以下の（１６）式及び（１７）式が得られる。 When a covariance matrix is obtained from the matrix X of the equation (15), the principal component is calculated, the T 2 statistic is calculated from the equation (13), and the Q statistic is calculated from the equation (14), the following ( Equations (16) and (17) are obtained.
行列Ｘには３時刻分のデータが含まれているため、（１６）式及び（１７）式にも３時刻分のＴ２統計量及びＱ統計量が含まれる。このように、各時刻におけるＴ２統計量及びＱ統計量を高炉の正常操業時の一定の期間で算出し、データベースとして蓄えておく。そして、得られた正常時におけるＴ２統計量の最大値で、操業中の各時刻におけるＴ２統計量のデータを除してＴ２統計量指数を求め、また、得られた正常時におけるＱ統計量の最大値で、操業中の各時刻におけるＱ統計量のデータを除してＱ統計量指数を求め、求めたＴ２統計量指数及び／またはＱ統計量指数により、操業中の正常・異常を判定する。 Since the matrix X includes data for three times, the equations (16) and (17) also include T 2 statistics and Q statistics for three times. In this way, the T 2 statistic and the Q statistic at each time are calculated for a certain period during normal operation of the blast furnace and stored as a database. Then, the maximum value of T 2 statistic in a normal obtained, determined the T 2 statistic index by dividing the data of the T 2 statistic at each time during the operation, also, Q in the resulting normal state the maximum value of the statistic, determine the Q statistic index by dividing the data of the Q statistic at each time during the operation, the T 2 statistic index and / or Q statistic index determined, normal-in operation Judge abnormalities.
即ち、本発明において、高炉内の操業異常を予知するにあたり、通常の条件下で行われている操業の測定データを用いてＴ２統計量及び／またはＱ統計量を算出し、算出したＴ２統計量を正常な操業範囲のシャフト圧測定データから算出したＴ２統計量の最大値で除したＴ２統計量指数を用いて異常現象を予知するか、及び／または、算出したＱ統計量を正常な操業範囲のシャフト圧測定データから算出したＱ統計量の最大値で除したＱ統計量指数を用いて異常現象を予知する。例えば、Ｔ２統計量指数またはＱ統計量指数が１．０よりも大きくなった場合に異常現象が発生すると予測するなどすることができる。 That is, in the present invention, when predicting an abnormal operation in the blast furnace, the T 2 statistic and / or the Q statistic are calculated using measurement data of the operation performed under normal conditions, and the calculated T 2 Predict abnormal phenomena using the T 2 statistic index divided by the maximum value of the T 2 statistic calculated from the shaft pressure measurement data in the normal operating range, and / or calculate the calculated Q statistic Abnormal phenomena are predicted using the Q statistic index divided by the maximum value of the Q statistic calculated from the shaft pressure measurement data in the normal operating range. For example, it can be predicted that an abnormal phenomenon will occur when the T 2 statistic index or the Q statistic index is greater than 1.0.
そして、Ｔ２統計量指数またはＱ統計量指数が１．０よりも大きくなって異常現象が予知された場合には、予知された異常現象の発生を未然に防止するために、例えば、高炉内部への送風量を減少させる或いは送風を停止させるなど、高炉操業条件を調整し、通気不良に起因する棚吊り、スリップ、吹き抜けなどの異常現象つまり炉況異常の発生を未然に防止する。 When the T 2 statistic index or the Q statistic index is larger than 1.0 and an abnormal phenomenon is predicted, for example, in order to prevent the occurrence of the predicted abnormal phenomenon, By adjusting the blast furnace operating conditions such as reducing the amount of air blown to the air flow or stopping the air flow, abnormal phenomena such as shelf suspending, slipping and blowout due to poor ventilation, that is, abnormal furnace conditions are prevented in advance.
このように、本発明によれば、高炉の通気不良に起因する炉内の異常現象、つまり、炉況異常を、高炉の複数箇所に設置されている圧力センサーによって測定されるシャフト圧の測定データを用いて精度良く予知することが可能となり、その結果、通気不良に起因する棚吊りや吹き抜けなどの異常現象の発生前に適切な操業アクションを適用することが可能となり、高炉の安定操業が実現される。 Thus, according to the present invention, abnormal pressure in the furnace due to poor ventilation of the blast furnace, that is, abnormal furnace conditions, shaft pressure measurement data measured by pressure sensors installed at a plurality of locations in the blast furnace. As a result, it is possible to apply appropriate operation actions before the occurrence of abnormal phenomena such as shelf hanging and blowout due to poor ventilation, resulting in stable operation of the blast furnace Is done.
実炉を対象とする前に、本発明で採用する異常現象予知手法が有効であることを確認するために、高炉内の装入物の降下挙動及びガス流れを模した数値シミュレーションによって得られた結果に対して、主成分分析を適用した。本実施例では、高炉内のガス流れ及び装入物の荷下がりをシミュレートするモデルとして、ガス流れに関しては数値流体力学に基づくモデルを用い、また、装入物の荷下がりに関しては離散要素法（または個別要素法）と呼称される個々の装入物粒子の運動を時々刻々と追跡するモデルを用い、最終的にはこの２つのモデルを連成したモデルを使用した。 In order to confirm that the abnormal phenomenon prediction method adopted in the present invention is effective before targeting the actual furnace, it was obtained by numerical simulation simulating the descending behavior of the charge in the blast furnace and the gas flow Principal component analysis was applied to the results. In this embodiment, as a model for simulating gas flow in the blast furnace and unloading of the charge, a model based on numerical fluid dynamics is used for the gas flow, and a discrete element method is used for unloading of the charge. We used a model that traces the movement of individual charge particles, called (or individual element method), and finally a model that couples the two models.
図３に計算対象の領域（計算格子）及び装入物充填状況を示す。高炉１は、通常、円筒型の形状を有しており、計算すべき領域もこれと同一とするべきであるが、モデルによる計算負荷を低減させるために、図３に示すような中心角度２０°の扇型領域を計算領域とした。図３に示す扇型領域に、上部から鉱石及びコークスを模した粒子を交互に層状に充填し、鉱石層２及びコークス層３を形成し、一方、下部の羽口を模した領域からガスを炉内へ流通させた。 FIG. 3 shows the calculation target area (calculation grid) and the charge filling status. The blast furnace 1 usually has a cylindrical shape, and the region to be calculated should be the same as this, but in order to reduce the calculation load by the model, a central angle 20 as shown in FIG. The fan-shaped region at ° was taken as the calculation region. The fan-shaped region shown in FIG. 3 is alternately filled with particles simulating ore and coke from the top to form the ore layer 2 and the coke layer 3, while gas from the region simulating the lower tuyere It was distributed in the furnace.
このモデルを用いて高炉内の通気性を悪化させ、操業異常、特に吹き抜けを意図的に発生させた。この通気性悪化から吹き抜け発生までの区間におけるシャフト圧の計算値について主成分分析を適用し、吹き抜けの予知が可能か検討した。 Using this model, the air permeability in the blast furnace was deteriorated, and abnormal operation, especially blow-through, was intentionally generated. The principal component analysis was applied to the calculated value of the shaft pressure in the section from the deterioration of the air permeability to the occurrence of the blow-through to examine whether the blow-through can be predicted.
鉱石の還元材として高炉内に装入されるコークスは、一方で炉内におけるガスパスの役割を果たす。そこで、高炉内の通気性を意図的に悪化させるために、図４に示すように、炉内に装入するコークス量を局所的に減少させた。以下、局所的に装入量を減少させたコークスによって形成されるコークス層３を、「薄層コークス層」と呼ぶ。 The coke charged into the blast furnace as the ore reducing material, on the other hand, serves as a gas path in the furnace. Therefore, in order to intentionally deteriorate the air permeability in the blast furnace, the amount of coke charged into the furnace was locally reduced as shown in FIG. Hereinafter, the coke layer 3 formed by coke having a locally reduced amount of charge is referred to as a “thin coke layer”.
図５に、局所的にコークス量を減らして炉の上部に形成させた薄層コークス層が、炉下部の融着帯に到達するまでの炉内の装入物の充填構造を示す。薄層コークス層が形成されたことにより炉内の通気は悪化し、融着帯の層構造は破壊されて吹き抜けが生じた。 FIG. 5 shows the charging structure of the charge in the furnace until the thin coke layer formed locally on the upper part of the furnace with a reduced amount of coke reaches the fusion zone in the lower part of the furnace. The formation of a thin coke layer deteriorated the ventilation in the furnace, destroying the layer structure of the cohesive zone and causing blow-through.
この一連の計算結果に対して、まず、薄層コークス層が形成されていない正常な期間のシャフト圧について主成分分析を行い、正常な期間におけるＴ２統計量及びＱ統計量の最大値を算出した。次に、薄層コークス層を形成した以降の期間についても同様に主成分分析を行い、各時刻において算出されたシャフト圧のＴ２統計量及びＱ統計量を算出し、算出されたＴ２統計量を正常な期間で算出したＴ２統計量の最大値で除したＴ２統計量指数、及び、算出されたＱ統計量を正常な期間で算出したＱ統計量の最大値で除したＱ統計量指数を求めた。 For this series of calculation results, first, the principal component analysis is performed for the shaft pressure in the normal period where the thin coke layer is not formed, and the maximum values of the T 2 statistic and the Q statistic in the normal period are calculated. did. Next, the same as principal component analysis also for the period after the formation of the thin layer coke layer, calculates the T 2 statistic and Q statistic of shaft pressure calculated at each time, the calculated T 2 statistic dividing the T 2 statistic index the amount at the maximum value of T 2 statistic calculated in the normal period, and, divided by Q statistics Q statistic calculated in the maximum value of the calculated Q statistic in normal periods The quantity index was determined.
求めたＴ２統計量指数及びＱ統計量指数が、１．０以下であれば正常な操業範囲内であり、１．０を上回れば炉内に何らかの異常が生じていることになる。 If the calculated T 2 statistic index and Q statistic index are 1.0 or less, they are within the normal operating range, and if they exceed 1.0, some abnormality has occurred in the furnace.
図６に、シミュレーション結果に主成分分析を適用して得られたＴ２統計量指数の分析結果を示し、また、図７に、シミュレーション結果に主成分分析を適用して得られたＱ統計量指数の分析結果を示す。Ｑ統計量指数については、薄層コークス層を形成して炉内の通気を悪化させた直後から大きく振動して１．０を上回り、異常現象の予兆が検知された。また、吹き抜けの発生直前には、Ｑ統計量指数は１．０を大きく上回っており、通気悪化に伴う炉内の異常を十分に検知していた。 FIG. 6 shows the analysis result of the T 2 statistic index obtained by applying the principal component analysis to the simulation result, and FIG. 7 shows the Q statistic obtained by applying the principal component analysis to the simulation result. The analysis result of the index is shown. The Q statistic index greatly exceeded 1.0 immediately after the thin coke layer was formed and the ventilation in the furnace was deteriorated, and a sign of an abnormal phenomenon was detected. Further, immediately before the occurrence of the blow-through, the Q statistic index greatly exceeded 1.0, and the abnormality in the furnace due to the deterioration of ventilation was sufficiently detected.
以上の結果から、シャフト圧への主成分分析の適用が、高炉内での異常現象つまり炉況異常の検知に有効であることを、シミュレーションによって確認することができた。 From the above results, it was confirmed by simulation that the application of principal component analysis to shaft pressure is effective in detecting abnormal phenomena in the blast furnace, that is, abnormal furnace conditions.
実施例１において、本発明で採用する異常現象予知手法が有効であることをシミュレーションによって確認した。以下、本発明で採用する異常現象予知手法を実炉の測定値に適用し、その有効性を確認する。 In Example 1, it was confirmed by simulation that the abnormal phenomenon prediction method employed in the present invention was effective. In the following, the abnormal phenomenon prediction method adopted in the present invention is applied to the measured value of the actual furnace, and its effectiveness is confirmed.
図８は、実機高炉に取り付けた圧力センサーによって測定された、高炉の炉高方向各点のシャフト圧の時間変化を表している。通常、高炉には東西南北の４箇所、高炉の高さ方向に７〜８箇所シャフト圧の圧力計が設置されているが、図８は実機高炉の北方向における高さ方向７箇所の圧力を時系列的に示している。尚、この高炉は、図８に示すように、午前６時付近で吹き抜けを起こしている。 FIG. 8 represents the time change of the shaft pressure at each point in the blast furnace height direction, measured by a pressure sensor attached to the actual blast furnace. Normally, blast furnaces are equipped with four pressure gauges in the east, west, north and south, and 7 to 8 shaft pressures in the height direction of the blast furnace. Shown in time series. In addition, as shown in FIG. 8, this blast furnace has blown through around 6 am.
図８に示すように、吹き抜けの直前の午前４時付近では、シャフト圧の一部に変動が確認されるが、それ以前の時間帯でのシャフト圧の挙動からは吹き抜けを予感させる大きな異常は見られない。また、図８に示すように、シャフト圧のデータは複数点存在するため、これら全てのシャフト圧データを監視しながら異常を判定するのは極めて困難である。 As shown in FIG. 8, a change in the shaft pressure is confirmed at around 4 am immediately before the blow-through, but from the behavior of the shaft pressure in the time zone before that, a large abnormality that predicts blow-through is can not see. Also, as shown in FIG. 8, since there are a plurality of shaft pressure data, it is extremely difficult to determine an abnormality while monitoring all the shaft pressure data.
このシャフト圧の時間変化の全データについて主成分分析を適用した結果が、図９及び図１０である。図９は、Ｔ２統計量指数の分析結果を示し、図１０は、Ｑ統計量指数の分析結果を示す。 The result of applying the principal component analysis to all the data of the change in shaft pressure over time is shown in FIGS. FIG. 9 shows the analysis result of the T 2 statistic index, and FIG. 10 shows the analysis result of the Q statistic index.
図９及び図１０に示すように、主成分分析の結果、シャフト圧のデータからは不明瞭であった炉内の異常の予兆が、主成分分析によるＴ２統計量指数及びＱ統計量指数の両者において、午前０時頃から確認され、特にＱ統計量指数において明確に予兆が確認される。また、吹き抜け発生直前の午前４時付近においてシャフト圧が大きく変動した際はＱ統計量指数が大きく変動し、吹き抜けの発生を予知している。 As shown in FIGS. 9 and 10, as a result of the principal component analysis, a sign of abnormality in the furnace that was unclear from the shaft pressure data is the T 2 statistic index and the Q statistic index by the principal component analysis. In both cases, it is confirmed from around midnight, and a clear sign is confirmed particularly in the Q statistic index. In addition, when the shaft pressure fluctuates greatly around 4 am just before the occurrence of the blow-through, the Q statistic index fluctuates greatly, and the occurrence of blow-through is predicted.
このように、シャフト圧に対して主成分分析を適用すれば、複数点のシャフト圧データを監視することなく、１つ或いは２つのデータから高炉内の異常現象を予知することが可能である。 Thus, if the principal component analysis is applied to the shaft pressure, it is possible to predict an abnormal phenomenon in the blast furnace from one or two data without monitoring the shaft pressure data at a plurality of points.
以上説明したように、高炉内の異常現象つまり炉況異常の検知に関して、本発明の適用は有効であり、主成分分析を適用して得られたＱ統計量が特に有効であることが確認できた。 As described above, it can be confirmed that the application of the present invention is effective in detecting abnormal phenomena in the blast furnace, that is, abnormal furnace conditions, and that the Q statistic obtained by applying principal component analysis is particularly effective. It was.
実施例２において、実機高炉に取り付けた圧力センサーの測定値から算出したＱ統計量指数が高炉の吹き抜けの検知に有効であることが確認された。本実施例では、吹き抜け発生直前のＱ統計量の挙動と吹き抜けの発生確率との関係について検討を行った。 In Example 2, it was confirmed that the Q statistic index calculated from the measurement value of the pressure sensor attached to the actual blast furnace was effective in detecting the blast furnace blow-through. In this example, the relationship between the behavior of the Q statistic immediately before the occurrence of the blow-through and the occurrence probability of the blow-through was examined.
吹き抜けの発生要因は種々考えられるが、炉頂から装入された装入物の強度低下に伴う粉の発生による通気性悪化、或いは、装入前の装入物中に占める粉の割合が元々高いことによる装入物の空隙率低下に伴う通気性悪化によって引き起こされることが多い。また、炉頂から装入された装入物が羽口に到達するまでには通常８時間程度を要するため、性状が悪化した装入物が装入されてから８時間以内に吹き抜けが起こる確率は高い。したがって、高炉の吹き抜けの予兆を事前に検知するためには、現在から過去８時間以内のＱ統計量の挙動を監視すればよい。 There are various possible causes of blow-by, but the deterioration of the air permeability due to the generation of powder accompanying the decrease in strength of the charge charged from the top of the furnace, or the ratio of the powder in the charge before charging is originally It is often caused by a deterioration in air permeability associated with a decrease in the porosity of the charge due to the high. In addition, since it usually takes about 8 hours for the charge charged from the top of the furnace to reach the tuyere, the probability that a blow-through will occur within 8 hours after the charge with deteriorated properties is charged. Is expensive. Therefore, in order to detect in advance a sign of a blast furnace blow-through, the behavior of Q statistics within the past 8 hours from the present time may be monitored.
図１１は、実機高炉における吹き抜け発生直前のＱ統計量指数の時間推移の一例を示す図である。図１１では、時刻２１：００付近で１．０を上回るＱ統計量指数の最初のピークが発生し、それ以降、１．０を上回るＱ統計量指数のピークが立て続けに発生し、１．０を上回るＱ統計量指数のピークが複数回発生した後、１．０を上回る最初のピークが発生してから約３時間４０分経過した時点で吹き抜けが発生した。このように、Ｑ統計量には、吹き抜けの予兆として１．０を上回るピークが離散的に発生することを発明者らは確認している。 FIG. 11 is a diagram showing an example of the time transition of the Q statistic index immediately before the occurrence of the blow-through in the actual blast furnace. In FIG. 11, the first peak of the Q statistic index exceeding 1.0 is generated around 21:00, and thereafter, the peak of the Q statistic index exceeding 1.0 is generated in succession. After multiple occurrences of the peak of the Q statistic index exceeding 1, the blow-through occurred at about 3 hours and 40 minutes after the first peak exceeding 1.0 was generated. As described above, the inventors have confirmed that peaks exceeding 1.0 are generated in the Q statistic as a sign of blow-through.
一方、図１２は、同様に、実機高炉に取り付けた圧力センサーの測定値から算出したＱ統計量指数の時間推移の一例を示す図である。図１２では、図１１と同様にＱ統計量指数が１．０より大となるピークの発生が確認されているものの、１．０を上回る２回目のピーク発生以降にはＱ統計量指数は低位を保ち、８時間経過後も吹き抜けは発生せず、炉内の通気性は安定したことが認められている。 On the other hand, FIG. 12 is also a diagram showing an example of the time transition of the Q statistic index calculated from the measured value of the pressure sensor attached to the actual blast furnace. In FIG. 12, the occurrence of a peak having a Q statistic index greater than 1.0 is confirmed as in FIG. 11, but the Q statistic index is low after the second peak exceeding 1.0. It was confirmed that the air permeability in the furnace was stable without blowing through after 8 hours.
尚、図１１及び図１２では、Ｑ統計量指数を計算する際に使用したシャフト圧の測定値は、実機高炉から１分毎にサンプリングした値を使用している。 11 and 12, the measured value of the shaft pressure used when calculating the Q statistic index is a value sampled every minute from the actual blast furnace.
そこで、１．０を上回るＱ統計量指数の最初のピークの発生以降８時間以内の１．０を上回るピークの発生回数と、最初のピーク発生以降の８時間以内に発生した吹き抜けの発生回数との関係を調査した。ここで、吹き抜けとは、高炉炉頂のガス温度が３００℃を超えた場合と定義している。 Therefore, the number of occurrences of peaks exceeding 1.0 within 8 hours after the occurrence of the first peak of the Q statistic index exceeding 1.0, and the occurrence number of blow-throughs occurring within 8 hours after the occurrence of the first peak, The relationship was investigated. Here, the blow-by is defined as the case where the gas temperature at the top of the blast furnace exceeds 300 ° C.
表２に、最初の１．０を上回るＱ統計量指数のピーク発生から８時間以内の１．０を上回るＱ統計量指数のピークの発生回数と、最初のピーク発生以降の８時間以内に発生した吹き抜けの発生回数とを示す。 Table 2 shows the number of occurrences of Q statistic index peaks exceeding 1.0 within 8 hours from the first occurrence of Q statistic index peaks exceeding 1.0 and the occurrence within 8 hours after the first peak occurrence. The number of occurrences of blow-throughs.
表２に示すように、１．０を上回るＱ統計量指数の最初のピークの発生から８時間以内の１．０を上回るピーク発生数が、最初のピークを含めて合計４回以上の場合に、最初のピーク発生以降の８時間以内で吹き抜けが発生していることがわかる。つまり、１．０を上回るＱ統計量指数のピークが発生し、その後８時間以内に１．０を上回るＱ統計量指数のピークが３回発生した場合には、その後に吹き抜けの発生する可能性が高くなっているといえる。 As shown in Table 2, when the number of peaks exceeding 1.0 within 8 hours from the occurrence of the first peak of the Q statistic index exceeding 1.0 is 4 or more in total including the first peak It can be seen that blow-through has occurred within 8 hours after the first peak occurrence. In other words, if a Q statistic index peak exceeding 1.0 occurs, and if a Q statistic index peak exceeding 1.0 occurs three times within 8 hours, a blow-through may occur after that. Can be said to be higher.
即ち、任意の８時間以内に１．０を上回るピークが合計４回以上にならないようにするために、１．０を上回る最初のＱ統計量指数のピークの発生後の８時間以内に、１．０を上回るピークが最初のピークを含めて合計３回発生した時点で直ちに高炉への送風量を減じれば（減風）、吹き抜けを回避できる可能性があることがわかった。 That is, within 8 hours after the occurrence of the first Q statistic index peak above 1.0, in order to ensure that there are no more than 4 peaks above 1.0 in any 8 hours. It was found that if a peak exceeding 0.0 was generated a total of three times including the first peak, the blow-through could be avoided if the air flow to the blast furnace was reduced immediately (wind reduction).
そこで、本知見に基づき、実機高炉の操業において、１．０を上回る最初のＱ統計量指数のピーク発生以降８時間以内の１．０を上回るピーク発生回数が３回に到達した時点で減風する試験を行った。 Therefore, based on this knowledge, in actual blast furnace operation, the wind was reduced when the number of peak occurrences exceeding 1.0 within 8 hours after the peak occurrence of the first Q statistic index exceeding 1.0 reached 3 times. A test was conducted.
図１３は、本試験期間における実機高炉のＱ統計量指数の時間推移を表す図である。図１３に示すように、時刻４：１０付近で１．０を上回る最初のＱ統計量指数のピークが発生し、その後、８時間以内に１．０を上回るピークの発生が最初のピークを含めて合計３回確認されたが、３回目のピークの発生後、直ちに減風を行った結果、１．０を上回るＱ統計量指数のピークはその後発生せず、吹き抜けも発生せずに安定操業が継続可能であった。 FIG. 13 is a diagram showing the time transition of the Q statistic index of the actual blast furnace during the test period. As shown in FIG. 13, the first Q statistic index peak exceeding 1.0 occurs at around 4:10, and thereafter, the peak exceeding 1.0 is included within 8 hours including the first peak. A total of three times were confirmed, but as soon as the third peak occurred, the wind was reduced immediately. As a result, a peak of Q statistic index exceeding 1.0 did not occur afterwards, and stable operation did not occur. Could continue.
以上の結果から、任意の８時間以内における前記Ｑ統計量指数の１．０超えの回数が３回を超えると炉況異常が発生する可能性があると判定し、任意の８時間以内における前記Ｑ統計量指数の１．０超えの３回目のピークの発生後に、予知された炉況異常の発生を未然に防止するために、直ちに高炉操業条件を調整することが好ましいことが確認できた。 From the above results, it is determined that there is a possibility that a furnace condition abnormality may occur when the number of times the Q statistic index exceeds 1.0 in any 8 hours exceeds 3, and the above in any 8 hours It has been confirmed that it is preferable to immediately adjust the blast furnace operating conditions in order to prevent the occurrence of a predicted furnace condition abnormality before the third peak exceeding 1.0 of the Q statistic index.
尚、Ｑ統計量指数の挙動によっては連続的に１．０超えをする場合も見られるが、そのような場合は炉況が著しく悪化しているために減風による吹き抜けが回避不可であり、Ｑ統計量が吹き抜けの事前検知の指標と成り得なかったので除外した。 In addition, depending on the behavior of the Q statistic index, it can be seen that it continuously exceeds 1.0, but in such a case, the furnace conditions are extremely deteriorated, so it is impossible to avoid blow-by due to wind reduction. The Q statistic was excluded because it could not be an indicator for prior detection of the stairwell.
１ 高炉
２ 鉱石層
３ コークス層
1 Blast furnace 2 Ore layer 3 Coke layer
Claims (3)
操業中に前記圧力センサーで測定された複数点のシャフト圧データに対して主成分分析を行い、主成分分析によってＴ２統計量及び／またはＱ統計量を算出し、算出したＴ２統計量を正常な操業範囲のシャフト圧測定データに基づいて算出したＴ２統計量の最大値で除した指数（Ｔ２統計量指数）、及び／または、算出したＱ統計量を正常な操業範囲のシャフト圧測定データに基づいて算出したＱ統計量の最大値で除した指数（Ｑ統計量指数）を用いて、高炉操業の異常を予知することを特徴とする、高炉の操業方法。 A method of operating a blast furnace that operates while predicting abnormal operation of the blast furnace based on shaft pressure data at multiple points measured by a plurality of pressure sensors installed in the blast furnace,
A principal component analysis is performed on the shaft pressure data at a plurality of points measured by the pressure sensor during operation, T 2 statistics and / or Q statistics are calculated by principal component analysis, and the calculated T 2 statistics are calculated. An index (T 2 statistic index) divided by the maximum value of T 2 statistic calculated based on shaft pressure measurement data in the normal operating range and / or the calculated Q statistic of shaft pressure in the normal operating range A method of operating a blast furnace, characterized by predicting abnormalities in blast furnace operation using an index (Q statistic index) divided by the maximum value of Q statistics calculated based on measurement data.
操業中に前記圧力センサーで測定された複数点のシャフト圧データに対して主成分分析を行い、主成分分析によってＴ２統計量及び／またはＱ統計量を算出し、算出したＴ２統計量を正常な操業範囲のシャフト圧測定データに基づいて算出したＴ２統計量の最大値で除した指数（Ｔ２統計量指数）、及び／または、算出したＱ統計量を正常な操業範囲のシャフト圧測定データに基づいて算出したＱ統計量の最大値で除した指数（Ｑ統計量指数）を用いて高炉の炉況異常を予知し、高炉操業条件を調整することを特徴とする、高炉の操業方法。 A method of operating a blast furnace that operates while predicting abnormal blast furnace conditions from multiple points of shaft pressure data measured by multiple pressure sensors installed in the blast furnace,
A principal component analysis is performed on the shaft pressure data at a plurality of points measured by the pressure sensor during operation, T 2 statistics and / or Q statistics are calculated by principal component analysis, and the calculated T 2 statistics are calculated. An index (T 2 statistic index) divided by the maximum value of T 2 statistic calculated based on shaft pressure measurement data in the normal operating range and / or the calculated Q statistic of shaft pressure in the normal operating range Blast furnace operation characterized by predicting abnormal blast furnace conditions and adjusting blast furnace operating conditions using an index (Q statistic index) divided by the maximum Q statistic calculated based on measurement data Method.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2016007653 | 2016-01-19 | ||
| JP2016007653 | 2016-01-19 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| JP2017128805A true JP2017128805A (en) | 2017-07-27 |
Family
ID=59395565
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP2017006316A Pending JP2017128805A (en) | 2016-01-19 | 2017-01-18 | Operation method of blast furnace |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP2017128805A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2020030072A (en) * | 2018-08-21 | 2020-02-27 | Ｊｆｅスチール株式会社 | Method and device for determining occurrence of particle dispersed in air, and method and device for measuring property of lumpy matter |
| WO2020204043A1 (en) | 2019-04-03 | 2020-10-08 | Ｊｆｅスチール株式会社 | Blast furnace abnormality assessment device, blast furnace abnormality assessment method, and blast furnace operation method |
| WO2021033721A1 (en) | 2019-08-22 | 2021-02-25 | Ｊｆｅスチール株式会社 | Blast furnace irregularity assessment device, blast furnace irregularity assessment method, blast furnace operation method, and molten pig iron production method |
| WO2021095595A1 (en) | 2019-11-13 | 2021-05-20 | Ｊｆｅスチール株式会社 | Method and system for operating production facility |
| WO2021095594A1 (en) | 2019-11-13 | 2021-05-20 | Ｊｆｅスチール株式会社 | Production equipment monitoring method, production equipment monitoring apparatus, and production equipment operating method |
| CN112859815A (en) * | 2021-01-21 | 2021-05-28 | 中南大学 | Method for monitoring and diagnosing abnormal furnace conditions in roasting process |
| WO2021182041A1 (en) * | 2020-03-12 | 2021-09-16 | Ｊｆｅスチール株式会社 | Abnormality determination method for blast furnace, training method for stabilization period model, operation method for blast furnace, and abnormality determination device for blast furnace |
| RU2794126C1 (en) * | 2019-04-03 | 2023-04-11 | ДжФЕ СТИЛ КОРПОРЕЙШН | Device for determining blast furnace failure, method for detecting blast furnace fault and method for operating blast furnace |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH01319615A (en) * | 1988-06-17 | 1989-12-25 | Kawasaki Steel Corp | Method of operating blast furnace |
| JP2005133115A (en) * | 2003-10-28 | 2005-05-26 | Nippon Steel Corp | Method, apparatus and computer program for monitoring operation of blast furnace |
| JP2012083993A (en) * | 2010-10-13 | 2012-04-26 | Jfe Steel Corp | Operation condition management device |
| WO2014203509A1 (en) * | 2013-06-19 | 2014-12-24 | Ｊｆｅスチール株式会社 | Method for detecting abnormality in blast furnace, and method for operating blast furnace |
| WO2015015936A1 (en) * | 2013-07-29 | 2015-02-05 | Ｊｆｅスチール株式会社 | Abnormality detection method and blast-furnace operation method |
-
2017
- 2017-01-18 JP JP2017006316A patent/JP2017128805A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH01319615A (en) * | 1988-06-17 | 1989-12-25 | Kawasaki Steel Corp | Method of operating blast furnace |
| JP2005133115A (en) * | 2003-10-28 | 2005-05-26 | Nippon Steel Corp | Method, apparatus and computer program for monitoring operation of blast furnace |
| JP2012083993A (en) * | 2010-10-13 | 2012-04-26 | Jfe Steel Corp | Operation condition management device |
| WO2014203509A1 (en) * | 2013-06-19 | 2014-12-24 | Ｊｆｅスチール株式会社 | Method for detecting abnormality in blast furnace, and method for operating blast furnace |
| WO2015015936A1 (en) * | 2013-07-29 | 2015-02-05 | Ｊｆｅスチール株式会社 | Abnormality detection method and blast-furnace operation method |
Cited By (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2020030072A (en) * | 2018-08-21 | 2020-02-27 | Ｊｆｅスチール株式会社 | Method and device for determining occurrence of particle dispersed in air, and method and device for measuring property of lumpy matter |
| EP3950966A4 (en) * | 2019-04-03 | 2022-05-11 | JFE Steel Corporation | Blast furnace abnormality assessment device, blast furnace abnormality assessment method, and blast furnace operation method |
| WO2020204043A1 (en) | 2019-04-03 | 2020-10-08 | Ｊｆｅスチール株式会社 | Blast furnace abnormality assessment device, blast furnace abnormality assessment method, and blast furnace operation method |
| JP6825753B1 (en) * | 2019-04-03 | 2021-02-03 | Ｊｆｅスチール株式会社 | Blast furnace abnormality determination device, blast furnace abnormality determination method, and blast furnace operation method |
| TWI745912B (en) * | 2019-04-03 | 2021-11-11 | 日商杰富意鋼鐵股份有限公司 | Blast furnace abnormality determination device, blast furnace abnormality determination method and blast furnace operation method |
| KR102574567B1 (en) * | 2019-04-03 | 2023-09-04 | 제이에프이 스틸 가부시키가이샤 | Blast furnace fault determination apparatus, method for determining fault in blast furnace, and method for operating blast furnace |
| RU2794126C1 (en) * | 2019-04-03 | 2023-04-11 | ДжФЕ СТИЛ КОРПОРЕЙШН | Device for determining blast furnace failure, method for detecting blast furnace fault and method for operating blast furnace |
| CN113614253A (en) * | 2019-04-03 | 2021-11-05 | 杰富意钢铁株式会社 | Blast furnace abnormality determination device, blast furnace abnormality determination method, and blast furnace operation method |
| KR20210127223A (en) | 2019-04-03 | 2021-10-21 | 제이에프이 스틸 가부시키가이샤 | Abnormality determination apparatus of a blast furnace, abnormality determination method of a blast furnace, and operation method of a blast furnace |
| WO2021033721A1 (en) | 2019-08-22 | 2021-02-25 | Ｊｆｅスチール株式会社 | Blast furnace irregularity assessment device, blast furnace irregularity assessment method, blast furnace operation method, and molten pig iron production method |
| JPWO2021033721A1 (en) * | 2019-08-22 | 2021-09-13 | Ｊｆｅスチール株式会社 | Blast furnace abnormality judgment device, blast furnace abnormality judgment method, blast furnace operation method and hot metal manufacturing method |
| EP3985132A4 (en) * | 2019-08-22 | 2022-09-07 | JFE Steel Corporation | Blast furnace irregularity assessment device, blast furnace irregularity assessment method, blast furnace operation method, and molten pig iron production method |
| CN114258433A (en) * | 2019-08-22 | 2022-03-29 | 杰富意钢铁株式会社 | Blast furnace abnormality determination device, blast furnace abnormality determination method, blast furnace operation method, and molten iron manufacturing method |
| TWI748604B (en) * | 2019-08-22 | 2021-12-01 | 日商杰富意鋼鐵股份有限公司 | Blast furnace abnormality judgment device, blast furnace abnormality judgment method, blast furnace operation method, and molten iron manufacturing method |
| KR20220035233A (en) | 2019-08-22 | 2022-03-21 | 제이에프이 스틸 가부시키가이샤 | Abnormality determination apparatus for a blast furnace, an abnormality determination method for a blast furnace, a method for operating a blast furnace, and a manufacturing method for molten iron |
| KR20220070519A (en) | 2019-11-13 | 2022-05-31 | 제이에프이 스틸 가부시키가이샤 | A production facility monitoring method, a production facility monitoring device, and a production facility operating method |
| KR20220073806A (en) | 2019-11-13 | 2022-06-03 | 제이에프이 스틸 가부시키가이샤 | Operation method and operation system of production equipment |
| WO2021095594A1 (en) | 2019-11-13 | 2021-05-20 | Ｊｆｅスチール株式会社 | Production equipment monitoring method, production equipment monitoring apparatus, and production equipment operating method |
| WO2021095595A1 (en) | 2019-11-13 | 2021-05-20 | Ｊｆｅスチール株式会社 | Method and system for operating production facility |
| JPWO2021182041A1 (en) * | 2020-03-12 | 2021-09-16 | ||
| WO2021182041A1 (en) * | 2020-03-12 | 2021-09-16 | Ｊｆｅスチール株式会社 | Abnormality determination method for blast furnace, training method for stabilization period model, operation method for blast furnace, and abnormality determination device for blast furnace |
| JP7192992B2 (en) | 2020-03-12 | 2022-12-20 | Ｊｆｅスチール株式会社 | Blast Furnace Abnormality Judgment Method, Stable Period Model Learning Method, Blast Furnace Operation Method, and Blast Furnace Abnormality Judgment Device |
| CN112859815B (en) * | 2021-01-21 | 2022-05-13 | 中南大学 | Method for monitoring and diagnosing abnormal furnace conditions in roasting process |
| CN112859815A (en) * | 2021-01-21 | 2021-05-28 | 中南大学 | Method for monitoring and diagnosing abnormal furnace conditions in roasting process |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP2017128805A (en) | Operation method of blast furnace | |
| US10527520B2 (en) | Operating wind motors and determining their remaining useful life | |
| US7774165B2 (en) | State monitoring of machines and technical installations | |
| US20140067327A1 (en) | Similarity curve-based equipment fault early detection and operation optimization methodology and system | |
| US20080288183A1 (en) | Systems and methods for monitoring energy system components | |
| JP2011033024A (en) | System and method for online monitoring of corrosion of gas turbine component | |
| JP2012221493A (en) | Method and system for monitoring operation of apparatus | |
| CN107710089B (en) | Plant equipment diagnosis device and plant equipment diagnosis method | |
| JP6607821B2 (en) | Blast furnace sensor failure detection method and abnormal situation prediction method | |
| CN104615123B (en) | K-nearest neighbor based sensor fault isolation method | |
| JP6686947B2 (en) | Blast furnace state determination device and blast furnace operating method | |
| EP2631724B1 (en) | Method for measuring health index of plant in which state of lower component is reflected, and computer-readable storage medium in which program for performing the method is stored | |
| CN112668794A (en) | Maintenance optimization method for aircraft engine under uncertain life prediction information | |
| JP2018113027A (en) | Process abnormality state diagnosis method and abnormality state diagnosis device | |
| JP6940030B2 (en) | Blast furnace abnormality judgment device, blast furnace abnormality judgment method, blast furnace operation method and hot metal manufacturing method | |
| CN104133437B (en) | Continuous-type chemical-engineering device and performance indicator real-time evaluation method and device thereof | |
| JP2010283274A (en) | Failure cause analyzer, failure cause analysis method, failure cause analysis program, and program recording medium | |
| Zhang et al. | Real-time residual life prediction based on kernel density estimation considering abrupt change point detection | |
| JP7307341B2 (en) | Furnace gas pressure fluctuation detection method | |
| TWI435936B (en) | Method for predicting channeling phenomenon of blast furnace | |
| Dutta et al. | Reliability analysis of defence vehicles gear box assembly under preventive maintenance | |
| JP7217593B2 (en) | Predictive diagnosis system | |
| CN112912807A (en) | Abnormality diagnosis device | |
| JP3487203B2 (en) | Blast furnace condition prediction method | |
| Yao et al. | Fault Detection and Diagnosis of Alumina Feeding System Using Individual Anode Current Measurement |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A621 | Written request for application examination |
Free format text: JAPANESE INTERMEDIATE CODE: A621 Effective date: 20170824 |
|
| RD03 | Notification of appointment of power of attorney |
Free format text: JAPANESE INTERMEDIATE CODE: A7423 Effective date: 20180502 |
|
| RD04 | Notification of resignation of power of attorney |
Free format text: JAPANESE INTERMEDIATE CODE: A7424 Effective date: 20180509 |
|
| A977 | Report on retrieval |
Free format text: JAPANESE INTERMEDIATE CODE: A971007 Effective date: 20180619 |
|
| A131 | Notification of reasons for refusal |
Free format text: JAPANESE INTERMEDIATE CODE: A131 Effective date: 20180717 |
|
| A521 | Request for written amendment filed |
Free format text: JAPANESE INTERMEDIATE CODE: A523 Effective date: 20180906 |
|
| A02 | Decision of refusal |
Free format text: JAPANESE INTERMEDIATE CODE: A02 Effective date: 20181002 |