JP4675523B2 - Method for calculating the terrace length of the raw material deposition layer at the top of the blast furnace furnace - Google Patents

Method for calculating the terrace length of the raw material deposition layer at the top of the blast furnace furnace Download PDF

Info

Publication number
JP4675523B2
JP4675523B2 JP2001268153A JP2001268153A JP4675523B2 JP 4675523 B2 JP4675523 B2 JP 4675523B2 JP 2001268153 A JP2001268153 A JP 2001268153A JP 2001268153 A JP2001268153 A JP 2001268153A JP 4675523 B2 JP4675523 B2 JP 4675523B2
Authority
JP
Japan
Prior art keywords
furnace
measurement point
value
measurement
terrace length
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.)
Expired - Fee Related
Application number
JP2001268153A
Other languages
Japanese (ja)
Other versions
JP2003073716A (en
Inventor
洋 大楠
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Nippon Steel Nisshin Co Ltd
Original Assignee
Nippon Steel Nisshin Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Nippon Steel Nisshin Co Ltd filed Critical Nippon Steel Nisshin Co Ltd
Priority to JP2001268153A priority Critical patent/JP4675523B2/en
Publication of JP2003073716A publication Critical patent/JP2003073716A/en
Application granted granted Critical
Publication of JP4675523B2 publication Critical patent/JP4675523B2/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Images

Landscapes

  • Manufacture Of Iron (AREA)

Description

【0001】
【発明が属する技術分野】
本発明は、高炉炉頂部の原料堆積層の表面形状をもとに炉壁近傍のテラス長さをコンピュータを用いて演算する方法に関する。
【0002】
【従来技術】
高炉の操業を安定維持させるためには、原料堆積層のプロフィルを最適形状に維持することが重要である。換言すれば、炉内のガス流が原料堆積層のプロフィルによって変化することから、プロフィルを最適形状に維持することによりガス流の分布が最適化し、もって高炉操業の安定化を図ることができる。
【0003】
原料堆積層のプロフィルデータ情報を高炉操業の管理に活用する事例として、具体的には、特開2000−212612号に示されるように、出銑口間で出銑、出滓量がアンバランス化する原因となる生鉱石下りの発生を未然に防止するために活用した事例、特開平2−225608号に示されるように、指尺乱れや炉頂圧力変動回数の低減に活用した事例がある。
こうした事例では、原料堆積層のプロフィルを定量的に示す管理値としてテラス長さ、すなわち炉中心から炉壁に向かって堆積形状の傾斜角度が特定の角度、例えば15°未満になる位置の肩部から炉壁までの水平距離が求められている。
【0004】
上記のテラス長さは、マイクロ波式プロフィール計等の測深装置を用い、炉の半径方向に沿って任意の間隔ごとに鉱石及びコークスの装入前後における、特定の基準レベルから堆積層表面までの深度を測定して得た表面形状データから得られる。具体的にはこのデータから人為的に図1に示すように傾斜部の直線bと、炉壁側への直線aを描き、その交点cから炉壁wまでの水平距離dを物指しを当てゝ読み取ったり、或いは上記基準レベルからの深度データ(以下、単に「深度データ」という)をコンピュータに入力して図2に示すように隣り合う2点間の深度データYn、Yn+1からその区間ΔX内での勾配θ=tan 1(ΔY/ΔX)を各測定点ごとにそれぞれ求める。図3は、こうして求めた各測定点での勾配(傾斜角)を示す。次に図3に示すように炉壁wから最初に+15°を越えるポイントeを見付けて、このポイントeから炉壁wまでの水平距離d’を演算処理してテラス長さを求めていた。
【0005】
【発明が解決しようとする課題】
テラス長さを人為的に求める前者の方法は、テラス長さを比較的精度よく求めることができる反面、手間と時間がかゝる難点がある。これに対し後者のコンピュータにより演算処理して求める方法は、テラス長さが自動的に算出されるが、ノイズによって精度が悪く、テラス長さの演算値と人為的に読み取られた値とが乖離してばらつきが大きくなる、という問題があった。
本発明は、テラス長さをコンピュータにより演算処理して求める方法において、ノイズの影響を除去して演算値の精度を向上させることができるようにしたものである。
【0006】
【課題の解決手段】
請求項1に係わる発明は、高炉炉頂部の原料堆積層の表面形状をもとに炉壁近傍のテラス長さをコンピュータを用いて演算する方法であって、
(1)測深装置、例えばマイクロ波式或いはレーザ式のプロフィール計を用いて原料堆積層表面までの深度を炉半径方向に沿って任意の間隔ごとに測定するプロセスと、
(2)上記(1)の工程で測定した各測定点のうち、炉壁側と炉中心側の測定点を除く各測定点において、その前後の複数の測定点の深度データを用いて多項式適用による平滑化データ処理を行うプロセスと、
(3)測定点iから前後にn点離れた測定点i−nとi+nにおける、上記(2)の工程で平滑化処理された深度データから中心差分による一次微分近似処理を行い、測定点iにおける堆積層表面の傾斜角を算出するプロセスと、
(4)上記(3)の工程で算出された各測定点の傾斜角に関するデータを炉半径方向にスキャニングし、傾斜角がしきい値未満となる測定点を検出するプロセスと、
(5)上記(4)の工程でしきい値未満として検出された測定点と、その近傍の測定点の上記傾斜角に関するデータより最小二乗法を用いて一次式を求め、この一次式としきい値が一致する位置を検出するプロセスと、
(6)上記(5)の工程で検出した、しきい値と一致する位置より炉壁までの水平距離を求めるプロセス
よりなることを特徴とする。
【0007】
以下、上記各プロセスの態様について詳述する。
(1)のプロセスにおいては、炉頂部に鉱石またはコークス原料を装入したのち、測深装置、例えばマイクロ波式プロフィール計を用い、炉壁から炉中心まで炉半径方向に移動させて一定間隔ごと、例えば10cm間隔で原料堆積層までの深度を測定し、得られたデータをプロセスコンピュータに入力する。
【0008】
図4は、各測定点における深度の一例を示すグラフである。
(2)のプロセスにおいてプロセスコンピュータは、図5に示すように任意の測定点iと、その前後各2点のi−2、i−1及びi+1、i+2を加えた計5点の深度データを用い、Sevitzky-Golay法により二次曲線y=ax2+bx+cで平滑化処理して、測定点iにおける上記二次曲線上の深度y(i)を算出する。図5の白抜き丸は、上記二次曲線により平滑化処理された測定点iの深度y(i)を示す。続いて計測点i+1について、その前後各2点のi−1、i及びi+2、i+3を加えた計5点の深度データを用いて同様にして平滑化処理を行い、測定点i+1における深度y(i+1)を算出する。この算出が、炉壁側の2点と、炉中心側の2点を除く各測定点について順次行われる。ここで炉壁側と炉中心側の2点の測定点を除いたのは、前後に二点ずつの測定点を確保できないからである。平滑化処理に用いるデータ数kを変更して、例えば前後の各一点を加えた計3点の深度データを用いて平滑化処理を行う場合には、二番目の測定点から二次曲線上の平滑化処理された深度が求められる。
【0009】
(3)のプロセスにおいてプロセスコンピュータは、図6に示すように任意の測定点iについて、その前後n点、図示する例においては2点離れた測定点における(2)のプロセスにより平滑化処理された深度データy(i−2)とy(i+2)からその間の傾斜角θを以下の数1式と数2式により算出する。
【0010】
【数1】

Figure 0004675523
【0011】
【数2】
Figure 0004675523
ここで、Δxは測定点のピッチを示す。
【0012】
コンピュータを用いた数値計算法に関して広く一般的に知られている“一次微分近似”の方法の一つに以下の数3式に示す“中心差分”と呼ばれる手法がある。
この手法は、測定データ内の注目点iでの傾斜角を求めるに当たって、その注目点iの両隣りの点(i+1とi−1)の区間内の平均的な勾配を中央に位置する注目点iの勾配として近似するものである。
【0013】
【数3】
Figure 0004675523
一方、本発明では、数1式に示すように、測定データ内の注目点iから前後にそれぞれn点(nは1以上の整数値)離れた点(i+nとi−n)区間内の平均的な勾配を、その中央に位置する注目点iの勾配であるとする。つまり、ランダムノイズを平滑化処理して除去する働きを持たせ、その加減を調整できるように従来の中心差分法に改良を加えたものである。なお、数1式においてn=1の場合は、数3式と同一となる。
【0014】
図7は、上記数1式及び数2式により算出された各測定点における傾斜角θを示す。
(4)のプロセスにおいてプロセスコンピュータは、各測定点での傾斜角θを炉中心から炉壁に向かってデータスキャニングする。そして傾斜角がしきい値、例えば15°未満となる測定点gを検出する。この場合、図8に示すように傾斜角が15°未満となる測定点g2 を検出したのち、更なるスキャニングによってしきい値、例えば25°を越える場合には更に炉壁側にしきい値15°未満の測定点が存在すると認識させ、再びしきい値を15°未満の測定点を探索する。 g1 はこれによって検出された測定点を示す。
【0015】
(5)のプロセスにおいてプロセスコンピュータは、図7でいえば、測定点gと、その前後のしきい値15°を挟む測定点のデータを基に最小二乗法を用いて一次式Y=AX+Bを求め、しきい値15°と一致する位置、すなわちX=G、Y=15となるGを検出する。
【0016】
(6)のプロセスにおいてプロセスコンピュータは、上記(5)のプロセスで求めたGより炉壁wまでの水平距離lを算出する。
以上のようなプロセスよりなる方法によると、平滑化処理によりランダムノイズが除去され、一次微分近似処理を行うことにより、データが更に平滑化され、ノイズをより一層除去することができる。
【0017】
請求項2に係わる発明は、請求項1に係わる発明における(2)のプロセスが省かれ、(3)のプロセスにおいて用いられる平滑化処理された深度データの代わりに(1)のプロセスで測定した深度データが用いられることを特徴とする。
本発明者らの計算結果によると、上記(1)のプロセスで測定したデータを平滑化処理しないで直接、中心差分による一次近似処理を行っても、請求項1に係わる発明とほゞ同様の精度でテラス長さを求めることができた。
【0018】
請求項3に係わる発明は、請求項1又は2に係わる発明において、上記(5)のプロセスを省き、上記(4)のプロセスにおいて検出された、しきい値未満となる測定点より炉壁までの水平距離を求めることを特徴とする。
測定点のピッチを小さくする程、上記(4)のプロセスで検出した測定点の傾斜角がしきい値に近くなり、(5)の工程で検出された位置と近似される。図6を例にとっていえば、測定点のピッチが小さくなる程、しきい値(15°)に近く、しきい値に近似した測定点が検出できるようになる。
【0019】
【実施例】
実施例1
高炉炉頂部にコークス装入後、マイクロ波式プロフィール計を用いて炉壁から炉中心まで半径方向に移動させ、10cm間隔で42か所、コークス層までの深度を測定し、プロセスコンピュータにデータ入力した。
【0020】
プロセスコンピュータには、初期条件として上述するプロセス(2)における深度データを平滑化処理する際に用いる測定点の数kを5、プロセス(3)の数1式と数2式に用いられるnを2、プロセス(4)のしきい値を15°に設定し、上述の測定データ入力からコークスのテラス長さである上記プロセス(6)の水平距離lを算出した。
これとは別にコークス層表面までの42か所の深度データを用いて図1に示すような方法で物差しを当てゝ水平距離dのテラス長さを読取り、これをプロセスコンピュータに入力してその差を求めた。
【0021】
以上のテラス長さの計算及び読み取りを、毎日1回、コークスを装入したのちに深度測定を行って得た73日分の深度データを用い、1日分を1回とカウントして、計73回行った。このうちコンピュータで計算でき、テラス長さが求められたのは69回であった。この69回について、演算値と読取値の差sの平均値と標準偏差をプロセスコンピュータで算出した。その結果、平均値は計算値が読取値より0.08m小さな値として得られた。また標準偏差は0.10mであった。
次にテラス長さの読取値と計算値の相関を調べたところ、読取値と計算値をプロットした図9において、一次式の傾きは0.906、相関係数Rの二乗R2は0.501であった。
【0022】
実施例2
プロセスコンピュータに入力された初期条件をn=3とする以外は実施例1と同じにし、実施例1で得られた測定データを基にして計算値と読取値より、その差の平均値と標準偏差を求めたところ、平均値は計算値が読取値より0.09m少なく、標準偏差は0.11mであった。
また、読取値と計算値との相関を実施例1と同様にして求めたところ、一次式の傾きは0.883、R2 は0.229で、テラス長さをコンピュータで求められたのは73回中、62回であった。
【0023】
実施例3
プロセスコンピュータに入力される初期条件を、平滑化処理する際に用いる深度データの測定点の数kを0、すなわち平滑化処理を行わないで、n=3とする以外は実施例1と同じにし、実施例1で得られた測定データを基にして計算値と読取値より、その差の平均値と標準偏差を求めたところ、平均値は計算値が読取値より0.10m少なく、また標準偏差は0.12mであった。
また、読取値と計算値との相関を実施例1と同様にして求めたところ、一次式の傾きは0.872、R2 は0.449で、テラス長さをコンピュータで求めることができたのは73回中、72回であった。
【0024】
実施例4
プロセスコンピュータに入力される初期条件を平滑化処理するための測定点の数kを5、n=2、しきい値を17.5°に設定し、実施例1で得られた測定データを基にして計算値と読取値より、その差の平均値と標準偏差を求めたところ、平均値は計算値が読取値より0.02m少なく、標準偏差は0.967mであった。
また読取値と計算値の相関を実施例1と同様にして求めたところ、図10に示すように一次式の傾きは0.967、R2 は0.666でテラス長さは73回共全てコンピュータで求めることができた。
【0025】
実施例5
プロセスコンピュータに入力される初期条件をn=3とする以外は、実施例4と同じにし、実施例1で得られた測定データを基にして計算値と読取値より、その差の平均値と標準偏差を求めたところ、平均値は計算値が読取値より0.03m少なく、標準偏差は0.11mであった。
また読取値と計算値の相関を実施例1と同様にして求めたところ、一次式の傾きは0.990、R2 は0.179で、テラス長さをコンピュータで求めることができたのは73回中、68回であった。
【0026】
実施例6
プロセスコンピュータに入力される初期条件を平滑化処理するための測定点の数kを0とし、n=3とする以外は実施例4と同じにし、実施例1で得られた測定データを基にして計算値と読取値より、その差の平均値と標準偏差を求めたところ、平均値は計算値が読取値より0.03m少なく、標準偏差は0.12mであった。
また読取値と計算値の相関を実施例1と同様にして求めたところ、一次式の傾きは0.952、R2 は0.422で、テラス長さは73回共、全てコンピュータで求めることができた。
【0027】
実施例7
プロセスコンピュータに入力される初期条件を、平滑化処理するための測定点の数kを5、n=2、しきい値を18.5°に設定し、実施例1で得られた測定データを基にして計算値と読取値より、その差の平均値と標準偏差を求めたところ、平均値は計算値と読取値が一致し、標準偏差は0.09mであった。
図11に示すように、一次式の傾きは0.990、R2 は0.620で、テラス長さは73回共、全てコンピュータで求めることができた。
【0028】
実施例8
プロセスコンピュータに入力される初期条件を、n=3とする以外は実施例7と同じにし、実施例1で得られた測定データを基にして計算値と読取値より、その差の平均値と標準偏差を求めたところ、平均値は計算値が読取値より0.01m少なく、標準偏差は0.11mであった。
また読取値と計算値の相関を実施例1と同様にして求めたところ、図12に示すように、一次式の傾きは0.982、R2 は0.243で、テラス長さをコンピュータで求めることができたのは、73回中、70回であった。
【0029】
実施例9
プロセスコンピュータに入力される初期条件を平滑化処理するための測定点の数kを0とし、n=3とする以外は実施例7と同じにし、実施例1で得られた測定データを基にして計算値と読取値より、その差の平均値と標準偏差を求めたところ、平均値は計算値と読取値が一致し、標準偏差は0.12であった。
また読取値と計算値の相関を実施例1と同様にして求めたところ、図13に示すように、一次式の傾きは0.987、R2 は0.406で、テラス長さは73回共、全てコンピュータで求めることができた。
以上の結果を以下の表1に示す。
【0030】
【表1】
Figure 0004675523
なお、表中、計算値−読取値の平均値を記載した欄の数値に付した−(マイナス)符号は、計測値が読取値より少ないことを示す。
【0031】
表1に見られるように、プロセスコンピュータに入力される初期条件k=5、n=2としたものが計算値と読取値の差が最も少なく、ばらつきが少なくなり、しきい値を15°→17.5°→18.5°と上げる程改善され、実施例7で最善の結果が得られること、k=5、n=3と平滑化する程、計算不能となるケースが増える傾向があること、実施例3、6、9のように、k=0の平滑化処理を行わない場合でも、計算値と読取値の差が少なく、ばらつきを少なくできること等が分かった。
【0032】
なお、上述するしきい値の設定は、図1に示す傾斜部における直線b領域の傾斜角から、炉壁近傍のフラット部における直線a領域の傾斜角(ゼロ)に変曲する点、即ち交点cの傾斜角を採用すると良い。
つまり、鉱石の場合は、傾斜部bの傾斜角は30°前後であるから、しきい値としてはその半分の15°を採用し、傾斜部bの傾斜角が37°程度であるコークスの場合は、しきい値に18.5°を採用すると良く、堆積原料によってしきい値の値を使い分けると、テラス長さの演算精度はより向上する。
【0033】
【発明の効果】
請求項1に係わる発明によると、測深装置により測定して得た深度データを平滑化処理し、かつ中心差分による一次近似処理を行うことにより、ノイズが除去され、テラス長さを精度よく求めることができる。
【0034】
請求項2に係わる発明によると、請求項1に係わる発明に比べ、平滑化処理するためのプロセス2がない分、プロセスが簡素化され、計算不能となるケースも少なくなり、テラス長さも請求項1に係わる発明とほゞ同様の精度で求めることができる。
【0035】
請求項3に係わる発明によると、請求項1に係わる発明に比べ、平滑化処理するためのプロセス(5)がない分、プロセスが簡素化され、測定点のピッチを小さくすることにより、テラス長さを請求項1に係わる発明と同様の精度で求めることができる。
【図面の簡単な説明】
【図1】テラス長さを人為的に求める場合について示す図。
【図2】勾配を計算するための説明図。
【図3】テラス長さをコンピュータを用いて計算して求める場合の従来法について示す図。
【図4】深度データのグラフ。
【図5】平滑化処理方法を示す図。
【図6】中心差分による傾斜角を求める方法について示す図。
【図7】テラス長さを求める方法について示す図。
【図8】しきい値未満の傾斜角が二か所ある場合のグラフ。
【図9】実施例1の読取値と計算値の相関を示すグラフ。
【図10】実施例4の読取値と計算値の相関を示すグラフ。
【図11】実施例7の読取値と計算値の相関を示すグラフ。
【図12】実施例8の読取値と計算値の相関を示すグラフ。
【図13】実施例9の読取値と計算値の相関を示すグラフ。[0001]
[Technical field to which the invention belongs]
The present invention relates to a method of calculating a terrace length in the vicinity of a furnace wall using a computer based on the surface shape of a raw material deposition layer at the top of a blast furnace furnace.
[0002]
[Prior art]
In order to stably maintain the operation of the blast furnace, it is important to maintain the profile of the raw material deposition layer in an optimum shape. In other words, since the gas flow in the furnace changes depending on the profile of the raw material deposition layer, the distribution of the gas flow can be optimized by maintaining the profile in the optimum shape, thereby stabilizing the blast furnace operation.
[0003]
As an example of utilizing raw material deposition layer profile data information for blast furnace operation management, specifically, as shown in Japanese Patent Laid-Open No. 2000-212612, the output and output amount are unbalanced between the output ports. There is a case where it is used to prevent the occurrence of raw ore descending, which is the cause of this, and a case where it is used to reduce finger scale disturbance and the number of fluctuations in top pressure as shown in JP-A-2-225608.
In such a case, as a control value that quantitatively shows the profile of the raw material deposition layer, the terrace length, that is, the shoulder at a position where the inclination angle of the deposition shape from the furnace center toward the furnace wall becomes a specific angle, for example, less than 15 °. The horizontal distance from the furnace wall to the furnace wall is required.
[0004]
The above terrace length is measured from a specific reference level to the surface of the deposited layer before and after charging of ore and coke at arbitrary intervals along the radial direction of the furnace using a depth measuring device such as a microwave profile meter. It is obtained from the surface shape data obtained by measuring the depth. Specifically, from this data, a straight line b of the inclined portion and a straight line a to the furnace wall side are artificially drawn as shown in FIG. 1, and the horizontal distance d from the intersection c to the furnace wall w is applied to the object. Read or input depth data from the above-mentioned reference level (hereinafter simply referred to as “depth data”) to the computer and, as shown in FIG. 2, from the depth data Yn, Yn + 1 between two adjacent points, the section ΔX The gradient θ = tan 1 (ΔY / ΔX) is obtained for each measurement point. FIG. 3 shows the gradient (inclination angle) at each measurement point thus obtained. Next, as shown in FIG. 3, a point e exceeding + 15 ° was first found from the furnace wall w, and the terrace length was obtained by calculating the horizontal distance d ′ from this point e to the furnace wall w.
[0005]
[Problems to be solved by the invention]
The former method for artificially determining the terrace length can determine the terrace length with relatively high accuracy, but has the disadvantage that it takes time and effort. On the other hand, in the latter method, the terrace length is calculated automatically by the computer, but the accuracy is poor due to noise, and the calculated terrace length and the artificially read value are different. As a result, there is a problem that the variation becomes large.
According to the present invention, in the method of calculating the terrace length by calculating with a computer, the influence of noise can be removed and the accuracy of the calculated value can be improved.
[0006]
[Means for solving problems]
The invention according to claim 1 is a method of calculating the terrace length near the furnace wall using a computer based on the surface shape of the raw material deposition layer at the top of the blast furnace furnace,
(1) a process of measuring the depth to the surface of the raw material deposition layer at arbitrary intervals along the furnace radial direction using a depth measuring device, for example, a microwave type or laser type profile meter;
(2) Among the measurement points measured in the above step (1), at each measurement point excluding the measurement points on the furnace wall side and the furnace center side, a polynomial is applied using depth data of a plurality of measurement points before and after the measurement point. Process for smoothing data processing by
(3) Perform a first-order differential approximation process using the center difference from the depth data smoothed in the step (2) at the measurement points i−n and i + n that are n points away from the measurement point i. A process of calculating the tilt angle of the surface of the deposited layer in
(4) A process of scanning data regarding the inclination angle of each measurement point calculated in the step (3) in the furnace radial direction to detect a measurement point at which the inclination angle is less than a threshold value;
(5) A linear equation is obtained by using the least square method from the measurement point detected as less than the threshold value in the step (4) and data on the inclination angle of the measurement point in the vicinity thereof, and this linear equation is set as the threshold. The process of finding the location where the values match,
(6) The method includes a process of obtaining a horizontal distance from the position coincident with the threshold value detected in the step (5) to the furnace wall.
[0007]
Hereinafter, the aspect of each said process is explained in full detail.
In the process of (1), ore or coke raw material is charged at the top of the furnace, and then moved to the furnace radial direction from the furnace wall to the furnace center using a depth measuring device, for example, a microwave profile meter, at regular intervals. For example, the depth to the raw material deposition layer is measured at intervals of 10 cm, and the obtained data is input to the process computer.
[0008]
FIG. 4 is a graph showing an example of the depth at each measurement point.
In the process (2), as shown in FIG. 5, the process computer obtains a total of five depth data including an arbitrary measurement point i and two points i-2, i-1 and i + 1, i + 2 before and after the arbitrary measurement point i. The depth y (i) on the quadratic curve at the measurement point i is calculated by using the Sevitzky-Golay method and smoothing with the quadratic curve y = ax 2 + bx + c. The white circles in FIG. 5 indicate the depth y (i) of the measurement point i smoothed by the quadratic curve. Subsequently, with respect to the measurement point i + 1, smoothing processing is similarly performed using depth data of a total of five points obtained by adding i−1, i, i + 2, and i + 3 at two points before and after the measurement point i + 1, and the depth y ( i + 1) is calculated. This calculation is sequentially performed for each measurement point excluding two points on the furnace wall side and two points on the furnace center side. The reason why the two measurement points on the furnace wall side and the furnace center side are excluded here is that two measurement points before and after cannot be secured. When the number of data k used in the smoothing process is changed, and the smoothing process is performed using, for example, a total of three points of depth data obtained by adding one point before and after, the second measurement point on the quadratic curve The smoothed depth is determined.
[0009]
In the process (3), the process computer is smoothed by the process (2) at an arbitrary measurement point i, as shown in FIG. 6, at n points before and after the measurement point, in the example shown, two measurement points apart. The inclination angle θ between the depth data y (i−2) and y (i + 2) is calculated by the following equations 1 and 2.
[0010]
[Expression 1]
Figure 0004675523
[0011]
[Expression 2]
Figure 0004675523
Here, Δx represents the pitch of the measurement points.
[0012]
One of the methods of “first-order differential approximation” that is widely known as a numerical calculation method using a computer is a method called “center difference” shown in the following equation (3).
In this method, when the inclination angle at the point of interest i in the measurement data is obtained, the point of interest is located at the center of the average gradient in the section of the points (i + 1 and i-1) adjacent to the point of interest i. It approximates as the gradient of i.
[0013]
[Equation 3]
Figure 0004675523
On the other hand, in the present invention, as shown in Equation 1, the average in the interval (i + n and i−n) that is n points away from the point of interest i in the measurement data (n is an integer value of 1 or more). The gradient is a gradient of the point of interest i located at the center. In other words, the conventional central difference method is improved so that random noise can be smoothed and removed, and the adjustment can be adjusted. When n = 1 in the formula 1, the formula is the same as the formula 3.
[0014]
FIG. 7 shows the inclination angle θ at each measurement point calculated by the above formulas 1 and 2.
In the process (4), the process computer scans the inclination angle θ at each measurement point from the furnace center toward the furnace wall. Then, a measurement point g whose inclination angle is less than a threshold value, for example, 15 ° is detected. In this case, as shown in FIG. 8, after detecting the measurement point g 2 at which the inclination angle is less than 15 °, if the threshold value is exceeded by further scanning, for example, 25 °, the threshold value 15 is further increased on the furnace wall side. Recognize that there is a measurement point of less than °, and search again for a measurement point with a threshold value of less than 15 °. g 1 indicates the measurement point detected by this.
[0015]
In the process of (5), in FIG. 7, the process computer uses the least square method based on the data of the measurement point g and the measurement points sandwiching the threshold value 15 ° before and after the measurement point g to obtain the linear expression Y = AX + B. The position that coincides with the threshold value of 15 °, that is, G that satisfies X = G and Y = 15 is detected.
[0016]
In the process (6), the process computer calculates a horizontal distance l from G obtained in the process (5) to the furnace wall w.
According to the method comprising the above process, random noise is removed by the smoothing process, and by performing the first-order differential approximation process, the data is further smoothed and the noise can be further removed.
[0017]
In the invention according to claim 2, the process (2) in the invention according to claim 1 is omitted, and measurement is performed by the process (1) instead of the smoothed depth data used in the process (3). Depth data is used.
According to the calculation results of the present inventors, even if the first-order approximation processing by the center difference is directly performed without smoothing the data measured in the process (1), it is almost the same as the invention according to claim 1. We were able to find the terrace length with accuracy.
[0018]
The invention according to claim 3 is the invention according to claim 1 or 2, wherein the process (5) is omitted, and from the measurement point detected in the process (4) that is less than the threshold value to the furnace wall. The horizontal distance is obtained.
As the pitch of the measurement points is made smaller, the inclination angle of the measurement point detected in the process (4) becomes closer to the threshold value and approximated to the position detected in the step (5). Taking FIG. 6 as an example, the smaller the pitch of the measurement points, the closer to the threshold value (15 °), and the measurement points that approximate the threshold value can be detected.
[0019]
【Example】
Example 1
After the coke is charged at the top of the blast furnace furnace, it is moved in the radial direction from the furnace wall to the furnace center using a microwave profile meter, and the depth to the coke layer is measured at 42 locations at 10 cm intervals, and data is input to the process computer. did.
[0020]
In the process computer, the number k of measurement points used when the depth data in the process (2) described above is smoothed as an initial condition is set to 5, and n used in the formulas 1 and 2 of the process (3) is set. 2. The threshold value of the process (4) was set to 15 °, and the horizontal distance 1 of the process (6), which is the terrace length of the coke, was calculated from the measurement data input described above.
Separately from this, using the depth data of 42 places to the surface of the coke layer, the terrace length of the horizontal distance d is read by applying a ruler as shown in FIG. Asked.
[0021]
The above-mentioned calculation and reading of the terrace length is carried out once a day, using the depth data for 73 days obtained by measuring the depth after charging the coke, and counting one day as one time. 73 times. Of these, 69 calculations were possible and the terrace length was calculated 69 times. For the 69 times, the average value and standard deviation of the difference s between the calculated value and the read value were calculated by the process computer. As a result, the average value was obtained as the calculated value being 0.08 m smaller than the read value. The standard deviation was 0.10 m.
Next, when the correlation between the read value of the terrace length and the calculated value was examined, in FIG. 9 where the read value and the calculated value were plotted, the slope of the linear expression was 0.906, and the square R 2 of the correlation coefficient R was 0. 501.
[0022]
Example 2
The same as Example 1 except that the initial condition input to the process computer is n = 3. Based on the measurement data obtained in Example 1, the average value of the difference and the standard are calculated from the calculated value and the read value. When the deviation was determined, the average value was 0.09 m less than the read value, and the standard deviation was 0.11 m.
Further, when the correlation between the read value and the calculated value was obtained in the same manner as in Example 1, the slope of the linear expression was 0.883, R 2 was 0.229, and the terrace length was obtained by the computer. It was 62 times out of 73 times.
[0023]
Example 3
The initial conditions input to the process computer are the same as those in the first embodiment except that the number k of the measurement points of the depth data used in the smoothing process is 0, that is, the smoothing process is not performed and n = 3. Based on the measurement data obtained in Example 1, the average value and the standard deviation of the difference were calculated from the calculated value and the reading value. The average value was 0.10 m less than the reading value, and the standard value The deviation was 0.12 m.
Further, when the correlation between the read value and the calculated value was obtained in the same manner as in Example 1, the slope of the linear expression was 0.872, R 2 was 0.449, and the terrace length could be obtained by a computer. The number was 72 out of 73.
[0024]
Example 4
Based on the measurement data obtained in Example 1, the number k of measurement points for smoothing the initial conditions input to the process computer is set to 5, n = 2, and the threshold value is set to 17.5 °. Then, when the average value and standard deviation of the difference were obtained from the calculated value and the read value, the average value was 0.02 m less than the read value and the standard deviation was 0.967 m.
The correlation between the read value and the calculated value was obtained in the same manner as in Example 1. As shown in FIG. 10, the slope of the linear expression was 0.967, R 2 was 0.666, and the terrace length was 73 times. I was able to find it with a computer.
[0025]
Example 5
The same as Example 4 except that the initial condition input to the process computer is n = 3. Based on the measurement data obtained in Example 1, the average value of the difference is calculated from the calculated value and the read value. When the standard deviation was obtained, the average value was 0.03 m less than the read value, and the standard deviation was 0.11 m.
When the correlation between the read value and the calculated value was obtained in the same manner as in Example 1, the slope of the linear expression was 0.990, R 2 was 0.179, and the terrace length could be obtained by a computer. It was 68 times out of 73 times.
[0026]
Example 6
The same as in Example 4 except that the number k of measurement points for smoothing the initial conditions input to the process computer is set to 0 and n = 3, and based on the measurement data obtained in Example 1. The average value and the standard deviation of the difference were calculated from the calculated value and the read value. The average value was 0.03 m less than the read value and the standard deviation was 0.12 m.
Further, when the correlation between the read value and the calculated value was obtained in the same manner as in Example 1, the slope of the linear expression was 0.952, R 2 was 0.422, and the terrace length was 73 times, and all were obtained by a computer. I was able to.
[0027]
Example 7
The initial conditions input to the process computer are set such that the number of measurement points k for smoothing processing is set to 5, n = 2, and the threshold value is set to 18.5 °. Based on the calculated value and the read value, the average value and the standard deviation of the difference were obtained, and the calculated value and the read value matched the average value, and the standard deviation was 0.09 m.
As shown in FIG. 11, the slope of the linear expression was 0.990, R 2 was 0.620, and the terrace length was 73 times.
[0028]
Example 8
The initial condition input to the process computer is the same as that of Example 7 except that n = 3. Based on the measurement data obtained in Example 1, the average value of the difference is calculated from the calculated value and the read value. When the standard deviation was obtained, the average value was 0.01 m less than the read value, and the standard deviation was 0.11 m.
The correlation between the read value and the calculated value was obtained in the same manner as in Example 1. As shown in FIG. 12, the slope of the linear expression was 0.982, R 2 was 0.243, and the terrace length was calculated by a computer. It was possible to obtain 70 out of 73 times.
[0029]
Example 9
The same as in Example 7 except that the number k of measurement points for smoothing the initial conditions input to the process computer is set to 0 and n = 3, and based on the measurement data obtained in Example 1. Then, when the average value and the standard deviation of the difference were obtained from the calculated value and the read value, the calculated value and the read value matched the average value, and the standard deviation was 0.12.
The correlation between the read value and the calculated value was obtained in the same manner as in Example 1. As shown in FIG. 13, the slope of the linear expression was 0.987, R 2 was 0.406, and the terrace length was 73 times. Both could be obtained by computer.
The above results are shown in Table 1 below.
[0030]
[Table 1]
Figure 0004675523
In the table, a minus (−) sign attached to a numerical value in a column in which the calculated value—the average value of the read values is written indicates that the measured value is smaller than the read value.
[0031]
As can be seen from Table 1, the initial conditions k = 5 and n = 2 input to the process computer have the smallest difference between the calculated value and the read value, the variation is small, and the threshold value is 15 ° → The better the result is from 17.5 ° → 18.5 °, the best result is obtained in Example 7, and the smoother k = 5 and n = 3, the more cases where calculation becomes impossible. In addition, as in Examples 3, 6, and 9, it was found that even when the smoothing process of k = 0 is not performed, the difference between the calculated value and the read value is small and the variation can be reduced.
[0032]
Note that the threshold value is set at the point where the inclination angle of the straight line b region in the inclined portion shown in FIG. 1 changes from the inclination angle (zero) of the straight line a region in the flat portion near the furnace wall, that is, the intersection point. It is preferable to adopt the inclination angle c.
That is, in the case of ore, since the inclination angle of the inclined portion b is around 30 °, 15 ° which is half of the threshold value is adopted, and in the case of coke where the inclined angle of the inclined portion b is about 37 °. In this case, it is preferable to adopt 18.5 ° as the threshold value. If the threshold value is properly used depending on the deposition material, the calculation accuracy of the terrace length is further improved.
[0033]
【The invention's effect】
According to the first aspect of the present invention, the depth data obtained by the measurement by the sounding device is smoothed, and the primary approximation process based on the center difference is performed, so that noise is removed and the terrace length is accurately obtained. Can do.
[0034]
According to the invention according to claim 2, compared to the invention according to claim 1, the process is simplified and the number of cases where calculation is impossible is reduced because the process 2 for smoothing is not performed, and the terrace length is also claimed. It can be obtained with the same accuracy as the invention according to 1.
[0035]
According to the invention according to claim 3, compared with the invention according to claim 1, since there is no process (5) for smoothing, the process is simplified, and the terrace length is reduced by reducing the pitch of the measurement points. This can be determined with the same accuracy as the invention according to claim 1.
[Brief description of the drawings]
FIG. 1 is a diagram showing a case where a terrace length is artificially obtained.
FIG. 2 is an explanatory diagram for calculating a gradient.
FIG. 3 is a diagram showing a conventional method in a case where a terrace length is calculated using a computer.
FIG. 4 is a graph of depth data.
FIG. 5 is a diagram showing a smoothing processing method.
FIG. 6 is a diagram illustrating a method for obtaining an inclination angle based on a center difference.
FIG. 7 is a view showing a method for obtaining a terrace length.
FIG. 8 is a graph when there are two inclination angles less than a threshold value.
9 is a graph showing the correlation between the read value and the calculated value in Example 1. FIG.
10 is a graph showing the correlation between the read value and the calculated value in Example 4. FIG.
11 is a graph showing the correlation between the read value and the calculated value in Example 7. FIG.
12 is a graph showing the correlation between the read value and the calculated value in Example 8. FIG.
13 is a graph showing the correlation between the read value and the calculated value in Example 9. FIG.

Claims (3)

高炉炉頂部の原料堆積層の表面形状をもとに炉壁近傍のテラス長さをコンピュータを用いて演算する方法であって、
(1)測深装置を用いて原料堆積層表面までの深度を炉半径方向に沿って任意の間隔ごとに測定するプロセスと、
(2)上記(1)の工程で測定した各測定点のうち、炉壁側と炉中心側の測定点を除く各測定点において、その前後の複数の測定点の深度データを用いて多項式適用による平滑化データ処理を行うプロセスと、
(3)測定点iから前後にn点離れた測定点i−nとi+nにおける、上記(2)の工程で平滑化処理された深度データから中心差分による一次微分近似処理を行い、測定点iにおける堆積層表面の傾斜角を算出するプロセスと、
(4)上記(3)の工程で算出された各測定点の傾斜角に関するデータを炉半径方向にスキャニングし、傾斜角がしきい値未満となる測定点を検出するプロセスと、
(5)上記(4)の工程でしきい値未満として検出された測定点と、その近傍の測定点の上記傾斜角に関するデータより最小二乗法を用いて一次式を求め、この一次式としきい値が一致する位置を検出するプロセスと、
(6)上記(5)の工程で検出した、しきい値と一致する位置より炉壁までの水平距離を求めるプロセス
よりなることを特徴とするテラス長さ演算方法。
A method of calculating the terrace length near the furnace wall using a computer based on the surface shape of the raw material deposition layer at the top of the blast furnace furnace,
(1) a process of measuring the depth to the surface of the raw material deposition layer at an arbitrary interval along the furnace radial direction using a depth measuring device;
(2) Among the measurement points measured in the above step (1), at each measurement point excluding the measurement points on the furnace wall side and the furnace center side, a polynomial is applied using depth data of a plurality of measurement points before and after the measurement point. Process for smoothing data processing by
(3) Perform a first-order differential approximation process using the center difference from the depth data smoothed in the step (2) at the measurement points i−n and i + n that are n points away from the measurement point i. A process of calculating the tilt angle of the surface of the deposited layer in
(4) A process of scanning data regarding the inclination angle of each measurement point calculated in the step (3) in the furnace radial direction to detect a measurement point at which the inclination angle is less than a threshold value;
(5) A linear equation is obtained by using the least square method from the measurement point detected as less than the threshold value in the step (4) and data on the inclination angle of the measurement point in the vicinity thereof, and this linear equation is set as the threshold. The process of finding the location where the values match,
(6) A terrace length calculation method comprising a process of obtaining a horizontal distance from the position coincident with the threshold detected in the step (5) to the furnace wall.
上記(2)のプロセスが省かれ、(3)のプロセスにおいて用いられる平滑化処理された深度データの代わりに(1)のプロセスで測定した深度データが用いられることを特徴とする請求項1記載のテラス長さ演算方法。2. The process (2) is omitted, and the depth data measured in the process (1) is used instead of the smoothed depth data used in the process (3). Terrace length calculation method. 上記(5)のプロセスを省き、上記(4)のプロセスにおいて検出された、しきい値未満とよりなる測定点より炉壁までの水平距離を求めることを特徴とする請求項1又は2記載のテラス長さ演算方法。3. The horizontal distance from the measurement point detected in the process (4) to a value less than a threshold value to the furnace wall is obtained by omitting the process (5). 4. Terrace length calculation method.
JP2001268153A 2001-09-05 2001-09-05 Method for calculating the terrace length of the raw material deposition layer at the top of the blast furnace furnace Expired - Fee Related JP4675523B2 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP2001268153A JP4675523B2 (en) 2001-09-05 2001-09-05 Method for calculating the terrace length of the raw material deposition layer at the top of the blast furnace furnace

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP2001268153A JP4675523B2 (en) 2001-09-05 2001-09-05 Method for calculating the terrace length of the raw material deposition layer at the top of the blast furnace furnace

Publications (2)

Publication Number Publication Date
JP2003073716A JP2003073716A (en) 2003-03-12
JP4675523B2 true JP4675523B2 (en) 2011-04-27

Family

ID=19094181

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2001268153A Expired - Fee Related JP4675523B2 (en) 2001-09-05 2001-09-05 Method for calculating the terrace length of the raw material deposition layer at the top of the blast furnace furnace

Country Status (1)

Country Link
JP (1) JP4675523B2 (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP5306665B2 (en) * 2008-02-05 2013-10-02 株式会社神戸製鋼所 Blast furnace charging abnormality monitoring method and monitoring apparatus using this method
JP5391021B2 (en) * 2008-11-14 2014-01-15 株式会社神戸製鋼所 Method of measuring terrace length of blast furnace and measuring device of terrace length of blast furnace
CN108801117B (en) * 2018-07-31 2024-02-23 南宁海关技术中心 Depth scale for single explosive package test and stacking test

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS5950102A (en) * 1982-09-14 1984-03-23 Kawasaki Steel Corp Operating method of blast furnace
JPH0694367A (en) * 1991-05-28 1994-04-05 Nippon Steel Corp Estimating method for depositing shape of material to be chaged in vertical furnace
JPH0860212A (en) * 1994-08-12 1996-03-05 Nkk Corp Method for supporting distribution control of charged material in bell-less blast furnace

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPS5950102A (en) * 1982-09-14 1984-03-23 Kawasaki Steel Corp Operating method of blast furnace
JPH0694367A (en) * 1991-05-28 1994-04-05 Nippon Steel Corp Estimating method for depositing shape of material to be chaged in vertical furnace
JPH0860212A (en) * 1994-08-12 1996-03-05 Nkk Corp Method for supporting distribution control of charged material in bell-less blast furnace

Also Published As

Publication number Publication date
JP2003073716A (en) 2003-03-12

Similar Documents

Publication Publication Date Title
CN110906875B (en) Visual processing method for aperture measurement
CN109029882A (en) Method for improving bridge deflection test precision based on inclinometer
US4025192A (en) Optical measuring method
CN112381781B (en) Glue road defect detection method based on 3D point cloud
Liu et al. Recognition and calibration of rail profile under affine-distortion-based point set mapping
CN110424945B (en) Shale oil horizontal well deployment method and device
CN105043390B (en) Gravitational field interpolation method based on universal Kriging
JP4675523B2 (en) Method for calculating the terrace length of the raw material deposition layer at the top of the blast furnace furnace
JP3126288B2 (en) Method of measuring rail rail shape
CN115540823B (en) Method for measuring and controlling vertical degree of variable-section ultrahigh pier
WO2019053773A1 (en) State determining device, state determining method, and computer readable recording medium
US20150219788A1 (en) Method of processing gravity gradient data
RU2466768C2 (en) Method of defining fluid boundary level
CN114355287B (en) Ultra-short baseline underwater sound distance measurement method and system
CN110308302B (en) Near-wall surface flow velocity measuring method and device
Jermak et al. Air gauge characteristics linearity improvement
JPH07225843A (en) Method for measuring center position of circle
CN104006781B (en) The computational methods of surface normal vector certainty of measurement
CN114065120A (en) Method for determining cycle threshold of linear-sigmoidal growth curve for PCR
CN109856681B (en) Process analysis method for describing morphological change of water channel along direction of material source
CN112733072A (en) Inverse distance square weighted spatial interpolation method
CN112113517A (en) Method for detecting flatness of sliding plate brick
CN109507730A (en) A kind of method of improved determining plane of refraction normal depth
JP4019039B2 (en) Capacitance type thickness measurement method
JP2881359B2 (en) Trace display method from flow field data

Legal Events

Date Code Title Description
A621 Written request for application examination

Free format text: JAPANESE INTERMEDIATE CODE: A621

Effective date: 20080808

A977 Report on retrieval

Free format text: JAPANESE INTERMEDIATE CODE: A971007

Effective date: 20110118

TRDD Decision of grant or rejection written
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20110125

A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

A61 First payment of annual fees (during grant procedure)

Free format text: JAPANESE INTERMEDIATE CODE: A61

Effective date: 20110126

FPAY Renewal fee payment (event date is renewal date of database)

Free format text: PAYMENT UNTIL: 20140204

Year of fee payment: 3

R150 Certificate of patent or registration of utility model

Free format text: JAPANESE INTERMEDIATE CODE: R150

LAPS Cancellation because of no payment of annual fees