JP3984392B2

JP3984392B2 - Quality judgment method for ladle replacement part of continuous cast slab

Info

Publication number: JP3984392B2
Application number: JP10349599A
Authority: JP
Inventors: 健夫井本; 弘昭飯星; 崇博片井
Original assignee: Nippon Steel Corp
Current assignee: Nippon Steel Corp
Priority date: 1999-04-12
Filing date: 1999-04-12
Publication date: 2007-10-03
Anticipated expiration: 2019-04-12
Also published as: JP2000292418A

Description

【０００１】
【発明の属する技術分野】
本発明は、連続鋳造における介在物起因の品質判定方法、特に多連鋳造時の取鍋交換部位における品質判定方法に関する。
【０００２】
【従来の技術】
連続鋳造工程では、単位鋳造量あたりのタンディッシュ（以下ＴＤと称す）の整備費や耐火物コストの低減を目的として、複数の鍋を連続して鋳造する多連鋳操業が指向されてきた。この多連鋳操業において取鍋を交換する際には、前鍋から注入終了時にＴＤ内に流入するスラグや、後鍋注入開始時のたたき込み等によって鋳片清浄度が悪化する傾向が見られ、取鍋交換部位の鋳片の清浄度はそれ以外の定常部位と比べてばらつきが極めて大きい。
【０００３】
そのため、自動車用鋼板や食缶用鋼板などのうち、特に加工時の変形率が大きく介在物に起因する割れが発生しやすいハイグレード材は取鍋交換部位に充当せず、取鍋交換部位の鋳片をあらかじめローグレード材に充当することを前提として製造されたり、鍋交換部の品質ばらつきも含めて品質を確保するため、前鍋終了時に流出するスラグの酸化度を予め低減するスラグ改質基準を設けるなど、生産性、コストの面で取鍋交換部の品質管理には大きな課題があった。
【０００４】
鋳片の介在物量を把握するために、例えば特開平２−１１２５７号公報に記載のように、浸漬ノズル内の圧力や湯面変動などの操業因子から間接的に品質を判定する方法があるが、取鍋からＴＤ、モールドまでの鍋交換部の介在物量を精度良く把握するためには、清浄度の直接分析値を含めた判定が必要である。
なお、清浄度の直接分析の方法としては、一般的なスライム法や検鏡法、トータル酸素分析の他、特開平１−７０１３４号公報に示されている電子ビーム法、特開平８−８２８７２９号公報に示されているコールドクルーシブル法、特開平１０−９６７２１号公報に用いられている高周波超音波法などもこれまで開発されてきた。
【０００５】
【発明が解決しようとする課題】
前述の清浄度直接分析法は採取サンプル分析可能サイズまで加工し清浄度の分析を行う。一般的に前述の清浄度直接分析方法では、数グラムから１ｋｇ程度のサンプルを分析する。製品の介在物を測定するためには、製鋼工程で溶鋼を鋳造後の凝固した鋳片から鋳片サンプルを採取し、分析を行うことは精度は高いが、鋳片サンプルの場合は凝固した鋳片から、断面サイズの数百ｋｇのサンプルを切り出し、前記のような小サイズまで加工する必要があるために、歩留りロスおよび、加工処理コストが多大になる。
【０００６】
一方、ＴＤ内の溶鋼段階で採取した溶鋼サンプルの場合は、サンプラーによって分析に必要な一般に１ｋｇ程度の溶鋼を汲み上げ、それを凝固させればよく、鋳片サンプルに比べ歩留りロスは軽微で加工も簡単である。しかし、該溶鋼サンプルの場合、モールド内で溶鋼流動変化により鋼中の介在物が移動するため、溶鋼サンプルの分析結果による鋳片の清浄度判定精度は、溶鋼凝固後に鋳片から採取した前記鋳片サンプル分析の判定比べ、精度は劣る可能性がある。
【０００７】
本発明では、低コストだが精度の劣るＴＤ溶鋼サンプル分析値に基づいた一次判定と、高コストだが高精度鋳片サンプル分析の二次判定を組み合わせることによって、従来方法に比べ低コストかつ高精度で鍋交換部の介在物品質判定方法を提供することを課題とする。
【０００８】
【課題を解決するための手段】
上記課題を解決するための手段である本発明は、鋼を連続鋳造機で多連鋳造する際の取鍋交換部位の鋳片品質判定方法において、取鍋交換部における取鍋溶鋼のタンディッシュ注入開始から、溶鋼注入積算量がタンディッシュ最大容量の１．５倍に達するまでの間に、タンディッシュ内より溶鋼サンプルを採取し、該溶鋼サンプルの品質分析結果を基に前記取鍋交換部位の鋳片品質を一次判定し、該品質判定結果が予め定められた判定精度の信頼区間の範囲外の場合には、前記鍋交換部における取鍋溶鋼のタンディッシュ注入開始から、溶鋼注入積算量がタンディッシュ最大容積の１．５倍に達するまでの間に、タンディッシュに注入した溶鋼の鋳造鋳片から採取した鋳片サンプルを基に、前記鍋交換部の鋳片品質を二次判定することを特徴とする連鋳鋳片の取鍋交換部位における品質判定方法である。
【０００９】
【発明の実施の形態】
鋼の連続鋳造の場合、取鍋交換点では前記したように前鍋からの溶鋼注入終了時にＴＤ内に流入するスラグや、次鍋注入開始時のたたき込み等によって清浄度が悪化し、その鍋交換部毎のバラツキも大きい。従って、数個の溶鋼あるいは鋳片サンプルにより鍋交換部全体の品質を判定する場合には、その採取サンプルの代表性が課題となる。
そこで清浄度判定用の溶鋼及び鋳片サンプルの分析結果に関して、次鍋交換点付近での採取タイミングにより当該溶鋼の鋳造鋳片を圧延加工後の製品品質の相関がどのように変化するかを調査したところ、前記サンプルの採取タイミングにより該サンプル分析結果は鍋交換部の製品品質と相関があることを突き止めた。
【００１０】
図２は鍋交換付近の溶鋼サンプル及び鋳片サンプルの採取タイミングによる各サンプルの介在物分析結果と、当該鍋交換付近の鋳片の圧延加工した製品品質との相関係数の関係を示す。
図２において横軸は鍋交換時の次鍋のＴＤ注入開始点を基準とし、取鍋からＴＤへの溶鋼注入量積算値をＴＤ最大容量で規格化した値を表し、縦軸は当該鋳片を圧延後の冷延板品質として内部欠陥指数（定義は後述）を表す。
【００１１】
図２より次鍋からＴＤへの取鍋注入開始点（溶鋼注入量積算値＝０）から溶鋼注入量積算値がＴＤ最大容量の１．５倍になるまでの間に採取したＴＤ溶鋼サンプル、および該溶鋼の鋳造鋳片の鋳片サンプル分析結果が当該製品品質との相関係数が高く、鍋交換部位における製品品質の代表性を得るのに適していることが発明者らの研究から判明した。
【００１２】
なお、溶鋼サンプル採取方法位置は、ＴＤ入り口から出口までの清浄度の変化に起因するばらつきを最小限にする意味からＴＤ出口（モールド注入口）近傍のＴＤ溶鋼表面から３００ｍｍ以上深い位置が望ましい。
【００１３】
また、介在物分析方法としては、検鏡法やスライム法などでも可能であるが迅速な分析方法としては、トータル酸素法やコールドクルーシブル法、電子ビーム法、高周波超音波法などが挙げられる。
【００１４】
次に、本発明による品質判定方法の一例を図１に示すフローチャートを基に説明する。図１に示すように一次判定では溶鋼サンプルを基に鍋交換部の鋳片品質を判定する。一次判定用の溶鋼サンプルは前記のように鍋交換部の次鍋のＴＤへの注入開始から溶鋼注入量積算値がＴＤ最大容量の１．５倍になるまでにＴＤから一次判定用の溶鋼サンプルを採取する。
【００１５】
一次判定では溶鋼サンプルの清浄度分析結果単独での品質判定も可能であるが、さらに高精度の品質判定には、その他の清浄度考慮因子として、例えば、平均モールド湯面変動量、平均モールド長片両端付近の湯面レベル差、鍋交換点からの鋳造長、前鍋の取鍋残湯量、鋳造中の最大湯面レベル変動量、ＴＤスライディングノズル変動量、鍋交換時のＴＤ最小溶鋼量、平均鋳造速度、モールド幅、モールド厚み、二次精錬後の取鍋スラグ厚み、二次精錬後のスラグ中トータルＦｅも考慮することが望ましい。
【００１６】
一次判定では、まず鍋交換部の溶鋼サンプル分析及び必要に応じ前記清浄度考慮因子を基に当該鋳片の圧延コイルの実績内部欠陥評点との相関係数を重回帰計算で求め、その結果を基に前記圧延コイル以外の他の圧延コイルの内部欠陥評点を予測する。
なお、内部欠陥評点とは冷延板段階における漏洩磁束検査の指標で、冷延板の重量に対してカウントされた介在物個数と比例した数値であり、例えばハイグレード材の場合には内部欠陥評点１．０以下の鋳片を合格材として判断している。
【００１７】
このようにして鍋交換部の溶鋼サンプルを基に予測した圧延コイルの内部欠陥評点と実績内部欠陥評点との比較の一例を図３に示す。例えばハイグレード材では前記のように内部欠陥評点基準値が１．０以下を合格材と判定するが、予測内部欠陥評点は実績内部欠陥評点に対してばらつきがあり、一義的に予測内部欠陥評点を基に対象鍋交換部の製品の合否判定した場合には実績内部欠陥評点が１．０未満であるにも関わらず、不合格と判定してしまう可能性があり、歩留が低下する恐れがある。
そこで、一次判定においては、溶鋼サンプル分析値による予測内部欠陥評点と実績内部欠陥評点の関係から、統計的手法を用いて品質合否判定を行う。
【００１８】
一般に回帰計算した場合の予測値の分布は真の値に対して正規分布と見なすことができ、正規分布曲線と標準偏差σから見た発生頻度は図４に示すようなの模式図となる。
図４に示すように予測内部欠陥評点を基にした品質判定では、真の値Ｘ（ここでは内部欠陥評点合格基準）、標準偏差σ及び正規分布曲線の信頼区間決定係数Ｋによって決まる信頼区間からその判定精度決まり、その信頼区間を基に下記（１）、（２）式を満足するように内部評点合格基準Ｘを補正して予測内部欠陥評点から鍋交換部の品質判定すれば良い。
▲１▼不合格材：予測内部欠陥評点＜Ｘ−Ｋ・σ ・・・・（１）
▲２▼合格材：予測内部欠陥評点＞Ｘ＋Ｋ・σ ・・・・（２）
【００１９】
例えば、前記のような内部欠陥評点合否基準＝１．０であるハイグレード材の場合、図４より信頼区間を約９８％（判定精度が約９８％）とする場合には、上記（１）式及び（２）式において内部欠陥評点合否基準＝１．０、信頼区間決定係数Ｋ＝２とする必要があり、予測内部欠陥評点が［１．０−２σ］未満の鋳片を不合格と判定し、また予測内部欠陥評点が［１．０＋２σ］超の鋳片を合格と判定すれば良い。
【００２０】
以上のように一次判定として溶鋼サンプル分析値を基に予め定められた判定精度（信頼区間）で品質の合否できなかた区間（前記一次判定の信頼区間９８％の場合には予測内部欠陥評点が、［１．０−２σ］〜［１．０＋２σ］の区間）の鋳片、即ち一次判定で合否判定ができなかった鍋交換部の鋳片は、分析自体の精度が高い鋳片サンプル分析による二次判定を行う。
【００２１】
二次判定用の鋳片サンプルは、図２に示すように鍋交換部において次鍋注入開始からＴＤへの溶鋼注入量積算値がＴＤ最大容量の１．５倍になるまでの間に注入された溶鋼の鋳造した鋳片から採取する。上記範囲内の溶鋼の鋳造した鋳片位置を特定するためには、通常、ＴＤへの溶鋼注入量積算値がＴＤ最大容量の１．５倍になるまでにモールド内のメニスカスを通過した鋳片部位を鋳造後に切り出すことで十分であるが、流動と凝固を考慮した各種数値シミュレーションや、タンディッシュ内へのトレーサー添加実験などにより、さらに鋳片サンプルの採取位置を適宜選択することが出来る。
【００２２】
二次判定についても一次判定と同様、分析値に基づく予測成績と圧延以降の製品成績との相関関係から、品質合否判定基準（ここでは内部欠陥評点１．０）の判定精度の信頼性区間を基に合否判定を実施する。
二次判定では内部欠陥評点予測値が前記（２）式を満足する場合を合格材と判定し、（２）式を満足しない場合はすべて不合格材と判定する。二次判定の鋳片サンプルは図２に示すように、溶鋼サンプル分析に比べ製品品質結果と相関性が高いことから、一次判定で合否判定できなかった予測内部欠陥評点の区間（例えば前記信頼性区間が約９８％とした場合の一次判定では［１．０−２σ］〜［１．０＋２σ］）のものでも合否判定ができる。
【００２３】
図１に示す品質判定フローでは一次判定、二次判定共に信頼区間決定係数Ｋ＝２としたが、Ｋの値は必要に応じ当業者が適宜判断できる。また、統計手法として重回帰分析法を用いたが、各種要因と欠陥発生率の関係が線形と見なせない場合はニューラルネットワーク法等を用いることで判定を行うことも可能である。
また、真の値に対する予測値のばらつきが正規分布に従わない特殊なケースの場合などはχ二乗分布や放物線など適宜フィッティングを行いその分布に見合った信頼区間を決定できる。
【００２４】
また、一次判定、二次判定に必要な取鍋交換点代表サンプルは、前記のように多連鋳造の２鍋以降の鍋注入開始点からＴＤへの溶鋼注入量がＴＤの最大容積の１．５倍までであれば１サンプルで判定可能である。しかし、試料不良による判定ミスや鋳造方向による清浄度の変化による誤差を低減させるためには複数のサンプルを採取して平均値や中央値を判定値とするなどは適宜行える。
また、判定指標は、製品成績と清浄度の分析データの単相関に基づいたものでも良いが、鍋交換点からの鋳片位置鋳造長、サンプルの採取タイミング、サンプル採取位置前後の湯面レベル変動などサンプル分析値以外の製品成績と相関関係が見られる操業指標も考慮した判定が望ましい。
【００２５】
【実施例】
実施例ではＴＤ最大容量４０ｔ、ヒートサイズ３４０ｔ、１ストランド連鋳機によってアルミキルド鋼を鋳造し、得られた鋳片を圧延により冷延コイルを製造した。鍋交換部においてはＴＤ内溶鋼から溶鋼サンプル、また当該溶鋼を鋳造した鋳片から鋳片サンプルをそれぞれ採取し、電子ビーム法により分析した。また鍋冷延コイルは漏洩磁束検査装置により内部欠陥評点を検査した。
表１に同一鍋交換部の鋳片１２１本について各方法（実施例１、従来例１〜２、比較例１）による品質判定結果を示す。なお本鋳片の合否判定基準は内部欠陥評点が１．０である。
【００２６】
表１において、実施例１は本発明の判定結果を表しており、前記図１に示す判定フローに従い、一次及び二次判定の信頼区間を９８％として鍋交換部の全鋳片から採取した溶鋼サンプル１２１本を基に一次判定し、一次判定で合否判定できなかった（信頼区間が９８％を外れた）５１本を二次判定で再度合否判定を行った。
なお溶鋼及び鋳片サンプルは本発明の規定内、即ち鍋注入開始点から注入積算量が６０ｔの範囲内からそれぞれ１サンプルずつ採取した。また従来例１及び２はそれぞれＴＤ溶鋼サンプル分析及び鋳片サンプル分析それぞれ単独で鍋交換部の鋳片の品質判定を行った。さらに、比較例１は溶鋼及び鋳片サンプルの採取を本発明の規定外位置から採取した場合の品質判定結果を示す。
【００２７】
【表１】

【００２８】
表１に示すように、本発明の実施例１では合格判定精度（採取サンプル分析結果を基に品質合格と判定した鋳片中、該鋳片の圧延後の製品品質実績も合格であった割合）が９８％であり、且つ総合合格率（鍋交換部の全１２１採取サンプル中、該鋳片の圧延後の製品品質実績も合格であった割合）が７８％であるのに対し、従来例１及び比較例１では合格判定精度は本発明実施例１と同程度であったが、総合合格率がそれぞれ４５％、２６％と低く、さらに残りの不合格と判定した鍋交換部の鋳片の中には、当該鋳片の圧延後の実績製品は合格基準をクリアしたの品質であるのに係わらず不合格と判定し、鋳片歩留が低下してしまった。また従来例２では、総合合格率は本発明の実施例１並の割合であるが、鋳片サンプル分析数は本発明に比べ２．４倍にも達し、全体の分析コストも本発明の実施例１に比べ２倍要した。
【００２９】
【発明の効果】
本発明により、以上のように本発明の場合には、従来並の高合格率で且つ低コストで鍋交換点の鋳片の品質判定が可能になった。
【図面の簡単な説明】
【図１】本発明例の品質判定フローチャートを示した図
【図２】サンプル採取タイミングと内部欠陥評点との相関係数の関係を示した図
【図３】ＴＤの溶鋼サンプル分析値に基づく冷延板成績の予測値と実績値を示した図
【図４】真の評点Ｘから見た予測内部欠陥評点のばらつきを示した図[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a quality determination method caused by inclusions in continuous casting, and more particularly, to a quality determination method in a ladle replacement site during multiple casting.
[0002]
[Prior art]
In the continuous casting process, for the purpose of reducing tundish (hereinafter referred to as TD) maintenance cost and refractory cost per unit casting amount, multiple casting operations for continuously casting a plurality of pans have been directed. When replacing the ladle in this multiple casting operation, there is a tendency for the slab cleanliness to deteriorate due to slag flowing into the TD from the front pan at the end of pouring, or slag at the start of the rear pan pouring, etc. The cleanliness of the slab at the ladle replacement site varies greatly compared to the other steady sites.
[0003]
Therefore, among steel plates for automobiles and steel plates for food cans, high-grade materials that have a large deformation rate during processing and are prone to cracking due to inclusions are not applied to the ladle replacement site. Slag modification that reduces the degree of oxidation of the slag flowing out at the end of the previous pan in order to ensure quality, including pre-applying slabs to low-grade materials, and to ensure quality, including quality variations in the pan replacement section There was a big problem in the quality control of the ladle exchange part in terms of productivity and cost, such as setting standards.
[0004]
In order to grasp the amount of inclusions in the slab, there is a method of indirectly determining the quality from operating factors such as pressure in the immersion nozzle and fluctuations in the molten metal surface, as described in JP-A-2-11257, for example. In order to accurately grasp the amount of inclusions in the pan changing part from the ladle to the TD and the mold, it is necessary to make a determination including the direct analysis value of the cleanliness.
As a method for direct analysis of cleanliness, in addition to a general slime method, speculum method, and total oxygen analysis, an electron beam method disclosed in JP-A-1-70134, JP-A-8-828729 The cold crucible method disclosed in the publication and the high-frequency ultrasonic method used in JP-A-10-96721 have been developed.
[0005]
[Problems to be solved by the invention]
In the above-described direct cleanliness analysis method, the collected sample is processed to a size that can be analyzed, and the cleanliness is analyzed. Generally, in the above-described direct cleanliness analysis method, a sample of several grams to 1 kg is analyzed. In order to measure product inclusions, it is highly accurate to collect and analyze a slab sample from a solidified slab after casting molten steel in the steelmaking process. Since it is necessary to cut out a sample having a cross-sectional size of several hundred kg from the piece and process it to the small size as described above, the yield loss and the processing cost are increased.
[0006]
On the other hand, in the case of a molten steel sample taken at the molten steel stage in TD, it is only necessary to pump up about 1 kg of molten steel required for analysis by a sampler and solidify it. Simple. However, in the case of the molten steel sample, inclusions in the steel move due to the change in the molten steel flow in the mold. Therefore, the accuracy of the slab cleanliness judgment based on the analysis result of the molten steel sample is the above-mentioned The accuracy may be inferior to that of single sample analysis.
[0007]
In the present invention, by combining the primary determination based on the TD molten steel sample analysis value, which is low-cost but inaccurate, and the secondary determination of the high-cost but high-accuracy slab sample analysis, it is less expensive and more accurate than the conventional method. It is an object of the present invention to provide a method for determining the quality of inclusions in a pot changing portion.
[0008]
[Means for Solving the Problems]
The present invention, which is a means for solving the above-mentioned problems, is a method for determining the quality of a slab at a ladle replacement site when steel is continuously cast by a continuous casting machine. During the period from the start until the cumulative amount of molten steel injection reaches 1.5 times the maximum capacity of the tundish, a molten steel sample is taken from within the tundish, and based on the result of quality analysis of the molten steel sample, When the quality of the slab is primarily determined and the quality determination result is out of the range of the confidence interval of the predetermined determination accuracy, from the start of the tundish injection of the ladle molten steel in the ladle replacement part, the molten steel injection integrated amount is Based on the slab sample taken from the cast slab of molten steel injected into the tundish until it reaches 1.5 times the maximum volume of the tundish, the slab quality of the pan replacement part is secondarily determined. Features A quality evaluation method in the ladle exchange sites continuous casting slab to be.
[0009]
DETAILED DESCRIPTION OF THE INVENTION
In the case of continuous casting of steel, at the ladle replacement point, as described above, the cleanliness deteriorates due to slag flowing into the TD at the end of pouring of molten steel from the previous pan, or knocking at the start of the next pan injection. There is also a large variation between the replacement parts. Therefore, when judging the quality of the whole pan exchange part with several molten steel or slab samples, the representativeness of the collected samples becomes an issue.
Therefore, regarding the analysis results of molten steel and slab samples for cleanliness determination, we investigated how the correlation of product quality after rolling of the cast slab of the molten steel changes depending on the sampling timing near the next pan replacement point. As a result, it was found out that the sample analysis result had a correlation with the product quality of the pan changing part by the sampling timing of the sample.
[0010]
FIG. 2 shows the relationship between the correlation analysis results of the inclusion analysis results of each sample according to the collection timing of the molten steel sample and slab sample near the pan change and the product quality of the slab rolled near the pan change.
In FIG. 2, the horizontal axis represents the TD injection start point of the next pan at the time of the pan replacement, and the integrated value of the molten steel injection amount from the ladle to the TD is normalized by the TD maximum capacity, and the vertical axis represents the slab. Represents the internal defect index (defined later) as the cold-rolled sheet quality after rolling.
[0011]
From FIG. 2 TD molten steel sample taken from the ladle pouring start point from the next ladle to the TD (molten steel injection amount integrated value = 0) until the molten steel injection amount integrated value becomes 1.5 times the TD maximum capacity, From the research conducted by the inventors, the slab sample analysis result of the cast slab of molten steel has a high correlation coefficient with the product quality, and is suitable for obtaining representativeness of product quality at the pan replacement site. did.
[0012]
The position of the molten steel sample collection method is desirably a position 300 mm or more deeper from the surface of the TD molten steel near the TD outlet (mold inlet) in order to minimize the variation caused by the change in cleanliness from the TD inlet to the outlet.
[0013]
The inclusion analysis method can be a speculum method or a slime method, but examples of a rapid analysis method include a total oxygen method, a cold crucible method, an electron beam method, and a high-frequency ultrasonic method.
[0014]
Next, an example of the quality determination method according to the present invention will be described based on the flowchart shown in FIG. As shown in FIG. 1, in the primary determination, the slab quality of the pan replacement part is determined based on the molten steel sample. As described above, the molten steel sample for primary judgment is from the TD until the integrated value of molten steel injection reaches 1.5 times the maximum capacity of the TD from the start of pouring into the TD of the next pan of the pan changing part. Collect.
[0015]
In the primary judgment, it is possible to judge the quality of the molten steel sample alone, but for more accurate quality judgment, other factors that consider cleanliness include, for example, average mold level fluctuation, average mold length, etc. Difference in molten metal level near both ends, casting length from the pot changing point, amount of remaining hot water in the previous pan, maximum molten metal level fluctuation during casting, TD sliding nozzle fluctuation, TD minimum molten steel quantity when changing the pot, It is desirable to consider the average casting speed, mold width, mold thickness, ladle slag thickness after secondary refining, and total Fe in the slag after secondary refining.
[0016]
In the primary determination, first, a correlation coefficient with the actual internal defect score of the rolled coil of the slab is obtained by multiple regression calculation based on the analysis of the molten steel sample in the pan replacement part and, as necessary, the cleanliness consideration factor, and the result is calculated. Based on this, the internal defect score of other rolled coils other than the rolled coil is predicted.
The internal defect score is an index of leakage magnetic flux inspection at the cold-rolled sheet stage, and is a value proportional to the number of inclusions counted against the weight of the cold-rolled sheet. A slab having a rating of 1.0 or less is judged as an acceptable material.
[0017]
FIG. 3 shows an example of the comparison between the internal defect score of the rolled coil and the actual internal defect score predicted based on the molten steel sample of the pan replacement part in this way. For example, in the high grade material, as described above, the internal defect score standard value is determined to be 1.0 or less as the acceptable material, but the predicted internal defect score varies with the actual internal defect score and is uniquely predicted internal defect score. When the pass / fail judgment of the product of the target pan replacement part is made based on this, there is a possibility that it may be judged as a failure even though the actual internal defect score is less than 1.0, and the yield may be lowered. There is.
Therefore, in the primary determination, the quality pass / fail determination is performed using a statistical method from the relationship between the predicted internal defect score and the actual internal defect score based on the analysis value of the molten steel sample.
[0018]
In general, the distribution of predicted values in the case of regression calculation can be regarded as a normal distribution with respect to a true value, and the occurrence frequency viewed from the normal distribution curve and the standard deviation σ is a schematic diagram as shown in FIG.
As shown in FIG. 4, in the quality judgment based on the predicted internal defect score, from the confidence interval determined by the true value X (in this case, the internal defect score pass criterion), the standard deviation σ, and the confidence interval determination coefficient K of the normal distribution curve. The determination accuracy is determined, and based on the confidence interval, the internal score passing standard X is corrected so as to satisfy the following formulas (1) and (2), and the quality of the pan replacement part may be determined from the predicted internal defect score.
(1) Rejected material: Predicted internal defect score <XK · σ (1)
(2) Accepted material: Predicted internal defect score> X + K · σ (2)
[0019]
For example, in the case of a high-grade material having an internal defect score acceptance standard of 1.0 as described above, when the confidence interval is about 98% (judgment accuracy is about 98%) from FIG. In the formula and (2), it is necessary to set the internal defect score acceptance criteria = 1.0 and the confidence interval determination coefficient K = 2, and reject the slab whose predicted internal defect score is less than [1.0-2σ]. It may be determined, and a slab whose predicted internal defect score exceeds [1.0 + 2σ] is determined to be acceptable.
[0020]
As described above, as the primary determination, the interval in which quality could not be accepted with a predetermined determination accuracy (confidence interval) based on the analysis value of the molten steel sample (in the case of the 98% reliability interval for the primary determination, the predicted internal defect score is The slab of [1.0-2σ] to [1.0 + 2σ]), that is, the slab of the pan changing portion that could not be accepted or rejected by the primary determination, was analyzed by slab sample analysis with high accuracy of analysis itself. Next decision is made.
[0021]
As shown in FIG. 2, the slab sample for secondary determination is injected in the pot changing section from the start of the next pot injection until the integrated value of the molten steel injection amount to TD becomes 1.5 times the TD maximum capacity. It is collected from the cast slab of molten steel. In order to specify the position of the cast slab of molten steel within the above range, the slab that has passed through the meniscus in the mold until the molten steel injection amount integrated value into the TD is 1.5 times the TD maximum capacity is usually obtained. It is sufficient to cut out the part after casting. However, the sampling position of the slab sample can be appropriately selected by various numerical simulations in consideration of flow and solidification, and by adding a tracer into the tundish.
[0022]
As with the primary determination, the reliability of the determination accuracy of the quality acceptance criterion (here, internal defect score 1.0) is determined from the correlation between the predicted result based on the analysis value and the product result after rolling. A pass / fail judgment is performed based on this.
In the secondary determination, when the predicted value of the internal defect score satisfies the above expression (2), it is determined as an acceptable material, and when it does not satisfy the expression (2), all are determined as unacceptable materials. As shown in FIG. 2, since the slab sample of the secondary determination has a high correlation with the product quality result compared to the molten steel sample analysis, the interval of the predicted internal defect score that could not be determined by the primary determination (for example, the reliability described above) In the primary determination when the section is about 98%, the pass / fail determination can be made even for [1.0-2σ] to [1.0 + 2σ]).
[0023]
In the quality determination flow shown in FIG. 1, the confidence interval determination coefficient K = 2 is set for both the primary determination and the secondary determination, but the value of K can be appropriately determined by those skilled in the art as necessary. In addition, although the multiple regression analysis method is used as a statistical method, when the relationship between various factors and the defect occurrence rate cannot be regarded as linear, it is possible to make a determination by using a neural network method or the like.
In addition, in the case of a special case where the variation of the predicted value with respect to the true value does not follow the normal distribution, it is possible to determine the confidence interval corresponding to the distribution by appropriately fitting a chi-square distribution or a parabola.
[0024]
In addition, the representative sample of the ladle exchange point required for the primary determination and the secondary determination is as follows. The amount of molten steel injected into the TD from the pot injection start point after the two or more pans of the multiple casting is 1. If it is up to 5 times, it can be determined by one sample. However, in order to reduce determination errors due to defective samples and errors due to changes in cleanliness depending on the casting direction, it is possible to appropriately collect a plurality of samples and use the average value or median value as a determination value.
In addition, the judgment index may be based on a simple correlation between product performance and cleanliness analysis data, but the slab position casting length from the pan change point, sample collection timing, and fluctuation of the molten metal level before and after the sample collection position It is desirable to make judgments that take into account operational indicators that are correlated with product performance other than sample analysis values.
[0025]
【Example】
In the examples, aluminum killed steel was cast with a TD maximum capacity of 40 t, a heat size of 340 t, and a 1-strand continuous casting machine, and a cold-rolled coil was manufactured by rolling the obtained slab. In the pan exchanging part, a molten steel sample was taken from the molten steel in the TD, and a slab sample was taken from the slab from which the molten steel was cast, and analyzed by the electron beam method. In addition, the pot cold rolled coil was inspected for internal defects using a leakage flux inspection device.
Table 1 shows the quality determination results by 121 methods (Example 1, Conventional Examples 1 and 2, Comparative Example 1) for 121 slabs of the same pan replacement part. In addition, the internal defect score is 1.0 as the pass / fail criterion for the slab.
[0026]
In Table 1, Example 1 represents the determination result of the present invention, and in accordance with the determination flow shown in FIG. 1, the molten steel collected from all the slabs of the pan replacement part with a 98% confidence interval for primary and secondary determination. The primary determination was made based on 121 samples, and the pass / fail determination was performed again by the secondary determination on 51 samples that could not be determined by the primary determination (the confidence interval was out of 98%).
The molten steel and slab samples were sampled one by one from the prescription of the present invention, that is, from the range where the total amount of injection was 60 t from the start point of the pot injection. Moreover, the prior art examples 1 and 2 performed the quality determination of the slab of the pan exchange part independently for the TD molten steel sample analysis and the slab sample analysis, respectively. Furthermore, the comparative example 1 shows the quality determination result at the time of extract | collecting molten steel and a slab sample from the position other than the prescription | regulation of this invention.
[0027]
[Table 1]

[0028]
As shown in Table 1, in Example 1 of the present invention, the acceptance judgment accuracy (in the slab judged as the quality acceptance based on the collected sample analysis result, the ratio of the product quality results after rolling the slab was also acceptable. ) Is 98%, and the overall acceptance rate (the ratio of product quality results after rolling of the slab in all 121 collected samples of the pan replacement part) is 78%, whereas the conventional example In 1 and Comparative Example 1, the accuracy of the pass judgment was the same as that of Example 1 of the present invention, but the overall pass rates were as low as 45% and 26%, respectively, and the slabs of the pan replacement part judged as the remaining rejects. Among them, the actual product after rolling the slab was judged to be rejected regardless of the quality that passed the acceptance criteria, and the slab yield was lowered. Further, in the conventional example 2, the overall pass rate is the same as that of the first embodiment of the present invention, but the number of slab samples analyzed is 2.4 times that of the present invention, and the overall analysis cost is also increased. It took twice as much as Example 1.
[0029]
【The invention's effect】
According to the present invention, as described above, in the case of the present invention, it is possible to judge the quality of the slab at the pan exchange point with a high pass rate and a low cost.
[Brief description of the drawings]
FIG. 1 is a diagram showing a quality determination flowchart of an example of the present invention. FIG. 2 is a diagram showing a correlation coefficient between sample collection timing and internal defect score. FIG. Fig. 4 shows the predicted value and actual value of the rolled sheet performance. Fig. 4 shows the variation in the predicted internal defect score as seen from the true score X.

Claims

In the slab quality judgment method at the ladle exchange site when multiple casts steel with a continuous caster, the accumulated molten steel injection amount is 1 of the maximum tundish capacity from the start of the tundish injection of ladle molten steel in the ladle exchange part. Collecting a molten steel sample from within the tundish until it reaches 5 times, firstly determining the slab quality of the ladle replacement site based on the quality analysis result of the molten steel sample, the quality determination result is determined in advance If the determination accuracy is outside the range of the confidence interval, the tundish injection of the ladle molten steel in the ladle replacement part is started until the cumulative amount of molten steel injection reaches 1.5 times the maximum tundish volume. Quality determination at the ladle replacement site of the continuous cast slab characterized by secondary determination of the slab quality of the ladle replacement part based on the slab sample taken from the cast slab of molten steel injected into the tundish Method