JP3882641B2 - Correction coefficient calculation method and orientation measurement method in online orientation measurement - Google Patents

Description

において補正係数初期値（α_{I １}、α_{I ２}、α_{I ３}…α_{I ｎ}）から徐々に各初期値を変化させて計算を繰り返し、ａ_kを最小にする最終的に定まる補正係数となるα_{L １}…α_{L ｎ}を算出するステップ。
【００１４】
本発明は、上記の補正係数算出方法を必要に応じて一つまたは二つ組み合わせて更に次のステップを備えて測定対象試料の配向を求めるオンライン配向測定方法である。
（ステップ１）走行する測定対象試料に測定ヘッドを接触または近接させた状態で、各誘電体共振器の共振周波数を測定するステップ。
（ステップ２）ステップ１で得られた各誘電体共振器の共振周波数とブランク時共振周波数の差から測定試料シフト量を算出するステップ。
（ステップ３）ステップ２で得られた周波数シフト量に請求項１記載の誘電体共振器ヘッドの補正係数を演算して補正後のシフト量とし、その値を元に測定対象試料の配向を求めるステップ。
【００１５】
【発明の実施の形態】
ソフトウエア的方法によって、複数の誘電体共振器の個体差を補正する方法の基本的な考え方としては以下の方法が考えられる。個々の誘電体共振器の検出部よりも一回り程度大きいサイズの標準試料を１枚用意し、これを順次各誘電体共振器の上に載せて各誘電体共振器のシフト量を測定し、その後各シフト量の平均値を各誘電体共振器のシフト量で除すことによって各誘電体共振器に対する補正係数を算出する方法である。実際の測定対象試料の測定では、時々刻々と測定される測定試料シフト量にこの補正係数を掛けることによって補正を行うという考え方である。
【００１６】
しかし、この方法を試行した場合に、実際的には全誘電体共振器の補正係数を求めるための時間がかかることや得られる補正係数の再現性も良くないという欠点があることが分かった。この原因としては、標準試料を複数回置き直す際の位置決め精度やその際の検出部との接触程度の問題、例えば常に同一の条件で均一に接触しているか、他の共振器との関連等の問題があると推察された。
【００１７】
そこで、本発明者らは時間をかけずに、再現性よく補正ができ、実用的でかつ効果的な２つの補正方法を発明するに至った。これによって誘電体共振器の個体差を無くし、高精度の測定ができることを確認した。
【００１８】
第一の方法は、無配向の大きな試料を用いる方法であり、ここで大きなとは実質的に全ての誘電体共振器の検出部を覆う程度の大きさを意味する。以後本方法を簡易型補正法と呼ぶ。等方性のシート（誘電率あるいは屈折率の面方向での異方性が実質的にないシート）を標準試料とし、これを測定したときの各誘電体共振器のシフト量（ブランク時の共振周波数と試料が有るときの共振周波数との差）が同じになるように予め補正係数を求めておき、測定時にはこの補正係数をシフト量に掛けて補正するという方法である。この方法は、無配向の標準試料があれば簡単に補正係数を求められる利点があり、日常のキャリブレーションには有効な手段である。特に、温度、湿度の経時的な変化に対して、短い周期で行う補正方法としては非常に有効である。しかし、厳密には完全に無配向の試料はあり得ないし、無配向に極めて近い物質も現実的には入手が難しいという問題がある。できるだけ配向を等方的にさせた試料を実質的に無配向の試料とし、これを標準試料とする。
【００１９】
そこで、このような厳密には少々配向がある標準試料を用いた場合にも精密な補正ができる方法として、第２の方法を案出した。この方法は、単純化して説明すれば標準試料を誘電体共振器の個数に対応して、一定の角度ずつ回転させながらその都度共振周波数を測定し、その結果に対して数値的処理を行うことによって各誘電体共振器の補正係数を求める方法である。以後、本方法を回転型補正法と呼ぶ。本補正法は、後に説明する逐次補正法と同様に測定装置が元来もっている各誘電体共振器の本質的な個体差を補正するために用いる方法として位置付けられる。これらの方法と測定中の外乱（温度、湿度など）の影響による個体差補正を行う前述の簡易型補正法とを併用した場合は、極めて高精度な測定が可能となるものである。
【００２０】
まず、簡易型補正法の考え方を図５に示す。図５は各々を角度＝７２°づつ変化させて配置した５個の誘電体共振器によって構成された測定ヘッドに標準試料１６を載せて測定する状態を示した図である。
このような複数の測定部を使用した場合、信号処理的には、図６のようにマイクロ波掃引発振器から出た信号を５個の誘電体共振器に分配し、透過強度を検波ダイオードで電圧に変換する。これを増幅、Ａ／Ｄ変換し、ピーク検出回路によってピーク位置が検出される。周波数の掃引は一定の周期で繰り返され、かつ掃引中のみハイレベルとなる同期信号が同時にマイクロ波掃引発振器から出ているため、この同期信号がハイレベルになる瞬間から透過強度が最大値をとるまでの時間を測定すれば、共振周波数が求められる。図６は５個の誘電体共振器からの信号を処理する回路のブロック図である。例えば、２５０ＭＨｚを１０ｍｓｅｃで掃引する場合のタイムチャートを図７に示す。図７は図６に示したブロック図における信号の処理を示すタイムチャートである。次に、補正係数を求める手順を示す。
【００２１】
１．測定ヘッド上に何も置かない状態で、各誘電体共振器の共振周波数を測定することによりブランク時共振周波数を求める。
２．少なくとも全ての誘電体共振器の測定面を覆う実質的に無配向の標準試料を測定ヘッド上に載せ、各誘電体共振器の共振周波数を測定する。
３．ステップ１および２にて得られた各誘電体共振器毎の共振周波数の差から標準試料シフト量を算出する。
４．得られた各誘電体共振器の標準試料シフト量を平均し、得られた平均値を各誘電体共振器の標準試料シフト量で除して各誘電体共振器毎の補正係数を算出する。
以上のステップで各誘電体共振器毎の補正係数が定まる。実際の測定では、その後以下のステップにしたがって配向を求めることになる。
【００２２】
１．走行する測定対象試料に測定ヘッドを接触または近接させた状態で、各誘電体共振器の共振周波数を測定する。
２．ステップ１で得られた各誘電体共振器の共振周波数とブランク時共振周波数の差から測定試料シフト量を算出する。
３．ステップ２で得られた周波数シフト量に先に求めた誘電体共振器ヘッドの各々の補正係数を掛けて補正後のシフト量とし、その値を元に測定対象試料の配向を求める。具体的には補正後のシフト量を極座標上にプロットし、配向パターンを求める。
【００２３】
以上の各誘電体共振器毎の補正係数を算出する手順の中のステップ４において「…得られた平均値を各誘電体共振器の標準シフト量で除して各誘電体共振器毎の補正係数」を得ているが、この補正係数は配向を求める手順の中のステップ３の「…各々の補正係数を掛けて…」に対応して「除し」て得られている。逆に「…各誘電体共振器の標準シフト量を得られた平均値で除して各誘電体共振器毎の補正係数」を得た場合は「…各々の補正係数によって割る…」ことによって補正後のシフト量が得られる。従ってこのことをまとめて表現するために補正係数を算出する手順の中のステップ４では「…得られた平均値と各誘電体共振器の標準シフト量を演算して各誘電体共振器毎の補正係数を算出する…」と表現した。また更に配向を求める手順の中のステップ３では「…各々の補正係数を演算して…」と表現した。
【００２４】
次に、第１の回転型補正法の試料測定段取りを図８に示す。図８は図５と同様に誘電体共振器の数が５個の場合を示している。以下に補正係数を求める手順を示す。
１．測定ヘッド上に何も置かない状態で、各誘電体共振器の共振周波数を測定することによりブランク時共振周波数を求める。
２．少なくとも全ての誘電体共振器の測定面を覆う実質的に無配向の標準試料を測定ヘッド上に載せ、各誘電体共振器の共振周波数を測定する。図８（２）参照、Pos１は標準試料のPos１領域の中心線を一点鎖線によって表わしている、Pos１領域を誘電体共振器１が測定している。以下各誘電体共振器も同様である。
３．測定ヘッド上で、標準試料を一定の角度θ（θ＝３６０°／誘電体共振器の数）、従ってこの場合は７２°回転させた後、同様にして各誘電体共振器の共振周波数を測定する。図８（３）参照。ここでは誘電体共振器２がPos１領域を測定している。この７２°の回転と測定を総計で誘電体共振器の数だけ繰り返し行う。従って、このステップ２と３では試料の５分割された扇形の各領域に対して５個の各誘電体共振器毎に共振周波数を測定していることになる。即ち標準試料が元の位置まで１回転すれば終了である。
【００２５】
４．全ての領域に対して各誘電体共振器毎に測定した共振周波数と各誘電体共振器のブランク時共振周波数との差から各領域毎に各誘電体共振器の標準試料シフト量を算出する。
５．以上の結果を図９に示した。図９では表計算ソフトウエアを用いて第１段階のような表に整理する。この表において、左上から右下に斜めも並んだデータ（標準試料シフト量）は、標準試料の同一領域を各誘電体共振器が測定していることになる。例えば図で網掛けの領域はPosＢの領域を各誘電体共振器が測定したデータである。同様に、一行下のデータについても左上から右下に沿ったデータは、標準試料の別の領域ではあるが、各誘電体共振器からみれば同じ領域を測定したデータとなっている。
６．そこで、各斜めのデータをそれぞれ縦に並び替えると、第２段階の表のようになる。この表で、各列は標準試料の同一領域のデータになっているので、誘電体共振器の個体差が無ければすべて同じ値になるはずである。そこで、各列毎に平均を計算する。
７．この平均を各列のデータで除すことにより、平均値で規格化した値が得られ、行毎でみると各誘電体共振器毎に規格化された値が並んでいることになり、各行毎の平均をとれば、これが各誘電体共振器の補正係数となる。この状態が第３段階の表に示されている。次の第４段階では第２段階の表を上記各補正係数により補正した後のデータを示している。
【００２６】
各誘電体共振器の補正係数が求まれば、先に説明した方法により同様に配向を求められる。また以上の例では円周上に角度を変えて配置した複数の誘電体共振器によって構成した測定ヘッド上に標準試料を同心になるようにおいて回転方向に位置を変えながら補正係数を求めた。
【００２７】
この考え方を直線上に角度を変えて配置した複数の誘電体共振器によって構成した測定ヘッドに適用することも可能である。このような測定ヘッドの例を図１０に示した。この場合標準試料は帯状の試料をこの誘電体共振器の列にそって一つの誘電体共振器の測定領域毎にずらしながら測定を繰り返して同様に補正係数を求められる。この際、回転型補正の場合は標準試料を回転させることにより全ての誘電体共振器が同一の各領域に対する測定を自然に行えるのに対して帯状試料を単にずらす場合は最初の測定以降別の領域が次々加わることになる。この場合はあまり良好な補正係数が得られにくいと考えられる。従ってその欠点を改善するために、帯状標準試料を誘電体共振器の数に合わせてずらす方向に領域毎に等分割し、それぞれの領域をテープ等でつないでおき、１回１回の測定後テープ等をはずして領域を順送りに並べ変えて測定するような方法をとることも可能である。このような測定方法を以降簡単のために逐次補正法と呼ぶ。この逐次補正方法と第１の回転型補正法の両方の補正係数算出方法の両方含むようにに表現すると以下のような方法になる。
【００２８】
次のステップを備えた補正係数算出方法。
（ステップ１）試料の一面側のみに配置された複数（ｎ個）の誘電体共振器を備える測定ヘッド上に何も置かない状態で、各誘電体共振器の共振周波数を測定することによりブランク時共振周波数を求めるステップ。
（ステップ２）少なくとも全ての誘電体共振器の測定面を覆う実質的に無配向の標準試料を測定ヘッド上に載せ、各誘電体共振器の共振周波数を測定するステップ。
（ステップ３）ステップ２において各誘電体共振器に対して定まった標準試料の特定の領域（１からｎ番目の領域）について、試料を少なくとも（ｎ−１）回置き換えて、全ての１からｎ番領域に対して各誘電体共振器毎に共振周波数を測定するステップ。
（ステップ４）全ての１からｎ番領域に対して各誘電体共振器毎に測定した共振周波数と各誘電体共振器のブランク時共振周波数との差から各領域（１からｎ番目の領域）毎に各誘電体共振器の標準試料シフト量を算出するステップ。
（ステップ５）ステップ４で得られた各誘電体共振器の標準試料シフト量の平均値を各領域（１からｎ番目の領域）毎に算出し、その平均値を各誘電体共振器の標準試料シフト量で除し、各領域毎に規格化した各誘電体共振器の標準試料シフト量を算出し、次にその各規格化した各誘電体共振器の標準試料シフト量を各誘電体共振器毎に平均化して、各誘電体共振器毎の補正係数を算出するステップ。
【００２９】
更に、簡易型補正法の項で説明したように補正係数を最終的にかけるかまたは補正係数で除するかについて両方を意味するように書き換えると以下のようになる。
次のステップを備えた補正係数算出方法。
（ステップ１）試料の一面側のみに配置された複数（ｎ個）の誘電体共振器を備える測定ヘッド上に何も置かない状態で、各誘電体共振器の共振周波数を測定することによりブランク時共振周波数を求めるステップ。
（ステップ２）少なくとも全ての誘電体共振器の測定面を覆う実質的に無配向の標準試料を測定ヘッド上に載せ、各誘電体共振器の共振周波数を測定するステップ。
（ステップ３）ステップ２において各誘電体共振器に対して定まった標準試料の特定の領域（１からｎ番目の領域）について、試料を少なくとも（ｎ−１）回置き換えて、全ての１からｎ番領域に対して各誘電体共振器毎に共振周波数を測定するステップ。
（ステップ４）全ての１からｎ番領域に対して各誘電体共振器毎に測定した共振周波数と各誘電体共振器のブランク時共振周波数との差から各領域（１からｎ番目の領域）毎に各誘電体共振器の標準試料シフト量を算出するステップ。
（ステップ５）ステップ４で得られた各誘電体共振器の標準試料シフト量の平均値を各領域（１からｎ番目の領域）毎に算出し、その平均値を各誘電体共振器の標準試料シフト量と演算し、各領域毎に規格化した各誘電体共振器の標準試料シフト量を算出し、次にその各規格化した各誘電体共振器の標準試料シフト量を各誘電体共振器毎に平均化して、各誘電体共振器毎の補正係数を算出するステップ。
【００３０】
回転型補正方法の別法として図１１に説明するような方法によって補正係数を求めることもできる。この方法を第２の回転型補正法と呼ぶ。図１１は表計算ソフトウエアを用いて第２の回転型補正法によりデータを処理する状態を示す図である。以下に手順を示す。
１．測定ヘッド上に何も置かない状態で、各誘電体共振器の共振周波数を測定することによりブランク時共振周波数を求める。
２．少なくとも全ての誘電体共振器の測定面を覆う実質的に無配向の標準試料を測定ヘッド上に載せ、各誘電体共振器の共振周波数を測定する。図８（２）参照、Pos１は標準試料のPos１領域の中心線を一点鎖線によって表わしている、Pos１領域を誘電体共振器１が測定している。以下各誘電体共振器も同様である。
３．測定ヘッド上で、標準試料を一定の角度θ（θ＝３６０°／誘電体共振器の数）、従ってこの場合は７２°回転させた後、同様にして各誘電体共振器の共振周波数を測定する。図８（３）参照。ここでは誘電体共振器２がPos１領域を測定している。この７２°の回転と測定を総計で誘電体共振器の数だけ繰り返し行う。従って、このステップ２と３では試料の５分割された扇形の各領域に対して５個の各誘電体共振器毎に共振周波数を測定していることになる。即ち標準試料が元の位置まで１回転すれば終了である。
【００３１】
４．全ての領域に対して各誘電体共振器毎に測定した共振周波数と各誘電体共振器のブランク時共振周波数との差から各領域毎に各誘電体共振器の標準試料シフト量を算出する。
５．以上の結果を図９に示した。図９では表計算を用いて第１段階のような表に整理する。この表において、左上から右下に引いた線に沿ったデータ（標準試料シフト量）は、標準試料の同一領域を各誘電体共振器が測定していることになる。例えば図で網掛けの領域はPosＢの領域を各誘電体共振器が測定したデータである。同様に、一行下のデータについても左上から右下に沿ったデータは、標準試料の別の領域ではあるが、各誘電体共振器からみれば同じ領域を測定したデータとなっている。
６．そこで、各斜めのデータをそれぞれ縦に並び替えると、第２段階の表のようになる。この表で、各列は標準試料の同一領域のデータになっているので、誘電体共振器の固体差が無ければすべて同じ値になるはずである。そこで、各列毎に平均を計算する。
７．第３段階に示すように、この平均を各列のデータ（標準試料シフト量）で除すことにより、平均値で規格化した値：規格化シフト量が得られる。
８．第４段階に示すように、各誘電体共振器毎の補正係数初期値を予め適当に決めておき、まずこの補正係数初期値と対応する各誘電体共振器の規格値を乗算し、その結果を５×５のマトリックスに入れる。２５個のデータができる限り１．０００に近づけばよいので、各誘電体共振器の補正係数初期値から少しずつ変えながら、その作業を繰り返す。これは各データと１．０００との差の２乗の和（２５個の２乗和）を最少にする最終的に定まる各誘電体共振器毎の補正係数を算出することになる。第５段階では、補正係数初期値を使用したときのこの２乗和を算出した状態を示している。最終的にこの２乗和を最小にする補正係数が算出されたとき、このときが最も全体としてながめた場合に、２５個のデータが同じ値（平均値で規格化した規格化シフト量の１．０００）に最も近づく時である。次の第６段階では第２段階の表を上記各補正係数により補正した後のデータを示している。
【００３２】
上記、８．のステップをもう少し詳細に説明すると以下のようなフローになる。
▲１▼各ヘッドの補正係数αｎ（ｎ＝1,2,3,…）の初期値（α_{I １}、α_{I ２}、α_{I ３}…α_{I ｎ}）を決める
▲２▼第３段階におけるマトリックスをｎ行×ｎ列として、各行について補正係数を掛け、以下の式（１）で定まるＸ_ijを求める。
Ｘ_ij＝Ｘ_nj×α_Ｉｎ（ｎ＝1,2,3,…ｎ、ｊ＝1,2,3,…ｎ）…（１）
（ｎ×ｎ）個のＸ_ijについて、次の式（２）を計算する
【００３３】
【数３】

▲３▼α_{I ｎ}を変えて同様の計算を繰り返し行い、このａ_kが最も最小値をとるときのα_{L １}…α_{L ｎ}を求める。その時のαが求める補正係数となる。つまり、この時がｎ×ｎ個の補正された規格化値が全体としてみた場合、最も１に近くなる。すなわち、各領域（列）での値が最もバラツキが小さくなる（ｎ個のシフト量が平均値に最も近づく）。
【００３４】
以上の例も第１の回転型補正法と同様に逐次補正方法にも適用できる。更には補正係数を最終的にかけるか補正係数で除するかについて両方を意味するように書き換えることもできる。そのように表現すると以下のようになる。
次のステップを備えた補正係数算出方法。
（ステップ１）試料の一面側のみに配置された複数（ｎ個）の誘電体共振器を備える測定ヘッド上に何も置かない状態で、各誘電体共振器の共振周波数を測定することによりブランク時共振周波数を求めるステップ。
（ステップ２）少なくとも全ての誘電体共振器の測定面を覆う実質的に無配向の標準試料を測定ヘッド上に載せ、各誘電体共振器の共振周波数を測定するステップ。
（ステップ３）ステップ２において各誘電体共振器に対して定まった標準試料の特定の領域（１からｎ番目の領域：これを仮にＡ，Ｂ，Ｃ…のｎ個の領域として以下Ａ，Ｂ，Ｃ…の記号を使用）について、試料を少なくとも（ｎ−１）回置き換えて、全ての１からｎ番領域に対して各誘電体共振器毎に各領域に対する共振周波数を測定するステップ。
（ステップ４）全ての１からｎ番領域に対して各誘電体共振器毎に測定した共振周波数と各誘電体共振器のブランク時共振周波数との差から各領域毎に各誘電体共振器の標準試料シフト量を算出するステップ。
（ステップ５）ステップ４で得られた各誘電体共振器の標準試料シフト量の平均値を各領域（Ａ，Ｂ，Ｃ…のｎ個の領域）毎に算出し、その平均値と各誘電体共振器の各領域における標準試料シフト量を演算し、各領域毎に規格化した各誘電体共振器の標準試料シフト量を算出するステップ。（この値を規格化シフト量とする）
（ステップ６）各誘電体共振器毎に予め決めた任意の補正係数初期値（α_{I １}、α_{I ２}、α_{I ３}…α_{I ｎ}）を規格化シフト量（ＳＶ_１Ａ，ＳＶ_１Ｂ，ＳＶ_１Ｃ，…ＳＶ_２Ａ，ＳＶ_２Ｂ…ＳＶ_３Ａ，ＳＶ_３Ｂ…）に演算し、補正後規格化シフト量初期値（ＳＶ_１Ａ×α_{I １}＝Ｘ_{１ A}，ＳＶ_１Ｂ×α_{I １}＝Ｘ_１Ｂ，_ＳＶ１Ｃ×α_{I １}＝Ｘ_１Ｃ，…ＳＶ_２Ａ×α_{I ２}＝Ｘ_{２ A}，ＳＶ_２Ｂ×α_{I ２}＝Ｘ_{２ B}…ＳＶ_３Ａ×α_{I ３}＝Ｘ_{３ A}，ＳＶ_３Ｂ×α_{I ３}＝Ｘ_{３ B}…）を得て、
この補正後規格化シフト量初期値についての下式（１）において補正係数初期値（α_{I １}、α_{I ２}、α_{I ３}…α_{I ｎ}）から徐々に各初期値から変化させて計算を繰り返し、ａ_kを最小にする最終的に定まる補正係数となるα_{L １}…α_{L ｎ}を算出するステップ。
【００３５】
【数４】

【００３６】
以上の説明中の式は最終的に定まる補正係数をかける場合に適用される式を説明上付け加えている。
以上いずれの方法にせよ得られた補正係数により簡易型補正法において説明したような配向を測定するステップを加えて最終的には配向を求めることができる。以上のような演算ステップをマイクロコンピュータ等によりプログラム化して補正係数の演算を自動化することは容易である。この場合、標準試料の測定が自動化されないときは、測定毎に手動操作が加わることになる。
以上のようにオンラインにて配向パターンが測定できれば、その測定結果をコンピュータによって演算し、抄紙機や延伸機のパラメータ、例えばＪ／Ｗ比（ジェットワイヤー比）や延伸倍率（延伸ロール回転速度差）を調整すること等によって配向パターンをオンライン制御できることはいうまでもない。
【００３７】
【実施例】
実施例１
図５に示した測定ヘッドを使用し、図６に示した回路構成をとって図７のようなタイムチャートにしたがって測定を行った。図１２に簡易型補正法を施した結果を示した。図１２は中程度の配向度を有する紙を３００ｍ／ｍｉｎで走行している時の繊維配向を測定した時の配向パターンを示す。
図１３には同一サンプルを同条件で測定しているときの経時変化を示すデータを示す。上半分が配向角度（−９０〜＋９０°表示）、下半分が配向度（０〜８００KHz）を示し、５秒毎にプロットしたものであり、安定して測定できているのがわかる。オフラインで測定した結果ともよく一致した。
【００３８】
実施例２
実施例１と同様の測定系によって測定を行った。図１４に回転型補正を施した測定データを示す。試料はほとんど配向のない紙を用い、同様に３００ｍ／ｍｉｎで走行中のものを測定した。配向度は１６KHzとほとんど無配向に近いデータが出ている。
図１５に同試料を同条件で測定したときの経時変化を示すデータを示す。ほとんど無配向に近いにもかかわらず、安定して測定されており、かつ配向角度もオフラインで従来の配向測定装置によって得られた測定結果とよく一致した。
【００３９】
【発明の効果】
以上のように、本発明によれば比較的異方性の小さい試料に対しても、精度良く配向性（配向角度および配向度）を測定することが可能となった。
【図面の簡単な説明】
【図１】誘電体共振器の（１）は平面図、（２）は垂直断面図である。
【図２】５個の誘電体共振器を一つの金属ケースに配置して得られた測定ヘッドの一例を示す平面図。
【図３】ブランク（試料の無い状態）時と試料測定時の共振周波数のシフト量を表わす図。
【図４】配向パターンの一例を示す図。
【図５】図５は各々を角度＝７２°づつ変化させて配置した５個の誘電体共振器によって構成された測定ヘッドに標準試料を載せて測定する状態を示した図。
【図６】５個の誘電体共振器からの信号を処理する回路のブロック図。
【図７】図６に示したブロック図における信号の処理を示すタイムチャート。
【図８】回転型補正法の試料測定段取りを示す図。
【図９】表計算ソフトウエアを用いて第１の回転型補正法によりデータを処理する状態を示す図。
【図１０】直線上に角度を変えて配置した複数の誘電体共振器によって構成した測定ヘッドの一例を示す図。
【図１１】表計算ソフトウエアを用いて第２の回転型補正法によりデータを処理する状態を示す図。
【図１２】中程度の配向度を有する紙を３００ｍ／ｍｉｎで走行している時の繊維配向を測定し簡易型補正法により処理した時のディスプレイを示す図。
【図１３】図１２のサンプルを測定しているときの経時変化を示す図。
【図１４】ほとんど配向のない紙を３００ｍ／ｍｉｎで走行中している時の繊維配向を測定し第１の回転型補正法により処理した時のディスプレイを示す図。
【図１５】図１４のサンプルを測定しているときの経時変化を示す図。
【符号の説明】
１〜５：誘電体共振器
１５：測定ヘッド
１６：標準試料[0001]
BACKGROUND OF THE INVENTION
The present invention includes not only sheet-like materials such as paper, non-woven fabrics, and films, but also three-dimensional articles such as molded articles such as plastics, rubbers, and ceramics. The present invention relates to a measurement method using a microwave dielectric resonator.
[0002]
[Prior art]
Conventionally, X-ray diffraction, infrared dichroism, mechanical breaking strength, ultrasonic wave propagation speed, birefringence, polarization fluorescence method, microwave method and the like have been used as methods for measuring the orientation of a sheet-like substance. Yes. Most of these are used in laboratories, and the only method that can be used as a so-called on-line measuring method that can measure a sample while traveling is the method based on birefringence. This is a method for measuring the anisotropy of the refractive index, that is, the birefringence or retardation (birefringence × thickness) in the sheet surface, and is disclosed in JP-A-4-89553.
[0003]
However, since this method using birefringence needs to be measured by transmitting visible light (polarized light), there is a problem in that it can be measured only by a substance that transmits light to some extent, such as a transparent film. Further, for a sample having a large wavelength dispersion and a large anisotropy such as a PET (polyethylene terephthalate) film, there is a problem that it is difficult to actually measure because it is difficult to determine the optical order.
[0004]
Therefore, the inventors use a microwave dielectric resonator to bring the detection part into contact with or close to one side of the sample, so that the orientation or dielectric anisotropy of the sample can be obtained online regardless of whether it is transparent or opaque. Has applied for a method to measure. See JP 10-325811 A and JP 2001-91476 A. This method uses the change in the resonance frequency when the dielectric resonator is in contact with or close to one side of the sample. The basic principle of measurement can be expressed in a word as “anisotropy of dielectric constant. I'm watching. " Both the refractive index in the birefringence method and the dielectric constant used in this method originate from electronic polarization, and in the high frequency region such as light, the square of the refractive index is equal to the dielectric constant. In the end, you will see the same thing.
[0005]
In practice, a plurality of dielectric resonators as shown in FIG. 1 are used. FIG. 1 is a plan view (1) and a cross-sectional view (2) of one dielectric resonator. A rectangular parallelepiped ceramic 11 serving as a dielectric is fixed in an aluminum block-made metal case 12 so that the upper surface thereof is flush with the upper surface of the metal case 12. Rod antennas 13a and 13b are provided in the vicinity of the ceramic, and microwaves are input to and output from these antennas to resonate the ceramic. A cover 14 made of tetrafluoroethylene resin is provided in the gap between the ceramic 11 and the metal case 12 to prevent entry of dust and the like. A plurality of such dielectric resonators 1 to 5 are arranged as shown in FIG. FIG. 2 is a plan view of the measuring head.
[0006]
As shown in FIG. 3, the fact that the shift amount of the resonance frequency in the presence or absence of the sample depends on the dielectric constant of the sample (strictly, it depends on the dielectric constant × thickness, but when the thickness is regarded as uniform) is used. Then, the difference (hereinafter referred to as the shift amount) between the resonance frequency when there is no sample and the resonance frequency when there is a sample is measured. FIG. 3 is a diagram showing the shift amount of the resonance frequency at the time of blank (without sample) and measurement of each sample. The sample shows two types of states installed in two directions of 0 ° and 90 °. If this shift amount is plotted on polar coordinates corresponding to the installation direction of each dielectric resonator and elliptical approximation is performed, if there is anisotropy, for example, an orientation pattern as shown in FIG. 4 is obtained. FIG. 4 is a diagram illustrating an example of an alignment pattern. In FIG. 4, if there is no anisotropy, ie, no orientation, this orientation pattern is a circle. Here, the major axis direction of the ellipse is the direction with the maximum shift amount, which indicates the direction with the maximum dielectric constant (or refractive index). That is, this direction is the direction in which molecular chains are arranged. The degree of orientation can be expressed by the difference between the major axis and the minor axis of the ellipse or the ratio of both.
[0007]
[Problems to be solved by the invention]
As a result of various improvements as described in the above publication, it has become possible to accurately measure an orientation pattern as shown in FIG. 4 for a stretched polymer film such as a PET film.
However, the dielectric anisotropy of paper, non-woven fabric and the like is relatively small as compared to a stretched polymer film such as a PET film. In addition, unlike PET films and polyethylene films, polymers such as polypropylene and polystyrene have a side chain structure with a relatively high polarizability, and therefore the macroscopic dielectric anisotropy tends to be small. Furthermore, the dielectric anisotropy tends to be small even when the degree of crystallinity is small even in the same polymer film or when the polymer is originally an amorphous polymer. In order to apply to such a measurement object having a relatively small dielectric anisotropy, the resonance frequency shift amount in a plurality of dielectric resonators (resonance frequency in a state where there is no sample and in a state where there is no sample) Of course, it is necessary to more accurately capture the difference between the dielectric resonators, but there is a slight difference between individual dielectric resonators (the same resonant frequency shift does not occur even if the same sample is contacted). It was found that it was necessary to make it as small as possible.
[0008]
As a means to solve this, of course, it is indispensable to manufacture with the same hardware specifications such as the external dimensions and dielectric constants of individual dielectric resonators, but as a matter of fact, the characteristics should be exactly the same. Is difficult. Eventually, the resonance frequency when the same sample is measured under the same conditions and the shift amount of the resonance frequency from the blank cannot be made the same.
[0009]
As described above, the individual differences that could not be made uniform by hardware are corrected by software to accurately measure orientation (orientation angle and orientation degree) even for samples with relatively small anisotropy. The purpose is to do. Such samples with relatively low anisotropy include paper or unstretched films, well-balanced biaxially stretched films, non-crystalline films, and molecular structures with highly polarized groups hanging from side chains. The film which has is mentioned.
[0010]
[Means for Solving the Problems]
  The present invention is a correction coefficient calculation method including the following steps in an online orientation measurement method for obtaining the orientation of a sample to be measured.
(Step 1) The blank resonance frequency is determined by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on a measurement head including a plurality of dielectric resonators arranged only on one side of the sample. Step to seek.
(Step 2) A substantially non-oriented standard sample covering the measurement surface of at least all dielectric resonatorsabout, Measuring the resonant frequency of each dielectric resonator.
(Step 3) The standard sample shift amount is calculated from the difference in resonance frequency for each dielectric resonator obtained in Steps 1 and 2, and the standard sample shift amount of each obtained dielectric resonator is averaged. A step of calculating a correction coefficient for each dielectric resonator by calculating the obtained average value and the standard shift amount of each dielectric resonator.
[0011]
  The present invention is a correction coefficient calculation method including the following steps in an online orientation measurement method for obtaining the orientation of a sample to be measured.
(Step 1) A blank is obtained by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on a measurement head including a plurality (n) of dielectric resonators disposed only on one side of the sample. Determining the resonance frequency.
(Step 2) Standard sample covering the measurement surface of at least all dielectric resonatorsabout, Measuring the resonant frequency of each dielectric resonator.
(Step 3) For the specific region (1st to nth region) of the standard sample determined for each dielectric resonator in Step 2, the sample is replaced at least (n−1) times, and all 1 to n Measuring a resonance frequency for each dielectric resonator with respect to the number region.
(Step 4) Each region (1st to nth region) from the difference between the resonance frequency measured for each dielectric resonator in all the 1st to nth regions and the blank resonance frequency of each dielectric resonator Calculating a standard sample shift amount of each dielectric resonator every time;
(Step 5) The average value of the standard sample shift amount of each dielectric resonator obtained in Step 4 is calculated for each region (1st to nth region), and the average value and the standard of each dielectric resonator are calculated. Calculate the sample shift amount, calculate the standard sample shift amount of each dielectric resonator normalized for each region, and then calculate the standard sample shift amount of each normalized dielectric resonator for each dielectric resonance Calculating a correction coefficient for each dielectric resonator by averaging each unit.
[0012]
  The present invention is a correction coefficient calculation method including the following steps in an online orientation measurement method for obtaining the orientation of a sample to be measured.
(Step 1) The blank resonance frequency is determined by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on a measurement head including a plurality of dielectric resonators arranged only on one side of the sample. Step to seek.
(Step 2) Standard sample covering the measurement surface of at least all dielectric resonatorsabout, Measuring the resonant frequency of each dielectric resonator.
(Step 3) A specific region of the standard sample determined for each dielectric resonator in Step 2 (1 to n-th region: suppose this is defined as n regions A, B, C... , C ...), and measuring the resonance frequency for each of the dielectric resonators for each of the 1st to nth regions by replacing the sample at least (n-1) times.
(Step 4) From the difference between the resonance frequency measured for each dielectric resonator in all the 1st to nth regions and the blank resonance frequency of each dielectric resonator, Calculating a standard sample shift amount;
(Step 5) The average value of the standard sample shift amount of each dielectric resonator obtained in Step 4 is calculated for each region (n regions of A, B, C...), And the average value and each dielectric A step of calculating a standard sample shift amount in each region of the body resonator and calculating a standard sample shift amount of each dielectric resonator normalized for each region (this value is referred to as a normalized shift amount).
(Step 6) Arbitrary correction coefficient initial value (α that is predetermined for each dielectric resonator)_{I 1}, Α_{I 2}, Α_{I 3}... α_{I n}) Normalized shift amount (SV)_1A, SV_1B, SV_1C, ... SV_2A, SV_2B... SV_3A, SV_3B...), and the corrected standardized shift amount initial value (X_{1 A}, X_1B, X_1C, ... X_{2 A}, X_{2 B}... X_{3 A}, X_{3 B}...), and the equation (1) for the initial value of the normalized shift after correction is obtained.
Correction coefficient initial value (α_{I 1}, Α_{I 2}, Α_{I 3}... α_{I n}), Gradually changing the initial values and repeating the calculation, a_kΑ, which is the final correction factor that minimizes_{L 1}... α_{L n}Calculating step.
[0013]
[Expression 2]

Correction coefficient initial value (α_{I 1}, Α_{I 2}, Α_{I 3}... α_{I n}), Gradually changing the initial values and repeating the calculation, a_kΑ, which is the final correction factor that minimizes_{L 1}... α_{L n}Calculating step.
[0014]
The present invention is an on-line orientation measurement method for obtaining the orientation of a sample to be measured by further comprising the following steps by combining one or two of the above correction coefficient calculation methods as required.
(Step 1) A step of measuring the resonance frequency of each dielectric resonator in a state where the measurement head is in contact with or close to the traveling measurement target sample.
(Step 2) A step of calculating a measurement sample shift amount from the difference between the resonance frequency of each dielectric resonator obtained in Step 1 and the resonance frequency at blank time.
(Step 3) The correction factor of the dielectric resonator head according to claim 1 is calculated from the frequency shift amount obtained in Step 2 to obtain the corrected shift amount, and the orientation of the sample to be measured is obtained based on the calculated value. Step.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
As a basic idea of a method for correcting individual differences among a plurality of dielectric resonators by a software method, the following method can be considered. Prepare one standard sample with a size about one size larger than the detection part of each dielectric resonator, place it on each dielectric resonator sequentially, and measure the shift amount of each dielectric resonator. Thereafter, a correction coefficient for each dielectric resonator is calculated by dividing the average value of each shift amount by the shift amount of each dielectric resonator. In the actual measurement of the sample to be measured, the idea is to perform correction by multiplying the measured sample shift amount measured every moment by this correction coefficient.
[0016]
However, when this method is tried, it has been found that in practice, it takes time to obtain the correction coefficient of the all dielectric resonator and the reproducibility of the obtained correction coefficient is not good. This can be caused by problems such as positioning accuracy when the standard sample is repositioned multiple times and the degree of contact with the detection unit at that time, for example, whether contact is always uniform under the same conditions, relation to other resonators, etc. It was guessed that there was a problem.
[0017]
Accordingly, the present inventors have invented two practical and effective correction methods that can be corrected with good reproducibility without taking time. As a result, it was confirmed that individual differences of dielectric resonators were eliminated and high-precision measurement was possible.
[0018]
The first method is a method using a large non-oriented sample, and the term “large” here means a size that covers substantially all the detectors of the dielectric resonators. Hereinafter, this method is referred to as a simplified correction method. An isotropic sheet (a sheet having substantially no anisotropy in the plane direction of dielectric constant or refractive index) is used as a standard sample, and the amount of shift of each dielectric resonator when this is measured (resonance at blank) In this method, a correction coefficient is obtained in advance so that the difference between the frequency and the resonance frequency when the sample is present is the same, and the correction coefficient is multiplied by the shift amount during measurement. This method has an advantage that a correction coefficient can be easily obtained if there is a non-oriented standard sample, and is an effective means for daily calibration. In particular, it is very effective as a correction method performed in a short cycle with respect to changes in temperature and humidity over time. However, strictly speaking, there is no completely unoriented sample, and there is a problem that a substance very close to non-oriented is actually difficult to obtain. A sample whose orientation is isotropic as much as possible is a substantially non-oriented sample, and this is a standard sample.
[0019]
Therefore, the second method has been devised as a method capable of performing a precise correction even when a standard sample having a slight orientation is used. In a simplified manner, this method involves measuring the resonance frequency each time the standard sample is rotated by a certain angle corresponding to the number of dielectric resonators, and performing numerical processing on the result. This is a method for obtaining the correction coefficient of each dielectric resonator. Hereinafter, this method is referred to as a rotational correction method. This correction method is positioned as a method used to correct an essential individual difference of each dielectric resonator, which the measurement apparatus originally has, as in the successive correction method described later. When these methods are used in combination with the above-described simplified correction method that corrects individual differences due to the influence of disturbance (temperature, humidity, etc.) during measurement, extremely high-precision measurement is possible.
[0020]
First, the concept of the simplified correction method is shown in FIG. FIG. 5 is a diagram showing a state in which the standard sample 16 is mounted on a measurement head constituted by five dielectric resonators arranged by changing each angle = 72 ° and measured.
When such a plurality of measuring units are used, in terms of signal processing, the signal emitted from the microwave sweep oscillator is distributed to five dielectric resonators as shown in FIG. Convert to This is amplified and A / D converted, and the peak position is detected by the peak detection circuit. The frequency sweep is repeated at a fixed period, and since the synchronization signal that becomes high level only during the sweep is output from the microwave sweep oscillator at the same time, the transmission intensity reaches the maximum value from the moment when this synchronization signal becomes high level. If the time until is measured, the resonance frequency can be obtained. FIG. 6 is a block diagram of a circuit for processing signals from five dielectric resonators. For example, FIG. 7 shows a time chart when sweeping 250 MHz at 10 msec. FIG. 7 is a time chart showing signal processing in the block diagram shown in FIG. Next, a procedure for obtaining a correction coefficient is shown.
[0021]
1. The blank resonance frequency is obtained by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on the measurement head.
2. A substantially non-oriented standard sample covering the measurement surfaces of at least all dielectric resonators is placed on the measurement head, and the resonance frequency of each dielectric resonator is measured.
3. The standard sample shift amount is calculated from the difference in resonance frequency for each dielectric resonator obtained in steps 1 and 2.
4). The obtained standard sample shift amount of each dielectric resonator is averaged, and the obtained average value is divided by the standard sample shift amount of each dielectric resonator to calculate a correction coefficient for each dielectric resonator.
The correction coefficient for each dielectric resonator is determined by the above steps. In actual measurement, the orientation is then determined according to the following steps.
[0022]
1. The resonance frequency of each dielectric resonator is measured in a state where the measurement head is in contact with or close to the traveling measurement target sample.
2. The measurement sample shift amount is calculated from the difference between the resonance frequency of each dielectric resonator obtained in step 1 and the blank resonance frequency.
3. The frequency shift amount obtained in step 2 is multiplied by the correction coefficient of each dielectric resonator head previously obtained to obtain the corrected shift amount, and the orientation of the sample to be measured is obtained based on that value. Specifically, the corrected shift amount is plotted on polar coordinates to obtain an orientation pattern.
[0023]
In step 4 in the procedure for calculating the correction coefficient for each dielectric resonator as described above, “... the average value obtained is divided by the standard shift amount of each dielectric resonator to correct for each dielectric resonator. The correction coefficient is obtained by “dividing” in correspondence with “... multiply each correction coefficient ...” in step 3 in the procedure for obtaining the orientation. On the other hand, when “… the correction coefficient for each dielectric resonator is obtained by dividing the standard shift amount of each dielectric resonator by the average value obtained”, “... divide by each correction factor…” A corrected shift amount is obtained. Therefore, in order to express this collectively, in step 4 in the procedure for calculating the correction coefficient, “... the average value obtained and the standard shift amount of each dielectric resonator are calculated and each dielectric resonator is calculated. The correction coefficient is calculated ... ". Further, in step 3 in the procedure for obtaining the orientation, “... each correction coefficient is calculated ...” is expressed.
[0024]
Next, the sample measurement setup of the first rotational correction method is shown in FIG. FIG. 8 shows a case where the number of dielectric resonators is five as in FIG. The procedure for obtaining the correction coefficient is shown below.
1. The blank resonance frequency is obtained by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on the measurement head.
2. A substantially non-oriented standard sample covering the measurement surfaces of at least all dielectric resonators is placed on the measurement head, and the resonance frequency of each dielectric resonator is measured. As shown in FIG. 8B, Pos1 represents the center line of the Pos1 region of the standard sample by a one-dot chain line, and the dielectric resonator 1 measures the Pos1 region. The same applies to each dielectric resonator hereinafter.
3. Measure the resonance frequency of each dielectric resonator in the same way after rotating the standard sample on the measuring head by a certain angle θ (θ = 360 ° / number of dielectric resonators), and in this case, rotating 72 °. To do. See FIG. 8 (3). Here, the dielectric resonator 2 measures the Pos1 region. This 72 ° rotation and measurement is repeated for the number of dielectric resonators in total. Therefore, in steps 2 and 3, the resonance frequency is measured for each of the five dielectric resonators for each of the five fan-shaped regions of the sample. That is, when the standard sample is rotated once to the original position, the process is completed.
[0025]
4). The standard sample shift amount of each dielectric resonator is calculated for each region from the difference between the resonance frequency measured for each dielectric resonator for all regions and the blank resonance frequency of each dielectric resonator.
5. The above results are shown in FIG. In FIG. 9, a spreadsheet such as the first stage is arranged using spreadsheet software. In this table, the data (standard sample shift amount) diagonally arranged from the upper left to the lower right means that each dielectric resonator measures the same region of the standard sample. For example, in the figure, the shaded area is data obtained by measuring each of the dielectric resonators in the PosB area. Similarly, the data along the line from the upper left to the lower right is another region of the standard sample, but is the data obtained by measuring the same region as viewed from each dielectric resonator.
6). Therefore, when the respective oblique data are rearranged vertically, a table in the second stage is obtained. In this table, since each column is data of the same region of the standard sample, all values should be the same if there is no individual difference of the dielectric resonator. Therefore, the average is calculated for each column.
7). By dividing this average by the data of each column, a value normalized by the average value is obtained, and when looking at each row, the values normalized for each dielectric resonator are arranged. If the average of each is taken, this becomes the correction coefficient of each dielectric resonator. This state is shown in the third stage table. The next fourth stage shows the data after the second stage table is corrected by the correction coefficients.
[0026]
Once the correction coefficient of each dielectric resonator is obtained, the orientation can be obtained in the same manner by the method described above. In the above example, the correction coefficient is obtained while changing the position in the rotation direction so that the standard sample is concentric on the measurement head constituted by a plurality of dielectric resonators arranged at different angles on the circumference.
[0027]
It is also possible to apply this concept to a measuring head constituted by a plurality of dielectric resonators arranged on a straight line at different angles. An example of such a measuring head is shown in FIG. In this case, the standard sample is obtained by repeating the measurement while shifting the strip-shaped sample for each measurement region of one dielectric resonator along the row of the dielectric resonators, and similarly obtaining the correction coefficient. At this time, in the case of the rotation type correction, all the dielectric resonators can naturally perform measurement for the same region by rotating the standard sample, but when the band sample is simply shifted, another measurement is performed after the first measurement. Areas will be added one after another. In this case, it is considered that a very good correction coefficient cannot be obtained. Therefore, in order to improve the defect, the band-shaped standard sample is equally divided for each region in the direction of shifting according to the number of dielectric resonators, and each region is connected with a tape or the like, and after one measurement. It is also possible to take a method of measuring by removing the tape and rearranging the areas in order. Such a measurement method is hereinafter referred to as a sequential correction method for simplicity. When expressed so as to include both the correction coefficient calculation methods of the sequential correction method and the first rotation type correction method, the following method is obtained.
[0028]
A correction coefficient calculation method comprising the following steps.
(Step 1) A blank is obtained by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on a measurement head including a plurality (n) of dielectric resonators disposed only on one side of the sample. Determining the resonance frequency.
(Step 2) A step of placing a substantially non-oriented standard sample covering the measurement surfaces of at least all dielectric resonators on the measurement head and measuring the resonance frequency of each dielectric resonator.
(Step 3) For the specific region (1st to nth region) of the standard sample determined for each dielectric resonator in Step 2, the sample is replaced at least (n−1) times, and all 1 to n Measuring a resonance frequency for each dielectric resonator with respect to the number region.
(Step 4) Each region (1st to nth region) from the difference between the resonance frequency measured for each dielectric resonator in all the 1st to nth regions and the blank resonance frequency of each dielectric resonator Calculating a standard sample shift amount of each dielectric resonator every time;
(Step 5) The average value of the standard sample shift amount of each dielectric resonator obtained in Step 4 is calculated for each region (1st to nth regions), and the average value is the standard of each dielectric resonator. Divide by the sample shift amount, calculate the standard sample shift amount of each dielectric resonator normalized for each region, and then calculate the standard sample shift amount of each normalized dielectric resonator for each dielectric resonance Calculating a correction coefficient for each dielectric resonator by averaging each unit.
[0029]
Furthermore, as described in the section of the simplified correction method, rewriting to mean both whether the correction coefficient is finally applied or divided by the correction coefficient is as follows.
A correction coefficient calculation method comprising the following steps.
(Step 1) A blank is obtained by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on a measurement head including a plurality (n) of dielectric resonators disposed only on one side of the sample. Determining the resonance frequency.
(Step 2) A step of placing a substantially non-oriented standard sample covering the measurement surfaces of at least all dielectric resonators on the measurement head and measuring the resonance frequency of each dielectric resonator.
(Step 3) For the specific region (1st to nth region) of the standard sample determined for each dielectric resonator in Step 2, the sample is replaced at least (n−1) times, and all 1 to n Measuring a resonance frequency for each dielectric resonator with respect to the number region.
(Step 4) Each region (1st to nth region) from the difference between the resonance frequency measured for each dielectric resonator in all the 1st to nth regions and the blank resonance frequency of each dielectric resonator Calculating a standard sample shift amount of each dielectric resonator every time;
(Step 5) The average value of the standard sample shift amount of each dielectric resonator obtained in Step 4 is calculated for each region (1st to nth regions), and the average value is the standard of each dielectric resonator. Calculate the standard sample shift amount of each dielectric resonator normalized for each region by calculating the sample shift amount, and then calculate the standard sample shift amount of each normalized dielectric resonator for each dielectric resonance. Calculating a correction coefficient for each dielectric resonator by averaging each unit.
[0030]
As another method of the rotational correction method, the correction coefficient can also be obtained by a method as described in FIG. This method is called a second rotational correction method. FIG. 11 is a diagram showing a state in which data is processed by the second rotational correction method using spreadsheet software. The procedure is shown below.
1. The blank resonance frequency is obtained by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on the measurement head.
2. A substantially non-oriented standard sample covering the measurement surfaces of at least all dielectric resonators is placed on the measurement head, and the resonance frequency of each dielectric resonator is measured. As shown in FIG. 8B, Pos1 represents the center line of the Pos1 region of the standard sample by a one-dot chain line, and the dielectric resonator 1 measures the Pos1 region. The same applies to each dielectric resonator hereinafter.
3. Measure the resonance frequency of each dielectric resonator in the same way after rotating the standard sample on the measuring head by a certain angle θ (θ = 360 ° / number of dielectric resonators), and in this case, rotating 72 °. To do. See FIG. 8 (3). Here, the dielectric resonator 2 measures the Pos1 region. This 72 ° rotation and measurement is repeated for the number of dielectric resonators in total. Therefore, in steps 2 and 3, the resonance frequency is measured for each of the five dielectric resonators for each of the five fan-shaped regions of the sample. That is, when the standard sample is rotated once to the original position, the process is completed.
[0031]
4). The standard sample shift amount of each dielectric resonator is calculated for each region from the difference between the resonance frequency measured for each dielectric resonator for all regions and the blank resonance frequency of each dielectric resonator.
5. The above results are shown in FIG. In FIG. 9, the table is arranged into a table like the first stage by using a spreadsheet. In this table, the data along the line drawn from the upper left to the lower right (standard sample shift amount) means that each dielectric resonator measures the same region of the standard sample. For example, in the figure, the shaded area is data obtained by measuring each of the dielectric resonators in the PosB area. Similarly, the data along the line from the upper left to the lower right is another region of the standard sample, but is the data obtained by measuring the same region as viewed from each dielectric resonator.
6). Therefore, when the respective oblique data are rearranged vertically, a table in the second stage is obtained. In this table, since each column is data of the same region of the standard sample, all the values should be the same unless there is a difference in the dielectric resonator. Therefore, the average is calculated for each column.
7). As shown in the third stage, by dividing this average by the data (standard sample shift amount) of each column, a value normalized by the average value: a normalized shift amount is obtained.
8). As shown in the fourth stage, a correction coefficient initial value for each dielectric resonator is appropriately determined in advance, and the correction coefficient initial value is first multiplied by the corresponding standard value of each dielectric resonator. Into a 5 × 5 matrix. Since the 25 data should be as close to 1.000 as possible, the operation is repeated while gradually changing the correction coefficient initial value of each dielectric resonator. This is to calculate a correction coefficient for each dielectric resonator that is finally determined to minimize the sum of squares of the difference between each data and 1.000 (25 square sums). The fifth stage shows a state where the square sum is calculated when the correction coefficient initial value is used. Finally, when the correction coefficient that minimizes the sum of squares is calculated, when this is the most overall, the 25 data have the same value (1 of the standardized shift amount normalized by the average value). .000). The next sixth stage shows data after the second stage table is corrected by the correction coefficients.
[0032]
Above, 8. This step will be described in a little more detail.
(1) The initial value (α of the correction coefficient αn (n = 1, 2, 3,...) Of each head._{I 1}, Α_{I 2}, Α_{I 3}... α_{I n}Decide
{Circle around (2)} The matrix in the third stage is n rows × n columns, multiplied by a correction coefficient for each row, and determined by the following equation (1)._ijAsk for.
X_ij= X_nj× α_In(N = 1,2,3, ... n, j = 1,2,3, ... n) (1)
(N × n) X_ijCalculate the following formula (2) for
[0033]
[Equation 3]

▲ 3 ▼ α_{I n}The same calculation is repeated while changing_kΑ when takes the smallest value_{L 1}... α_{L n}Ask for. Α at that time is a correction coefficient to be obtained. In other words, at this time, when n × n corrected standardized values are viewed as a whole, it is closest to 1. That is, the value in each region (row) has the smallest variation (n shift amounts are closest to the average value).
[0034]
The above example can also be applied to the sequential correction method as in the first rotational correction method. Furthermore, it can be rewritten to mean both whether the correction coefficient is finally applied or divided by the correction coefficient. This is expressed as follows.
A correction coefficient calculation method comprising the following steps.
(Step 1) A blank is obtained by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on a measurement head including a plurality (n) of dielectric resonators disposed only on one side of the sample. Determining the resonance frequency.
(Step 2) A step of placing a substantially non-oriented standard sample covering the measurement surfaces of at least all dielectric resonators on the measurement head and measuring the resonance frequency of each dielectric resonator.
(Step 3) A specific region of the standard sample determined for each dielectric resonator in Step 2 (1 to n-th region: suppose this is defined as n regions A, B, C... , C ...), and measuring the resonance frequency for each of the dielectric resonators for each of the 1st to nth regions by replacing the sample at least (n-1) times.
(Step 4) From the difference between the resonance frequency measured for each dielectric resonator in all the 1st to nth regions and the blank resonance frequency of each dielectric resonator, Calculating a standard sample shift amount;
(Step 5) The average value of the standard sample shift amount of each dielectric resonator obtained in Step 4 is calculated for each region (n regions of A, B, C...), And the average value and each dielectric Calculating a standard sample shift amount in each region of the body resonator, and calculating a standard sample shift amount of each dielectric resonator normalized for each region; (This value is the standardized shift amount.)
(Step 6) Arbitrary correction coefficient initial value (α that is predetermined for each dielectric resonator)_{I 1}, Α_{I 2}, Α_{I 3}... α_{I n}) Normalized shift amount (SV)_1A, SV_1B, SV_1C, ... SV_2A, SV_2B... SV_3A, SV_3B...), and the corrected standardized shift amount initial value (SV)_1A× α_{I 1}= X_{1 A}, SV_1B× α_{I 1}= X_1B,_SV1C× α_{I 1}= X_1C, ... SV_2A× α_{I 2}= X_{2 A}, SV_2B× α_{I 2}= X_{2 B}... SV_3A× α_{I 3}= X_{3 A}, SV_3B× α_{I 3}= X_{3 B}...)
The correction coefficient initial value (α in the following formula (1) for the initial value of the normalized shift amount after correction_{I 1}, Α_{I 2}, Α_{I 3}... α_{I n}) Gradually from each initial value to repeat the calculation, and a_kΑ, which is the final correction factor that minimizes_{L 1}... α_{L n}Calculating step.
[0035]
[Expression 4]

[0036]
The formulas in the above description are added for the purpose of explanation when the final correction coefficient is applied.
By adding the step of measuring the orientation as described in the simplified correction method using the correction coefficient obtained by any of the above methods, the orientation can be finally obtained. It is easy to program the calculation steps as described above by a microcomputer or the like to automate the calculation of the correction coefficient. In this case, when the measurement of the standard sample is not automated, a manual operation is added for each measurement.
If the orientation pattern can be measured online as described above, the measurement result is calculated by a computer, and parameters of the paper machine and stretching machine, such as J / W ratio (jet wire ratio) and stretching ratio (stretching roll rotational speed difference). Needless to say, the alignment pattern can be controlled on-line by adjusting the angle.
[0037]
【Example】
Example 1
Using the measuring head shown in FIG. 5, the circuit configuration shown in FIG. 6 was taken, and the measurement was performed according to the time chart as shown in FIG. FIG. 12 shows the result of applying the simple correction method. FIG. 12 shows an orientation pattern when the fiber orientation is measured when running a paper having a medium orientation degree at 300 m / min.
FIG. 13 shows data showing changes over time when the same sample is measured under the same conditions. The upper half shows the orientation angle (-90 to + 90 ° display) and the lower half shows the orientation degree (0 to 800 KHz), which is plotted every 5 seconds. It was also in good agreement with the results measured offline.
[0038]
Example 2
The measurement was performed by the same measurement system as in Example 1. FIG. 14 shows measurement data subjected to rotational correction. As the sample, paper having almost no orientation was used, and the sample while running at 300 m / min was measured in the same manner. The degree of orientation is 16 KHz, which is almost non-oriented data.
FIG. 15 shows data showing the change over time when the same sample was measured under the same conditions. Although it was almost non-oriented, it was measured stably, and the orientation angle was also in good agreement with the measurement result obtained by a conventional orientation measuring device offline.
[0039]
【The invention's effect】
As described above, according to the present invention, it is possible to accurately measure the orientation (orientation angle and orientation degree) even for a sample having a relatively small anisotropy.
[Brief description of the drawings]
1A is a plan view of a dielectric resonator, and FIG. 1B is a vertical sectional view thereof.
FIG. 2 is a plan view showing an example of a measurement head obtained by arranging five dielectric resonators in one metal case.
FIG. 3 is a diagram showing a shift amount of a resonance frequency between a blank (without a sample) and a sample measurement.
FIG. 4 is a diagram showing an example of an alignment pattern.
FIG. 5 is a view showing a state in which a standard sample is placed on a measurement head constituted by five dielectric resonators arranged by changing each angle = 72 ° and measurement is performed.
FIG. 6 is a block diagram of a circuit for processing signals from five dielectric resonators.
7 is a time chart showing signal processing in the block diagram shown in FIG. 6. FIG.
FIG. 8 is a diagram showing a sample measurement setup of a rotation type correction method.
FIG. 9 is a diagram showing a state in which data is processed by a first rotational correction method using spreadsheet software.
FIG. 10 is a diagram showing an example of a measurement head configured by a plurality of dielectric resonators arranged on a straight line at different angles.
FIG. 11 is a diagram showing a state in which data is processed by a second rotational correction method using spreadsheet software.
FIG. 12 is a diagram showing a display when a fiber orientation is measured when paper having a medium orientation degree is running at 300 m / min and processed by a simple correction method.
13 is a graph showing a change with time when the sample of FIG. 12 is measured. FIG.
FIG. 14 is a view showing a display when the fiber orientation is measured when paper having almost no orientation is running at 300 m / min and processed by the first rotational correction method.
15 is a diagram showing a change with time when the sample of FIG. 14 is measured.
[Explanation of symbols]
1 to 5: Dielectric resonator
15: Measuring head
16: Standard sample

Claims (4)

A correction coefficient calculation method comprising the following steps in an online orientation measurement method for obtaining the orientation of a sample to be measured.
(Step 1) The blank resonance frequency is determined by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on a measurement head including a plurality of dielectric resonators arranged only on one side of the sample. Step to seek.
(Step 2) for substantially non-oriented standard sample covers the measurement surface of at least all of the dielectric resonators, the step of measuring the resonant frequency of each dielectric resonator.
(Step 3) The standard sample shift amount is calculated from the difference in resonance frequency for each dielectric resonator obtained in Steps 1 and 2, and the standard sample shift amount of each obtained dielectric resonator is averaged. Calculating a correction coefficient for each dielectric resonator by calculating the obtained average value and the standard shift amount of each dielectric resonator; A correction coefficient calculation method comprising the following steps in an online orientation measurement method for obtaining the orientation of a sample to be measured.
(Step 1) A blank is obtained by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on a measurement head including a plurality (n) of dielectric resonators disposed only on one side of the sample. Determining the resonance frequency.
(Step 2) A step of measuring the resonance frequency of each dielectric resonator with respect to a standard sample covering the measurement surfaces of at least all the dielectric resonators.
(Step 3) For the specific region (1st to nth region) of the standard sample determined for each dielectric resonator in Step 2, the sample is replaced at least (n−1) times, and all 1 to n Measuring a resonance frequency for each dielectric resonator with respect to the number region.
(Step 4) Each region (1st to nth region) from the difference between the resonance frequency measured for each dielectric resonator in all the 1st to nth regions and the blank resonance frequency of each dielectric resonator Calculating a standard sample shift amount of each dielectric resonator every time;
(Step 5) The average value of the standard sample shift amount of each dielectric resonator obtained in Step 4 is calculated for each region (1st to nth region), and the average value and the standard of each dielectric resonator are calculated. Calculate the sample shift amount, calculate the standard sample shift amount of each dielectric resonator normalized for each region, and then calculate the standard sample shift amount of each normalized dielectric resonator for each dielectric resonance Calculating a correction coefficient for each dielectric resonator by averaging each unit. A correction coefficient calculation method comprising the following steps in an online orientation measurement method for obtaining the orientation of a sample to be measured.
(Step 1) The blank resonance frequency is determined by measuring the resonance frequency of each dielectric resonator in a state where nothing is placed on a measurement head including a plurality of dielectric resonators arranged only on one side of the sample. Step to seek.
(Step 2) A step of measuring the resonance frequency of each dielectric resonator with respect to a standard sample covering the measurement surfaces of at least all the dielectric resonators.
(Step 3) A specific region of the standard sample determined for each dielectric resonator in Step 2 (1 to n-th region: suppose this is defined as n regions A, B, C... , C ...), and measuring the resonance frequency for each of the dielectric resonators for each of the 1st to nth regions by replacing the sample at least (n-1) times.
(Step 4) From the difference between the resonance frequency measured for each dielectric resonator in all the 1st to nth regions and the blank resonance frequency of each dielectric resonator, Calculating a standard sample shift amount;
(Step 5) The average value of the standard sample shift amount of each dielectric resonator obtained in Step 4 is calculated for each region (n regions of A, B, C...), And the average value and each dielectric A step of calculating a standard sample shift amount in each region of the body resonator and calculating a standard sample shift amount of each dielectric resonator normalized for each region (this value is referred to as a normalized shift amount).
(Step 6) Arbitrary correction coefficient initial values (α _{I 1} , α _{I 2} , α _{I 3} ... Α _{I n} ) predetermined for each dielectric resonator are normalized shift amounts (SV _1A , SV _1B , SV). _{1C, ... SV 2A, SV 2B} ... SV 3A, calculated on the _{SV 3B} ...), the corrected normalization shift amount initial value _{(X 1 a, X 1B,} X 1C, ... X 2 a, X 2 B ... X 3 _A , X _{3 B} ...), And an equation (1) for the initial value of the normalized shift amount after correction

, The calculation is repeated by gradually changing the initial values from the correction coefficient initial values (α _{I 1} , α _{I 2} , α _{I 3} ... Α _{I n} ), and the correction coefficient is finally determined to minimize a _k. α _{L 1} ... Step for calculating α _{L n} . An on-line orientation measurement method for obtaining the orientation of a sample to be measured by further comprising the following steps in the correction coefficient calculation method according to claim 1, 2 or 3.
(Step 1) A step of measuring the resonance frequency of each dielectric resonator in a state where the measurement head is in contact with or close to the traveling measurement target sample.
(Step 2) A step of calculating the measurement sample shift amount from the difference between the resonance frequency of each dielectric resonator obtained in step 1 and the blank resonance frequency.
(Step 3) The correction factor of the dielectric resonator head according to claim 1 is calculated from the frequency shift amount obtained in Step 2 to obtain the corrected shift amount, and the orientation of the sample to be measured is obtained based on the calculated value. Step.

Images