JPH1031511A - Method and device for predicting abnormal fluctuation - Google Patents
Method and device for predicting abnormal fluctuationInfo
- Publication number
- JPH1031511A JPH1031511A JP18478796A JP18478796A JPH1031511A JP H1031511 A JPH1031511 A JP H1031511A JP 18478796 A JP18478796 A JP 18478796A JP 18478796 A JP18478796 A JP 18478796A JP H1031511 A JPH1031511 A JP H1031511A
- Authority
- JP
- Japan
- Prior art keywords
- change amount
- frequency component
- average value
- complex number
- real part
- 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.)
- Granted
Links
Abstract
Description
【０００１】[0001]
【発明の属する技術分野】本発明は、半導体等の製造時
に測定対象とするデータの値が急激に変動することを予
測できる異常変動予測方法及び異常変動予測装置に関す
る。BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an abnormal fluctuation prediction method and an abnormal fluctuation prediction apparatus capable of predicting a sudden change in the value of data to be measured during the manufacture of a semiconductor or the like.
【０００２】[0002]
【従来の技術】監視対象の状態を示すデータがどのよう
に変動するかを予測する方法において、その変動を近似
直線を用いて表わせないほど複雑であり、且つ、変動の
要因が判明していない場合は、測定時の変動と類似した
過去の変動を用いて、近い未来も類似した過去の変動と
同じ値になると推測する予測方法が有効である。2. Description of the Related Art A method for predicting how data indicating the state of a monitored object fluctuates is so complicated that the fluctuation cannot be expressed using an approximate line, and the cause of the fluctuation is not known. In this case, it is effective to use a past variation similar to the variation at the time of measurement, and to use a prediction method that estimates that the near future will have the same value as the similar past variation.
【０００３】その１つの方法として、公開番号 特開平
０８−０１６２０９「予測制御方法及びその装置」があ
る。One of the methods is disclosed in Japanese Patent Application Laid-Open No. 08-016209 "Predictive control method and apparatus".
【０００４】以下、前記従来の予測制御方法及びその装
置を図面を参照しながら概要を説明する。The outline of the conventional predictive control method and its apparatus will be described below with reference to the drawings.
【０００５】図７は従来の予測制御装置を示す機能ブロ
ック図である。図７に示すように、半導体製造装置１０
０は制御対象であり、測定器１０１は半導体製造装置１
００の状態等を示すパラメータを測定する。パターン抽
出手段１０２はパラメータ値を外部記憶装置１０６に記
録すると共に変動しているパラメータ値の変化率等を抽
出する。類似度算出手段１０３は現在測定しているパラ
メータ値の変動が過去の変動と比べてどのくらい類似し
ているかを算出する。評価手段１０４は類似度と測定値
及び予測値の差とを比較して、予測値を求めるために使
用したデータが適切であるかを判断し、不適当なデータ
を外部記憶装置１０６に記録する。また、現在測定して
いるパラメータの変動と類似した過去のデータとから半
導体製造装置１００の状態を予測する。制御器１０５は
評価手段１０４の予測結果に基づいて半導体製造装置１
００を制御する。FIG. 7 is a functional block diagram showing a conventional predictive control device. As shown in FIG.
0 is a control target, and the measuring instrument 101 is a semiconductor manufacturing apparatus 1
A parameter indicating the state of 00 and the like is measured. The pattern extracting means 102 records the parameter values in the external storage device 106 and extracts the rate of change of the changing parameter values. The similarity calculation means 103 calculates how similar the variation of the parameter value currently measured is compared to the past variation. The evaluation unit 104 compares the similarity with the difference between the measured value and the predicted value, determines whether the data used for obtaining the predicted value is appropriate, and records the inappropriate data in the external storage device 106. . Further, the state of the semiconductor manufacturing apparatus 100 is predicted from the past data similar to the variation of the parameter currently measured. The controller 105 controls the semiconductor manufacturing apparatus 1 based on the prediction result of the
00 is controlled.
【０００６】以上のように構成された予測制御装置は、
過去に測定したパラメータの変動パターンを蓄積し、蓄
積している変動パターンの中から、測定中のパラメータ
の変動と類似しているものを抽出し、該測定中のパラメ
ータが、近い未来に、抽出した変動パターンが示すとお
りに変動すると予測し、その予測結果を用いて半導体製
造装置１００を制御する。[0006] The predictive control device configured as described above comprises:
The variation patterns of the parameters measured in the past are accumulated, and those similar to the variation of the parameter being measured are extracted from the accumulated variation patterns, and the parameter being measured is extracted in the near future. The semiconductor manufacturing apparatus 100 is predicted to fluctuate as indicated by the obtained fluctuation pattern, and the semiconductor manufacturing apparatus 100 is controlled using the prediction result.
【０００７】[0007]
【発明が解決しようとする課題】しかしながら、前記従
来の予測制御装置は、過去に測定したデータの変動パタ
ーンが少ない場合は、測定中のデータが示す変動と類似
しない場合が多いため、近い未来の変動を予測できない
という問題を有していた。従って、比較対象の変動パタ
ーンが十分に得られない環境下では、比較対象の変動パ
ターンが少ない場合でも予測できる方法、又は比較対象
の変動パターンを必要としない予測方法が必要となる。However, in the conventional predictive control device, when the fluctuation pattern of the data measured in the past is small, the fluctuation is often not similar to the fluctuation indicated by the data being measured. There was a problem that fluctuation could not be predicted. Therefore, in an environment where the variation pattern of the comparison target is not sufficiently obtained, a method that can predict even when the variation pattern of the comparison target is small or a prediction method that does not require the variation pattern of the comparison target is required.
【０００８】また、監視対象の状態を示すデータを管理
する場合は、監視対象の状態が大きく変化して問題が発
生することを未然に防ぐことが重要であり、小さな変化
を予測しても問題を未然に防止する効果は極めて小さ
い。Further, when managing data indicating the state of a monitoring target, it is important to prevent a state of the monitoring target from greatly changing and causing a problem beforehand. Is very small.
【０００９】本発明は、前記従来の問題を解決し、近い
未来における監視対象の状態を予測する際に比較対象の
変動パターンを用いることなく、近い未来に監視対象の
状態が大きく変化することを予測できるようにすること
を目的とする。The present invention solves the above-mentioned conventional problem, and makes it possible to predict that the state of a monitoring target changes greatly in the near future without using a fluctuation pattern of a comparison target when predicting the state of the monitoring target in the near future. The aim is to be predictable.
【００１０】[0010]
【課題を解決するための手段】請求項１の発明が講じた
解決手段は、測定対象とするデータの値が大きく変化す
ることを予測する異常変動予測方法を対象とし、測定値
が時間とともに変化する時系列データにおける変動周期
の周波数成分を算出する周波数成分算出工程と、前記周
波数成分の変化量を検出し、検出した周波数成分の変化
量である検出変化量があらかじめ設定した変化量である
許容変化量の範囲内であるか否かを判断し、前記検出変
化量が前記許容変化量の範囲内にないときには、前記測
定値が大きく変動する兆候であると判定する異常変動検
出工程とを備えている構成とするものである。Means for Solving the Problems According to a first aspect of the present invention, there is provided a method for predicting an abnormal fluctuation in which a value of data to be measured is largely changed, wherein a measured value changes with time. A frequency component calculating step of calculating a frequency component of a fluctuation cycle in the time-series data to be detected, and detecting a change amount of the frequency component, and detecting a change amount of the detected frequency component as a predetermined change amount. Determining whether the change is within a range of the change amount, and when the detected change amount is not within the range of the allowable change amount, determining an abnormal change detection step of determining that the measured value is a sign of a large change. Configuration.
【００１１】請求項１の構成により、測定値が時間とと
もに変化する時系列データにおける変動周期の周波数成
分を算出した後、周波数成分の変化量を検出し、検出変
化量が許容変化量の範囲内であるか否かを判断し、検出
変化量が許容変化量の範囲内にないときには、測定値が
大きく変動する兆候であると判定するため、比較対象と
する変動パターンを用いる必要がない。According to the first aspect of the present invention, after calculating the frequency component of the fluctuation period in the time-series data in which the measured value changes with time, the change amount of the frequency component is detected, and the detected change amount is within the allowable change amount. Is determined, and when the detected change amount is not within the range of the allowable change amount, it is determined that the measured value is a sign of a large change, so that there is no need to use a change pattern to be compared.
【００１２】請求項２の発明は、請求項１の構成に、前
記周波数成分算出工程は、前記時系列データをフーリエ
変換することにより導出された複素数の実部及び虚部の
値よりなる前記周波数成分を算出する工程を含む構成を
付加するものである。According to a second aspect of the present invention, in the configuration of the first aspect, the frequency component calculating step includes the step of performing the Fourier transform of the time series data, wherein the frequency component comprises values of a real part and an imaginary part of a complex number. A configuration including a step of calculating a component is added.
【００１３】請求項３の発明は、請求項２の構成に、前
記異常変動検出工程は、前記複素数の実部の平均値及び
虚部の平均値を用いて前記周波数成分の変化量を検出す
る工程を含む構成を付加するものである。According to a third aspect of the present invention, in the configuration of the second aspect, the abnormal fluctuation detecting step detects the amount of change in the frequency component using an average value of a real part and an average value of an imaginary part of the complex number. A configuration including a step is added.
【００１４】請求項４の発明は、請求項３の構成に、前
記異常変動検出工程は、前記複素数の実部の平均値と虚
部の平均値とにより表わされる２次元の座標系を導入
し、前記許容変化量の範囲を前記座標系における長円形
の円周及び内部よりなる領域として表わす工程を含み、
前記長円形の長軸は、前記複素数の実部及び虚部の各平
均値から算出された近似直線に平行であり、前記長軸か
ら垂直方向への前記長円形の円周までの距離は、前記長
軸に垂直な方向に対して前記複素数の実部及び虚部の各
平均値が示す位置の分散値に依存する構成を付加するも
のである。According to a fourth aspect of the present invention, in the configuration of the third aspect, the abnormal fluctuation detecting step introduces a two-dimensional coordinate system represented by an average value of a real part and an average value of an imaginary part of the complex number. Representing the range of the permissible variation as a region consisting of the circumference and the interior of an oval in the coordinate system,
The major axis of the oval is parallel to the approximate straight line calculated from the average value of the real part and the imaginary part of the complex number, and the distance from the major axis to the circumference of the oval in the vertical direction is: A configuration that depends on a variance value of a position indicated by each average value of the real part and the imaginary part of the complex number in a direction perpendicular to the major axis is added.
【００１５】請求項５の発明が講じた解決手段は、異常
変動予測装置を、測定値が時間とともに変化する時系列
データをフーリエ変換することにより導出された複素数
の実部及び虚部の値を用いて前記時系列データにおける
変動周期の周波数成分を算出する周波数成分算出手段
と、前記複素数の実部の平均値及び虚部の平均値を用い
て前記周波数成分の変化量を検出し、検出した周波数成
分の変化量である検出変化量があらかじめ設定した変化
量である許容変化量の範囲内であるか否かを判断し、前
記検出変化量が前記許容変化量の範囲内にないときに
は、測定値が大きく変動する兆候であると判定する異常
変動検出手段とを備え、前記異常変動検出手段におい
て、前記複素数の実部の平均値と虚部の平均値とにより
表わされる２次元の座標系を導入され、前記許容変化量
の範囲を前記座標系における長円形の円周及び内部より
なる領域として表わされると共に、前記長円形の長軸
は、前記複素数の実部及び虚部の各平均値から算出され
た近似直線に平行であり、前記長軸から垂直方向への前
記長円形の円周までの距離は、前記長軸に垂直な方向に
対して前記複素数の実部及び虚部の各平均値が示す位置
の分散値に依存する構成とするものである。According to a fifth aspect of the present invention, there is provided a means for predicting an abnormal fluctuation, wherein a value of a real part and an imaginary part of a complex number derived by performing a Fourier transform on time series data whose measured values change with time is calculated. Frequency component calculating means for calculating the frequency component of the fluctuation period in the time-series data using the average value of the real part and the average value of the imaginary part of the complex number, the amount of change in the frequency component was detected and detected. It is determined whether or not the detected change amount that is the change amount of the frequency component is within a range of an allowable change amount that is a preset change amount, and when the detected change amount is not within the range of the allowable change amount, the measurement is performed. Abnormal fluctuation detecting means for determining that the value is a sign of a large fluctuation, wherein the abnormal fluctuation detecting means includes two-dimensional coordinates represented by an average value of a real part and an average value of an imaginary part of the complex number. Is introduced, the range of the permissible variation is represented as an area consisting of the circumference and the interior of the oval in the coordinate system, and the major axis of the oval is the average value of the real part and the imaginary part of the complex number. Is parallel to the approximate straight line calculated from, and the distance from the major axis to the circumference of the oval in the vertical direction is the real part and the imaginary part of the complex number with respect to the direction perpendicular to the major axis. The configuration depends on the variance value of the position indicated by the average value.
【００１６】請求項５の構成により、測定値が時間とと
もに変化する時系列データをフーリエ変換することによ
り導出された複素数の実部及び虚部の値を用いて前記時
系列データにおける変動周期の周波数成分を算出した
後、複素数の実部の平均値及び虚部の平均値を用いて周
波数成分の変化量を検出し、検出変化量が許容変化量の
範囲内であるか否かを判断し、検出変化量が許容変化量
の範囲内にないときには、測定値が大きく変動する兆候
であると判定するため、比較対象とする変動パターンを
用いる必要がない。According to the fifth aspect of the present invention, the frequency of the fluctuation period in the time series data is obtained by using the values of the real part and the imaginary part of the complex number derived by subjecting the time series data whose measured value changes with time to Fourier transform. After calculating the components, the change amount of the frequency component is detected using the average value of the real part and the average value of the imaginary part of the complex number, and it is determined whether the detected change amount is within the range of the allowable change amount, When the detected change amount is not within the range of the allowable change amount, it is determined that the measured value is a sign of a large change, so that there is no need to use a change pattern to be compared.
【００１７】請求項６の発明が講じた解決手段は、測定
対象のデータの値が大きく変化することを予測するプロ
グラム製品を対象とし、以下の手順を実現するプログラ
ム記録媒体を含む：測定値が時間とともに変化する時系
列データをフーリエ変換することにより導出された複素
数の実部及び虚部の値を用いて前記時系列データにおけ
る変動周期の周波数成分を算出する周波数成分算出手段
と、前記複素数の実部の平均値及び虚部の平均値を用い
て前記周波数成分の変化量を検出し、検出した周波数成
分の変化量である検出変化量があらかじめ設定した変化
量である許容変化量の範囲内であるか否かを判断し、前
記検出変化量が前記許容変化量の範囲内にないときに
は、測定値が大きく変動する兆候であると判定する異常
変動検出手段とを備え、前記異常変動検出手段におい
て、前記複素数の実部の平均値と虚部の平均値とにより
表わされる２次元の座標系を導入され、前記許容変化量
の範囲を前記座標系における長円形の円周及び内部より
なる領域として表わすと共に、前記長円形の長軸は、前
記複素数の実部及び虚部の各平均値から算出された近似
直線に平行であり、前記長軸から垂直方向への前記長円
形の円周までの距離は、前記長軸に垂直な方向に対して
前記複素数の実部及び虚部の各平均値が示す位置の分散
値に依存する構成とするものである。According to a sixth aspect of the present invention, there is provided a solution for a program product which predicts that a value of data to be measured is largely changed, and includes a program recording medium which realizes the following procedure: Frequency component calculating means for calculating a frequency component of a fluctuation period in the time-series data using values of a real part and an imaginary part of a complex number derived by performing Fourier transform on time-series data that changes with time; and Using the average value of the real part and the average value of the imaginary part, the amount of change in the frequency component is detected, and the detected amount of change in the detected frequency component is within a range of an allowable amount of change that is a predetermined amount of change. Abnormal fluctuation detecting means for judging whether the detected change amount is not within the range of the allowable change amount, and determining that the measured value is a sign of a large change. In the abnormal variation detecting means, a two-dimensional coordinate system represented by an average value of a real part and an average value of an imaginary part of the complex number is introduced, and the range of the allowable variation is defined by an oval circle in the coordinate system. The long axis of the oval is parallel to an approximate straight line calculated from the average value of the real part and the imaginary part of the complex number, and is expressed in a vertical direction from the long axis. The distance to the circumference of the oval depends on the variance of the position indicated by each average value of the real part and the imaginary part of the complex number in a direction perpendicular to the long axis.
【００１８】[0018]
【発明の実施の形態】本発明の一実施形態に係る異常変
動予測方法及びその装置を図面を参照しながら説明す
る。本実施形態において、ＣＶＤ（Ｃｈｅｍｉｃａｌ
Ｖａｐｏｒ Ｄｅｐｏｓｉｔｉｏｎ）装置内に発生する
ダスト等の異物の個数（以下、異物数と記す。）の異常
増加に関する予測を一例として挙げる。DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An abnormal fluctuation prediction method and apparatus according to one embodiment of the present invention will be described with reference to the drawings. In the present embodiment, a CVD (Chemical
As an example, a prediction relating to an abnormal increase in the number of foreign substances such as dust (hereinafter, referred to as the number of foreign substances) generated in a Vapor Deposition apparatus will be described.
【００１９】図１は本発明の一実施形態に係る異常変動
予測方法を用いた異常変動予測装置を示す機能ブロック
図である。図１に示すように、本装置は、ＣＶＤ装置１
０において製造中のウェハの表面に付着した異物数を検
出する測定器１１と、異物数等を蓄積する外部記憶装置
１２と、異物数が急激に増加することを予測するための
異物数の変動を解析する解析装置１３と、解析装置１３
が出力する予測結果を表示する表示手段１４とを備え、
解析装置１３は、異物数の時間的変動を示す時系列デー
タを作成する時系列データ作成手段１３ａと、時系列デ
ータが示す変動の周波数成分を算出する周波数成分算出
手段１３ｂと、異物数が大きく増加する兆候として周波
数成分の特徴的な変化を検出する異常変動検出手段１３
ｃとを有している。FIG. 1 is a functional block diagram showing an abnormal fluctuation prediction device using an abnormal fluctuation prediction method according to one embodiment of the present invention. As shown in FIG. 1, this apparatus is a CVD apparatus 1
0, a measuring device 11 for detecting the number of foreign substances attached to the surface of the wafer being manufactured, an external storage device 12 for storing the number of foreign substances, and a change in the number of foreign substances for estimating that the number of foreign substances increases rapidly. Analysis device 13 for analyzing the
And display means 14 for displaying the prediction result output by
The analysis device 13 includes a time-series data creation unit 13a that creates time-series data indicating a temporal change in the number of foreign objects, a frequency component calculation unit 13b that calculates a frequency component of the change indicated by the time-series data, and a large number of foreign objects. Abnormal fluctuation detecting means 13 for detecting a characteristic change of a frequency component as a sign of increase
c.
【００２０】以下、前記のように構成された異常変動予
測装置の動作を図面を参照しながら説明する。図１に示
すように、時系列データ作成手段１３ａは、測定器１１
から新たな測定値を抽出する度に、抽出された測定値を
含めた時系列データを作成し、周波数成分算出手段１３
ｂ及び異常変動検出手段１３ｃは、時系列データ作成手
段１３ａが作成した時系列データを用いて近い未来に測
定値が大きく変動するか否かを予測する。Hereinafter, the operation of the abnormal fluctuation prediction apparatus configured as described above will be described with reference to the drawings. As shown in FIG. 1, the time-series data creating means 13a
Each time a new measurement value is extracted from the data, time-series data including the extracted measurement value is created, and the frequency component calculation means 13
b and the abnormal fluctuation detecting means 13c predict whether or not the measured value will fluctuate greatly in the near future using the time series data created by the time series data creating means 13a.
【００２１】図２は時系列データの一例であり、図２に
示すように、時系列データ作成手段１３ａは、測定器１
１から異物数を抽出する度に、異物数と測定時の累積処
理ウェハ枚数とを対にした２次元の配列からなる時系列
データを作成する。２０は異物数を格納する異物数格納
フィールドであり、２１は累積処理ウェハ枚数を格納す
る累積処理ウェハ枚数格納フィールドである。FIG. 2 shows an example of the time-series data. As shown in FIG.
Each time the number of foreign particles is extracted from No. 1, time-series data consisting of a two-dimensional array in which the number of foreign particles and the cumulative number of wafers processed during measurement are paired is created. Reference numeral 20 denotes a foreign object number storage field for storing the number of foreign objects, and reference numeral 21 denotes an accumulated processed wafer number storage field for storing the accumulated processed wafer number.
【００２２】以下、時系列データの作成方法を説明す
る。時系列データ作成手段１３ａは、時系列データを作
成する度に、時系列データを外部記憶装置１２に格納し
ており、測定器１１から新たに測定された異物数を収集
すると、その異物数と測定時の累積処理ウェハ枚数とを
対にしたデータを外部記憶装置１２から抽出した時系列
データの最後に追加している。Hereinafter, a method of creating time-series data will be described. The time-series data creating means 13a stores the time-series data in the external storage device 12 every time the time-series data is created. Data corresponding to the cumulative number of processed wafers at the time of measurement is added to the end of the time-series data extracted from the external storage device 12.
【００２３】図３及び図４はＣＶＤ装置内において発生
した異物数を測定し時系列データ作成手段１３ａにより
作成された時系列データを表わすトレンドグラフであっ
て、累積処理ウェハ枚数に対する異物数の変動を表わし
ており、グラフの横軸は累積処理ウェハ枚数、縦軸はウ
ェハの表面に付着した異物数を表わしている。FIG. 3 and FIG. 4 are trend graphs showing the time series data created by the time series data creation means 13a by measuring the number of foreign matters generated in the CVD apparatus. The horizontal axis of the graph represents the cumulative number of processed wafers, and the vertical axis represents the number of foreign particles attached to the surface of the wafer.
【００２４】ここで、累積処理ウェハ枚数は、ＣＶＤ装
置１０で処理したウェハの枚数を、ＣＶＤ装置１０内を
掃除して内壁に付着している異物を取り除いた時点から
数えた値であり、ＣＶＤ装置１０内を掃除するたびに累
積処理ウェハ枚数は０に戻る。関数ｆτ（ｔ）は、累積
処理ウェハ枚数ｔに対する異物数を示した非連続な関数
であり、累積処理ウェハ枚数が０枚からτ枚における異
物数の変動を示している。すなわち、図３（ａ）に示す
関数ｆτ（ｔ）は累積処理ウェハ枚数ｔ＝０からｔ＝ｔ
a までに測定した異物数の変動を、図３（ｂ）に示す関
数ｆτ（ｔ）は累積処理ウェハ枚数ｔ＝ｔb までに測定
した異物数の変動を、図３（ｃ）に示す関数ｆτ（ｔ）
は累積処理ウェハ枚数ｔ＝ｔc までに測定した異物数の
変動を、図４（ａ）に示す関数ｆτ（ｔ）は累積処理ウ
ェハ枚数ｔ＝ｔd までに測定した異物数の変動を、図４
（ｂ）に示す関数ｆτ（ｔ）は累積処理ウェハ枚数ｔ＝
ｔe までに測定した異物数の変動をそれぞれ表わしてい
る。Here, the cumulative number of processed wafers is a value obtained by counting the number of wafers processed by the CVD apparatus 10 from the point in time when the inside of the CVD apparatus 10 is cleaned to remove foreign matters adhering to the inner wall. Each time the inside of the apparatus 10 is cleaned, the cumulative number of processed wafers returns to zero. The function fτ (t) is a discontinuous function indicating the number of foreign particles with respect to the cumulative number of processed wafers t, and indicates a change in the number of foreign particles when the cumulative number of processed wafers is 0 to τ. That is, the function fτ (t) shown in FIG.
The function fτ (t) shown in FIG. 3 (b) is a function fτ (t) shown in FIG. 3 (b), and the function fτ (t) shown in FIG. (T)
FIG. 4A shows the variation in the number of foreign particles measured up to the cumulative number of processed wafers t = tc, and the function fτ (t) shown in FIG.
The function fτ (t) shown in FIG.
The fluctuations in the number of foreign substances measured up to te are respectively shown.
【００２５】次に、周波数成分算出手段１３ｂは、時系
列データが作成される度に、その時系列データが示す異
物数の変動の周波数成分を算出する。すなわち、［数
１］に示すように、時系列データに対応した関数ｆτ
（ｔ）に対して離散フーリエ変換を行なう。Next, each time the time series data is created, the frequency component calculation means 13b calculates the frequency component of the change in the number of foreign substances indicated by the time series data. That is, as shown in [Equation 1], the function fτ corresponding to the time-series data
Perform discrete Fourier transform on (t).
【００２６】[0026]
【数１】 (Equation 1)
【００２７】ここで、Ｆτ（ｍω0 ）はｆτ（ｔ）のフ
ーリエ積分関数、ｍは高調波の番号、ｘτ（ｍ）はｍ高
調波の実部、ｙτ（ｍ）は第ｍ高調波の虚部、ωは周波
数、Ｔはサンプリング周期、ω0 は１／Ｔ、ｍとｋは０
以上のシーケンス番号を表わす。第ｍ高調波の周波数は
ｍω0 である。また、ｘτ（ｍ）とｙτ（ｍ）とは、次
の関係を有する周波数成分であって、｛ｘτ（ｍ）2 ＋
ｙτ（ｍ）2 ｝1/2 は第ｍ高調波のスペクトル強度を表
わすと共に、ｔａｎ-1｛ｙτ（ｍ）／ｘτ（ｍ）｝は第
ｍ高調波の位相を表わす。Where Fτ (mω 0 ) is the Fourier integral function of fτ (t), m is the harmonic number, xτ (m) is the real part of the m harmonic, and yτ (m) is the m-th harmonic. Imaginary part, ω is frequency, T is sampling period, ω 0 is 1 / T, m and k are 0
These represent the above sequence numbers. The frequency of the m-th harmonic is mω 0 . Further, xτ (m) and yτ (m) are frequency components having the following relationship, and ｛xτ (m) 2 +
yτ (m) 2 ｝ 1/2 represents the spectral intensity of the m-th harmonic, and tan −1 {yτ (m) / xτ (m)} represents the phase of the m-th harmonic.
【００２８】なお、ｆτ（ｔ）に［数１］を適用すると
きに、サンプリングデータが不足している場合は、離散
的な実データの間を最小自乗法を用いて補間した数値を
用いる。When applying [Equation 1] to fτ (t), if sampling data is insufficient, a value obtained by interpolating discrete real data using the least squares method is used.
【００２９】次に、異常変動検出手段１３ｃは、周波数
成分算出手段１３ｂにより求められた全高調波の実部及
び虚部の平均値を外部記憶装置１２に記録する。ただ
し、ＣＶＤ装置１０内の掃除が行なわれ、累積処理ウェ
ハ枚数が前回よりも少なくなった場合（例えば０枚）
は、外部記憶装置１２に蓄積してきた実部及び虚部の平
均値を削除した後、周波数成分算出手段１３ｂにより求
められた全高調波の実部及び虚部の平均値を外部記憶装
置１２にそのまま記録する。Next, the abnormal fluctuation detecting means 13c records the average value of the real part and the imaginary part of all harmonics obtained by the frequency component calculating means 13b in the external storage device 12. However, when the inside of the CVD apparatus 10 is cleaned and the cumulative number of processed wafers becomes smaller than the previous number (for example, 0).
Deletes the average value of the real part and the imaginary part accumulated in the external storage device 12 and then stores the average value of the real part and the imaginary part of all the harmonics obtained by the frequency component calculation means 13b in the external storage device 12. Record as it is.
【００３０】その後、これまでに外部記憶装置１２に蓄
積してきた実部及び虚部の平均値を用いて異物数が異常
に増加する兆候を検出し、検出した場合には表示手段１
４に近い未来に異物数が異常に増加する旨を表示する。Thereafter, using the average value of the real part and the imaginary part accumulated in the external storage device 12 until then, a sign indicating an abnormal increase in the number of foreign substances is detected.
A message indicating that the number of foreign substances abnormally increases in the near future of 4 is displayed.
【００３１】以下、周波数成分の実部及び虚部の平均値
の変化において、異物数が異常に増加する兆候を示す特
徴的な変化を図３、図４及び図５を用いて説明する。A characteristic change indicating a sign of an abnormal increase in the number of foreign substances in the change of the average value of the real part and the imaginary part of the frequency component will be described below with reference to FIGS. 3, 4 and 5.
【００３２】図５は周波数成分の累積処理ウェハ枚数の
増加に伴って変化する実部及び虚部の平均値を示す散布
図である。散布図の横軸は実部の平均値を示し、縦軸は
虚部の平均値を示す。各プロットは、異物数を測定する
度に作成した時系列データから［数１］を用いて算出さ
れた実部及び虚部の平均値を示す。また、累積処理ウェ
ハ枚数の昇順に各プロットを曲線でトレースしており、
プロットに添えて記述されているτ＝ｔa ，・・・，ｔ
e は、図３及び図４に示すτ＝ｔa 、ｔb 、ｔc 、ｔd
、ｔe のときの時系列データから算出された値のプロ
ットであることを表わしている。FIG. 5 is a scatter diagram showing the average value of the real part and the imaginary part which change with the increase in the number of frequency component accumulated wafers. The horizontal axis of the scatter diagram shows the average value of the real part, and the vertical axis shows the average value of the imaginary part. Each plot shows the average value of the real part and the imaginary part calculated using [Equation 1] from the time series data created each time the number of foreign substances is measured. Also, each plot is traced by a curve in ascending order of the cumulative number of processed wafers,
Τ = ta,..., T described along with the plot
e is τ = ta, tb, tc, td shown in FIGS.
, Te, plots the values calculated from the time series data.
【００３３】図３、図４及び図５に示す各グラフを解析
すると次のことが分かる。Analysis of the graphs shown in FIGS. 3, 4 and 5 reveals the following.
【００３４】すなわち、図４（ａ）に示すτ＜ｔd の領
域において、異物数は大きな変動を示しておらず、τ＜
ｔd の領域に対応する図５のプロットの軌跡は長円形の
ような軌道を描く曲線である。しかし、異物数が大きく
増加する図４（ｂ）に示すτ≧ｔd の領域に対応した図
５のプロットは、長円形が占める領域になく、該領域か
ら外れた値となる。That is, in the region of τ <td shown in FIG.
The locus of the plot of FIG. 5 corresponding to the region of td is a curve that draws a locus like an oval. However, the plot of FIG. 5 corresponding to the region of τ ≧ td shown in FIG. 4B where the number of foreign substances greatly increases is not in the region occupied by the oval, and is a value outside the region.
【００３５】従って、図５のプロットの軌跡が描く長円
形のような領域を許容変化量として設定し、この許容変
化量を示す領域から最初に外れたデータを検出すること
により、異物数が大きく増加することを予測することが
可能になる。Therefore, an area such as an oval drawn by the locus of the plot in FIG. 5 is set as an allowable change amount, and data which first deviates from the area indicating the allowable change amount is detected, thereby increasing the number of foreign substances. It is possible to predict the increase.
【００３６】以下、図５のプロットの軌跡が描く長円形
のような領域から最初に外れたデータを検出する方法を
図６を用いて説明する。図６は長円形の領域から外れた
プロットを検出するために用いる長円形３０を図５に追
加した散布図である。長円形３０は、該長円形３０の大
きさ、位置又は向きが実部及び虚部にそれぞれ新しい平
均値が算出される度に変化する。長円形３０の長軸は、
実部と虚部との平均値から算出した近似直線に平行であ
り、長軸から垂直方向に位置する長円形３０の円周まで
の距離は、長軸に垂直な方向に対するプロットの分散に
依存する。Hereinafter, a method for detecting data that first deviates from an oval-like region drawn by the locus of the plot in FIG. 5 will be described with reference to FIG. FIG. 6 is a scatter diagram in which an oval 30 used for detecting a plot deviating from the oval region is added to FIG. The size, position, or orientation of the oval 30 changes each time a new average value is calculated for the real part and the imaginary part. The major axis of the oval 30 is
The distance from the major axis to the circumference of the oval 30 located in the vertical direction, which is parallel to the approximate straight line calculated from the average value of the real part and the imaginary part, depends on the variance of the plot in the direction perpendicular to the major axis. I do.
【００３７】図６に示す長円形３０は、周波数成分算出
手段１３ｂがτ＝ｔd のプロットが示す値を求めた後
に、異常変動検出手段１３ｃが用いる長円形である。The ellipse 30 shown in FIG. 6 is an ellipse used by the abnormal fluctuation detecting means 13c after the frequency component calculating means 13b obtains the value indicated by the plot of τ = td.
【００３８】このように、異常変動検出手段１３ｃは、
最新のτに対応したプロットが長円形３０が占める領域
から外れているか否かを調べ、該長円形３０が占める許
容変化領域から大きく外れるプロットを検出したとき
に、近い未来に異物数が大きく増加すると予測する。As described above, the abnormal fluctuation detecting means 13c
It is checked whether or not the plot corresponding to the latest τ is out of the area occupied by the oval 30. When a plot that greatly deviates from the allowable change area occupied by the oval 30 is detected, the number of foreign substances greatly increases in the near future. Then predict.
【００３９】以上説明したように、本実施形態による
と、異物数が変動する周波数成分を離散フーリエ変換を
用いて算出し、異物数が大きく増加する兆候を示す周波
数成分の変化を検出することにより、異物数の異常な増
加を予測することができる。As described above, according to the present embodiment, the frequency component in which the number of foreign matters fluctuates is calculated using the discrete Fourier transform, and the change in the frequency component showing a sign that the number of foreign matters greatly increases is detected. In addition, an abnormal increase in the number of foreign substances can be predicted.
【００４０】なお、本実施形態において、異常変動検出
手段１３ｃは異物数が異常に増加する兆候を示すデータ
を検出するために長円形３０を用いたが、長円形３０の
代わりに、あらかじめ設定した値以上にプロットが密集
している領域を用いて、その領域外に最初に現われるτ
に対応するプロットを検出してもよい。In this embodiment, the abnormal fluctuation detecting means 13c uses the oval 30 to detect data indicating a sign of an abnormal increase in the number of foreign substances. Using the region where the plots are denser than the value, τ that first appears outside that region
May be detected.
【００４１】また、本実施形態はＣＶＤ装置１０内にお
いて発生する異物数を予測する場合を説明したが、ＣＶ
Ｄ装置１０以外の洗浄装置、スパッタ装置等における異
物数であってもよく、異物数以外のエッチングレートな
どであっても予測することができる。また、半導体製造
に関わるデータ以外に、周期的な変動を示す多くの要因
が重なり合って観測されるデータに対しても、データ値
が急激に大きくなることを予測することができる。In this embodiment, the case of estimating the number of foreign substances generated in the CVD apparatus 10 has been described.
The number of foreign substances in a cleaning apparatus other than the D apparatus 10, a sputtering apparatus, or the like may be used, and even an etching rate other than the number of foreign substances can be predicted. In addition to data relating to semiconductor manufacturing, it is possible to predict that a data value will suddenly increase even for data that is observed by overlapping many factors indicating periodic fluctuations.
【００４２】[0042]
【発明の効果】請求項１の異常変動予測方法によると、
過去に測定した多くの変動パターンが存在しない場合で
も、測定値が大きく変動することを正確に予測すること
ができる。According to the abnormal fluctuation prediction method of the first aspect,
Even when there are not many fluctuation patterns measured in the past, it is possible to accurately predict that the measured value largely fluctuates.
【００４３】請求項２の異常変動予測方法によると、請
求項１の異常変動予測方法の効果が得られる上に、周波
数成分算出工程において、時系列データをフーリエ変換
することにより導出された複素数の実部及び虚部よりな
る周波数成分を算出するため、時系列データの周波数成
分を確実に算出することができる。According to the abnormal fluctuation prediction method of the second aspect, the effect of the abnormal fluctuation prediction method of the first aspect is obtained, and in the frequency component calculation step, the complex number derived by performing Fourier transform of the time series data is obtained. Since the frequency component including the real part and the imaginary part is calculated, the frequency component of the time-series data can be reliably calculated.
【００４４】請求項３の異常変動予測方法によると、請
求項２の異常変動予測方法の効果が得られる上に、異常
変動検出工程において、複素数の実部の平均値及び虚部
の平均値を用いて周波数成分の変化量を検出するため、
時系列データが示す変動周期に関わる周波数成分の変化
から測定値が大きく変化する兆候を確実に検出すること
ができる。According to the abnormal fluctuation prediction method of the third aspect, the effect of the abnormal fluctuation prediction method of the second aspect can be obtained, and in the abnormal fluctuation detecting step, the average value of the real part and the average value of the imaginary part of the complex number are calculated. To detect the amount of change in the frequency component using
It is possible to reliably detect a sign that the measured value greatly changes from a change in the frequency component related to the fluctuation cycle indicated by the time-series data.
【００４５】請求項４の異常変動予測方法によると、請
求項３の異常変動予測方法の効果が得られる上に、異常
変動検出工程は、複素数の実部の平均値と虚部の平均値
とにより表わされる２次元の座標系を導入し、許容変化
量の範囲を座標系における長円形の円周及び内部よりな
る領域として表わす工程を含み、長円形の長軸は、複素
数の実部及び虚部の各平均値から算出された近似直線に
平行であり、長軸から垂直方向への長円形の円周までの
距離は、長軸に垂直な方向に対して複素数の実部及び虚
部の各平均値が示す位置の分散値に依存するため、時系
列データが示す変動周期に関わる周波数成分の変化から
測定値が大きく変化する兆候をより一層確実に検出する
ことができる。According to the abnormal fluctuation predicting method of the fourth aspect, the effect of the abnormal fluctuation predicting method of the third aspect can be obtained. In addition, the abnormal fluctuation detecting step includes the step of calculating the average value of the real part and the average value of the imaginary part of the complex number. And expressing the range of the permissible variation as an area consisting of the circumference and interior of the ellipse in the coordinate system, wherein the major axis of the ellipse is the real part and the imaginary part of the complex number. The distance from the major axis to the circumference of the ellipse in the vertical direction is parallel to the approximate straight line calculated from each average value of the parts, and the distance between the real part and the imaginary part of the complex number with respect to the direction perpendicular to the major axis. Since it depends on the variance value of the position indicated by each average value, it is possible to more reliably detect a sign that the measured value greatly changes from a change in the frequency component related to the fluctuation cycle indicated by the time-series data.
【００４６】請求項５の異常変動予測装置によると、過
去に測定した多くの変動パターンが存在しない場合で
も、測定値が大きく変動することを正確に予測すること
ができる。According to the abnormal fluctuation predicting device of the present invention, it is possible to accurately predict that the measured value largely fluctuates even when there are not many fluctuation patterns measured in the past.
【００４７】また、周波数成分算出手段は、周波数成分
に時系列データをフーリエ変換することにより導出され
た複素数の実部及び虚部の値を用いるため、時系列デー
タの周波数成分を確実に算出することができる。Further, the frequency component calculating means uses the values of the real part and the imaginary part of the complex number derived by Fourier-transforming the time series data into the frequency component, so that the frequency component of the time series data is reliably calculated. be able to.
【００４８】さらに、異常変動検出手段は、複素数の実
部の平均値と虚部の平均値とにより表わされる２次元の
座標系を導入され、許容変化量の範囲を座標系における
長円形の円周及び内部よりなる領域として表わされ、長
円形の長軸は、複素数の実部及び虚部の各平均値から算
出された近似直線に平行であり、長軸から垂直方向への
長円形の円周までの距離は、長軸に垂直な方向に対して
複素数の実部及び虚部の各平均値が示す位置の分散値に
依存するため、時系列データが示す変動周期に関わる周
波数成分の変化から測定値が大きく変化する兆候を確実
に検出することができる。Further, the abnormal fluctuation detecting means is provided with a two-dimensional coordinate system represented by an average value of a real part and an average value of an imaginary part of a complex number. The long axis of the ellipse is parallel to the approximate straight line calculated from the average values of the real part and the imaginary part of the complex number, and is the ellipse in the vertical direction from the long axis. Since the distance to the circumference depends on the variance of the position indicated by each average of the real part and imaginary part of the complex number in the direction perpendicular to the long axis, the frequency component related to the fluctuation period indicated by the time-series data A sign that the measured value greatly changes from the change can be reliably detected.
【００４９】請求項６のプログラム製品によると、時系
列データをフーリエ変換することにより導出された複素
数の実部及び虚部の値を用いる周波数成分算出手段と、
複素数の実部の平均値及び虚部の平均値を用いて周波数
成分の変化量を検出し、検出した周波数成分の変化量で
ある検出変化量があらかじめ設定した変化量である許容
変化量の範囲内であるか否かを判断し、検出変化量が許
容変化量の範囲内にないときには、前記測定値が大きく
変動する兆候であると判定する異常変動検出手段とを有
する異常変動予測手段が記載されているため、過去に測
定した多くの変動パターンが存在しない場合でも、測定
値が大きく変動することを正確に予測することができ
る。According to the program product of claim 6, frequency component calculating means using values of a real part and an imaginary part of a complex number derived by performing a Fourier transform on the time series data,
The variation of the frequency component is detected using the average value of the real part and the average value of the imaginary part of the complex number, and the range of the allowable variation in which the detected variation that is the variation of the detected frequency component is a preset variation is set. It is determined whether or not the abnormal change is within the range of the allowable change amount.If the detected change amount is not within the range of the allowable change amount, the abnormal change detecting unit has an abnormal change detecting unit that determines that the measured value is a sign of a large change. Therefore, even when there are not many fluctuation patterns measured in the past, it is possible to accurately predict that the measured value largely fluctuates.
【図１】本発明の一実施形態に係る異常変動予測方法を
用いた異常変動予測装置を示す機能ブロック図である。FIG. 1 is a functional block diagram showing an abnormal fluctuation prediction device using an abnormal fluctuation prediction method according to an embodiment of the present invention.
【図２】本発明の一実施形態に係る異常変動予測方法に
おける時系列データのデータフォーマットを示す図であ
る。FIG. 2 is a diagram showing a data format of time-series data in the abnormal fluctuation prediction method according to one embodiment of the present invention.
【図３】（ａ）〜（ｃ）は本発明の一実施形態に係る異
常変動予測方法における時系列データを表わすトレンド
グラフである。FIGS. 3A to 3C are trend graphs showing time-series data in the abnormal fluctuation prediction method according to one embodiment of the present invention.
【図４】（ａ）、（ｂ）は本発明の一実施形態に係る異
常変動予測方法における時系列データを表わすトレンド
グラフである。FIGS. 4A and 4B are trend graphs showing time-series data in the abnormal fluctuation prediction method according to one embodiment of the present invention.
【図５】本発明の一実施形態に係る異常変動予測方法に
おける周波数成分の累積処理ウェハ枚数の増加に伴って
変化する実部及び虚部の平均値を示す散布図である。FIG. 5 is a scatter diagram showing an average value of a real part and an imaginary part that changes with an increase in the number of wafers to be cumulatively processed for frequency components in the abnormal fluctuation prediction method according to the embodiment of the present invention.
【図６】本発明の一実施形態に係る異常変動予測方法に
おける周波数成分の累積処理ウェハ枚数の増加に伴って
変化する実部及び虚部の平均値を示す散布図であって、
異常の兆候が現われる周波数成分の特徴的な変化を検出
する方法を示した図である。FIG. 6 is a scatter diagram showing an average value of a real part and an imaginary part that changes with an increase in the number of processed wafers of the frequency component in the abnormal fluctuation prediction method according to the embodiment of the present invention;
FIG. 7 is a diagram illustrating a method of detecting a characteristic change of a frequency component in which a sign of abnormality appears.
【図７】従来の予測制御装置を示す機能ブロック図であ
る。FIG. 7 is a functional block diagram showing a conventional predictive control device.
１０ ＣＶＤ装置 １１ 測定器１ １２ 外部記憶装置 １３ 解析装置 １３ａ 時系列データ作成手段 １３ｂ 周波数成分算出手段 １３ｃ 異常変動検出手段 １４ 表示手段 ２０ 異物数格納フィールド ２１ 累積処理ウェハ枚数格納フィールド ３０ 長円形 DESCRIPTION OF SYMBOLS 10 CVD apparatus 11 Measuring device 1 12 External storage device 13 Analysis device 13a Time series data creation means 13b Frequency component calculation means 13c Abnormal fluctuation detection means 14 Display means 20 Foreign matter number storage field 21 Cumulative processing wafer number storage field 30 Oval
Claims (6)
することを予測する異常変動予測方法であって、 測定値が時間とともに変化する時系列データにおける変
動周期の周波数成分を算出する周波数成分算出工程と、 前記周波数成分の変化量を検出し、検出した周波数成分
の変化量である検出変化量があらかじめ設定した変化量
である許容変化量の範囲内であるか否かを判断し、前記
検出変化量が前記許容変化量の範囲内にないときには、
測定値が大きく変動する兆候であると判定する異常変動
検出工程とを備えていることを特徴とする異常変動予測
方法。An abnormal fluctuation prediction method for predicting a large change in the value of data to be measured, comprising: calculating a frequency component of a fluctuation cycle in time-series data in which a measured value changes with time. Detecting the change amount of the frequency component, and determining whether the detected change amount that is the detected change amount of the frequency component is within a range of an allowable change amount that is a previously set change amount, and performing the detection. When the change amount is not within the range of the allowable change amount,
An abnormal fluctuation detecting step of determining that the measured value is a sign of a large fluctuation.
データをフーリエ変換することにより導出された複素数
の実部及び虚部の値よりなる前記周波数成分を算出する
工程を含むことを特徴とする請求項１に記載の異常変動
予測方法。2. The frequency component calculating step includes a step of calculating the frequency component comprising values of a real part and an imaginary part of a complex number derived by performing a Fourier transform on the time series data. The abnormal fluctuation prediction method according to claim 1.
実部の平均値及び虚部の平均値を用いて前記周波数成分
の変化量を検出する工程を含むことを特徴とする請求項
２に記載の異常変動予測方法。3. The method according to claim 2, wherein the abnormal fluctuation detecting step includes a step of detecting a change amount of the frequency component using an average value of a real part and an average value of an imaginary part of the complex number. Abnormal fluctuation prediction method described.
実部の平均値と虚部の平均値とにより表わされる２次元
の座標系を導入し、前記許容変化量の範囲を前記座標系
における長円形の円周及び内部よりなる領域として表わ
す工程を含み、 前記長円形の長軸は、前記複素数の実部及び虚部の各平
均値から算出された近似直線に平行であり、 前記長軸から垂直方向への前記長円形の円周までの距離
は、前記長軸に垂直な方向に対して前記複素数の実部及
び虚部の各平均値が示す位置の分散値に依存することを
特徴とする請求項３に記載の異常変動予測方法。4. The abnormal variation detecting step introduces a two-dimensional coordinate system represented by an average value of a real part and an average value of an imaginary part of the complex number, and sets a range of the allowable change amount in the coordinate system. A step of representing the region as a region consisting of the circumference and the interior of the oval, wherein the long axis of the oval is parallel to an approximate straight line calculated from respective average values of the real part and the imaginary part of the complex number, and The distance from the vertical to the circumference of the oval depends on the variance of the position indicated by each average value of the real part and the imaginary part of the complex number with respect to the direction perpendicular to the major axis. The abnormal fluctuation prediction method according to claim 3, wherein
ータをフーリエ変換することにより導出された複素数の
実部及び虚部の値を用いて前記時系列データにおける変
動周期の周波数成分を算出する周波数成分算出手段と、 前記複素数の実部の平均値及び虚部の平均値を用いて前
記周波数成分の変化量を検出し、検出した周波数成分の
変化量である検出変化量があらかじめ設定した変化量で
ある許容変化量の範囲内であるか否かを判断し、前記検
出変化量が前記許容変化量の範囲内にないときには、測
定値が大きく変動する兆候であると判定する異常変動検
出手段とを備え、 前記異常変動検出手段において、前記複素数の実部の平
均値と虚部の平均値とにより表わされる２次元の座標系
を導入され、前記許容変化量の範囲を前記座標系におけ
る長円形の円周及び内部よりなる領域として表わされる
と共に、 前記長円形の長軸は、前記複素数の実部及び虚部の各平
均値から算出された近似直線に平行であり、 前記長軸から垂直方向への前記長円形の円周までの距離
は、前記長軸に垂直な方向に対して前記複素数の実部及
び虚部の各平均値が示す位置の分散値に依存することを
特徴とする異常変動予測装置。5. A frequency for calculating a frequency component of a fluctuation period in the time-series data using values of a real part and an imaginary part of a complex number derived by performing a Fourier transform on the time-series data whose measured value changes with time. A component calculating means for detecting a change amount of the frequency component using an average value of a real part and an average value of an imaginary part of the complex number, and a detected change amount which is a detected change amount of the frequency component is a predetermined change amount. It is determined whether or not within the range of the allowable change amount, and when the detected change amount is not within the range of the allowable change amount, an abnormal change detection unit that determines that the measured value is a sign of a large change. In the abnormal fluctuation detecting means, a two-dimensional coordinate system represented by an average value of a real part and an average value of an imaginary part of the complex number is introduced, and a range of the allowable change amount is set in the coordinate system. Expressed as a region consisting of the circumference and the interior of the oval, the long axis of the oval is parallel to an approximate straight line calculated from each average value of the real part and the imaginary part of the complex number, and from the long axis The distance to the circumference of the oval in the vertical direction is dependent on the variance of the position indicated by each average value of the real part and the imaginary part of the complex number with respect to the direction perpendicular to the long axis. Abnormal fluctuation prediction device.
ことを予測するプログラム製品であって、以下の手順を
実現するプログラム記録媒体を含む：測定値が時間とと
もに変化する時系列データをフーリエ変換することによ
り導出された複素数の実部及び虚部の値を用いて前記時
系列データにおける変動周期の周波数成分を算出する周
波数成分算出手段と、 前記複素数の実部の平均値及び虚部の平均値を用いて前
記周波数成分の変化量を検出し、検出した周波数成分の
変化量である検出変化量があらかじめ設定した変化量で
ある許容変化量の範囲内であるか否かを判断し、前記検
出変化量が前記許容変化量の範囲内にないときには、測
定値が大きく変動する兆候であると判定する異常変動検
出手段とを備え、 前記異常変動検出手段において、前記複素数の実部の平
均値と虚部の平均値とにより表わされる２次元の座標系
を導入され、前記許容変化量の範囲を前記座標系におけ
る長円形の円周及び内部よりなる領域として表わすと共
に、 前記長円形の長軸は、前記複素数の実部及び虚部の各平
均値から算出された近似直線に平行であり、 前記長軸から垂直方向への前記長円形の円周までの距離
は、前記長軸に垂直な方向に対して前記複素数の実部及
び虚部の各平均値が示す位置の分散値に依存することを
特徴とする異常変動予測手段を記載したプログラム製
品。6. A program product for predicting a large change in the value of data to be measured, including a program recording medium for realizing the following procedure: Fourier transform of time-series data in which a measured value changes with time Frequency component calculating means for calculating the frequency component of the fluctuation period in the time-series data using the values of the real part and the imaginary part of the complex number derived by the calculation, and the average of the real part and the imaginary part of the complex number The change amount of the frequency component is detected using the value, and it is determined whether the detected change amount which is the detected change amount of the frequency component is within a range of an allowable change amount which is a preset change amount, and When the detected change amount is not within the range of the permissible change amount, the device includes an abnormal change detecting unit that determines that the measured value is a sign of a large change. A two-dimensional coordinate system represented by the average value of the real part and the average value of the imaginary part of the complex number is introduced, and the range of the allowable variation is expressed as an area consisting of the circumference and the inside of an oval in the coordinate system. Along with the major axis of the ellipse is parallel to an approximate straight line calculated from the average value of the real part and the imaginary part of the complex number, and the distance from the major axis to the circumference of the ellipse in the vertical direction. Is a program product which describes an abnormal fluctuation predicting means, which depends on a variance value of a position indicated by each average value of a real part and an imaginary part of the complex number in a direction perpendicular to the major axis.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP18478796A JP3628812B2 (en) | 1996-07-15 | 1996-07-15 | Abnormal fluctuation prediction method and abnormal fluctuation prediction apparatus |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP18478796A JP3628812B2 (en) | 1996-07-15 | 1996-07-15 | Abnormal fluctuation prediction method and abnormal fluctuation prediction apparatus |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JPH1031511A true JPH1031511A (en) | 1998-02-03 |
| JP3628812B2 JP3628812B2 (en) | 2005-03-16 |
Family
ID=16159292
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP18478796A Expired - Fee Related JP3628812B2 (en) | 1996-07-15 | 1996-07-15 | Abnormal fluctuation prediction method and abnormal fluctuation prediction apparatus |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP3628812B2 (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008186248A (en) * | 2007-01-30 | 2008-08-14 | Mizuho Information & Research Institute Inc | Discrete data processing system, discrete data processing method and discrete data processing program |
| JP2008282172A (en) * | 2007-05-09 | 2008-11-20 | National Maritime Research Institute | Structure monitoring device by means of sound |
| JP2011104341A (en) * | 2009-10-22 | 2011-06-02 | Nippon Koden Corp | Biological parameter displaying apparatus |
| KR20160079571A (en) * | 2014-12-26 | 2016-07-06 | 삼성전자주식회사 | Method for processing location information and method for processing measurement information including the same |
| WO2024004216A1 (en) * | 2022-07-01 | 2024-01-04 | 株式会社Synspective | Ground deformation analysis device and ground deformation analysis method |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH01245122A (en) * | 1988-03-28 | 1989-09-29 | Toshiba Corp | Abnormality detecting method |
| JPH0463662A (en) * | 1990-07-03 | 1992-02-28 | Komatsu Ltd | Tool abnormality detecting device |
| JPH0560596A (en) * | 1991-09-04 | 1993-03-09 | Hitachi Ltd | Abnormality diagnostic unit for rotary equipment |
| JPH064789A (en) * | 1992-06-22 | 1994-01-14 | Hitachi Ltd | Method and device for monitoring abnormality of equipment |
| JPH08159890A (en) * | 1994-12-06 | 1996-06-21 | Dai Ichi Seiyaku Co Ltd | Load alteration measuring method |
| JPH102843A (en) * | 1996-06-14 | 1998-01-06 | Syst Sogo Kaihatsu Kk | Production process control system |
| JPH10511177A (en) * | 1994-12-09 | 1998-10-27 | エクソン・ケミカル・パテンツ・インク | Plant parameter detection by power spectral density monitoring |
-
1996
- 1996-07-15 JP JP18478796A patent/JP3628812B2/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH01245122A (en) * | 1988-03-28 | 1989-09-29 | Toshiba Corp | Abnormality detecting method |
| JPH0463662A (en) * | 1990-07-03 | 1992-02-28 | Komatsu Ltd | Tool abnormality detecting device |
| JPH0560596A (en) * | 1991-09-04 | 1993-03-09 | Hitachi Ltd | Abnormality diagnostic unit for rotary equipment |
| JPH064789A (en) * | 1992-06-22 | 1994-01-14 | Hitachi Ltd | Method and device for monitoring abnormality of equipment |
| JPH08159890A (en) * | 1994-12-06 | 1996-06-21 | Dai Ichi Seiyaku Co Ltd | Load alteration measuring method |
| JPH10511177A (en) * | 1994-12-09 | 1998-10-27 | エクソン・ケミカル・パテンツ・インク | Plant parameter detection by power spectral density monitoring |
| JPH102843A (en) * | 1996-06-14 | 1998-01-06 | Syst Sogo Kaihatsu Kk | Production process control system |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008186248A (en) * | 2007-01-30 | 2008-08-14 | Mizuho Information & Research Institute Inc | Discrete data processing system, discrete data processing method and discrete data processing program |
| JP4669849B2 (en) * | 2007-01-30 | 2011-04-13 | みずほ情報総研株式会社 | Discrete data processing system, discrete data processing method, and discrete data processing program |
| JP2008282172A (en) * | 2007-05-09 | 2008-11-20 | National Maritime Research Institute | Structure monitoring device by means of sound |
| JP2011104341A (en) * | 2009-10-22 | 2011-06-02 | Nippon Koden Corp | Biological parameter displaying apparatus |
| US9743842B2 (en) | 2009-10-22 | 2017-08-29 | Nihon Kohden Corporation | Biological parameter displaying apparatus |
| KR20160079571A (en) * | 2014-12-26 | 2016-07-06 | 삼성전자주식회사 | Method for processing location information and method for processing measurement information including the same |
| WO2024004216A1 (en) * | 2022-07-01 | 2024-01-04 | 株式会社Synspective | Ground deformation analysis device and ground deformation analysis method |
Also Published As
| Publication number | Publication date |
|---|---|
| JP3628812B2 (en) | 2005-03-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP3495705B2 (en) | Ultra-sensitive monitoring of sensors and processes | |
| KR100990845B1 (en) | Plasma abnormal discharge diagnosis method, plasma abnormal discharge diagnosis system, and computer-readable recording medium having computer program thereon | |
| JP4850857B2 (en) | Adaptive modeling of change states in predictive state monitoring | |
| US20020046004A1 (en) | General method for tracking the evolution of hidden damage or other unwanted changes in machinery components and predicting remaining useful life | |
| CN110068435B (en) | Vibration analysis system and method | |
| US7949497B2 (en) | Machine condition monitoring using discontinuity detection | |
| JP2000259223A (en) | Plant monitoring device | |
| JP2020052740A (en) | Abnormality detection device, abnormality detection method, and program | |
| US20110196820A1 (en) | Robust Filtering And Prediction Using Switching Models For Machine Condition Monitoring | |
| CN112246469B (en) | Method and apparatus for controlling spray coating of vehicle | |
| EP3975077A1 (en) | Monitoring device and method for segmenting different times series of sensor data points | |
| EP3703250A1 (en) | Servo driver and state change detecting method | |
| JP3628812B2 (en) | Abnormal fluctuation prediction method and abnormal fluctuation prediction apparatus | |
| EP3349084B1 (en) | Process monitoring device, method and program | |
| JP2010218301A (en) | Failure diagnosis apparatus, failure diagnosis method, and failure diagnosis program | |
| CN113721086A (en) | Method for monitoring the electrical insulation condition of a piece of equipment of an MV or HV electrical system | |
| US10922858B2 (en) | Display apparatus, display method, and recording medium | |
| JPH08314530A (en) | Fault prediction device | |
| JP2003068717A (en) | Semiconductor processing apparatus | |
| JP7020565B2 (en) | Process control equipment, process control method and process control program | |
| US20070225926A1 (en) | Method and apparatus for quantitatively determining severity of degradation in a signal | |
| JP4102019B2 (en) | Analysis apparatus, method, and computer-readable storage medium in manufacturing process | |
| EP3754446A1 (en) | Method and assistance system for detecting an abnormal behaviour of a device | |
| JP3319297B2 (en) | Blast furnace abnormal furnace condition prediction method | |
| KR100446926B1 (en) | Method of monitoring and/or controlling a semiconductor manufacturing apparatus and a system therefor |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A977 | Report on retrieval |
Free format text: JAPANESE INTERMEDIATE CODE: A971007 Effective date: 20040318 |
|
| A131 | Notification of reasons for refusal |
Free format text: JAPANESE INTERMEDIATE CODE: A131 Effective date: 20040629 |
|
| A521 | Written amendment |
Free format text: JAPANESE INTERMEDIATE CODE: A523 Effective date: 20040825 |
|
| 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: 20041124 |
|
| A61 | First payment of annual fees (during grant procedure) |
Free format text: JAPANESE INTERMEDIATE CODE: A61 Effective date: 20041209 |
|
| R150 | Certificate of patent or registration of utility model |
Free format text: JAPANESE INTERMEDIATE CODE: R150 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20071217 Year of fee payment: 3 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20081217 Year of fee payment: 4 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20091217 Year of fee payment: 5 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20091217 Year of fee payment: 5 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20101217 Year of fee payment: 6 |
|
| LAPS | Cancellation because of no payment of annual fees |