JPS595913A - Method for detecting sensor whose output is abnormal - Google Patents
Method for detecting sensor whose output is abnormalInfo
- Publication number
- JPS595913A JPS595913A JP11520282A JP11520282A JPS595913A JP S595913 A JPS595913 A JP S595913A JP 11520282 A JP11520282 A JP 11520282A JP 11520282 A JP11520282 A JP 11520282A JP S595913 A JPS595913 A JP S595913A
- Authority
- JP
- Japan
- Prior art keywords
- value
- sensor
- abnormal
- sensors
- estimated
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D3/00—Indicating or recording apparatus with provision for the special purposes referred to in the subgroups
- G01D3/08—Indicating or recording apparatus with provision for the special purposes referred to in the subgroups with provision for safeguarding the apparatus, e.g. against abnormal operation, against breakdown
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
Abstract
Description
【発明の詳細な説明】
〔発明の技術分野〕
本発明ば、プｃｘｔス制御寺において使用されるでンサ
について出力値の異常を検知する方法に関する。DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to a method for detecting an abnormality in the output value of a sensor used in a CXT control system.
生産工程などの工業プロヤスにおいて各種の状態量を計
算機制御により自動制御を行うプロヤス制御においては
、制御対象となる状態量を検出して制御装置に送るｔン
サの機能が重要となっている。ところが、ｔンサ自身の
異常やでンサの取付は方の誤り等によってｔンサ出力値
が必ずしも正しい値を示さないことがある。このような
異常値にもとづいて計算機制御を行った場合には適正な
制御が行われない。したがって、適正な制御のためには
異常なでンサ出力を検知し、その値を補正するか無視す
る等の処理が必要となる。BACKGROUND ART In industrial control in which various state quantities are automatically controlled by computer control in industrial processes such as production processes, the function of a sensor that detects state quantities to be controlled and sends them to a control device is important. However, the output value of the sensor may not always show the correct value due to an abnormality in the sensor itself or an error in the installation of the sensor. If computer control is performed based on such abnormal values, proper control will not be performed. Therefore, for proper control, it is necessary to detect an abnormal sensor output and perform processing such as correcting or ignoring the value.
この異常なでンサの検知に関しては従来、１つの状態量
に対して複数の七ンサを取付け、大多数のでンサが示し
た値を正しい値とするいわゆる多数決に基づく方法、あ
るいは状態量がとり得ないかとってはならない値をあら
かじめ設定しておきそのような値をｔンサが示した場合
にｔンサを異常と判定する方法などが採用されている。Conventionally, the method for detecting this abnormal sensor is based on so-called majority voting, in which multiple sensors are attached to one state quantity and the value indicated by the majority of sensors is the correct value, or the state quantity can be used. A method has been adopted in which a value that must not be present is set in advance, and when the sensor shows such a value, the sensor is determined to be abnormal.
しかしながら複数センサを用いる方法に、多くのセンサ
を用意しなければならないためコストの増加を招く欠点
があり、また１ンサを取付ける場所によっては構造上あ
るいは面積上複数でンサの取付けが不可能なこともある
。またせっかく複数のセンサを取付けたとしても、流量
測定において曲管部や断面積急変部にてンサを取付ける
等のセンサの取付自体の誤りに起因する測定値異常につ
いてＩ′ｊニーｔｒンサ自体に異常がない限りすべての
センサが同じ値を示してしまうため多数決により正しい
値と誤認してしまうことになる。However, the method of using multiple sensors has the disadvantage of increasing costs because many sensors must be prepared, and depending on the location where one sensor is installed, it may be impossible to install multiple sensors due to the structure or area. There is also. In addition, even if multiple sensors are installed, the I'j knee sensor itself will not be able to handle measurement abnormalities caused by incorrect installation of the sensors, such as installing the sensor at a curved pipe section or a section where the cross-sectional area suddenly changes. As long as there is no abnormality, all the sensors will show the same value, so the majority vote will result in a false recognition of the correct value.
一方、状態ｔがとり得ない値をもってセンサ異常とする
方法では、あらかじめ異常とに見ない範囲の値を示した
軽微な異常のあるセンサは正常と一１４１断されてしま
う欠点がある。On the other hand, the method of determining a sensor abnormality when a value that cannot be taken by the state t has the disadvantage that a sensor with a slight abnormality that shows a value in a range that is not seen as an abnormality is determined to be normal.
そこで本発明は、従来の出力値異常センサの検知方法に
おける欠点を除去し、１つの状態量に対して１つのセン
サを用いながらセンサ自身の異常及びセンサの取付は時
の異常を容易に検出することのできる異常センサの検知
方法を提供することを目的とする。Therefore, the present invention eliminates the shortcomings in the conventional detection method of an output value abnormality sensor, and easily detects an abnormality in the sensor itself and an abnormality in time when the sensor is attached, while using one sensor for one state quantity. The purpose of the present invention is to provide a method for detecting an abnormality using a sensor.
本発明げ、でンサで検出すべき各種状態量間に存在する
既知の物理的関係を利用して最も本来の値に近い状態ｔ
を計算によって求め、その値とセンサ自身が示している
測定値との偏差によりセンサの異常を判断するものであ
り、■状態につき１ｔンサを用意することによって確実
に異常な出力値を示すセンサの検知を可能ならしめるも
のである。The present invention utilizes the known physical relationships between various state quantities to be detected by the sensor to detect the state t that is closest to the original value.
is determined by calculation, and the abnormality of the sensor is determined based on the deviation between that value and the measured value indicated by the sensor itself. ■ By preparing a 1-ton sensor for each condition, it is possible to ensure that the sensor exhibits an abnormal output value. This makes detection possible.
まず、本発明の基本原理となっている、最も本来の値に
近い状態量を推定する数学的手法につき説明する。First, a mathematical method for estimating the state quantity closest to the original value, which is the basic principle of the present invention, will be explained.
あるプロでスを仮定すると、そのプロでスには流量、ｍ
闇、圧力等の各稽状態童が存在し、これらの状態量間に
は相互に物理的な関係が成立している。この物理的関係
が既知であるとすれば、各（３）
状態を並びに各状態量間に成立する関係式の係数及び定
数を要素として次のような行列ベクトル形式の方程式が
成立する。Assuming a certain process, the process has a flow rate, m
There are various physical states such as darkness and pressure, and there is a physical relationship between these state quantities. Assuming that this physical relationship is known, the following matrix-vector equation is established using each (3) state and the coefficients and constants of the relational expression established between each state quantity as elements.
Ａ−ｘ＝ｂ ・・・・・・・・・・・
・・・・（１）ここで、ベクトルｘＨ各状態量を表わす
ｎ次元空間を有する列ベクトル、・ベクトル八及びｂｉ
関係式の定数及び係数よりなるベクトルでベクトルＡに
ｍＸｎ次元空間を有する行列ベクトル、・ベクトルｂ
Ｉｔｓ ｍ次元空間を有する列ベクトルであり、これら
を次のように表わすことができる。A-x=b・・・・・・・・・・・・
...(1) Here, vector xH is a column vector having an n-dimensional space representing each state quantity, ・Vector 8 and bi
A vector consisting of constants and coefficients of the relational expression, a matrix vector having an mXn-dimensional space in vector A, ・Vector b
Its is a column vector with m-dimensional space, and these can be expressed as follows.
ＡεＨｍＸｆｌ
ＸＥＲ”
ｉＥ Ｈｍ
各状態量げセンサによって測定されており、既知である
から、ベクトルｘＨ各測定値により構成されるベクトル
ｃｖｃｘり表わされる。すなわち、Ｘ＝Ｃ・・・・・・
・・・・・・・・・（２）でありＣεＨｎ
である。AεHm
......(2) and CεHn.
（２式の両辺にｎＸｎ次元の単位行列１をかけれ（４）
ば
Ｘ ・ ｌ＝ｃ ・・・・・・・・
・・・・・・・（り′ただし ｌ ｅ ＲｎＸ ｎ
となるから（１）式と（２）２式はまとめることができ
と表わすことができる。(Multiply both sides of equation 2 by nXn-dimensional unit matrix 1 (4) If X ・ l=c ・・・・・・・・・
......(R' However, since le RnX n , equations (1) and (2) can be combined and expressed as follows.
とすれば
Ｄ−ｘ＝ｅ ・・・・・・・・・・・
・・・・（３）と表わされる。この方程式は、未知数の
数ｎよりも方程式の数、（ｍ十ｎ）の方が多い不能方程
式である。この方程式は完全には解けないが誤差を最小
とする最小２乗解ｘｈ次のような方程式を満たすことが
知られている。Then D-x=e・・・・・・・・・・・・
...It is expressed as (3). This equation is an impossible equation in which the number of equations (m1n) is greater than the number of unknowns n. Although this equation cannot be completely solved, it is known that a least squares solution xh that minimizes the error satisfies the following equation.
ＤＴＤｘ＝ＤＴｅ
ここで、ＤＴ はベクトルＤの転置行列であり、上記方
程式に正規方程式として知られているものである。もし
ベクトルＤが線形独立な列を待てばＤの階数ｒと空間次
元数ｎに等しく、ＤＴＤは正方対称で逆ｏＪ能となり、
−意の最小２乗解が存在し、×−〔ＤＴＤ〕−１ＤＴｅ
として得られる。すなわち、この解にそれがとり得る倶
差を最も少なくするような値を与える性質を有するため
、測定値に当てはめると、本来あるべき値より異常に太
きくはずれた場合にも他の正常な測定値及びこれらの間
に存在する関係式から最も本来の値に近い値を得ること
が可能となる。DTDx=DTe where DT is the transposed matrix of vector D, which is known as the normal equation to the above equation. If the vector D is a linearly independent sequence, the rank r of D is equal to the spatial dimension n, and the DTD is square symmetric and has an inverse oJ function.
- There exists a least squares solution, which is obtained as ×-[DTD]-1DTe. In other words, this solution has the property of giving a value that minimizes the possible difference, so when applied to a measured value, even if the value deviates abnormally from the original value, it will be compared to other normal measurements. It becomes possible to obtain the value closest to the original value from the values and the relational expressions that exist between them.
本発明げこの手法を応用したものであり、以下図面を参
照しながら本発明の一笑施例を説明する。This method is an application of the method of the present invention, and a simple embodiment of the present invention will be described below with reference to the drawings.
第１図は本発明に係る異常センサの検知に使用するシス
テムの概略を示しり構成図であって、■は制御対象とな
るプロセスであり、これには制御要素ごとにセンサ２が
覗り付けられている。これらのセンサ２からの出力にゾ
ロヤス入力装置１３を経て計算機４に取り込まれ、本発
明にかかる異常センサの検知が行われる。この結果は必
要に応じラインプリンタなどの出力機器に出力されると
共に、この異常値に検知過程で得られた最小勝差を与え
る推定値に補正されて計算制御の有効データとなる。FIG. 1 is a block diagram showing an outline of a system used for detecting an abnormality sensor according to the present invention, where ■ indicates a process to be controlled, and a sensor 2 looks into it for each control element. It is being Outputs from these sensors 2 are input to the computer 4 via the Zoroyas input device 13, and abnormal sensor detection according to the present invention is performed. This result is output to an output device such as a line printer as necessary, and is corrected to an estimated value that gives the minimum winning difference obtained in the detection process to this abnormal value, and becomes effective data for calculation control.
第２図に異常センサの検知方法の詳細を示すフローチャ
ートであって、まず各センサよりの信号を読み込み、こ
れとあらかじめ計算機の記憶装着内に記憶されている各
状態量間に成立する関係式の係数及び定数から行列方程
式を構成し、この方程式の最小２乗解を求める状態量値
の推定計算を行う。次にこの推定値と測定値とを比較し
、それらの差分が一定の基準値、例えば５％を超えるも
のについては異常値と判断して選び出す。このようにし
て異常値と判断された出力値については。FIG. 2 is a flowchart showing the details of the abnormality sensor detection method, in which the signals from each sensor are first read, and the relational expression established between this and each state quantity stored in advance in the memory of the computer is calculated. A matrix equation is constructed from coefficients and constants, and state quantity values are estimated and calculated to find the least squares solution of this equation. Next, this estimated value and the measured value are compared, and if the difference between them exceeds a certain reference value, for example 5%, it is determined to be an abnormal value and selected. Regarding output values that are determined to be abnormal values in this way.
必要に応じディスプレイ装置やプリンタ等の出力装酸に
、１ンサの名称や推定値とともに表示される。これＫよ
り操作員は異常値を推定値に＃き換える等の手動操作を
行い、プロセス制御をより適切に行うことができる。な
お、このような操作を計算機が自動的に行うようにする
ことも可能である。If necessary, it will be displayed on the output device of a display device, printer, etc. together with the name and estimated value of the sensor. This allows the operator to perform manual operations such as replacing abnormal values with estimated values, thereby allowing more appropriate process control. Note that it is also possible to have a computer automatically perform such operations.
センサで検出すべき状態量間に存在する既知の関係を利
用して最も本来の値に近い状態量を関係式の係数、定数
及びセンサ出力値を要素とする行列方程式より最小２乗
解を求めることによって推定し、この推定値とセンサ出
力値との差が一定基準より多い場合にばセンサ出力値を
異常と判断する本発明にかかる出力値異常センサの検知
方法を用いれば、１状態量については１ｔンサを準備す
れば十分であるため既存プロでスへの導入が容易である
他ｔンサの数を最小限にすることができ制御システムを
安価に構成することができる。またでンサ自身に異常が
ないにもかかわらｆ−ンサの取付状態が不適切である場
合にも出力値の異常を検知することができる。Using the known relationship between the state quantities to be detected by the sensor, find the least squares solution for the state quantity closest to the original value from a matrix equation whose elements are the coefficients of the relational expression, constants, and sensor output values. If the method for detecting an abnormal output value sensor according to the present invention is used, in which the sensor output value is determined to be abnormal if the difference between this estimated value and the sensor output value is larger than a certain standard, Since it is sufficient to prepare a 1 ton sensor, it is easy to introduce it into an existing professional system.In addition, the number of ton sensors can be minimized, and the control system can be constructed at low cost. Further, even if there is no abnormality in the f-sensor itself, but the f-sensor is not installed properly, it is possible to detect an abnormality in the output value.
さらに異常値検出過程で得られた推定値を補正値として
制御計算に使用できるため、制御を適正に行うことがで
き、プロセス制御の確実性安全性を向上させることがで
きる。Furthermore, since the estimated value obtained in the abnormal value detection process can be used as a correction value in control calculations, control can be performed appropriately and the reliability and safety of process control can be improved.
第１図は本発明に係る異常センサ出力値の検知を行うた
めのシステムの概略を示す構成図、第２図は本発明に係
る異常センサ出力値の検知方法における処理の概略を示
すフローチャートである。
１・・・プロセス、２・・・センサ、３・・・プローｔ
ス入力装首、４・・・計算機、５・・・出力機器。
出願人代理人 猪 股 清
第２図FIG. 1 is a configuration diagram showing an outline of a system for detecting an abnormal sensor output value according to the present invention, and FIG. 2 is a flowchart showing an outline of processing in the method for detecting an abnormal sensor output value according to the present invention. . 1... process, 2... sensor, 3... probe t
4. Computer, 5. Output device. Applicant's agent Kiyoshi Inomata Figure 2
Claims (1)
介して計算機に取り込み、この信号に演算処理を施して
その演算結果を操作機器に与え、プロｔスを制御する方
法において、前記プロヤス内の各制御要素の状態量ＭＩ
Ｃ成立する関係式の係数、定数及びｔンサ出力値を要素
とする行列方程式より最／Ｊ・２乗解を計算機により推
定し、この推定値と前記ｔンサ出力値との差が一定基準
より多い場合には前記ｔンサ出力値を異常と判断するよ
うにした出力値異常センサの検知方法。In the method of inputting signals from a plurality of temperature sensors into a computer via a process signal input device, performing arithmetic processing on the signals, and giving the result of the calculation to an operating device to control the processor, State quantity MI of each control element
The most /J squared solution is estimated by a computer from the matrix equation whose elements are the coefficients, constants, and t-sensor output values of the relational expression that holds C, and the difference between this estimated value and the t-sensor output value is calculated based on a certain standard. A method for detecting an abnormal output value of a sensor, in which the output value of the t-sensor is determined to be abnormal if the output value is abnormal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP11520282A JPS595913A (en) | 1982-07-02 | 1982-07-02 | Method for detecting sensor whose output is abnormal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP11520282A JPS595913A (en) | 1982-07-02 | 1982-07-02 | Method for detecting sensor whose output is abnormal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| JPS595913A true JPS595913A (en) | 1984-01-12 |
Family
ID=14656879
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP11520282A Pending JPS595913A (en) | 1982-07-02 | 1982-07-02 | Method for detecting sensor whose output is abnormal |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JPS595913A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS6227672A (en) * | 1985-07-29 | 1987-02-05 | Ishikawajima Harima Heavy Ind Co Ltd | Apparatus for self-diagnosis of sensor |
| US5680409A (en) * | 1995-08-11 | 1997-10-21 | Fisher-Rosemount Systems, Inc. | Method and apparatus for detecting and identifying faulty sensors in a process |
| WO2002088850A1 (en) | 2001-04-26 | 2002-11-07 | Abb As | Method for detecting and correcting sensor failure in oil and gas production system |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS5348545A (en) * | 1976-10-15 | 1978-05-02 | Hitachi Ltd | Monitor for equipment condition |
| JPS54142470A (en) * | 1978-04-27 | 1979-11-06 | Tohoku Electric Power Co | Forecast monitor |
-
1982
- 1982-07-02 JP JP11520282A patent/JPS595913A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS5348545A (en) * | 1976-10-15 | 1978-05-02 | Hitachi Ltd | Monitor for equipment condition |
| JPS54142470A (en) * | 1978-04-27 | 1979-11-06 | Tohoku Electric Power Co | Forecast monitor |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS6227672A (en) * | 1985-07-29 | 1987-02-05 | Ishikawajima Harima Heavy Ind Co Ltd | Apparatus for self-diagnosis of sensor |
| US5680409A (en) * | 1995-08-11 | 1997-10-21 | Fisher-Rosemount Systems, Inc. | Method and apparatus for detecting and identifying faulty sensors in a process |
| WO2002088850A1 (en) | 2001-04-26 | 2002-11-07 | Abb As | Method for detecting and correcting sensor failure in oil and gas production system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP3698444B2 (en) | Method and apparatus for detecting and identifying defect sensors in a process | |
| KR102025145B1 (en) | Apparatus and Method for Predicting Plant Data | |
| Eubank | Approximate regression models and splines | |
| CN102692248B (en) | The deteriorated sensor realized in transmitter detects | |
| US10884404B2 (en) | Method of predicting plant data and apparatus using the same | |
| WO1997021157A1 (en) | A method of predictive maintenance of a process control system haivng fluid movement | |
| Mammadova et al. | Profile monitoring for count data using Poisson and Conway–Maxwell–Poisson regression-based control charts under multicollinearity problem | |
| CN112463646A (en) | Sensor abnormity detection method and device | |
| Amand et al. | Plant monitoring and fault detection: Synergy between data reconciliation and principal component analysis | |
| JPS595913A (en) | Method for detecting sensor whose output is abnormal | |
| US20050174073A1 (en) | System for controlling the operation of at least one aircraft engine | |
| JP4636428B2 (en) | Multivariable transmitter and arithmetic processing method of multivariable transmitter | |
| US20170159486A1 (en) | Method for Validation of an Investigated Sensor and Corresponding Machine | |
| JP2022020940A (en) | Abnormality diagnosis system | |
| JP4296888B2 (en) | Electronic control unit | |
| JPH06109498A (en) | Detector for unsteady and abnormal state | |
| JPS6348300B2 (en) | ||
| KR102199695B1 (en) | Apparatus and method for online signal data validation via auto associative bilateral kernel regression | |
| WO2024105727A1 (en) | Factor inference apparatus, factor inference method, and factor inference program | |
| JP2691243B2 (en) | Process abnormality monitoring method and apparatus thereof | |
| JP2557885B2 (en) | Differential pressure reactor water level measuring device | |
| JPS5833529Y2 (en) | Measured value correction method | |
| TW202036191A (en) | Monitoring device, display device, monitoring method, and monitoring program | |
| JPH04296620A (en) | Parameter display device | |
| JPH04186119A (en) | Method and device for monitoring abnormality |