TW202210977A - Abnormality detection device - Google Patents
Abnormality detection device Download PDFInfo
- Publication number
- TW202210977A TW202210977A TW109144679A TW109144679A TW202210977A TW 202210977 A TW202210977 A TW 202210977A TW 109144679 A TW109144679 A TW 109144679A TW 109144679 A TW109144679 A TW 109144679A TW 202210977 A TW202210977 A TW 202210977A
- Authority
- TW
- Taiwan
- Prior art keywords
- state
- state transition
- series data
- abnormality detection
- unit
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B23/00—Testing or monitoring of control systems or parts thereof
- G05B23/02—Electric testing or monitoring
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Automation & Control Theory (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
Description
本發明係關於一種異常檢知裝置。The present invention relates to an abnormality detection device.
檢出製造機械等各種裝置之異常之技術係眾所周知的。例如在專利文獻1中,提案有一種依據自關連於某特定製造機械之偵知器所取得之物理量數據、及做為表示該製造機械之動作有無異常之資訊之動作狀態數據之相關性,以學習異常之預兆,使用此學習結果以檢知異常之預兆之異常檢出器。製造機械是否正常動作之假設結果,係例如藉輸出製造機械係正在正常動作之機率以進行之。
[專利文獻]Techniques for detecting abnormalities in various devices such as manufacturing machines are well known. For example, in
[專利文獻1]日本特開2019-204155號公報[Patent Document 1] Japanese Patent Application Laid-Open No. 2019-204155
一般說來,針對取得複數動作狀態之製造機械,當由該複數動作狀態所構成之製程之流程係異常時,可認為該製造機械係異常。但是,當依據專利文獻1之技術時,係僅依據針對個別之動作狀態之學習結果,判定製造機械有無異常,未考慮製程之流程。因此,有無法考慮製程之流程以檢知異常之問題。Generally speaking, for a manufacturing machine that obtains a plurality of operation states, when the process flow of the process constituted by the plurality of operation states is abnormal, the manufacturing machine can be considered to be abnormal. However, according to the technique of
本發明係用於解決這種問題所研發出者,其目的係在於提供一種可考慮由複數動作狀態所構成之製程之流程,而進行異常檢知之異常檢知裝置。The present invention was developed to solve such a problem, and an object of the present invention is to provide an abnormality detection device that can perform abnormality detection in consideration of a process flow composed of a plurality of operating states.
實施形態之異常檢知裝置之一側面係包括:特徵量抽出部,接收對象裝置之時間序列數據,以抽出該時間序列數據的特徵量;狀態轉換假設部,自該被抽出之特徵量,特定該對象裝置之動作狀態,同時參照界定複數動作狀態間之狀態轉換模式後之狀態轉換模式資訊,自該被抽出之特徵量,假設該對象裝置之狀態轉換;以及異常檢知部,參照界定針對該狀態轉換模式資訊及各動作狀態之正常範圍後之正常範圍資訊,以自該被特定之動作狀態、該被假設之狀態轉換、及該被抽出之時間序列數據的特徵量,判定該時間序列數據是否係異常。 [發明效果]One aspect of the abnormality detection apparatus according to the embodiment includes: a feature value extraction unit that receives time-series data of the target device and extracts a feature value of the time-series data; and a state transition assumption unit that specifies a feature value from the extracted feature value. The operation state of the target device also refers to the state transition pattern information after defining the state transition pattern between the plurality of operation states, and assumes the state transition of the target device from the extracted feature value; and the abnormality detection unit refers to the definition for The state transition pattern information and the normal range information after the normal range of each operation state are used to determine the time series based on the characteristic quantities from the specified operation state, the assumed state transition, and the extracted time series data Whether the data is abnormal. [Inventive effect]
當依據上述異常檢知裝置之該一側面時,參照狀態轉換模式資訊及正常範圍資訊,而進行異常判定,所以,可使異常判定在狀態轉換之階段與各動作狀態之階段之兩階段進行之。因此,可考慮由複數動作狀態所構成之製程之流程,以進行異常檢知。When referring to the state transition mode information and the normal range information, the abnormality determination is performed according to the one aspect of the abnormality detection device, so that the abnormality determination can be performed in two stages: the state transition stage and the operation state stage. . Therefore, a process flow composed of a plurality of action states can be considered for abnormality detection.
以下,針對本發明之實施形態,參照附圖以說明之。而且,在用於說明實施形態之全部圖面中,係藉對於同一之構造部賦予同一編號,省略重複說明。Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In addition, in all the drawings for explaining the embodiment, the same structural parts are given the same reference numerals, and overlapping descriptions are omitted.
實施形態1.
<推理階段>
(異常檢知系統1000)
參照圖1,說明包括異常檢知裝置100之異常檢知系統1000之整體性構造。圖1係表示包括實施形態1中之異常檢知裝置100及異常檢知裝置100之異常檢知系統1000之構造之方塊圖。如圖1所示,異常檢知系統1000係由做為監視對象之n台(n係1以上之整數)之對象裝置OD1、OD2、...、ODn、使分別由這些對象裝置0D1、OD2、...,ODn所派送之時間序列數據,透過通訊網路NW以接收之異常檢知裝置100、及顯示接受由使用者所做之各種設定,或者,由異常檢知裝置100所做之輸出結果之外部裝置200所構成。在此,所謂之對象裝置可以係單一之裝置,也可以係包含一群複數裝置之系統或設備。
在對象裝置ODn或其附近,係配置有用於取得關於對象裝置ODn之動作之數據,或關於對象裝置ODn之動作環境之數據之未圖示之偵知器。被配置於一個對象裝置之偵知器之數量,可以係一個或複數個。當配置複數偵知器時,複數偵知器可以係同種(例如軸向、水平方向、及鉛直方向之三個加速度偵知器),也可以係異種(例如電壓偵知器與電流偵知器)。取得之數據只要係可當作時間序列數據以取得者即可,其並未特別侷限。在時間序列數據之例中,係包含電流、電壓、電力、速度、加速度、角速度、壓力、磁力、扭力、溫度、濕度、生產數、出貨數、股價、及網際網路之流量數據。時間序列數據可以係偵知器之檢出值本身,也可以係檢出值之統計值(例如平均、最大、最小),也可以係複數偵知器之檢出值之計算值(例如電力)。自對象裝置ODn或其附近所取得之時間序列數據,係被關連到機種或設置處所等之對象裝置ODn之識別資訊,以透過有線或無線之通訊網路NW,而被派送到異常檢知裝置100。The target device ODn or its vicinity is provided with a not-shown detector for acquiring data on the operation of the target device ODn or data on the operating environment of the target device ODn. The number of listeners arranged in a target device may be one or more. When configuring multiple detectors, the multiple detectors can be of the same type (such as three acceleration detectors in the axial, horizontal, and vertical directions), or they can be of different types (such as voltage detectors and current detectors) ). The acquired data is not particularly limited as long as it can be acquired as time series data. Examples of time series data include current, voltage, electricity, velocity, acceleration, angular velocity, pressure, magnetism, torque, temperature, humidity, production numbers, shipments, stock prices, and Internet traffic data. Time series data can be the detected value of the detector itself, or the statistical value of the detected value (such as average, maximum, minimum), or the calculated value of the detected value of a complex detector (such as electricity) . The time-series data obtained from the target device ODn or its vicinity is related to the identification information of the target device ODn of the model or installation location, etc., and is sent to the
異常檢知裝置100係依據自對象裝置OD1、OD2、...、或ODn所接收之時間序列數據,針對接收有時間序列數據之對象裝置,進行狀態轉換假設及異常檢知。又,異常檢知裝置100係被關連到狀態轉換假設及異常檢知之結果與對象裝置之識別資訊,以累積到記憶部4,對應來自外部裝置200之輸入,自記憶部4搜尋已取得之偵知器值之時間序列數據(例如電流值之時間序列數據、網際網路之流量數據之時間序列數據),取得某觀測時刻中之偵知器值、綿延某觀測期間之偵知器值之時間序列數據、及對象裝置之設置處所等資訊。The
(異常檢知裝置100)
異常檢知裝置100係包括接收部1、時間序列數據分析部2、數據紀錄控制部3、記憶部4、界面部5、及時間序列數據匹配部6。(abnormality detection device 100 )
The
(接收部1)
接收部1係接收自對象裝置OD1、OD2、...、ODn所派送之派送數據D1,自接收之派送數據D1,抽出包含偵知器值之種類數、偵知器值之時間序列數據、偵知器項目、及對象裝置之使用環境數據之對象裝置資訊D2。接收部1係使抽出之對象裝置資訊D2,輸出到時間序列數據分析部2與數據紀錄控制部3。(Receiver 1)
The
在此,所謂偵知器值之種類數,係意味自被搭載於對象裝置之偵知器所取得之偵知器數據之種類數。一個偵知器所可取得之數據之種類係僅一個,例如當係僅電壓時,偵知器數據之種類係1。當一個偵知器所可取得之數據之種類係兩個時,例如當係電壓與電流時,偵知器數據之種類係2。而且,即使係同種之偵知器數據,當由物理性不同之偵知器取得時,係對應偵知器之數量,而計數種類數者。例如當使用軸向、水平方向、及鉛直方向之三個加速度偵知器以取得數據時,偵知器數據之種類數係算成3。Here, the number of types of the sensor value means the number of types of sensor data acquired from the sensor mounted on the target device. There is only one type of data that can be obtained by a detector. For example, when it is only voltage, the type of detector data is 1. When there are two types of data that can be obtained by a sensor, for example, when it is voltage and current, the type of sensor data is
所謂偵知器值之時間序列數據,係意味在由偵知器所取得之時間序列之偵知器數據。又,所謂偵知器項目,係意味用於識別自偵知器所取得之偵知器數據之種類或偵知器之設置處所等,被搭載於對象裝置之偵知器之項目。在偵知器數據之種類,係包含關於電流、電壓、扭力、溫度等之對象裝置之運作方式之項目,或關於生產數量或出貨數量等之對象裝置所製造之製品之項目。The so-called time series data of the detector value means the detector data in the time series obtained by the detector. In addition, the so-called sensor item means an item for identifying the type of sensor data acquired from the sensor, the installation location of the sensor, and the like, which are installed in the sensor of the target device. The types of sensor data include items related to the operation of the target device such as current, voltage, torque, temperature, etc., or items related to the products manufactured by the target device such as the production quantity or shipment quantity.
(時間序列數據分析部2)
時間序列數據分析部2係對於自接收部1所接收之對象裝置資訊D2,進行時間序列數據分析處理。時間序列數據分析部2係輸出對象裝置資訊D2的分析結果D3到數據紀錄控制部3。在分析結果D3係包含項目分類資訊、及後述之偵知器值的特徵量。針對時間序列數據分析部2之詳細,係詳述於後。(Time Series Data Analysis Section 2)
The time-series
(數據紀錄控制部3)
數據紀錄控制部3係對應自接收部1所輸入之對象裝置資訊D2、及自時間序列數據分析部2所輸入之分析結果D3,以儲存到記憶部4,藉此,構築資料庫。在此,當在一個對象裝置資訊D2包含複數偵知器值時,也可以在每個偵知器值之項目,生成分析結果D3(D3-1、D3-2、D3-3、...),對應複數分析結果D3(D3-1、D3-2、D3-3、...)到一個對象裝置資訊D2。例如當一個對象裝置資訊D2係包含電壓值及電流值時,也可以生成電壓值之分析結果D3-1、及電流值之分析結果D3-2,對應複數分析結果D3-1及D3-2到一個對象裝置資訊D2。或者,也可以使複數偵知器值之加工數據,總結為一個分析結果D3,對應一個分析結果D3到一個對象裝置資訊D2。(Data logging control section 3)
The data logging control unit 3 stores the target device information D2 input from the receiving
(記憶部4)
記憶部4係記憶各種數據。在實施形態1中,記憶部4係對應於對象裝置資訊D2與分析結果D3以記憶之。又,在實施形態1中,記憶部4係記憶後述之時間序列數據匹配部6所使用之學習過之假設模型D4。(memory section 4)
The memory unit 4 stores various data. In the first embodiment, the storage unit 4 stores the target device information D2 and the analysis result D3 in correspondence with them. Furthermore, in the first embodiment, the memory unit 4 memorizes the learned hypothesis model D4 used by the time-series
在學習過之假設模型D4,係包含對應於複數偵知器值的各項目之特徵量之複數偵知器值的各項目之動作狀態之正常範圍(正常範圍資訊)及狀態轉換模式(狀態轉換模式資訊)。動作狀態之正常範圍,係分別針對複數動作狀態以決定之。The learned hypothetical model D4 includes the normal range (normal range information) and the state transition pattern (state transition) of the operation state of each item of the complex detector value corresponding to the feature quantity of each item of the complex detector value mode information). The normal range of the action state is determined for each of the plurality of action states.
動作狀態之正常範圍,係使用例如以回歸、貝葉斯假設、由統計手法所做之可靠區間、狀態空間模型、3σ法、由CNN等之機器學習所做之手法學習過之範圍。或者,異常檢知裝置100之使用者,也可以定義各動作狀態之正常範圍。The normal range of the action state is the range that has been learned using methods such as regression, Bayesian assumptions, reliability intervals by statistical methods, state space models, 3σ method, and methods by machine learning such as CNN. Alternatively, the user of the
狀態轉換模式係使用例如以因果推論、多元分析、由CNN等之機器學習所做之分類手法等學習過之範圍。或者,當對象裝置之狀態轉換模式,係藉例如對象裝置之規格書而事先瞭解時,異常檢知裝置100之使用者也可以定義狀態轉換模式。狀態轉換模式也可以係複數。The state transition pattern is a range that has been learned using, for example, causal inference, multivariate analysis, classification by machine learning such as CNN, and the like. Alternatively, when the state transition mode of the target device is known in advance by, for example, the specification of the target device, the user of the
又,在實施形態1中,係表示記憶部4記憶對象裝置資訊D2及分析結果D3之構造,但是,也可以採用其他構造。例如也可以取代記憶部4,而被配置於通訊網路NW上之單數或複數之網路儲存裝置(未圖示),記憶對象裝置資訊D2及分析結果D3,數據紀錄控制部3存取該網路儲存裝置之構造。藉此,數據紀錄控制部3係使對象裝置資訊D2與分析結果D3,儲存於外部的網路儲存裝置,可在異常檢知裝置100的外部構築資料庫。又,也可以係時間序列數據匹配部6所使用之學習過之假設模型D4,也不記憶於記憶部4,而記憶於外部的網路儲存裝置之構造。In addition, in
(界面部5)
界面部5係連接外部裝置200與異常檢知裝置100的各部,使溝通或各種控制成為可能。異常檢知裝置100之使用者係可使用外部裝置200,以設定監視數據取得部61搜尋時間序列數據之監視條件D5。又,使用者係可使用外部裝置200,確認時間序列數據匹配部6之匹配結果D6。(Interface part 5)
The
(時間序列數據匹配部6)
時間序列數據匹配部6係進行複數偵知器值的時間序列數據之狀態轉換假設及異常檢知之匹配。在實施形態1中,時間序列數據匹配部6係自記憶部4取得學習過之假設模型D4,同時依據自界面部5輸入之監視條件D5,自記憶部4取得對象裝置資訊D2及分析結果D3,進行偵知器值之匹配。又,時間序列數據匹配部6係輸出偵知器值之匹配結果D6到界面部5,匹配結果D6係透過外部裝置200,以被傳遞到使用者。(Time Series Data Matching Section 6)
The time-series
接著,針對時間序列數據分析部2與時間序列數據匹配部6之詳細構造,參照圖2以說明之。圖2係表示異常檢知裝置100之詳細構造之構造圖。Next, the detailed structures of the time-series
(時間序列數據分析部2之詳細說明)
在實施形態1中,時間序列數據分析部2係包括偵知器項目檢出部21、特徵量抽出部22、及偵知器項目分類部23。(Detailed description of time series data analysis section 2)
In the first embodiment, the time-series
(偵知器項目檢出部21)
偵知器項目檢出部21係參照自接收部1輸入之關於對象裝置資訊D2的偵知器項目之數據、及對象裝置之使用環境數據,以檢出出現於對象裝置資訊D2之對象裝置之各使用環境之偵知器項目。(Detector item detection unit 21)
The sniffer
(特徵量抽出部22)
特徵量抽出部22係自對象裝置資訊的時間序列數據,抽出用於掌握現在之動作狀態、及可自現在之動作狀態轉換往下一動作狀態之狀態轉換之偵知器值的特徵量。在此,於偵知器值的特徵量,雖然包含被包含於偵知器值之波形之在長時間或短時間之傾向、波形之長度、頻率、偵知器值之更新頻率、及某波形(時間序列數據)之複數部分波形間之類似度等之指標,但是,本發明並不侷限於此。偵知器值的特徵量,也可以係在對象裝置資訊於追加而被取得時,決定一定之時間寬度,以如移動平均地連續算出,或者,指定任意之時間點或區間以算出。在此,所謂傾向,係表示最大值或最小值、平均值、分散或標準偏差、相關、斜率、或誤差或殘差等數據之特徵之指標。特徵量也可以自頻率領域抽出。(feature extraction unit 22)
The feature
(偵知器項目分類部23)
偵知器項目分類部23係使被抽出之偵知器值的特徵量,根據偵知器項目以進行分類。在實施形態1中,偵知器項目分類部23係依據偵知器項目檢出部21所檢出之對象裝置之各使用環境之偵知器項目、及特徵量抽出部22所抽出之偵知器值的特徵量,使偵知器值的特徵量根據偵知器項目以進行分類。(Detector Item Classification Section 23)
The snooper
(時間序列數據匹配部6之詳細說明)
接著,說明時間序列數據匹配部6之詳細構造。時間序列數據匹配部6係針對複數偵知器值的時間序列數據以進行匹配,判定各動作狀態之偵知器值的時間序列數據的觀測值是否正常。時間序列數據匹配部6係在記憶部4之對象裝置資訊D2及分析結果D3被更新時,使用學習過之假設模型D4以分析異常,當檢出異常時,輸出發現異常之結果。又,當透過界面部5以自外部裝置200設定監視條件D5時,依據監視條件D5以輸出分析結果。在此,於監視條件D5,係包含例如當作監視對象之區域資訊、當作搜尋對象之時刻資訊、當作搜尋對象之偵知器項目、及當作搜尋對象之偵知器值的特徵量。在實施形態1中,時間序列數據匹配部6係包括監視數據取得部61、加工部62、狀態轉換假設部63、及異常檢知部64。以下,說明各構造部。(Detailed description of time series data matching section 6)
Next, the detailed structure of the time-series
(監視數據取得部61)
監視數據取得部61係取得匹配之一個或複數對象裝置的對象裝置資訊D2、對應於對象裝置資訊D2之分析結果D3、及學習過之假設模型D4。在實施形態1中,監視數據取得部61係自記憶部4搜尋匹配於藉外部裝置200而被設定之監視條件D5之偵知器項目、及綁在此偵知器項目之偵知器值的時間序列數據及特徵量,取得對應之對象裝置資訊D2及分析結果D3。(Monitoring data acquisition unit 61 )
The monitoring
(狀態轉換假設部63)
狀態轉換假設部63係對應由監視數據取得部61所取得之對象裝置資訊D2及分析結果D3、及由記憶部4所取得之學習過之假設模型D4,特定動作中之對象裝置之動作狀態,同時假設狀態轉換。狀態轉換係藉例如做為自某動作狀態往其他動作狀態轉換之機率之狀態轉換機率而被表示。在實施形態1中,狀態轉換假設部63係由包含於對象裝置資訊D2之偵知器值的時間序列數據、偵知器項目、及對象裝置的使用環境數據、及包含於分析結果D3之偵知器值的特徵量,特定現在之動作狀態(第1動作狀態)。又,狀態轉換假設部63係參照包含於學習過之假設模型D4之狀態轉換之模式,使用偵知器值的特徵量,算出自現在之動作狀態(第1動作狀態)往下一動作狀態(第2動作狀態)之狀態轉換機率。自現在之動作狀態轉換之下一動作狀態(第2動作狀態),係如果有一個之情形時,也有複數之情形。狀態轉換之模式,也可由異常檢知裝置100之使用者定義。狀態轉換之模式係例如被保存為複數之動作狀態及表示與複數之動作狀態之轉換關係之狀態轉換表。狀態轉換假設部63係使自監視數據取得部61取得之對象裝置資訊D2及分析結果D3、自記憶部4取得之學習過之假設模型D4、及由狀態轉換假設部63所進行之處理結果(特定結果或假設結果),交付給加工部62。而且,對象裝置資訊D2及分析結果D3,也可以係自監視數據取得部61交付給加工部62。(State Transition Assumption Unit 63)
The state
(加工部62)
加工部62係取得被輸入之對象裝置資訊D2、包含於對應於對象裝置資訊D2之分析結果D3之偵知器值的特徵量、及狀態轉換之機率等之狀態轉換假設部63所進行過之處理結果。又,加工部62係也取得學習過之假設模型D4。加工部62係對應這些取得之數據或資訊與學習過之假設模型D4,針對出現於對象裝置資訊D2之動作中之對象裝置之偵知器值的時間序列數據,將一定機率以上之動作狀態之時間序列數據,視為各動作狀態,實施正常化、往動作狀態單位之拆分、及狀態轉換之模式之算出。加工部62係使這些處理,如圖4A~圖4D所示地進行。(Processing Section 62)
The
正常化之處理,如圖4A所示,也可以對於一連串偵知器值的時間序列數據全體實施,或者,如圖4C所示,在各動作狀態實施。當在時間序列數據之大小有不同時,有時最好實施正常化,但是,也有可能省略正常化。在數據之正常化之例中,係包含例如以公式(1)所表示之min-max正常化、以公式(2)所表示之z正常化、以公式(3)所表示之準位正常化。min-max正常化係使部分列的值域,轉換為0~1。z正常化係將部分列的值域,當作平均0、標準偏差1之轉換。準位正常化係使部分列之平均當作0之轉換。The normalization process may be performed on the entire time-series data of a series of detector values as shown in FIG. 4A , or may be performed in each operating state as shown in FIG. 4C . When there are differences in the size of the time series data, normalization is sometimes preferable, but it is possible to omit normalization. Examples of data normalization include, for example, min-max normalization represented by equation (1), z normalization represented by equation (2), and level normalization represented by equation (3). . The min-max normalization system converts the range of some columns into 0-1. The z normalization system treats the range of some columns as a transformation of
但是,使正常化時間序列數據T後之結果之時間序列數據,標示為TN 。又,i=1,...,n,函數min、max、mean、std分別係Ti,w 之最小值、最大值、平均值、及標準偏差。However, the time series data resulting from normalizing the time series data T is denoted as T N . Also, i=1,...,n, and the functions min, max, mean, and std are the minimum value, maximum value, average value, and standard deviation of T i,w , respectively.
如圖4B所示,加工部62係當對於偵知器值的時間序列數據,未進行正常化時,依據偵知器值的時間序列數據,拆分偵知器值的時間序列數據往各動作狀態A~D。如圖4C所示,加工部62係當對於偵知器值的時間序列數據,進行正常化後,依據被正常化之偵知器值的時間序列數據,拆分被正常化之偵知器值的時間序列數據往各動作狀態A~D。被拆分之動作狀態可以係三個以下,也可以係五個以上。如圖4D所示,加工部62係依據被拆分之動作狀態A~D,實施狀態轉換之模式之算出。As shown in FIG. 4B , the
(異常檢知部64)
異常檢知部64係使用被輸入之對象裝置資訊D2、以加工部62加工過之動作中之對象裝置之偵知器值之動作狀態、及以狀態轉換假設部63假設過之狀態轉換及狀態轉換機率,判定動作中之對象裝置之偵知器值之異常。在實施形態1中,異常檢知部64係包括狀態轉換偏差度算出部641、動作狀態偏差度算出部642、及判定部643。(abnormality detection unit 64)
The
(狀態轉換偏差度算出部641)
狀態轉換偏差度算出部641係對應以監視數據取得部61取得之對象裝置資訊D2、以狀態轉換假設部63假設之狀態轉換及狀態轉換機率、分析結果D3、及學習過之假設模型D4,以算出在動作中之對象裝置之偵知器值之狀態轉換之離開正常之偏差度。在實施形態1中,狀態轉換偏差度算出部641係使用包含於對象裝置資訊D2之偵知器值的時間序列數據、偵知器項目、及對象裝置之使用環境數據、以狀態轉換假設部63假設之狀態轉換及狀態轉換機率、包含於分析結果D3之偵知器值的特徵量、及包含於學習過之假設模型D4之狀態轉換模式及複數偵知器值之各偵知器項目之特徵量,以算出在狀態轉換之離開正常之偏差度。此偏差度之指標,係使用比較被狀態轉換假設部63所假設(自即將被特定之動作狀態)之往現在之動作狀態之轉換機率,與包含於學習過之假設模型D4之狀態轉換模式中之轉換機率,隨著進行與正常不同之狀態轉換之轉換機率變高而變高之指標。「被假設之往現在之動作狀態之轉換機率」之表現,係包含緊接之前之動作狀態係確定,但是,現在之動作狀態係未確定之意味。當一般性地表現時,第1動作狀態係確定,但是,時間性地後續第1動作狀態之第2動作狀態係未確定。因此,「被假設之往現在之動作狀態之轉換機率」之表現,也可表現為「自被特定之現在之動作狀態,往被假設之下一動作狀態之轉換機率」。而且,狀態轉換模式係也可以取代學習過之假設模型D4,而由異常檢知裝置100之使用者設定。(State transition deviation degree calculation unit 641 )
The state transition deviation
(動作狀態偏差度算出部642)
動作狀態偏差度算出部642係對應以監視數據取得部61取得之對象裝置資訊D2、以加工部62加工過之動作中之對象裝置之偵知器值的波形數據及動作狀態、分析結果D3、以狀態轉換假設部63假設之狀態轉換、及學習過之假設模型D4,以算出動作中之對象裝置之偵知器值之動作狀態之離開正常之偏差度。在實施形態1中,動作狀態偏差度算出部642係使用包含於對象裝置資訊D2之偵知器值的時間序列數據、偵知器項目、及對象裝置之使用環境數據、包含於分析結果D3之偵知器值的特徵量、對應於以狀態轉換假設部63假設之狀態轉換之學習過之假設模型D4所包含之各動作狀態之正常範圍、狀態轉換模式、及複數偵知器值之各項目之特徵量,以算出動作狀態之離開正常之偏差度。此偏差度之指標係使用:1)比較自檢知對象逐次取得之波形數據本身,與對應於依據此被逐次取得之波形數據,以狀態轉換假設部63假設之狀態轉換之學習過之假設模型D4所包含之各動作狀態之正常範圍;或2)比較依據自檢知對象逐次取得之波形數據,加工部62加工過之動作中之對象裝置之偵知器值的波形數據,與對應於依據自檢知對象逐次取得之波形數據,而加工部62加工過之動作狀態之學習過之假設模型D4所包含之波形之正常範圍,隨著離開正常範圍之乖離之程度變高而變高之指標。而且,各動作狀態之正常範圍,也可以取代學習過之假設模型D4,而由異常檢知裝置100之使用者設定。(Operation state deviation degree calculation unit 642 )
The operation state deviation
(判定部643)
判定部643係使用以狀態轉換偏差度算出部641與動作狀態偏差度算出部642所算出之結果,判定對象裝置有無異常。在實施形態1中,判定部643係使用以狀態轉換偏差度算出部641與動作狀態偏差度算出部642所算出之結果,判定對象裝置有無異常。判定方法可以針對以狀態轉換偏差度算出部641與動作狀態偏差度算出部642所分別算出之結果,進行AND條件、OR條件等之邏輯運算,以正常或異常之兩值算出,或者,計算離開正常之距離等,以當作異常度之數值算出。(determination unit 643)
The
圖5A~圖5C、圖6A~圖6C、及圖7A~圖7C係表示由判定部643所做之判定之具體例之圖。在學習過之假設模型D4中,針對對象裝置,可能會出現之狀態轉換,係被保持圖5A之狀態轉換圖之關係。在圖5A之狀態轉換圖中,針對對象裝置,可能會出現之狀態轉換之模式,係包含圖5B之模式1與圖5C之模式2。模式1係動作狀態為動作狀態A→動作狀態B→動作狀態C→動作狀態D→動作狀態A地轉換之模式。模式2係動作狀態為動作狀態A→動作狀態C→動作狀態D→動作狀態A地轉換之模式。FIGS. 5A to 5C , FIGS. 6A to 6C , and FIGS. 7A to 7C are diagrams showing specific examples of the determination made by the
圖6A~圖6C係用於說明做為在狀態轉換之異常之例之異常檢知例1之圖。其當作以圖6A之實線表示之時間序列數據係被取得。此時間序列數據之分析之結果,係發現有動作狀態A→動作狀態C之狀態轉換,現在係接著動作狀態C以取得時間序列數據之中途。而且,在本發明實施形態之異常檢知中,係只要瞭解緊接之前之動作狀態即可,所以,只要發現係動作狀態C即足夠,無須發現有動作狀態A→動作狀態C之狀態轉換。在模式1與模式2兩者之中,在動作狀態C之後,可能取得之動作狀態係動作狀態D。動作狀態D在圖6A中,係以虛線表示。但是,現在取得之中途之時間序列數據,係展現追蹤動作狀態A之波形之上升。動作狀態A本身,係對象裝置之可能取得之動作狀態中之一個,所以,為動作狀態A其本身並非異常。但是,動作狀態A係在被學習過之狀態轉換模式中,並非在動作狀態C之後,可能取得之動作狀態,所以,在動作狀態C之後,動作狀態A繼續之情事,係可判定為異常。更具體說來,在動作狀態C之後繼續之現在之時間序列數據,係可確認自動作狀態D之波形有既定乖離之時間點上,係可判定為異常。異常之判定係可以使用如圖6B之各動作狀態之狀態機率,也可以使用如圖6C之現在之時間序列數據之異常度。在異常檢知例1中,做為對象裝置之可能會出現之狀態轉換之模式之做為皆不對應模式1與模式2之轉換之機率(例如往動作狀態A之轉換機率),係變得大於往動作狀態D之轉換機率,所以,轉換被判定為異常。6A to 6C are diagrams for explaining the abnormality detection example 1 which is an example of the abnormality in the state transition. It is taken as time series data represented by the solid line in Fig. 6A. As a result of the analysis of the time series data, it is found that there is a state transition from the action state A to the action state C, and now the action state C is followed to obtain the time series data. Furthermore, in the abnormality detection of the embodiment of the present invention, it is only necessary to know the operation state immediately before, so it is enough to find the operation state C, and it is not necessary to find the state transition from the operation state A to the operation state C. In both
圖7A~圖7C係用於說明做為在某動作狀態之異常之例之異常檢知例2之圖。其當作以圖7A之實線表示之時間序列數據係被取得。此時間序列數據之分析之結果,係發現現在之時間序列數據之緊接之前之動作狀態為動作狀態A,現在係以虛線表示之動作狀態B之波形,與以虛線表示之動作狀態C之波形間之數據被取得之中途。在學習過之假設模型D4中,針對各動作狀態,被判定為正常之範圍係被定義,所以,時間序列數據如果屬於針對動作狀態B之波形,被判定為正常之範圍,或針對動作狀態C之波形,被判定為正常之範圍時,動作狀態B或動作狀態C係可判定為正常。又,圖7A之現在之時間序列數據,係在可確認為並非在這些之任一範圍內之時間點上,可判定為異常。異常之判定,可以使用如圖7B之各動作狀態之狀態機率,或者,如圖7C之現在之時間序列數據之異常度。在異常檢知例2中,狀態轉換雖然1係被判定為做為對象裝置之可能會出現之狀態轉換之模式之模式1與模式2之一者,但是,偵知器值的時間序列數據之波形,係也自模式1與模式2之無論何者之動作狀態之正常範圍偏離,所以,被判定為動作狀態之異常。7A to 7C are diagrams for explaining an abnormality detection example 2 as an example of an abnormality in a certain operating state. It is taken as time series data represented by the solid line in Fig. 7A. As a result of the analysis of the time series data, it is found that the action state immediately before the current time series data is the action state A, the current waveform of the action state B represented by the dotted line, and the waveform of the action state C represented by the dotted line. The data was obtained in between. In the learned hypothetical model D4, the range determined to be normal for each action state is defined. Therefore, if the time-series data belongs to the waveform for the action state B, it is determined to be in the normal range, or for the action state C. When the waveform is judged to be within the normal range, the action state B or the action state C can be judged to be normal. Furthermore, the current time-series data in FIG. 7A can be determined to be abnormal at a time point where it can be confirmed that it does not fall within any of these ranges. The abnormality can be determined by using the state probability of each action state as shown in FIG. 7B , or the abnormality degree of the current time-series data as shown in FIG. 7C . In the abnormality detection example 2, although the
如此一來,異常檢知部64係參照學習過針對狀態轉換模式及各動作狀態之正常範圍之假設模型D4,自被加工部62加工過之動作狀態、被狀態轉換假設部63特定過之動作狀態、被狀態轉換假設部63假設過之狀態轉換、及被抽出之時間序列數據的特徵量,判定時間序列數據是否異常。也可以使用被狀態轉換假設部63假設過之狀態轉換機率。異常檢知部64係參照針對狀態轉換模式(狀態轉換模式資訊)及各動作狀態之正常範圍(正常範圍資訊),以進行異常判定,所以,可使異常判定以狀態轉換之階段與各動作狀態之階段之兩階段進行。因此,當依據實施形態1之異常檢知裝置100時,可考慮由複數動作狀態所構成之製程之流程,以進行異常檢知。In this way, the
又,藉整合這種兩階段中之各判斷結果,即使係自現在之動作狀態轉換之下一動作狀態尚未確定時,也可檢知對象裝置之異常。亦即,可參照被學習過之狀態轉換模式,以假設自現在之動作狀態轉換之下一動作狀態,同時參照被學習過之正常範圍,以假設該被假設之下一動作狀態之正常範圍,所以,在時間序列數據係自該被假設之下一動作狀態之正常範圍偏離後之時間點上,可判定時間序列數據係異常。Furthermore, by integrating the judgment results in these two stages, it is possible to detect an abnormality of the target device even when the next operation state has not been determined since the current operation state transition. That is, you can refer to the learned state transition pattern to assume the next action state from the current action state, and at the same time refer to the learned normal range to assume the normal range of the assumed next action state, Therefore, at a time point after the time series data deviates from the normal range of the assumed next action state, it can be determined that the time series data is abnormal.
又,成為藉算出動作狀態及狀態轉換之機率,可定量性或視覺性地提示異常度,所以,成為可藉異常檢知裝置100之使用者之決策,適切地支援。In addition, the degree of abnormality can be quantitatively or visually indicated by calculating the probability of the operation state and state transition, so that the decision of the user of the
即使係取代學習過之假設模型D4,而使用異常檢知裝置100之使用者所定義之針對狀態轉換模式及各動作狀態之正常範圍時,也同樣可獲得這些優點。These advantages can also be obtained even when the user-defined
接著,參照圖3A及圖3B,說明異常檢知裝置100之硬體構造。茲舉一例,如圖3A所示,異常檢知裝置100係例如藉接收界面101、輸出入界面102、及處理迴路103而被實現。接收界面101係實現接收部1,輸出入界面102係實現界面部5,處理迴路103係實現時間序列數據分析部2、數據紀錄控制部3、記憶部4、及時間序列數據匹配部6。當藉這種硬體構造,而異常檢知裝置100被實現時,後述之圖8之流程圖之步驟S1~S9、及圖10之流程圖之步驟S21~S23,係由處理迴路103執行。處理迴路103係例如單一迴路、複合迴路、程式化過之處理器、並列程式化過之處理器、ASIC(App1ication Specific Integrated Circuit)、FPGA(Fie1d-Programmab1e Gate Array)、或這些之組合。也可使時間序列數據分析部2、數據紀錄控制部3、記憶部4、及時間序列數據匹配部6之功能,以別個之處理迴路實現,或者,集合這些功能而以一個處理迴路實現。Next, the hardware structure of the
茲舉另一例,如圖3B所示,異常檢知裝置100係藉例如接收界面101、輸出入界面102、處理器104、及記憶體105而被實現。接收界面101係實現接收部1,輸出入界面102係實現界面部5。又,藉被儲存於記憶體105之程式被處理器104所讀出以執行,時間序列數據分析部2、數據紀錄控制部3、及時間序列數據匹配部6係被實現。記憶部4係藉記憶體105而被實現。當藉這種硬體構造,而異常檢知裝置100被實現時,後述之圖8之流程圖之步驟S1~S9、及圖10之流程圖之步驟S21~S23,係由處理器104執行。程式係由軟體、韌體、或軟體與韌體之組合而被實現。在記憶體105之例中,例如係包含RAM(Random Access Memory)、ROM(Read Only Memory)、快閃記憶體、EPROM(Erasable Programmable Read Only Memory)、EEPR0M(E1ectrica11y-EPR0M)等非揮發性或揮發性之半導體記憶體、磁碟、軟碟、光碟、CD、MD、DVD。As another example, as shown in FIG. 3B , the
關於時間序列數據分析部2、數據紀錄控制部3、記憶部4、及時間序列數據匹配部6之功能,也可以係以處理迴路實現一部份,或者,以與軟體或韌體協同運作之處理器實現一部份。例如使數據紀錄控制部3以處理迴路實現,使時間序列數據分析部2及時間序列數據匹配部6,藉處理器104讀出被記憶於記憶體105之程式以執行而實現,使記憶部4以記憶體105實現。如此一來,時間序列數據分析部2、數據紀錄控制部3、記憶部4、及時間序列數據匹配部6之功能,係藉處理迴路、與軟體或韌體協同運作之處理器、或這些之組合而被實現。The functions of the time-series
2.動作
接著,針對異常檢知裝置100及異常檢知系統1000之動作,參照圖8以說明之。圖8係表示實施形態1中之異常檢知裝置100之異常檢知處理之流程圖。2. Action
Next, the operation of the
(步驟S1)
在步驟S1中,係藉異常檢知裝置100之使用者,透過外部裝置200而設定監視條件D5。異常檢知裝置100係自外部裝置200取得透過界面部5而被設定之監視條件D5,交付取得之監視條件D5到時間序列數據匹配部6。時間序列數據匹配部6的監視數據取得部61,係依據接收之監視條件D5,決定進行異常檢出之對象裝置。依據此決定,異常檢知裝置100係自進行異常檢出之對象裝置ODn,取得時間序列數據,特徵量抽出部22係自對象裝置ODn之時間序列數據,抽出偵知器值的特徵量,以輸出分析結果D3到紀錄控制部,數據紀錄控制部3係輸出分析結果D3到記憶部。(step S1)
In step S1 , the monitoring condition D5 is set by the user of the
(步驟S2)
在步驟S2中,監視數據取得部61係使用由特徵量抽出部22所算出之記憶部4內的分析結果D3、及記憶部4內之學習過之假設模型D4,取得依據監視條件D5之特徵量。(step S2)
In step S2, the monitoring
(步驟S3)
在步驟S3中,狀態轉換假設部63係以在步驟S2,監視數據取得部61所取得之特徵量為基礎,針對依據監視條件D5之分析結果D3,算出轉換機率。(step S3)
In step S3, the state
(步驟S4)
在步驟S4中,狀態轉換假設部63係以步驟S2,監視數據取得部61所取得之特徵量為基礎,針對依據監視條件D5之分析結果D3,算出波形之類似度。(step S4)
In step S4, the state
(步驟S5)
在步驟S5中,異常檢知部64係以在步驟S3所算出之轉換機率為基礎,針對依據監視條件D5之分析結果D3,算出轉換之異常度。(step S5)
In step S5, the
(步驟S6)
在步驟S6中,異常檢知部64係以在步驟S4所算出之波形之類似度為基礎,針對依據監視條件D5之分析結果D3,算出波形之異常度。(step S6)
In step S6, the
(步驟S7)
在步驟S7中,異常檢知部64係整合在步驟S5及步驟S6所算出之轉換之異常度與波形之異常度。(step S7)
In step S7, the
(步驟S8)
在步驟S8中,異常檢知部64係以在步驟S7所整合之異常度為基礎,針對分析結果D3,判定有無異常。(step S8)
In step S8, the
(步驟S9)
在步驟S9中,異常檢知部64係輸出在步驟S8所判定之判定結果,判定結果係透過界面部5而顯示於外部裝置200。(step S9)
In step S9 , the
<學習階段>
圖9係用於學習異常檢知裝置100所使用之假設模型之學習裝置300之構造圖。學習裝置300係包括學習用數據取得部301及模型生成部302。<Learning stage>
FIG. 9 is a configuration diagram of a
(學習用數據取得部301)
學習用數據取得部301係取得賦予關連到對象裝置之電壓或電流等之各偵知器項目之偵知器值的時間序列數據、及各偵知器項目的特徵量之學習用數據D11。時間序列數據可以係偵知器之檢出值本身,也可以係檢出值之統計值(例如平均、最大、最小),也可以係複數偵知器之檢出值之計算值(例如電力)。(
(模型生成部302)
模型生成部302係依據被學習用數據取得部301取得之學習用數據D11,學習表示對象裝置所取得之複數動作狀態間之轉換之方法或順序之狀態轉換模式,與各動作狀態中之時間序列數據之正常範圍。(Model generation unit 302 )
The
因為進行此學習,模型生成部302係分析時間序列數據,使時間序列數據(以下,有時稱做「波形」。),拆分為對應複數動作狀態之複數波形(以下,有時稱做「部分波形」。)。在此,所謂動作狀態,係意味在大致分類對象裝置之動作時之對象裝置所取得之動作之狀態。茲舉一例,如果係致動器之動作狀態時,其係例如驅動致動器之電流之上升、峰值持續、及下降等之狀態。在時間序列數據之分析中,例如係藉取得表示數據之波形之突變點之事件數據而進行。在突變點之檢出中,例如可使用Ramer-Douglas-Peucker(RDP)算法。又,也可以使用自我相關或動態時間伸縮法(Dynamic Time Warping:DTW)等之手法、主成分分析或判別分析等之多元分析、或支持向量機(SVM)等之假設手法。In order to perform this learning, the
在狀態轉換之模式之學習中,也可以取得表示動作狀態之突變點之事件數據,也可以使用因果推論、多元分析、及由CNN等之機器學習所做之分類手法等。又,也可以可視化動作狀態之出現順序,以提示異常檢知裝置100之使用者,使用者定義狀態轉換模式。In the learning of state transition patterns, event data representing mutation points of action states can also be obtained, and causal inference, multivariate analysis, and classification methods by machine learning such as CNN can also be used. In addition, the appearance sequence of the action states can also be visualized to prompt the user of the
茲舉一例,因為學習狀態轉換之模式,模型生成部302係對於被拆分之部分波形,首先分配用於識別動作狀態之動作狀態ID(例如A、B、C、D等)。在製品製造數據中,因為連續製造同一或類似之製品,所以,想成在同一動作之部分波形之形狀係類似。在此,動作狀態ID之分配係可當作聚類問題而加以捕捉。也想成製造機械之動作之順序係非一定,而係依據複數狀態轉換模式而動作之情形,所以,也可以在各動作狀態,亦即,在各部分波形之形狀聚類。具體說來,可利用做為無教師學習之一手法之k-means法。藉使用此手法,可在各類似形狀聚類部分波形,在各類似形狀,編號動作狀態ID。For example, to learn the mode of state transition, the
如此一來,藉動作狀態ID被編號,複數動作狀態間之狀態轉換模式、及自某動作狀態往其他動作狀態之狀態轉換之機率,也成為可自動學習。In this way, with the action state IDs being numbered, the state transition pattern between plural action states and the probability of state transition from one action state to another action state can also be learned automatically.
不僅動作狀態ID,也一併學習狀態轉換,藉此,即使表示有某部分波形之動作狀態之部分波形係正常時,也可檢知動作狀態之轉換之方法係異常,亦即,採取該動作狀態之本身係異常。Not only the action state ID, but also state transitions are learned, so that even if a part of the waveform indicating the action state of a certain part of the waveform is normal, it is possible to detect that the transition of the action state is abnormal, that is, take the action. The state itself is abnormal.
在各動作狀態之正常範圍之學習,也可使用回歸、或由貝葉斯假設或統計手法所做之可靠區間、狀態空間模型、3σ法或由CNN等之機器學習所做之手法。茲舉一例,針對拆分而可得之各動作狀態之部分波形,求出在各時刻之數據分佈,學習發生機率成為α%(例如α=1%)以上之部分係正常範圍。學習之結果,也可以學習針對全部之動作狀態,正常範圍成為發生機率係α=1%以上之部分,也可以學習針對某動作狀態,正常範圍成為發生機率係α=1%以上之部分,其中,針對另一動作狀態,正常範圍成為發生機率係α=2%以上之部分。數據之分佈情況係在各動作狀態有所不同,所以,針對各動作狀態,即使統計量被定義為同一之正常範圍,物理量也在各動作狀態,分別被設定範圍。如此一來,在各動作狀態學習正常範圍,藉此,成為可設定對應各動作狀態之動作之特徵之正常範圍,成為可高精度地異常檢知。In the learning of the normal range of each action state, regression, or reliable intervals by Bayesian assumptions or statistical methods, state space models, 3σ methods, or methods by machine learning such as CNN can also be used. As an example, for the partial waveform of each action state obtained by splitting, the data distribution at each time is obtained, and the part where the probability of learning occurrence is more than α% (for example, α=1%) is in the normal range. As a result of the learning, it is also possible to learn that for all action states, the normal range becomes the part where the probability of occurrence is α=1% or more, and it is also possible to learn that for a certain action state, the normal range becomes the part where the probability of occurrence is α=1% or more, among which , for another action state, the normal range is the part where the probability of occurrence is α=2% or more. The distribution of data differs in each operating state. Therefore, even if the statistic is defined in the same normal range for each operating state, the physical quantity is also set within each operating state, and the range is set separately. In this way, the normal range is learned for each operation state, whereby a normal range can be set in which the characteristics of the operation corresponding to each operation state can be set, and abnormality can be detected with high accuracy.
例如針對某動作狀態1,正常範圍係R1,針對另一動作狀態2,正常範圍係R2,假設R1係比R2還要窄。For example, for a
如果,當針對兩動作狀態,正常範圍一律定義為R1時,針對動作狀態1,雖然可適切地異常判定,但是,針對動作狀態2,係無法適切地異常判定。因為針對動作狀態2,判定為正常之範圍係被設定為比本來之R2還要窄(針對動作狀態2,正常範圍過小)。因此,針對動作狀態2,係不管是否在正常地動作,有時係被判定為異常。If the normal range is uniformly defined as R1 for both operation states, the
反之,當針對同動作狀態,正常範圍一律定義為R2時,針對動作狀態2,雖然可適切地異常判定,但是,針對動作狀態1,係無法適切地異常判定。因為針對動作狀態1,被判定為正常之範圍係被設定為比本來之R1還要寬(針對動作狀態1,正常範圍過大)。因此,針對動作狀態1,係不管是否在異常地動作,有時係被判定為正常。Conversely, when the normal range is uniformly defined as R2 for the same operation state, although the
如此一來,當針對複數動作狀態,欲一律設定正常範圍時,在針對某動作狀態之正常範圍與針對另一動作狀態之正常範圍之間,產生權衡,而針對複數動作狀態之全部,設定正常範圍係較困難。In this way, when it is desired to uniformly set the normal range for the plural action states, a trade-off occurs between the normal range for one action state and the normal range for another action state, and for all the plural action states, the normal range is set. Scope is more difficult.
如本發明之實施形態1所示,藉針對複數之各動作狀態,個別學習正常範圍,成為可解決該種權衡,以針對複數動作狀態之全部,適切地設定正常範圍。As shown in the first embodiment of the present invention, by learning the normal range individually for each of the plural operation states, such a trade-off can be solved, and the normal range can be appropriately set for all the plural operation states.
而且,學習裝置300係為了學習異常檢知裝置100所使用之學習模型而被使用,但是,例如也可以係透過網路以被連接於異常檢知裝置100,與此異常檢知裝置100係別個之裝置。或者,學習裝置300也可以係被內建於異常檢知裝置100。Furthermore, the
模型生成部302係藉執行如以上之學習,生成做為狀態轉換模式(狀態轉換模式資訊)及各動作狀態之正常範圍(正常範圍資訊)之學習結果之學習過之假設模型D12,輸出到異常檢知裝置100的記憶部4。The
(記憶部4)
記憶部4係記憶自模型生成部302輸出之學習過之假設模型D12。(memory section 4)
The memory unit 4 stores the learned hypothesis model D12 output from the
接著,使用圖10,說明學習裝置300之學習處理。圖10係表示學習裝置300之學習處理之流程圖。Next, the learning process of the
(步驟S21)
在步驟S21中,學習用數據取得部301係取得賦予關連到對象裝置之各偵知器項目之偵知器值的時間序列數據,與各偵知器項目的特徵量之學習用數據D11。(step S21)
In step S21, the
(步驟S22)
在步驟S22中,模型生成部302係依據被學習用數據取得部301所取得之學習用數據D11,分析對象裝置之各偵知器項目之偵知器值的時間序列數據,拆分時間序列數據往複數動作狀態,學習被拆分後之動作狀態間之狀態轉換模式、及各動作狀態中之時間序列數據之正常範圍,以生成假設模型。(step S22)
In step S22, the
(步驟S23)
在步驟S23中,記憶部4係記憶模型生成部302所生成之學習過之假設模型D12。(step S23)
In step S23, the memory unit 4 memorizes the learned hypothesis model D12 generated by the
<變形例>
在狀態轉換假設及異常檢知中,時間序列數據匹配部6所匹配之時間序列數據,可以係被即時取得之數據,或者,過去已取得之數據。又,也可以對於取得之偵知器值的時間序列數據之全部,進行異常檢知。<Variation>
In the state transition assumption and abnormality detection, the time-series data matched by the time-series
<附記>
以下,針對實施形態之一部份之側面做整理。
(附記1)
異常檢知裝置(100)係包括:特徵量抽出部(22),接收對象裝置之時間序列數據,以抽出該時間序列數據的特徵量;狀態轉換假設部(63),自該被抽出之特徵量,特定該對象裝置之動作狀態,同時參照界定複數動作狀態間之狀態轉換模式之狀態轉換模式資訊,以自該被抽出之特徵量,假設該對象裝置之狀態轉換;以及異常檢知部(64),參照界定針對該狀態轉換模式資訊及各動作狀態之正常範圍後之正常範圍資訊,自該被特定之動作狀態、該被假設之狀態轉換、及該被抽出之時間序列數據的特徵量,判定該時間序列數據是否係異常。
(附記2)
附記2之異常檢知裝置,附記1所述之異常檢知裝置,其中該狀態轉換假設部係藉算出狀態轉換機率,進行該狀態轉換之假設。
(附記3)
附記3之異常檢知裝置,係附記2所述之異常檢知裝置,其中該異常檢知部係參照該狀態轉換模式資訊,取得自該被特定之動作狀態轉換之動作狀態,當該被假設之狀態轉換之狀態轉換機率,係高於往該被取得之動作狀態轉換之機率時,判定該被假設之狀態轉換係異常。
(附記4)
附記4之異常檢知裝置,係附記1~3之任一者所述之異常檢知裝置,其中該異常檢知部係參照該正常範圍資訊,以取得該被特定之動作狀態中之正常範圍,當對應於該時間序列數據之該被特定之動作狀態之波形,自該正常範圍偏離後,該時間序列數據係當作波形數據而判定為異常。
(附記5)
附記5之異常檢知裝置,係附記1~4之任一者所述之異常檢知裝置,其中該異常檢知部係包括:狀態轉換偏差度算出部(641),算出離開該被假設之狀態轉換之正常狀態轉換之狀態轉換偏差度;動作狀態偏差度算出部(642),算出自該被特定之動作狀態中之該時間序列數據之波形之正常範圍之動作狀態偏差度;以及判定部,依據該狀態轉換偏差度及該動作狀態偏差度,判定該對象裝置是否係異常。
(附記6)
附記6之異常檢知裝置,係附記1~5之任一者所述之異常檢知裝置,其中該狀態轉換模式資訊及該正常範圍資訊,係當作由機器學習所做之學習過之假設模型所得之資訊。
(附記7)
附記7之異常檢知裝置,係附記6所述之異常檢知裝置,其中還包括學習該假設模型之學習裝置(300),該學習裝置係包括自時間序列數據的特徵量,算出各狀態轉換模式及動作狀態之正常範圍之模型生成部(302)。<Additional Notes>
Hereinafter, the side surface of a part of the embodiment will be sorted out.
(Note 1)
An abnormality detection device (100) includes: a feature value extraction unit (22) that receives time-series data of a target device and extracts a feature value of the time-series data; a state transition assumption unit (63) that extracts features from the extracted features Quantity, specifying the operating state of the target device, while referring to the state transition pattern information that defines the state transition pattern between the plurality of operating states, and assuming the state transition of the target device from the extracted feature quantity; and the abnormality detection section ( 64), with reference to the normal range information after defining the normal range for the state transition pattern information and each action state, from the specified action state, the assumed state transition, and the extracted time-series data feature quantities , to determine whether the time series data is abnormal.
(Supplement 2)
The abnormality detection device of
而且,也可以進行實施形態之組合或變形,也可以省略任意之構造部。 [產業上之利用可能性]Furthermore, a combination or modification of the embodiment may be performed, and any structural portion may be omitted. [Possibilities of Industrial Use]
本發明之異常檢知裝置,係可檢知時間序列數據之異常,所以,例如可當作工廠的生產系統之用於預防保養之裝置、鐵路或車站等之社會基礎設施所使用之系統或裝置之用於預防保養之裝置、股價等之經濟指標之監視裝置使用。The abnormality detection device of the present invention can detect abnormality of time-series data, so, for example, it can be used as a preventive maintenance device in a production system of a factory, a system or device used in social infrastructure such as railways or stations, etc. It is used for monitoring equipment of economic indicators such as equipment for preventive maintenance and stock price.
1:接收部 2:時間序列數據分析部 3:數據紀錄控制部 4:記憶部 5:界面部 6:時間序列數據匹配部 21:偵知器項目檢出部 22:特徵量抽出部 23:偵知器項目分類部 61:監視數據取得部 62:加工部 63:狀態轉換假設部 64:異常檢知部 100:異常檢知裝置 101:接收界面 102:輸出入界面 103:處理迴路 104:處理器 105:記憶體 200:外部裝置 300:學習裝置 301:學習用數據取得部 302:模型生成部 641:狀態轉換偏差度算出部 642:動作狀態偏差度算出部 643:判定部 1000:異常檢知系統1: Receiving Department 2: Time Series Data Analysis Department 3: Data Recording Control Department 4: Memory Department 5: Interface Department 6: Time Series Data Matching Section 21: Detector Item Checkout Department 22: Feature extraction part 23: Detector Item Classification Department 61: Monitoring data acquisition department 62: Processing Department 63: State Transition Assumption Department 64: Anomaly Detection Department 100: Abnormal detection device 101: Receive interface 102: I/O interface 103: Processing Loops 104: Processor 105: Memory 200: External device 300: Learning Devices 301: Data Acquisition Department for Learning 302: Model Generation Department 641: State transition deviation degree calculator 642: Operation state deviation degree calculator 643: Judgment Department 1000: Anomaly Detection System
〔圖1〕係顯示實施形態1之異常檢知裝置及異常檢知裝置系統之構造之方塊圖。
〔圖2〕係顯示實施形態1之異常檢知裝置之詳細構造之方塊圖。
〔圖3A〕係顯示實施形態1之異常檢知裝置之硬體構造例之方塊圖。
〔圖3B〕係顯示實施形態1之異常檢知裝置之其他硬體構造例之方塊圖。
〔圖4A〕係顯示時間序列數據全體之正常化之圖。
〔圖4B〕係顯示時間序列數據之往各動作狀態之拆分之例之圖。
〔圖4C〕係顯示時間序列數據之在各動作狀態之正常化之圖。
〔圖4D〕係顯示狀態轉換之模式之算出例之圖。
〔圖5A〕係顯示可能會出現之狀態轉換之圖。
〔圖5B〕係顯示可能會出現之狀態轉換所包含之模式1之圖。
〔圖5C〕係顯示可能會出現之狀態轉換所包含之模式2之圖。
〔圖6A〕係某時間序列數據之時間表。
〔圖6B〕係顯示圖6A之時間序列數據之動作狀態之狀態機率之圖。
〔圖6C〕係顯示圖6A之時間序列數據之異常度之判定之圖。
〔圖7A〕係某時間序列數據之時間表。
〔圖7B〕係顯示圖7A之時間序列數據之動作狀態之狀態機率之圖。
〔圖7C〕係顯示圖7A之時間序列數據之異常度之判定之圖。
〔圖8〕係實施形態1中之異常檢知裝置之異常檢知處理之流程圖。
〔圖9〕係用於學習實施形態1之異常檢知裝置所使用之假設模型之學習裝置之構造圖。
〔圖10〕係顯示學習裝置之學習處理之流程圖。[FIG. 1] is a block diagram showing the structure of the abnormality detection apparatus and the abnormality detection apparatus system of
1:接收部1: Receiving Department
2:時間序列數據分析部2: Time Series Data Analysis Department
3:數據紀錄控制部3: Data Recording Control Department
4:記憶部4: Memory Department
5:界面部5: Interface Department
6:時間序列數據匹配部6: Time Series Data Matching Section
21:偵知器項目檢出部21: Detector Item Checkout Department
22:特徵量抽出部22: Feature extraction part
23:偵知器項目分類部23: Detector Item Classification Department
61:監視數據取得部61: Monitoring data acquisition department
62:加工部62: Processing Department
63:狀態轉換假設部63: State Transition Assumption Department
64:異常檢知部64: Anomaly Detection Department
100:異常檢知裝置100: Abnormal detection device
641:狀態轉換偏差度算出部641: State transition deviation degree calculator
642:動作狀態偏差度算出部642: Operation state deviation degree calculator
643:判定部643: Judgment Department
D2:對象裝置資訊D2: Target device information
D3:分析結果D3: Analysis results
Claims (7)
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2020/034581 WO2022054256A1 (en) | 2020-09-11 | 2020-09-11 | Abnormality detection device |
WOPCT/JP2020/034581 | 2020-09-11 |
Publications (1)
Publication Number | Publication Date |
---|---|
TW202210977A true TW202210977A (en) | 2022-03-16 |
Family
ID=80631449
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
TW109144679A TW202210977A (en) | 2020-09-11 | 2020-12-17 | Abnormality detection device |
Country Status (3)
Country | Link |
---|---|
JP (1) | JP7158624B2 (en) |
TW (1) | TW202210977A (en) |
WO (1) | WO2022054256A1 (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2023218550A1 (en) * | 2022-05-11 | 2023-11-16 | 三菱電機株式会社 | Correction equipment, processing method, and processing program |
WO2024047694A1 (en) * | 2022-08-29 | 2024-03-07 | 三菱電機株式会社 | Operation assistance device, operation assistance system, and operation assistance method |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013030984A1 (en) * | 2011-08-31 | 2013-03-07 | 株式会社日立エンジニアリング・アンド・サービス | Facility state monitoring method and device for same |
JP6561980B2 (en) * | 2014-03-24 | 2019-08-21 | 日本電気株式会社 | Monitoring device, monitoring system, monitoring method and program |
US10712733B2 (en) * | 2016-12-12 | 2020-07-14 | Mitsubishi Electric Research Laboratories, Inc. | Methods and systems for discovery of prognostic subsequences in time series |
JP6810097B2 (en) * | 2018-05-21 | 2021-01-06 | ファナック株式会社 | Anomaly detector |
US11782395B2 (en) * | 2018-06-18 | 2023-10-10 | Mitsubishi Electric Corporation | Diagnostic device, diagnostic method and program |
-
2020
- 2020-09-11 WO PCT/JP2020/034581 patent/WO2022054256A1/en active Application Filing
- 2020-09-11 JP JP2022510985A patent/JP7158624B2/en active Active
- 2020-12-17 TW TW109144679A patent/TW202210977A/en unknown
Also Published As
Publication number | Publication date |
---|---|
JPWO2022054256A1 (en) | 2022-03-17 |
JP7158624B2 (en) | 2022-10-21 |
WO2022054256A1 (en) | 2022-03-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Cook et al. | Anomaly detection for IoT time-series data: A survey | |
EP3460611B1 (en) | System and method for aircraft fault detection | |
EP3373089B1 (en) | Operating state classification device | |
US8682824B2 (en) | Method and device for monitoring the state of a facility | |
US20190310979A1 (en) | Anomaly data priority assessment device and anomaly data priority assessment method | |
US8732100B2 (en) | Method and apparatus for event detection permitting per event adjustment of false alarm rate | |
WO2012032812A1 (en) | Malfunction detection method and system thereof | |
US20140046878A1 (en) | Method and system for detecting sound events in a given environment | |
TW202210977A (en) | Abnormality detection device | |
CN112818066A (en) | Time sequence data anomaly detection method and device, electronic equipment and storage medium | |
US20120296606A1 (en) | Method, computer program, and system for performing interpolation on sensor data for high system availability | |
US20070239629A1 (en) | Cluster Trending Method for Abnormal Events Detection | |
JP2004531815A (en) | Diagnostic system and method for predictive condition monitoring | |
KR102472637B1 (en) | Method for analyzing time series data, determining a key influence variable and apparatus supporting the same | |
US11755448B2 (en) | Event monitoring apparatus, method and program recording medium | |
KR102433598B1 (en) | A System and Method for Deriving Data Boundary | |
KR20200074825A (en) | Method and System for Power Data Outlier Detection Based on Pattern Tagging | |
JPWO2019012653A1 (en) | Learning system, analysis system, learning method and storage medium | |
CN113330379A (en) | Abnormality cause estimation device, abnormality cause estimation method, and program | |
US20230251167A1 (en) | Learning device, defect detection device, and defect detection method | |
US11954131B2 (en) | Time-series data processing method | |
US20230102000A1 (en) | Time-series pattern explanatory information generating apparatus | |
US12019433B2 (en) | Periodicity analysis apparatus, method and program recording medium | |
Manca et al. | Convolutional kernel-based transformation and clustering of similar industrial alarm floods | |
US20220138624A1 (en) | Time-series data processing method |