TWI795719B - data processing device - Google Patents

data processing device Download PDF

Info

Publication number
TWI795719B
TWI795719B TW110102460A TW110102460A TWI795719B TW I795719 B TWI795719 B TW I795719B TW 110102460 A TW110102460 A TW 110102460A TW 110102460 A TW110102460 A TW 110102460A TW I795719 B TWI795719 B TW I795719B
Authority
TW
Taiwan
Prior art keywords
time
series data
aforementioned
data
series
Prior art date
Application number
TW110102460A
Other languages
Chinese (zh)
Other versions
TW202219770A (en
Inventor
中村隆顕
Original Assignee
日商三菱電機股份有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 日商三菱電機股份有限公司 filed Critical 日商三菱電機股份有限公司
Publication of TW202219770A publication Critical patent/TW202219770A/en
Application granted granted Critical
Publication of TWI795719B publication Critical patent/TWI795719B/en

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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3065—Monitoring arrangements determined by the means or processing involved in reporting the monitored data
    • G06F11/3072—Monitoring arrangements determined by the means or processing involved in reporting the monitored data where the reporting involves data filtering, e.g. pattern matching, time or event triggered, adaptive or policy-based reporting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Automation & Control Theory (AREA)
  • Quality & Reliability (AREA)
  • Testing And Monitoring For Control Systems (AREA)
  • Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
  • Electrical Discharge Machining, Electrochemical Machining, And Combined Machining (AREA)
  • Memory System Of A Hierarchy Structure (AREA)

Abstract

資料處理裝置(100)包含：時序取得部(110)，取得時序資料；特徵取出部(120)，基於時序取得部(110)取得的時序資料，取出時序資料的特徵量；第1分割部(130)，基於特徵取出部(120)取出的特徵量，將時序資料分割為複數個第1部分時序資料；以及第2分割部(140)，基於特徵取出部(120)取出的特徵量，將第1分割部(130)分割的複數個第1部分時序資料的每一個分割為複數個第2部分時序資料。 The data processing device (100) comprises: a time-series acquisition part (110), which obtains time-series data; a feature extraction part (120), which extracts the feature quantity of the time-series data based on the time-series data obtained by the time-series acquisition part (110); the first division part ( 130), based on the feature quantity extracted by the feature extraction unit (120), the time series data is divided into a plurality of first part time series data; and the second division unit (140), based on the feature quantity extracted by the feature extraction unit (120), the Each of the plurality of first partial time-series data divided by the first dividing unit (130) is divided into a plurality of second partial time-series data.

Description

資料處理裝置 data processing device

本揭露關於資料處理裝置以及資料處理方法。 This disclosure relates to a data processing device and a data processing method.

將時序資料分割為複數個部分時序資料的技術已為人所知。 Techniques for dividing time-series data into a plurality of partial time-series data are known.

舉例來說，專利文獻1揭露了一種資料處理裝置，包含：取出條件輸入部，受理包含機器的狀態的變化點的波形資料、該波形資料的參數資訊、以及機器的遷移資訊的輸入；相似度算出部，算出機器的時序資料與波形資料的相似度；運轉模式判定部，基於機器的遷移資訊，設定機器的狀態；變化點檢測部，基於算出的相似度、以及判定的機器的狀態，從機器的時序資料之中檢測變化點，設定時序資料的部分列，也就是設定程序段(Segment)的開始時刻以及程序段的結束時刻；以及資訊輸出部，輸出程序段的開始時刻以及程序段的結束時刻，作為程序段資訊。 For example, Patent Document 1 discloses a data processing device, including: an extraction condition input unit that accepts input of waveform data including change points of the state of the machine, parameter information of the waveform data, and transition information of the machine; similarity The calculation part calculates the similarity between the time series data and the waveform data of the machine; the operation mode determination part sets the state of the machine based on the migration information of the machine; the change point detection part based on the calculated similarity and the determined state of the machine, from Detect change points in the time series data of the machine, set the partial column of the time series data, that is, set the start time of the program segment (Segment) and the end time of the program segment; and the information output part, output the start time of the program segment and the end time of the program segment The end time is used as the block information.

專利文獻1記載的資料處理裝置(以下稱為「既有的資料處理裝置」)，使用事先準備好作為用來分割的指標的波形資料以及波形資料的參數資訊，從時序資料之中檢測波型(Wave Pattern)，藉以針對每一個在時序資料中反覆出現的波形資料，將時序資料分割為程序段，也就是分割為部分時序資料。 The data processing device described in Patent Document 1 (hereinafter referred to as "existing data processing device") detects waveforms from time-series data using waveform data prepared in advance as indicators for division and parameter information of the waveform data. (Wave Pattern), so as to divide the time-series data into program segments for each waveform data repeatedly appearing in the time-series data, that is, divide it into partial time-series data.

[先前技術文獻] [Prior Art Literature] [專利文獻] [Patent Document]

專利文獻1：國際公開2020/008533號公報Patent Document 1: International Publication No. 2020/008533

[發明所欲解決的課題][Problems to be Solved by the Invention]

既有的資料處理裝置有以下問題點：為了將時序資料分割為程序段，也就是分割為部分時序資料，因此必須事先準備好作為用來分割的指標的波形資料、以及波形資料的參數資訊（以下稱為「特別資訊」）。Existing data processing devices have the following problems: In order to divide the time series data into program segments, that is, into partial time series data, it is necessary to prepare in advance the waveform data used as the index for division and the parameter information of the waveform data ( hereinafter referred to as "Special Information").

本揭露是為了解決上述問題點，目的在於提供一種資料處理裝置，在沒有事先準備好特別資訊的情況下，也可以針對每一個在時序資料中反覆出現的波型，將時序資料分割為部分時序資料。 [用以解決課題的手段] The purpose of this disclosure is to solve the above problems, and the purpose is to provide a data processing device that can divide the time-series data into partial time-series for each wave pattern that repeatedly appears in the time-series data without preparing special information in advance material. [Means to solve the problem]

關於本揭露的資料處理裝置，包含：時序取得部，取得時序資料；特徵取出部，基於時序取得部取得的時序資料，取出時序資料的特徵量；第1分割部，基於特徵取出部取出的特徵量，將時序資料分割為複數個第1部分時序資料；以及第2分割部，基於特徵取出部取出的特徵量，將第1分割部分割的複數個第1部分時序資料的每一個分割為複數個第2部分時序資料。 [發明的效果] The data processing device disclosed in this disclosure includes: a time-series acquisition unit, which acquires time-series data; a feature extraction unit, which extracts the feature quantity of the time-series data based on the time-series data obtained by the time-series acquisition unit; a first segmentation unit, based on the features extracted by the feature extraction unit The amount of time series data is divided into a plurality of first part time series data; and the second division part is based on the feature quantity extracted by the feature extraction part, and each of the plurality of first part time series data divided by the first division part is divided into a plurality of Part 2 timing data. [Effect of the invention]

根據本揭露，在沒有事先準備好特別資訊的情況下，也可以針對每一個在時序資料中反覆出現的波型，將時序資料分割為部分時序資料。According to the present disclosure, without preparing special information in advance, the time-series data can also be divided into partial time-series data for each waveform that appears repeatedly in the time-series data.

以下，針對本揭露的實施形態，一邊參照圖式一邊詳細說明。Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings.

實施形態1. 參照第1圖至第5圖，針對關於實施形態1的資料處理裝置100、以及應用資料處理裝置100的資料處理系統1進行說明。 第1圖為一方塊圖，示意關於實施形態1的資料處理裝置100、以及應用資料處理裝置100的資料處理系統1的重點的構成的一例。 關於實施形態1的資料處理系統1，包含記憶裝置10、顯示裝置20、以及資料處理裝置100。 Implementation form 1. Referring to Fig. 1 to Fig. 5, the data processing device 100 of Embodiment 1 and the data processing system 1 to which the data processing device 100 is applied will be described. Fig. 1 is a block diagram showing an example of a key configuration of a data processing device 100 according to Embodiment 1 and a data processing system 1 to which the data processing device 100 is applied. The data processing system 1 of Embodiment 1 includes a memory device 10 , a display device 20 , and a data processing device 100 .

資料處理裝置100取得時序資料，針對取得的時序資料，執行事先設定的資料處理。針對資料處理裝置100的詳細將於後面描述。 記憶裝置10是記憶資料處理裝置100執行資料處理時必要的資訊的裝置。舉例來說，記憶裝置10事先記憶時序資料。 顯示裝置20是顯示器等的裝置，顯示輸出資料處理裝置100輸出的顯示資訊所示的顯示影像。換言之，資料處理裝置100將顯示資訊輸出至顯示裝置20，讓顯示資訊所示的顯示影像在顯示裝置20顯示。 The data processing device 100 acquires time-series data, and executes preset data processing on the acquired time-series data. The details of the data processing device 100 will be described later. The memory device 10 is a device that stores information necessary for data processing performed by the data processing device 100 . For example, the memory device 10 stores time series data in advance. The display device 20 is a device such as a monitor, and displays a display image indicated by the display information output by the output data processing device 100 . In other words, the data processing device 100 outputs the display information to the display device 20 , so that the display image indicated by the display information is displayed on the display device 20 .

關於實施形態1的資料處理裝置100，包含：時序取得部110、特徵取出部120、第1分割部130、第2分割部140、以及異常偵測部150及輸出部190。 另外，異常偵測部150在資料處理裝置100當中並非必要的構成。 實施形態1中，是以資料處理裝置100包含異常偵測部150為前提下進行說明。 The data processing device 100 of Embodiment 1 includes: a sequence acquisition unit 110 , a feature extraction unit 120 , a first division unit 130 , a second division unit 140 , an abnormality detection unit 150 , and an output unit 190 . In addition, the abnormality detection unit 150 is not an essential structure in the data processing device 100 . Embodiment 1 is described on the premise that the data processing device 100 includes the abnormality detection unit 150 .

時序取得部110取得時序資料。 所謂「時序取得部110取得的時序資料」，是表示以事先設定的時間間隔，進行測量、測定、觀測、或統計等的物理量的時序資訊。 具體來說，時序取得部110取得的時序資料，是將振動感測器、旋轉感測器、陀螺儀感測器、溫度感測器、或聲音感測器等感測器輸出的訊號轉換為時序資訊的訊號。若時序資料是表示以事先設定的時間間隔，進行測量、測定、觀測、或統計等的物理量的時序資訊，則不限於表示將感測器輸出的訊號轉換為時序資料之資料。另外，事先設定的時間間隔，並不需要為均一間隔，時間間隔包含了任意間隔。 The time-series acquisition unit 110 acquires time-series data. The "time-series data acquired by the time-series acquisition unit 110" is time-series information representing physical quantities such as measurement, measurement, observation, or statistics at predetermined time intervals. Specifically, the timing data acquired by the timing acquisition unit 110 is to convert signals output from sensors such as vibration sensors, rotation sensors, gyroscope sensors, temperature sensors, or sound sensors into Signal of timing information. If time-series data is time-series information representing physical quantities that are measured, measured, observed, or counted at predetermined time intervals, it is not limited to data that represents the conversion of signals output by sensors into time-series data. In addition, the time interval set in advance does not need to be a uniform interval, and the time interval includes any interval.

具體而言，舉例來說，時序取得部110藉由讀取事先記憶於記憶裝置10的時序資料，以取得時序資料。 由於時序取得部110只要是可以取得時序資料的元件即可，因此，針對時序取得部110取得的時序資料的取得來源，或是時序取得部110取得時序資料的方法並沒有限定。 以下，在實施形態1中，假設時序取得部110取得的時序資料，是設置於加工裝置的振動感測器輸出的訊號轉換為時序資訊的訊號進行說明。 Specifically, for example, the timing acquisition unit 110 acquires the timing data by reading the timing data previously stored in the memory device 10 . As long as the time series acquisition unit 110 is a component capable of acquiring time series data, there is no limitation on the source of the time series data acquired by the time series acquisition unit 110 or the method by which the time series acquisition unit 110 acquires the time series data. Hereinafter, in Embodiment 1, it will be described assuming that the timing data acquired by the timing acquisition unit 110 is a signal converted from a signal output by a vibration sensor provided in a processing device into timing information.

特徵取出部120基於時序取得部110取得的時序資料，取出時序資料的特徵量。 具體而言，舉例來說，特徵取出部120使用視窗，從時序取得部110取得的時序資料之中取出特徵量。換言之，特徵取出部120輸出的特徵量，成為表示滑動視窗對應的時序資料當中的區域的特徵的值以時序表示之資訊。 更具體而言，舉例來說，特徵取出部120使用滑動視窗，將滑動視窗內的時序資料的最大值減掉最小值的值（以下稱為「振幅值」）作為特徵量並取出。特徵取出部120輸出的特徵量，成為滑動視窗對應的時序資料當中的區域的振幅值以時序表示之資訊。 由於使用滑動視窗從時序資料之中取出特徵量的方法已為眾所皆知，故省略說明。 The feature extraction unit 120 extracts feature quantities of the time-series data based on the time-series data acquired by the time-series acquisition unit 110 . Specifically, for example, the feature extraction unit 120 extracts feature quantities from the time-series data acquired by the time-series acquisition unit 110 using a window. In other words, the feature quantity output by the feature extraction unit 120 becomes information expressed in time series by values representing the features of the region in the time series data corresponding to the sliding window. More specifically, for example, the feature extraction unit 120 uses a sliding window, and extracts a value obtained by subtracting the minimum value from the maximum value of the time-series data within the sliding window (hereinafter referred to as “amplitude value”) as a feature value. The feature quantity output by the feature extraction unit 120 becomes information expressed in time series by the amplitude value of the region in the time series data corresponding to the sliding window. Since the method of extracting feature quantities from time-series data using sliding windows is well known, the description is omitted.

第1分割部130基於特徵取出部120取出的特徵量，將時序資料分割為複數個第1部分時序資料。 舉例來說，特徵取出部120使用事先設定的第1時間長度的第1滑動視窗，從時序取得部110取得的時序資料之中，取出第1特徵量。第1分割部130基於特徵取出部120取出的第1特徵量，將時序資料分割為複數個第1部分時序資料。 具體而言，舉例來說，第1分割部130基於第1特徵量，將第1特徵量所示的振幅值與事先設定的臨界值進行比較，特定臨界值以下的振幅值比臨界值還大的第1特徵量中的位置。第1分割部130將該位置對應的滑動視窗的結束期視為時序資料的變化點（以下稱為「第1變化點」），藉由在該第1變化點分割時序資料，將時序資料分割為複數個第1部分時序資料。 舉例來說，第1分割部130產生第1分割資訊，該第1分割資訊表示第1分割部130分割的各第1部分時序資料的時序資料當中的分割位置。 The first division unit 130 divides the time-series data into a plurality of first partial time-series data based on the feature quantities extracted by the feature extraction unit 120 . For example, the feature extraction unit 120 extracts the first feature quantity from the time-series data acquired by the time-series acquisition unit 110 using a first sliding window of a preset first time length. The first division unit 130 divides the time-series data into a plurality of first partial time-series data based on the first feature quantity extracted by the feature extraction unit 120 . Specifically, for example, the first dividing unit 130 compares the amplitude value indicated by the first feature value with a preset threshold value based on the first feature value, and the amplitude value below the specific threshold value is larger than the threshold value. The position in the first feature quantity of . The first division unit 130 regards the end date of the sliding window corresponding to this position as a change point of the time-series data (hereinafter referred to as "the first change point"), and divides the time-series data by dividing the time-series data at the first change point. It is a plurality of Part 1 time series data. For example, the first division unit 130 generates first division information, and the first division information indicates the division position in the time-series data of each first part of the time-series data divided by the first division unit 130 .

另外，第1分割部130特定比臨界值還大的振幅值變得在臨界值以下的第1特徵量中的位置。第1分割部130也可以將該位置對應的滑動視窗的開始期視為時序資料的變化點（以下稱為「第2變化點」）藉由在該第2變化點分割各第1部分時序資料，將第1分割部130分割的複數個第1部分時序資料的每一個，分割為第1振動區間時序資料與第1安定區間時序資料。 若第1分割部130對複數個第1部分時序資料的每一個，分割為第1振動區間時序資料與第1安定區間時序資料時，第1分割部130分割的複數個第1部分時序資料的每一個，都包含第1振動區間時序資料與第1安定區間時序資料。 In addition, the first dividing unit 130 specifies a position in the first feature quantity where the amplitude value larger than the critical value is equal to or smaller than the critical value. The first division unit 130 may also regard the start period of the sliding window corresponding to the position as a change point of the time-series data (hereinafter referred to as "the second change point") by dividing each first part of the time-series data at the second change point Each of the plurality of first partial time-series data divided by the first dividing unit 130 is divided into first vibration interval time-series data and first stable interval time-series data. If the first division unit 130 divides each of the plurality of first partial time-series data into the first vibration interval time-series data and the first stable interval time-series data, the plurality of first partial time-series data divided by the first division unit 130 Each of them includes the timing data of the first vibration interval and the timing data of the first stable interval.

該情況下，舉例來說，第1分割部130針對第1分割部130分割的各第1部分時序資料，產生第1分割資訊，該第1分割資訊表示第1振動區間時序資料以及第1安定區間時序資料的時序資料當中的分割位置。 另外，該情況下，該臨界值如以上所述，是為了將第1部分時序資料分割為第1振動區間時序資料與第1安定區間時序資料，所事先設定的值。因此，舉例來說，若時序取得部110取得的時序資料，為振動感測器等的感測器輸出的訊號轉換為時序資訊的訊號，則可以藉由將該臨界值定為感測器的分解能的1倍至2倍程度的精度，從第1部分時序資料之中取出第1安定區間時序資料。另外，若感測器，或設有感測器的加工裝置等的測量對象的裝置等為相同的規則時，該臨界值可以重複使用相同的值。因此，可以簡單地設定該臨界值。 In this case, for example, the first division unit 130 generates first division information for each first partial time-series data divided by the first division unit 130, and the first division information represents the first vibration interval time-series data and the first stable The division position in the time series data of interval time series data. In this case, the critical value is a value set in advance for dividing the first partial time-series data into the first vibration interval time-series data and the first stable interval time-series data as described above. Therefore, for example, if the timing data obtained by the timing acquisition unit 110 is a signal output from a sensor such as a vibration sensor, which is converted into a signal of timing information, it can be determined by setting the threshold as the sensor's Decompose the accuracy of about 1 to 2 times the energy, and extract the timing data of the first stable interval from the timing data of the first part. In addition, if the sensor, or the device to be measured, such as a processing device provided with the sensor, has the same rules, the same threshold value may be used repeatedly. Therefore, the critical value can be easily set.

第2分割部140基於特徵取出部120取出的特徵量，將第1分割部130分割的複數個第1部分時序資料的每一個分割為複數個第2部分時序資料。 舉例來說，特徵取出部120使用比第1時間長度還短並且事先設定的第2時間長度的第2滑動視窗，從第1分割部130分割的複數個第1部分時序資料的每一個之中，取出第2特徵量。第2分割部140基於取出的第2特徵量，將第1部分時序資料分割為複數個第2部分時序資料。 具體而言，舉例來說，第2分割部140基於第2特徵量，將第2特徵量所示的振幅值與事先設定的臨界值進行比較，特定臨界值以下的振幅值變得比臨界值還大的第2特徵量中的位置。第2分割部140將該位置對應的滑動視窗的結束期視為第1部分時序資料的變化點（以下稱為「第3變化點」），藉由在該第3變化點分割時序資料，將第1部分時序資料分割為複數個第2部分時序資料。 舉例來說，第2分割部140對複數個第1部分時序資料的每一個，產生第2分割資訊，該第2分割資訊表示第2分割部140分割的各第2部分時序資料的時序資料當中的分割位置。 The second dividing unit 140 divides each of the plurality of first partial time-series data divided by the first dividing unit 130 into a plurality of second partial time-series data based on the feature quantities extracted by the feature extracting unit 120 . For example, the feature extracting unit 120 uses a second sliding window of a preset second time length that is shorter than the first time length, and from each of the plurality of first partial time-series data divided by the first dividing unit 130 , take out the second feature quantity. The second division unit 140 divides the first partial time-series data into a plurality of second partial time-series data based on the extracted second feature quantity. Specifically, for example, based on the second feature quantity, the second dividing unit 140 compares the amplitude value indicated by the second feature quantity with a preset threshold value, and the amplitude value below the specific threshold value becomes larger than the threshold value. The position in the second feature quantity that is still larger. The second division unit 140 regards the end date of the sliding window corresponding to this position as a change point of the first part of the time-series data (hereinafter referred to as "the third change point"), and divides the time-series data at the third change point to divide The first part of time series data is divided into a plurality of second part time series data. For example, the second division unit 140 generates second division information for each of the plurality of first partial time-series data, and the second division information indicates that among the time-series data of each second partial time-series data divided by the second division unit 140 split position.

另外，第2分割部140特定比臨界值還大的振幅值變得在臨界值以下的第2特徵量中的位置。第2分割部140也可以將該位置對應的滑動視窗的開始期視為時序資料的變化點（以下稱為「第4變化點」）藉由在該第4變化點分割各第2部分時序資料，將第2分割部140分割的複數個第2部分時序資料的每一個，分割為第2振動區間時序資料與第2安定區間時序資料。 若第2分割部140對複數個第2部分時序資料的每一個，分割為第2振動區間時序資料與第2安定區間時序資料時，第2分割部140分割的複數個第2部分時序資料的每一個，都包含第2振動區間時序資料與第2安定區間時序資料。 In addition, the second dividing unit 140 specifies a position in the second feature quantity at which the amplitude value larger than the critical value is equal to or smaller than the critical value. The second division unit 140 may also regard the start period of the sliding window corresponding to the position as a change point of the time-series data (hereinafter referred to as "the fourth change point") by dividing each second part of the time-series data at the fourth change point Each of the plurality of second partial time-series data divided by the second division unit 140 is divided into the second vibration interval time-series data and the second stable interval time-series data. If the second division unit 140 divides each of the plurality of second partial time-series data into the second vibration interval time-series data and the second stable interval time-series data, the plurality of second partial time-series data divided by the second division unit 140 Each of them includes the time-series data of the second vibration interval and the time-series data of the second stable interval.

該情況下，舉例來說，第2分割部140針對第2分割部140分割的各第2部分時序資料，產生第2分割資訊，該第2分割資訊表示第2振動區間時序資料以及第2安定區間時序資料的時序資料當中的分割位置。 另外，該情況下，該臨界值如以上所述，是為了將第2部分時序資料分割為第2振動區間時序資料與第2安定區間時序資料，所事先設定的值。因此，舉例來說，若時序取得部110取得的時序資料，為振動感測器等的感測器輸出的訊號轉換為時序資訊的訊號，則可以藉由將該臨界值定為感測器的分解能的1倍至2倍程度的精度，從第2部分時序資料之中取出第2安定區間時序資料。另外，若感測器，或設有感測器的加工裝置等的測量對象的裝置等為相同的規則時，該臨界值可以重複使用相同的值。因此，可以簡單地設定該臨界值。 In this case, for example, the second division unit 140 generates second division information for each of the second partial time-series data divided by the second division unit 140, and the second division information represents the second vibration interval time-series data and the second stable The division position in the time series data of interval time series data. In addition, in this case, the critical value is a value set in advance for dividing the second partial time-series data into the second vibration interval time-series data and the second stable interval time-series data as described above. Therefore, for example, if the timing data obtained by the timing acquisition unit 110 is a signal output from a sensor such as a vibration sensor, which is converted into a signal of timing information, it can be determined by setting the threshold as the sensor's The precision of 1 to 2 times of the decomposition energy is taken out from the second part of the time series data in the second stable interval. In addition, if the sensor, or the device to be measured, such as a processing device provided with the sensor, has the same rules, the same threshold value may be used repeatedly. Therefore, the critical value can be easily set.

參照第2圖，針對關於實施形態1的資料處理裝置100包含的時序取得部110取得的時序資料、以及第1分割部130分割時序資料得到的第1部分時序資料進行說明。 第2圖為一說明圖，示意關於實施形態1的資料處理裝置100包含的時序取得部110取得的時序資料、以及第1分割部130分割時序資料得到的第1部分時序資料的一例。 Referring to FIG. 2, the time-series data obtained by the time-series acquisition unit 110 included in the data processing apparatus 100 of the first embodiment and the first partial time-series data obtained by dividing the time-series data by the first dividing unit 130 will be described. FIG. 2 is an explanatory diagram showing an example of the time-series data obtained by the time-series acquisition unit 110 included in the data processing device 100 of the first embodiment and the first partial time-series data obtained by dividing the time-series data by the first dividing unit 130.

若時序取得部110取得的時序資料，為反覆製造相同加工製品的加工裝置當中設置的振動感測器輸出的訊號轉換為時序資訊而得時，時序資料當中，同樣的時序值的推移會在每次製造加工時反覆出現。 第1分割部130可以藉由將時序資料分割為複數個第1部分時序資料，將時序資料分割為製造複數個加工製品的每一個的期間(以下稱為「製造期間」)對應的第1部分時序資料。 製造期間中，存在有將身為加工對象的零件之加工零件加工為加工製品的加工期間、以及某個加工製品的加工結束的時間點，到下一個加工製品的加工開始的時間點之間的閒置期間。 第1分割部130可以藉由將第1部分時序資料分割為第1振動區間時序資料以及第1安定區間時序資料，將第1部分時序資料分割為加工期間對應的第1振動區間時序資料、以及閒置期間對應的第1安定區間時序資料。 If the time series data obtained by the time series acquisition unit 110 is obtained by converting the signal output from the vibration sensor installed in the processing device repeatedly manufacturing the same processed product into time series information, in the time series data, the transition of the same time series value will occur every Repeatedly appear during each manufacturing process. The first division unit 130 may divide the time-series data into a plurality of first parts of the time-series data, and divide the time-series data into first parts corresponding to the period of manufacturing each of the plurality of processed products (hereinafter referred to as "manufacturing period") time series data. In the manufacturing period, there is a processing period in which a processed part of a part to be processed is processed into a processed product, and the period between the time when the processing of a certain processed product ends and the time when the processing of the next processed product starts idle period. The first division unit 130 may divide the first part of the time-series data into the first vibration interval time-series data corresponding to the processing period, and The timing data of the first stable interval corresponding to the idle period.

參照第3圖，針對關於實施形態1的資料處理裝置100包含的第1分割部130分割得到的第1部分時序資料、以及第2分割部140分割第1部分時序資料得到的第2部分時序資料進行說明。 第3圖為一說明圖，示意關於實施形態1的資料處理裝置100包含的第1分割部130分割得到的第1部分時序資料、以及第2分割部140分割第1部分時序資料得到的第2部分時序資料的一例。 Referring to FIG. 3, regarding the first part of time-series data obtained by dividing the first part of time-series data obtained by the first division unit 130 included in the data processing device 100 of Embodiment 1, and the second part of time-series data obtained by dividing the first part of time-series data by the second division unit 140 Be explained. Fig. 3 is an explanatory diagram showing the first partial time-series data obtained by dividing the first part 130 included in the data processing device 100 of the first embodiment, and the second partial time-series data obtained by dividing the first part time-series data by the second dividing part 140. An example of partial time series data.

若時序取得部110取得的時序資料，為反覆製造相同加工製品的加工裝置當中設置的振動感測器輸出的訊號轉換為時序資訊而得時，第1部分時序資料中，複數個加工程序的每一個對應的時序值的推移會在1個加工期間當中出現。 此處，所謂的「加工程序」，是將加工零件設置於台座的程序，偵測設置於台座的加工零件的形狀的程序，切削設置於台座的加工零件的程序，檢查切削後的加工零件的形狀的程序，或是將切削後的加工零件從台座取下的程序。 If the time-series data obtained by the time-series acquisition unit 110 is obtained by converting the signal output from a vibration sensor installed in a processing device repeatedly manufacturing the same processed product into time-series information, in the first part of the time-series data, each of the plurality of processing procedures A corresponding timing value transition occurs during 1 processing period. Here, the so-called "processing program" is a program for setting the machined part on the pedestal, a program for detecting the shape of the machined part set on the pedestal, a program for cutting the machined part set on the pedestal, and a program for inspecting the machined part after cutting. The program of the shape, or the program of removing the machined part from the base after cutting.

第2分割部140可以藉由將複數個第1部分時序資料當中的第1振動區間時序資料的每一個分割為第2部分時序資料，將第1振動區間時序資料分割為複數個加工程序的每一個對應的期間(以下稱為「程序期間」)對應的第2部分時序資料。 複數個程序期間的每一個當中，存在有加工裝置實際運作的程序運作期間、以及在某個程序中加工裝置的運作結束的時間點，到下一個加工程序中加工裝置的運作開始的時間點之間的程序閒置期間。 第2分割部140可以藉由將第2部分時序資料分割為第2振動區間時序資料以及第2安定區間時序資料，將第2部分時序資料分割為程序運作期間對應的第2振動區間時序資料、以及程序閒置期間對應的第2安定區間時序資料。 The second division unit 140 may divide the first vibration interval time-series data into each of the plurality of processing programs by dividing each of the first vibration interval time-series data among the plurality of first partial time-series data into the second partial time-series data. Part 2 time-series data corresponding to a corresponding period (hereinafter referred to as "program period"). For each of the plurality of program periods, there is a program operation period in which the processing device is actually operated, a time point when the operation of the processing device in a certain program ends, and a time point between the time point when the operation of the processing device starts in the next processing program. The program idle period in between. The second division unit 140 can divide the second part of the time-series data into the second vibration interval time-series data and the second stable interval time-series data, and divide the second part of the time-series data into the second vibration interval time-series data corresponding to the program operation period, And the timing data of the second stable interval corresponding to the idle period of the program.

根據以上那樣的構成，若事先設定第1時間長度以及第2時間長度，則資料處理裝置100就可以針對每一個在時序資料中反覆出現的波型，將時序資料分割為部分時序資料。 既有的資料處理裝置中，使用者必須要事先準備好波形資料、以及波形資料的參數資訊（以下稱為「特別資訊」），作為用來分割過去取得的時序資料並分割時序資料的指標。時序資料的分析需要一定程度的知識，對使用者而言，事先準備好特別資訊並不容易。 According to the above configuration, if the first time length and the second time length are set in advance, the data processing device 100 can divide the time-series data into partial time-series data for each waveform that appears repeatedly in the time-series data. In the existing data processing device, the user must prepare the waveform data and the parameter information of the waveform data (hereinafter referred to as "special information") in advance as indicators for dividing the time-series data obtained in the past and dividing the time-series data. The analysis of time series data requires a certain degree of knowledge, and it is not easy for users to prepare special information in advance.

對照之下，舉例來說，若時序取得部110取得的時序資料，為反覆製造相同加工製品的加工裝置當中設置的振動感測器輸出的訊號轉換為時序資訊而得時，由於閒置期間與程序閒置期間任何一者皆為已知，因此使用者可以基於閒置期間與程序閒置期間，輕易地決定第1時間長度與第2時間長度。 結果，在沒有事先準備好特別資訊的情況下，資料處理裝置100也可以針對每一個在時序資料中反覆出現的波型，將時序資料分割為部分時序資料。 In contrast, for example, if the timing data acquired by the timing acquisition unit 110 is obtained by converting the signal output from the vibration sensor installed in the processing device repeatedly manufacturing the same processed product into timing information, due to the idle period and the program Any of the idle periods is known, so the user can easily determine the first time length and the second time length based on the idle period and the program idle period. As a result, without preparing special information in advance, the data processing device 100 can also divide the time-series data into partial time-series data for each waveform that appears repeatedly in the time-series data.

輸出部190舉例來說，基於第1分割部130產生的第1分割資訊、以及第2分割部140產生的第2分割資訊，產生將第1分割資訊與第2分割資訊視覺化的顯示影像，並輸出示意顯示影像的顯示資訊。 具體而言，舉例來說，輸出部190將顯示資訊輸出至顯示裝置20，讓顯示資訊所示的顯示影像在顯示裝置20顯示。 輸出部190並不限於輸出顯示資訊的元件。舉例來說，輸出部190也可以將第1分割資訊與第2分割資訊輸出至記憶裝置10，讓記憶裝置10記憶第1分割資訊與第2分割資訊。 The output unit 190, for example, generates a display image that visualizes the first division information and the second division information based on the first division information generated by the first division unit 130 and the second division information generated by the second division unit 140, And output the display information indicating the display image. Specifically, for example, the output unit 190 outputs the display information to the display device 20 , so that the display image indicated by the display information is displayed on the display device 20 . The output unit 190 is not limited to an element for outputting and displaying information. For example, the output unit 190 may also output the first division information and the second division information to the memory device 10, so that the memory device 10 stores the first division information and the second division information.

上述那樣的資料處理裝置100也可以包含異常偵測部150。 異常偵測部150基於第2分割部140分割的複數個第2部分時序資料之中的至少1個第2部分時序資料，偵測時序取得部110取得的時序資料的異常。舉例來說，異常偵測部150產生示意偵測結果的偵測資訊。 具體而言，舉例來說，基於第2分割部140分割的複數個第2部分時序資料之中的第2安定區間時序資料，偵測時序取得部110取得的時序資料的異常。 The above-mentioned data processing device 100 may also include an abnormality detection unit 150 . The abnormality detecting unit 150 detects abnormality of the time-series data obtained by the time-series obtaining unit 110 based on at least one second partial time-series data among the plurality of second partial time-series data divided by the second dividing unit 140 . For example, the anomaly detection unit 150 generates detection information indicating detection results. Specifically, for example, based on the second stable interval time-series data among the plurality of second partial time-series data divided by the second dividing unit 140 , the abnormality of the time-series data obtained by the time-series obtaining unit 110 is detected.

更具體而言，舉例來說，異常偵測部150藉由判定第1分割部130分割的複數個第1部分時序資料的每一個對應的第2安定區間時序資料的個數，在所有的第1部分時序資料中是否相同，偵測時序資料的異常。 假設，若複數個第1部分時序資料的每一個對應的第2安定區間時序資料的個數，在所有的第1部分時序資料中並不相同，舉例來說，有可能是2個加工程序彼此之間的期間，比事先設定的期間還要短。另外，有可能是2個加工程序彼此之間的期間因為某種原因，而在加工裝置中產生了沒有預期的振動。 因此，藉由以上那樣的構成，異常偵測部150可以用高精度偵測時序資料的異常。 More specifically, for example, by determining the number of second stable interval time-series data corresponding to each of the plurality of first partial time-series data divided by the first division unit 130, the anomaly detection unit 150, among all the first partial time-series data, 1 Whether part of the time-series data is the same, detect the abnormality of the time-series data. Assume that if the number of the second stable interval time-series data corresponding to each of the plurality of first-part time-series data is not the same among all the first-part time-series data, for example, it is possible that two processing procedures are mutually The period in between is shorter than the pre-set period. In addition, there is a possibility that unexpected vibration occurs in the processing device for some reason between the two processing programs. Therefore, with the above configuration, the anomaly detection unit 150 can detect an anomaly of the time-series data with high precision.

另外，舉例來說，異常偵測部150也可以基於第2分割部140分割的複數個第2部分時序資料之中的第2安定區間時序資料、以及示意該第2安定區間時序資料是否為第1振動區間時序資料當中的第幾號的第2安定區間時序資料之第2安定區間位置資訊，偵測時序取得部110取得的時序資料的異常。 舉例來說，異常偵測部150將第2安定區間時序資料與第2安定區間位置資訊，輸入至事先準備好的已學習模型。異常偵測部150基於已學習模型輸出的推論結果，偵測時序資料的異常。舉例來說，已學習模型事先記憶於記憶裝置10，異常偵測部150藉由從記憶裝置10讀取已學習模型，來取得已學習模型。 藉由以上那樣的構成，異常偵測部150可以用高精度偵測時序資料的異常。 In addition, for example, the abnormality detection unit 150 may also indicate whether the second stable interval time series data is the second stable interval time series data based on the second stable interval time series data among the plurality of second partial time series data divided by the second dividing unit 140. The position information of the second stable interval of the second stable interval time-series data, which is the number in the first vibration interval time-series data, detects the abnormality of the time-series data acquired by the timing acquisition unit 110 . For example, the anomaly detection unit 150 inputs the time-series data of the second stable interval and the position information of the second stable interval into the pre-prepared learned model. The anomaly detection unit 150 detects anomalies in time series data based on inference results output by the learned model. For example, the learned model is stored in the memory device 10 in advance, and the anomaly detection unit 150 obtains the learned model by reading the learned model from the memory device 10 . With the above configuration, the anomaly detection unit 150 can detect an anomaly of the time-series data with high precision.

另外，已學習模型是不圖示的學習裝置基於第2安定區間時序資料與第2安定區間位置資訊，讓事先準備好的學習模型進行監督式學習或無監督學習等的學習所產生。針對已學習模型的產生方法，將省略說明。In addition, the learned model is generated by a learning device not shown in the figure, based on the time-series data of the second stable interval and the position information of the second stable interval, and the pre-prepared learning model is subjected to learning such as supervised learning or unsupervised learning. Regarding the generation method of the learned model, description will be omitted.

另外，舉例來說，異常偵測部150也可以基於第2安定區間時序資料以及第2安定區間位置資訊，將該第2安定區間時序資料、以及第2分割部140對其他第1部分時序資料的第1振動區間時序資料進行分割後的複數個第2部分時序資料之中的第2安定區間時序資料，也就是與第2安定區間位置資訊相同的第2安定區間時序資料進行比較，偵測時序取得部110取得的時序資料的異常。 具體而言，舉例來說，異常偵測部150將第2安定區間時序資料、以及與第2安定區間位置資訊相同的第2安定區間時序資料進行比較，藉由判定相似度，偵測時序取得部110取得的時序資料的異常。由於判定時序資料彼此之間的相似度的方法已為眾所皆知，故省略說明。 藉由以上那樣的構成，異常偵測部150可以用高精度偵測時序資料的異常。 In addition, for example, the anomaly detection unit 150 may also, based on the second stable interval time series data and the second stable interval position information, the second stable interval time series data and the second division unit 140 compare other first partial time series data The second stable interval timing data among the plurality of second part timing data after the first vibration interval timing data is divided, that is, the second stable interval timing data with the same position information as the second stable interval is compared and detected The time-series data acquired by the time-series acquisition unit 110 is abnormal. Specifically, for example, the anomaly detection unit 150 compares the second stable interval timing data with the second stable interval timing data that is the same as the second stable interval location information, and by determining the similarity, the detection timing is obtained. The time series data acquired by the unit 110 is abnormal. Since the method of determining the similarity between time-series data is well known, the description is omitted. With the above configuration, the anomaly detection unit 150 can detect an anomaly of the time-series data with high precision.

異常偵測部150偵測時序資料的異常之際，也可以算出時序資料的異常度，並產生示意算出的異常度的偵測資訊。 藉由異常偵測部150算出時序資料的異常度，資料處理裝置100舉例來說，可以在加工裝置故障前等的異常發生之前，確認加工裝置的劣化等的異常發生的徵兆。 When the anomaly detection unit 150 detects an anomaly in the time-series data, it may also calculate the anomaly degree of the time-series data, and generate detection information indicating the calculated anomaly degree. By calculating the abnormality degree of the time-series data by the abnormality detection unit 150, the data processing device 100, for example, can confirm the signs of abnormalities such as deterioration of the processing device before the abnormality occurs before the processing device fails.

資料處理裝置100包含異常偵測部150時，輸出部190舉例來說，連同第1分割資訊以及第2分割資訊，產生將異常偵測部150產生的偵測資訊視覺化的顯示影像，並輸出示意顯示影像的顯示資訊。 輸出部190可以連同第1分割資訊與第2分割資訊，將偵測資訊輸出至記憶裝置10，讓記憶裝置10記憶第1分割資訊、第2分割資訊、以及偵測資訊。 When the data processing device 100 includes the abnormality detection unit 150, the output unit 190, for example, together with the first division information and the second division information, generates a display image that visualizes the detection information generated by the abnormality detection unit 150, and outputs Indicates the display information of the displayed image. The output unit 190 can output the detection information to the memory device 10 together with the first division information and the second division information, so that the memory device 10 can store the first division information, the second division information, and the detection information.

參照第4A圖以及第4B圖，針對關於實施形態1的資料處理裝置100的重點的硬體構成進行說明。 第4A圖以及第4B圖為一示意圖，示意關於實施形態1的資料處理裝置100的重點的硬體構成的一例。 The essential hardware configuration of the data processing device 100 according to Embodiment 1 will be described with reference to FIGS. 4A and 4B. FIG. 4A and FIG. 4B are schematic diagrams showing an example of the essential hardware configuration of the data processing device 100 according to the first embodiment.

如第4A圖所示，資料處理裝置100由電腦所構成，該電腦包含處理器401以及記憶體402。 記憶體402當中記憶程式，該程式用以使該電腦發揮時序取得部110、特徵取出部120、第1分割部130、第2分割部140、異常偵測部150、以及輸出部190的功能。藉由處理器401讀取並執行記憶於記憶體402當中的程式，以實現時序取得部110、特徵取出部120、第1分割部130、第2分割部140、異常偵測部150、以及輸出部190。 As shown in FIG. 4A , the data processing device 100 is composed of a computer, and the computer includes a processor 401 and a memory 402 . The memory 402 stores a program for making the computer perform the functions of the timing acquisition unit 110 , the feature extraction unit 120 , the first division unit 130 , the second division unit 140 , the abnormality detection unit 150 , and the output unit 190 . The program stored in the memory 402 is read and executed by the processor 401 to realize the timing acquisition unit 110, the feature extraction unit 120, the first segmentation unit 130, the second segmentation unit 140, the abnormality detection unit 150, and the output Section 190.

另外，如第4B圖所示，資料處理裝置100也可以由處理電路403所構成。該情況下，時序取得部110、特徵取出部120、第1分割部130、第2分割部140、異常偵測部150、以及輸出部190的功能可以由處理電路403來實現。In addition, as shown in FIG. 4B , the data processing device 100 may also be composed of a processing circuit 403 . In this case, the functions of the sequence acquisition unit 110 , the feature extraction unit 120 , the first division unit 130 , the second division unit 140 , the abnormality detection unit 150 , and the output unit 190 can be realized by the processing circuit 403 .

另外，資料處理裝置100也可以由處理器401、記憶體402、以及處理電路403所構成。該情況下，時序取得部110、特徵取出部120、第1分割部130、第2分割部140、異常偵測部150、以及輸出部190的功能當中的一部分的功能，可以由處理器401以及記憶體402來實現；而剩餘的功能可以由處理電路403來實現。In addition, the data processing device 100 may also be composed of a processor 401 , a memory 402 , and a processing circuit 403 . In this case, part of the functions of the sequence acquisition unit 110, the feature extraction unit 120, the first division unit 130, the second division unit 140, the abnormality detection unit 150, and the output unit 190 may be performed by the processor 401 and The memory 402 is implemented; and the remaining functions can be implemented by the processing circuit 403 .

處理器401舉例來說，使用中央處理器(Central Processing Unit，CPU)、圖形處理器(Graphics Processing Unit，GPU)、微處理器、微控制器、或數位訊號處理器(Digital Signal Processor，DSP)。The processor 401 is, for example, a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU), a microprocessor, a microcontroller, or a digital signal processor (Digital Signal Processor, DSP). .

記憶體402舉例來說，使用半導體記憶體或磁碟。更具體來說，記憶體402使用隨機存取記憶體(Random Access Memory，RAM)、唯讀記憶體(Read Only Memory，ROM)、快閃記憶體、可抹除可程式唯讀記憶體(Erasable Programming Read Only Memory，EPROM)、電子式可抹除可程式唯讀記憶體(Electrically Erasable Programming Read Only Memory，EEPROM)、固態硬碟(Solid State Drive，SSD)、或硬式磁碟機(Hard Disk Drive，HDD)。The memory 402 is, for example, a semiconductor memory or a magnetic disk. More specifically, the memory 402 uses random access memory (Random Access Memory, RAM), read only memory (Read Only Memory, ROM), flash memory, erasable programmable read only memory (Erasable Programming Read Only Memory, EPROM), Electronically Erasable Programmable Read Only Memory (EEPROM), Solid State Drive (SSD), or Hard Disk Drive (Hard Disk Drive , HDD).

處理電路403舉例來說，使用應用特定積體電路(Application Specific Integrated Circuit，ASIC)、可程式邏輯裝置(Programmable Logic Device，PLD)、場效可程式閘陣列(Field-Programmable Gate Array，FPGA)、系統單晶片(System-on-a-Chip，SoC)、或系統大型積體電路(Large-Scale Integration，LSI)。The processing circuit 403, for example, uses an Application Specific Integrated Circuit (ASIC), a Programmable Logic Device (PLD), a Field-Programmable Gate Array (FPGA), System-on-a-Chip (SoC), or system large-scale integrated circuits (Large-Scale Integration, LSI).

參照第5圖，針對關於實施形態1的資料處理裝置100的運作進行說明。 第5圖為一流程圖，說明關於實施形態1的資料處理裝置100的處理的一例。 Referring to Fig. 5, the operation of the data processing device 100 according to the first embodiment will be described. Fig. 5 is a flowchart illustrating an example of the processing of the data processing device 100 according to the first embodiment.

首先，在步驟ST501，時序取得部110取得時序資料。 接著，在步驟ST502，特徵取出部120取出第1特徵量。 接著，在步驟ST503，第1分割部130將時序資料分割為複數個第1部分時序資料。 接著，在步驟ST504，特徵取出部120取出第2特徵量。 接著，在步驟ST505，第2分割部140將第1分割部130分割的複數個第1部分時序資料的每一個，分割為複數個第2部分時序資料。 接著，在步驟ST506，異常偵測部150偵測時序資料的異常。 接著，在步驟ST507，輸出部190輸出顯示資訊。 步驟ST507之後，資料處理裝置100結束該流程圖的處理。 First, in step ST501, the time-series acquisition unit 110 acquires time-series data. Next, in step ST502, the feature extraction unit 120 extracts the first feature amount. Next, in step ST503, the first dividing unit 130 divides the time-series data into a plurality of first partial time-series data. Next, in step ST504, the feature extraction unit 120 extracts the second feature amount. Next, in step ST505 , the second dividing unit 140 divides each of the plurality of first partial time-series data divided by the first dividing unit 130 into a plurality of second partial time-series data. Next, in step ST506, the anomaly detection unit 150 detects an anomaly of the time series data. Next, in step ST507, the output unit 190 outputs display information. After step ST507, the data processing device 100 ends the processing of this flowchart.

如以上所述，資料處理裝置100包含：時序取得部110，取得時序資料；特徵取出部120，基於時序取得部110取得的時序資料，取出時序資料的特徵量；第1分割部130，基於特徵取出部120取出的特徵量，將時序資料分割為複數個第1部分時序資料；以及第2分割部140，基於特徵取出部120取出的特徵量，將第1分割部130分割的複數個第1部分時序資料的每一個分割為複數個第2部分時序資料。 藉由這樣的構成，在沒有事先準備好特別資訊的情況下，資料處理裝置100也可以針對每一個在時序資料中反覆出現的波型，將時序資料分割為部分時序資料。 As described above, the data processing device 100 includes: a time-series acquisition unit 110, which acquires time-series data; a feature extraction unit 120, which extracts the feature quantity of the time-series data based on the time-series data acquired by the time-series acquisition unit 110; the first segmentation unit 130, based on the feature The feature quantity extracted by the extraction unit 120 divides the time-series data into a plurality of first partial time-series data; Each of the partial time-series data is divided into a plurality of second partial time-series data. With such a configuration, without preparing special information in advance, the data processing device 100 can also divide the time-series data into partial time-series data for each waveform that appears repeatedly in the time-series data.

另外，資料處理裝置100在上述構成中，特徵取出部120使用事先設定的第1時間長度的第1滑動視窗，從時序取得部110取得的時序資料之中，取出第1特徵量；第1分割部130基於特徵取出部120取出的第1特徵量，將時序資料分割為第1部分時序資料；特徵取出部120使用比第1時間長度還短並且事先設定的第2時間長度的第2滑動視窗，從第1分割部130分割的複數個第1部分時序資料的每一個之中，取出第2特徵量；第2分割部140基於取出的第2特徵量，將第1部分時序資料分割為複數個第2部分時序資料。 藉由這樣的構成，在沒有事先準備好特別資訊的情況下，資料處理裝置100也可以針對每一個在時序資料中反覆出現的波型，將時序資料分割為部分時序資料。 In addition, in the data processing device 100, in the above configuration, the feature extraction unit 120 extracts the first feature quantity from the time series data acquired by the time series acquisition unit 110 using the first sliding window of the first time length set in advance; the first division The part 130 divides the time-series data into the first part of time-series data based on the first feature quantity extracted by the feature extraction part 120; the feature extraction part 120 uses the second sliding window of the second time length which is shorter than the first time length and preset , from each of the plurality of first partial time-series data divided by the first division unit 130, the second feature quantity is extracted; the second division unit 140 divides the first part of the time-series data into plural parts based on the extracted second feature quantity Part 2 timing data. With such a configuration, without preparing special information in advance, the data processing device 100 can also divide the time-series data into partial time-series data for each waveform that appears repeatedly in the time-series data.

另外，資料處理裝置100除了上述構成之外，包含：異常偵測部150，基於第2分割部140分割的複數個第2部分時序資料之中，至少1個第2部分時序資料，偵測時序取得部110取得的時序資料的異常。 藉由這樣的構成，資料處理裝置100可以用高精度偵測時序資料的異常。 In addition, the data processing device 100 includes, in addition to the above-mentioned configuration, an abnormality detection unit 150, which detects the timing based on at least one second partial time-series data among the plurality of second partial time-series data divided by the second dividing unit 140. The time-series data acquired by the acquisition unit 110 is abnormal. With such a configuration, the data processing device 100 can detect anomalies in time-series data with high precision.

另外，資料處理裝置100在上述構成中，第1分割部130分割的複數個第1部分時序資料的每一個，都包含第1振動區間時序資料以及第1安定區間時序資料；第2分割部140針對第1分割部130分割的複數個第1部分時序資料的每一個，將第1部分時序資料之中的第1振動區間時序資料，分割為複數個第2部分時序資料；第2分割部140分割的複數個第2部分時序資料的每一個，都包含第2振動區間時序資料以及第2安定區間時序資料；異常偵測部150基於第2分割部140分割的複數個第2部分時序資料之中的第2安定區間時序資料，偵測時序取得部110取得的時序資料的異常。 藉由這樣的構成，資料處理裝置100可以用高精度偵測時序資料的異常。 In addition, in the data processing device 100 in the above configuration, each of the plurality of first partial time-series data divided by the first division unit 130 includes the first vibration interval time-series data and the first stable interval time-series data; the second division unit 140 For each of the plurality of first partial time-series data divided by the first division unit 130, the first vibration interval time-series data among the first partial time-series data is divided into a plurality of second partial time-series data; the second division unit 140 Each of the divided plurality of second partial time-series data includes the second vibration interval time-series data and the second stable interval time-series data; In the second stable interval time series data, abnormality of the time series data acquired by the time series acquisition unit 110 is detected. With such a configuration, the data processing device 100 can detect anomalies in time-series data with high precision.

另外，資料處理裝置100在上述構成中，異常偵測部150基於第2分割部140分割的複數個第2部分時序資料之中的第2安定區間時序資料、以及示意該第2安定區間時序資料為第1振動區間時序資料當中的第幾號第2安定區間時序資料之第2安定區間位置資訊，偵測時序取得部110取得的時序資料的異常。 藉由這樣的構成，資料處理裝置100可以用高精度偵測時序資料的異常。 In addition, in the above configuration of the data processing device 100, the abnormality detection unit 150 is based on the second stable interval time-series data among the plurality of second partial time-series data divided by the second division unit 140, and the time-series data indicating the second stable interval. It is the position information of the second stable interval of the second stable interval time-series data of the second stable interval among the first vibration interval time-series data, and detects the abnormality of the time-series data obtained by the time-series acquisition unit 110 . With such a configuration, the data processing device 100 can detect anomalies in time-series data with high precision.

另外，資料處理裝置100在上述構成中，異常偵測部150基於第2安定區間時序資料以及第2安定區間位置資訊，將該第2安定區間時序資料、以及第2分割部140對其他第1部分時序資料的第1振動區間時序資料進行分割後的複數個第2部分時序資料之中的第2安定區間時序資料，也就是與第2安定區間位置資訊相同的第2安定區間時序資料進行比較，偵測時序取得部110取得的時序資料的異常。 藉由這樣的構成，資料處理裝置100可以用高精度偵測時序資料的異常。 In addition, in the above configuration of the data processing device 100, the abnormality detection unit 150 compares the second stable interval time-series data and the second dividing unit 140 to other first Partial time-series data of the first vibration interval time-series data is divided into the second stable interval time-series data among the plurality of second partial time-series data, that is, the second stable interval time-series data with the same location information as the second stable interval , to detect anomalies in the time-series data acquired by the time-series acquisition unit 110 . With such a configuration, the data processing device 100 can detect anomalies in time-series data with high precision.

另外，本揭露在該揭露的範圍內，可以進行實施形態的任意構成元件的變形，或是實施形態中任意構成元件的省略。 [產業可利用性] In addition, within the scope of the disclosure, the present disclosure can modify arbitrary constituent elements of the embodiments, or omit arbitrary constituent elements in the embodiments. [Industrial availability]

關於本揭露的資料處理裝置，可以應用在偵測時序資料異常的資料處理系統。The data processing device disclosed in this disclosure can be applied to a data processing system for detecting anomalies in time series data.

1:資料處理系統 10:記憶裝置 20:顯示裝置 100:資料處理裝置 110:時序取得部 120:特徵取出部 130:第1分割部 140:第2分割部 150:異常偵測部 190:輸出部 401:處理器 402:記憶體 403:處理電路 ST501~ST507:步驟 1: Data processing system 10: memory device 20: Display device 100: data processing device 110: Timing Acquisition Department 120: Feature extraction part 130: The first division 140: The second division 150: Anomaly Detection Department 190: output part 401: Processor 402: Memory 403: processing circuit ST501~ST507: Steps

第1圖為一方塊圖，示意關於實施形態1的資料處理裝置、以及應用資料處理裝置的資料處理系統的重點的構成的一例。 第2圖為一說明圖，示意關於實施形態1的資料處理裝置包含的時序取得部取得的時序資料、以及第1分割部分割時序資料得到的第1部分時序資料的一例。 第3圖為一說明圖，示意關於實施形態1的資料處理裝置包含的第1分割部分割得到的第1部分時序資料、以及第2分割部分割第1部分時序資料得到的第2部分時序資料的一例。 第4A圖以及第4B圖為一示意圖，示意關於實施形態1的資料處理裝置的重點的硬體構成的一例。 第5圖為一流程圖，說明關於實施形態1的資料處理裝置的處理的一例。 Fig. 1 is a block diagram showing an example of the key configuration of the data processing device and the data processing system using the data processing device according to the first embodiment. Fig. 2 is an explanatory diagram showing an example of the time-series data obtained by the time-series acquisition unit included in the data processing device of the first embodiment and the first partial time-series data obtained by dividing the time-series data by the first division unit. Fig. 3 is an explanatory diagram showing the first part of time-series data obtained by dividing the first part and the second part of time-series data obtained by dividing the first part of time-series data by the second division part included in the data processing device of Embodiment 1. An example of FIG. 4A and FIG. 4B are diagrams showing an example of the essential hardware configuration of the data processing device of the first embodiment. Fig. 5 is a flowchart illustrating an example of the processing of the data processing device according to the first embodiment.

1:資料處理系統 1: Data processing system

10:記憶裝置 10: memory device

20:顯示裝置 20: Display device

100:資料處理裝置 100: data processing device

110:時序取得部 110: Timing Acquisition Department

120:特徵取出部 120: Feature extraction part

130:第1分割部 130: The first division

140:第2分割部 140: The second division

150:異常偵測部 150: Anomaly Detection Department

190:輸出部 190: output part

Claims (4)

一種資料處理裝置，包含：時序取得部，取得時序資料；特徵取出部，基於前述時序取得部取得的前述時序資料，取出前述時序資料的特徵量；第1分割部，基於前述特徵取出部取出的前述特徵量，將前述時序資料分割為複數個第1部分時序資料；第2分割部，基於前述特徵取出部取出的前述特徵量，將前述第1分割部分割的複數個前述第1部分時序資料的每一個分割為複數個第2部分時序資料；以及異常偵測部，基於前述第2分割部分割的複數個前述第2部分時序資料之中，至少1個前述第2部分時序資料，偵測前述時序取得部取得的前述時序資料的異常；其中，前述第1分割部分割的複數個前述第1部分時序資料的每一個，都包含第1振動區間時序資料以及第1安定區間時序資料；其中，前述第2分割部針對前述第1分割部分割的複數個前述第1部分時序資料的每一個，將前述第1部分時序資料之中的前述第1振動區間時序資料，分割為複數個前述第2部分時序資料；其中，前述第2分割部分割的複數個前述第2部分時序資料的每一個，都包含第2振動區間時序資料以及第2安定區間時序資料；其中，前述異常偵測部基於前述第2分割部分割的複數個前述第2部分時序資料之中的前述第2安定區間時序資料，偵測前述時序取得部取得的前述時序資料的異常。 A data processing device, comprising: a time-series acquisition unit for acquiring time-series data; a feature extraction unit for extracting feature quantities of the time-series data based on the time-series data acquired by the time-series acquisition unit; The aforementioned feature quantity divides the aforementioned time-series data into a plurality of first part time-series data; the second division part divides the plurality of aforementioned first part time-series data divided by the aforementioned first division part based on the aforementioned feature quantity extracted by the aforementioned feature extraction part Each of the divisions is divided into a plurality of second part time series data; and the anomaly detection part detects at least one of the aforementioned second part time series data based on the plurality of the aforementioned second part time series data divided by the aforementioned second division part. The abnormality of the aforementioned timing data obtained by the aforementioned timing acquisition unit; wherein, each of the plurality of aforementioned first part timing data divided by the aforementioned first dividing unit includes the first vibration interval timing data and the first stable interval timing data; wherein For each of the plurality of first partial time-series data divided by the first division unit, the second division unit divides the first vibration interval time-series data among the first partial time-series data into a plurality of the first partial time-series data. 2 parts of time-series data; wherein, each of the plurality of aforementioned second-part time-series data divided by the aforementioned second division part includes the second vibration interval time-series data and the second stable interval time-series data; wherein, the aforementioned anomaly detection part is based on The second stable interval time-series data among the plurality of second part time-series data divided by the second division unit detects the abnormality of the time-series data obtained by the time-series acquisition unit. 如請求項1之資料處理裝置， 其中，前述特徵取出部使用事先設定的第1時間長度的第1滑動視窗，從前述時序取得部取得的前述時序資料之中，取出第1特徵量；其中，前述第1分割部基於前述特徵取出部取出的前述第1特徵量，將前述時序資料分割為複數個前述第1部分時序資料；其中，前述特徵取出部使用比前述第1時間長度還短並且事先設定的第2時間長度的第2滑動視窗，從前述第1分割部分割的複數個前述第1部分時序資料的每一個之中，取出第2特徵量；其中，前述第2分割部基於取出的前述第2特徵量，將前述第1部分時序資料分割為複數個前述第2部分時序資料。 For the data processing device of claim 1, Wherein, the above-mentioned feature extraction part uses the first sliding window of the first time length set in advance to extract the first feature value from the above-mentioned time-series data obtained by the above-mentioned time-series acquisition part; Partially extracted first feature quantity, divides the aforementioned time-series data into a plurality of first partial time-series data; wherein, the aforementioned feature extraction unit uses a second time-sequence that is shorter than the first time length and is set in advance. The sliding window extracts the second feature value from each of the plurality of first partial time-series data divided by the first segment; wherein, the second segment extracts the aforementioned second feature based on the extracted second feature Part 1 time-series data is divided into a plurality of the aforementioned second part time-series data. 如請求項1之資料處理裝置，其中，前述異常偵測部基於前述第2分割部分割的複數個前述第2部分時序資料之中的前述第2安定區間時序資料、以及示意該第2安定區間時序資料為前述第1振動區間時序資料當中的第幾號前述第2安定區間時序資料之第2安定區間位置資訊，偵測前述時序取得部取得的前述時序資料的異常。 The data processing device according to claim 1, wherein the abnormality detection unit is based on the second stable interval time-series data among the plurality of second part time-series data divided by the second division unit, and indicates the second stable interval The time-series data is the position information of the second stable interval of the second stable interval data of the number in the first vibration interval time-series data, and detects the abnormality of the aforementioned timing data acquired by the aforementioned timing acquisition unit. 如請求項3之資料處理裝置，其中，前述異常偵測部基於前述第2安定區間時序資料以及前述第2安定區間位置資訊，將該第2安定區間時序資料、以及前述第2分割部對其他前述第1部分時序資料的前述第1振動區間時序資料進行分割後的複數個前述第2部分時序資料之中的前述第2安定區間時序資料，也就是與前述第2安定區間位置資訊相同的前述第2安定區間時序資料進行比較，偵測前述時序取得部取得的前述時序資料的異常。 The data processing device according to claim 3, wherein, based on the timing data of the second stable interval and the position information of the second stable interval, the abnormality detection unit performs the timing data of the second stable interval and the second division unit to other The above-mentioned second stable interval time-series data among the plurality of aforementioned second part time-series data obtained by dividing the aforementioned first vibration interval time-series data of the aforementioned first part of time-series data, that is, the aforementioned second stable interval time-series data that is the same as the aforementioned second stable interval position information The time series data in the second stable interval is compared to detect anomalies in the time series data acquired by the time series acquisition unit.
TW110102460A 2020-07-03 2021-01-22 data processing device TWI795719B (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
WOPCT/JP2020/026255 2020-07-03
PCT/JP2020/026255 WO2022003969A1 (en) 2020-07-03 2020-07-03 Data processing device and data processing method

Publications (2)

Publication Number Publication Date
TW202219770A TW202219770A (en) 2022-05-16
TWI795719B true TWI795719B (en) 2023-03-11

Family

ID=79314973

Family Applications (1)

Application Number Title Priority Date Filing Date
TW110102460A TWI795719B (en) 2020-07-03 2021-01-22 data processing device

Country Status (5)

Country Link
JP (1) JP7114008B2 (en)
KR (1) KR102553397B1 (en)
CN (1) CN115698956A (en)
TW (1) TWI795719B (en)
WO (1) WO2022003969A1 (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2016046705A (en) * 2014-08-25 2016-04-04 コニカミノルタ株式会社 Conference record editing apparatus, method and program for the same, conference record reproduction apparatus, and conference system
JP2018149183A (en) * 2017-03-14 2018-09-27 オムロンヘルスケア株式会社 Blood pressure data processing device, blood pressure data processing method, and program
TW201931466A (en) * 2017-12-28 2019-08-01 日商東京威力科創股份有限公司 Data processing device, data processing method, and program
JP2020067786A (en) * 2018-10-24 2020-04-30 オムロン株式会社 Control device and control program

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP6057786B2 (en) * 2013-03-13 2017-01-11 ヤフー株式会社 Time-series data analysis device, time-series data analysis method, and program
JP2016110534A (en) * 2014-12-10 2016-06-20 日本電信電話株式会社 Information input device using biological sensor, information input method using biological sensor, and information input program using biological sensor
US11137750B2 (en) * 2016-10-06 2021-10-05 Mitsubishi Electric Corporation Time-series data processing device
CN112368683B (en) 2018-07-03 2024-03-26 三菱电机株式会社 Data processing apparatus and data processing method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2016046705A (en) * 2014-08-25 2016-04-04 コニカミノルタ株式会社 Conference record editing apparatus, method and program for the same, conference record reproduction apparatus, and conference system
JP2018149183A (en) * 2017-03-14 2018-09-27 オムロンヘルスケア株式会社 Blood pressure data processing device, blood pressure data processing method, and program
TW201931466A (en) * 2017-12-28 2019-08-01 日商東京威力科創股份有限公司 Data processing device, data processing method, and program
JP2020067786A (en) * 2018-10-24 2020-04-30 オムロン株式会社 Control device and control program

Also Published As

Publication number Publication date
JP7114008B2 (en) 2022-08-05
TW202219770A (en) 2022-05-16
KR20230010773A (en) 2023-01-19
WO2022003969A1 (en) 2022-01-06
KR102553397B1 (en) 2023-07-07
CN115698956A (en) 2023-02-03
JPWO2022003969A1 (en) 2022-01-06

Similar Documents

Publication Publication Date Title
TWI620046B (en) Time-series data processing method, computer readable recording medium having recorded thereon time-series data processing program, and time-series data processing device
US10281360B2 (en) Method and system for the vibration analysis of an engine
JP7340265B2 (en) Abnormality detection device, abnormality detection method, and program
US11209487B2 (en) Rotor diagnostic apparatus, rotor diagnostic method, and rotor diagnostic program
US11782430B2 (en) Abnormality diagnosis method, abnormality diagnosis device and non-transitory computer readable storage medium
RU2724455C1 (en) Diagnostic equipment of deterioration characteristics due to aging and diagnostics method of deterioration due to aging
JP2008166644A (en) Integrated circuit device abnormality detection apparatus, method and program
TW202006488A (en) Data processing device and data processing method
JP2020071624A (en) Abnormality diagnosing apparatus, abnormality diagnosing method and program
JP5535954B2 (en) Soundness evaluation apparatus, method and program
JP6169473B2 (en) Work machine management device and work machine abnormality determination method
TWI795719B (en) data processing device
EP3416012A1 (en) Monitoring device, and method for controlling monitoring device
JP2015161506A (en) Signal analyzer, signal analysis method, degradation diagnosis apparatus, and degradation diagnosis method
JP6459345B2 (en) Fluctuation data management system and its specificity detection method
JP2019113970A (en) State diagnostic method of process and state diagnostic device of process
JP7230371B2 (en) Abnormality detection device, abnormality detection method, abnormality detection program and abnormality detection system
JP6302523B1 (en) Product diagnostic device, product diagnostic method, and product diagnostic program
US20240144101A1 (en) Generation system, computer-readable storage medium, and method for generating waveform evaluation model
CN114862283B (en) Quality inspection method and quality inspection device for machine tool machining
US20240077865A1 (en) Information processing apparatus, information processing method, and computer-readable recording medium
JP5597166B2 (en) Plant monitoring device
WO2023105725A1 (en) Time series data processing method
JP3765480B2 (en) Rotational accuracy measuring device
CN117896301A (en) Baud rate analysis method, device, electronic equipment and storage medium