WO2023218550A1 - Correction equipment, processing method, and processing program - Google Patents

Correction equipment, processing method, and processing program Download PDF

Info

Publication number
WO2023218550A1
WO2023218550A1 PCT/JP2022/019918 JP2022019918W WO2023218550A1 WO 2023218550 A1 WO2023218550 A1 WO 2023218550A1 JP 2022019918 W JP2022019918 W JP 2022019918W WO 2023218550 A1 WO2023218550 A1 WO 2023218550A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
sensor information
correction
normal
waveform data
Prior art date
Application number
PCT/JP2022/019918
Other languages
French (fr)
Japanese (ja)
Inventor
督 那須
Original Assignee
三菱電機株式会社
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 三菱電機株式会社 filed Critical 三菱電機株式会社
Priority to JP2022562766A priority Critical patent/JPWO2023218550A1/ja
Priority to PCT/JP2022/019918 priority patent/WO2023218550A1/en
Publication of WO2023218550A1 publication Critical patent/WO2023218550A1/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
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B9/00—Safety arrangements
    • G05B9/02—Safety arrangements electric

Definitions

  • the present disclosure relates to a correction device, a processing method, and a processing program.
  • Patent Document 1 a technique for estimating the cause of an abnormality in a target has been proposed.
  • the cause of the abnormality is estimated.
  • the abnormal state cannot be corrected and the device cannot be returned to a normal state.
  • the purpose of the present disclosure is to transition the device to a normal state.
  • the correction device includes a plurality of sensor information including first sensor information corresponding to the one or more sensors, which is obtained by measuring a device in which one or more sensors are operating, and whether the first sensor information is normal or not.
  • the apparatus includes a correction information generation section that inputs information regarding an abnormality output by an abnormality determination model and generates correction information for transitioning the device to a normal state, and an output section that outputs the correction information.
  • FIG. 2 is a block diagram showing the functions of the correction device according to the first embodiment.
  • FIG. 3 is a diagram showing hardware included in the correction device according to the first embodiment.
  • 5 is a flowchart illustrating an example of processing executed by the normal data generation unit of the first embodiment.
  • 3 is a diagram illustrating a specific example of processing executed by the normal data generation unit of Embodiment 1.
  • FIG. 5 is a flowchart illustrating an example of processing executed by the waveform data generation section and the abnormality determination section of the first embodiment.
  • 3 is a diagram illustrating a specific example of processing executed by the abnormality determination unit of the first embodiment.
  • FIG. 7 is a flowchart illustrating an example of processing executed by a correction information generation unit and an output unit according to the first embodiment.
  • 7 is a block diagram showing the functions of a correction device according to a second embodiment.
  • FIG. FIG. 7 is a block diagram showing the functions of a correction device according to a third embodiment.
  • FIG. 1 is a block diagram showing the functions of the correction device according to the first embodiment.
  • the correction device 100 is a device that executes a processing method. First, the correction device 100 will be briefly explained.
  • the correction device 100 determines whether the device 200 is in an abnormal state using a plurality of sensor information obtained by one or more sensors measuring the device 200 in operation. For example, when the device 200 is in an abnormal state, the correction device 100 outputs correction information to the stopped device 200. Thereby, the device 200 transitions to a normal state.
  • the device 200 is a train.
  • the plurality of sensor information includes a mascon notch, overhead wire voltage, train acceleration, train speed, torque, and the like.
  • the sensor is an acceleration sensor, a speed sensor, or the like.
  • FIG. 2 is a diagram showing hardware included in the correction device according to the first embodiment.
  • the correction device 100 includes a processor 101, a volatile storage device 102, and a nonvolatile storage device 103.
  • the processor 101 controls the entire correction device 100.
  • the processor 101 is a CPU (Central Processing Unit), an FPGA (Field Programmable Gate Array), or the like.
  • Processor 101 may be a multiprocessor.
  • the correction device 100 may include a processing circuit.
  • Volatile storage device 102 is the main storage device of correction device 100.
  • the volatile storage device 102 is a RAM (Random Access Memory).
  • the nonvolatile storage device 103 is an auxiliary storage device of the correction device 100.
  • the nonvolatile storage device 103 is an HDD (Hard Disk Drive) or an SSD (Solid State Drive).
  • the storage area secured in the volatile storage device 102 or the nonvolatile storage device 103 is called a storage section.
  • the correction device 100 includes an acquisition section 110, a normal data generation section 120, a waveform data generation section 130, an abnormality determination section 140, a correction information generation section 150, and an output section 160.
  • a part or all of the acquisition section 110, the normal data generation section 120, the waveform data generation section 130, the abnormality determination section 140, the correction information generation section 150, and the output section 160 may be realized by a processing circuit. Additionally, some or all of the acquisition unit 110, normal data generation unit 120, waveform data generation unit 130, abnormality determination unit 140, correction information generation unit 150, and output unit 160 are realized as modules of a program executed by the processor 101. You may. For example, the program executed by the processor 101 is also referred to as a processing program. For example, the processing program is recorded on a recording medium.
  • the acquisition unit 110 acquires a plurality of sensor information.
  • the acquisition unit 110 acquires a plurality of sensor information via a network.
  • the plurality of sensor information includes first sensor information corresponding to one or more sensors, which is obtained by measuring the device 200 in which one or more sensors are in operation. That is, the plurality of sensor information includes one or more first sensor information.
  • the plurality of sensor information is assumed to be a mass control notch, overhead wire voltage, train acceleration, train speed, and torque.
  • the first sensor information is the acceleration of the train, the speed of the train, and the torque. Note that although the plurality of sensor information and the first sensor information are defined as described above, the plurality of sensor information and the first sensor information may be data other than the above.
  • the acquisition unit 110 acquires the normal generation information 11. For example, the acquisition unit 110 acquires the normal generation information 11 from the storage unit. Further, for example, the acquisition unit 110 acquires the normal generation information 11 from an external device. Note that the external device is a device that can be connected to the correction device 100. Illustrations of external devices are omitted.
  • the normal generation information 11 is information used to generate normal data of the first sensor information.
  • the normal generation information 11 is information for generating normal data of the first sensor information.
  • the normal generation information 11 is a mathematical model.
  • the normal generation information 11 is generated based on the design information of the device 200.
  • the acquisition unit 110 acquires the abnormality determination model 12.
  • the acquisition unit 110 acquires the abnormality determination model 12 from a storage unit or an external device.
  • the abnormality determination model 12 may be a trained model.
  • the abnormality determination model 12 may be expressed by a predetermined mathematical formula.
  • the acquisition unit 110 acquires the correction information generation model 13.
  • the acquisition unit 110 acquires the correction information generation model 13 from a storage unit or an external device.
  • the correction information generation model 13 may be a learned model. Further, the correction information generation model 13 may be expressed by a predetermined mathematical formula.
  • the normal data generation unit 120 generates normal time series data of the first sensor information using sensor information other than the first sensor information among the plurality of sensor information and the normal generation information 11. Then, the normal data generation unit 120 generates waveform data based on the normal time series data of the first sensor information. Note that the waveform data is also referred to as first waveform data.
  • FIG. 3 is a flowchart illustrating an example of processing executed by the normal data generation unit of the first embodiment.
  • the normal data generation unit 120 acquires one piece of sensor information other than the first sensor information in order of oldest sensor information. Specifically, the normal data generation unit 120 acquires one mascon notch and one overhead wire voltage.
  • the normal data generation unit 120 generates normal data of the first sensor information using sensor information other than the first sensor information and the normal generation information 11. Specifically, the normal data generation unit 120 generates normal data of train acceleration, train speed, and torque.
  • Step S13 The normal data generation unit 120 determines whether the following sensor information exists. If the next sensor information exists, the process advances to step S11. If the next sensor information does not exist, the process proceeds to step S14. Note that when the process proceeds to step S14, normal time-series data of the first sensor information has been generated.
  • the time series data is a plurality of data obtained by repeatedly executing step S12.
  • the normal data generation unit 120 generates waveform data based on the normal time series data of the first sensor information. Specifically, the normal data generation unit 120 generates acceleration waveform data based on normal time series data of train acceleration. The normal data generation unit 120 generates speed waveform data based on normal time series data of train speed. The normal data generation unit 120 generates torque waveform data based on normal torque time series data.
  • FIG. 4 is a diagram illustrating a specific example of processing executed by the normal data generation unit of the first embodiment.
  • the waveform data generation unit 130 generates waveform data based on time-series data of first sensor information among the plurality of sensor information. Note that the waveform data is also referred to as second waveform data.
  • the abnormality determination unit 140 compares the waveform data generated by the normal data generation unit 120 and the waveform data generated by the waveform data generation unit 130.
  • the abnormality determination unit 140 uses the comparison result and the abnormality determination model 12 to determine whether the device 200 is in an abnormal state.
  • the abnormality determination model 12 may be expressed as follows.
  • the abnormality determination model 12 is a model that can input the comparison result and output information regarding an abnormality of the device 200.
  • FIG. 5 is a flowchart illustrating an example of processing executed by the waveform data generation section and the abnormality determination section of the first embodiment.
  • the waveform data generation unit 130 acquires one piece of first sensor information in order of oldest sensor information. Specifically, the waveform data generation unit 130 obtains one acceleration, one velocity, and one torque.
  • Step S22 The waveform data generation unit 130 determines whether the following sensor information exists. If the next sensor information exists, the process advances to step S21. If the next sensor information does not exist, the process proceeds to step S23. Note that when the process proceeds to step S23, acceleration time series data, speed time series data, and torque time series data have been acquired.
  • the waveform data generation unit 130 generates waveform data based on the time series data of the first sensor information. Specifically, the waveform data generation unit 130 generates acceleration waveform data based on acceleration time series data. The waveform data generation unit 130 generates velocity waveform data based on the velocity time series data. The waveform data generation unit 130 generates torque waveform data based on the torque time series data. (Step S24) The abnormality determination unit 140 compares the waveform data generated by the normal data generation unit 120 and the waveform data generated by the waveform data generation unit 130.
  • the abnormality determination unit 140 uses the comparison result and the abnormality determination model 12 to determine whether the device 200 is in an abnormal state. Specifically, the abnormality determination unit 140 inputs the comparison result to the abnormality determination model 12, and the abnormality determination model 12 outputs information indicating whether or not the device 200 is in an abnormal state. Furthermore, when the device 200 is in an abnormal state, the abnormality determination model 12 outputs the cause of the abnormality, the waveform data in which the abnormality was detected, and the abnormal period during which the abnormality is occurring.
  • FIG. 6 is a diagram illustrating a specific example of processing executed by the abnormality determination unit of the first embodiment.
  • the abnormality determination unit 140 compares the torque waveform data generated by the normal data generation unit 120 and the torque waveform data generated by the waveform data generation unit 130.
  • the abnormality determination unit 140 compares the acceleration waveform data generated by the normal data generation unit 120 and the acceleration waveform data generated by the waveform data generation unit 130.
  • the abnormality determination unit 140 compares the velocity waveform data generated by the normal data generation unit 120 and the velocity waveform data generated by the waveform data generation unit 130.
  • the abnormality determination unit 140 may compare the values at each time. Further, the abnormality determination unit 140 may compare the sum of values in a predetermined period.
  • the abnormality determination model 12 determines that the device 200 is normal if there is a difference between the torque waveform data and the speed waveform data in the entire interval, and if the difference between the speed waveform data is smaller than a threshold value. It is determined that the state is
  • the correction information generation unit 150 When the device 200 is in an abnormal state, the correction information generation unit 150 receives at least the information regarding the abnormality output by the abnormality determination model 12 and generates correction information for transitioning the device 200 to a normal state. Further, when the device 200 is in an abnormal state, the correction information generation unit 150 generates the correction information using at least the information regarding the abnormality output by the abnormality determination model 12 and the correction information generation model 13. good.
  • This sentence can be expressed as follows.
  • the correction information generation unit 150 When the device 200 is in an abnormal state, the correction information generation unit 150 generates information regarding the abnormality, which is at least one of the cause of the abnormality, the waveform data in which the abnormality was detected, and the abnormal period during which the abnormality is occurring;
  • the correction information is generated using the correction information generation model 13. Note that, for example, the correction information is a parameter.
  • the correction information generation unit 150 When the device 200 is in an abnormal state, the correction information generation unit 150 generates the correction information using sensor information other than the first sensor information, information regarding the abnormality, and the correction information generation model 13. Good too.
  • the sensor information other than the first sensor information is the sensor information of the mascon notch.
  • the waveform data in which an abnormality is detected is torque waveform data.
  • the parameter of the inverter that receives the input of the mascon notch and supplies power to the motor shifts to a normal state.
  • the waveform data in which an abnormality has been detected is velocity waveform data.
  • the correction information generation model 13 may generate correction information based on data in a section after the abnormal period.
  • the output unit 160 outputs correction information.
  • the output unit 160 outputs correction information and correction information setting instructions to the device 200. Accordingly, the correction information is set in the device 200. Then, the device 200 transitions to a normal state. Further, for example, the output unit 160 outputs the correction information to a terminal owned by a maintenance worker. This allows the maintenance worker to set the correction information in the device 200. Then, the device 200 transitions to a normal state through the setting work performed by the maintenance worker.
  • FIG. 7 is a flowchart illustrating an example of processing executed by the correction information generation unit and output unit of the first embodiment.
  • the correction information generation unit 150 generates correction information using the cause of the abnormality, the waveform data in which the abnormality was detected, the abnormal period during which the abnormality has occurred, and the correction information generation model 13. Specifically, the correction information generation unit 150 inputs the cause of the abnormality, the waveform data in which the abnormality was detected, and the abnormal period in which the abnormality has occurred to the correction information generation model 13, so that the correction information generation model 13 Output correction information.
  • the output unit 160 outputs the correction information.
  • the correction device 100 generates correction information when the device 200 is in an abnormal state. Correction device 100 outputs correction information. Thereby, the device 200 transitions to a normal state. Therefore, the correction device 100 can shift the device 200 to a normal state.
  • Embodiment 2 Next, a second embodiment will be described. In the second embodiment, matters that are different from the first embodiment will be mainly explained. In the second embodiment, explanations of matters common to the first embodiment will be omitted.
  • correction information is output to the stopped device 200 in operation.
  • the correction information is limited to information that can be set in the device 200 in operation.
  • the information is a parameter within a changeable range. Furthermore, the range may be determined based on design information.
  • the correction information includes information that indirectly changes information whose settings are not allowed to be changed while the device 200 is in operation. For example, if the information whose settings cannot be changed while the device 200 is in operation is a voltage parameter, the correction information is a resistance parameter. Specifically, when increasing the voltage, the correction information is a parameter of the resistance, which is a low value. By setting the parameters of the resistance in the device 200, the current increases and the voltage increases.
  • the information that indirectly changes the information whose settings are not allowed to be changed while the device 200 is in operation may be one parameter or may be multiple parameters.
  • FIG. 8 is a block diagram showing the functions of the correction device according to the second embodiment. Components in FIG. 8 that are the same as those shown in FIG. 1 are designated by the same reference numerals as those shown in FIG.
  • the correction device 100a includes a correction information generation section 150a.
  • the acquisition unit 110 acquires a plurality of sensor information.
  • the plurality of sensor information may be expressed as real-time information.
  • the correction information generation unit 150a returns the operating device 200 to a normal state using at least the information regarding the abnormality output by the abnormality determination model 12 and the correction information generation model 13. Generate correction information for migration.
  • the correction device 100a generates correction information for shifting the operating device 200 to a normal state.
  • the correction device 100a outputs correction information.
  • the operating device 200 transitions to a normal state. Therefore, the correction device 100a can shift the operating device 200 to a normal state.
  • Embodiment 3 Next, Embodiment 3 will be described. In the third embodiment, matters that are different from the first embodiment will be mainly explained. In the third embodiment, explanations of matters common to the first embodiment will be omitted.
  • FIG. 9 is a block diagram showing the functions of the correction device according to the third embodiment. Components in FIG. 9 that are the same as those shown in FIG. 1 are designated by the same reference numerals as those shown in FIG.
  • the correction device 100 further includes a relearning section 170. Part or all of the relearning unit 170 may be realized by a processing circuit. Further, part or all of the relearning unit 170 may be realized as a module of a program executed by the processor 101.
  • the acquisition unit 110 acquires the correction results. For example, the acquisition unit 110 acquires the correction result from the device 200. Further, for example, the acquisition unit 110 acquires the correction result from an external device.
  • the relearning unit 170 relearns the abnormality determination model 12 and the correction information generation model 13 when the correction result indicates an abnormality. Further, the relearning unit 170 may relearn the abnormality determination model 12 and the correction information generation model 13 when a correction result indicating an abnormality is obtained equal to or greater than a threshold value. Relearning will be explained in detail below.
  • the relearning unit 170 uses the comparison result between the waveform data generated by the normal data generation unit 120 and the waveform data generated by the waveform data generation unit 130 as learning data.
  • the relearning unit 170 uses information regarding the abnormality (for example, the cause of the abnormality, etc.) used when generating the correction information as learning data. Further, the learning data may be labeled as "abnormal".
  • part or all of previously used learning data may be used. Further, similar data among previously used learning data may be excluded. When part or all of previously used learning data is used, the same number of learning data as the previous learning data may be used in relearning.
  • the correction device 100 can generate correction information for transitioning the device 200 to a normal state. .

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

Correction equipment (100) comprises: an acquisition unit (110) that acquires normal generation information (11) and a plurality of sets of sensor information including first sensor information obtained by measuring a device (200); a normal data generation unit (120) that generates first waveform data on the basis of normal time-series data for the first sensor information, said normal time-series data being generated using the normal generation information (11) and sensor information other than the first sensor information; a waveform data generation unit (130) that generates second waveform data on the basis of time-series data of the first sensor information; an abnormality determination unit (140) that determines whether or not the device (200) is in an abnormal state using an abnormality determination model (12) that can receive input of the result of a comparison between the first waveform data and the second waveform data and output information relating to an abnormality in the device (200); a correction information generation unit (150) that generates correction information if the device (200) is in the abnormal state; and an output unit (160) that outputs the correction information.

Description

補正装置、処理方法、及び処理プログラムCorrection device, processing method, and processing program
　本開示は、補正装置、処理方法、及び処理プログラムに関する。 The present disclosure relates to a correction device, a processing method, and a processing program.
　機器が異常な動作を行うことは、防ぐ必要がある。特に、機器が列車である場合、列車が異常な動作を行うことは、人命に大きな影響を与える。ここで、対象の異常の原因を推定する技術が提案されている（特許文献１を参照）。 It is necessary to prevent equipment from operating abnormally. In particular, when the device is a train, abnormal movement of the train has a great impact on human life. Here, a technique for estimating the cause of an abnormality in a target has been proposed (see Patent Document 1).
国際公開第２０１６／１９５０９２号International Publication No. 2016/195092
　上記の技術では、異常の原因が推定される。しかし、上記の技術では、異常な状態が是正されて、機器が、正常な状態に移行できていない。 With the above technology, the cause of the abnormality is estimated. However, with the above techniques, the abnormal state cannot be corrected and the device cannot be returned to a normal state.
　本開示の目的は、機器を正常な状態に移行させることである。 The purpose of the present disclosure is to transition the device to a normal state.
　本開示の一態様に係る補正装置が提供される。補正装置は、１以上のセンサが動作中の機器を計測することで得られる、前記１以上のセンサと対応した第１のセンサ情報を含む複数のセンサ情報、及び前記第１のセンサ情報の正常なデータの生成に用いられる正常生成情報を取得する取得部と、前記複数のセンサ情報のうち、前記第１のセンサ情報以外のセンサ情報と、前記正常生成情報とを用いて、前記第１のセンサ情報の正常な時系列データを生成し、前記第１のセンサ情報の正常な時系列データに基づいて、第１の波形データを生成する正常データ生成部と、前記複数のセンサ情報のうちの前記第１のセンサ情報の時系列データに基づいて、第２の波形データを生成する波形データ生成部と、前記第１の波形データと前記第２の波形データとの比較結果を入力して前記機器の異常に関する情報を出力することが可能な異常判定モデルを用いて、前記機器が異常な状態であるか否かを判定する異常判定部と、前記機器が異常な状態である場合、少なくとも前記異常判定モデルが出力した異常に関する情報を入力して前記機器を正常な状態に移行させるための補正情報を生成する補正情報生成部と、前記補正情報を出力する出力部と、を有する。 A correction device according to one aspect of the present disclosure is provided. The correction device includes a plurality of sensor information including first sensor information corresponding to the one or more sensors, which is obtained by measuring a device in which one or more sensors are operating, and whether the first sensor information is normal or not. an acquisition unit that acquires normal generation information used to generate data; and sensor information other than the first sensor information among the plurality of sensor information, and the normal generation information to generate the first a normal data generation unit that generates normal time series data of the sensor information and generates first waveform data based on the normal time series data of the first sensor information; a waveform data generation unit that generates second waveform data based on time series data of the first sensor information; and a waveform data generation unit that inputs a comparison result between the first waveform data and the second waveform data, and an abnormality determination unit that determines whether the device is in an abnormal state using an abnormality determination model capable of outputting information regarding an abnormality of the device; The apparatus includes a correction information generation section that inputs information regarding an abnormality output by an abnormality determination model and generates correction information for transitioning the device to a normal state, and an output section that outputs the correction information.
　本開示によれば、機器を正常な状態に移行させることができる。 According to the present disclosure, it is possible to transition the device to a normal state.
実施の形態１の補正装置の機能を示すブロック図である。FIG. 2 is a block diagram showing the functions of the correction device according to the first embodiment. 実施の形態１の補正装置が有するハードウェアを示す図である。FIG. 3 is a diagram showing hardware included in the correction device according to the first embodiment. 実施の形態１の正常データ生成部が実行する処理の例を示すフローチャートである。5 is a flowchart illustrating an example of processing executed by the normal data generation unit of the first embodiment. 実施の形態１の正常データ生成部が実行する処理の具体例を示す図である。3 is a diagram illustrating a specific example of processing executed by the normal data generation unit of Embodiment 1. FIG. 実施の形態１の波形データ生成部と異常判定部が実行する処理の例を示すフローチャートである。5 is a flowchart illustrating an example of processing executed by the waveform data generation section and the abnormality determination section of the first embodiment. 実施の形態１の異常判定部が実行する処理の具体例を示す図である。3 is a diagram illustrating a specific example of processing executed by the abnormality determination unit of the first embodiment. FIG. 実施の形態１の補正情報生成部と出力部が実行する処理の例を示すフローチャートである。7 is a flowchart illustrating an example of processing executed by a correction information generation unit and an output unit according to the first embodiment. 実施の形態２の補正装置の機能を示すブロック図である。7 is a block diagram showing the functions of a correction device according to a second embodiment. FIG. 実施の形態３の補正装置の機能を示すブロック図である。FIG. 7 is a block diagram showing the functions of a correction device according to a third embodiment.
　以下、図面を参照しながら実施の形態を説明する。以下の実施の形態は、例にすぎず、本開示の範囲内で種々の変更が可能である。 Hereinafter, embodiments will be described with reference to the drawings. The following embodiments are merely examples, and various modifications can be made within the scope of the present disclosure.
実施の形態１．
　図１は、実施の形態１の補正装置の機能を示すブロック図である。補正装置１００は、処理方法を実行する装置である。
　まず、補正装置１００を簡単に説明する。補正装置１００は、１以上のセンサが動作中の機器２００を計測することで得られた複数のセンサ情報を用いて、機器２００が異常な状態であるか否かを判定する。例えば、補正装置１００は、機器２００が異常な状態である場合、停止中の機器２００に補正情報を出力する。これにより、機器２００は、正常な状態に移行する。 Embodiment 1.
FIG. 1 is a block diagram showing the functions of the correction device according to the first embodiment. The correction device 100 is a device that executes a processing method.
First, the correction device 100 will be briefly explained. The correction device 100 determines whether the device 200 is in an abnormal state using a plurality of sensor information obtained by one or more sensors measuring the device 200 in operation. For example, when the device 200 is in an abnormal state, the correction device 100 outputs correction information to the stopped device 200. Thereby, the device 200 transitions to a normal state.
　ここで、例えば、機器２００は、列車である。例えば、機器２００が列車である場合、複数のセンサ情報は、マスコンノッチ、架線電圧、列車の加速度、列車の速度、トルクなどである。また、例えば、センサは、加速度センサ、速度センサなどである。 Here, for example, the device 200 is a train. For example, when the device 200 is a train, the plurality of sensor information includes a mascon notch, overhead wire voltage, train acceleration, train speed, torque, and the like. Further, for example, the sensor is an acceleration sensor, a speed sensor, or the like.
　次に、補正装置１００が有するハードウェアを説明する。
　図２は、実施の形態１の補正装置が有するハードウェアを示す図である。補正装置１００は、プロセッサ１０１、揮発性記憶装置１０２、及び不揮発性記憶装置１０３を有する。 Next, the hardware included in the correction device 100 will be explained.
FIG. 2 is a diagram showing hardware included in the correction device according to the first embodiment. The correction device 100 includes a processor 101, a volatile storage device 102, and a nonvolatile storage device 103.
　プロセッサ１０１は、補正装置１００全体を制御する。例えば、プロセッサ１０１は、ＣＰＵ（Ｃｅｎｔｒａｌ　Ｐｒｏｃｅｓｓｉｎｇ　Ｕｎｉｔ）、ＦＰＧＡ（Ｆｉｅｌｄ　Ｐｒｏｇｒａｍｍａｂｌｅ　Ｇａｔｅ　Ａｒｒａｙ）などである。プロセッサ１０１は、マルチプロセッサでもよい。また、補正装置１００は、処理回路を有してもよい。 The processor 101 controls the entire correction device 100. For example, the processor 101 is a CPU (Central Processing Unit), an FPGA (Field Programmable Gate Array), or the like. Processor 101 may be a multiprocessor. Further, the correction device 100 may include a processing circuit.
　揮発性記憶装置１０２は、補正装置１００の主記憶装置である。例えば、揮発性記憶装置１０２は、ＲＡＭ（Ｒａｎｄｏｍ　Ａｃｃｅｓｓ　Ｍｅｍｏｒｙ）である。不揮発性記憶装置１０３は、補正装置１００の補助記憶装置である。例えば、不揮発性記憶装置１０３は、ＨＤＤ（Ｈａｒｄ　Ｄｉｓｋ　Ｄｒｉｖｅ）、又はＳＳＤ（Ｓｏｌｉｄ　Ｓｔａｔｅ　Ｄｒｉｖｅ）である。
　また、揮発性記憶装置１０２又は不揮発性記憶装置１０３に確保した記憶領域は、記憶部と呼ぶ。 Volatile storage device 102 is the main storage device of correction device 100. For example, the volatile storage device 102 is a RAM (Random Access Memory). The nonvolatile storage device 103 is an auxiliary storage device of the correction device 100. For example, the nonvolatile storage device 103 is an HDD (Hard Disk Drive) or an SSD (Solid State Drive).
Further, the storage area secured in the volatile storage device 102 or the nonvolatile storage device 103 is called a storage section.
　図１に戻って、補正装置１００が有する機能を説明する。
　補正装置１００は、取得部１１０、正常データ生成部１２０、波形データ生成部１３０、異常判定部１４０、補正情報生成部１５０、及び出力部１６０を有する。 Returning to FIG. 1, the functions of the correction device 100 will be explained.
The correction device 100 includes an acquisition section 110, a normal data generation section 120, a waveform data generation section 130, an abnormality determination section 140, a correction information generation section 150, and an output section 160.
　取得部１１０、正常データ生成部１２０、波形データ生成部１３０、異常判定部１４０、補正情報生成部１５０、及び出力部１６０の一部又は全部は、処理回路によって実現してもよい。また、取得部１１０、正常データ生成部１２０、波形データ生成部１３０、異常判定部１４０、補正情報生成部１５０、及び出力部１６０の一部又は全部は、プロセッサ１０１が実行するプログラムのモジュールとして実現してもよい。例えば、プロセッサ１０１が実行するプログラムは、処理プログラムとも言う。例えば、処理プログラムは、記録媒体に記録されている。 A part or all of the acquisition section 110, the normal data generation section 120, the waveform data generation section 130, the abnormality determination section 140, the correction information generation section 150, and the output section 160 may be realized by a processing circuit. Additionally, some or all of the acquisition unit 110, normal data generation unit 120, waveform data generation unit 130, abnormality determination unit 140, correction information generation unit 150, and output unit 160 are realized as modules of a program executed by the processor 101. You may. For example, the program executed by the processor 101 is also referred to as a processing program. For example, the processing program is recorded on a recording medium.
　取得部１１０は、複数のセンサ情報を取得する。例えば、取得部１１０は、ネットワークを介して、複数のセンサ情報を取得する。
　複数のセンサ情報は、１以上のセンサが動作中の機器２００を計測することで得られる、１以上のセンサと対応した第１のセンサ情報を含む。すなわち、複数のセンサ情報は、１以上の第１のセンサ情報を含む。以下の説明では、複数のセンサ情報は、マスコンノッチ、架線電圧、列車の加速度、列車の速度、及びトルクとする。また、第１のセンサ情報は、列車の加速度、列車の速度、及びトルクとする。なお、複数のセンサ情報及び第１のセンサ情報は、上記のように定められるが、複数のセンサ情報及び第１のセンサ情報は、上記以外のデータでもよい。 The acquisition unit 110 acquires a plurality of sensor information. For example, the acquisition unit 110 acquires a plurality of sensor information via a network.
The plurality of sensor information includes first sensor information corresponding to one or more sensors, which is obtained by measuring the device 200 in which one or more sensors are in operation. That is, the plurality of sensor information includes one or more first sensor information. In the following description, the plurality of sensor information is assumed to be a mass control notch, overhead wire voltage, train acceleration, train speed, and torque. Further, the first sensor information is the acceleration of the train, the speed of the train, and the torque. Note that although the plurality of sensor information and the first sensor information are defined as described above, the plurality of sensor information and the first sensor information may be data other than the above.
　取得部１１０は、正常生成情報１１を取得する。例えば、取得部１１０は、正常生成情報１１を記憶部から取得する。また、例えば、取得部１１０は、正常生成情報１１を外部装置から取得する。なお、外部装置は、補正装置１００に接続可能な装置である。外部装置の図は、省略されている。 The acquisition unit 110 acquires the normal generation information 11. For example, the acquisition unit 110 acquires the normal generation information 11 from the storage unit. Further, for example, the acquisition unit 110 acquires the normal generation information 11 from an external device. Note that the external device is a device that can be connected to the correction device 100. Illustrations of external devices are omitted.
　正常生成情報１１は、第１のセンサ情報の正常なデータの生成に用いられる情報である。言い換えれば、正常生成情報１１は、第１のセンサ情報の正常なデータを生成するための情報である。詳細には、正常生成情報１１は、数理モデルである。例えば、正常生成情報１１は、機器２００の設計情報に基づいて、生成される。 The normal generation information 11 is information used to generate normal data of the first sensor information. In other words, the normal generation information 11 is information for generating normal data of the first sensor information. Specifically, the normal generation information 11 is a mathematical model. For example, the normal generation information 11 is generated based on the design information of the device 200.
　取得部１１０は、異常判定モデル１２を取得する。例えば、取得部１１０は、異常判定モデル１２を記憶部又は外部装置から取得する。なお、異常判定モデル１２は、学習済モデルでもよい。また、異常判定モデル１２は、予め決められた数式などで表現されるものでもよい。 The acquisition unit 110 acquires the abnormality determination model 12. For example, the acquisition unit 110 acquires the abnormality determination model 12 from a storage unit or an external device. Note that the abnormality determination model 12 may be a trained model. Further, the abnormality determination model 12 may be expressed by a predetermined mathematical formula.
　取得部１１０は、補正情報生成モデル１３を取得する。例えば、取得部１１０は、補正情報生成モデル１３を記憶部又は外部装置から取得する。なお、補正情報生成モデル１３は、学習済モデルでもよい。また、補正情報生成モデル１３は、予め決められた数式などで表現されるものでもよい。 The acquisition unit 110 acquires the correction information generation model 13. For example, the acquisition unit 110 acquires the correction information generation model 13 from a storage unit or an external device. Note that the correction information generation model 13 may be a learned model. Further, the correction information generation model 13 may be expressed by a predetermined mathematical formula.
　正常データ生成部１２０は、複数のセンサ情報のうち、第１のセンサ情報以外のセンサ情報と、正常生成情報１１とを用いて、第１のセンサ情報の正常な時系列データを生成する。そして、正常データ生成部１２０は、第１のセンサ情報の正常な時系列データに基づいて、波形データを生成する。なお、当該波形データは、第１の波形データとも言う。 The normal data generation unit 120 generates normal time series data of the first sensor information using sensor information other than the first sensor information among the plurality of sensor information and the normal generation information 11. Then, the normal data generation unit 120 generates waveform data based on the normal time series data of the first sensor information. Note that the waveform data is also referred to as first waveform data.
　次に、正常データ生成部１２０が実行する処理を、フローチャートを用いて説明する。
　図３は、実施の形態１の正常データ生成部が実行する処理の例を示すフローチャートである。
　（ステップＳ１１）正常データ生成部１２０は、古い順に、第１のセンサ情報以外のセンサ情報を１つ取得する。具体的には、正常データ生成部１２０は、１つのマスコンノッチ、及び１つの架線電圧を取得する。
　（ステップＳ１２）正常データ生成部１２０は、第１のセンサ情報以外のセンサ情報と、正常生成情報１１とを用いて、第１のセンサ情報の正常なデータを生成する。具体的には、正常データ生成部１２０は、列車の加速度、列車の速度、及びトルクの正常なデータを生成する。 Next, the processing executed by the normal data generation unit 120 will be explained using a flowchart.
FIG. 3 is a flowchart illustrating an example of processing executed by the normal data generation unit of the first embodiment.
(Step S11) The normal data generation unit 120 acquires one piece of sensor information other than the first sensor information in order of oldest sensor information. Specifically, the normal data generation unit 120 acquires one mascon notch and one overhead wire voltage.
(Step S12) The normal data generation unit 120 generates normal data of the first sensor information using sensor information other than the first sensor information and the normal generation information 11. Specifically, the normal data generation unit 120 generates normal data of train acceleration, train speed, and torque.
　（ステップＳ１３）正常データ生成部１２０は、次のセンサ情報が存在するか否かを判定する。次のセンサ情報が存在する場合、処理は、ステップＳ１１に進む。次のセンサ情報が存在しない場合、処理は、ステップＳ１４に進む。なお、処理が、ステップＳ１４に進む場合、第１のセンサ情報の正常な時系列データが生成されている。当該時系列データは、ステップＳ１２が繰り返し実行されることで得られた複数のデータである。 (Step S13) The normal data generation unit 120 determines whether the following sensor information exists. If the next sensor information exists, the process advances to step S11. If the next sensor information does not exist, the process proceeds to step S14. Note that when the process proceeds to step S14, normal time-series data of the first sensor information has been generated. The time series data is a plurality of data obtained by repeatedly executing step S12.
　（ステップＳ１４）正常データ生成部１２０は、第１のセンサ情報の正常な時系列データに基づいて、波形データを生成する。具体的には、正常データ生成部１２０は、列車の加速度の正常な時系列データに基づいて、加速度の波形データを生成する。正常データ生成部１２０は、列車の速度の正常な時系列データに基づいて、速度の波形データを生成する。正常データ生成部１２０は、トルクの正常な時系列データに基づいて、トルクの波形データを生成する。 (Step S14) The normal data generation unit 120 generates waveform data based on the normal time series data of the first sensor information. Specifically, the normal data generation unit 120 generates acceleration waveform data based on normal time series data of train acceleration. The normal data generation unit 120 generates speed waveform data based on normal time series data of train speed. The normal data generation unit 120 generates torque waveform data based on normal torque time series data.
　次に、正常データ生成部１２０が実行する処理を具体的に説明する。
　図４は、実施の形態１の正常データ生成部が実行する処理の具体例を示す図である。正常データ生成部１２０は、“ｔ＝０”のマスコンノッチと、“ｔ＝０”の架線電圧とを取得する。正常データ生成部１２０は、“ｔ＝０”のマスコンノッチと、“ｔ＝０”の架線電圧と、正常生成情報１１とを用いて、“ｔ＝０”の加速度の正常なデータ、“ｔ＝０”の速度の正常なデータ、及び“ｔ＝０”のトルクの正常なデータを生成する。 Next, the processing executed by the normal data generation unit 120 will be specifically explained.
FIG. 4 is a diagram illustrating a specific example of processing executed by the normal data generation unit of the first embodiment. The normal data generation unit 120 acquires the mask notch at "t=0" and the overhead wire voltage at "t=0". The normal data generation unit 120 uses the mask notch at “t=0”, the overhead wire voltage at “t=0”, and the normal generation information 11 to generate normal data for the acceleration at “t=0”, “t Normal speed data at "t=0" and normal torque data at "t=0" are generated.
　正常データ生成部１２０は、“ｔ＝１”のマスコンノッチと、“ｔ＝１”の架線電圧とを取得する。正常データ生成部１２０は、“ｔ＝１”のマスコンノッチと、“ｔ＝１”の架線電圧と、正常生成情報１１とを用いて、“ｔ＝１”の加速度の正常なデータ、“ｔ＝１”の速度の正常なデータ、及び“ｔ＝１”のトルクの正常なデータを生成する。正常データ生成部１２０は、“ｔ＝２～６００”のデータを用いて、同様の処理を繰り返す。 The normal data generation unit 120 acquires the mask notch at “t=1” and the overhead wire voltage at “t=1”. The normal data generation unit 120 uses the mask notch at “t=1”, the overhead line voltage at “t=1”, and the normal generation information 11 to generate normal data for the acceleration at “t=1”, “t Normal speed data of "t=1" and normal torque data of "t=1" are generated. The normal data generation unit 120 repeats the same process using data of “t=2 to 600”.
　正常データ生成部１２０は、“ｔ＝０～６００”のトルクの正常なデータに基づいて、トルクの波形データを生成する。正常データ生成部１２０は、“ｔ＝０～６００”の加速度の正常なデータに基づいて、加速度の波形データを生成する。正常データ生成部１２０は、“ｔ＝０～６００”の速度の正常なデータに基づいて、速度の波形データを生成する。 The normal data generation unit 120 generates torque waveform data based on the normal torque data of “t=0 to 600”. The normal data generation unit 120 generates acceleration waveform data based on normal acceleration data of "t=0 to 600". The normal data generation unit 120 generates velocity waveform data based on the normal velocity data of "t=0 to 600".
　また、正常データ生成部１２０は、“ｔ＝ｎ（ｎは、整数）”の加速度の正常なデータ、“ｔ＝ｎ”の速度の正常なデータ、及び“ｔ＝ｎ”のトルクの正常なデータを生成する場合、“ｔ＝ｎ－１”のマスコンノッチと、“ｔ＝ｎ－１”の架線電圧とを用いてもよい。さらに、正常データ生成部１２０は、“ｔ＝０～ｎ－２”のマスコンノッチと、“ｔ＝０～ｎ－２”の架線電圧とを用いてもよい。正常データ生成部１２０は、“ｔ＝０～ｎ－１”の加速度、“ｔ＝０～ｎ－１”の速度、及び“ｔ＝０～ｎ－１”のトルクを用いてもよい。 The normal data generation unit 120 also generates normal acceleration data at “t=n” (n is an integer), normal speed data at “t=n”, and normal torque data at “t=n”. When generating data, a mask notch of "t=n-1" and an overhead line voltage of "t=n-1" may be used. Further, the normal data generation unit 120 may use a mask notch of "t=0 to n-2" and an overhead line voltage of "t=0 to n-2". The normal data generation unit 120 may use acceleration of "t=0 to n-1", speed of "t=0 to n-1", and torque of "t=0 to n-1".
　図１に戻って、波形データ生成部１３０を説明する。
　波形データ生成部１３０は、複数のセンサ情報のうちの第１のセンサ情報の時系列データに基づいて、波形データを生成する。なお、当該波形データは、第２の波形データとも言う。 Returning to FIG. 1, the waveform data generation section 130 will be explained.
The waveform data generation unit 130 generates waveform data based on time-series data of first sensor information among the plurality of sensor information. Note that the waveform data is also referred to as second waveform data.
　異常判定部１４０は、正常データ生成部１２０が生成した波形データと、波形データ生成部１３０が生成した波形データとを比較する。異常判定部１４０は、比較結果と異常判定モデル１２とを用いて、機器２００が異常な状態であるか否かを判定する。ここで、異常判定モデル１２は、次のように表現してもよい。異常判定モデル１２は、当該比較結果を入力して機器２００の異常に関する情報を出力することが可能なモデルである。 The abnormality determination unit 140 compares the waveform data generated by the normal data generation unit 120 and the waveform data generated by the waveform data generation unit 130. The abnormality determination unit 140 uses the comparison result and the abnormality determination model 12 to determine whether the device 200 is in an abnormal state. Here, the abnormality determination model 12 may be expressed as follows. The abnormality determination model 12 is a model that can input the comparison result and output information regarding an abnormality of the device 200.
　波形データ生成部１３０と異常判定部１４０が実行する処理を、フローチャートを用いて説明する。
　図５は、実施の形態１の波形データ生成部と異常判定部が実行する処理の例を示すフローチャートである。
　（ステップＳ２１）波形データ生成部１３０は、古い順に、第１のセンサ情報を１つ取得する。具体的には、波形データ生成部１３０は、１つの加速度、１つの速度、及び１つのトルクを取得する。 The processing executed by the waveform data generation section 130 and the abnormality determination section 140 will be explained using a flowchart.
FIG. 5 is a flowchart illustrating an example of processing executed by the waveform data generation section and the abnormality determination section of the first embodiment.
(Step S21) The waveform data generation unit 130 acquires one piece of first sensor information in order of oldest sensor information. Specifically, the waveform data generation unit 130 obtains one acceleration, one velocity, and one torque.
　（ステップＳ２２）波形データ生成部１３０は、次のセンサ情報が存在するか否かを判定する。次のセンサ情報が存在する場合、処理は、ステップＳ２１に進む。次のセンサ情報が存在しない場合、処理は、ステップＳ２３に進む。なお、処理が、ステップＳ２３に進む場合、加速度の時系列データ、速度の時系列データ、及びトルクの時系列データが取得されている。 (Step S22) The waveform data generation unit 130 determines whether the following sensor information exists. If the next sensor information exists, the process advances to step S21. If the next sensor information does not exist, the process proceeds to step S23. Note that when the process proceeds to step S23, acceleration time series data, speed time series data, and torque time series data have been acquired.
　（ステップＳ２３）波形データ生成部１３０は、第１のセンサ情報の時系列データに基づいて、波形データを生成する。具体的には、波形データ生成部１３０は、加速度の時系列データに基づいて、加速度の波形データを生成する。波形データ生成部１３０は、速度の時系列データに基づいて、速度の波形データを生成する。波形データ生成部１３０は、トルクの時系列データに基づいて、トルクの波形データを生成する。
　（ステップＳ２４）異常判定部１４０は、正常データ生成部１２０が生成した波形データと、波形データ生成部１３０が生成した波形データとを比較する。 (Step S23) The waveform data generation unit 130 generates waveform data based on the time series data of the first sensor information. Specifically, the waveform data generation unit 130 generates acceleration waveform data based on acceleration time series data. The waveform data generation unit 130 generates velocity waveform data based on the velocity time series data. The waveform data generation unit 130 generates torque waveform data based on the torque time series data.
(Step S24) The abnormality determination unit 140 compares the waveform data generated by the normal data generation unit 120 and the waveform data generated by the waveform data generation unit 130.
　（ステップＳ２５）異常判定部１４０は、比較結果と異常判定モデル１２とを用いて、機器２００が異常な状態であるか否かを判定する。詳細には、異常判定部１４０が、比較結果を異常判定モデル１２に入力することで、異常判定モデル１２は、機器２００が異常な状態であるか否かを示す情報を出力する。
　また、機器２００が異常な状態である場合、異常判定モデル１２は、異常原因、異常が検出された波形データ、及び異常が発生している異常期間を出力する。 (Step S25) The abnormality determination unit 140 uses the comparison result and the abnormality determination model 12 to determine whether the device 200 is in an abnormal state. Specifically, the abnormality determination unit 140 inputs the comparison result to the abnormality determination model 12, and the abnormality determination model 12 outputs information indicating whether or not the device 200 is in an abnormal state.
Furthermore, when the device 200 is in an abnormal state, the abnormality determination model 12 outputs the cause of the abnormality, the waveform data in which the abnormality was detected, and the abnormal period during which the abnormality is occurring.
　次に、異常判定部１４０が実行する処理を具体的に説明する。
　図６は、実施の形態１の異常判定部が実行する処理の具体例を示す図である。異常判定部１４０は、正常データ生成部１２０が生成したトルクの波形データと、波形データ生成部１３０が生成したトルクの波形データとを比較する。異常判定部１４０は、正常データ生成部１２０が生成した加速度の波形データと、波形データ生成部１３０が生成した加速度の波形データとを比較する。異常判定部１４０は、正常データ生成部１２０が生成した速度の波形データと、波形データ生成部１３０が生成した速度の波形データとを比較する。 Next, the processing executed by the abnormality determination unit 140 will be specifically explained.
FIG. 6 is a diagram illustrating a specific example of processing executed by the abnormality determination unit of the first embodiment. The abnormality determination unit 140 compares the torque waveform data generated by the normal data generation unit 120 and the torque waveform data generated by the waveform data generation unit 130. The abnormality determination unit 140 compares the acceleration waveform data generated by the normal data generation unit 120 and the acceleration waveform data generated by the waveform data generation unit 130. The abnormality determination unit 140 compares the velocity waveform data generated by the normal data generation unit 120 and the velocity waveform data generated by the waveform data generation unit 130.
　異常判定部１４０は、比較を行う場合、時刻毎に値を比較してもよい。また、異常判定部１４０は、予め定められた期間における値の総和を比較してもよい。 When performing the comparison, the abnormality determination unit 140 may compare the values at each time. Further, the abnormality determination unit 140 may compare the sum of values in a predetermined period.
　異常判定部１４０は、比較結果を異常判定モデル１２に入力する。例えば、異常判定モデル１２は、“ｔ＝２４０～２６０”における、トルクの波形データと加速度の波形データとに差が生じている場合、機器２００が異常な状態であると判定する。また、例えば、異常判定モデル１２は、“ｔ＝２４０～２６０”における、トルクの波形データと速度の波形データとに差が生じている場合、かつ速度の波形データの差が閾値よりも小さい場合、機器２００が正常な状態であると判定する。また、例えば、異常判定モデル１２は、全区間における、トルクの波形データと速度の波形データとに差が生じている場合、かつ速度の波形データの差が閾値よりも小さい場合、機器２００が正常な状態であると判定する。 The abnormality determination unit 140 inputs the comparison results to the abnormality determination model 12. For example, the abnormality determination model 12 determines that the device 200 is in an abnormal state when there is a difference between the torque waveform data and the acceleration waveform data at "t=240 to 260". Further, for example, the abnormality determination model 12 is configured such that when there is a difference between the torque waveform data and the speed waveform data at "t=240 to 260", and when the difference between the speed waveform data is smaller than the threshold value, , it is determined that the device 200 is in a normal state. For example, the abnormality determination model 12 determines that the device 200 is normal if there is a difference between the torque waveform data and the speed waveform data in the entire interval, and if the difference between the speed waveform data is smaller than a threshold value. It is determined that the state is
　図１に戻って、補正情報生成部１５０を説明する。
　補正情報生成部１５０は、機器２００が異常な状態である場合、少なくとも異常判定モデル１２が出力した異常に関する情報を入力して機器２００を正常な状態に移行させるための補正情報を生成する。
　また、補正情報生成部１５０は、機器２００が異常な状態である場合、少なくとも異常判定モデル１２が出力した異常に関する情報と、補正情報生成モデル１３とを用いて、当該補正情報を生成してもよい。この文章は、次のように表現してもよい。補正情報生成部１５０は、機器２００が異常な状態である場合、異常原因、異常が検出された波形データ、及び異常が発生している異常期間のうちの少なくとも１つである異常に関する情報と、補正情報生成モデル１３とを用いて、当該補正情報を生成する。なお、例えば、補正情報は、パラメータである。 Returning to FIG. 1, the correction information generation unit 150 will be explained.
When the device 200 is in an abnormal state, the correction information generation unit 150 receives at least the information regarding the abnormality output by the abnormality determination model 12 and generates correction information for transitioning the device 200 to a normal state.
Further, when the device 200 is in an abnormal state, the correction information generation unit 150 generates the correction information using at least the information regarding the abnormality output by the abnormality determination model 12 and the correction information generation model 13. good. This sentence can be expressed as follows. When the device 200 is in an abnormal state, the correction information generation unit 150 generates information regarding the abnormality, which is at least one of the cause of the abnormality, the waveform data in which the abnormality was detected, and the abnormal period during which the abnormality is occurring; The correction information is generated using the correction information generation model 13. Note that, for example, the correction information is a parameter.
　補正情報生成部１５０は、機器２００が異常な状態である場合、第１のセンサ情報以外のセンサ情報と、異常に関する情報と、補正情報生成モデル１３とを用いて、当該補正情報を生成してもよい。 When the device 200 is in an abnormal state, the correction information generation unit 150 generates the correction information using sensor information other than the first sensor information, information regarding the abnormality, and the correction information generation model 13. Good too.
　例えば、第１のセンサ情報以外のセンサ情報は、マスコンノッチのセンサ情報とする。異常期間は、“ｔ＝２４０～２６０”とする。異常が検出された波形データは、トルクの波形データとする。補正情報生成モデル１３は、“ｔ＝２４０～２６０”におけるマスコンノッチが“５”であるため、マスコンノッチの入力を受けてモータに電力を供給するインバータのパラメータを正常な状態に移行させるための補正情報を生成する。さらに、異常が検出された波形データは、速度の波形データとする。補正情報生成モデル１３は、異常期間でない“ｔ＝０～２４０”におけるマスコンノッチが“５”であり、“ｔ＝２４０～２６０”における速度が“５０～６０”であるため、速度を抑制するための補正情報を生成する。 For example, the sensor information other than the first sensor information is the sensor information of the mascon notch. The abnormal period is "t=240 to 260". The waveform data in which an abnormality is detected is torque waveform data. In the correction information generation model 13, since the mascon notch at "t=240 to 260" is "5", the parameter of the inverter that receives the input of the mascon notch and supplies power to the motor shifts to a normal state. Generate correction information. Further, the waveform data in which an abnormality has been detected is velocity waveform data. The correction information generation model 13 suppresses the speed because the mask notch is “5” in “t=0 to 240” which is not an abnormal period, and the speed in “t=240 to 260” is “50 to 60”. Generate correction information for
　補正情報生成モデル１３は、異常期間よりも後の区間のデータに基づいて、補正情報を生成してもよい。 The correction information generation model 13 may generate correction information based on data in a section after the abnormal period.
　出力部１６０は、補正情報を出力する。例えば、出力部１６０は、補正情報と、補正情報の設定指示とを機器２００に出力する。これにより、機器２００では、補正情報が設定される。そして、機器２００は、正常な状態に移行する。また、例えば、出力部１６０は、保守作業員が有する端末に補正情報を出力する。これにより、保守作業員は、補正情報を機器２００に設定することができる。そして、機器２００は、保守作業員の設定作業により、正常な状態に移行する。 The output unit 160 outputs correction information. For example, the output unit 160 outputs correction information and correction information setting instructions to the device 200. Accordingly, the correction information is set in the device 200. Then, the device 200 transitions to a normal state. Further, for example, the output unit 160 outputs the correction information to a terminal owned by a maintenance worker. This allows the maintenance worker to set the correction information in the device 200. Then, the device 200 transitions to a normal state through the setting work performed by the maintenance worker.
　補正情報生成部１５０と出力部１６０が実行する処理を、フローチャートを用いて説明する。
　図７は、実施の形態１の補正情報生成部と出力部が実行する処理の例を示すフローチャートである。
　（ステップＳ３１）補正情報生成部１５０は、異常原因、異常が検出された波形データ、異常が発生している異常期間、及び補正情報生成モデル１３を用いて、補正情報を生成する。詳細には、補正情報生成部１５０が、異常原因、異常が検出された波形データ、異常が発生している異常期間を補正情報生成モデル１３に入力することで、補正情報生成モデル１３は、当該補正情報を出力する。
　（ステップＳ３２）出力部１６０は、当該補正情報を出力する。 The processing executed by the correction information generation section 150 and the output section 160 will be explained using a flowchart.
FIG. 7 is a flowchart illustrating an example of processing executed by the correction information generation unit and output unit of the first embodiment.
(Step S31) The correction information generation unit 150 generates correction information using the cause of the abnormality, the waveform data in which the abnormality was detected, the abnormal period during which the abnormality has occurred, and the correction information generation model 13. Specifically, the correction information generation unit 150 inputs the cause of the abnormality, the waveform data in which the abnormality was detected, and the abnormal period in which the abnormality has occurred to the correction information generation model 13, so that the correction information generation model 13 Output correction information.
(Step S32) The output unit 160 outputs the correction information.
　実施の形態１によれば、補正装置１００は、機器２００が異常な状態である場合、補正情報を生成する。補正装置１００は、補正情報を出力する。これにより、機器２００は、正常な状態に移行する。よって、補正装置１００は、機器２００を正常な状態に移行させることができる。 According to the first embodiment, the correction device 100 generates correction information when the device 200 is in an abnormal state. Correction device 100 outputs correction information. Thereby, the device 200 transitions to a normal state. Therefore, the correction device 100 can shift the device 200 to a normal state.
実施の形態２．
　次に、実施の形態２を説明する。実施の形態２では、実施の形態１と相違する事項を主に説明する。そして、実施の形態２では、実施の形態１と共通する事項の説明を省略する。 Embodiment 2.
Next, a second embodiment will be described. In the second embodiment, matters that are different from the first embodiment will be mainly explained. In the second embodiment, explanations of matters common to the first embodiment will be omitted.
　実施の形態１では、停止中の機器２００に補正情報が出力される場合を説明した。実施の形態２では、動作中の機器２００に補正情報が出力される。このように、動作中の機器２００に補正情報が出力されるため、補正情報は、動作中の機器２００に設定することが可能な情報に限られる。また、当該情報は、変更が可能な範囲内のパラメータである。さらに、当該範囲は、設計情報に基づいて、決定されてもよい。 In the first embodiment, the case where correction information is output to the stopped device 200 has been described. In the second embodiment, correction information is output to the device 200 in operation. In this way, since the correction information is output to the device 200 in operation, the correction information is limited to information that can be set in the device 200 in operation. Further, the information is a parameter within a changeable range. Furthermore, the range may be determined based on design information.
　また、補正情報は、機器２００が動作中に設定変更が認められない情報を間接的に変更する情報を含む。例えば、機器２００が動作中に設定変更が認められない情報が電圧のパラメータである場合、補正情報は、抵抗のパラメータである。具体的には、電圧を上げる場合、補正情報は、低い値である、抵抗のパラメータである。当該抵抗のパラメータが機器２００に設定されることで、電流が上がり、そして、電圧が上がる。 Further, the correction information includes information that indirectly changes information whose settings are not allowed to be changed while the device 200 is in operation. For example, if the information whose settings cannot be changed while the device 200 is in operation is a voltage parameter, the correction information is a resistance parameter. Specifically, when increasing the voltage, the correction information is a parameter of the resistance, which is a low value. By setting the parameters of the resistance in the device 200, the current increases and the voltage increases.
　機器２００が動作中に設定変更が認められない情報を間接的に変更する情報は、１つのパラメータでもよいし、複数のパラメータでもよい。 The information that indirectly changes the information whose settings are not allowed to be changed while the device 200 is in operation may be one parameter or may be multiple parameters.
　図８は、実施の形態２の補正装置の機能を示すブロック図である。図１に示される構成と同じ図８の構成は、図１に示される符号と同じ符号を付している。
　補正装置１００ａは、補正情報生成部１５０ａを有する。 FIG. 8 is a block diagram showing the functions of the correction device according to the second embodiment. Components in FIG. 8 that are the same as those shown in FIG. 1 are designated by the same reference numerals as those shown in FIG.
The correction device 100a includes a correction information generation section 150a.
　取得部１１０は、複数のセンサ情報を取得する。なお、複数のセンサ情報は、リアルタイムの情報であると表現してもよい。
　補正情報生成部１５０ａは、機器２００が異常な状態である場合、少なくとも異常判定モデル１２が出力した異常に関する情報と、補正情報生成モデル１３とを用いて、動作中の機器２００を正常な状態に移行させるための補正情報を生成する。 The acquisition unit 110 acquires a plurality of sensor information. Note that the plurality of sensor information may be expressed as real-time information.
When the device 200 is in an abnormal state, the correction information generation unit 150a returns the operating device 200 to a normal state using at least the information regarding the abnormality output by the abnormality determination model 12 and the correction information generation model 13. Generate correction information for migration.
　実施の形態２によれば、補正装置１００ａは、動作中の機器２００を正常な状態に移行させるための補正情報を生成する。補正装置１００ａは、補正情報を出力する。これにより、動作中の機器２００は、正常な状態に移行する。よって、補正装置１００ａは、動作中の機器２００を正常な状態に移行させることができる。 According to the second embodiment, the correction device 100a generates correction information for shifting the operating device 200 to a normal state. The correction device 100a outputs correction information. As a result, the operating device 200 transitions to a normal state. Therefore, the correction device 100a can shift the operating device 200 to a normal state.
実施の形態３．
　次に、実施の形態３を説明する。実施の形態３では、実施の形態１と相違する事項を主に説明する。そして、実施の形態３では、実施の形態１と共通する事項の説明を省略する。 Embodiment 3.
Next, Embodiment 3 will be described. In the third embodiment, matters that are different from the first embodiment will be mainly explained. In the third embodiment, explanations of matters common to the first embodiment will be omitted.
　図９は、実施の形態３の補正装置の機能を示すブロック図である。図１に示される構成と同じ図９の構成は、図１に示される符号と同じ符号を付している。
　補正装置１００は、さらに再学習部１７０を有する。
　再学習部１７０の一部又は全部は、処理回路によって実現してもよい。また、再学習部１７０の一部又は全部は、プロセッサ１０１が実行するプログラムのモジュールとして実現してもよい。 FIG. 9 is a block diagram showing the functions of the correction device according to the third embodiment. Components in FIG. 9 that are the same as those shown in FIG. 1 are designated by the same reference numerals as those shown in FIG.
The correction device 100 further includes a relearning section 170.
Part or all of the relearning unit 170 may be realized by a processing circuit. Further, part or all of the relearning unit 170 may be realized as a module of a program executed by the processor 101.
　取得部１１０は、補正結果を取得する。例えば、取得部１１０は、補正結果を機器２００から取得する。また、例えば、取得部１１０は、補正結果を外部装置から取得する。
　再学習部１７０は、補正結果が異常であることを示している場合、異常判定モデル１２及び補正情報生成モデル１３を再学習する。また、再学習部１７０は、異常を示す補正結果が閾値以上取得された場合、異常判定モデル１２及び補正情報生成モデル１３を再学習してもよい。以下、再学習を詳細に説明する。 The acquisition unit 110 acquires the correction results. For example, the acquisition unit 110 acquires the correction result from the device 200. Further, for example, the acquisition unit 110 acquires the correction result from an external device.
The relearning unit 170 relearns the abnormality determination model 12 and the correction information generation model 13 when the correction result indicates an abnormality. Further, the relearning unit 170 may relearn the abnormality determination model 12 and the correction information generation model 13 when a correction result indicating an abnormality is obtained equal to or greater than a threshold value. Relearning will be explained in detail below.
　異常判定モデル１２が再学習される場合、再学習部１７０は、正常データ生成部１２０が生成した波形データと、波形データ生成部１３０が生成した波形データとの比較結果を、学習データとして用いる。 When the abnormality determination model 12 is retrained, the relearning unit 170 uses the comparison result between the waveform data generated by the normal data generation unit 120 and the waveform data generated by the waveform data generation unit 130 as learning data.
　補正情報生成モデル１３が再学習される場合、再学習部１７０は、補正情報を生成する際に用いられた異常に関する情報（例えば、異常原因など）を、学習データとして用いる。また、当該学習データには、“異常”のラベルが付されてもよい。 When the correction information generation model 13 is retrained, the relearning unit 170 uses information regarding the abnormality (for example, the cause of the abnormality, etc.) used when generating the correction information as learning data. Further, the learning data may be labeled as "abnormal".
　再学習では、以前用いられた学習データの一部又は全部が、用いられてもよい。また、以前用いられた学習データの中で類似するデータは、除外されてもよい。
　以前用いられた学習データの一部又は全部が、用いられる場合、再学習では、前回の学習データの数と同じ数の学習データが、用いられてもよい。 In relearning, part or all of previously used learning data may be used. Further, similar data among previously used learning data may be excluded.
When part or all of previously used learning data is used, the same number of learning data as the previous learning data may be used in relearning.
　実施の形態３によれば、異常判定モデル１２及び補正情報生成モデル１３が再学習されることで、補正装置１００は、機器２００を正常な状態に移行させるための補正情報を生成できるようになる。 According to the third embodiment, by relearning the abnormality determination model 12 and the correction information generation model 13, the correction device 100 can generate correction information for transitioning the device 200 to a normal state. .
　以上に説明した各実施の形態における特徴は、互いに適宜組み合わせることができる。 The features of each embodiment described above can be combined with each other as appropriate.
　１１　正常生成情報、　１２　異常判定モデル、　１３　補正情報生成モデル、　１００，１００ａ　補正装置、　１０１　プロセッサ、　１０２　揮発性記憶装置、　１０３　不揮発性記憶装置、　１１０　取得部、　１２０　正常データ生成部、　１３０　波形データ生成部、　１４０　異常判定部、　１５０，１５０ａ　補正情報生成部、　１６０　出力部、　１７０　再学習部、　２００　機器。 11 normal generation information, 12 abnormality determination model, 13 correction information generation model, 100, 100a correction device, 101 processor, 102 volatile storage device, 103 non-volatile storage device, 110 acquisition unit, 120 normal data generation unit, 130 Waveform data Generation unit, 140 abnormality determination unit, 150, 150a correction information generation unit, 160 output unit, 170 relearning unit, 200 equipment.

Claims (6)

  1. 　１以上のセンサが動作中の機器を計測することで得られる、前記１以上のセンサと対応した第１のセンサ情報を含む複数のセンサ情報、及び前記第１のセンサ情報の正常なデータの生成に用いられる正常生成情報を取得する取得部と、
    　前記複数のセンサ情報のうち、前記第１のセンサ情報以外のセンサ情報と、前記正常生成情報とを用いて、前記第１のセンサ情報の正常な時系列データを生成し、前記第１のセンサ情報の正常な時系列データに基づいて、第１の波形データを生成する正常データ生成部と、
    　前記複数のセンサ情報のうちの前記第１のセンサ情報の時系列データに基づいて、第２の波形データを生成する波形データ生成部と、
    　前記第１の波形データと前記第２の波形データとの比較結果を入力して前記機器の異常に関する情報を出力することが可能な異常判定モデルを用いて、前記機器が異常な状態であるか否かを判定する異常判定部と、
    　前記機器が異常な状態である場合、少なくとも前記異常判定モデルが出力した異常に関する情報を入力して前記機器を正常な状態に移行させるための補正情報を生成する補正情報生成部と、
    　前記補正情報を出力する出力部と、
    　を有する補正装置。     Generation of a plurality of sensor information including first sensor information corresponding to the one or more sensors, which is obtained by measuring a device in which one or more sensors are operating, and normal data of the first sensor information. an acquisition unit that acquires normal generation information used for
    Among the plurality of sensor information, sensor information other than the first sensor information and the normal generation information are used to generate normal time series data of the first sensor information, and the first sensor information is a normal data generation unit that generates first waveform data based on normal time series data of the information;
    a waveform data generation unit that generates second waveform data based on time-series data of the first sensor information among the plurality of sensor information;
    Using an abnormality determination model that is capable of inputting a comparison result between the first waveform data and the second waveform data and outputting information regarding an abnormality of the device, determine whether the device is in an abnormal state. an abnormality determination unit that determines whether or not the
    a correction information generation unit that inputs at least information regarding the abnormality output by the abnormality determination model and generates correction information for transitioning the device to a normal state when the device is in an abnormal state;
    an output unit that outputs the correction information;
    A correction device having.
  2. 　前記補正情報生成部は、前記機器が異常な状態である場合、少なくとも前記異常判定モデルが出力した異常に関する情報を用いて、動作中の前記機器を正常な状態に移行させるための補正情報を生成する、
    　請求項１に記載の補正装置。     When the device is in an abnormal state, the correction information generation unit generates correction information for transitioning the operating device to a normal state, using at least information regarding the abnormality output by the abnormality determination model. do,
    A correction device according to claim 1.
  3. 　前記補正情報は、前記機器が動作中に設定変更が認められない情報を間接的に変更する情報を含む、
    　請求項２に記載の補正装置。     The correction information includes information that indirectly changes information whose settings are not allowed to be changed while the device is in operation.
    A correction device according to claim 2.
  4. 　再学習部をさらに有し、
    　前記取得部は、前記異常判定モデル、補正情報生成モデル、及び補正結果を取得し、
    　前記再学習部は、前記補正結果が異常であることを示している場合、前記異常判定モデル及び前記補正情報生成モデルを再学習する、
    　請求項１から３のいずれか１項に記載の補正装置。     It also has a relearning department,
    The acquisition unit acquires the abnormality determination model, the correction information generation model, and the correction result,
    The relearning unit relearns the abnormality determination model and the correction information generation model when the correction result indicates that the correction result is abnormal.
    A correction device according to any one of claims 1 to 3.
  5. 　補正装置が、
    　１以上のセンサが動作中の機器を計測することで得られる、前記１以上のセンサと対応した第１のセンサ情報を含む複数のセンサ情報、及び前記第１のセンサ情報の正常なデータの生成に用いられる正常生成情報を取得し、前記複数のセンサ情報のうち、前記第１のセンサ情報以外のセンサ情報と、前記正常生成情報とを用いて、前記第１のセンサ情報の正常な時系列データを生成し、前記第１のセンサ情報の正常な時系列データに基づいて、第１の波形データを生成し、前記複数のセンサ情報のうちの前記第１のセンサ情報の時系列データに基づいて、第２の波形データを生成し、
    　前記第１の波形データと前記第２の波形データとの比較結果を入力して前記機器の異常に関する情報を出力することが可能な異常判定モデルを用いて、前記機器が異常な状態であるか否かを判定し、
    　前記機器が異常な状態である場合、少なくとも前記異常判定モデルが出力した異常に関する情報を入力して前記機器を正常な状態に移行させるための補正情報を生成し、
    　前記補正情報を出力する、
    　処理方法。     The correction device
    Generation of a plurality of sensor information including first sensor information corresponding to the one or more sensors, which is obtained by measuring a device in which one or more sensors are operating, and normal data of the first sensor information. The normal generation information used for the first sensor information is acquired, and the normal time series of the first sensor information is obtained using sensor information other than the first sensor information among the plurality of sensor information and the normal generation information. generate data, generate first waveform data based on normal time-series data of the first sensor information, and generate first waveform data based on the time-series data of the first sensor information among the plurality of sensor information; to generate second waveform data,
    Using an abnormality determination model that is capable of inputting a comparison result between the first waveform data and the second waveform data and outputting information regarding an abnormality of the device, determine whether the device is in an abnormal state. Determine whether or not
    when the device is in an abnormal state, inputting at least information regarding the abnormality output by the abnormality determination model to generate correction information for transitioning the device to a normal state;
    outputting the correction information;
    Processing method.
  6. 　補正装置に、
    　１以上のセンサが動作中の機器を計測することで得られる、前記１以上のセンサと対応した第１のセンサ情報を含む複数のセンサ情報、及び前記第１のセンサ情報の正常なデータの生成に用いられる正常生成情報を取得し、前記複数のセンサ情報のうち、前記第１のセンサ情報以外のセンサ情報と、前記正常生成情報とを用いて、前記第１のセンサ情報の正常な時系列データを生成し、前記第１のセンサ情報の正常な時系列データに基づいて、第１の波形データを生成し、前記複数のセンサ情報のうちの前記第１のセンサ情報の時系列データに基づいて、第２の波形データを生成し、
    　前記第１の波形データと前記第２の波形データとの比較結果を入力して前記機器の異常に関する情報を出力することが可能な異常判定モデルを用いて、前記機器が異常な状態であるか否かを判定し、
    　前記機器が異常な状態である場合、少なくとも前記異常判定モデルが出力した異常に関する情報を入力して前記機器を正常な状態に移行させるための補正情報を生成し、
    　前記補正情報を出力する、
    　処理を実行させる処理プログラム。
          In the correction device,
    Generation of a plurality of sensor information including first sensor information corresponding to the one or more sensors, which is obtained by measuring a device in which one or more sensors are operating, and normal data of the first sensor information. The normal generation information used for the first sensor information is acquired, and the normal time series of the first sensor information is obtained using sensor information other than the first sensor information among the plurality of sensor information and the normal generation information. generate data, generate first waveform data based on normal time-series data of the first sensor information, and generate first waveform data based on the time-series data of the first sensor information among the plurality of sensor information; to generate second waveform data,
    Using an abnormality determination model that is capable of inputting a comparison result between the first waveform data and the second waveform data and outputting information regarding an abnormality of the device, determine whether the device is in an abnormal state. Determine whether or not
    when the device is in an abnormal state, inputting at least information regarding the abnormality output by the abnormality determination model to generate correction information for transitioning the device to a normal state;
    outputting the correction information;
    A processing program that executes processing.
PCT/JP2022/019918 2022-05-11 2022-05-11 Correction equipment, processing method, and processing program WO2023218550A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
JP2022562766A JPWO2023218550A1 (en) 2022-05-11 2022-05-11
PCT/JP2022/019918 WO2023218550A1 (en) 2022-05-11 2022-05-11 Correction equipment, processing method, and processing program

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2022/019918 WO2023218550A1 (en) 2022-05-11 2022-05-11 Correction equipment, processing method, and processing program

Publications (1)

Publication Number Publication Date
WO2023218550A1 true WO2023218550A1 (en) 2023-11-16

Family

ID=88729985

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2022/019918 WO2023218550A1 (en) 2022-05-11 2022-05-11 Correction equipment, processing method, and processing program

Country Status (2)

Country Link
JP (1) JPWO2023218550A1 (en)
WO (1) WO2023218550A1 (en)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016195092A1 (en) * 2015-06-05 2016-12-08 株式会社日立製作所 Anomaly sensing device
WO2019207767A1 (en) * 2018-04-27 2019-10-31 株式会社日立製作所 Control device and control method
WO2022054256A1 (en) * 2020-09-11 2022-03-17 三菱電機株式会社 Abnormality detection device

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016195092A1 (en) * 2015-06-05 2016-12-08 株式会社日立製作所 Anomaly sensing device
WO2019207767A1 (en) * 2018-04-27 2019-10-31 株式会社日立製作所 Control device and control method
WO2022054256A1 (en) * 2020-09-11 2022-03-17 三菱電機株式会社 Abnormality detection device

Also Published As

Publication number Publication date
JPWO2023218550A1 (en) 2023-11-16

Similar Documents

Publication Publication Date Title
JPWO2019142331A1 (en) Failure prediction system and failure prediction method
US9915925B2 (en) Initiated test health management system and method
JP2022552424A (en) Methods and apparatus for detecting faults, methods and apparatus for training models, and devices and storage media thereof
JPWO2019106875A1 (en) Abnormality determination system, motor control device, and abnormality determination device
Sedghi et al. Fault detection of a seven level modular multilevel inverter via voltage histogram and neural network
CN111684475B (en) Learning support device, processing system, learning support method, and storage medium
CN111008732B (en) Stack model-based fault prediction method and system
JP2006250818A (en) Diagnosis system for electrical installation
Wang et al. Condition monitoring of wind turbine based on incremental learning and multivariate state estimation technique
CN115453356A (en) Power equipment running state monitoring and analyzing method, system, terminal and medium
JP2001206296A (en) Method and device for generating maintenance message
WO2023218550A1 (en) Correction equipment, processing method, and processing program
Danesh et al. Out-of-Distribution Dynamics Detection: RL-Relevant Benchmarks and Results
JP2018025936A (en) Machine tool
CN114060228B (en) Method and device for detecting abnormal torque jump of wind generating set and controller
Salfner Predicting failures with hidden Markov models
CN110163389B (en) Neural network method applied to predictive maintenance management
CN103065060A (en) Computing method for transient stability limit of transmission section of power system
Tian et al. Health condition prognostics of gears using a recurrent neural network approach
CN113377595B (en) Fault diagnosis method, device, electronic equipment and storage medium
RU2626338C1 (en) Adaptive digital smoothing and predictive device
Radaideh et al. Early fault detection in particle accelerator power electronics using ensemble learning
JP2020187616A (en) Plant monitoring model creation device, plant monitoring model creation method, and plant monitoring model creation program
Momtaz et al. On-line diagnosis and compensation for parametric failures in linear state variable circuits and systems using time-domain checksum observers
WO2021207369A1 (en) Sensor attribution for anomaly detection

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2022562766

Country of ref document: JP

Kind code of ref document: A

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22941631

Country of ref document: EP

Kind code of ref document: A1