WO2023032201A1 - 健全度診断装置および健全度診断方法 - Google Patents
健全度診断装置および健全度診断方法 Download PDFInfo
- Publication number
- WO2023032201A1 WO2023032201A1 PCT/JP2021/032651 JP2021032651W WO2023032201A1 WO 2023032201 A1 WO2023032201 A1 WO 2023032201A1 JP 2021032651 W JP2021032651 W JP 2021032651W WO 2023032201 A1 WO2023032201 A1 WO 2023032201A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- feature amount
- data
- soundness
- unit
- amount data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
- B60L3/00—Electric devices on electrically-propelled vehicles for safety purposes; Monitoring operating variables, e.g. speed, deceleration or energy consumption
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C5/00—Registering or indicating the working of vehicles
- G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
- G07C5/0808—Diagnosing performance data
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B60—VEHICLES IN GENERAL
- B60L—PROPULSION OF ELECTRICALLY-PROPELLED VEHICLES; SUPPLYING ELECTRIC POWER FOR AUXILIARY EQUIPMENT OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRODYNAMIC BRAKE SYSTEMS FOR VEHICLES IN GENERAL; MAGNETIC SUSPENSION OR LEVITATION FOR VEHICLES; MONITORING OPERATING VARIABLES OF ELECTRICALLY-PROPELLED VEHICLES; ELECTRIC SAFETY DEVICES FOR ELECTRICALLY-PROPELLED VEHICLES
- B60L3/00—Electric devices on electrically-propelled vehicles for safety purposes; Monitoring operating variables, e.g. speed, deceleration or energy consumption
- B60L3/0023—Detecting, eliminating, remedying or compensating for drive train abnormalities, e.g. failures within the drive train
Definitions
- the present disclosure relates to a soundness diagnosis device and a soundness diagnosis method for diagnosing the soundness of equipment.
- Patent Literature 1 discloses a technique for accurately diagnosing an abnormality of a brake in consideration of various conditions such as a running section in the case where the diagnosis target is a brake mounted on a railroad vehicle.
- the device to be diagnosed is a device operated by a person, such as a brake mounted on a railroad vehicle
- the data detected by the sensor varies among individuals depending on the characteristics of the person operating the device. easier to get out. Therefore, even if the soundness of equipment is diagnosed using data that includes individual differences, there is a problem that the accuracy of the diagnosis decreases.
- the present disclosure has been made in view of the above, and an object thereof is to obtain a soundness diagnosis device capable of suppressing deterioration in accuracy when diagnosing the soundness of equipment.
- the present disclosure is a soundness diagnosis device for diagnosing the soundness of equipment.
- the soundness diagnosis device has a data loading unit that acquires the operation data of the device during the diagnosis target period, and based on the physical characteristics of the device, extracts the data part to be the target of the feature amount data from the operation data as sample data, and extracts the sample data
- a feature data generation unit that generates feature data using
- an inference unit that performs soundness diagnosis on the feature data that is sequentially generated using a trained model that has undergone model learning of the normal state of the device.
- a visualization unit that visualizes the transition of the soundness diagnosis result obtained by the inference unit.
- the soundness diagnostic device of the present disclosure has the effect of suppressing a decrease in accuracy when diagnosing the soundness of equipment.
- FIG. 1 is a first diagram showing a configuration example of a soundness diagnosis device according to an embodiment
- Flowchart showing the operation of the soundness diagnosis device according to the embodiment 4 is a flow chart showing the operation of generating feature data by the feature data generator of the soundness diagnostic device according to the embodiment
- FIG. 4 is a diagram showing an example of a configuration of a processing circuit provided in the soundness diagnosis device according to the embodiment when the processing circuit is realized by a processor and a memory
- FIG. 4 is a diagram showing an example of a configuration of a processing circuit provided in the soundness diagnosis device according to the embodiment when the processing circuit is configured by dedicated hardware
- FIG. 1 is a first diagram showing a configuration example of a soundness diagnosis device 20 according to this embodiment.
- the soundness diagnosis device 20 is a device that diagnoses the soundness of the equipment mounted on the railway vehicle 10 using the operation data 31 on the equipment mounted on the railway vehicle 10 .
- the operation data 31 is data indicating the operation state of equipment detected by sensors (not shown) mounted on the railcar 10 .
- the brake system 11 is a system that includes a brake cylinder (not shown) and controls braking force by air pressure.
- the soundness diagnosis device 20 includes a condition setting unit 21, a data loading unit 22, a feature data generation unit 23, a learning unit 24, an inference unit 25, and a visualization unit 26. Prepare.
- the condition setting section 21 and the visualization section 26 are included in an operation UI (User Interface) 27 .
- FIG. 2 is a flow chart showing the operation of the soundness diagnosis device 20 according to this embodiment.
- the condition setting unit 21 receives various condition settings for the data loading unit 22 and the feature amount data generation unit 23 from the user 40 of the soundness diagnosis device 20, and sets conditions for the data loading unit 22 and the feature amount data generation unit 23. Settings are made (step S11). Specifically, the condition setting unit 21 receives from the user 40 a diagnosis target period for which the data loading unit 22 acquires the operation data 31 . Further, the condition setting unit 21 receives from the user 40 the setting of various conditions when the feature amount data generation unit 23 generates the feature amount data 32 based on the physical characteristics of the brake system 11 .
- the condition setting unit 21 is, for example, an interface such as a mouse and a keyboard. When the condition setting unit 21 and the visualization unit 26 are integrated in the operation UI 27, the condition setting unit 21 may be a touch panel or the like.
- the data loading unit 22 acquires the operation data 31 of the brake system 11 during the diagnostic period (step S12).
- the data loading unit 22 acquires the operation data 31 of the diagnosis target period from the operation data 31 of the entire period detected by the railroad vehicle 10, but is not limited to this.
- the data loading unit 22 may acquire the operation data 31 for the entire period output from the railroad vehicle 10 and extract the operation data 31 for the diagnosis target period from the operation data 31 for the entire period.
- the operation data 31 for the entire period detected by the railcar 10 may be stored in the storage unit, and the data loading unit 22 may read the operation data 31 from the storage unit.
- the data loading unit 22 outputs the operation data 31 of the brake system 11 during the diagnosis target period to the feature amount data generating unit 23 .
- the feature amount data generation unit 23 cuts out the target data portion of the feature amount data 32 from the operation data 31 as sample data based on the physical characteristics of the equipment, that is, the brake system 11, and generates the feature amount data using the cut out sample data. 32 is generated (step S13).
- the operation of the feature amount data generation unit 23 to generate the feature amount data 32 will be described in detail.
- FIG. 3 is a flow chart showing the operation of generating the feature data 32 by the feature data generator 23 of the soundness diagnostic device 20 according to the present embodiment.
- the feature amount data generation unit 23 detects the brake release timing when the railroad vehicle 10 departs from the brake cylinder pressure data included in the operation data 31. (Step S21).
- the feature amount data 32 should always show the same behavior and the tendency should change little by little with changes in soundness, that is, progress of deterioration.
- the operation data 31 of the brake system 11 is always acquired while the railroad vehicle 10 is running, and the value varies depending on the complicated control during running. I need to cut out the quantity.
- the brake system 11 of the railway vehicle 10 is mainly used during deceleration, stopping, and the like. However, the brake system 11 tends to vary depending on the characteristics of the railway vehicle 10 driver, that is, individual differences.
- the data of the part where the brake cylinder pressure decreases as the brakes are released is cut out and used as the feature amount data 32.
- the portion where the brake cylinder pressure drops as the brake is released is the BC (Brake Cylinder) pressure drop portion where the air in the brake cylinder is released. Since the BC pressure fall has little individual difference among drivers and always behaves the same as a physical characteristic of the brake system 11, it meets the requirements of the feature amount data 32 having the characteristic of deterioration progression described above.
- the feature amount data generation unit 23 extracts sample data from the brake cylinder pressure data included in the operation data 31 for a specified time-series range including the brake release timing when the railroad vehicle 10 departs, that is, the BC pressure drop point. (step S22).
- the feature amount data generator 23 may cut out the sample data with the BC pressure falling time as a starting point, or may cut out the sample data with a starting point at a specified past time after the BC pressure falling time. .
- the feature data generation unit 23 cleanses the extracted sample data (step S23).
- the feature amount data generation unit 23 can remove elements that become noise when calculating the feature amount data 32 by removing irregular sample data that is clearly different from aging.
- the feature data generation unit 23 performs filtering to extract sample data that matches the set conditions from the cleansed sample data (step S24).
- the feature amount data generating unit 23 narrows down the environmental conditions during running of the railway vehicle 10 so that the environmental conditions during running of the railway vehicle 10 are unified, that is, the analysis conditions for the soundness diagnosis are unified.
- the feature amount data generation unit 23 performs filtering based on the position of the brake notch at the time of stop, the boarding rate of the railway vehicle 10, the brake loosening speed, and the like. Note that it is also assumed that the tendency of the occupancy rate differs depending on the route on which the railway vehicle 10 is used. Therefore, the feature amount data generation unit 23 may change the filtering conditions depending on the route on which the railroad vehicle 10 is used.
- the feature amount data generation unit 23 may perform filtering so as to classify the analysis conditions into a plurality of conditions instead of narrowing down to one condition so as to unify the analysis conditions in the soundness diagnosis. For example, when the feature amount data generating unit 23 performs filtering based on the boarding rate of the railway vehicle 10, classification is performed into groups such as boarding rate of less than 30%, boarding rate of 30% or more and less than 70%, boarding rate of 70% or more, and the like. Filtering may be performed to
- the feature data generation unit 23 processes the filtered sample data to generate feature data 32 (step S25).
- the feature amount data generation unit 23 processes the sample data after filtering based on domain knowledge, that is, the physical characteristics of the brake system 11, which is the equipment to be diagnosed. For example, the feature amount data generation unit 23 may obtain a temporary difference by one-time differentiation, may obtain a cumulative sum by one-time integration, or may combine a plurality of feature amounts.
- the feature amount data generation unit 23 outputs the generated feature amount data 32 to the learning unit 24 and the inference unit 25 .
- the feature amount data generation unit 23 may store the generated feature amount data 32 in the storage unit. In this case, the learning unit 24 and the inference unit 25 read the feature amount data 32 from the storage unit.
- the feature amount data generation unit 23 determines the timing at which the brakes are released when the railroad vehicle 10 departs from the brake cylinder pressure included in the operation data 31, based on the vehicle speed information and the brake notch information of the railroad vehicle 10. Cut out the range including as sample data. It can also be said that the feature amount data generator 23 extracts a range including the timing at which the brakes of the railway vehicle 10 are released from the operation data 31 as sample data. The feature amount data generation unit 23 cleans the extracted sample data, and performs filtering so that the environmental conditions during running of the railroad vehicle 10 are unified. The feature data generator 23 processes the filtered sample data based on the physical characteristics of the brake system 11 to generate feature data 32 .
- the learning unit 24 performs model learning of the normal state using the feature amount data 32 acquired in the normal state among the feature amount data 32 (step S14).
- the learning unit 24 obtains a learned model 33 as a result of the model learning.
- the feature amount data 32 acquired during normal operation may be based on data obtained during testing at the factory where the railway vehicle 10 was manufactured, or may be based on data that has been specified after the operation of the railway vehicle 10 is started. It may be based on data obtained early in the introduction of the period.
- the learning unit 24 performs model learning of the normal state using machine learning. For example, an outlier detection technique can be used for the machine learning model.
- outlier detection techniques include algorithms such as OCSVM (One Class Support Vector Machine), but are not limited to these.
- OCSVM One Class Support Vector Machine
- a technique other than the outlier detection technique, such as deep learning may be used.
- the learning unit 24 outputs the learned model 33 obtained as a result of model learning to the inference unit 25 .
- the learning unit 24 may store the learned model 33 in the storage unit.
- the inference unit 25 reads the learned model 33 from the storage unit.
- the inference unit 25 uses the learned model 33 obtained by model learning of the normal state of the device, that is, the learned model 33 obtained as a result of the model learning in the learning unit 24, to the feature amount data 32 that is sequentially generated.
- a soundness diagnosis is performed (step S15).
- the inference unit 25 may perform a soundness diagnosis on the feature amount data 32 that is sequentially generated at predetermined intervals, for example, once a week. You may perform a soundness diagnosis at timing.
- the inference unit 25 calculates the degree of deviation from the normal state as a score for the feature amount data 32 that is sequentially generated, and normalizes the scored degree of deviation from the normal state.
- the obtained result is defined as the soundness diagnosis result 34 .
- the inference unit 25 calculates the distance from the boundary plane of the learned model 33 for each feature amount data 32 and normalizes it.
- the inference unit 25 outputs the soundness diagnosis result 34 obtained as a result of the soundness diagnosis to the visualization unit 26 .
- the inference unit 25 may store the soundness diagnosis result 34 in the storage unit.
- the visualization unit 26 reads the soundness diagnosis result 34 from the storage unit.
- the visualization unit 26 visualizes transitions in the soundness diagnosis results 34 obtained by the inference unit 25 (step S16).
- the visualization unit 26 utilizes, for example, a scatter diagram, a line graph, etc., for the soundness diagnosis results 34 based on the plurality of feature amount data 32, and plots the soundness diagnosis results 34 from the past to the present in time series to visualize the transition trend. visualize.
- the soundness diagnostic device 20 is assumed to be used by the user 40, such as an engineer of an equipment manufacturer who manufactures equipment mounted on the railroad vehicle 10, such as the brake system 11 in the example of FIG. .
- the equipment manufacturer user 40 estimates the future soundness trend from the transition trend of the soundness diagnosis result 34 from the past to the present, and provides the railway operator with maintenance, equipment replacement, etc. of the brake system 11 at the optimum timing. Recommend.
- the user 40 may set a plurality of thresholds for the soundness diagnosis result 34 to determine the timing of maintenance of the brake system 11, the timing of device replacement, and the like.
- the user 40 can statistically analyze the health degree diagnosis results 34 from a microscopic point of view based on each sample point and the health degree diagnosis results 34 from a macro point of view based on the periodic health degree diagnosis results 34 such as monthly. Check and judge the results. For example, when maintenance of the railway vehicle 10 is performed by the railway operator, it is assumed that the tendency of the feature amount data 32 obtained by the health degree diagnosis device 20 will differ before and after the maintenance. In general, it is considered that the feature amount data 32 tend to improve when the railcar 10 is maintained by the railroad company.
- the user 40 confirms the soundness diagnosis result 34 by assuming that maintenance or the like has been performed by the railway operator when the tendency of the feature amount data 32 changes after a certain point. When the maintenance information of the railway vehicle 10 is obtained from the railway operator, the user 40 confirms the soundness diagnosis result 34 based on the maintenance information.
- the soundness diagnostic device 20 may be installed at the equipment manufacturer described above. It may be installed in an operating railroad operator, railcar 10, or the like. If the operation UI 27 is installed by the device manufacturer and the rest is installed in the railway operator operating the railway vehicle 10 or the railway vehicle 10, the operation UI 27 may be a terminal device such as a tablet. .
- the soundness diagnostic device 20 can remotely diagnose the soundness of the brake system 11 even if it is installed in a different place from the brake system 11 to be diagnosed.
- the user 40 of the soundness diagnosis device 20 can use the condition setting unit 21 to load the data load unit 22 by using the operation UI 27 regardless of where the parts other than the operation UI 27 of the soundness diagnosis device 20 are installed. And it is possible to set conditions for the feature amount data generation unit 23 and check the soundness diagnosis result 34 visualized by the visualization unit 26 .
- the condition setting unit 21 receives various condition settings for the data loading unit 22 and the feature amount data generation unit 23 from the user 40, and sets conditions for the data loading unit 22 and the feature amount data generation unit 23. However, it is not limited to this.
- the condition setting unit 21 also accepts condition settings from the user 40 regarding the learning method in the learning unit 24 and the soundness diagnosis timing in the inference unit 25, and sets conditions for the learning unit 24 and the inference unit 25. good.
- the condition setting section 21 is an interface that receives operations from the user 40 .
- the part that displays the soundness diagnosis result 34 is a display such as an LCD (Liquid Crystal Display).
- the processing circuit may be a memory that stores a program and a processor that executes the program stored in the memory, or may be dedicated hardware. Processing circuitry is also called control circuitry.
- FIG. 4 is a diagram showing an example of the configuration of the processing circuit 90 when the processing circuit included in the soundness diagnostic device 20 according to the present embodiment is realized by the processor 91 and the memory 92.
- a processing circuit 90 shown in FIG. 4 is a control circuit and includes a processor 91 and a memory 92 .
- each function of the processing circuit 90 is implemented by software, firmware, or a combination of software and firmware.
- Software or firmware is written as a program and stored in memory 92 .
- each function is realized by the processor 91 reading and executing the program stored in the memory 92.
- the processing circuit 90 has a memory 92 for storing a program that results in the execution of the processing of the soundness diagnosis device 20 .
- This program can also be said to be a program for causing the soundness diagnosis device 20 to execute each function realized by the processing circuit 90 .
- This program may be provided by a storage medium storing the program, or may be provided by other means such as a communication medium.
- the program includes an acquisition step in which the data loading unit 22 acquires the operation data 31 of the device during the diagnosis target period, and the feature amount data generation unit 23 extracts the feature amount data 32 from the operation data 31 based on the physical characteristics of the device.
- a visualization step in which the visualization unit 26 visualizes the transition of the health degree diagnosis result 34 obtained by the inference unit 25 .
- the processor 91 is, for example, a CPU (Central Processing Unit), a processing device, an arithmetic device, a microprocessor, a microcomputer, or a DSP (Digital Signal Processor).
- the memory 92 is a non-volatile or volatile memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable ROM), EEPROM (registered trademark) (Electrically EPROM), etc.
- RAM Random Access Memory
- ROM Read Only Memory
- flash memory EPROM (Erasable Programmable ROM), EEPROM (registered trademark) (Electrically EPROM), etc.
- a semiconductor memory, a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, a DVD (Digital Versatile Disc), a Blu-ray disc, or a HDD (Hard Disk Drive) is applicable.
- FIG. 5 is a diagram showing an example of the configuration of the processing circuit 93 when the processing circuit included in the soundness diagnostic device 20 according to the present embodiment is configured with dedicated hardware.
- the processing circuit 93 shown in FIG. 5 is, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof. thing applies.
- the processing circuit 93 may be partially realized by dedicated hardware and partially realized by software or firmware.
- the processing circuitry 93 can implement each of the functions described above by dedicated hardware, software, firmware, or a combination thereof.
- the soundness diagnostic device 20 controls the brake system 11, which is a device mounted on the railroad vehicle 10, at the time when the railroad vehicle 10 departs, that is, when the brakes are released.
- the data of the BC pressure falling portion where the brake cylinder pressure drops is extracted from the operation data 31, and cleansing, filtering, etc. are performed to generate the feature amount data 32.
- the soundness diagnosis device 20 uses the feature amount data 32 to learn normal data through machine learning, and diagnoses the soundness of the brake system 11 based on the degree of deviation from the normal state.
- the soundness diagnosis device 20 suppresses the influence of individual differences among drivers even for the equipment that tends to cause individual differences depending on the driver of the railway vehicle 10, and prevents a decrease in accuracy when diagnosing the soundness of the equipment. can be suppressed.
- the soundness diagnosis device 20 can obtain the feature amount data 32 based on the already obtainable operation data 31, the soundness diagnosis can be performed without additionally introducing a special sensor, detection device, or the like. be able to.
- a user 40 of the soundness diagnostic device 20 can remotely confirm the current soundness of the brake system 11 without inspecting the actual brake system 11 of the railcar 10 on site, and judge the timing of maintenance and equipment replacement. be able to.
- the device whose health degree is to be diagnosed by the health degree diagnosis device 20 is the brake system 11. is not limited to the braking system 11 .
- the portion of the BC pressure drop when the railroad vehicle 10 departs was cut out. If there is a timing of data that has a small number of physical characteristics and similar behavior is detected, the soundness diagnostic device 20 can be applied.
- FIG. 6 is a second diagram showing a configuration example of the soundness diagnosis device 20A according to the present embodiment.
- model learning for a normal state is performed in advance using the feature amount data 32 acquired during normal operation among the feature amount data 32
- a learned model 33 obtained as a result of model learning is stored in a storage unit inside the soundness diagnosis device 20A.
- the inference unit 25 reads a learned model 33 obtained by performing model learning on the normal state of the device from the storage unit, and uses the read-out learned model 33 to diagnose the soundness of the feature amount data 32 that is sequentially generated. be able to.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Medical Informatics (AREA)
- Evolutionary Computation (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Sustainable Development (AREA)
- Life Sciences & Earth Sciences (AREA)
- Sustainable Energy (AREA)
- Power Engineering (AREA)
- Transportation (AREA)
- Mechanical Engineering (AREA)
- Electric Propulsion And Braking For Vehicles (AREA)
- Regulating Braking Force (AREA)
- Valves And Accessory Devices For Braking Systems (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
Description
図1は、本実施の形態に係る健全度診断装置20の構成例を示す第1の図である。健全度診断装置20は、鉄道車両10に搭載されている機器についての稼働データ31を用いて、鉄道車両10に搭載されている機器の健全度を診断する装置である。稼働データ31は、鉄道車両10に搭載される図示しないセンサなどによって検出された、機器の稼働状態を示すデータである。本実施の形態では、具体的に、鉄道車両10に搭載されている機器がブレーキシステム11の場合について説明する。ブレーキシステム11は、図示しないブレーキシリンダーを備え、空気圧によってブレーキ力を制御するシステムとする。
Claims (20)
- 機器の健全度を診断する健全度診断装置であって、
診断対象期間における前記機器の稼働データを取得するデータロード部と、
前記機器の物理特性に基づいて、前記稼働データから特徴量データの対象とするデータ箇所をサンプルデータとして切り出し、前記サンプルデータを用いて前記特徴量データを生成する特徴量データ生成部と、
前記機器の正常時状態をモデル学習した学習済モデルを用いて、逐次生成される前記特徴量データに対して健全度診断を行う推論部と、
前記推論部によって得られた健全度診断結果の推移を可視化する可視化部と、
を備えることを特徴とする健全度診断装置。 - 前記特徴量データのうち正常時に取得された前記特徴量データを用いて前記正常時状態をモデル学習し、当該モデル学習の結果、前記学習済モデルを得る学習部、
を備えることを特徴とする請求項1に記載の健全度診断装置。 - 前記学習部は、機械学習を用いて、前記正常時状態をモデル学習し、
前記推論部は、逐次生成される前記特徴量データについて、前記正常時状態からの乖離度合いをスコアとして算出し、スコア化した前記正常時状態からの乖離度合いを正規化したものを前記健全度診断結果とする、
ことを特徴とする請求項2に記載の健全度診断装置。 - 前記機器は鉄道車両に搭載されるブレーキシステムであり、
前記特徴量データ生成部は、前記稼働データから、前記鉄道車両のブレーキが緩解されるタイミングを含む範囲を前記サンプルデータとして切り出す、
ことを特徴とする請求項1から3のいずれか1つに記載の健全度診断装置。 - 前記機器は鉄道車両に搭載されるブレーキシステムであり、
前記特徴量データ生成部は、前記鉄道車両の車速情報およびブレーキノッチ情報に基づいて、前記稼働データに含まれるブレーキシリンダー圧力から、前記鉄道車両の発車時のブレーキが緩解されるタイミングを含む範囲を前記サンプルデータとして切り出す、
ことを特徴とする請求項1から3のいずれか1つに記載の健全度診断装置。 - 前記機器は鉄道車両に搭載されるブレーキシステムであり、
前記特徴量データ生成部は、切り出した前記サンプルデータに対してクレンジングを行い、前記鉄道車両の走行時の環境条件が統一されるようにフィルタリングを行う、
ことを特徴とする請求項1から5のいずれか1つに記載の健全度診断装置。 - 前記特徴量データ生成部は、前記機器の前記物理特性に基づいて、フィルタリング後の前記サンプルデータに加工を施して前記特徴量データを生成する、
ことを特徴とする請求項6に記載の健全度診断装置。 - 前記可視化部は、過去から現在までの前記健全度診断結果を時系列プロットすることで推移傾向を可視化する、
ことを特徴とする請求項1から7のいずれか1つに記載の健全度診断装置。 - 前記機器と異なる場所に設置され、遠隔で前記機器の健全度を診断する、
ことを特徴とする請求項1から8のいずれか1つに記載の健全度診断装置。 - 前記データロード部が前記稼働データを取得する対象となる前記診断対象期間、および前記特徴量データ生成部が前記機器の物理特性に基づいて前記特徴量データを生成する際の各種の条件の設定を前記健全度診断装置のユーザから受け付ける条件設定部、
を備え、
前記健全度診断装置のユーザは、前記条件設定部を用いて前記条件の設定を行い、前記可視化部で可視化された前記健全度診断結果を確認する、
ことを特徴とする請求項9に記載の健全度診断装置。 - 機器の健全度を診断する健全度診断装置の健全度診断方法であって、
データロード部が、診断対象期間における前記機器の稼働データを取得する取得ステップと、
特徴量データ生成部が、前記機器の物理特性に基づいて、前記稼働データから特徴量データの対象とするデータ箇所をサンプルデータとして切り出し、前記サンプルデータを用いて前記特徴量データを生成する生成ステップと、
推論部が、前記機器の正常時状態をモデル学習した学習済モデルを用いて、逐次生成される前記特徴量データに対して健全度診断を行う推論ステップと、
可視化部が、前記推論部によって得られた健全度診断結果の推移を可視化する可視化ステップと、
を含むことを特徴とする健全度診断方法。 - 学習部が、前記特徴量データのうち正常時に取得された前記特徴量データを用いて前記正常時状態をモデル学習し、当該モデル学習の結果、前記学習済モデルを得る学習ステップ、
を含むことを特徴とする請求項11に記載の健全度診断方法。 - 前記学習ステップにおいて、前記学習部は、機械学習を用いて、前記正常時状態をモデル学習し、
前記推論ステップにおいて、前記推論部は、逐次生成される前記特徴量データについて、前記正常時状態からの乖離度合いをスコアとして算出し、スコア化した前記正常時状態からの乖離度合いを正規化したものを前記健全度診断結果とする、
ことを特徴とする請求項12に記載の健全度診断方法。 - 前記機器は鉄道車両に搭載されるブレーキシステムであり、
前記生成ステップにおいて、前記特徴量データ生成部は、前記稼働データから、前記鉄道車両のブレーキが緩解されるタイミングを含む範囲を前記サンプルデータとして切り出す、
ことを特徴とする請求項11から13のいずれか1つに記載の健全度診断方法。 - 前記機器は鉄道車両に搭載されるブレーキシステムであり、
前記生成ステップにおいて、前記特徴量データ生成部は、前記鉄道車両の車速情報およびブレーキノッチ情報に基づいて、前記稼働データに含まれるブレーキシリンダー圧力から、前記鉄道車両の発車時のブレーキが緩解されるタイミングを含む範囲を前記サンプルデータとして切り出す、
ことを特徴とする請求項11から13のいずれか1つに記載の健全度診断方法。 - 前記機器は鉄道車両に搭載されるブレーキシステムであり、
前記生成ステップにおいて、前記特徴量データ生成部は、切り出した前記サンプルデータに対してクレンジングを行い、前記鉄道車両の走行時の環境条件が統一されるようにフィルタリングを行う、
ことを特徴とする請求項11から15のいずれか1つに記載の健全度診断方法。 - 前記生成ステップにおいて、前記特徴量データ生成部は、前記機器の前記物理特性に基づいて、フィルタリング後の前記サンプルデータに加工を施して前記特徴量データを生成する、
ことを特徴とする請求項16に記載の健全度診断方法。 - 前記可視化ステップにおいて、前記可視化部は、過去から現在までの前記健全度診断結果を時系列プロットすることで推移傾向を可視化する、
ことを特徴とする請求項11から17のいずれか1つに記載の健全度診断方法。 - 前記健全度診断装置は前記機器と異なる場所に設置され、遠隔で前記機器の健全度を診断する、
ことを特徴とする請求項11から18のいずれか1つに記載の健全度診断方法。 - 条件設定部が、前記データロード部が前記稼働データを取得する対象となる前記診断対象期間、および前記特徴量データ生成部が前記機器の物理特性に基づいて前記特徴量データを生成する際の各種の条件の設定を前記健全度診断装置のユーザから受け付ける条件設定ステップ、
を含み、
前記健全度診断装置のユーザは、前記条件設定部を用いて前記条件の設定を行い、前記可視化部で可視化された前記健全度診断結果を確認する、
ことを特徴とする請求項19に記載の健全度診断方法。
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| DE112021008198.5T DE112021008198T5 (de) | 2021-09-06 | 2021-09-06 | Unversehrtheitsdiagnose-Vorrichtung und Unversehrtheitsdiagnose-Verfahren |
| JP2023544977A JP7475553B2 (ja) | 2021-09-06 | 2021-09-06 | 健全度診断装置および健全度診断方法 |
| US18/682,164 US20240378928A1 (en) | 2021-09-06 | 2021-09-06 | Soundness diagnosis apparatus and soundness diagnosis method |
| PCT/JP2021/032651 WO2023032201A1 (ja) | 2021-09-06 | 2021-09-06 | 健全度診断装置および健全度診断方法 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/JP2021/032651 WO2023032201A1 (ja) | 2021-09-06 | 2021-09-06 | 健全度診断装置および健全度診断方法 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2023032201A1 true WO2023032201A1 (ja) | 2023-03-09 |
Family
ID=85411089
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2021/032651 Ceased WO2023032201A1 (ja) | 2021-09-06 | 2021-09-06 | 健全度診断装置および健全度診断方法 |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US20240378928A1 (ja) |
| JP (1) | JP7475553B2 (ja) |
| DE (1) | DE112021008198T5 (ja) |
| WO (1) | WO2023032201A1 (ja) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003237561A (ja) * | 2002-02-18 | 2003-08-27 | Nabco Ltd | 鉄道車両用ブレーキ装置及びその制御方法 |
| JP2020093770A (ja) * | 2015-12-17 | 2020-06-18 | 株式会社東芝 | 状態診断装置及び方法 |
| JP6851558B1 (ja) * | 2020-04-27 | 2021-03-31 | 三菱電機株式会社 | 異常診断方法、異常診断装置および異常診断プログラム |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6889057B2 (ja) * | 2017-07-14 | 2021-06-18 | 株式会社東芝 | 情報処理装置、情報処理方法及びコンピュータプログラム |
-
2021
- 2021-09-06 DE DE112021008198.5T patent/DE112021008198T5/de active Pending
- 2021-09-06 JP JP2023544977A patent/JP7475553B2/ja active Active
- 2021-09-06 WO PCT/JP2021/032651 patent/WO2023032201A1/ja not_active Ceased
- 2021-09-06 US US18/682,164 patent/US20240378928A1/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003237561A (ja) * | 2002-02-18 | 2003-08-27 | Nabco Ltd | 鉄道車両用ブレーキ装置及びその制御方法 |
| JP2020093770A (ja) * | 2015-12-17 | 2020-06-18 | 株式会社東芝 | 状態診断装置及び方法 |
| JP6851558B1 (ja) * | 2020-04-27 | 2021-03-31 | 三菱電機株式会社 | 異常診断方法、異常診断装置および異常診断プログラム |
Also Published As
| Publication number | Publication date |
|---|---|
| US20240378928A1 (en) | 2024-11-14 |
| JPWO2023032201A1 (ja) | 2023-03-09 |
| DE112021008198T5 (de) | 2024-07-11 |
| JP7475553B2 (ja) | 2024-04-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3038048A1 (en) | System and method for determining vehicle component conditions | |
| CN106740441B (zh) | 用于诊断刹车灯的方法和装置 | |
| ITCO20090068A1 (it) | Metodo e sistema per diagnosticare compressori | |
| CN114254449B (zh) | 信息处理方法和装置、显示方法和装置、记录介质、产品制造方法、以及学习数据获取方法 | |
| JP2014174983A (ja) | ターゲットシステムにおいて障害を特定する方法、該方法によりターゲットシステムを修復する手順、該方法を実行するコンピュータシステムおよびコンピュータプログラムならびに該プログラムを格納するコンピュータ可読媒体 | |
| KR102248732B1 (ko) | 종합주의력 검사 데이터에 기초하여 주의력 결핍 및 과잉 행동 장애를 분류 및 치료반응을 예측하는 시스템 및 방법 | |
| JP7166839B2 (ja) | ヘルスモニタリングシステム | |
| TWI861824B (zh) | 異常診斷方法、異常診斷裝置以及異常診斷程式 | |
| US20260110599A1 (en) | Vibration monitoring in order to detect an error during a process automation | |
| WO2020204043A1 (ja) | 高炉の異常判定装置、高炉の異常判定方法、及び高炉の操業方法 | |
| JP2001318031A (ja) | 装置の異常診断方法 | |
| CN103508303A (zh) | 异常诊断方法、异常诊断装置以及具有异常诊断装置的乘客传送设备 | |
| WO2023032201A1 (ja) | 健全度診断装置および健全度診断方法 | |
| JP7153585B2 (ja) | 異常原因推定方法、および、異常原因推定装置 | |
| JP2024109377A (ja) | 車両診断システム | |
| JP2016222387A (ja) | 昇降機の診断方式 | |
| US12204397B2 (en) | Fault diagnosis apparatus, non-transitory computer-readable recording medium, and fault diagnosis method | |
| JP2019191799A5 (ja) | ||
| EP3743816A1 (en) | Maintenance intervention predicting | |
| CN112685827B (zh) | 数据记录装置以及数据记录方法 | |
| JP2025019460A (ja) | 電動機振動モデル生成装置、電動機診断システム及び電動機診断プログラム | |
| JP2008256442A (ja) | タイヤ耐久力性能予測方法、タイヤ耐久力性能予測装置、及びタイヤ耐久力性能予測プログラム | |
| JP2007322377A (ja) | 車載故障診断装置およびその試験方法 | |
| KR102385534B1 (ko) | 데이터 간 상호 유사도에 기반한 가상수명데이터 생성장치 및 방법 | |
| JP7551031B1 (ja) | 転がり軸受の異常検出装置、及びロジスティック回帰モデルの生成方法 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21956086 Country of ref document: EP Kind code of ref document: A1 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2023544977 Country of ref document: JP |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 18682164 Country of ref document: US |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 202427011513 Country of ref document: IN |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 112021008198 Country of ref document: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 21956086 Country of ref document: EP Kind code of ref document: A1 |