WO2023000482A1 - 一种基于机理分析的站台门异常检测方法及装置 - Google Patents

一种基于机理分析的站台门异常检测方法及装置 Download PDF

Info

Publication number
WO2023000482A1
WO2023000482A1 PCT/CN2021/118886 CN2021118886W WO2023000482A1 WO 2023000482 A1 WO2023000482 A1 WO 2023000482A1 CN 2021118886 W CN2021118886 W CN 2021118886W WO 2023000482 A1 WO2023000482 A1 WO 2023000482A1
Authority
WO
WIPO (PCT)
Prior art keywords
curve
door
data
curves
abnormal
Prior art date
Application number
PCT/CN2021/118886
Other languages
English (en)
French (fr)
Inventor
刘文凯
李鸿飞
王玥邈
贾建平
Original Assignee
广州新科佳都科技有限公司
广州华佳软件有限公司
佳都科技集团股份有限公司
广东华之源信息工程有限公司
广州佳都城轨智慧运维服务有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 广州新科佳都科技有限公司, 广州华佳软件有限公司, 佳都科技集团股份有限公司, 广东华之源信息工程有限公司, 广州佳都城轨智慧运维服务有限公司 filed Critical 广州新科佳都科技有限公司
Publication of WO2023000482A1 publication Critical patent/WO2023000482A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/243—Classification techniques relating to the number of classes
    • G06F18/24323—Tree-organised classifiers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/088—Non-supervised learning, e.g. competitive learning

Definitions

  • the embodiments of the present application relate to the technical field of platform door operation, and in particular to a platform door abnormality detection method and device based on mechanism analysis.
  • Abnormal detection of platform doors is an important part of intelligent operation and maintenance, and it is also the premise of fault identification and prediction of platform doors.
  • In the actual operation and maintenance scenario due to the diversity of platform door fault types, it also leads to abnormal diversity.
  • the anomaly detection of platform doors is still based on the threshold judgment method, and the anomaly detection method based on threshold judgment does not have the ability to detect complex anomalies.
  • the embodiment of the present application provides a platform door abnormality detection method and device based on mechanism analysis, which can solve the problem of detecting complex abnormal situations, improve the accuracy of platform door abnormality detection, and improve the abnormal early warning ability of platform door related equipment.
  • the embodiment of the present application provides a platform door abnormality detection method based on mechanism analysis, including:
  • the operation data curves of each equipment are respectively segmented to obtain corresponding segmented curves
  • the detection index of the operation data of each equipment input the detection index data corresponding to each segmented curve into the abnormal detection algorithm model, and obtain the predicted abnormal curve;
  • Anomalies are predicted based on the predicted anomaly curve.
  • operation data of each equipment is processed to obtain the operation data curve of each equipment, specifically:
  • Data processing is performed on the detected electromagnet current data to obtain a current-time curve of the electromagnet.
  • the speed-time curve of the door body includes an acceleration phase, a constant speed phase, a deceleration phase and a crawling phase;
  • the current-time curve of the motor includes an acceleration phase, a constant speed phase, a deceleration phase and a creeping phase;
  • the current-time curve of the electromagnet includes a suction lock phase, an engagement lock phase and a stable phase.
  • the operation data curves of each equipment are respectively segmented to obtain corresponding segmented curves, specifically:
  • the current-time curve of the electromagnet is divided into corresponding segmental curves of sucking lock, segmental curve of engaging lock and segmental curve of stability.
  • the detection indicators of the operation data of each device include:
  • the abnormality detection algorithm model is an isolated forest algorithm model
  • the detection index data corresponding to each segmented curve is input into the abnormality detection algorithm model to obtain the predicted abnormality curve, specifically:
  • Machine learning is performed on the input data through the isolation forest algorithm, and the corresponding predicted abnormal data curve is output.
  • the prediction of the abnormal phenomenon according to the prediction abnormality curve is specifically:
  • the embodiment of the present application provides a platform door abnormality detection device based on mechanism analysis, including:
  • the detection module is used to detect the operation data of each equipment of the platform door;
  • the data processing module is used to process the operating data of each device to obtain the operating data curve of each device;
  • a segmentation module configured to segment the operating data curves of each device to obtain corresponding segmented curves based on the state change mechanism during the operation of the device;
  • the abnormality prediction module is used to preset the detection index of each equipment operation data, and the detection index data corresponding to each segmented curve is input into the abnormality detection algorithm model to obtain the predicted abnormality curve;
  • the abnormal phenomenon query module is used to predict the abnormal phenomenon according to the predicted abnormal phenomenon curve.
  • an electronic device including:
  • the memory is used to store one or more programs
  • the one or more processors When the one or more programs are executed by the one or more processors, the one or more processors implement the platform door abnormality detection method based on mechanism analysis as described in the first aspect.
  • the embodiment of the present application provides a storage medium containing computer-executable instructions, the computer-executable instructions are used to perform the mechanism-based analysis platform as described in the first aspect when executed by a computer processor Door anomaly detection method.
  • the operating data curves of each equipment are obtained by processing the detected operating data of each equipment at the platform door, and the operating data curves of each equipment are respectively segmented on the basis of the state change mechanism during the operation of each equipment to obtain the corresponding Segmented curves, and the detection index data corresponding to each segmented curve is input into the abnormal detection algorithm model to obtain the predicted abnormal curve to predict the abnormal phenomenon.
  • Fig. 1 is a flow chart of a platform door abnormality detection method based on mechanism analysis provided by Embodiment 1 of the present application;
  • Fig. 2 is the speed-time graph of the door body in the door opening process in Embodiment 1 of the present application;
  • Fig. 3 is a schematic diagram of the speed state change mechanism of the door body during the door opening process in Embodiment 1 of the present application;
  • Fig. 4 is the speed-time graph of the door body in the process of closing the door in Embodiment 1 of the present application;
  • FIG. 5 is a schematic diagram of the speed state change mechanism of the door body during the door closing process in Embodiment 1 of the present application;
  • Fig. 6 is the current-time curve of the door opening process motor in the first embodiment of the present application.
  • Fig. 7 is a schematic diagram of the current state change mechanism of the motor in the door opening process in Embodiment 1 of the present application;
  • Fig. 8 is the current-time curve of the motor in the door closing process in Embodiment 1 of the present application.
  • Fig. 9 is a schematic diagram of the current state change mechanism of the motor in the door closing process in Embodiment 1 of the present application.
  • Fig. 10 is the current-time curve of the electromagnet in Embodiment 1 of the present application.
  • Fig. 11 is a schematic diagram of the current state change mechanism of the electromagnet in Embodiment 1 of the present application.
  • Fig. 12 is a schematic diagram of the abnormal curve of incomplete locking in Embodiment 1 of the present application.
  • Fig. 13 is a schematic diagram of the abnormal curve of the bottom of the door in the first embodiment of the present application.
  • Fig. 14 is a schematic diagram of the abnormal curve of the column scratching the glass in the first embodiment of the present application during the crawling stage;
  • Fig. 15 is a schematic structural diagram of a platform door abnormality detection device based on mechanism analysis provided in Embodiment 2 of the present application;
  • FIG. 16 is a schematic structural diagram of an electronic device provided in Embodiment 3 of the present application.
  • the platform door abnormality detection method and device based on mechanism analysis provided by this application aims to process the detected operating data of each equipment of the platform door to obtain the operating data curve of each equipment, based on the mechanism basis of state changes during the operation of each equipment
  • the operation data curve of each equipment is segmented to obtain the corresponding segmented curve, and the detection index data corresponding to each segmented curve is input into the abnormal detection algorithm model to obtain the predicted abnormal curve to predict the abnormal phenomenon, so as to analyze the complex abnormal situation Detection, improve the accuracy of platform door abnormality detection, and improve the abnormal early warning ability of platform door related equipment.
  • the machine learning training of data is carried out through the abnormal detection algorithm model to improve the work efficiency of finding abnormal points.
  • the abnormal situation is usually more complicated, and in actual situations, there are often noises and abnormal points mixed together, which are difficult to distinguish.
  • methods based on threshold judgment are usually used for anomaly detection, and anomaly detection methods based on threshold judgment do not have the ability to detect complex anomalies.
  • the platform door anomaly detection method based on mechanism analysis according to the embodiment of the present application is provided to solve the existing problem that it is difficult to detect complex anomalies.
  • FIG 1 shows a flow chart of a platform door anomaly detection method based on mechanism analysis provided in Embodiment 1 of the present application.
  • the platform door anomaly detection method based on mechanism analysis provided in this embodiment can be composed of Execution by abnormal detection equipment, the platform door abnormal detection equipment based on mechanism analysis can be realized by software and/or hardware, the platform door abnormal detection equipment based on mechanism analysis can be composed of two or more physical entities, or can be A physical entity constitutes.
  • the platform door anomaly detection device based on mechanism analysis can be an intelligent terminal or the like.
  • the platform door anomaly detection method based on mechanism analysis specifically includes:
  • the various equipment of the platform door are specifically equipment such as a door body, a motor, and an electromagnet lock.
  • the sensor detects data such as door moving speed, door moving acceleration, motor voltage and current values, electromagnet current values, and magnetic flux.
  • data processing is performed on the detected data such as door moving speed, door moving acceleration, motor voltage and current value, electromagnet current value, magnetic flux, etc., to obtain the operation data of door body equipment, motor equipment and electromagnet lock lifting equipment curve.
  • data processing is performed on the detected moving speed data of the door body to obtain a speed-time curve of the door body. Detect the speed data of the door body during the door opening process and the door closing process, and perform data processing on the detected door moving speed data and closing door moving speed data to obtain the speed-time curve of the door body during the door opening process and the door closing process The speed-time curve of the door body.
  • the door speed-time curve of the door body includes an acceleration phase, a constant speed phase, a deceleration phase and a crawling phase.
  • the speed-time curves of the door body during the door opening process and the door closing process are fitted, and the door body speed-time fitting curve of the door opening process and the door body speed-time fitting process of the door closing process are obtained
  • the curve is segmented according to the mechanism of the door moving speed change during the door opening process and the door closing process.
  • the door body speed-time fitting curve in the door opening process is divided into four stages, the first stage is the acceleration stage, and the acceleration stage includes slow Acceleration phase and fast acceleration phase.
  • the slow acceleration stage is to open the door after receiving the start door opening command.
  • the door opening speed gradually increases, and the acceleration remains unchanged;
  • the speed increases continuously up to the preset door opening speed threshold.
  • the second stage is the constant speed stage. After the door opening speed reaches the preset door opening speed threshold, the door moves at a constant speed at the speed threshold.
  • the constant speed stage ends and enters the third stage.
  • the third stage is the deceleration stage.
  • the fourth stage is the crawling stage, the door crawls at a lower speed until the door is opened. Divide the time points according to the four stages of the door body speed-time fitting curve in the door opening process, divide the corresponding time points in the door body speed-time curve in the door opening process, and divide the door body speed-time curve in the door opening process into corresponding The acceleration phase, constant speed phase, deceleration phase and crawling phase.
  • the door body speed-time fitting curve is also divided into four stages. Except for the first stage without slow acceleration, the other stages are the same as the door opening process, and will not be repeated here.
  • Divide the time points according to the four stages of the door body speed-time fitting curve in the door closing process divide the time points corresponding to the door body speed-time curve in the door closing process, and divide the door body speed-time curve in the door closing process into corresponding The acceleration phase, constant speed phase, deceleration phase and crawling phase.
  • the crawling stage distance is known to be x 1
  • the deceleration distance is set to x 2
  • the total movement distance of door opening is x
  • the calculation method of the specified positions ⁇ 1 and ⁇ 2 is:
  • data processing is performed on the detected motor current data to obtain a current-time curve of the motor.
  • Detect the motor current data during the door opening process and the door closing process perform data processing according to the detected current data of the door opening motor and the door closing motor, and obtain the current-time curve of the motor during the door opening process and the current-time curve of the motor during the door closing process.
  • the motor current-time curve is divided to obtain segmented curves.
  • the current-time curve of the motor includes an acceleration phase, a constant speed phase, a deceleration phase and a creeping phase.
  • the current-time curve of the door opening process and the door closing process motor is fitted, and the door opening process motor current-time fitting curve and the door closing process motor current-time fitting curve are obtained, according to The change mechanism of the motor current during the door opening process and the door closing process is segmented into curves.
  • the motor current is controlled and input by the gate control unit DCU. Detect the current position of the door body and door speed data; input the detected door body position and door speed data into the PID (PID is the abbreviation of Proportional (proportional), Integral (integral), Differential (differential)) algorithm, and output the corresponding current value .
  • PID is the abbreviation of Proportional (proportional), Integral (integral), Differential (differential)) algorithm, and output the corresponding current value .
  • the motor current controls the door body speed by controlling the motor speed, so the phase division is the same as the door body speed.
  • the motor current-time fitting curve of the door opening process is divided into four stages according to the variation mechanism of the motor current, the first stage is the acceleration stage, and the acceleration stage includes the slow acceleration stage and the Fast acceleration phase.
  • the slow acceleration stage is to open the door after receiving the start door opening command, the door opening speed gradually increases, and the current is a certain value; after a certain period of slow acceleration, the motor current increases and enters the fast acceleration stage, and the current in the fast acceleration stage continues to increase to the preset value. door open motor current threshold.
  • the second stage is the constant speed stage. The current transmission tends to a constant value below the maximum value, which is used to resist resistance.
  • the third stage is the deceleration stage.
  • the fourth stage is the crawling stage, the door crawls at a lower speed until the door is opened.
  • the current value will rise at the end. When the rising current is too large, it means that the resistance is too large and abnormal.
  • the motor current-time fitting curve is also divided into four stages. Except for the first stage without slow acceleration, the other stages are the same as the door opening process, and will not be repeated here. Since the gate lock needs to be hit at the final stage of closing the door, the current value will rise in the end. When the rising current is too large, it means that the resistance is too large and abnormal. Divide the time points according to the four stages of the motor current-time fitting curve in the door closing process, divide the corresponding time points in the motor current-time curve in the door closing process, and divide the motor current-time curve in the door closing process into corresponding acceleration stages , constant speed phase, deceleration phase and crawling phase.
  • data processing is performed on the detected electromagnet current data to obtain a current-time curve of the electromagnet. Detect the current data of the electromagnet lock during the door opening process and the door closing process, and perform data processing according to the detected electromagnet current data of door opening and closing to obtain the current-time curve of the electromagnet.
  • the electromagnet current-time curve includes suction lock phase, lock lock phase and stable phase.
  • the current-time curve of the electromagnet in the process of opening the door and the process of closing the door is fitted, and the fitting curve of the electromagnet current-time is obtained, which is carried out according to the mechanism of the electromagnet current change in the process of opening the door and the process of closing the door Segments of the curve.
  • the set lock coefficient is positively correlated with the input voltage
  • the door control unit DCU transmits voltage to the electromagnet
  • the coil in the electromagnet generates electromagnetic force, which attracts the armature and rises to achieve the purpose of lifting the lock.
  • the change process of the current in the electromagnet is as follows: after the coil is connected to the power supply voltage, the generated current increases from zero. Since there is not only resistance but also a certain inductance in the coil, the increase of current cannot reach a stable value immediately, but Index rises. When the current and magnetic flux reach a certain value, the generated suction makes the armature start to move.
  • the gap between the armature and the electromagnet decreases, the inductance in the coil changes, and a counter electromotive force is generated, thus making the coil
  • the current in the coil decreases until the armature moves to the end position.
  • the gap and inductance between the armature and the electromagnet no longer change, that is, the counter electromotive force no longer exists, and the current in the coil continues to rise until it reaches a stable value.
  • the voltage input is disconnected when the door moves to the end position of closing or opening, and the whole process is less than 1 second.
  • the current-time fitting curve of the electromagnet is divided into three stages, the first stage is the suction lock stage, and the suction lock stage receives The voltage starts and ends when the armature moves to the end position; the second stage is the locking stage, and the locking stage starts from the armature movement to the end position and ends when the current reaches a stable value; the third stage is the stable stage, and the stable stage is when the current reaches a stable value The value starts when the door body moves to the closing or opening end position and ends when the voltage input is disconnected.
  • the time from the time when the voltage is received to the time when the armature moves to the end position is the suction lock time, and the time from the time when the armature moves to the end position to the current stable value is the lock time.
  • the time points are divided into three stages, and the corresponding division time points in the current-time curve of the electromagnet during the door opening process and the door closing process are divided into stages, and the door will be opened.
  • the current-time curve of the electromagnet in the process and the door closing process is divided into the corresponding sucking lock phase, locking phase and stable phase.
  • the door body speed-time curve during the door opening process and the door closing process is divided into an acceleration phase, a constant speed phase, a deceleration phase, and a crawling phase.
  • the speed-time curve of the door body is divided into corresponding acceleration segment curves, uniform velocity segment curves, deceleration segment curves and creep segment curves.
  • the motor current-time curves during the door opening process and the door closing process are divided into acceleration phase, constant speed phase, deceleration phase and crawling phase, and the motor current is divided according to the time point of the phase division
  • the current-time curve is divided into corresponding acceleration segment curves, constant speed segment curves, deceleration segment curves and creep segment curves.
  • the electromagnet current-time curves in the door-opening process and the door-closing process are divided into the suction lock stage, the lock lock stage, and the stable stage.
  • the time point divides the current-time curve of the electromagnet into corresponding suction-lock segmental curves, engaging-lock segmental curves and stable segmental curves.
  • the key process is segmented and feature extracted, which is highly interpretable.
  • set corresponding detection indicators for the operation data of each operating equipment in advance include: acceleration time, deceleration time, maximum speed value of the door body, speed average value and variance in the constant speed stage; motor current Acceleration time, deceleration time, maximum current value and maximum current value in creeping phase; electromagnet sucking and locking time and current value in stable phase.
  • the abnormal detection algorithm model is an isolated forest algorithm model; correspondingly, the detection index data corresponding to each segmented curve is input into the abnormal detection algorithm model to obtain the predicted abnormal curve, specifically: through the isolated forest algorithm The input data is subjected to machine learning, and the corresponding predicted abnormal data curve is output.
  • the anomaly detection algorithm model is an isolated forest algorithm model.
  • the isolated forest is an unsupervised anomaly detection method suitable for continuous data, that is, no labeled samples are required for training, but the features need to be continuous.
  • the isolation forest uses a very efficient strategy. In isolation forests, the dataset is randomly split recursively until all sample points are isolated. Under this random split strategy, outliers usually have shorter paths. Intuitively speaking, those clusters with high density need to be cut many times to be isolated, but those points with low density can be easily isolated.
  • the unsupervised learning algorithm of the isolation forest is used to predict the fault of the platform door equipment curve without relying on the historical obstacle information.
  • the isolation forest algorithm is used to detect the abnormality of the equipment curve of the platform door, which is suitable for the situation that the failure frequency of the platform door is small and there is basically no failure record. Based on the fact that the number of obstacles in the existing platform doors is small, using the isolation forest algorithm to detect the abnormality of the equipment curve of the platform door improves the effectiveness of the detection, and it is more applicable to the current situation that the number of common obstacles in the platform door is less.
  • a comparison table of predicted abnormal curves and abnormal phenomena is formed, and a corresponding comparison database is established.
  • the abnormal curve shows that the falling time of the magnet current lock feature is short, the overall lock time is short, and the corresponding abnormal phenomenon is that the lock is not fully locked and the lock gap becomes smaller.
  • the abnormal curve shows that the motor current does not conform to the four-stage door opening and closing process, and the corresponding abnormal phenomenon is that a foreign object is stuck or blocked at the bottom of the door body on site.
  • the abnormal curve shows that the value of the motor current in the creeping stage is greater than the limit value of 1000 mA, and the corresponding abnormal phenomenon is that the pillar scratches the glass during the door opening crawling stage.
  • a comparison table can be formed through abnormal curves and abnormal phenomena, and a database can be established to facilitate operation and maintenance personnel to accurately locate the fault location and simplify the maintenance process.
  • the operation data curves of each equipment are obtained by processing the detected operation data of each equipment, and the operation data curves of each equipment are respectively segmented on the basis of the state change mechanism during the operation of each equipment to obtain corresponding segmented curves.
  • the detection index data corresponding to each segmented curve is input into the abnormal detection algorithm model to obtain the predicted abnormal curve to predict the abnormal phenomenon.
  • FIG. 15 is a schematic structural diagram of a platform door abnormality detection device based on mechanism analysis provided in Embodiment 2 of the present application.
  • the platform door abnormality detection device based on mechanism analysis provided by this embodiment specifically includes: a detection module 21 , a data processing module 22 , a segmentation module 23 , an abnormality prediction module 24 and an abnormality query module 25 .
  • the detection module 21 is used for detecting the operation data of each equipment of the platform door;
  • the data processing module 22 is used for performing data processing on the operation data of each equipment to obtain the operation data curve of each equipment;
  • the segmentation module 23 is used to segment the operating data curves of each equipment to obtain corresponding segmented curves based on the state change mechanism during the operation of the equipment;
  • the abnormal prediction module 24 is used to preset the detection index of each equipment operation data, and input the detection index data corresponding to each segmented curve into the abnormal detection algorithm model to obtain the predicted abnormal curve;
  • the abnormal phenomenon query module 25 is used to predict the abnormal phenomenon according to the predicted abnormal phenomenon curve.
  • the operation data curves of each equipment are obtained by processing the detected operation data of each equipment, and the operation data curves of each equipment are respectively segmented on the basis of the state change mechanism during the operation of each equipment to obtain corresponding segmented curves.
  • the detection index data corresponding to each segmented curve is input into the abnormal detection algorithm model to obtain the predicted abnormal curve to predict the abnormal phenomenon.
  • the platform door anomaly detection device based on mechanism analysis provided in Embodiment 2 of the present application can be used to implement the mechanism analysis-based platform door anomaly detection method provided in Embodiment 1 above, and has corresponding functions and beneficial effects.
  • Embodiment 3 of the present application provides an electronic device.
  • the electronic device includes: a processor 31 , a memory 32 , a communication module 33 , an input device 34 and an output device 35 .
  • the number of processors in the electronic device may be one or more, and the number of memories in the electronic device may be one or more.
  • the processor, memory, communication module, input device and output device of the electronic device can be connected through a bus or in other ways.
  • the memory 32 can be used to store software programs, computer-executable programs and modules, such as the program instructions/modules corresponding to the platform door abnormality detection method based on mechanism analysis described in any embodiment of the present application (such as , the detection module, data processing module, segmentation module, abnormality prediction module and abnormal phenomenon query module in the platform door abnormality detection device based on mechanism analysis).
  • the memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required by a function; the data storage area may store data created according to the use of the device, and the like.
  • the memory may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage devices.
  • the memory may further include memory located remotely from the processor, which remote memory may be connected to the device via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.
  • the communication module 33 is used for data transmission.
  • the processor 31 executes various functional applications and data processing of the device by running the software programs, instructions and modules stored in the memory, that is, to realize the above-mentioned platform door abnormality detection method based on mechanism analysis.
  • the input device 34 can be used for receiving inputted numerical or character information, and generating key signal input related to user setting and function control of the device.
  • the output device 35 may include a display device such as a display screen.
  • the electronic device provided above can be used to execute the mechanism analysis-based abnormality detection method for platform doors provided in the first embodiment above, and has corresponding functions and beneficial effects.
  • the embodiment of the present application also provides a storage medium containing computer-executable instructions.
  • the method for detecting an abnormality of a platform door based on mechanism analysis is provided.
  • the platform door abnormality detection method includes performing data processing on the detected operation data of each equipment to obtain the operation data curve of each equipment, and dividing the operation curve of each equipment based on the state change mechanism in the operation process of each equipment to obtain the corresponding segment curve, and input the detection index data corresponding to each segmented curve into the abnormal detection algorithm model to obtain the predicted abnormal curve to predict the abnormal phenomenon.
  • storage medium any of various types of memory devices or storage devices.
  • storage medium is intended to include: installation media such as CD-ROMs, floppy disks, or tape drives; computer system memory or random access memory such as DRAM, DDR RAM, SRAM, EDO RAM, Rambus RAM, etc. ; non-volatile memory, such as flash memory, magnetic media (eg hard disk or optical storage); registers or other similar types of memory elements, etc.
  • the storage medium may also include other types of memory or combinations thereof.
  • the storage medium may be located in a first computer system in which the program is executed, or may be located in a different second computer system connected to the first computer system through a network such as the Internet.
  • the second computer system may provide program instructions to the first computer for execution.
  • storage medium may include two or more storage media residing in different locations, for example in different computer systems connected by a network.
  • the storage medium may store program instructions (eg embodied as computer programs) executable by one or more processors.
  • the storage medium containing computer-executable instructions provided by the embodiments of the present application is not limited to the above-mentioned platform door abnormality detection method based on mechanism analysis, and can also execute the method described in any embodiment of the present application. Related operations in the provided platform door anomaly detection method based on mechanism analysis.
  • the platform door anomaly detection device based on mechanism analysis, storage medium and electronic equipment provided in the above embodiments can execute the platform door anomaly detection method based on mechanism analysis provided in any embodiment of the present application, which is not described in detail in the above embodiments For technical details, see the platform door anomaly detection method based on mechanism analysis provided by any embodiment of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Mathematical Physics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computing Systems (AREA)
  • Medical Informatics (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Power-Operated Mechanisms For Wings (AREA)

Abstract

一种基于机理分析的站台门异常检测方法及装置。该方法主要包括通过检测站台门各设备运行数据(S101)；将各设备运行数据进行数据处理，得到各设备运行数据曲线(S102)；基于设备运行过程中的状态变化机理，将各设备运行数据曲线分别进行分割得到对应的分段曲线(S103)；预设各设备运行数据的检测指标，将各分段曲线对应的检测指标数据输入异常检测算法模型中，得到预测异常曲线(S104)；根据预测异常曲线预测异常现象(S105)。该方法能够解决对复杂异常情况进行检测的问题，提升站台门异常检测的准确率，提高对站台门相关设备的异常预警能力。

Description

一种基于机理分析的站台门异常检测方法及装置 技术领域
本申请实施例涉及站台门运营技术领域，尤其涉及一种基于机理分析的站台门异常检测方法及装置。
背景技术
随着我国高速铁路和城市轨道交通运营规模的迅速扩大，整个轨道交通行业对于保障运营安全、提高服务质量及降低运营成本，开始显现巨大的刚性需求，对设施和设备的可靠性、可用性、可维修性和安全性提出了越来越高的要求。而目前，我国轨道交通运维呈现出人员分布不均、线路个性化、技术水平差异化、设备制式多样化、客流量持续攀升、拥挤度超标以及需要高效应对突发事件的局面，具体表现在：大量使用人工操作，运维效率较低、运维数据不够细化，频度不够高、缺少处理分析设备设施大数据的系统平台和智能化应用。
站台门异常检测是智能运维的重要内容，也是站台门故障识别及预测的前提。在实际的运维场景中，由于站台门故障类型的多样性，从而也导致了异常多样性。此外，受现场环境干扰，常常存在噪音和异常点混杂在一起难以区分的情况。然而目前对站台门异常检测还停留在基于阈值判断方法中，基于阈值判断的异常检测方法并不具备对复杂异常进行检测的能力。
发明内容
本申请实施例提供一种基于机理分析的站台门异常检测方法及装置，能够解决对复杂异常情况进行检测的问题，提升站台门异常检测的准确率，提高对站台门相关设备的异常预警能力。
在第一方面，本申请实施例提供了一种基于机理分析的站台门异常检测方法，包括：
检测站台门各设备运行数据；
将各设备运行数据进行数据处理，得到各设备运行数据曲线；
基于所述设备运行过程中的状态变化机理，将各设备运行数据曲线分别进行分割得到对应的分段曲线；
预设各设备运行数据的检测指标，将各分段曲线对应的检测指标数据输入异常检测算法模型中，得到预测异常曲线；
根据预测异常曲线预测异常现象。
进一步的，所述将各设备运行数据进行数据处理，得到各设备运行数据曲线，具体为：
将检测到的门体速度数据进行数据处理，得到门体的速度-时间曲线；
将检测到的电机电流数据进行数据处理，得到电机的电流-时间曲线；
将检测到的电磁铁电流数据进行数据处理，得到电磁铁的电流-时间曲线。
进一步的，所述门体的速度-时间曲线包括加速阶段、匀速阶段、减速阶段和爬行阶段；
所述电机的电流-时间曲线包括加速阶段、匀速阶段、减速阶段和爬行阶段；
所述电磁铁的电流-时间曲线包括吸锁阶段、衔锁阶段和稳定阶段。
进一步的，所述基于所述设备运行过程中的状态变化机理，将各设备运行数据曲线分别进行分割得到对应的分段曲线，具体为：
将门体的速度-时间曲线分割成对应的加速分段曲线、匀速分段曲线、减速分段曲线和爬行分段曲线；
将电机的电流-时间曲线分割成对应的加速分段曲线、匀速分段曲线、减速分段曲线和爬行分段曲线；
将电磁铁的电流-时间曲线分割成对应的吸锁分段曲线、衔锁分段曲线和稳定分段曲线。
进一步的，所述各设备运行数据的检测指标，包括：
门体的加速时间、减速时间、最大速度值、匀速阶段的速度均值和方差；
电机电流加速时间、减速时间、最大电流值和爬行阶段的最大电流值；
电磁体吸锁和衔锁时间及稳定阶段的电流值。
进一步的，所述异常检测算法模型为孤立森林算法模型；
对应的，所述将各分段曲线对应的检测指标数据输入异常检测算法模型中，得到预测异常曲线，具体为：
通过孤立森林算法对输入的数据进行机器学习，输出相应的预测异常数据曲线。
进一步的，所述根据预测异常曲线预测异常现象，具体为：
形成预测异常曲线与异常现象对照表，建立相应对照数据库。
在第二方面，本申请实施例提供了一种基于机理分析的站台门异常检测装置，包括：
检测模块，用于检测站台门各设备运行数据；
数据处理模块，用于将各设备运行数据进行数据处理，得到各设备运行数据曲线；
分割模块，用于基于所述设备运行过程中的状态变化机理，将各设备运行数据曲线分别进行分割得到对应的分段曲线；
异常预测模块，用于预设各设备运行数据的检测指标，将各分段曲线对应的检测指标数 据输入异常检测算法模型中，得到预测异常曲线；
异常现象查询模块，用于根据预测异常曲线预测异常现象。
在第三方面，本申请实施例提供了一种电子设备，包括：
存储器以及一个或多个处理器；
所述存储器，用于存储一个或多个程序；
当所述一个或多个程序被所述一个或多个处理器执行，使得所述一个或多个处理器实现如第一方面所述的基于机理分析的站台门异常检测方法。
在第四方面，本申请实施例提供了一种包含计算机可执行指令的存储介质，所述计算机可执行指令在由计算机处理器执行时用于执行如第一方面所述的基于机理分析的站台门异常检测方法。
本申请实施例通过将检测到的站台门各设备运行数据进行数据处理得到各设备运行数据曲线，在基于各设备运行过程中的状态变化机理基础上将各设备运行数据曲线分别进行分割得到对应的分段曲线，并且将各分段曲线对应的检测指标数据输入异常检测算法模型中得到预测异常曲线以预测异常现象。采用上述技术手段，可以通过对各设备运行数据曲线分割后得到的分段曲线进行预设检测指标数据的异常算法训练，输出异常曲线以预测异常现象，以此对复杂异常情况进行检测，提升站台门异常检测的准确率，提高对站台门相关设备的异常预警能力。此外，通过异常检测算法模型进行数据的机器学习训练，提高对异常点查找的工作效率。
附图说明
图1是本申请实施例一提供的一种基于机理分析的站台门异常检测方法的流程图；
图2是本申请实施例一中开门过程门体的速度-时间曲线图；
图3是本申请实施例一中开门过程门体的速度状态变化机理示意图；
图4是本申请实施例一中关门过程门体的速度-时间曲线图；
图5是本申请实施例一中关门过程门体的速度状态变化机理示意图；
图6是本申请实施例一中开门过程电机的电流-时间曲线；
图7是本申请实施例一中开门过程电机的电流状态变化机理示意图；
图8是本申请实施例一中关门过程电机的电流-时间曲线；
图9是本申请实施例一中关门过程电机的电流状态变化机理示意图；
图10是本申请实施例一中电磁铁的电流-时间曲线；
图11是本申请实施例一中电磁铁的电流状态变化机理示意图；
图12是本申请实施例一中的未完全落锁异常曲线示意图；
图13是本申请实施例一中的门体底部遇阻异常曲线示意图；
图14是本申请实施例一中的开门爬行阶段立柱刮玻璃遇阻异常曲线示意图；
图15是本申请实施例二提供的一种基于机理分析的站台门异常检测装置的结构示意图；
图16是本申请实施例三提供的一种电子设备的结构示意图。
具体实施方式
为了使本申请的目的、技术方案和优点更加清楚，下面结合附图对本申请具体实施例作进一步的详细描述。可以理解的是，此处所描述的具体实施例仅仅用于解释本申请，而非对本申请的限定。另外还需要说明的是，为了便于描述，附图中仅示出了与本申请相关的部分而非全部内容。在更加详细地讨论示例性实施例之前应当提到的是，一些示例性实施例被描述成作为流程图描绘的处理或方法。虽然流程图将各项操作(或步骤)描述成顺序的处理，但是其中的许多操作可以被并行地、并发地或者同时实施。此外，各项操作的顺序可以被重新安排。当其操作完成时所述处理可以被终止，但是还可以具有未包括在附图中的附加步骤。所述处理可以对应于方法、函数、规程、子例程、子程序等等。
本申请提供的基于机理分析的站台门异常检测方法及装置，旨在将检测到的站台门各设备运行数据进行数据处理得到各设备运行数据曲线，在基于各设备运行过程中的状态变化机理基础上将各设备运行数据曲线分别进行分割得到对应的分段曲线，并且将各分段曲线对应的检测指标数据输入异常检测算法模型中得到预测异常曲线以预测异常现象，以此对复杂异常情况进行检测，提升站台门异常检测的准确率，提高对站台门相关设备的异常预警能力。此外，通过异常检测算法模型进行数据的机器学习训练，提高对异常点查找的工作效率。相对于传统的站台门异常检测方式，其通常异常情况比较复杂，实际情况下常常存在噪声和异常点混杂一起难于区分。目前通常用基于阈值判断的方法进行异常检测，基于阈值判断的异常检测方法并不具备对复杂异常进行检测的能力。基于此，提供本申请实施例的基于机理分析的站台门异常检测方法，以解决现有难以对复杂异常情况进行检测的问题。
实施例一：
图1给出了本申请实施例一提供的一种基于机理分析的站台门异常检测方法的流程图，本实施例中提供的基于机理分析的站台门异常检测方法可以由基于机理分析的站台门异常检测设备执行，该基于机理分析的站台门异常检测设备可以通过软件和/或硬件的方式实现，该基于机理分析的站台门异常检测设备可以是两个或多个物理实体构成，也可以是一个物理实 体构成。一般而言，该基于机理分析的站台门异常检测设备可以是智能终端等。
下述以智能终端为执行基于机理分析的站台门异常检测方法的主体为例，进行描述。参照图1，该基于机理分析的站台门异常检测方法具体包括：
S101、检测站台门各设备运行数据。
具体的，站台门各设备具体为门体、电机和电磁铁提锁等设备。通过传感器对门体移动速度、门体移动加速度、电机电压电流值和电磁铁电流值、磁通量等数据进行检测。
S102、将各设备运行数据进行数据处理，得到各设备运行数据曲线。
具体的，对检测到的门体移动速度、门体移动加速度、电机电压电流值和电磁铁电流值、磁通量等数据进行数据处理，得到门体设备、电机设备和电磁铁提锁设备的运行数据曲线。
具体的，参照图2和图4，将检测到的门体移动速度数据进行数据处理，得到门体的速度-时间曲线。检测门体开门过程和关门过程中门体移动的速度数据，对检测到的开门门体移动速度数据和关门门体移动速度数据进行数据处理，得到开门过程门体的速度-时间曲线和关门过程门体的速度-时间曲线。
进一步的，基于门体运行过程中的门体移动速度变化机理，将门体速度-时间曲线进行分割得到分段曲线。所述门体的速度-时间曲线包括加速阶段、匀速阶段、减速阶段和爬行阶段。
示例性的，参照图3和图5，对开门过程和关门过程门体的速度-时间曲线进行拟合处理，得到开门过程门体速度-时间拟合曲线和关门过程门体速度-时间拟合曲线，根据开门过程和关门过程中的门体移动速度变化机理进行曲线的分段。
示例性的，参照图3，在开门的过程中，根据门体移动速度的变化机理将开门过程门体速度-时间拟合曲线分为四个阶段，第一阶段为加速阶段，加速阶段包括缓加速阶段和快加速阶段。其中，缓加速阶段为接收到启动开门指令后进行开门操作，开门速度逐渐提高，加速度不变；缓加速一定时间后，加速度增加，进入快加速阶段，快加速阶段加速度保持不变，快加速阶段速度不断增加至预设的开门速度阈值。第二阶段为匀速阶段，开门速度达到预设的开门速度阈值后，门体以该速度阈值进行匀速移动。门体移动到预设的指定位置ω 1时匀速阶段结束，进入第三阶段。第三阶段为减速阶段，门体移动到预设的指定位置ω 1时，开始减速，减速移动至指定位置ω 2时，减速阶段结束，进入第四阶段。第四阶段为爬行阶段，门体以较低速度爬行至开门完成。根据开门过程门体速度-时间拟合曲线四个阶段划分时间点，在开门过程门体速度-时间曲线中相对应的划分时间点进行阶段划分，将开门过程门体速度-时间曲线划分为对应的加速阶段、匀速阶段、减速阶段和爬行阶段。
示例性的，参照图5，在关门过程，门体速度-时间拟合曲线同样分为四个阶段，除第一阶段无缓加速阶段外，其他阶段与开门过程相同，在此不再赘述。根据关门过程门体速度- 时间拟合曲线四个阶段划分时间点，在关门过程门体速度-时间曲线中相对应的划分时间点进行阶段划分，将关门过程门体速度-时间曲线划分为对应的加速阶段、匀速阶段、减速阶段和爬行阶段。
进一步的，在开门和关门过程中，已知爬行阶段距离为x 1，设定减速距离为x 2和开门总运动距离为x，指定位置ω 1和ω 2的计算方式为：
w 1＝x-x 1-x 2，w 2＝x-x 1。
进一步的，参照图6和图8，将检测到的电机电流数据进行数据处理，得到电机的电流-时间曲线。检测门体开门过程和关门过程中电机电流数据，根据检测到的开门电机电流数据和关门电机电流数据进行数据处理，得到开门过程电机的电流-时间曲线和关门过程电机的电流-时间曲线。
进一步的，基于门体运行过程中的电机电流变化机理，将电机电流-时间曲线进行分割得到分段曲线。所述电机的电流-时间曲线包括加速阶段、匀速阶段、减速阶段和爬行阶段。
示例性的，参照图7和图9，对开门过程和关门过程电机的电流-时间曲线进行拟合处理，得到开门过程电机电流-时间拟合曲线和关门过程电机电流-时间拟合曲线，根据开门过程和关门过程中的电机电流的变化机理进行曲线的分段。
具体的，电机电流由门控单元DCU控制输入。检测门体当前位置和门速数据；将检测到门体位置和门速数据输入PID(PID即：Proportional(比例)、Integral(积分)、Differential(微分)的缩写)算法中，输出相应电流值。电机电流通过控制电机转速来控制门体速度，因此和门体速度的阶段划分相同。
示例性的，参照图7，在开门的过程中，根据电机电流的变化机理将开门过程电机电流-时间拟合曲线分为四个阶段，第一阶段为加速阶段，加速阶段包括缓加速阶段和快加速阶段。其中，缓加速阶段为接收到启动开门指令后进行开门操作，开门速度逐渐提高，电流为一定值；缓加速一定时间后，电机电流增加，进入快加速阶段，快加速阶段电流不断增加至预设的开门电机电流阈值。第二阶段为匀速阶段，电流传输趋于最大值以下的恒定值，用于对抗阻力，门体移动到的指定位置ω 1时匀速阶段结束，进入第三阶段。第三阶段为减速阶段，门体移动到预设的指定位置ω 1时，开始减速，电机电流持续减少；门体移动至指定位置ω 2时减速阶段结束，进入第四阶段。第四阶段为爬行阶段，门体以较低速度爬行至开门完成。此外，由于开门最后阶段可能产生机械摩擦，最后电流值会有所回升，当回升电流过大时表示阻力太大异常。根据开门过程电机电流-时间拟合曲线四个阶段划分时间点，在开门过程电机电流-时间曲线中相对应的划分时间点进行阶段划分，将开门过程电机电流-时间曲线划分为对应的加速阶段、匀速阶段、减速阶段和爬行阶段。
示例性的，参照图9，在关门过程中，电机电流-时间拟合曲线同样分为四个阶段，除第一阶段无缓加速阶段外，其他阶段与开门过程相同，在此不再赘述。由于关门最后阶段需要撞击闸锁，最后电流值会有所回升，当回升电流过大时表示阻力太大异常。根据关门过程电机电流-时间拟合曲线四个阶段划分时间点，在关门过程电机电流-时间曲线中相对应的划分时间点进行阶段划分，将关门过程电机电流-时间曲线划分为对应的加速阶段、匀速阶段、减速阶段和爬行阶段。
进一步的，参照图10，将检测到的电磁铁电流数据进行数据处理，得到电磁铁的电流-时间曲线。检测门体开门过程和关门过程中电磁铁提锁的电流数据，根据检测到的开门和关门的电磁铁电流数据进行数据处理，得到电磁铁的电流-时间曲线。
进一步的，基于门体运行过程中的电磁铁电流的变化机理，将电磁铁电流-时间曲线进行分割得到分段曲线。所述电磁铁电流-时间曲线包括吸锁阶段、衔锁阶段、稳定阶段。
示例性的，参照图11，对开门过程和关门过程电磁铁的电流-时间曲线进行拟合处理，得到电磁铁电流-时间拟合曲线，根据开门过程和关门过程中的电磁铁电流变化机理进行曲线的分段。
具体的，设定的提锁系数和输入电压正相关，门控单元DCU传输电压至电磁铁，电磁铁中的线圈产生电磁力，吸附衔铁上升，达到提锁的目的。电磁铁中电流的变化过程如下：线圈接入电源电压以后，产生的电流从零值增加，由于线圈中不仅有电阻，而且有一定的电感存在，所以电流的增加不能立即达到稳定值，而是指数上升。当电流和磁通达到一定数值时，产生的吸力使衔铁开始运动，随着衔铁的运动，衔铁和电磁铁之间的缝隙减小，线圈中的电感发生变化，产生一个反电动势，因而使线圈中的电流减小，直到衔铁运动到终点位置为止，此时衔铁和电磁铁之间的缝隙和电感不再发生变化，即反电动势不再存在，线圈中电流又继续上升，直到稳定值，持续至门体移动至关门或开门终点位置时断开电压输入，整个过程小于1秒。
进一步的，根据开门过程和关门过程中电磁铁电流变化机理，将所述电磁铁的电流-时间拟合曲线分为三个阶段，第一阶段为吸锁阶段，吸锁阶段从电磁铁接收到电压开始到衔铁运动到终点位置时结束；第二阶段是衔锁阶段，衔锁阶段从衔铁运动到终点位置开始到电流达到稳定值时结束；第三阶段为稳定阶段，稳定阶段为电流达到稳定值时开始到门体移动至关门或开门终点位置断开电压输入时结束。从接收到电压的时间开始到衔铁运动到终点位置的时间为吸锁时间，从衔铁运动到终点位置的时间到达到电流稳定值的时间为衔锁时间。根据开门过程和关门过程的电磁铁的电流-时间拟合曲线三个阶段划分时间点，在开门过程和关门过程的电磁铁的电流-时间曲线中相对应的划分时间点进行阶段划分，将开门过程和关门过程 的电磁铁的电流-时间曲线划分为对应的吸锁阶段、衔锁阶段和稳定阶段。
S103、基于所述设备运行过程中的状态变化机理，将各设备运行数据曲线分别进行分割得到对应的分段曲线。
具体的，根据开门过程和关门过程中的门体移动速度变化机理，将开门过程和关门过程门体速度-时间曲线划分为加速阶段、匀速阶段、减速阶段和爬行阶段，根据阶段划分的时间点将门体的速度-时间曲线分割成对应的加速分段曲线、匀速分段曲线、减速分段曲线和爬行分段曲线。
具体的，根据开门过程和关门过程中的电机电流变化机理，将开门过程和关门过程电机电流-时间曲线划分为加速阶段、匀速阶段、减速阶段和爬行阶段，根据阶段划分的时间点将电机的电流-时间曲线分割成对应的加速分段曲线、匀速分段曲线、减速分段曲线和爬行分段曲线。
具体的，根据开门过程和关门过程中的电磁铁提锁过程的电流变化机理，将开门过程和关门过程电磁铁电流-时间曲线划分为吸锁阶段、衔锁阶段和稳定阶段，根据阶段划分的时间点将电磁铁的电流-时间曲线分割成对应的吸锁分段曲线、衔锁分段曲线和稳定分段曲线。
进一步的，基于开关门过程中的各设备运行过程中的状态变化机理，对关键过程进行分割和特征提取，可解释性强。
S104、预设各设备运行数据的检测指标，将各分段曲线对应的检测指标数据输入异常检测算法模型中，得到预测异常曲线。
具体的，预先对各运行设备的运行数据设置对应的检测指标，各设备运行数据的检测指标，包括：门体的加速时间、减速时间、最大速度值、匀速阶段的速度均值和方差；电机电流加速时间、减速时间、最大电流值和爬行阶段的最大电流值；电磁体吸锁和衔锁时间及稳定阶段的电流值。
进一步的，所述异常检测算法模型为孤立森林算法模型；对应的，所述将各分段曲线对应的检测指标数据输入异常检测算法模型中，得到预测异常曲线，具体为：通过孤立森林算法对输入的数据进行机器学习，输出相应的预测异常数据曲线。
具体的，通过将预设的各运行设备运行数据的检测指标数据输入异常检测算法模型，输出异常曲线。所述异常检测算法模型为孤立森林算法模型，孤立森林是一种适用于连续数据的无监督异常检测方法，即不需要有标记的样本来训练，但特征需要是连续的。对于如何查找哪些点容易被孤立，孤立森林使用了一套非常高效的策略。在孤立森林中，递归地随机分割数据集，直到所有的样本点都是孤立的。在这种随机分割的策略下，异常点通常具有较短的路径。直观上来讲，那些密度很高的簇是需要被切很多次才能被孤立，但是那些密度很低 的点很容易就可以被孤立。
通过孤立森林的无监督学习算法对站台门设备曲线进行故障预测，可以不依赖历史障碍信息。采用孤立森林算法对站台门设备曲线，进行异常检测，适合站台门故障发生次数少，基本无故障记录的情况。基于现有的站台门障碍次数较少，运用孤立森林算法对站台门设备曲线进行异常检测提高了检测的有效性，对现有的站台门普遍障碍发生次数较少的现状的适用性更强。
S105、根据预测异常曲线预测异常现象。
具体的，形成预测异常曲线与异常现象对照表，建立相应对照数据库。
示例性的，参照图12，异常曲线表现为磁铁电流提锁特征下落时间短，整体提锁时间短，对应的异常现象为未完全落锁，闸锁间隙变小。
示例性的，参照图13，异常曲线表现为电机电流不符合四阶段开关门过程，对应异常现象为现场门体底部卡异物或遇阻。
示例性的，参照图14，,异常曲线表现为电机电流爬行阶段数值大于限定值1000毫安，对应异常现象为开门爬行阶段立柱刮玻璃遇阻。
进一步的，通过异常曲线和异常现象可形成对照表，建立数据库，方便运维人员准确定位故障位置，简化检修流程。
上述，通过将检测到的各设备运行数据进行数据处理得到各设备运行数据曲线，在基于各设备运行过程中的状态变化机理基础上将各设备运行数据曲线分别进行分割得到对应的分段曲线，并且将各分段曲线对应的检测指标数据输入异常检测算法模型中得到预测异常曲线以预测异常现象。采用上述技术手段，可以通过对各设备运行数据曲线分割后得到的分段曲线进行预设检测指标数据的异常算法训练，输出异常曲线以预测异常现象，以此对复杂异常情况进行检测，提升站台门异常检测的准确率，提高对站台门相关设备的异常预警能力。此外，通过异常检测算法模型进行数据的机器学习训练，提高对异常点查找的工作效率。
实施例二：
在上述实施例的基础上，图15为本申请实施例二提供的一种基于机理分析的站台门异常检测装置的结构示意图。参考图15，本实施例提供的基于机理分析的站台门异常检测装置具体包括：检测模块21、数据处理模块22、分割模块23、异常预测模块24和异常现象查询模块25。
其中，检测模块21用于检测站台门各设备运行数据；
数据处理模块22用于将各设备运行数据进行数据处理，得到各设备运行数据曲线；
分割模块23用于基于所述设备运行过程中的状态变化机理，将各设备运行数据曲线分别进行分割得到对应的分段曲线；
异常预测模块24用于预设各设备运行数据的检测指标，将各分段曲线对应的检测指标数据输入异常检测算法模型中，得到预测异常曲线；
异常现象查询模块25用于根据预测异常曲线预测异常现象。
上述，通过将检测到的各设备运行数据进行数据处理得到各设备运行数据曲线，在基于各设备运行过程中的状态变化机理基础上将各设备运行数据曲线分别进行分割得到对应的分段曲线，并且将各分段曲线对应的检测指标数据输入异常检测算法模型中得到预测异常曲线以预测异常现象。采用上述技术手段，可以通过对各设备运行数据曲线分割后得到的分段曲线进行预设检测指标数据的异常算法训练，输出异常曲线以预测异常现象，以此对复杂异常情况进行检测，提升站台门异常检测的准确率，提高对站台门相关设备的异常预警能力。此外，通过异常检测算法模型进行数据的机器学习训练，提高对异常点查找的工作效率。
本申请实施例二提供的基于机理分析的站台门异常检测装置可以用于执行上述实施例一提供的基于机理分析的站台门异常检测方法，具备相应的功能和有益效果。
实施例三：
本申请实施例三提供了一种电子设备，参照图16，该电子设备包括：处理器31、存储器32、通信模块33、输入装置34及输出装置35。该电子设备中处理器的数量可以是一个或者多个，该电子设备中的存储器的数量可以是一个或者多个。该电子设备的处理器、存储器、通信模块、输入装置及输出装置可以通过总线或者其他方式连接。
存储器32作为一种计算机可读存储介质，可用于存储软件程序、计算机可执行程序以及模块，如本申请任意实施例所述的基于机理分析的站台门异常检测方法对应的程序指令/模块(例如，基于机理分析的站台门异常检测装置中的检测模块、数据处理模块、分割模块、异常预测模块和异常现象查询模块)。存储器可主要包括存储程序区和存储数据区，其中，存储程序区可存储操作系统、至少一个功能所需的应用程序；存储数据区可存储根据设备的使用所创建的数据等。此外，存储器可以包括高速随机存取存储器，还可以包括非易失性存储器，例如至少一个磁盘存储器件、闪存器件、或其他非易失性固态存储器件。在一些实例中，存储器可进一步包括相对于处理器远程设置的存储器，这些远程存储器可以通过网络连接至设备。上述网络的实例包括但不限于互联网、企业内部网、局域网、移动通信网及其组合。
通信模块33用于进行数据传输。
处理器31通过运行存储在存储器中的软件程序、指令以及模块，从而执行设备的各种功 能应用以及数据处理，即实现上述的基于机理分析的站台门异常检测方法。
输入装置34可用于接收输入的数字或字符信息，以及产生与设备的用户设置以及功能控制有关的键信号输入。输出装置35可包括显示屏等显示设备。
上述提供的电子设备可用于执行上述实施例一提供的基于机理分析的站台门异常检测方法，具备相应的功能和有益效果。
实施例四：
本申请实施例还提供一种包含计算机可执行指令的存储介质，所述计算机可执行指令在由计算机处理器执行时用于执行一种基于机理分析的站台门异常检测方法，该基于机理分析的站台门异常检测方法包括将检测到的各设备运行数据进行数据处理得到各设备运行数据曲线，在基于各设备运行过程中的状态变化机理基础上将各设备运行曲线分别进行分割得到对应的分段曲线，并且将各分段曲线对应的检测指标数据输入异常检测算法模型中得到预测异常曲线以预测异常现象。
存储介质——任何的各种类型的存储器设备或存储设备。术语“存储介质”旨在包括：安装介质，例如CD-ROM、软盘或磁带装置；计算机系统存储器或随机存取存储器，诸如DRAM、DDR RAM、SRAM、EDO RAM，兰巴斯(Rambus)RAM等；非易失性存储器，诸如闪存、磁介质(例如硬盘或光存储)；寄存器或其它相似类型的存储器元件等。存储介质可以还包括其它类型的存储器或其组合。另外，存储介质可以位于程序在其中被执行的第一计算机系统中，或者可以位于不同的第二计算机系统中，第二计算机系统通过网络(诸如因特网)连接到第一计算机系统。第二计算机系统可以提供程序指令给第一计算机用于执行。术语“存储介质”可以包括驻留在不同位置中(例如在通过网络连接的不同计算机系统中)的两个或更多存储介质。存储介质可以存储可由一个或多个处理器执行的程序指令(例如具体实现为计算机程序)。
当然，本申请实施例所提供的一种包含计算机可执行指令的存储介质，其计算机可执行指令不限于如上所述的基于机理分析的站台门异常检测方法，还可以执行本申请任意实施例所提供的基于机理分析的站台门异常检测方法中的相关操作。
上述实施例中提供的基于机理分析的站台门异常检测装置、存储介质及电子设备可执行本申请任意实施例所提供的基于机理分析的站台门异常检测方法，未在上述实施例中详尽描述的技术细节，可见本申请任意实施例所提供的基于机理分析的站台门异常检测方法。
上述仅为本申请的较佳实施例及所运用的技术原理。本申请不限于这里所述的特定实施例，对本领域技术人员来说能够进行的各种明显变化、重新调整及替代均不会脱离本申请的 保护范围。因此，虽然通过以上实施例对本申请进行了较为详细的说明，但是本申请不仅仅限于以上实施例，在不脱离本申请构思的情况下，还可以包括更多其他等效实施例，而本申请的范围由权利要求的范围决定。
[根据细则91更正 27.10.2021]　
[根据细则91更正 27.10.2021]　

Claims (10)

  1. 一种基于机理分析的站台门异常检测方法，其特征在于，包括：
    检测站台门各设备运行数据；
    将各设备运行数据进行数据处理，得到各设备运行数据曲线；
    基于所述设备运行过程中的状态变化机理，将各设备运行数据曲线分别进行分割得到对应的分段曲线；
    预设各设备运行数据的检测指标，将各分段曲线对应的检测指标数据输入异常检测算法模型中，得到预测异常曲线；
    根据预测异常曲线预测异常现象。
  2. 根据权利要求1所述的基于机理分析的站台门异常检测方法，其特征在于，所述将各设备运行数据进行数据处理，得到各设备运行数据曲线，具体为：
    将检测到的门体速度数据进行数据处理，得到门体的速度-时间曲线；
    将检测到的电机电流数据进行数据处理，得到电机的电流-时间曲线；
    将检测到的电磁铁电流数据进行数据处理，得到电磁铁的电流-时间曲线。
  3. 根据权利要求2所述的基于机理分析的站台门异常检测方法，其特征在于，所述门体的速度-时间曲线包括加速阶段、匀速阶段、减速阶段和爬行阶段；
    所述电机的电流-时间曲线包括加速阶段、匀速阶段、减速阶段和爬行阶段；
    所述电磁铁的电流-时间曲线包括吸锁阶段、衔锁阶段和稳定阶段。
  4. 根据权利要求2所述的基于机理分析的站台门异常检测方法，其特征在于，所述基于所述设备运行过程中的状态变化机理，将各设备运行数据曲线分别进行分割得到对应的分段曲线，具体为：
    将门体的速度-时间曲线分割成对应的加速分段曲线、匀速分段曲线、减速分段曲线和爬行分段曲线；
    将电机的电流-时间曲线分割成对应的加速分段曲线、匀速分段曲线、减速分段曲线和爬行分段曲线；
    将电磁铁的电流-时间曲线分割成对应的吸锁分段曲线、衔锁分段曲线和稳定分段曲线。
  5. 根据权利要求3所述的基于机理分析的站台门异常检测方法，其特征在于，所述各设备运行数据的检测指标，包括：
    门体的加速时间、减速时间、最大速度值、匀速阶段的速度均值和方差；
    电机电流加速时间、减速时间、最大电流值和爬行阶段的最大电流值；
    电磁体吸锁和衔锁时间及稳定阶段的电流值。
  6. 根据权利要求1所述的基于机理分析的站台门异常检测方法，其特征在于，所述异常 检测算法模型为孤立森林算法模型；
    对应的，所述将各分段曲线对应的检测指标数据输入异常检测算法模型中，得到预测异常曲线，具体为：
    通过孤立森林算法对输入的数据进行机器学习，输出相应的预测异常数据曲线。
  7. 根据权利要求1所述的基于机理分析的站台门异常检测方法，其特征在于，所述根据预测异常曲线预测异常现象，具体为：
    形成预测异常曲线与异常现象对照表，建立相应对照数据库。
  8. 一种基于机理分析的站台门异常检测装置，其特征在于，包括：
    检测模块，用于检测站台门各设备运行数据；
    数据处理模块，用于将各设备运行数据进行数据处理，得到各设备运行数据曲线；
    分割模块，用于基于所述设备运行过程中的状态变化机理，将各设备运行数据曲线分别进行分割得到对应的分段曲线；
    异常预测模块，用于预设各设备运行数据的检测指标，将各分段曲线对应的检测指标数据输入异常检测算法模型中，得到预测异常曲线；
    异常现象查询模块，用于根据预测异常曲线预测异常现象。
  9. 一种电子设备，其特征在于，包括：
    存储器以及一个或多个处理器；
    所述存储器，用于存储一个或多个程序；
    当所述一个或多个程序被所述一个或多个处理器执行，使得所述一个或多个处理器实现如权利要求1-7任一所述的基于机理分析的站台门异常检测方法。
  10. 一种包含计算机可执行指令的存储介质，其特征在于，所述计算机可执行指令在由计算机处理器执行时用于执行如权利要求1-7任一所述的基于机理分析的站台门异常检测方法。
PCT/CN2021/118886 2021-07-23 2021-09-16 一种基于机理分析的站台门异常检测方法及装置 WO2023000482A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110837962.2A CN113591377B (zh) 2021-07-23 2021-07-23 一种基于机理分析的站台门异常检测方法及装置
CN202110837962.2 2021-07-23

Publications (1)

Publication Number Publication Date
WO2023000482A1 true WO2023000482A1 (zh) 2023-01-26

Family

ID=78249341

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/118886 WO2023000482A1 (zh) 2021-07-23 2021-09-16 一种基于机理分析的站台门异常检测方法及装置

Country Status (2)

Country Link
CN (1) CN113591377B (zh)
WO (1) WO2023000482A1 (zh)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113807678A (zh) * 2021-09-01 2021-12-17 日立楼宇技术（广州）有限公司 站台安全门能效管理方法、装置、计算机设备和存储介质
CN116930749A (zh) * 2023-06-15 2023-10-24 浙江联大科技有限公司 管状电机的遇阻检测系统及其方法
CN117113261A (zh) * 2023-10-20 2023-11-24 国网江西省电力有限公司电力科学研究院 一种电力物联网异常检测方法及系统
CN117118907A (zh) * 2023-10-25 2023-11-24 深圳市亲邻科技有限公司 一种门禁流量动态监测系统及其方法
CN117391481A (zh) * 2023-12-11 2024-01-12 阿尔卑斯系统集成(大连)有限公司 一种基于大数据的电力数据监视方法及系统
CN117408537A (zh) * 2023-12-15 2024-01-16 安徽科派自动化技术有限公司 一种能够实现实时风险预测的电能质量监测系统
CN117454096A (zh) * 2023-12-25 2024-01-26 西安高商智能科技有限责任公司 一种电机生产质量检测方法及系统
CN117951616A (zh) * 2024-03-27 2024-04-30 山东海纳智能装备科技股份有限公司 一种局部通风机智能运维分析方法

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000052976A (ja) * 1998-08-18 2000-02-22 Nippon Signal Co Ltd:The プラットホームドアの再開閉システム
CN108416362A (zh) * 2018-01-29 2018-08-17 同济大学 一种道岔异常预警与故障诊断方法
CN110399935A (zh) * 2019-08-02 2019-11-01 哈工大机器人（合肥）国际创新研究院 基于孤立森林机器学习的机器人实时异常监测方法及系统
CN110626900A (zh) * 2019-09-23 2019-12-31 猫岐智能科技（上海）有限公司 设备运行异常判断方法
CN111277459A (zh) * 2020-01-16 2020-06-12 新华三信息安全技术有限公司 一种设备异常检测方法、装置和机器可读存储介质
CN111612040A (zh) * 2020-04-24 2020-09-01 平安直通咨询有限公司上海分公司 基于孤立森林算法的财务数据异常检测方法及相关装置
CN113123955A (zh) * 2019-12-30 2021-07-16 北京国双科技有限公司 柱塞泵异常检测方法、装置、存储介质及电子设备

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP7063600B2 (ja) * 2017-12-19 2022-05-09 東日本旅客鉄道株式会社 戸挟み検知装置及び戸挟み検知システム
CN108382304B (zh) * 2018-02-09 2020-12-11 上海市城市建设设计研究总院（集团）有限公司 基于lte通信的地铁站台门工作监测装置的运行方法
US11562565B2 (en) * 2019-01-03 2023-01-24 Lucomm Technologies, Inc. System for physical-virtual environment fusion
CN111291096B (zh) * 2020-03-03 2023-07-28 腾讯科技（深圳）有限公司 数据集构建方法、装置和存储介质及异常指标检测方法
CN111967532A (zh) * 2020-09-02 2020-11-20 深圳市佳运通电子有限公司 油田加热炉异常检测方法和系统及相关设备
CN112381308A (zh) * 2020-11-20 2021-02-19 广州新科佳都科技有限公司 一种电流预测模型的训练及其电流预测方法与相关装置

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2000052976A (ja) * 1998-08-18 2000-02-22 Nippon Signal Co Ltd:The プラットホームドアの再開閉システム
CN108416362A (zh) * 2018-01-29 2018-08-17 同济大学 一种道岔异常预警与故障诊断方法
CN110399935A (zh) * 2019-08-02 2019-11-01 哈工大机器人（合肥）国际创新研究院 基于孤立森林机器学习的机器人实时异常监测方法及系统
CN110626900A (zh) * 2019-09-23 2019-12-31 猫岐智能科技（上海）有限公司 设备运行异常判断方法
CN113123955A (zh) * 2019-12-30 2021-07-16 北京国双科技有限公司 柱塞泵异常检测方法、装置、存储介质及电子设备
CN111277459A (zh) * 2020-01-16 2020-06-12 新华三信息安全技术有限公司 一种设备异常检测方法、装置和机器可读存储介质
CN111612040A (zh) * 2020-04-24 2020-09-01 平安直通咨询有限公司上海分公司 基于孤立森林算法的财务数据异常检测方法及相关装置

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113807678A (zh) * 2021-09-01 2021-12-17 日立楼宇技术（广州）有限公司 站台安全门能效管理方法、装置、计算机设备和存储介质
CN113807678B (zh) * 2021-09-01 2024-03-22 日立楼宇技术（广州）有限公司 站台安全门能效管理方法、装置、计算机设备和存储介质
CN116930749A (zh) * 2023-06-15 2023-10-24 浙江联大科技有限公司 管状电机的遇阻检测系统及其方法
CN116930749B (zh) * 2023-06-15 2024-02-13 浙江联大科技有限公司 管状电机的遇阻检测系统及其方法
CN117113261B (zh) * 2023-10-20 2024-02-06 国网江西省电力有限公司电力科学研究院 一种电力物联网异常检测方法及系统
CN117113261A (zh) * 2023-10-20 2023-11-24 国网江西省电力有限公司电力科学研究院 一种电力物联网异常检测方法及系统
CN117118907B (zh) * 2023-10-25 2024-02-02 深圳市亲邻科技有限公司 一种门禁流量动态监测系统及其方法
CN117118907A (zh) * 2023-10-25 2023-11-24 深圳市亲邻科技有限公司 一种门禁流量动态监测系统及其方法
CN117391481A (zh) * 2023-12-11 2024-01-12 阿尔卑斯系统集成(大连)有限公司 一种基于大数据的电力数据监视方法及系统
CN117391481B (zh) * 2023-12-11 2024-03-01 阿尔卑斯系统集成(大连)有限公司 一种基于大数据的电力数据监视方法及系统
CN117408537A (zh) * 2023-12-15 2024-01-16 安徽科派自动化技术有限公司 一种能够实现实时风险预测的电能质量监测系统
CN117408537B (zh) * 2023-12-15 2024-05-07 安徽科派自动化技术有限公司 一种能够实现实时风险预测的电能质量监测系统
CN117454096A (zh) * 2023-12-25 2024-01-26 西安高商智能科技有限责任公司 一种电机生产质量检测方法及系统
CN117454096B (zh) * 2023-12-25 2024-03-01 西安高商智能科技有限责任公司 一种电机生产质量检测方法及系统
CN117951616A (zh) * 2024-03-27 2024-04-30 山东海纳智能装备科技股份有限公司 一种局部通风机智能运维分析方法
CN117951616B (zh) * 2024-03-27 2024-05-28 山东海纳智能装备科技股份有限公司 一种局部通风机智能运维分析方法

Also Published As

Publication number Publication date
CN113591377A (zh) 2021-11-02
CN113591377B (zh) 2023-01-03

Similar Documents

Publication Publication Date Title
WO2023000482A1 (zh) 一种基于机理分析的站台门异常检测方法及装置
CN109178001B (zh) 一种地铁车门控制方法及装置、计算机可读存储介质
WO2018049695A1 (zh) 集成数据采集传输装置的门控器及其处理传输方法
CN113807678B (zh) 站台安全门能效管理方法、装置、计算机设备和存储介质
JP2013018577A (ja) エレベータドアの点検装置およびエレベータドアの点検方法
CN110002373A (zh) 变电站巡检机器人室内外自动传送系统及控制方法
CN103670114A (zh) 列车车门控制方法及装置
CN112127718B (zh) 站台门系统滑动门门锁运行状态检测方法
CN109572732A (zh) 一种轨道车辆站台补偿机构的控制方法、系统及轨道车辆
CN113591376B (zh) 一种基于曲线关联分割机制的站台门异常检测方法及装置
Jin et al. High-speed train emergency brake modeling and online identification of time-varying parameters
CN113374377B (zh) 一种库门控制方法、装置、系统及存储介质
CN105484604A (zh) 一种机库大门控制系统及方法
US20210238010A1 (en) Method and elevator controller for detecting a malfunction in an elevator
CN112478973B (zh) 电梯错层处理方法、装置、电梯系统及存储介质
CN113568393B (zh) 站台门故障检测方法、装置、计算机设备和存储介质
CN111045341B (zh) 一种摆闸同步控制方法
WO2019060110A1 (en) SYSTEM FOR MONITORING AND CONTROLLING ELEVATOR TRAFFIC
Torres Perez et al. Automatic Fault Detection and Classification in Lift Door Systems Using Vibration Signal Features
KR20190012709A (ko) 자동문 제어 장치 및 그의 동작 방법
US20240133226A1 (en) Method for operating a door actuator
CN205445276U (zh) 一种机库大门控制系统
CN115526231A (zh) 基于交互式门控单元的系统运维处理方法及系统
KR102514398B1 (ko) 철도차량의 출입문 고장 예측 장치 및 그 방법
JP2005206314A (ja) エレベータの制御装置

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE