WO2023098753A1 - 一种配电终端故障诊断方法、系统、装置及存储介质 - Google Patents

一种配电终端故障诊断方法、系统、装置及存储介质 Download PDF

Info

Publication number
WO2023098753A1
WO2023098753A1 PCT/CN2022/135595 CN2022135595W WO2023098753A1 WO 2023098753 A1 WO2023098753 A1 WO 2023098753A1 CN 2022135595 W CN2022135595 W CN 2022135595W WO 2023098753 A1 WO2023098753 A1 WO 2023098753A1
Authority
WO
WIPO (PCT)
Prior art keywords
power distribution
measurement data
distribution terminal
fault
status
Prior art date
Application number
PCT/CN2022/135595
Other languages
English (en)
French (fr)
Inventor
赖奎
林浩泉
黄景云
赵国荣
苏博波
戴雄杰
李志娟
施明
李岳锋
苏海林
Original Assignee
广东电网有限责任公司江门供电局
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 广东电网有限责任公司江门供电局 filed Critical 广东电网有限责任公司江门供电局
Publication of WO2023098753A1 publication Critical patent/WO2023098753A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/14—Fourier, Walsh or analogous domain transformations, e.g. Laplace, Hilbert, Karhunen-Loeve, transforms
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R23/00—Arrangements for measuring frequencies; Arrangements for analysing frequency spectra
    • G01R23/16—Spectrum analysis; Fourier analysis
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • 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
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
    • Y04S10/52—Outage or fault management, e.g. fault detection or location

Definitions

  • the present invention relates to the technical field of power fault diagnosis, in particular to a method, system, device and storage medium for fault diagnosis of power distribution terminals.
  • the power distribution terminal is an important device to solve the problem of power distribution reliability. It can monitor related equipment, signals and electric energy in the power distribution network. Power distribution terminals are susceptible to failures in actual operation due to the influence of the installation environment. For the distribution automation system, maintenance personnel can only realize the fault of the distribution terminal when the distribution terminal is completely disconnected or refuses to operate, and cannot quickly diagnose the hidden dangers leading to the fault. In addition, the troubleshooting of power distribution terminals is often carried out offline when they are completely disabled. This passive and delayed troubleshooting method makes it difficult for power distribution terminals to troubleshoot early, which greatly increases the cost of operation and maintenance.
  • the invention provides a power distribution terminal fault diagnosis method, system, device and storage medium, which solves the technical problem of fast identification of power distribution terminal faults.
  • the first aspect of the present invention provides a power distribution terminal fault diagnosis method, including:
  • the measurement data model corresponding to each fault type includes the measurement data model in the normal state, the measurement data model corresponding to the precision distortion fault, and the attenuation shock fault corresponding
  • the plurality of measurement data models are processed by short-time Fourier transform to convert the time-domain data type in the measurement data model into a time-frequency diagram;
  • the collecting the operating state of the power distribution terminal and generating a state sequence includes:
  • a state sequence is generated according to the running state evaluation vector of each monitoring point.
  • each functional module of the power distribution terminal includes a central processing unit, an operation control loop, a communication module, a power supply module, and an acquisition module.
  • each function of the power distribution terminal Set the corresponding monitoring points for the fault type of the module, including:
  • Setting the monitoring points of the central processing unit includes task running status, system parameter setting status and global positioning system time synchronization status;
  • Setting monitoring points of said operational control loop includes telemetry status
  • Setting the monitoring points of the communication module includes uplink communication status and downlink communication status
  • Setting the monitoring points of the power module includes power status and battery status
  • Setting the monitoring points of the collection module includes the state of the collection unit and the state of the sensor probe.
  • the input of the processed measurement data model to a one-dimensional neural network for training includes:
  • y i is the matching value of the distribution terminal operating status
  • ⁇ is the balance factor
  • ⁇ (0,1) is the modulation factor, which is related to The weighting factors formed are used to reduce the loss function value to a single sample that is easy to classify.
  • the second aspect of the present invention provides a power distribution terminal fault diagnosis system, including:
  • the collection module is used to collect the operating state of the power distribution terminal and generate a state sequence
  • the measurement data model building module is used to establish a measurement data model corresponding to each fault type according to the state sequence, and the measurement data model corresponding to each fault type includes a measurement data model in a normal state, and a measurement data model corresponding to a precision distortion type fault.
  • a time-frequency transform module configured to process the various measurement data models by short-time Fourier transform, so as to convert the time-domain data types in the measurement data models into time-frequency diagrams;
  • the training module is used to input the processed measurement data model into the one-dimensional neural network for training, and obtain the trained fault diagnosis model;
  • the fault diagnosis module is used to input the state sequence to be tested, use the fault diagnosis model to diagnose the fault type and output the diagnosis result.
  • the collection module includes:
  • the monitoring point setting unit is used to set corresponding monitoring points according to the fault type of each functional module of the power distribution terminal;
  • An operation state evaluation vector acquisition unit configured to collect the operation state of each of the monitoring points, judge the operation state of each of the monitoring points through the self-diagnosis system, and obtain a corresponding operation state evaluation vector;
  • the state sequence generation unit is used to generate a state sequence according to the evaluation vector of the running state of each monitoring point.
  • each functional module of the power distribution terminal includes a central processing unit, an operation control loop, a communication module, a power supply module, and an acquisition module
  • the monitoring point setting unit includes:
  • the first setting subunit is used to set the monitoring points of the central processing unit including task running status, system parameter setting status and global positioning system time synchronization status;
  • the second setting subunit is used to set the monitoring point of the operation control loop including the telemetry state
  • the third setting subunit is used to set the monitoring points of the communication module including uplink communication status and downlink communication status;
  • the fourth setting subunit is used to set the monitoring points of the power module including power status and battery status;
  • the fifth setting subunit is used to set the monitoring points of the collection module including the state of the collection unit and the state of the sensor probe.
  • the training module is specifically used for:
  • y i is the matching value of the distribution terminal operating status
  • ⁇ is the balance factor
  • ⁇ (0,1) is the modulation factor, which is related to The weighting factors formed are used to reduce the loss function value to a single sample that is easy to classify.
  • the third aspect of the present invention provides a power distribution terminal fault diagnosis device, including:
  • the memory is used to store instructions; wherein, the instructions are instructions that can implement the power distribution terminal fault diagnosis method described in any one of the above methods that can be realized;
  • a processor configured to execute instructions in the memory.
  • the fourth aspect of the present invention is a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, it realizes the power distribution terminal described in any of the ways that can be realized above Fault diagnosis method.
  • the present invention has the following advantages:
  • the invention collects the operating state of the power distribution terminal and generates a state sequence, establishes a measurement data model corresponding to each fault type according to the state sequence, and processes the various measurement data models through short-time Fourier transform, so as to The time-domain data type is converted into a time-frequency diagram, and then the processed measurement data model is input to the one-dimensional neural network for training, and the trained fault diagnosis model is obtained, so that the fault of the power distribution terminal can be identified through the fault diagnosis model ;
  • the present invention can realize the rapid diagnosis of the running state of the power distribution terminal, solve the current problems of outstanding quality of the power distribution terminal, the difficulty of operation and maintenance, and the high cost of operation and maintenance, and is beneficial to improve the intelligence level of the inspection, operation and maintenance of the power distribution terminal It has high practicality and applicability to practical engineering problems.
  • Fig. 1 is a flow chart of a method for diagnosing a fault in a power distribution terminal provided in an optional embodiment of the present invention
  • Fig. 2 is a structural block diagram of a power distribution terminal fault diagnosis system provided by an optional embodiment of the present invention.
  • 1-acquisition module 2-measurement data model building module; 3-time-frequency transformation module; 4-training module; 5-fault diagnosis module.
  • Embodiments of the present invention provide a power distribution terminal fault diagnosis method, system, device and storage medium, which are used to solve the technical problem of fast identification of power distribution terminal faults.
  • FIG. 1 shows a flow chart of a method for diagnosing a fault in a power distribution terminal provided by an embodiment of the present invention.
  • a power distribution terminal fault diagnosis method provided by an embodiment of the present invention includes steps S1-S5.
  • Step S1 collecting the operating status of the power distribution terminal and generating a status sequence.
  • the collection of the operating state of the power distribution terminal and generating a state sequence includes:
  • a state sequence is generated according to the running state evaluation vector of each monitoring point.
  • each functional module of the power distribution terminal is divided into a central processing unit, an operation control loop, a communication module, a power supply module and an acquisition module.
  • the central processing unit is the core component of the power distribution terminal, which is mainly responsible for data collection, error analysis, remote control command transmission and power distribution terminal communication processing.
  • the fault description information of each module is as follows:
  • the fault phenomenon includes PT burning and bursting, frequent start-up of backup power supply and abnormal voltage of backup power supply; the fault types caused by this fault phenomenon include power supply PT (voltage transformer) failure, power supply CT (current transformer) failure and backup power supply failure;
  • the fault phenomenon includes that the collected data is far from the mean value and no collected data, and the corresponding fault types include the aging of the measuring equipment, the fault of the transformer, and the fault of the filter circuit;
  • the fault phenomenon includes weak communication signal, no response, repeated link establishment or frequent initialization, and the corresponding fault types include communication interface fault and wireless communication fault;
  • the fault phenomena include switch refusal and switch malfunction, and the corresponding fault types include control circuit faults;
  • the fault phenomena include the terminal being offline, the terminal software restarting, and the switch refusing to move.
  • the corresponding fault types include circuit board aging faults and terminal software crashes.
  • the corresponding monitoring points can be set as follows:
  • Setting the monitoring points of the central processing unit includes task running status, system parameter setting status and global positioning system time synchronization status;
  • Setting monitoring points of said operational control loop includes telemetry status
  • Setting the monitoring points of the communication module includes uplink communication status and downlink communication status
  • Setting the monitoring points of the power module includes power status and battery status
  • Setting the monitoring points of the collection module includes the state of the collection unit and the state of the sensor probe.
  • the self-diagnosis method is used to judge the running state of each monitoring point, and obtain the corresponding running state judgment vector, including:
  • Task running state self-diagnosis judge whether the software task is executed normally through the software execution monitoring program; the corresponding number of the running state evaluation vector in the state sequence is M[1];
  • Parameter self-diagnosis mainly includes self-diagnosis of power distribution terminal system parameters and limit setting parameters.
  • the power distribution terminal system parameters include IP address and device address.
  • the limit setting parameters include anti-shake time; its operating status evaluation vector The corresponding number in the state sequence is M[2];
  • Time synchronization self-diagnosis By designing a self-diagnosis synchronization interrupt program, compare the time synchronization result with the program after the interruption recovery. If the comparison results are consistent, it means that the global positioning system has no faults in time synchronization; its operating state evaluation vector is in the state sequence The corresponding number is M[3];
  • the telemetry state is realized by collecting the remote control instructions issued by the central processor and the collection module, making an intelligent comparison, and determining the causal relationship between the two hits.
  • the corresponding number of the running state evaluation vector in the state sequence is M[4];
  • the self-diagnosis of the distribution terminal communication module includes two parts: uplink communication and downlink communication.
  • the self-diagnosis method of the communication module failure is realized by returning the confirmation command; the self-diagnosis module determines whether the module is faulty by sending a query command to the master station or upstream and downstream intelligent devices and checking whether the module can receive a response; its operating status evaluation vector
  • the corresponding numbers in the state sequence are M[5] and M[6] respectively.
  • the self-diagnosis of the power module includes detecting whether the power supply is cut off, and detecting whether the connected battery is missing or causing undervoltage; the corresponding numbers of its operating state evaluation vectors in the state sequence are M[7], M[8], M[ 9];
  • the data collected for the power distribution terminal includes telemetry and external signal data, and the external measurement includes voltage and current values. If the value is greater than ⁇ , it is outlier data, and the self-diagnosis judgment principle of the acquisition module is as follows:
  • the corresponding numbers of the operating state evaluation vectors corresponding to the above three fault types in the state sequence are M[10], M[11], and M[12].
  • the value of the running state evaluation vector is 1 or 0.
  • Step S2 establish the measurement data model corresponding to each fault type according to the state sequence, the measurement data model corresponding to each fault type includes the measurement data model in the normal state, the measurement data model corresponding to the accuracy distortion type fault, the attenuation shock The measurement data model corresponding to the type fault and the measurement data model corresponding to the harmonic interference type fault.
  • the measurement data model when the setting is in a normal state is:
  • N(0, ⁇ 2 ) is the measurement noise of Gaussian distribution with mean value 0 and variance ⁇ 2
  • A is the amplitude of the measured voltage or current
  • ⁇ is the angular frequency of the measured voltage or current
  • the measurement data model corresponding to the precision distortion fault is:
  • N f (0, ⁇ f 2 ) is the measurement noise of Gaussian distribution with mean value 0 and variance ⁇ f 2 ;
  • the measurement data model is the measurement data model corresponding to the attenuation oscillation fault.
  • the measurement data model is set as :
  • b is the oscillation attenuation constant
  • t is the oscillation measurement time
  • the measurement data model is the measurement data model corresponding to the harmonic interference fault.
  • the measurement data model is set to for:
  • a 2 represent the amplitude and phase angle of the second harmonic component respectively
  • a z Represent the amplitude and phase angle of the zth harmonic component, respectively.
  • Step S3 process the various measurement data models by short-time Fourier transform, so as to convert the time-domain data type in the measurement data model into a time-frequency diagram.
  • the measurement data model of the power distribution terminal established above is a time-domain model.
  • this step introduces Short-time Fourier Transform (STFT) to convert digital data samples is a time-frequency spectrum containing time domain and frequency domain information.
  • STFT Short-time Fourier Transform
  • STFT is a joint time-frequency analysis method for time-varying and non-stationary signals. It can convert one-dimensional fault signal data into a characteristic spectrum containing time domain and frequency domain information.
  • the conversion formula is shown in the following formula:
  • f(t) is the measurement data model to be transformed
  • g(t- ⁇ ) is the window function centered at the time ⁇ ;
  • T is the measurement time length of the window function
  • t is the time
  • Step S4 inputting the processed measurement data model into a one-dimensional neural network for training to obtain a trained fault diagnosis model
  • Step S5 input the state sequence to be tested, use the fault diagnosis model to diagnose the fault type and output the diagnosis result.
  • the input of the processed measurement data model to a one-dimensional neural network for training includes:
  • y i is the matching value of the distribution terminal operating status
  • ⁇ is the balance factor
  • ⁇ (0,1) is the modulation factor, which is related to The weighting factors formed are used to reduce the loss function value to a single sample that is easy to classify.
  • the traditional one-dimensional neural network can handle sequence data, there are still the following problems: 1) As the depth of the neural network increases, the number of parameters increases sharply, the scattering of gradients or the explosion of gradients become more serious and the model is difficult to form and optimize ; 2) The size of the simple convolution kernel is difficult to capture the temporal information of different temporal granularities, which affects the performance of the evaluation. In view of this, it is necessary to improve the structure of the traditional one-dimensional neural network.
  • the embodiment of the present invention improves the cross-entropy, and uses the improved cross-entropy to construct a loss function to train a one-dimensional neural network model, which can effectively improve the accuracy of training the fault diagnosis model.
  • the invention also provides a power distribution terminal fault diagnosis system.
  • FIG. 2 shows a structural block diagram of a power distribution terminal fault diagnosis system provided by an embodiment of the present invention.
  • the collection module 1 is used to collect the operating state of the power distribution terminal and generate a state sequence
  • the measurement data model building module 2 is used to establish a measurement data model corresponding to each fault type according to the state sequence, and the measurement data model corresponding to each fault type includes a measurement data model in a normal state, and a measurement data model corresponding to a precision distortion type fault. Measurement data model, measurement data model corresponding to attenuation and oscillation faults, and measurement data model corresponding to harmonic interference faults;
  • a time-frequency transformation module 3 configured to process the various measurement data models by short-time Fourier transform, so as to convert the time-domain data type in the measurement data model into a time-frequency diagram;
  • the training module 4 is used for inputting the processed measurement data model into the one-dimensional neural network for training to obtain a trained fault diagnosis model
  • the fault diagnosis module 5 is used to input the state sequence to be tested, use the fault diagnosis model to diagnose the fault type and output the diagnosis result.
  • the collection module 1 includes:
  • the monitoring point setting unit is used to set corresponding monitoring points according to the fault type of each functional module of the power distribution terminal;
  • the running state evaluation vector acquisition unit is used to collect the running state of each of the monitoring points, and judge the running state of each of the monitoring points through the self-diagnosis system to obtain a corresponding running state evaluation vector;
  • the state sequence generation unit is used to generate a state sequence according to the evaluation vector of the running state of each monitoring point.
  • each functional module of the power distribution terminal includes a central processing unit, an operation control loop, a communication module, a power supply module, and an acquisition module 1, and the monitoring point setting unit includes:
  • the first setting subunit is used to set the monitoring points of the central processing unit including task running status, system parameter setting status and global positioning system time synchronization status;
  • the second setting subunit is used to set the monitoring point of the operation control loop including the telemetry state
  • the third setting subunit is used to set the monitoring points of the communication module including uplink communication status and downlink communication status;
  • the fourth setting subunit is used to set the monitoring points of the power module including power status and battery status;
  • the fifth setting subunit is used to set the monitoring points of the collection module 1 including the state of the collection unit and the state of the sensor probe.
  • the training module 4 is specifically used for:
  • y i is the matching value of the distribution terminal operating status
  • ⁇ is the balance factor
  • ⁇ (0,1) is the modulation factor, which is related to The weighting factors formed are used to reduce the loss function value to a single sample that is easy to classify.
  • the present invention also provides a power distribution terminal fault diagnosis device, including:
  • the memory is used to store instructions; wherein, the instructions are instructions that can implement the fault diagnosis method for a power distribution terminal as described in any one of the above embodiments;
  • a processor configured to execute instructions in the memory.
  • the present invention also provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the power distribution terminal fault diagnosis as described in any one of the above embodiments is realized method.
  • the typical faults of the power distribution terminals are analyzed for each module, and the prior knowledge of the power distribution terminals is utilized to the greatest extent, which can greatly improve the The success rate of fault diagnosis and identification.
  • the method, system and device of the present invention can be applied to solve the fault problem of complex distribution network terminal equipment, and can quickly obtain comprehensive fault diagnosis results, thereby better improving the load recovery rate.
  • the disclosed systems, devices and methods may be implemented in other ways.
  • the device embodiments described above are only illustrative.
  • the division of the units is only a logical function division. In actual implementation, there may be other division methods.
  • multiple units or components can be combined or May be integrated into another system, or some features may be ignored, or not implemented.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • each functional unit in each embodiment of the present invention may be integrated into one processing unit, each unit may exist separately physically, or two or more units may be integrated into one unit.
  • the above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.
  • the integrated unit is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.
  • the computer software product is stored in a storage medium , including several instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the method described in each embodiment of the present invention.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), magnetic disk or optical disc, etc., which can store program codes. .

Landscapes

  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Analysis (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Computational Mathematics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Algebra (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Remote Monitoring And Control Of Power-Distribution Networks (AREA)

Abstract

一种配电终端故障诊断方法、系统、装置及存储介质，涉及电力故障诊断技术领域。本发明对配电终端的运行状态进行采集并生成状态序列，根据状态序列建立各故障类型对应的测量数据模型，并通过短时傅里叶变换对所述多种测量数据模型进行处理，以将时域数据类型转换为时频图，进而将处理后的测量数据模型输入至一维神经网络进行训练，得到训练完毕的故障诊断模型，从而可以通过该故障诊断模型对配电终端的故障进行识别；本发明可以实现配电终端运行时状态的快速诊断。

Description

一种配电终端故障诊断方法、系统、装置及存储介质
本申请要求于2021年12月02日提交中国专利局、申请号为202111463158.9、发明名称为“一种配电终端故障诊断方法、系统、装置及存储介质”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本发明涉及电力故障诊断技术领域，尤其涉及一种配电终端故障诊断方法、系统、装置及存储介质。
背景技术
配电终端是解决配电可靠性问题的重要设备，其可以监控配电网络中的相关设备、信号和电能。配电终端容易受到安装环境的影响而在实际运行中出现故障。对于配电自动化系统，维护人员只有在配电终端完全断开或拒绝运行的情况下才能意识到配电终端故障，无法快速诊断出导致故障的隐患。此外，配电终端的故障排查往往是在完全禁用的情况下离线进行的，这种被动、延迟的故障检修方式，使得配电终端难以在早期进行故障排查，极大增加了运维开销。
发明内容
本发明提供了一种配电终端故障诊断方法、系统、装置及存储介质，解决了配电终端故障的快速识别的技术问题。
本发明第一方面提供一种配电终端故障诊断方法，包括：
采集配电终端的运行状态并生成状态序列；
根据所述状态序列建立各故障类型对应的测量数据模型，所述各故障类型对应的测量数据模型包括处于正常状态时的测量数据模型、精度失真类故障对应的测量数据模型、衰减震荡类故障对应的测量数据模型以及谐波干扰类故障对应的测量数据模型；
通过短时傅里叶变换对所述多种测量数据模型进行处理，以将测量数 据模型中的时域数据类型转换为时频图；
将处理后的测量数据模型输入至一维神经网络进行训练，得到训练完毕的故障诊断模型；
输入待测状态序列，利用所述故障诊断模型进行故障类型诊断并输出诊断结果。
根据本发明第一方面的一种能够实现的方式，所述采集配电终端的运行状态并生成状态序列，包括：
根据配电终端的各功能模块的故障类型设置相应的监测点；
采集各所述监测点的运行状态，通过自诊断方法对各所述监控点的运行状态进行评判，得到对应的运行状态评判向量；
根据各所述监控点的运行状态评判向量生成状态序列。
根据本发明第一方面的一种能够实现的方式，所述配电终端的各功能模块包括中央处理单元、操作控制回路、通信模块、电源模块及采集模块，所述根据配电终端的各功能模块的故障类型设置相应的监测点，包括：
设置所述中央处理单元的监测点包括任务运行状态、系统参数设置状态及全球定位系统对时状态；
设置所述操作控制回路的监测点包括遥测状态；
设置所述通信模块的监测点包括上行通信状态和下行通信状态；
设置所述电源模块的监测点包括电源状态和电池状态；
设置所述采集模块的监测点包括采集单元状态和传感器探头状态。
根据本发明第一方面的一种能够实现的方式，所述将处理后的测量数据模型输入至一维神经网络进行训练，包括：
利用改进的交叉熵构建损失函数，利用所述损失函数对所述一维神经网络的模型进行训练；
其中所述改进的交叉熵的计算式为：
Figure PCTCN2022135595-appb-000001
式中，
Figure PCTCN2022135595-appb-000002
为故障模型的匹配度函数区间，y i为配电终端运行状态类匹配值，α为平衡因子，α∈(0，1)，γ为调制因子，其与
Figure PCTCN2022135595-appb-000003
构成的权重因子用 于将损失函数值降低到一个易于分类的单个样本。
本发明第二方面提供一种配电终端故障诊断系统，包括：
采集模块，用于采集配电终端的运行状态并生成状态序列；
测量数据模型构建模块，用于根据所述状态序列建立各故障类型对应的测量数据模型，所述各故障类型对应的测量数据模型包括处于正常状态时的测量数据模型、精度失真类故障对应的测量数据模型、衰减震荡类故障对应的测量数据模型以及谐波干扰类故障对应的测量数据模型；
时频变换模块，用于通过短时傅里叶变换对所述多种测量数据模型进行处理，以将测量数据模型中的时域数据类型转换为时频图；
训练模块，用于将处理后的测量数据模型输入至一维神经网络进行训练，得到训练完毕的故障诊断模型；
故障诊断模块，用于输入待测状态序列，利用所述故障诊断模型进行故障类型诊断并输出诊断结果。
根据本发明第二方面的一种能够实现的方式，所述采集模块包括：
监测点设置单元，用于根据配电终端的各功能模块的故障类型设置相应的监测点；
运行状态评判向量获取单元，用于采集各所述监测点的运行状态，通过自诊断系统对各所述监控点的运行状态进行评判，得到对应的运行状态评判向量；
状态序列生成单元，用于根据各所述监控点的运行状态评判向量生成状态序列。
根据本发明第二方面的一种能够实现的方式，所述配电终端的各功能模块包括中央处理单元、操作控制回路、通信模块、电源模块及采集模块，所述监测点设置单元包括：
第一设置子单元，用于设置所述中央处理单元的监测点包括任务运行状态、系统参数设置状态及全球定位系统对时状态；
第二设置子单元，用于设置所述操作控制回路的监测点包括遥测状态；
第三设置子单元，用于设置所述通信模块的监测点包括上行通信状态和下行通信状态；
第四设置子单元，用于设置所述电源模块的监测点包括电源状态和电池状态；
第五设置子单元，用于设置所述采集模块的监测点包括采集单元状态和传感器探头状态。
根据本发明第二方面的一种能够实现的方式，所述训练模块具体用于：
利用改进的交叉熵构建损失函数，利用所述损失函数对所述一维神经网络的模型进行训练；
其中所述改进的交叉熵的计算式为：
Figure PCTCN2022135595-appb-000004
式中，
Figure PCTCN2022135595-appb-000005
为故障模型的匹配度函数区间，y i为配电终端运行状态类匹配值，α为平衡因子，α∈(0，1)，γ为调制因子，其与
Figure PCTCN2022135595-appb-000006
构成的权重因子用于将损失函数值降低到一个易于分类的单个样本。
本发明第三方面提供了一种配电终端故障诊断装置，包括：
存储器，用于存储指令；其中，所述指令为可实现如上任意一项能够实现的方式所述的配电终端故障诊断方法的指令；
处理器，用于执行所述存储器中的指令。
本发明第四方面一种计算机可读存储介质，所述计算机可读存储介质上存储有计算机程序，所述计算机程序被处理器执行时实现如上任意一项能够实现的方式所述的配电终端故障诊断方法。
从以上技术方案可以看出，本发明具有以下优点：
本发明对配电终端的运行状态进行采集并生成状态序列，根据状态序列建立各故障类型对应的测量数据模型，并通过短时傅里叶变换对所述多种测量数据模型进行处理，以将时域数据类型转换为时频图，进而将处理后的测量数据模型输入至一维神经网络进行训练，得到训练完毕的故障诊断模型，从而可以通过该故障诊断模型对配电终端的故障进行识别；本发明可以实现配电终端运行时状态的快速诊断，解决目前配电终端质量问题突出、运维难度大、运维开销多的问题，有益于提升配电终端检运维工作的智能化水平及工作效率，对工程实际问题有着较高的实用性和适用性。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动性的前提下，还可以根据这些附图获得其它的附图。
图1为本发明一个可选实施例提供的一种配电终端故障诊断方法的流程图；
图2为本发明一个可选实施例提供的一种配电终端故障诊断系统的结构框图。
附图说明：
1-采集模块；2-测量数据模型构建模块；3-时频变换模块；4-训练模块；5-故障诊断模块。
具体实施方式
本发明实施例提供了一种配电终端故障诊断方法、系统、装置及存储介质，用于解决配电终端故障的快速识别的技术问题。
为使得本发明的发明目的、特征、优点能够更加的明显和易懂，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，下面所描述的实施例仅仅是本发明一部分实施例，而非全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其它实施例，都属于本发明保护的范围。
请参阅图1，图1示出了本发明实施例提供的一种配电终端故障诊断方法的流程图。
本发明实施例提供的一种配电终端故障诊断方法，包括步骤S1-S5。
步骤S1，采集配电终端的运行状态并生成状态序列。
在一种能够实现的方式中，所述采集配电终端的运行状态并生成状态 序列，包括：
根据配电终端的各功能模块的故障类型设置相应的监测点；
采集各所述监测点的运行状态，通过自诊断方法对各所述监控点的运行状态进行评判，得到对应的运行状态评判向量；
根据各所述监控点的运行状态评判向量生成状态序列。
其中，为实现配电终端故障诊断过程的标准化，将所述配电终端的各功能模块分为中央处理单元、操作控制回路、通信模块、电源模块及采集模块。
其中，中央处理单元为配电终端的核心组成部件，主要负责数据采集、差错分析、远程控制命令传输和配电终端通信处理等。
为设置相应的监测点，需要对各模块的故障进行分析，以得到故障描述信息。各模块的故障描述信息如下：
(1)电源模块：
其故障现象包括PT烧毁与爆裂、后备电源频繁启动以及后备电源电压异常；根据该故障现象引起的故障类型包括电源PT(电压互感器)故障、电源CT(电流互感器)故障及后备电源故障；
(2)采集模块：
其故障现象包括采集数据远离均值和无采集数据，相应的故障类型包括测量设备老化、互感器故障、滤波电路故障；
(3)通信模块：
其故障现象包括通信信号弱、不响应、反复建立链路或频繁初始化，相应的故障类型包括通信接口故障、无线通信故障；
(4)操作控制回路：
其故障现象包括开关拒动和开关误动，相应的故障类型包括控制回路故障；
(5)中央处理单元：
其故障现象包括终端处于离线状态、终端软件重启和开关拒动，相应的故障类型包括电路板老化故障和终端软件死机。
基于上述的故障描述信息，作为一种优选的实施方式，可以按照如下 方式设置相应的监测点：
设置所述中央处理单元的监测点包括任务运行状态、系统参数设置状态及全球定位系统对时状态；
设置所述操作控制回路的监测点包括遥测状态；
设置所述通信模块的监测点包括上行通信状态和下行通信状态；
设置所述电源模块的监测点包括电源状态和电池状态；
设置所述采集模块的监测点包括采集单元状态和传感器探头状态。
其中，所述通过自诊断方法对各所述监控点的运行状态进行评判，得到对应的运行状态评判向量，包括：
(1)中央处理单元：
任务运行状态自诊断：通过软件执行监控程序判断软件任务是否正常执行；其运行状态评判向量在状态序列中对应的编号为M[1]；
参数自诊断：参数自诊断主要包括对配电终端系统参数和限制设置参数进行自诊断，该配电终端系统参数包括IP地址及设备地址，该限制设置参数包括防抖时间；其运行状态评判向量在状态序列中对应的编号为M[2]；
对时自诊断：通过设计自诊断同步中断程序，将时间同步结果与中断恢复后的程序进行比较，比较结果如果一致，则表示全球定位系统对时无故障；其运行状态评判向量在状态序列中对应的编号为M[3]；
(2)操作控制回路：
遥测状态是通过采集中央处理器和采集模块发出的遥控指令，进行智能比较，确定两次命中的因果关系来实现的，其运行状态评判向量在状态序列中对应的编号为M[4]；
(3)通信模块自诊断：
配电终端通信模块的自诊断包括上行通信和下行通信两部分。通信模块故障的自诊断方法是通过确认指令的返回实现；自诊断模块通过向主站或上下游智能设备发出查询命令并检查模块是否能收到响应来确定模块是否有故障；其运行状态评判向量在状态序列中对应的编号分别为M[5]、M[6]。
(4)电源模块自诊断：
电源模块的自诊断包括检测电源是否断电，以及检测连接的电池是否缺失或导致欠电压；其运行状态评判向量在状态序列中对应的编号分别为M[7]、M[8]、M[9]；
(5)采集模块自诊断：
为配电终端收集的数据包括遥测和外部信号数据，外部测量包括电压和电流值；为方便描述采集模块的自诊断模型，设遥测数据为ξ、阈值为ε、平均值为θ，设定差值大于δ为离群数据，则采集模块的自诊断判定原则如下：
若|ξ-θ|＞0，表示数据远离均值，则判定采集单元故障；
若ξ＞δ，表示数据大于阀值，则判定采集单元故障；
若δ＞0，表示数据为0，则判定传感器探头失效故障；
上述三种故障类型对应的运行状态评判向量在状态序列中对应的编号分别为M[10]、M[11]、M[12]。
在上述的自诊断中，运行状态评判向量的值为1或0。可以根据实际情况设置每个运行状态评判向量的值的定义规则。例如，当<[10]＝0时，表示采集单元无故障，当M[10]＝1时，表示采集单元故障。又例如，当M[1]＝1时，表示软件任务正常进行，M[1]＝0时，表示软件任务执行异常。所有监测点的运行状态评判向量的值构成了给定时刻的状态序列M，其中M＝(M[1]，M[2]，...，M[n])，M[k]为监测点k在当前时刻的运行状态评判向量的值。
步骤S2，根据所述状态序列建立各故障类型对应的测量数据模型，所述各故障类型对应的测量数据模型包括处于正常状态时的测量数据模型、精度失真类故障对应的测量数据模型、衰减震荡类故障对应的测量数据模型以及谐波干扰类故障对应的测量数据模型。
由于配电终端中互感器、模数转换器等测量装置实际运行中会时常受到各种因素的干扰，实际中采集到的交流电压或电流会加有一定的随机噪声，一般可认为噪声服从高斯分布。因此，作为优选的实施方式，设置处于正常状态时的测量数据模型为：
Figure PCTCN2022135595-appb-000007
式中，N(0，δ 2)为均值为0，方差为δ 2的高斯分布的测量噪声，A为测量电压或电流的幅值，ω为测量电压或电流的角频率，
Figure PCTCN2022135595-appb-000008
测量电压或电流的相角。
当配电终端采集系统中的互感器等信号处理模块与信号传输单元等发生故障均会导致测量值的精度失真，测量值含有的噪声有偏大的特征，也被称为大噪声故障数据。在发生精度失真故障时，电压、电流测量平均值不变，测量方差发生改变，此时的精度失真类故障对应的测量数据模型为：
Figure PCTCN2022135595-appb-000009
式中，N f(0，δ f 2)为均值为0，方差为δ f 2的高斯分布的测量噪声；
当配电终端中的测量设备老化或性能衰减时，往往使得测量数据也具有一定的震荡衰减特征，此时的测量数据模型即为衰减震荡类故障对应的测量数据模型，该测量数据模型设置为：
Figure PCTCN2022135595-appb-000010
式中，b为振荡衰减常数，t为振荡测量时间；
当配电终端中的测量元件受到环境干扰或元件故障会使得测量的交流量偏离基频。如电磁式互感器的励磁特性劣化时会发生铁磁谐振，致使测量中含有高次谐波分量，此时的测量数据模型即为谐波干扰类故障对应的测量数据模型，该测量数据模型设置为：
Figure PCTCN2022135595-appb-000011
式中，A 2、
Figure PCTCN2022135595-appb-000012
分别代表第2次谐波分量的幅值与相角，A z、
Figure PCTCN2022135595-appb-000013
分别代表第z次谐波分量的幅值与相角。
步骤S3，通过短时傅里叶变换对所述多种测量数据模型进行处理，以将测量数据模型中的时域数据类型转换为时频图。
以上建立的配电终端的测量数据模型为时域模型，为强化不同数据类型的类别特征，本步骤引入短时傅里叶变换(Short-time Fourier Transform，STFT)，将数字式的数据样本转换为含有时域与频域信息的时频图谱。
STFT是针对时变、非平稳信号的一种联合时频分析方法，能将一维的故障信号数据转换为包含时域和频域信息的特征谱，其转换公式见下式：
Figure PCTCN2022135595-appb-000014
式中，f(t)为待变换的测量数据模型，g(t-μ)为中心位于μ时刻的窗函数；
本实施例选取常用的海明窗函数，其表达式见下式：
Figure PCTCN2022135595-appb-000015
式中，T为窗函数的量测时间长度，t为时间。
步骤S4，将处理后的测量数据模型输入至一维神经网络进行训练，得到训练完毕的故障诊断模型；
步骤S5，输入待测状态序列，利用所述故障诊断模型进行故障类型诊断并输出诊断结果。
在一种能够实现的方式中，所述将处理后的测量数据模型输入至一维神经网络进行训练，包括：
利用改进的交叉熵构建损失函数，利用所述损失函数对所述一维神经网络的模型进行训练；
其中所述改进的交叉熵的计算式为：
Figure PCTCN2022135595-appb-000016
式中，
Figure PCTCN2022135595-appb-000017
为故障模型的匹配度函数区间，y i为配电终端运行状态类匹配值，α为平衡因子，α∈(0，1)，γ为调制因子，其与
Figure PCTCN2022135595-appb-000018
构成的权重因子用于将损失函数值降低到一个易于分类的单个样本。
传统的一维神经网络虽然可以处理序列数据，但仍然存在以下问题：1) 随着神经网络深度的增加，参数量急剧增加，梯度的散射或梯度的爆炸变得更加严重和模型难以形成和优化；2)简单卷积核的大小难以捕捉不同时间粒度的时间信息，影响评估的性能。有鉴于此，需要改进传统一维神经网络的结构。
由于故障相较于正常运行状态时的情况较少，因此配电终端稳定运行的状态数据要多于配电终端发生故障的状态数据，这导致模型更加关注稳定样本，从而导致对不稳定样本的误判。
为解决上述问题，本发明实施例对交叉熵进行改进，并利用改进的交叉熵构建损失函数来训练一维神经网络的模型，能够有效提高训练故障诊断模型的精度。
本发明还提供了一种配电终端故障诊断系统。
请参阅图2，图2示出了本发明实施例提供的一种配电终端故障诊断系统的结构框图。
本发明实施例提供的一种配电终端故障诊断系统，包括：
采集模块1，用于采集配电终端的运行状态并生成状态序列；
测量数据模型构建模块2，用于根据所述状态序列建立各故障类型对应的测量数据模型，所述各故障类型对应的测量数据模型包括处于正常状态时的测量数据模型、精度失真类故障对应的测量数据模型、衰减震荡类故障对应的测量数据模型以及谐波干扰类故障对应的测量数据模型；
时频变换模块3，用于通过短时傅里叶变换对所述多种测量数据模型进行处理，以将测量数据模型中的时域数据类型转换为时频图；
训练模块4，用于将处理后的测量数据模型输入至一维神经网络进行训练，得到训练完毕的故障诊断模型；
故障诊断模块5，用于输入待测状态序列，利用所述故障诊断模型进行故障类型诊断并输出诊断结果。
在一种能够实现的方式中，所述采集模块1包括：
监测点设置单元，用于根据配电终端的各功能模块的故障类型设置相应的监测点；
运行状态评判向量获取单元，用于采集各所述监测点的运行状态，通 过自诊断系统对各所述监控点的运行状态进行评判，得到对应的运行状态评判向量；
状态序列生成单元，用于根据各所述监控点的运行状态评判向量生成状态序列。
在一种能够实现的方式中，所述配电终端的各功能模块包括中央处理单元、操作控制回路、通信模块、电源模块及采集模块1，所述监测点设置单元包括：
第一设置子单元，用于设置所述中央处理单元的监测点包括任务运行状态、系统参数设置状态及全球定位系统对时状态；
第二设置子单元，用于设置所述操作控制回路的监测点包括遥测状态；
第三设置子单元，用于设置所述通信模块的监测点包括上行通信状态和下行通信状态；
第四设置子单元，用于设置所述电源模块的监测点包括电源状态和电池状态；
第五设置子单元，用于设置所述采集模块1的监测点包括采集单元状态和传感器探头状态。
在一种能够实现的方式中，所述训练模块4具体用于：
利用改进的交叉熵构建损失函数，利用所述损失函数对所述一维神经网络的模型进行训练；
其中所述改进的交叉熵的计算式为：
Figure PCTCN2022135595-appb-000019
式中，
Figure PCTCN2022135595-appb-000020
为故障模型的匹配度函数区间，y i为配电终端运行状态类匹配值，α为平衡因子，α∈(0，1)，γ为调制因子，其与
Figure PCTCN2022135595-appb-000021
构成的权重因子用于将损失函数值降低到一个易于分类的单个样本。
本发明还提供了一种配电终端故障诊断装置，包括：
存储器，用于存储指令；其中，所述指令为可实现如上任意一项实施例所述的配电终端故障诊断方法的指令；
处理器，用于执行所述存储器中的指令。
本发明还提供了一种计算机可读存储介质，所述计算机可读存储介质上存储有计算机程序，所述计算机程序被处理器执行时实现如上任意一项实施例所述的配电终端故障诊断方法。
本发明上述实施例，在分析各类配电终端现场运行故障的基础上，针对各模块分析了配电终端的典型故障，最大程度地利用了配电终端的先验知识，能够极大地提高了故障诊断识别的成功率。本发明的方法、系统即装置可应用于求解复杂配电网终端设备故障问题，且可快速得到全面的故障诊断结果，从而更好地提高负荷恢复率。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和模块的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本发明各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方 案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，以上实施例仅用以说明本发明的技术方案，而非对其限制；尽管参照前述实施例对本发明进行了详细的说明，本领域的普通技术人员应当理解：其依然可以对前述各实施例所记载的技术方案进行修改，或者对其中部分技术特征进行等同替换；而这些修改或者替换，并不使相应技术方案的本质脱离本发明各实施例技术方案的精神和范围。

Claims (10)

  1. 一种配电终端故障诊断方法，其特征在于，包括：
    采集配电终端的运行状态并生成状态序列；
    根据所述状态序列建立各故障类型对应的测量数据模型，所述各故障类型对应的测量数据模型包括处于正常状态时的测量数据模型、精度失真类故障对应的测量数据模型、衰减震荡类故障对应的测量数据模型以及谐波干扰类故障对应的测量数据模型；
    通过短时傅里叶变换对所述多种测量数据模型进行处理，以将测量数据模型中的时域数据类型转换为时频图；
    将处理后的测量数据模型输入至一维神经网络进行训练，得到训练完毕的故障诊断模型；
    输入待测状态序列，利用所述故障诊断模型进行故障类型诊断并输出诊断结果。
  2. 根据权利要求1所述的配电终端故障诊断方法，其特征在于，所述采集配电终端的运行状态并生成状态序列，包括：
    根据配电终端的各功能模块的故障类型设置相应的监测点；
    采集各所述监测点的运行状态，通过自诊断方法对各所述监控点的运行状态进行评判，得到对应的运行状态评判向量；
    根据各所述监控点的运行状态评判向量生成状态序列。
  3. 根据权利要求2所述的配电终端故障诊断方法，其特征在于，所述配电终端的各功能模块包括中央处理单元、操作控制回路、通信模块、电源模块及采集模块，所述根据配电终端的各功能模块的故障类型设置相应的监测点，包括：
    设置所述中央处理单元的监测点包括任务运行状态、系统参数设置状态及全球定位系统对时状态；
    设置所述操作控制回路的监测点包括遥测状态；
    设置所述通信模块的监测点包括上行通信状态和下行通信状态；
    设置所述电源模块的监测点包括电源状态和电池状态；
    设置所述采集模块的监测点包括采集单元状态和传感器探头状态。
  4. 根据权利要求1所述的配电终端故障诊断方法，其特征在于，所述将处理后的测量数据模型输入至一维神经网络进行训练，包括：
    利用改进的交叉熵构建损失函数，利用所述损失函数对所述一维神经网络的模型进行训练；
    其中所述改进的交叉熵的计算式为：
    Figure PCTCN2022135595-appb-100001
    式中，
    Figure PCTCN2022135595-appb-100002
    为故障模型的匹配度函数区间，y i为配电终端运行状态类匹配值，α为平衡因子，α∈(0，1)，γ为调制因子，其与
    Figure PCTCN2022135595-appb-100003
    构成的权重因子用于将损失函数值降低到一个易于分类的单个样本。
  5. 一种配电终端故障诊断系统，其特征在于，包括：
    采集模块，用于采集配电终端的运行状态并生成状态序列；
    测量数据模型构建模块，用于根据所述状态序列建立各故障类型对应的测量数据模型，所述各故障类型对应的测量数据模型包括处于正常状态时的测量数据模型、精度失真类故障对应的测量数据模型、衰减震荡类故障对应的测量数据模型以及谐波干扰类故障对应的测量数据模型；
    时频变换模块，用于通过短时傅里叶变换对所述多种测量数据模型进行处理，以将测量数据模型中的时域数据类型转换为时频图；
    训练模块，用于将处理后的测量数据模型输入至一维神经网络进行训练，得到训练完毕的故障诊断模型；
    故障诊断模块，用于输入待测状态序列，利用所述故障诊断模型进行故障类型诊断并输出诊断结果。
  6. 根据权利要求5所述的配电终端故障诊断系统，其特征在于，所述采集模块包括：
    监测点设置单元，用于根据配电终端的各功能模块的故障类型设置相应的监测点；
    运行状态评判向量获取单元，用于采集各所述监测点的运行状态，通过自诊断系统对各所述监控点的运行状态进行评判，得到对应的运行状态评判向量；
    状态序列生成单元，用于根据各所述监控点的运行状态评判向量生成状态序列。
  7. 根据权利要求6所述的配电终端故障诊断系统，其特征在于，所述配电终端的各功能模块包括中央处理单元、操作控制回路、通信模块、电源模块及采集模块，所述监测点设置单元包括：
    第一设置子单元，用于设置所述中央处理单元的监测点包括任务运行状态、系统参数设置状态及全球定位系统对时状态；
    第二设置子单元，用于设置所述操作控制回路的监测点包括遥测状态；
    第三设置子单元，用于设置所述通信模块的监测点包括上行通信状态和下行通信状态；
    第四设置子单元，用于设置所述电源模块的监测点包括电源状态和电池状态；
    第五设置子单元，用于设置所述采集模块的监测点包括采集单元状态和传感器探头状态。
  8. 根据权利要求5所述的配电终端故障诊断系统，其特征在于，所述训练模块具体用于：
    利用改进的交叉熵构建损失函数，利用所述损失函数对所述一维神经网络的模型进行训练；
    其中所述改进的交叉熵的计算式为：
    Figure PCTCN2022135595-appb-100004
    式中，
    Figure PCTCN2022135595-appb-100005
    为故障模型的匹配度函数区间，y i为配电终端运行状态类匹配值，α为平衡因子，α∈(0，1)，γ为调制因子，其与
    Figure PCTCN2022135595-appb-100006
    构成的权重因子用于将损失函数值降低到一个易于分类的单个样本。
  9. 一种配电终端故障诊断装置，其特征在于，包括：
    存储器，用于存储指令；其中，所述指令为可实现如权利要求1-4任意一项所述的配电终端故障诊断方法的指令；
    处理器，用于执行所述存储器中的指令。
  10. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介 质上存储有计算机程序，所述计算机程序被处理器执行时实现如权利要求1-4任意一项所述的配电终端故障诊断方法。
PCT/CN2022/135595 2021-12-02 2022-11-30 一种配电终端故障诊断方法、系统、装置及存储介质 WO2023098753A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111463158.9A CN114186184A (zh) 2021-12-02 2021-12-02 一种配电终端故障诊断方法、系统、装置及存储介质
CN202111463158.9 2021-12-02

Publications (1)

Publication Number Publication Date
WO2023098753A1 true WO2023098753A1 (zh) 2023-06-08

Family

ID=80603318

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/135595 WO2023098753A1 (zh) 2021-12-02 2022-11-30 一种配电终端故障诊断方法、系统、装置及存储介质

Country Status (2)

Country Link
CN (1) CN114186184A (zh)
WO (1) WO2023098753A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116540006A (zh) * 2023-07-06 2023-08-04 国网天津市电力公司滨海供电分公司 一种配电自动化终端故障检测装置及方法
CN117290803A (zh) * 2023-11-27 2023-12-26 深圳鹏城新能科技有限公司 一种储能逆变器远程故障诊断方法、系统及介质
CN117849512A (zh) * 2024-01-19 2024-04-09 深圳市轩宇软件开发有限公司 摩托车电气系统故障检测系统
CN118101413A (zh) * 2024-04-25 2024-05-28 中国电建集团江西省电力设计院有限公司 基于神经网络的数据通信网设备故障诊断方法及设备
CN118509261A (zh) * 2024-07-18 2024-08-16 湖南大学 天地一体化电力信息网络的零信任安全防护实现方法及系统

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114186184A (zh) * 2021-12-02 2022-03-15 广东电网有限责任公司江门供电局 一种配电终端故障诊断方法、系统、装置及存储介质
TWI800365B (zh) 2022-04-29 2023-04-21 華碩電腦股份有限公司 電源供應器的健康檢測系統及其檢測方法

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011056144A1 (en) * 2009-11-06 2011-05-12 Nanyang Polytechnic Fault signature-based case library system and method for power system trouble-shooting and recovery
CN107102633A (zh) * 2017-04-26 2017-08-29 上海金智晟东电力科技有限公司 一种配电终端故障自诊断方法及系统
CN110516742A (zh) * 2019-08-28 2019-11-29 广东电网有限责任公司 一种基于组合神经网络模型的配电终端故障判别方法及系统
KR20200014118A (ko) * 2018-07-31 2020-02-10 주식회사 비엠티 이상진단 기능을 갖는 수배전반
CN113111591A (zh) * 2021-04-29 2021-07-13 南方电网电力科技股份有限公司 基于模块化配电终端内部故障自动诊断方法、装置及设备
CN114186184A (zh) * 2021-12-02 2022-03-15 广东电网有限责任公司江门供电局 一种配电终端故障诊断方法、系统、装置及存储介质

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2011056144A1 (en) * 2009-11-06 2011-05-12 Nanyang Polytechnic Fault signature-based case library system and method for power system trouble-shooting and recovery
CN107102633A (zh) * 2017-04-26 2017-08-29 上海金智晟东电力科技有限公司 一种配电终端故障自诊断方法及系统
KR20200014118A (ko) * 2018-07-31 2020-02-10 주식회사 비엠티 이상진단 기능을 갖는 수배전반
CN110516742A (zh) * 2019-08-28 2019-11-29 广东电网有限责任公司 一种基于组合神经网络模型的配电终端故障判别方法及系统
CN113111591A (zh) * 2021-04-29 2021-07-13 南方电网电力科技股份有限公司 基于模块化配电终端内部故障自动诊断方法、装置及设备
CN114186184A (zh) * 2021-12-02 2022-03-15 广东电网有限责任公司江门供电局 一种配电终端故障诊断方法、系统、装置及存储介质

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116540006A (zh) * 2023-07-06 2023-08-04 国网天津市电力公司滨海供电分公司 一种配电自动化终端故障检测装置及方法
CN116540006B (zh) * 2023-07-06 2023-10-17 国网天津市电力公司滨海供电分公司 一种配电自动化终端故障检测装置及方法
CN117290803A (zh) * 2023-11-27 2023-12-26 深圳鹏城新能科技有限公司 一种储能逆变器远程故障诊断方法、系统及介质
CN117290803B (zh) * 2023-11-27 2024-03-26 深圳鹏城新能科技有限公司 一种储能逆变器远程故障诊断方法、系统及介质
CN117849512A (zh) * 2024-01-19 2024-04-09 深圳市轩宇软件开发有限公司 摩托车电气系统故障检测系统
CN118101413A (zh) * 2024-04-25 2024-05-28 中国电建集团江西省电力设计院有限公司 基于神经网络的数据通信网设备故障诊断方法及设备
CN118509261A (zh) * 2024-07-18 2024-08-16 湖南大学 天地一体化电力信息网络的零信任安全防护实现方法及系统

Also Published As

Publication number Publication date
CN114186184A (zh) 2022-03-15

Similar Documents

Publication Publication Date Title
WO2023098753A1 (zh) 一种配电终端故障诊断方法、系统、装置及存储介质
CN116879662B (zh) 基于数据分析的变压器故障检测方法
CN111474510B (zh) 一种非平稳输出的电压互感器的误差评估方法及系统
CN110361686B (zh) 基于多参数的电容式电压互感器故障检测方法
CN109613428A (zh) 一种能像系统及其在电机设备故障检测方法中的应用
CN102998545A (zh) 一种变压器绕组工作状态的在线监测方法
CN109284933B (zh) 一种基于数理统计的电子式互感器状态评估系统及方法
CN107270970A (zh) 高耸电力设备振动监测装置及其进行故障诊断的方法
CN117310353B (zh) 一种变电站一、二次回路通流加压故障测试方法及系统
CN108490323A (zh) 一种用于对变压器故障进行处理的系统及方法
CN109270482A (zh) 电压互感器计量准确度在线评价方法及终端设备
CN109828168A (zh) 基于核密度估计的变换器故障诊断方法
CN114255784A (zh) 一种基于声纹识别的变电站设备故障诊断方法及相关装置
CN108414877A (zh) 一种用于对变压器故障进行预警的系统及方法
CN111627199A (zh) 一种水电站大坝安全监测系统及监测方法
CN117849691B (zh) 一种电容式电压互感器多维协同运行监测预警系统和方法
CN110703149B (zh) 一种利用字符间距的变压器运行状态振声检测方法和系统
CN117233687A (zh) 一种基于历史数据的cvt初始误差评估方法、介质及终端
CN116885716A (zh) 基于边缘计算的低压配电网电能质量态势感知方法及系统
CN111123009A (zh) 一种基于scd文件的智能变电站雪崩测试方法、系统及终端设备
CN117406048B (zh) 一种变压器放电故障诊断方法及装置
CN117176199B (zh) 一种hplc通信单元故障诊断方法及装置
CN113672658B (zh) 基于复相关系数的电力设备在线监测错误数据识别方法
CN101710161A (zh) 基于自适应频率方波电压的转子接地在线监测装置及方法
CN108919041A (zh) 一种基于聚类分析的变压器绕组状态在线监测方法

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: 22900557

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE