CN115906942A - Arc fault identification method, device, equipment and readable storage medium - Google Patents
Arc fault identification method, device, equipment and readable storage medium Download PDFInfo
- Publication number
- CN115906942A CN115906942A CN202211500540.7A CN202211500540A CN115906942A CN 115906942 A CN115906942 A CN 115906942A CN 202211500540 A CN202211500540 A CN 202211500540A CN 115906942 A CN115906942 A CN 115906942A
- Authority
- CN
- China
- Prior art keywords
- arc
- sample
- detection model
- fault
- current data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E10/00—Energy generation through renewable energy sources
- Y02E10/50—Photovoltaic [PV] energy
-
- 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
Landscapes
- Testing Relating To Insulation (AREA)
Abstract
The application relates to an arc fault identification method, in particular to the technical field of fault identification. The method comprises the following steps: acquiring target current data; processing the target current data through an arc detection model to obtain an arc fault probability value; the arc detection model is a machine learning model obtained by training an arc sample data set; the arc sample dataset comprises sample current data and an arc label on the sample current data; when the arc fault probability value is larger than a first fault threshold value, determining the target current data as fault arc current; when the arc fault probability value is smaller than a first fault threshold value and larger than a second fault threshold value, updating an arc sample data set according to target current data; wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met. Based on the scheme, the accuracy is high when the arc fault recognition function is realized.
Description
Technical Field
The application relates to the technical field of fault identification, in particular to an arc fault identification method, device, equipment and a readable storage medium.
Background
With the rapid development of new energy technologies, photovoltaic systems are increasingly widely used. However, due to aging of the solar photovoltaic panel, a direct current arc fault is easily generated inside the solar photovoltaic panel, and thus a fire is caused. Therefore, detection of a dc arc is necessary. The electrical characteristics of the dc arc include current, voltage, etc., and most of the existing techniques collect dc arc current data for processing and analyzing.
The existing arc detection method mainly utilizes current, voltage time domain and frequency domain characteristics to carry out analysis and judgment, generally relates to debugging of different threshold parameters of photovoltaic field environment, temperature, noise and other conditions, defines a threshold through summarizing data rules, and distinguishes whether an arc occurs through the threshold.
However, the existing arc detection method has poor anti-interference performance, easy misjudgment and low accuracy.
Disclosure of Invention
The application provides an arc fault identification method, an arc fault identification device, an arc fault identification equipment and a readable storage medium, which improve the accuracy of arc fault identification.
In one aspect, a method of arc fault identification is provided, the method comprising:
acquiring target current data;
processing the target current data through an arc detection model to obtain an arc fault probability value; the arc detection model is a machine learning model obtained by training through an arc sample data set; the arc sample dataset comprises sample current data and an arc label on the sample current data;
when the arc fault probability value is greater than a first fault threshold value, determining the target current data as a fault arc current;
when the arc fault probability value is smaller than a first fault threshold value and larger than a second fault threshold value, updating the arc sample data set according to the target current data;
wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met.
In yet another aspect, an arc fault identification apparatus is provided, the apparatus comprising:
the data acquisition module is used for acquiring target current data;
the fault detection module is used for processing the target current data through an arc detection model to obtain an arc fault probability value; the arc detection model is a machine learning model obtained by training through an arc sample data set; the arc sample dataset comprises sample current data and an arc label on the sample current data;
the fault detection module is further configured to determine the target current data as a fault arc current when the arc fault probability value is greater than a first fault threshold;
the data updating module is used for updating the arc sample data set according to the target current data when the arc fault probability value is smaller than a first fault threshold value and larger than a second fault threshold value;
wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met.
In one possible implementation, the failure detection module is further configured to,
determining the target current data as normal current data when the arc fault probability value is less than a second fault threshold.
In one possible implementation, the data update module is further configured to,
acquiring an electric arc label corresponding to the target current data;
and updating the arc sample data set according to the target current data and the arc label in the target current data to obtain the updated arc sample data set.
In a possible implementation manner, the specified condition is that a difference between the updated arc sample data set and a sample number of an arc sample data set for training the arc detection model is greater than a number threshold;
or the specified condition is that the difference between the current time and the last training time of the arc detection model is larger than a time threshold.
In one possible implementation, the apparatus further includes:
the data set acquisition module is used for acquiring the arc sample data set;
the sample prediction module is used for processing the sample current data through an arc detection model to obtain a sample prediction probability;
and the model updating module is used for carrying out back propagation updating on the arc detection model based on the sample prediction probability and the arc label on the sample current data to obtain an updated arc detection model.
In one possible implementation, the model update module is further configured to,
obtaining a loss function value according to the sample prediction probability and the electric arc label on the sample current data;
and performing back propagation through a neural network based on the loss function value to update the weight of each layer of the arc detection model, so as to obtain an updated arc detection model.
In one possible implementation, the model update module is further configured to:
carrying out back propagation updating on the arc detection model to obtain a candidate arc detection model;
quantizing the weight parameters of each layer in the candidate arc detection model into a data format with specified precision to obtain an updated arc detection model;
wherein the data format of the specified precision comprises at least one of:
a floating point number of 8-bit precision;
a floating point number of 16-bit precision;
fixed point number of 8-bit precision.
In yet another aspect, a computer device is provided, which includes a processor and a memory, where at least one instruction is stored in the memory, and the at least one instruction is loaded and executed by the processor to implement the arc fault identification method described above.
In yet another aspect, a computer-readable storage medium having stored therein at least one instruction, which is loaded and executed by a processor to implement the arc fault identification method described above, is provided.
In yet another aspect, a computer program product or computer program is provided, the computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the arc fault identification method.
The technical scheme provided by the application can comprise the following beneficial effects:
the method comprises the steps of firstly, acquiring target current data; processing the target current data through an arc detection model to obtain an arc fault probability value; the arc detection model is a machine learning model obtained by training an arc sample data set; the arc sample dataset comprises sample current data and an arc label on the sample current data; when the arc fault probability value is larger than a first fault threshold value, determining the target current data as fault arc current; when the arc fault probability value is smaller than a first fault threshold value and larger than a second fault threshold value, updating an arc sample data set according to target current data; wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met. According to the scheme, the arc fault probability value corresponding to the target current data is obtained through the arc detection model, the corresponding target current data is screened according to the range of the arc fault probability value to update the arc sample data set, then the arc detection model is retrained, the arc detection model after retraining is completed has better identification capability, and the arc fault identification accuracy is improved.
Drawings
In order to more clearly illustrate the detailed description of the present application or the technical solutions in the prior art, the drawings needed to be used in the detailed description of the present application or the prior art description will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a schematic diagram illustrating the structure of an arc fault identification system according to an exemplary embodiment.
FIG. 2 is a flow chart illustrating a method of arc fault identification in accordance with an exemplary embodiment.
FIG. 3 is a flow chart illustrating a method of arc fault identification in accordance with an exemplary embodiment.
Fig. 4 is a block diagram illustrating the structure of an arc fault identification apparatus according to an exemplary embodiment.
FIG. 5 is a block diagram illustrating the structure of a computer device according to an example embodiment.
Detailed Description
The technical solutions of the present application will be described clearly and completely with reference to the accompanying drawings, and it should be understood that the described embodiments are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It should be understood that "indication" mentioned in the embodiments of the present application may be a direct indication, an indirect indication, or an indication of an association relationship. For example, a indicates B, which may mean that a directly indicates B, e.g., B may be obtained by a; it may also mean that a indicates B indirectly, for example, a indicates C, and B may be obtained by C; it can also mean that there is an association between a and B.
In the description of the embodiments of the present application, the term "correspond" may indicate that there is a direct correspondence or an indirect correspondence between the two, may also indicate that there is an association between the two, and may also indicate and be indicated, configure and configured, and so on.
In the embodiment of the present application, "predefining" may be implemented by pre-saving a corresponding code, table or other means that can be used to indicate related information in a device (for example, including a terminal device and a network device), and the present application is not limited to a specific implementation manner thereof.
Fig. 1 is a schematic diagram illustrating the structure of an arc fault identification system according to an exemplary embodiment. The arc fault identification system includes a data processing device 110 and a current collection device 120.
Optionally, the current collecting device 120 includes a data storage, and when the current collecting device collects the target current to obtain target current data, the current data may be stored in the data storage. For example, the current collection device may be a current sensor, a current collector, and a current collection current.
Alternatively, the data processing device 110 may be a computer device with high computational power, and the data processing device is configured to analyze the collected target current data to obtain the characteristics of the target current data.
Optionally, the data processing device 110 may be a terminal device installed with current analysis software, and when the terminal device receives an instruction for analyzing current data, the terminal device may read corresponding current data from a data storage in the current collecting device 120 and analyze the current data, so as to obtain characteristics of the target current data.
Optionally, the terminal device may process the target current data through the arc detection model to obtain characteristics of the target current data.
Optionally, the data processing device 110 may also be a server installed with current analysis software, the current collecting device may be a terminal device, and after the terminal device collects the target current data, the target current data may be transmitted to the server to complete characteristic analysis of the target current data.
Optionally, the data processing device 110 and the current collecting device 120 may be connected in communication through a wired or wireless network.
Optionally, the server may be a server cluster or a distributed system formed by a plurality of physical servers, and may also be a cloud server providing technical computing services such as cloud service, a cloud database, cloud computing, a cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN, and a big data and artificial intelligence platform.
Optionally, the system may further include a management device, where the management device is configured to manage the system (e.g., manage connection states between the modules and the server, and the management device is connected to the server through a communication network. Optionally, the communication network is a wired network or a wireless network.
Optionally, the wireless network or wired network described above uses standard communication techniques and/or protocols. The network is typically the internet, but may be any other network including, but not limited to, a local area network, a metropolitan area network, a wide area network, a mobile, a limited or wireless network, a private network, or any combination of virtual private networks. In some embodiments, data exchanged over the network is represented using techniques and/or formats including hypertext markup language, extensible markup language, and the like. All or some of the links may also be encrypted using conventional encryption techniques such as secure sockets layer, transport layer security, virtual private network, internet protocol security, and the like. In other embodiments, custom and/or dedicated data communication techniques may also be used in place of, or in addition to, the data communication techniques described above.
FIG. 2 is a flow chart illustrating a method of arc fault identification in accordance with an exemplary embodiment. The method is performed by a computer device, which may be the data processing device 110 as shown in fig. 1. As shown in fig. 2, the arc fault identification method may include the steps of:
Arcing is a gas discharge phenomenon, which is a momentary spark produced by the passage of current through some insulating medium, such as air. High temperature and high heat generated by fault electric arc easily ignite the circuit insulating layer to cause the circuit to be on fire, thereby causing fire. Therefore, there is a need to identify arc faults in a timely manner to eliminate potential safety hazards.
Optionally, the current signal of the power station is collected in real time by the current collecting device and transmitted to the computer device, so as to perform the following steps.
Before the target current data is processed by the arc detection model, the arc detection model needs to be trained.
The arc detection model is a machine learning model obtained by training an arc sample data set; the arc sample dataset includes sample current data and an arc signature on the sample current data.
Further, after the arc detection model is trained, the target current data can be input into the arc detection model and processed to obtain the arc fault probability value.
Optionally, the arc fault probability value e [0,1], the first fault threshold value e [0,1].
Optionally, if the target current data is determined to be the fault arc current, the computer device may send an instruction to the alarm and disconnect the circuit to reduce the safety risk.
Optionally, an alarm function can be realized by accessing an inverter.
Optionally, when the arc fault probability value is smaller than the first fault threshold, the target current data is determined as a normal current.
That is, in the embodiment of the present application, when the arc fault probability value output by the arc detection model is smaller than the first fault threshold, the target current data may be regarded as normal current data temporarily, and no arc fault is generated.
Optionally, when the arc fault probability value is smaller than the first fault threshold and larger than the second fault threshold, the arc sample data set is updated according to the target current data.
In an actual application scenario, because the arc fault probability value is a prediction of whether an arc occurs, when the arc fault probability value does not reach the first fault threshold, an arc may occur in a circuit, and the arc fault probability value at this time is very close to the first fault threshold (that is, the arc fault probability value at this time is also a higher probability value). Meanwhile, because the arc fault probability value when the above condition occurs is a value within a certain range, a second fault threshold value can be set, and the target current data of which the corresponding arc fault probability value is between the first fault threshold value and the second fault threshold value is added into the arc sample data set, so that the arc detection model can more accurately predict the condition that the arc occurs in the circuit when the arc fault probability value does not reach the first fault threshold value. For example, when the arc detection model detects an arc fault probability value within the range, the arc fault probability value may be increased again as appropriate.
Wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met.
Optionally, the specified condition may be set according to an actual situation, for example, the arc detection model is retrained each time the arc sample data set is updated, or the arc detection model is retrained when the number of times the arc sample data set is updated reaches a certain value, or a fixed time is set every day to retrain the arc detection model.
After the arc detection model is retrained each time, the arc detection model after retraining has better recognition capability on the target current data of which the arc fault probability is between the first fault threshold and the second fault threshold, so that the arc fault recognition accuracy of the arc detection model on the input target current data is higher after retraining.
In summary, the present application first obtains target current data; processing the target current data through an arc detection model to obtain an arc fault probability value; the arc detection model is a machine learning model obtained by training an arc sample data set; the arc sample dataset comprises sample current data and an arc label on the sample current data; when the arc fault probability value is larger than a first fault threshold value, determining the target current data as fault arc current; when the arc fault probability value is smaller than a first fault threshold value and larger than a second fault threshold value, updating an arc sample data set according to target current data; wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met. According to the scheme, the arc fault probability value corresponding to the target current data is obtained through the arc detection model, the corresponding target current data is screened according to the range of the arc fault probability value to update the arc sample data set, then the arc detection model is retrained, the arc detection model after retraining is completed has better identification capability, and the arc fault identification accuracy is improved.
FIG. 3 is a flow chart illustrating a method of arc fault identification in accordance with an exemplary embodiment. The method is performed by a computer device, which may be a data processing device in an arc fault identification system as shown in fig. 1. As shown in fig. 3, the arc fault identification method may include the steps of:
Optionally, the arc detection model is deployed at the edge terminal. And detecting the current signal of the power station in real time to obtain target current data, and inputting the target current data into an arc detection model to identify the arc fault.
The arc detection model is a machine learning model obtained by training an arc sample data set; the arc sample dataset includes sample current data and an arc signature on the sample current data.
Optionally, the arc sample data set may be a set of sample current data obtained by building an arc data acquisition system including a current source, an arc generation device, and a data acquisition circuit, where the sample current data includes a normal sample when the current is at a normal value and a fault sample when an arc fault occurs.
Optionally, the sample current data further includes an arc label, which may indicate whether the sample current data is data at the time of an arc fault. The arc marking of the sample current data can be carried out manually, and the arc marking of the sample current data can also be recorded by equipment with an arc detection function. For example, when an arc fault occurs, a component in the equipment with the arc detection function is damaged and alarms, and then the chip in the equipment is used for carrying out arc marking on the sample current data.
Optionally, a sequence of points in a window in the arc data acquisition system is taken as a sample. And collecting a plurality of samples, wherein 80% of the collected samples are used as a training set, and 20% of the collected samples are used as a testing set and are respectively used for training and testing the arc detection model.
Optionally, the sample current data includes a time-domain feature signal, the sample current data may be directly input into the arc detection model to train the arc detection model, or the time-domain feature signal may be preprocessed, the time-domain feature signal is subjected to fourier transform to obtain a frequency-domain feature, and the frequency-domain feature is screened, so that the input size is reduced while the interference part is removed, and the speed and the accuracy of the arc detection model are improved.
Optionally, the arc detection model may be trained directly through the sample current data, or a pseudo quantization node may be added in the process of training the arc detection model to simulate rounding and clamping operations of a quantization effect, so that the adaptability of the arc detection model to the quantization effect is improved in the training process, and higher accuracy is obtained. In the process, all calculations (including forward and backward propagation calculations of the model and pseudo-quantization node calculations) are realized by floating point calculations, and are quantized into a real int8 model after training is completed, so that the inference speed of the model is increased.
For example, a pseudo quantization node is inserted into the weight, activation, input and output of the arc detection model, quantization perception training is performed on the arc detection model in the training process, the accuracy loss when the arc detection model is quantized to low accuracy is simulated, and the accuracy loss is transferred to a loss function, so that the arc detection model trained according to the loss function value is less influenced by weight quantization when the weight parameter is quantized, and the accuracy loss caused by model quantization is reduced. Optionally, after the arc detection model is trained, the arc detection model is quantized to a specified precision, so that the weight parameters of each layer in the arc detection model are quantized to a data format with the specified precision.
Wherein, the data format with the specified precision comprises at least one of the following:
a floating point number of 8-bit precision;
a floating point number of 16-bit precision;
fixed point number of 8-bit precision.
Optionally, the arc detection model is subjected to dynamic range quantization, and weights are converted into 8-bit precision by using a tensrflow Lite (a machine learning system), so that the size of the model is reduced to one fourth of the original size, and a calculation speed higher than that of a pure floating point number is provided.
Optionally, the arc detection model is quantified by float16, that is, weights are converted into 16-bit floating point numbers by using TensorFlow Lite, so that the size of the model can be reduced to one half of the original size.
Optionally, the arc detection model is subjected to integer quantization, and 32-bit floating point numbers (such as weight and activation output) are converted into 8-bit fixed point numbers, so that the size of the model can be reduced, and the inference speed can be increased.
GPUs can compute in this reduced precision arithmetic in a native manner, thereby achieving faster computation speeds than traditional floating point operations.
Optionally, the quantization perception training may be combined with the quantization after training, the quantization perception training is performed in the training process, and the quantization after training is performed after training.
Because the weight parameters in the trained model are quantized, and a part of information amount is lost, a quantitative perception training mode can be adopted in the training process, namely, part of information is deleted in the training process, so that the trained model can better adapt to the influence of information amount loss caused by weight parameter quantization, the size of the model is further reduced, the detection time of the model is shortened, and the prediction accuracy of the model is ensured as much as possible.
After the arc detection model is trained by the scheme, the trained arc detection model can process the input current data and generate a corresponding prediction result (namely, an arc fault probability value), and the prediction result can indicate whether the input current data is the fault arc current or not.
When the arc fault probability value is less than a second fault threshold, the target current data is determined to be normal current data.
Optionally, the first fault threshold and the second fault threshold may be adjusted according to actual conditions, so as to improve the accuracy of the arc detection model.
Wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met.
The specified condition is that a difference between the updated arc sample data set and a sample number of an arc sample data set for training the arc detection model is greater than a number threshold.
Alternatively, the specified condition is that a difference between a current time and a last training time of the arc detection model is greater than a time threshold.
Optionally, the number threshold and the time threshold may be adjusted according to actual conditions to improve the accuracy of the arc detection model. For example, when the number threshold is 50, that is, the difference between the updated arc sample data set and the number of samples in the arc sample data set for training the arc detection model is greater than 50, retraining the arc detection model through the updated arc sample data set; or, when the time threshold is two hours, retraining the arc detection model through the updated arc sample data set every two hours.
In one possible implementation, the computer device may obtain an arc label corresponding to the target current data;
and updating the arc sample data set according to the target current data and the arc label in the target current data to obtain the updated arc sample data set.
That is, after the target current data is processed by the arc detection model, when the arc fault probability value is between the first fault threshold and the second fault threshold, it represents that the probability that the target current data belongs to the fault arc current is still high, and the arc detection model needs to improve the judgment capability of the target current data, so that the computer device can put the target current data and the corresponding arc label into the arc sample data set to update the arc sample data set. At this time, the updated arc sample data contains newly added current data which is more difficult to judge, and the judgment capability of the fault arc can be obviously improved through the arc detection model trained by the updated arc sample data.
In one possible implementation, the computer device obtains a loss function value based on the sample prediction probability and an arc label on the sample current data;
based on the loss function value, performing back propagation through a neural network to update the weight of each layer of the arc detection model, and obtaining an updated arc detection model;
carrying out back propagation updating on the arc detection model to obtain a candidate arc detection model;
and quantizing the weight parameters of each layer in the candidate arc detection model into a data format with specified precision to obtain an updated arc detection model.
In other words, in the retraining process, the computer device performs quantization perception training on the arc detection model through the pseudo quantization model in the training process, and performs quantization processing on the weight parameters of each layer in the arc detection model after the training is completed.
In summary, the present application first obtains target current data; processing the target current data through an arc detection model to obtain an arc fault probability value; the arc detection model is a machine learning model obtained by training an arc sample data set; the arc sample dataset comprises sample current data and an arc label on the sample current data; when the arc fault probability value is greater than a first fault threshold value, determining the target current data as fault arc current; when the arc fault probability value is smaller than a first fault threshold value and larger than a second fault threshold value, updating an arc sample data set according to the target current data; wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met. According to the scheme, the arc fault probability value corresponding to the target current data is obtained through the arc detection model, the corresponding target current data is screened according to the range of the arc fault probability value to update the arc sample data set, then the arc detection model is retrained, the arc detection model after retraining is completed has better identification capability, and the arc fault identification accuracy is improved.
Fig. 4 is a block diagram illustrating the structure of an arc fault identification apparatus according to an exemplary embodiment. The arc fault recognition apparatus includes:
a data acquisition module 401, configured to acquire target current data;
a fault detection module 402, configured to process the target current data through an arc detection model to obtain an arc fault probability value; the arc detection model is a machine learning model obtained by training an arc sample data set; the arc sample dataset includes sample current data and an arc label on the sample current data;
the fault detection module is further used for determining the target current data as a fault arc current when the arc fault probability value is larger than a first fault threshold value;
a data updating module 403, configured to update the arc sample data set according to the target current data when the arc fault probability value is smaller than a first fault threshold and larger than a second fault threshold;
wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met.
In one possible implementation, the fault detection module is further configured to,
when the arc fault probability value is less than a second fault threshold, the target current data is determined to be normal current data.
In one possible implementation, the data update module is further configured to,
acquiring an electric arc label corresponding to the target current data;
and updating the arc sample data set according to the target current data and the arc label in the target current data to obtain the updated arc sample data set.
In one possible implementation, the specified condition is that a difference between the updated arc sample data set and a sample number of an arc sample data set for training the arc detection model is greater than a number threshold;
alternatively, the specified condition is that a difference between a current time and a last training time of the arc detection model is greater than a time threshold.
In one possible implementation, the apparatus further includes:
a data set acquisition module for acquiring the arc sample data set;
the sample prediction module is used for processing the sample current data through the arc detection model to obtain the sample prediction probability;
and the model updating module is used for carrying out back propagation updating on the arc detection model based on the sample prediction probability and the arc label on the sample current data to obtain an updated arc detection model.
In one possible implementation, the model update module is further configured to,
obtaining a loss function value according to the sample prediction probability and the electric arc label on the sample current data;
and performing back propagation through a neural network based on the loss function value to update the weight of each layer of the arc detection model, so as to obtain an updated arc detection model.
In one possible implementation, the model update module is further configured to,
carrying out back propagation updating on the arc detection model to obtain a candidate arc detection model;
quantizing each layer of weight parameters in the candidate arc detection model into a data format with specified precision to obtain an updated arc detection model;
wherein, the data format with the specified precision comprises at least one of the following:
a floating point number of 8-bit precision;
a floating point number of 16-bit precision;
fixed point number of 8-bit precision.
In summary, the present application first obtains target current data; processing the target current data through an arc detection model to obtain an arc fault probability value; the arc detection model is a machine learning model obtained by training an arc sample data set; the arc sample dataset comprises sample current data and an arc label on the sample current data; when the arc fault probability value is larger than a first fault threshold value, determining the target current data as fault arc current; when the arc fault probability value is smaller than a first fault threshold value and larger than a second fault threshold value, updating an arc sample data set according to target current data; wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met. According to the scheme, the arc fault probability value corresponding to the target current data is obtained through the arc detection model, the corresponding target current data is screened according to the range of the arc fault probability value to update the arc sample data set, then the arc detection model is retrained, the arc detection model after retraining is completed has better identification capability, and the arc fault identification accuracy is improved.
Fig. 5 shows a block diagram of a computer device 500 according to an exemplary embodiment of the present application. The computer device may be implemented as a server in the above-mentioned solution of the present application. The computer apparatus 500 includes a Central Processing Unit (CPU) 501, a system Memory 504 including a Random Access Memory (RAM) 502 and a Read-Only Memory (ROM) 503, and a system bus 505 connecting the system Memory 504 and the CPU 501. The computer device 500 also includes a mass storage device 506 for storing an operating system 509, application programs 510, and other program modules 511.
The mass storage device 506 is connected to the central processing unit 501 through a mass storage controller (not shown) connected to the system bus 505. The mass storage device 506 and its associated computer-readable media provide non-volatile storage for the computer device 500. That is, the mass storage device 506 may include a computer-readable medium (not shown) such as a hard disk or Compact disk-Only Memory (CD-ROM) drive.
Without loss of generality, the computer readable media may comprise computer storage media and communication media. Computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes RAM, ROM, erasable Programmable Read-Only Memory (EPROM), electrically Erasable Programmable Read-Only Memory (EEPROM), flash Memory or other solid state Memory technology, CD-ROM, digital Versatile Disks (DVD), or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage, or other magnetic storage devices. Of course, those skilled in the art will appreciate that the computer storage media is not limited to the foregoing. The system memory 504 and mass storage device 506 described above may be collectively referred to as memory.
The computer device 500 may also operate as a remote computer connected to a network through a network, such as the internet, in accordance with various embodiments of the present disclosure. That is, the computer device 500 may be connected to a network 508 through the network interface unit 507 connected to the system bus 505, or may be connected to another type of network or remote computer system (not shown) using the network interface unit 507.
The memory further includes at least one computer program, the at least one computer program is stored in the memory, and the central processing unit 501 executes the at least one computer program to implement all or part of the steps of the methods shown in the above embodiments.
In an exemplary embodiment, a computer readable storage medium is also provided for storing at least one computer program, which is loaded and executed by a processor to implement all or part of the steps of the above method. For example, the computer-readable storage medium may be a Read-Only Memory (ROM), a Random Access Memory (RAM), a Compact Disc Read-Only Memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, and the like.
In an exemplary embodiment, a computer program product or a computer program is also provided, which comprises computer instructions, which are stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to perform all or part of the steps of the method described in any of the embodiments of fig. 2 or fig. 3.
Other embodiments of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.
It will be understood that the present application is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the application is limited only by the appended claims.
Claims (10)
1. A method of arc fault identification, the method comprising:
acquiring target current data;
processing the target current data through an arc detection model to obtain an arc fault probability value; the arc detection model is a machine learning model obtained by training through an arc sample data set; the arc sample dataset comprises sample current data and an arc label on the sample current data;
determining the target current data as a fault arc current when the arc fault probability value is greater than a first fault threshold;
when the arc fault probability value is smaller than a first fault threshold value and larger than a second fault threshold value, updating the arc sample data set according to the target current data;
wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met.
2. The method of claim 1, further comprising:
determining the target current data as normal current data when the arc fault probability value is less than a second fault threshold.
3. The method of claim 2, wherein the updating the arc sample dataset according to the target current data comprises:
acquiring an electric arc label corresponding to the target current data;
and updating the arc sample data set according to the target current data and the arc label in the target current data to obtain the updated arc sample data set.
4. The method of claim 2, wherein the specified condition is that a difference between a number of samples of the updated arc sample data set and an arc sample data set for training the arc detection model is greater than a number threshold;
or the specified condition is that the difference between the current time and the last training time of the arc detection model is larger than a time threshold.
5. The method of any of claims 1 to 4, further comprising:
acquiring the arc sample data set;
processing the sample current data through an arc detection model to obtain a sample prediction probability;
and carrying out back propagation updating on the arc detection model based on the sample prediction probability and the arc label on the sample current data to obtain an updated arc detection model.
6. The method of claim 5, wherein the back-propagation updating the arc detection model based on the sample prediction probabilities and the arc labels on the sample current data to obtain an updated arc detection model comprises:
obtaining a loss function value according to the sample prediction probability and the electric arc label on the sample current data;
and performing back propagation through a neural network based on the loss function value to update the weight of each layer of the arc detection model, so as to obtain an updated arc detection model.
7. The method of claim 5, wherein the back-propagation updating the arc detection model to obtain an updated arc detection model, further comprises:
carrying out back propagation updating on the arc detection model to obtain a candidate arc detection model;
quantizing each layer of weight parameters in the candidate arc detection model into a data format with specified precision to obtain an updated arc detection model;
wherein the data format of the specified precision comprises at least one of:
a floating point number of 8-bit precision;
a floating point number of 16-bit precision;
fixed point number of 8 bit precision.
8. An arc fault identification device, the device comprising:
the data acquisition module is used for acquiring target current data;
the fault detection module is used for processing the target current data through an electric arc detection model to obtain an electric arc fault probability value; the arc detection model is a machine learning model obtained by training through an arc sample data set; the arc sample dataset comprises sample current data and an arc label on the sample current data;
the fault detection module is further configured to determine the target current data as a fault arc current when the arc fault probability value is greater than a first fault threshold;
the data updating module is used for updating the arc sample data set according to the target current data when the arc fault probability value is smaller than a first fault threshold value and larger than a second fault threshold value;
wherein the updated arc sample data set is used to retrain the arc detection model when specified conditions are met.
9. A computer device comprising a processor and a memory, the memory having stored therein at least one instruction that is loaded and executed by the processor to implement the arc fault identification method of any of claims 1 to 7.
10. A computer readable storage medium having stored therein at least one instruction, which is loaded and executed by a processor to implement the arc fault identification method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211500540.7A CN115906942A (en) | 2022-11-28 | 2022-11-28 | Arc fault identification method, device, equipment and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211500540.7A CN115906942A (en) | 2022-11-28 | 2022-11-28 | Arc fault identification method, device, equipment and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115906942A true CN115906942A (en) | 2023-04-04 |
Family
ID=86493045
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211500540.7A Pending CN115906942A (en) | 2022-11-28 | 2022-11-28 | Arc fault identification method, device, equipment and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115906942A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117828275A (en) * | 2024-01-10 | 2024-04-05 | 江苏神州半导体科技有限公司 | Prediction method and device for plasma chamber alignment |
-
2022
- 2022-11-28 CN CN202211500540.7A patent/CN115906942A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117828275A (en) * | 2024-01-10 | 2024-04-05 | 江苏神州半导体科技有限公司 | Prediction method and device for plasma chamber alignment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11868906B2 (en) | System and method for fault detection of components using information fusion technique | |
| AU2019413432B2 (en) | Scalable system and engine for forecasting wind turbine failure | |
| CN104951990B (en) | Power grid data processing method and equipment | |
| JP2022552449A (en) | Methods and apparatus for inspecting wind turbine blades, and equipment and storage media therefor | |
| CN109143094B (en) | Abnormal data detection method and device for power battery | |
| CN114692926A (en) | Safety evaluation method, system and equipment for vehicle battery and readable storage medium | |
| CN117110798B (en) | Fault detection method and system for intelligent power distribution network | |
| CN115906942A (en) | Arc fault identification method, device, equipment and readable storage medium | |
| CN116578869A (en) | Fault diagnosis method, fault diagnosis device and electronic device for power system | |
| CN116148656A (en) | Portable analog breaker fault detection method | |
| Hou et al. | A Federated Learning‐Based Fault Detection Algorithm for Power Terminals | |
| CN110704614B (en) | Information processing method and device for predicting user group type in application | |
| CN116298844B (en) | Semi-dynamic arrangement-based substation high-voltage circuit breaker state monitoring system and method | |
| CN114697230B (en) | Zero trust-based energy station safety monitoring system and method | |
| CN113805105A (en) | Three-phase transformer detection method and system | |
| CN113268844A (en) | Fault equipment acquisition method, device and equipment for power line | |
| CN116304846B (en) | CVT internal insulation abnormality online assessment method based on self-supervision learning | |
| CN116708135B (en) | Network service fault monitoring method and device, electronic equipment and storage medium | |
| CN116055351B (en) | Intelligent substation modeling consistency checking device and method | |
| CN117952600B (en) | New energy automobile motor evaluation method and system based on acoustic data | |
| Qiu et al. | Fan sound fault detection algorithm based on the fusion of convolutional neural network and gated recurrent unit network | |
| CN117972386A (en) | Centralized area positioning method and system for power supply quality problems | |
| CN115935166A (en) | Detection model training method and device and related equipment | |
| Sun et al. | Intelligent Fault Diagnosis of Transformer Based on Infrared Image and Mask RCNN | |
| CN117972419A (en) | Anomaly-transporter photovoltaic power generation abnormality detection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |