Arc fault discrimination method based on transition characteristics and related equipment

Abstract

The invention belongs to the field of intelligent power grids, and discloses an arc fault discrimination method based on transition characteristics and related equipment.The method comprises the steps that power grid signals to be discriminated in a preset time period are collected and input into two neural network models respectively, a first discrimination model is constructed based on normal signal characteristics and arc signal characteristics, and a second discrimination model is constructed based on normal signal characteristics and arc signal characteristics; and the second discrimination model focuses on the characteristics of transition from the normal signal to the normal signal and transition from the normal signal to the arc signal. By adopting the method, the accuracy and reliability of arc fault detection are remarkably improved. According to the method, the static characteristic difference between the arc fault and the normal signal is utilized, and the dynamic characteristic of the signal transition stage is deeply analyzed, so that the problems of misjudgment and missed judgment caused by insufficient information amount and simple algorithm strategy in the traditional technology are effectively solved; the distinguishing capability of the system between normal load work and arc fault occurrence is enhanced, and a powerful guarantee is provided for safe and stable operation of a power grid.

Description

Technical Field

[0001] The present invention belongs to the field of arc fault detection, and in particular relates to an arc fault identification method based on transition characteristics and related equipment. Background Art

[0002] With the continuous progress and innovation of smart grid technology, arc fault detection technology plays an increasingly critical role in ensuring the safe and stable operation of the power grid. However, the current arc fault detection technology still faces many challenges. Traditional signal processing methods constitute the core of existing technologies. They mainly rely on low-frequency sampling technology and widely use general-purpose chips for processing; this method has exposed some significant limitations in practical applications: the collected arc information is relatively scarce, and the processing algorithm strategy is relatively simple and direct.

[0003] The constraints of existing technologies have led to a key problem: it is extremely difficult to distinguish between normal load operation and arc faults. Due to the lack of information and algorithm complexity, the system often finds it difficult to effectively capture and identify the unique characteristics of arc faults in the transition phase; this lack of recognition ability directly affects the accuracy of arc fault detection, resulting in frequent misjudgments and missed judgments, which in turn poses a potential threat to the safety and stability of the power grid.

[0004] It can be seen that the existing arc fault detection measures have poor detection accuracy because they cannot effectively capture and identify the unique characteristics of arc faults in the transition stage. Summary of the invention

[0005] The present invention provides an arc fault identification method based on transition characteristics and related equipment to solve the technical problem that existing arc fault detection measures cannot effectively capture and identify the unique characteristics exhibited by the arc fault in the transition stage, resulting in poor detection accuracy.

[0006] In order to achieve the above object, the present invention adopts the following technical solution: In a first aspect, the present invention provides an arc fault identification method based on transition characteristics, comprising: Collecting power grid signals to be identified within a preset time period; Inputting all collected power grid signals to be judged into the first judgment model and the second judgment model respectively to obtain the first judgment result and the second judgment result; Based on the first discrimination result and the second discrimination result, an arc fault discrimination result corresponding to the power grid signal to be discriminated is obtained; Wherein, the first discrimination model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; The second discriminant model adopts a neural network model constructed based on the first transition feature and the second transition feature; The first transition characteristic includes a characteristic of transitioning from a normal signal to a normal signal; and the second transition characteristic includes a characteristic of transitioning from a normal signal to an arc signal.

[0007] A further improvement of the present invention is that the training process of the first discriminant model includes: Collect normal signals and arc signals in the power grid under various working conditions; The short-time FFT feature extraction method is used to extract the features of the normal signal and the arc signal to obtain the features of the normal signal and the arc signal; Construct a binary classification data set based on normal signal features and arc signal features; The pre-built one-dimensional convolutional neural network model is trained and tested based on the binary classification data set, and the trained first discriminant model is output.

[0008] A further improvement of the present invention is that the training process of the second discriminant model includes: Collect normal signals and arc signals in the power grid under various working conditions; The short-time FFT feature extraction method is used to extract the features of the normal signal and the arc signal, and the features of the transition from the normal signal to the normal signal and the features of the transition from the normal signal to the arc signal are obtained; Constructing a transition data set based on the characteristics of transition from normal signal to normal signal and the characteristics of transition from normal signal to arc signal; The pre-built one-dimensional convolutional neural network model is trained and tested based on the transition data set, and the trained second discriminant model is output.

[0009] A further improvement of the present invention is that the characteristics of the transition from normal signal to normal signal include transition characteristics corresponding to the start and stop of equipment, normal operation, and switching of different equipment; among the characteristics of the transition from normal signal to arc signal, the normal signal characteristics account for P1 of 30~70%, and the arc signal characteristics account for P2 of 1-P1.

[0010] 5. The arc fault identification method based on transition characteristics according to claim 1, characterized in that the first identification result is 0 or 1; the second identification result is 0 or 1.

[0011] A further improvement of the present invention is that if the first discrimination result and the second discrimination result are both 1, an arc prediction result is output.

[0012] A further improvement of the present invention is that if a preset number of arc prediction results are output cumulatively within a preset judgment time, the arc fault judgment result corresponding to the power grid signal to be judged is an arc fault.

[0013] In a second aspect, the present invention provides an arc fault identification system based on transition characteristics, comprising: A data acquisition module, used to collect power grid signals to be identified within a preset time period; A data discrimination module, used for inputting all collected power grid signals to be discriminated into the first discrimination model and the second discrimination model respectively, to obtain a first discrimination result and a second discrimination result; A discrimination result output module, used for obtaining an arc fault discrimination result corresponding to the power grid signal to be discriminated based on the first discrimination result and the second discrimination result; Wherein, the first discrimination model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; The second discriminant model adopts a neural network model constructed based on the first transition feature and the second transition feature; The first transition characteristic includes a characteristic of transitioning from a normal signal to a normal signal; and the second transition characteristic includes a characteristic of transitioning from a normal signal to an arc signal.

[0014] In a third aspect, the present invention provides a device, comprising: Memory for storing computer programs; A processor is used to implement the steps of the above-mentioned arc fault identification method based on transition characteristics when executing the computer program.

[0015] In a fourth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it is used to implement the steps of the above-mentioned arc fault identification method based on transition characteristics.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The present invention provides a multi-threshold active and passive balancing control method for lithium batteries. The method collects the power grid signals to be judged within a preset time period and inputs them into two neural network models respectively. The first discrimination model is constructed based on the normal signal characteristics and the arc signal characteristics, and the second discrimination model focuses on the characteristics of the transition from normal signal to normal signal and from normal signal to arc signal. The method significantly improves the accuracy and reliability of arc fault detection. In the method, not only the static characteristic difference between arc fault and normal signal is utilized, but also the dynamic characteristics of the signal transition stage are deeply analyzed, thereby effectively overcoming the misjudgment and missed judgment problems caused by insufficient information and simple algorithm strategies in traditional technologies, enhancing the system's ability to distinguish between normal load operation and arc fault occurrence, and providing a strong guarantee for the safe and stable operation of the power grid.

[0017] Preferably, in the present invention, by collecting normal signals and arc signals under various working conditions in the power grid and using the short-time FFT feature extraction method, key features can be efficiently extracted and a high-quality binary classification data set can be constructed; the one-dimensional convolutional neural network model trained based on the data set can learn the essential differences between normal signals and arc signals, thereby improving the discrimination ability of the first discriminant model.

[0018] Preferably, in the present invention, the transition features between normal signals and arc signals are collected and extracted, especially focusing on the features of the transition from normal signals to arc signals, which helps to capture early warning signals before arc faults occur; the neural network model trained based on these transition features can more sensitively identify early signs of arc faults, thereby improving the timeliness of detection.

[0019] Preferably, in the present invention, the characteristics of the transition from normal signal to normal signal are clarified, such as equipment start and stop, normal operation, switching of different equipment, etc.; as well as the characteristics of the transition from normal signal to arc signal, and the ratio of normal signal to arc signal characteristics is designed, providing a clear goal and basis for model training; this helps the model to learn and identify these characteristics more accurately, thereby improving the discrimination accuracy.

[0020] Preferably, in the present invention, the discrimination result is simplified to 0 or 1, which is convenient for subsequent logical judgment and result output. This concise and clear representation method helps to improve the efficiency and accuracy of the discrimination process.

[0021] Preferably, in the present invention, the arc prediction result is output only when both discrimination models output 1; this dual judgment mechanism further enhances the reliability of the discrimination result and reduces the possibility of misjudgment.

[0022] Preferably, in the present invention, by setting a preset judgment time and number, the existence of an arc fault can be confirmed more robustly; this cumulative judgment mechanism helps to accurately identify arc faults in a complex and changeable power grid environment and improve the stability and safety of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 A flow chart of an arc fault identification method based on transition characteristics provided by an embodiment of the present invention; Figure 2 A transition characteristic data graph provided by an embodiment of the present invention; Figure 3 A flow chart of an arc fault identification method based on transition characteristics provided by the present invention; Figure 4 A schematic structural diagram of an arc fault identification system based on transition characteristics provided by the present invention. DETAILED DESCRIPTION

[0024] In order to further understand the content of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the embodiments are only for explaining the present invention and are not intended to limit it.

[0025] Example 1 This embodiment provides an arc fault identification method based on transition characteristics, including: Collecting power grid signals to be identified within a preset time period; Inputting all collected power grid signals to be judged into the first judgment model and the second judgment model respectively to obtain the first judgment result and the second judgment result; Based on the first discrimination result and the second discrimination result, an arc fault discrimination result corresponding to the power grid signal to be discriminated is obtained; Wherein, the first discrimination model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; The second discriminant model adopts a neural network model constructed based on the first transition feature and the second transition feature; The first transition characteristic includes a characteristic of transitioning from a normal signal to a normal signal; and the second transition characteristic includes a characteristic of transitioning from a normal signal to an arc signal.

[0026] This embodiment is further explained with reference to the accompanying drawings: like Figure 1 As shown, this embodiment provides an arc fault identification method based on transition characteristics, and the specific steps are as follows: Step 1: Data collection: Collect arc fault and normal signals in the power grid, including data of various loads operating separately or mixed, various line lengths, and various scenarios, using 50Msps data acquisition equipment to collect data; in this way, the features extracted from high sampling rate data improve the discrimination of arc fault detection.

[0027] Step 2: Feature extraction: Use the time-frequency extraction strategy short-time FFT to extract features. In this step, it can be achieved by only properly selecting some window functions and window size parameters.

[0028] (1) For the collected data, the data is basically 10ms, starting from the zero-crossing position of the 50Hz power grid voltage until the next zero-crossing. (2) Perform short-time FFT feature extraction on every 10 ms of data, with a total of 40 frequency channels and 200 points in each frequency channel (the two values ​​of 40 and 200 can be set to other values); It should be noted that short-time FFT feature extraction, that is, short-time fast Fourier transform (STFT) feature extraction, is a time-frequency analysis method widely used in the field of signal processing.

[0029] 3. Dataset construction: The extracted features are used to construct arc fault datasets, including pure normal and arc datasets (called binary classification datasets) and transition datasets from normal to arc (called transition datasets).

[0030] In this embodiment, regarding the features extracted in step 2, each sample is 10ms and has a dimension of 40×200, and maximum value scaling normalization is performed to divide normal data into one category (marked as 0) and arc data into one category (marked as 1), thereby constructing a binary classification data set; In this embodiment, a transition data set is constructed for the normal and arc data of the same working condition in the binary classification data set. The transition data set is also divided into two categories, 0 and 1, and is also scaled and normalized to the maximum value. Note that each sample time of the binary classification data set is 10ms, while each sample time of the transition data set is 100ms. Figure 2 As shown, the description is as follows: Category 0: The structure requires a transition from normal to normal characteristics. This normal to normal includes the start and stop of equipment, normal operation, switching of different equipment and other working conditions, so there are also some transition characteristics.

[0031] Category 1: The structure needs to be a transition from normal to arc characteristics, and the proportion of normal feature data is between 30% and 70% to ensure the transition characteristics of the transition data set.

[0032] Step 4: Model training: The training strategy uses general technology to build a one-dimensional convolutional neural network based on the constructed data set for training, as described below: (1) Binary classification model training: Train a one-dimensional convolutional neural network model based on a binary classification dataset.

[0033] (2) Transition model training: Train a one-dimensional convolutional neural network model based on the transition dataset.

[0034] Step 5, fault detection: Apply the two trained models to the real-time power grid signal, that is, perform arc fault detection on the power grid signal to be identified within a preset time period. The joint detection of the two models is an innovation of this embodiment.

[0035] In this embodiment, the program operation is triggered by the zero crossing point. The binary classification model will give a discrimination result every 10ms (arc is recorded as 1, normal is recorded as 0), and the transition model will give a discrimination result every 100ms (arc is recorded as 1, normal is recorded as 0); In this embodiment, an arc is determined only when both models are judged as 1, an arc is determined M times (M can be set) within a time of N ms (N can be set), and an arc is determined only when the grid signal is finally determined.

[0036] It can be seen that in this embodiment, the arc fault identification measure is mainly based on the high-frequency strategy, using high-frequency (50Msps) ADC for data acquisition, performing common time-frequency analysis to extract features based on the collected data, and training a neural network model as a basic discrimination model based on the accumulated arc and normal data sets. The basic discrimination model has a very high accuracy rate; in order to further solve a small number of missed judgments and misjudgments, another discrimination model constructed with transition data is introduced, and the two discrimination models work together; a data set from normal to arc occurrence time period is constructed, and another convolutional neural network model is trained as a transition discrimination model. Finally, the two models are jointly judged, thereby ensuring the high accuracy of arc detection.

[0037] Furthermore, the construction of transition data is to solve the scenarios that are easy to misjudge in the binary classification model (such as equipment start and stop, external interference, and multi-device switching). The construction of the binary classification data set is to identify arcs from the static numerical difference, but in fact, some background noises in different environments are different, so the construction strategy of introducing the transition data set is to enable AI to better identify the difference between arcs and normal conditions from the perspective of dynamic changes; the joint discrimination effect of static and dynamic will inevitably be more stable.

[0038] Example 2 like Figure 3 As shown, this embodiment provides an arc fault identification method based on transition characteristics, comprising the following steps: Collecting power grid signals to be identified within a preset time period; Inputting all collected power grid signals to be judged into the first judgment model and the second judgment model respectively to obtain the first judgment result and the second judgment result; Based on the first discrimination result and the second discrimination result, an arc fault discrimination result corresponding to the power grid signal to be discriminated is obtained; The first discrimination model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; The second discriminant model adopts a neural network model constructed based on the first transition feature and the second transition feature; The first transition characteristic includes a characteristic of transition from a normal signal to a normal signal; the second transition characteristic includes a characteristic of transition from a normal signal to an arc signal.

[0039] Specifically, the training process of the first discriminant model includes: Collect normal signals and arc signals in the power grid under various working conditions; The short-time FFT feature extraction method is used to extract the features of the normal signal and the arc signal to obtain the features of the normal signal and the arc signal; Construct a binary classification data set based on normal signal features and arc signal features; The pre-built one-dimensional convolutional neural network model is trained and tested based on the binary classification data set, and the trained first discriminant model is output.

[0040] The training process of the second discriminant model includes: Collect normal signals and arc signals in the power grid under various working conditions; The short-time FFT feature extraction method is used to extract the features of the normal signal and the arc signal, and the features of the transition from the normal signal to the normal signal and the features of the transition from the normal signal to the arc signal are obtained; Constructing a transition data set based on the characteristics of transition from normal signal to normal signal and the characteristics of transition from normal signal to arc signal; The pre-built one-dimensional convolutional neural network model is trained and tested based on the transition data set, and the trained second discriminant model is output.

[0041] The characteristics of the transition from normal signal to normal signal include the start and stop of the equipment, normal operation, and the transition characteristics corresponding to the switching of different equipment; among the characteristics of the transition from normal signal to arc signal, the normal signal characteristics account for P1 30~70%, and the arc signal characteristics account for P2 1-P1.

[0042] The first discrimination result is 0 or 1; the second discrimination result is 0 or 1.

[0043] If the first judgment result and the second judgment result are both 1, the arc prediction result is output.

[0044] Here, if a preset number of arc prediction results are output cumulatively within a preset judgment time, the arc fault judgment result corresponding to the power grid signal to be judged is an arc fault.

[0045] like Figure 4As shown, the present embodiment also provides an arc fault identification system based on transition characteristics, including: a data acquisition module, used to collect power grid signals to be identified within a preset time period; a data identification module, used to input all the collected power grid signals to be identified into a first identification model and a second identification model respectively, to obtain a first identification result and a second identification result; a identification result output module, used to obtain an arc fault identification result corresponding to the power grid signal to be identified based on the first identification result and the second identification result; wherein, the first identification model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; the second identification model adopts a neural network model constructed based on the first transition characteristics and the second transition characteristics; the first transition characteristics include characteristics of transition from normal signal to normal signal; the second transition characteristics include characteristics of transition from normal signal to arc signal.

[0046] The present invention also provides a device, comprising: a memory for storing a computer program; and a processor for implementing the steps of the arc fault identification method based on transition characteristics when executing the computer program.

[0047] When the processor executes the computer program, the above-mentioned arc fault identification steps based on transition characteristics are implemented, for example: collecting the grid signal to be identified within a preset time period; inputting all the collected grid signals to be identified into the first discrimination model and the second discrimination model respectively to obtain the first discrimination result and the second discrimination result; based on the first discrimination result and the second discrimination result, obtaining the arc fault discrimination result corresponding to the grid signal to be identified; wherein, the first discrimination model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; the second discrimination model adopts a neural network model constructed based on the first transition characteristic and the second transition characteristic; the first transition characteristic includes the characteristic of transition from normal signal to normal signal; the second transition characteristic includes the characteristic of transition from normal signal to arc signal.

[0048] Alternatively, when the processor executes the computer program, the functions of each module in the above-mentioned system are implemented, for example: a data acquisition module, which is used to collect the power grid signals to be judged within a preset time period; a data judgment module, which is used to input all the collected power grid signals to be judged into the first judgment model and the second judgment model respectively, to obtain the first judgment result and the second judgment result; a judgment result output module, which is used to obtain the arc fault judgment result corresponding to the power grid signal to be judged based on the first judgment result and the second judgment result; wherein, the first judgment model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; the second judgment model adopts a neural network model constructed based on first transition characteristics and second transition characteristics; the first transition characteristic includes the characteristic of transition from normal signal to normal signal; the second transition characteristic includes the characteristic of transition from normal signal to arc signal.

[0049] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of completing preset functions, and the instruction segments are used to describe the execution process of the computer program in the arc fault discrimination device based on transition characteristics. For example, the computer program can be divided into a data acquisition module, a data discrimination module and a discrimination result output module; the specific functions of each module are as follows: a data acquisition module, used to collect the power grid signals to be judged within a preset time period; a data discrimination module, used to input all the collected power grid signals to be judged into the first discrimination model and the second discrimination model respectively, to obtain the first discrimination result and the second discrimination result; a discrimination result output module, used to obtain the arc fault discrimination result corresponding to the power grid signal to be judged based on the first discrimination result and the second discrimination result; wherein, the first discrimination model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; the second discrimination model adopts a neural network model constructed based on first transition characteristics and second transition characteristics; the first transition characteristic includes the characteristic of transition from normal signal to normal signal; the second transition characteristic includes the characteristic of transition from normal signal to arc signal.

[0050] The arc fault identification device based on transition characteristics can be a computing device such as a desktop computer, a notebook, a PDA, and a cloud server. The arc fault identification device based on transition characteristics can include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above is an example of an arc fault identification device based on transition characteristics, and does not constitute a limitation on the arc fault identification device based on transition characteristics, and can include more components than the above, or a combination of certain components, or different components. For example, the arc fault identification device based on transition characteristics can also include input and output devices, network access devices, buses, etc.

[0051] The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The processor is the control center of the arc fault discrimination based on transition characteristics, and uses various interfaces and lines to connect various parts of the entire arc fault discrimination device based on transition characteristics.

[0052] The memory may be used to store the computer program and / or module, and the processor implements various functions of the arc fault identification device based on transition characteristics by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory.

[0053] The memory may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory may include a high-speed random access memory, and may also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0054] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the arc fault identification method based on transition characteristics are implemented.

[0055] If the module / unit integrated in the arc fault identification system based on transition characteristics is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.

[0056] Based on such understanding, the present invention implements all or part of the processes in the above-mentioned arc fault identification method based on transition characteristics, and can also be completed by instructing related hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above-mentioned arc fault identification method based on transition characteristics can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or preset intermediate form, etc.

[0057] The computer-readable storage medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0058] It should be noted that the content contained in the computer-readable storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable storage media do not include electrical carrier signals and telecommunication signals.

[0059] The above embodiment is only one of the implementation methods that can realize the technical solution of the present invention. The scope of protection claimed by the present invention is not limited to this embodiment alone, but also includes changes, replacements and other implementation methods that can be easily thought of by any technician familiar with the technical field within the technical scope disclosed by the present invention. In summary, the present invention provides an arc fault identification method based on transition characteristics and related equipment, which has the following advantages over existing arc fault detection measures: This discrimination method is mainly reflected in the accuracy and reliability of arc fault discrimination. This method collects the grid signal to be judged within a preset time period, and inputs it into the first discrimination model constructed based on the normal signal characteristics and the arc signal characteristics, and the second discrimination model constructed based on the transition characteristics from normal signal to normal signal and from normal signal to arc signal. This scheme can comprehensively consider the normal state, abnormal state and transition state of the grid signal, so as to more comprehensively capture the characteristics of arc faults. The first discrimination model uses the short-time FFT feature extraction method to extract features of normal signals and arc signals, and trains and tests them through a one-dimensional convolutional neural network, which improves the initial discrimination ability of arc faults. The second discrimination model further considers the transition characteristics of the signal, including the transition characteristics corresponding to the start and stop of the equipment, normal operation, switching of different equipment, etc., as well as the feature proportion when the normal signal transitions to the arc signal, which helps to more accurately identify the precursor signal of the arc fault. Finally, the results of the two discrimination models are combined. When both are judged as arc faults, the arc prediction result is output, and the arc fault judgment result corresponding to the grid signal to be judged is determined to be an arc fault only after the preset number of prediction results are accumulated within the preset judgment time. This double discrimination and cumulative judgment method effectively improves the accuracy and reliability of arc fault judgment, reduces the possibility of misjudgment and missed judgment, and provides a strong guarantee for the safe and stable operation of the power grid.

[0060] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for distinguishing arc faults based on transition characteristics, characterized in that: include: Collecting power grid signals to be identified within a preset time period; Inputting all collected power grid signals to be judged into the first judgment model and the second judgment model respectively to obtain the first judgment result and the second judgment result; Based on the first discrimination result and the second discrimination result, an arc fault discrimination result corresponding to the power grid signal to be discriminated is obtained; Wherein, the first discrimination model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; The second discriminant model adopts a neural network model constructed based on the first transition feature and the second transition feature; The first transition characteristic includes a characteristic of transitioning from a normal signal to a normal signal; and the second transition characteristic includes a characteristic of transitioning from a normal signal to an arc signal.

2. The arc fault identification method based on transition characteristics according to claim 1, characterized in that: The training process of the first discriminant model includes: Collect normal signals and arc signals in the power grid under various working conditions; The short-time FFT feature extraction method is used to extract the features of the normal signal and the arc signal to obtain the features of the normal signal and the arc signal; Construct a binary classification data set based on normal signal features and arc signal features; The pre-built one-dimensional convolutional neural network model is trained and tested based on the binary classification data set, and the trained first discriminant model is output.

3. The arc fault identification method based on transition characteristics according to claim 1, characterized in that: The training process of the second discriminant model includes: Collect normal signals and arc signals in the power grid under various working conditions; The short-time FFT feature extraction method is used to extract the features of the normal signal and the arc signal, and the features of the transition from the normal signal to the normal signal and the features of the transition from the normal signal to the arc signal are obtained; Constructing a transition data set based on the characteristics of transition from normal signal to normal signal and the characteristics of transition from normal signal to arc signal; The pre-built one-dimensional convolutional neural network model is trained and tested based on the transition data set, and the trained second discriminant model is output.

4. The arc fault identification method based on transition characteristics according to claim 3 is characterized in that: The characteristics of the transition from normal signal to normal signal include the transition characteristics corresponding to the start and stop of equipment, normal operation, and switching of different equipment; among the characteristics of the transition from normal signal to arc signal, the normal signal characteristics account for P1 of 30~70%, and the arc signal characteristics account for P2 of 1-P1.

5. The arc fault identification method based on transition characteristics according to claim 1, characterized in that: The first discrimination result is 0 or 1; the second discrimination result is 0 or 1.

6. The arc fault identification method based on transition characteristics according to claim 5, characterized in that: If the first discrimination result and the second discrimination result are both 1, an arc prediction result is output.

7. The arc fault identification method based on transition characteristics according to claim 6, characterized in that: If a preset number of arc prediction results are output cumulatively within a preset judgment time, the arc fault judgment result corresponding to the power grid signal to be judged is an arc fault.

8. An arc fault identification system based on transition characteristics, characterized in that: include: A data acquisition module, used to collect power grid signals to be identified within a preset time period; A data discrimination module, used for inputting all collected power grid signals to be discriminated into the first discrimination model and the second discrimination model respectively, to obtain a first discrimination result and a second discrimination result; A discrimination result output module, used for obtaining an arc fault discrimination result corresponding to the power grid signal to be discriminated based on the first discrimination result and the second discrimination result; Wherein, the first discrimination model adopts a neural network model constructed based on normal signal characteristics and arc signal characteristics; The second discriminant model adopts a neural network model constructed based on the first transition feature and the second transition feature; The first transition characteristic includes a characteristic of transitioning from a normal signal to a normal signal; and the second transition characteristic includes a characteristic of transitioning from a normal signal to an arc signal.

9. A device, characterized in that: include: Memory for storing computer programs; A processor is used to implement the steps of the arc fault identification method based on transition characteristics according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, it is used to implement the steps of the arc fault identification method based on transition characteristics as described in any one of claims 1 to 7.

Images