CN114113926B

CN114113926B - Series arc fault diagnosis method, apparatus, computer device and storage medium

Info

Publication number: CN114113926B
Application number: CN202111282675.6A
Authority: CN
Inventors: 颜天佑; 蔡蒂; 刘福来; 卢灏; 王学良
Original assignee: Guangzhou Power Supply Bureau of Guangdong Power Grid Co Ltd
Current assignee: Guangzhou Power Supply Bureau of Guangdong Power Grid Co Ltd
Priority date: 2021-11-01
Filing date: 2021-11-01
Publication date: 2024-05-28
Anticipated expiration: 2041-11-01
Also published as: CN114113926A

Abstract

The application relates to a series arc fault diagnosis method, a series arc fault diagnosis device, a series arc fault diagnosis computer device, a series arc fault diagnosis storage medium and a series arc fault diagnosis computer program product. The method comprises the following steps: obtaining a current signal to be detected in a photovoltaic power station direct current system, and mapping the current signal to be detected to a digital twin model of the photovoltaic power station; carrying out wavelet pretreatment on a current signal to be detected to obtain a pretreatment signal; extracting a time domain feature quantity to be detected of the pretreatment signal, extracting a frequency domain feature quantity to be detected of the pretreatment signal, and extracting an energy entropy feature quantity to be detected of the pretreatment signal; acquiring a three-dimensional feature space constructed in a digital twin model of a photovoltaic power station in advance; and if the fault area of the current signal to be detected in the three-dimensional characteristic space is determined based on the time domain characteristic quantity to be detected, the frequency domain characteristic quantity to be detected and the energy entropy characteristic quantity to be detected, triggering alarm operation. The method can improve the accuracy of diagnosis of the series arc faults of the direct current system of the photovoltaic power station.

Description

Series arc fault diagnosis method, apparatus, computer device and storage medium

Technical Field

The present application relates to the field of power technology, and in particular, to a series arc fault diagnosis method, apparatus, computer device, storage medium, and computer program product.

Background

With the increasing decrease of global primary energy sources, solar photovoltaic power generation systems are applied on a larger scale. The mass distributed photovoltaic power station has the characteristics of complex and various application scenes, different meteorological conditions, different grid structures of access points and the like, and is faced with a plurality of difficulties in the aspects of operation and maintenance information acquisition, decision model customization, result evaluation and the like. Along with the long-time operation of the photovoltaic system and the influence of various factors such as barren mountains, deserts and the like of the input environment, the system is extremely easy to generate direct current arc faults, so that the power generation efficiency of the system is reduced, and even the safety of the system and the whole power grid is jeopardized. The direct current arc faults of the photovoltaic system mainly comprise series arc faults caused by poor contact among photovoltaic plates, between the photovoltaic plates and a guide frame, between two junction boxes, between damaged connecting wires and the like, and parallel arc faults mainly caused by short circuit of positive and negative polarity conductors, short circuit of line grounding and the like caused by line damage. The probability of parallel arc faults is low, the current is high, the parallel arc faults are easy to detect by the protection device, the probability of series arc faults is high, the fault current is low, and the parallel arc faults are difficult to detect by the protection device.

In the conventional technology, the detection of the series direct current arc fault of the photovoltaic system is generally performed based on a single evaluation index, and the single evaluation standard comprises considering the time domain characteristic or the frequency domain characteristic of the current signal of the photovoltaic system. However, a single evaluation criterion may cause erroneous judgment or missed judgment, and there is a problem that the accuracy of the diagnosis result is low.

Disclosure of Invention

In view of the foregoing, it is desirable to provide a series arc fault diagnosis method, apparatus, computer device, computer readable storage medium, and computer program product that can improve the accuracy of the diagnosis results.

In a first aspect, the present application provides a method of series arc fault diagnosis. The method comprises the following steps:

Obtaining a current signal to be detected in a photovoltaic power station direct current system, and mapping the current signal to be detected to a digital twin model of the photovoltaic power station;

Carrying out wavelet pretreatment on the current signal to be detected to obtain a pretreatment signal;

extracting a time domain feature quantity to be detected of the pretreatment signal, extracting a frequency domain feature quantity to be detected of the pretreatment signal, and extracting an energy entropy feature quantity to be detected of the pretreatment signal;

Acquiring a three-dimensional feature space constructed in the digital twin model of the photovoltaic power station in advance;

And if the fault area of the current signal to be detected in the three-dimensional characteristic space is determined based on the time domain characteristic quantity to be detected, the frequency domain characteristic quantity to be detected and the energy entropy characteristic quantity to be detected, triggering alarm operation.

In one embodiment, the method further comprises:

acquiring normal series current data and fault series current data in a direct current system of a photovoltaic power station;

Respectively inputting the normal series current data and the fault series current data into the photovoltaic digital twin model;

Performing wavelet preprocessing on the normal series current data and the fault series current data respectively to obtain preprocessed normal current and preprocessed fault current;

Extracting time domain characteristic quantity, frequency domain characteristic quantity and energy entropy characteristic quantity of the preprocessed normal current respectively to obtain corresponding normal time domain characteristic quantity, normal frequency domain characteristic quantity and normal energy entropy characteristic quantity;

extracting time domain feature quantity, frequency domain feature quantity and energy entropy feature quantity of the preprocessed fault current respectively to obtain corresponding fault time domain feature quantity, fault frequency domain feature quantity and fault energy entropy feature quantity;

Determining a normal region based on the normal time domain feature quantity, the normal frequency domain feature quantity and the normal energy entropy feature quantity, and determining a fault region based on the fault time domain feature quantity, the fault frequency domain feature quantity and the fault energy entropy feature quantity;

Constructing a three-dimensional feature space in a photovoltaic digital twin model according to the normal region and the fault region; the three-dimensional feature space comprises a normal area, a fault area and an interference area.

In one embodiment, the extracting the time domain feature quantity to be detected of the preprocessed signal includes:

Calculating the current signal mean change rate and the current period maximum difference of the preprocessing signal;

and taking the difference between the current signal mean change rate and the current period maximum value as the time domain characteristic quantity to be detected of the preprocessing signal.

In one embodiment, the extracting the frequency domain feature to be detected of the preprocessed signal includes:

Performing Fourier transform on the preprocessing signal to obtain a frequency domain current to be detected;

determining a frequency band interval according to the sampling frequency and the sampling time of the current signal to be detected, and carrying out frequency band division on the frequency domain current to be detected according to the frequency band interval;

calculating frequency band energy corresponding to a preset target frequency band, and taking the calculated frequency band energy as a frequency domain feature quantity to be detected of the preprocessing signal; the preset target frequency band is a frequency band of which the frequency band energy comparison value of the frequency domain current corresponding to the normal series current data and the fault series current data respectively meets the difference condition.

In one embodiment, the extracting the energy entropy feature quantity to be detected of the preprocessing signal includes:

Acquiring a preset group of different random white noise sequences;

respectively adding each group of random white noise sequences into the preprocessing signals to obtain a plurality of groups of noise processing signals;

For each group of noise processing signals, performing empirical mode decomposition on the corresponding noise processing signals to obtain connotation mode components corresponding to each hierarchy respectively;

determining target connotation mode components corresponding to the same hierarchy based on connotation mode components corresponding to the same hierarchy in a plurality of groups of noise processing signals so as to obtain target connotation mode components respectively corresponding to each hierarchy;

and determining the energy entropy corresponding to the target hierarchy based on the target content modal components respectively corresponding to the target hierarchy, and taking the energy entropy corresponding to the target hierarchy as the characteristic quantity of the energy entropy to be detected of the preprocessing signal.

In one embodiment, the performing empirical mode decomposition on the corresponding noise processing signal to obtain content modal components corresponding to each hierarchy level includes:

obtaining a signal to be decomposed corresponding to the current iteration; the signal to be decomposed in the first iteration is a corresponding noise processing signal;

Acquiring all local maximum value points and local minimum value points in the signal to be decomposed, and determining an upper envelope line and a lower envelope line of the noise processing signal based on the local maximum value points and the local minimum value points;

calculating the average value of the upper envelope curve and the lower envelope curve to obtain an average value curve, and determining a response sequence according to the noise processing signal and the average value curve;

Decomposing based on the response sequence to obtain an connotation modal component of the current level;

Separating the connotation modal component of the current level from the noise processing signal to obtain a residual processing signal;

If more than two extreme values exist in the residual processing signals, the residual processing signals are used as signals to be decomposed of the next iteration, and the step of obtaining all local maximum values and minimum value points in the signals to be decomposed is returned to be continuously executed until stopping when stopping conditions are reached, so that content modal components corresponding to all layers respectively are obtained.

In a second aspect, the application also provides a series arc fault diagnosis device. The device comprises:

the acquisition module is used for acquiring a current signal to be detected in a direct current system of the photovoltaic power station and mapping the current signal to be detected to a digital twin model of the photovoltaic power station;

the processing module is used for carrying out wavelet pretreatment on the current signal to be detected to obtain a pretreatment signal;

The extraction module is used for extracting the time domain feature quantity to be detected of the pretreatment signal, extracting the frequency domain feature quantity to be detected of the pretreatment signal and extracting the energy entropy feature quantity to be detected of the pretreatment signal;

The acquisition module is also used for acquiring a three-dimensional feature space constructed in the digital twin model of the photovoltaic power station in advance;

The determining module is used for triggering alarm operation if determining a fault area of the current signal to be detected in the three-dimensional feature space based on the time domain feature quantity to be detected, the frequency domain feature quantity to be detected and the energy entropy feature quantity to be detected.

In a third aspect, the present application also provides a computer device. The computer device comprises a memory storing a computer program and a processor which when executing the computer program performs the steps of:

In a fourth aspect, the present application also provides a computer-readable storage medium. The computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of:

In a fifth aspect, the present application also provides a computer program product. The computer program product comprises a computer program which, when executed by a processor, implements the steps of:

The series arc fault diagnosis method, the series arc fault diagnosis device, the series arc fault diagnosis computer equipment, the series arc fault diagnosis storage medium and the series arc fault diagnosis computer program product firstly acquire a current signal to be detected in a direct current system of a photovoltaic power station, and map the current signal to be detected into a built digital twin model of the photovoltaic power station. Carrying out wavelet preprocessing on the current signal to be detected to obtain a preprocessed signal, and respectively extracting a time domain characteristic quantity, a frequency domain characteristic quantity and an energy entropy characteristic quantity of the preprocessed signal based on the obtained preprocessed signal, wherein the corresponding characteristic quantities are called a time domain characteristic quantity to be detected, a frequency domain characteristic quantity to be detected and an energy entropy characteristic quantity to be detected. The method comprises the steps of obtaining a three-dimensional feature space constructed in a digital twin model of a photovoltaic power station in advance, determining the position of a current signal to be detected in the three-dimensional feature space corresponding to three feature quantities according to the obtained three feature quantities of a time domain feature quantity to be detected, a frequency domain feature quantity to be detected and an energy entropy feature quantity to be detected, and if the position is in a fault area in the three-dimensional feature space, indicating that the current signal to be detected is an arc fault signal, triggering alarm operation at the moment and reminding workers. Therefore, diagnosis of the series arc faults in the direct current system of the photovoltaic power station is carried out through triple criteria of the time domain feature quantity, the frequency domain feature quantity and the energy entropy feature quantity, accuracy of diagnosis of the series arc faults of the direct current system of the photovoltaic power station is improved, and safe and stable operation of the photovoltaic power station is guaranteed.

Drawings

FIG. 1 is an environmental diagram of an application of a series arc fault diagnosis method in one embodiment;

FIG. 2 is a flow diagram of a series arc fault diagnosis method in one embodiment;

FIG. 3 is a flow diagram of three-dimensional feature space construction in one embodiment;

FIG. 4 is a time domain signal diagram of preprocessing normal current and preprocessing fault current in one embodiment;

FIG. 5 is a graph of the 10 th order EEMD decomposition of normal series current data and fault series current data for an embodiment;

FIG. 6 is a schematic diagram of a three-dimensional feature space in one embodiment;

FIG. 7 is a block diagram of a series arc fault diagnostic device in one embodiment;

fig. 8 is an internal structural diagram of a computer device in one embodiment.

Detailed Description

The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.

The series arc fault diagnosis method provided by the embodiment of the application can be applied to an application environment shown in figure 1. The application environment comprises a photovoltaic power station direct current system 102 and a computer device 104. Wherein the photovoltaic power plant dc system 102 communicates with the computer device 104 via a network. The computer device 104 may be a terminal or a server in particular. The terminal may be, but not limited to, various personal computers, notebook computers, smartphones, tablet computers and portable wearable devices, and the server may be implemented as a stand-alone server or a server cluster composed of a plurality of servers.

Taking a serial arc fault diagnosis method implemented by computer equipment as an example, the computer equipment acquires a current signal to be detected in a photovoltaic power station direct current system, maps the current signal to be detected into a digital twin model of the photovoltaic power station, performs wavelet preprocessing on the current signal to be detected to obtain a preprocessed signal, extracts a time domain feature quantity to be detected, a frequency domain feature quantity to be detected and an energy entropy feature quantity to be detected of the preprocessed signal, inputs the time domain feature quantity to be detected, the frequency domain feature quantity to be detected and the energy entropy feature quantity to be detected into a three-dimensional feature space acquired in advance, and determines whether the corresponding current signal to be detected falls into a fault area in the three-dimensional feature space, if so, triggers alarm operation.

In one embodiment, as shown in fig. 2, a method for diagnosing arc faults in series is provided, and the method is applied to the computer device in fig. 1 for illustration, and includes the following steps:

Step S202, a current signal to be detected in a photovoltaic power station direct current system is obtained, and the current signal to be detected is mapped to a digital twin model of the photovoltaic power station.

The photovoltaic power station digital twin model is obtained by mapping a three-dimensional model of equipment in a computer virtual interface based on equipment actual measurement state information in a photovoltaic power station, and when the photovoltaic power station digital twin model is constructed, the internal structure of the equipment is thinned, split and mapped into the three-dimensional model in the computer virtual interface, so that each element in the equipment can be observed in the three-dimensional model, and all-dimensional information such as data information, environmental information, personnel dynamics and the like is collected through a sensor, so that three-dimensional visualization of the model is achieved.

Specifically, a current transformer, a signal conditioning module and a data acquisition card with a certain sampling frequency are installed in a photovoltaic power station direct-current system, a computer device acquires current signals, the current signals to be detected acquired according to the certain sampling frequency are transmitted to a photovoltaic power station digital twin model, and the current signals to be detected are displayed in real time in the photovoltaic power station digital twin model.

Step S204, carrying out wavelet pretreatment on the current signal to be detected to obtain a pretreatment signal.

The wavelet preprocessing is used for filtering direct current interference signals in the current signals to be detected, and reducing the interference of noise signals.

Specifically, the obtained current signal to be detected is accompanied with the interference of alternating current noise of a direct current signal and a low-frequency signal, so that the computer equipment firstly performs wavelet preprocessing on the obtained current signal to be detected in data processing software, and filters the direct current signal interference signal to obtain a preprocessed signal.

Step S206, extracting the time domain feature quantity to be detected of the pre-processing signal, extracting the frequency domain feature quantity to be detected of the pre-processing signal, and extracting the energy entropy feature quantity to be detected of the pre-processing signal.

The time domain feature quantity to be detected is a feature quantity obtained by calculating a preprocessing signal in a time domain; the frequency domain characteristic quantity to be detected is a characteristic quantity obtained by calculating a preprocessing signal in a frequency domain; the time domain feature quantity to be detected is a feature quantity obtained by calculating a preprocessing signal in the time domain; the energy entropy feature quantity to be detected represents the degree of energy confusion of the preprocessing signal.

Specifically, after the preprocessing signal is obtained, the computer equipment calculates the characteristic quantity of the preprocessing signal in the time domain, the characteristic quantity of the preprocessing signal in the frequency domain and the energy entropy characteristic quantity respectively, so as to obtain the corresponding time domain characteristic quantity to be detected, the frequency domain characteristic quantity to be detected and the energy entropy characteristic quantity to be detected.

And step S208, acquiring a three-dimensional feature space constructed in the digital twin model of the photovoltaic power station in advance.

The three-dimensional feature space comprises a normal region, a fault region and an interference region, and three dimensions of the three-dimensional feature space are a time domain feature, a frequency domain feature and an energy entropy feature respectively.

Specifically, according to the sampling frequency of the current signal to be detected, the computer equipment acquires a corresponding three-dimensional feature space from the digital twin model of the photovoltaic power station, so that fault detection is carried out on the current signal to be detected.

Step S210, if a fault area where the current signal to be detected falls in the three-dimensional feature space is determined based on the time domain feature quantity to be detected, the frequency domain feature quantity to be detected and the energy entropy feature quantity to be detected, an alarm operation is triggered.

The alarming operation is used for reminding operators of arc faults of the photovoltaic power station direct-current system, and the alarming operation can be carried out by flashing an alarm lamp and accompanying a prompt tone, so that the method is not limited.

Specifically, according to the obtained three characteristic parameters of the time domain characteristic quantity to be detected, the frequency domain characteristic quantity to be detected and the energy entropy characteristic quantity to be detected, the computer equipment can determine the position of the current signal to be detected in the three-dimensional characteristic space, the computer equipment judges whether the position is in a fault area in the three-dimensional characteristic space, if the position is in the fault area in the three-dimensional characteristic space, the current signal to be detected is an arc fault signal, and an arc fault occurs in a photovoltaic power station direct current system when the current signal to be detected is collected, the computer equipment triggers an alarm operation in a digital twin model of the photovoltaic power station, and accordingly the personnel is reminded to execute corresponding measures on the photovoltaic power station.

In the series arc fault diagnosis method, a current signal to be detected in a direct current system of a photovoltaic power station is firstly obtained, and the current signal to be detected is mapped into a built digital twin model of the photovoltaic power station. Carrying out wavelet preprocessing on the current signal to be detected to obtain a preprocessed signal, and respectively extracting a time domain characteristic quantity, a frequency domain characteristic quantity and an energy entropy characteristic quantity of the preprocessed signal based on the obtained preprocessed signal, wherein the corresponding characteristic quantities are called a time domain characteristic quantity to be detected, a frequency domain characteristic quantity to be detected and an energy entropy characteristic quantity to be detected. The method comprises the steps of obtaining a three-dimensional feature space constructed in a digital twin model of a photovoltaic power station in advance, determining the position of a current signal to be detected in the three-dimensional feature space corresponding to three feature quantities according to the obtained three feature quantities of a time domain feature quantity to be detected, a frequency domain feature quantity to be detected and an energy entropy feature quantity to be detected, and if the position is in a fault area in the three-dimensional feature space, indicating that the current signal to be detected is an arc fault signal, triggering alarm operation at the moment and reminding workers. Therefore, diagnosis of the series arc faults in the direct current system of the photovoltaic power station is carried out through triple criteria of the time domain feature quantity, the frequency domain feature quantity and the energy entropy feature quantity, accuracy of diagnosis of the series arc faults of the direct current system of the photovoltaic power station is improved, and safe and stable operation of the photovoltaic power station is guaranteed.

In one embodiment, as shown in fig. 3, the method further comprises:

And step S302, acquiring normal series current data and fault series current data in a direct current system of the photovoltaic power station.

The normal series current data are current data acquired when the photovoltaic power station direct current system does not have arc faults, and the fault series current data are current data acquired when the photovoltaic power station direct current system has arc faults.

Specifically, a data acquisition card with a specified sampling frequency is used for acquiring series current when an arc fault does not occur in a direct current system of a photovoltaic power station, and the series current is normal series current data; and collecting serial current when an arc fault occurs in a direct current system of the photovoltaic power station by using a data collection card with the same sampling frequency, wherein the current is fault serial current data, and the normal serial current data and the fault serial current data are obtained by computer equipment.

For example, a data acquisition card with the sampling frequency of 1MHz is used for acquiring normal series current data when an arc fault does not occur in a direct current system of a photovoltaic power station; and the data acquisition card with the sampling frequency of 1MHz is also used for acquiring fault series current data when the direct current system of the photovoltaic power station has arc faults.

Step S304, the normal series current data and the fault series current data are respectively input into a photovoltaic digital twin model.

Specifically, the computer equipment inputs the obtained normal series current data and fault series current data which are collected according to the specific sampling frequency into the corresponding photovoltaic digital twin model.

In one embodiment, the computer device inputs the normal series current data and the fault series current data with the sampling frequency of f ₀ into the photovoltaic digital twin model respectively, and f ₀ may be 1MHz.

Step S306, wavelet preprocessing is respectively carried out on the normal series current data and the fault series current data, and preprocessed normal current and preprocessed fault current are obtained.

The wavelet preprocessing is used for filtering direct current interference signals in normal series current data and fault series current data, and reducing interference of noise signals on the normal series current data and the fault series current data.

Specifically, the computer device performs wavelet preprocessing on the normal series current data and the fault series current data respectively to obtain corresponding preprocessed normal current and preprocessed fault current.

In one embodiment, the normal series current data with the sampling frequency f ₀ and the sampling time t ₀ is subjected to wavelet pretreatment to obtain a pretreated normal current, and the fault series current data with the sampling frequency f ₀ MHz and the sampling time t ₀ ms is subjected to wavelet pretreatment to obtain a pretreated fault current, wherein the f ₀ can be 1MHz, and the t ₀ can be 200ms. As shown in fig. 4, which is a time domain signal diagram of the preprocessing normal current and the preprocessing fault current, in the diagram, "normal" indicates the preprocessing normal current when the photovoltaic system is in a normal state, and "series arc" indicates the preprocessing fault current when the photovoltaic system has an arc fault, it can be seen that the amplitude of the current waveform is not greatly changed when the photovoltaic system is in a normal state, and when the photovoltaic system has an arc fault, the time domain diagram starts to have a large amplitude of up-down jitter.

Step S308, extracting the normal time domain feature quantity of the preprocessed normal current and the fault time domain feature quantity of the preprocessed fault current respectively.

Specifically, the computer device extracts a normal time domain feature quantity of the pre-processing normal current and a fault time domain feature quantity of the pre-processing fault current, respectively.

In one embodiment, the preprocessing normal current and the preprocessing fault current are both time-domain currents, the normal time-domain feature quantity comprises a mean change rate and a current period worst value, and the fault time-domain feature quantity also comprises a mean change rate and a current period worst value. The calculation formula of the mean change rate I _var is as follows:

Wherein N is the number of current data in a single time window T, and I (N) ^T is the current amplitude of each point in the time window T; i (n) ^T+1 is the current amplitude at each point in the next time window (T+1); i (n) can be the current amplitude of each point in the window T, or the current amplitude of each point in the time window (T+1), or the average value of the current amplitudes in the time window T and the time window (T+1).

The calculation formula of the current period worst value I _diff is:

Wherein, The maximum value and the minimum value of the current amplitude of all sampling points in the same time window T are respectively.

For preprocessing normal current and preprocessing fault current, according to a calculation formula of a mean change rate and a worst value of a current period, a normal time domain characteristic quantity of the preprocessing normal current and a fault time domain characteristic quantity of the preprocessing fault current can be obtained respectively.

Step S310, extracting a normal frequency domain feature of the pre-processing normal current and a fault frequency domain feature of the pre-processing fault current, respectively.

Specifically, the computer device extracts a normal frequency domain feature quantity of the pre-processing normal current and a fault frequency domain feature quantity of the pre-processing fault current, respectively.

In one embodiment, the sampling frequency of the normal series current data and the fault series current data is f ₀, and the sampling time is t ₀. The computer equipment performs fast Fourier transform on the preprocessed normal current and the preprocessed fault current which are obtained after the wavelet preprocessing of the normal serial current data and the fault serial current data respectively to obtain a normal frequency domain current and a fault frequency domain current, the computer equipment equally divides the whole frequency band of the normal frequency domain current and the fault frequency domain current into n frequency bands (n is a positive integer) respectively by taking K (K is a frequency value) as an interval, and n frequency bands corresponding to the normal frequency domain current and the fault frequency domain current are respectively marked as f ₁、f₂, … and fn. And calculating the square sum of the amplitude values of all frequency points corresponding to each frequency segment corresponding to the normal frequency domain current and the fault frequency domain current as the energy of the frequency segment, wherein the energy of each frequency segment of the normal frequency domain current and the fault frequency domain current is respectively marked as E _{0 i} and E _{arc i} (i=1, 2, …, n).

Calculating the energy ratio beta _i of the normal frequency domain current and the fault frequency domain current corresponding to each frequency band, namely:

β_i＝E_{arc i}/E_{0 i}

And analyzing the energy ratio result, and taking the energy E _{0 i} and E _{arc i} (i=1, 2, …, n) with obvious frequency ranges as corresponding normal frequency domain characteristic quantity and fault frequency domain characteristic quantity.

Such as: the frequency band f ₀ can be 1mhz, the frequency band t ₀ can be 200ms, the whole frequency band of the normal frequency domain current and the fault frequency domain current is 0-500kHz, k=50 kHz is used as an interval, the whole frequency band of the normal frequency domain current and the fault frequency domain current is divided into n=10 frequency bands, and the 10 frequency bands corresponding to the normal frequency domain current and the fault frequency domain current are respectively marked as f ₁、f₂、…、f₁₀. And calculating the square sum of the amplitude values of all frequency points corresponding to each frequency segment corresponding to the normal frequency domain current and the fault frequency domain current as the energy of the frequency segment, wherein the energy of each frequency segment of the normal frequency domain current and the fault frequency domain current is respectively marked as E _{0 i} and E _{arc i} (i=1, 2, …, 10).

β_i＝E_{arc i}/E_{0 i}

And analyzing the energy ratio result, and taking the energy E _{0 i} and E _{arc i} (i=1, 2, …, 10) with obvious frequency ranges as corresponding normal frequency domain characteristic quantity and fault frequency domain characteristic quantity.

In this embodiment, a large number of experiments show that the energy in the low frequency range 0-50 kHz and 50-100 kHz is more obvious than the energy in the normal state after the arc fault occurs, so the energy E _{0 1}、E_{0 2} in the frequency range corresponding to f ₁ and f ₂ is used as the normal frequency domain characteristic quantity, and the energy E _{arc 1}、E_{arc 2} is used as the fault frequency domain characteristic quantity.

Step S312, extracting the normal energy entropy characteristic quantity of the preprocessed normal current and the fault energy entropy characteristic quantity of the preprocessed fault current respectively.

Specifically, the computer device extracts a normal energy entropy feature quantity of the preprocessing normal current and a fault energy entropy feature quantity of the preprocessing fault current, respectively.

In one embodiment, p (p is a positive integer) sets of random white noise are added to the preprocessing normal current and the preprocessing fault current respectively, and each time a set of random white noise is added, for example, p sets of random white noise with equal variance are sequentially added to the preprocessing normal current, a first set of random white noise sequences with zero mean value can be added to the first set of random white noise sequences, so as to obtain a first set of normal noise processing signals, and the like, so as to obtain p sets of normal noise processing signals. Similarly, p groups of random white noise with equal variance are sequentially added into the preprocessing fault current, a group of random white noise sequences with zero mean value can be added into the first group, a first group of fault noise processing signals are obtained, and p groups of fault noise processing signals are obtained in the same way.

And respectively carrying out empirical mode (EMPIRICAL MODE DECOMPOSITION, EMD) decomposition on each group of normal noise processing signals and each group of fault noise processing signals to obtain normal connotation mode components (INTRINSIC MODE FUNCTIONS, IMF) respectively corresponding to different levels of each group of normal noise processing signals and fault connotation mode components respectively corresponding to different levels of each group of fault noise processing signals. And calculating the average value of the normal connotation modal components of the same level in the p groups of normal noise processing signals to obtain the normal target connotation modal components corresponding to each level, and similarly, calculating the average value of the fault connotation modal components of the same level in the p groups of fault noise processing signals to obtain the fault target connotation modal components corresponding to each level.

According to the obtained normal target connotation modal component and fault target connotation modal component corresponding to each hierarchy, the cosine similarity S _j between the fault target connotation modal component and the normal target connotation modal component of each hierarchy is calculated, and the calculation formula is as follows:

Wherein j is the order of the target connotation modal component, F _j (t) is the j-order normal target connotation modal component, and G _j (t) is the j-order fault target connotation modal component; n is the data point quantity.

According to EEMD (Ensemble Empirical Mode Decomposition ) algorithm, defining energy entropy to represent probability of occurrence of certain information, and solving the energy entropy of the corresponding layer of the target connotation modal component as normal energy entropy characteristic quantity and fault energy entropy characteristic quantity when cosine similarity is smaller than 0.15, wherein the calculation formula of the energy entropy H is as follows:

Wherein H _j is the energy entropy of the j-order target connotation modal component, m is the total order of the target connotation modal components, and p _j is the proportion of the energy of the j-th target connotation modal component to the energy of the whole target connotation modal component.

The method for respectively carrying out empirical mode decomposition on each group of normal noise processing signals and each group of fault noise processing signals comprises the following steps:

1) Obtaining signals to be decomposed corresponding to the current iteration, wherein the signals to be decomposed of the first iteration are a first group of normal noise processing signals and a first group of fault noise processing signals;

2) Determining all local maximum value points and local minimum value points of a first group of normal noise processing signals, determining all local maximum value points and local minimum value points of a first group of fault noise processing signals, connecting all local maximum value points and local minimum value points of the first group of normal noise processing signals by a cubic spline interpolation method to form an upper envelope line and a lower envelope line of the first group of normal noise processing signals, and connecting all local maximum value points and local minimum value points of the first group of fault noise processing signals by a cubic spline interpolation method to form an upper envelope line and a lower envelope line of the first group of fault noise processing signals;

3) Calculating the average value of the upper envelope curve and the lower envelope curve of the first group of normal noise processing signals to obtain a first group of normal average value curves, calculating the average value of the upper envelope curve and the lower envelope curve of the first group of fault noise processing signals to obtain a first group of fault average value curves, differentiating the first group of normal noise processing signals from the first group of normal average value curves to obtain a first group of normal response sequences, and differentiating the first group of fault noise processing signals from the first group of fault average value curves to obtain a first group of fault response sequences;

4) Judging whether the first group of normal response sequences and the first group of fault response sequences meet IMF conditions or not respectively, wherein the IMF conditions are that the number of local extreme points and zero crossing points of the corresponding sequences is equal or at most one phase difference in the whole time range, and the average value of the envelope determined by the local maximum points and the envelope determined by the local minimum points is zero;

5) If the first group of normal response sequences and the first group of fault response sequences meet the IMF conditions, corresponding normal connotation modal components and fault connotation modal components can be obtained according to the IMF conditions, and the corresponding normal connotation modal components and fault connotation modal components are separated from corresponding normal noise processing signals and fault noise processing signals respectively, so that normal residual processing signals and fault residual processing signals are obtained;

6) If more than two extreme values exist in the normal residual processing signal and the fault residual processing signal respectively, the normal residual processing signal and the fault residual processing signal are used as signals to be decomposed of the next iteration, and the step 1) is returned to be continuously executed until the number of the extreme values existing in the normal residual processing signal and the fault residual processing signal respectively is not more than 2, and decomposition is stopped to obtain normal connotation modal components and fault connotation modal components of layers corresponding to each decomposition;

7) And if the first group of normal response sequences and the first group of fault response sequences do not meet the IMF condition, respectively taking the first group of normal response sequences and the first group of fault response sequences as a normal noise processing signal and a fault noise processing signal, and returning to the step 1) for continuous execution.

In a specific embodiment, as shown in fig. 5, an connotation mode component diagram of normal series current data and fault series current data after 10-order EEMD decomposition is shown, where the ordinate current amplitude variation range of the normal series current data is [ -0.1,0.1] a, and the ordinate variation of the fault series current data is divided into [ -0.2,0.2] a. After the cosine similarity of each order fault target content modal component and the normal target content modal component is calculated, the cosine similarity of the fifth order (IMF 5) fault series current data and the sixth order (IMF 6) fault series current data and the normal series current data is found to be lower and lower than 0.15, which indicates that the energy changes of the fifth order and the sixth order content modal components before and after the fault occurs are obvious, so that the energy entropy of the fifth order and the sixth order normal target content modal components can be used as normal energy entropy characteristic quantity, and the energy entropy of the fifth order and the sixth order fault target content modal components can be used as fault energy entropy characteristic quantity.

And step S314, determining a normal region based on the normal time domain feature quantity, the normal frequency domain feature quantity and the normal energy entropy feature quantity, and determining a fault region based on the fault time domain feature quantity, the fault frequency domain feature quantity and the fault energy entropy feature quantity.

Specifically, according to the normal time domain feature quantity, the normal frequency domain feature quantity and the normal energy entropy feature quantity of the normal series current data, which are respectively located at the original points, the computer device can determine the position of the normal series current data in the three-dimensional feature space, the position where the normal series current data falls is the normal area, and similarly, according to the fault time domain feature quantity, the fault frequency domain feature quantity and the fault energy entropy feature quantity of the fault series current data, which are respectively located at the original points, the computer device can determine the position of the fault series current data in the three-dimensional feature space, and the position where the fault series current data falls is the fault area.

In one embodiment, the position of both current signals in the three-dimensional feature space of the normal series current data or the fault series current data can be determined by the following formula:

Wherein d _T is the distance of the current signal in the time domain feature direction in the three-dimensional feature space, and is determined by two time domain feature quantities of the mean change rate and the period maximum difference; d _F is the distance between the current signal in the frequency characteristic direction in the three-dimensional characteristic space, and is determined by the energy corresponding to the frequency band with obvious energy change of the current signal; d _H is the distance between the current signal and the energy entropy characteristic direction in the three-dimensional characteristic space, and is determined by the energy entropy of the target connotation modal component when the cosine degree of the current signal is less than 0.15; o is the feature origin in each direction.

Step S316, constructing a three-dimensional feature space in the photovoltaic digital twin model according to the normal region and the fault region; the three-dimensional feature space comprises a normal area, a fault area and an interference area.

In particular, based on the determined normal and fault regions, the computer device may construct a three-dimensional feature space in the photovoltaic digital twin model, with other regions outside the normal and fault regions referred to as disturbance regions.

In one embodiment, as shown in fig. 6, T _m、F_m and H _m are boundaries of normal regions of the three-dimensional feature space, and T _n、F_n and H _n are boundaries of fault regions of the three-dimensional feature space. The smaller the values of Δt, Δf and Δh, the more sensitive but less reliable the determination of the fault, whereas the larger the values of Δt, Δf and Δh, the higher but less sensitive the determination of the fault. Through multiple experiments, it is determined that the time domain feature criterion is T _n＝T_m +Δt=0.76, the frequency domain feature criterion is F _n＝F_m +Δf=6.48, the energy entropy feature criterion is H _n＝H_m +Δh=7.55, and for different photovoltaic power stations, each feature criterion can be properly adjusted according to actual measurement values, and the method is not limited.

In the above embodiment, the normal series current data and the fault series current data in the photovoltaic power station direct current system are obtained, the normal series current data and the fault series current data are input into the photovoltaic digital twin model, the normal series current data and the fault series current data are subjected to wavelet preprocessing to obtain corresponding preprocessed normal current and preprocessed fault current, the normal time domain feature quantity, the normal frequency domain feature quantity and the normal energy entropy feature quantity are extracted for the corresponding preprocessed normal current, the fault time domain feature quantity, the fault frequency domain feature quantity and the fault energy entropy feature quantity are extracted for the corresponding preprocessed fault current, a normal area is determined according to the normal time domain feature quantity, the normal frequency domain feature quantity and the normal energy entropy feature quantity, and the fault area is determined according to the fault time domain feature quantity, the fault frequency domain feature quantity and the fault energy entropy feature quantity, so that the three-dimensional feature space is constructed in the photovoltaic digital twin model. By constructing a three-dimensional feature space in the photovoltaic digital twin model and determining feature criteria, whether the current signal to be detected is an arc fault signal or not can be accurately judged, and fault diagnosis results can be displayed in the twin model more clearly and conveniently so as to guide operation and maintenance.

In one embodiment, extracting the time domain feature to be detected of the pre-processed signal includes: calculating the current signal mean change rate and the current period maximum difference of the preprocessing signal; and taking the difference between the average change rate of the current signal and the current period maximum value as a time domain characteristic quantity to be detected of the preprocessing signal.

The preprocessing signal is a time domain signal, the time domain feature quantity to be detected comprises a mean change rate and a current period worst value, and a calculation formula of the mean change rate I _var is as follows:

The calculation formula of the current period worst value I _diff is:

Specifically, the computer equipment calculates the current signal mean change rate and the current period worst value difference of the preprocessing signal according to a calculation formula of the mean change rate and the current period worst value, and takes the current signal mean change rate and the current period worst value difference as the time domain feature quantity to be detected of the preprocessing signal.

In this embodiment, when the dc system of the photovoltaic power station is in a normal state, the amplitude of the collected series current data waveform is not greatly changed, and when the dc system of the photovoltaic power station has an arc fault, the collected fault series current data waveform is greatly dithered, so that the average value change rate of the current signal and the current period maximum value difference are taken as the time domain feature quantity to be detected, and the abrupt change characteristic of the current amplitude before and after the arc fault of the dc system of the photovoltaic power station can be further reflected.

In one embodiment, extracting the frequency domain feature to be detected of the pre-processed signal includes: performing Fourier transform on the preprocessed signals to obtain frequency domain current to be detected; determining a frequency band interval according to the sampling frequency and the sampling time of the current signal to be detected, and carrying out frequency band division on the frequency domain current to be detected according to the frequency band interval; calculating frequency band energy corresponding to a preset target frequency band, and taking the calculated frequency band energy as a frequency domain feature quantity to be detected of the preprocessing signal; the preset target frequency band is a frequency band of which the frequency band energy comparison value of the frequency domain current corresponding to the normal series current data and the fault series current data respectively meets the difference condition.

Specifically, the computer equipment performs fourier transform on the preprocessing signal to obtain a frequency domain current to be detected, determines a frequency band interval according to the sampling frequency and the sampling time of the current signal to be detected, performs frequency band division on the frequency domain current to be detected according to the frequency band interval, calculates frequency band energy corresponding to a preset target frequency band, and takes the calculated frequency band energy as a frequency domain feature quantity to be detected of the preprocessing signal.

In one embodiment, the sampling frequency of the current signal to be detected is f ₀, the sampling time is t ₀, the computer device performs fast fourier transform on the preprocessed signal obtained after the wavelet preprocessing of the current signal to be detected to obtain the frequency domain current to be detected, the whole frequency band in the frequency domain current to be detected is divided into n (n is a positive integer) frequency bands with K as a frequency value as an interval, and the n frequency bands are respectively marked as f ₁、f₂、…、f_n. The computer equipment calculates the square sum of the amplitude values of all frequency points corresponding to each frequency segment of the frequency domain current to be detected as the energy of the frequency segment, and according to a large number of experiments of normal frequency domain current and fault frequency domain current, the energy in a certain frequency band after the occurrence of an arc fault is more obvious than the energy in a normal state, so that for a current signal to be detected, the energy in the frequency band with obvious energy change is used as the characteristic quantity of the frequency domain to be detected.

For example, f ₀ may be 1mhz, t ₀ may be 200ms, and the computer device performs fast fourier transform on a preprocessed signal obtained after performing wavelet preprocessing on a current signal to be detected to obtain a frequency domain current to be detected, and equally divide the frequency domain current to be detected into 10 frequency bands with a frequency band of 0-500kHz being 50kHz, where each of the 10 frequency bands is denoted as f ₁、f₂、…、f₁₀. The computer equipment calculates the square sum of the amplitude values of all frequency points corresponding to each frequency segment of the frequency domain current to be detected as the energy of the frequency segment, and according to a large number of experiments of normal frequency domain current and fault frequency domain current, the energy in the low frequency segments of 0-50 kHz and 50-100 kHz after the occurrence of arc faults is more obvious than the energy in the normal state, so that for a current signal to be detected, the energy of the frequency segment corresponding to f ₁ and f ₂ is used as the frequency domain characteristic quantity to be detected.

In this embodiment, according to the frequency band determined by a large amount of experimental data, the energy of the current signal to be detected in the frequency band is calculated as the frequency domain feature quantity to be detected, so that the difference of the frequency domain features of the photovoltaic power station direct current system before and after the arc fault can be more accurately reflected.

In one embodiment, extracting the energy entropy feature to be detected of the pre-processed signal includes: acquiring a preset group of different random white noise sequences; each group of random white noise sequences are respectively added into the preprocessing signals to obtain a plurality of groups of noise processing signals; for each group of noise processing signals, performing empirical mode decomposition on the corresponding noise processing signals to obtain connotation mode components corresponding to each hierarchy respectively; determining target connotation mode components corresponding to the same hierarchy based on connotation mode components corresponding to the same hierarchy in the plurality of groups of noise processing signals so as to obtain target connotation mode components respectively corresponding to each hierarchy; and determining the energy entropy corresponding to the target hierarchy based on the target content modal components respectively corresponding to the target hierarchy, and taking the energy entropy corresponding to the target hierarchy as the characteristic quantity of the energy entropy to be detected of the preprocessing signal.

Specifically, the computer device obtains a preset group of different random white noise sequences, adds each group of random white noise sequences to the preprocessing signals respectively to obtain a plurality of groups of noise processing signals, carries out empirical mode decomposition on the corresponding noise processing signals for each group of noise processing signals to obtain connotation mode components corresponding to each hierarchy respectively, determines target connotation mode components corresponding to the same hierarchy based on connotation mode components corresponding to the same hierarchy in the plurality of groups of noise processing signals to obtain target connotation mode components corresponding to each hierarchy respectively, determines energy entropy corresponding to the target hierarchy based on the target connotation mode components corresponding to each hierarchy respectively, and takes the energy entropy corresponding to the target hierarchy as a characteristic quantity of energy entropy to be detected of the preprocessing signals.

In one embodiment, p groups of random white noise with equal variance are added to the preprocessed signals, one group is added at a time, the first group can be a group of random white noise sequences with zero mean value, and finally p groups of noise processed signals are obtained. And performing empirical mode decomposition on each group of noise processing signals to obtain content mode components corresponding to different layers of each group of noise processing signals. And calculating the mean value of the connotation modal components of the same layer in the p groups of noise processing signals to obtain the target connotation modal component corresponding to each layer.

According to EEMD algorithm, defining energy entropy to represent probability of occurrence of certain information, and according to target content modal components corresponding to each layer respectively, calculating energy entropy corresponding to the target layer, wherein the calculation formula is as follows:

For example, when p is specifically 10 and the order m is also 10, according to a large number of experiments of normal series current data and fault series current data, it is known that the energy entropy of the target connotation modal components of the fifth order (IMF 5) and the sixth order (IMF 6) can be used as the frequency domain feature quantity to be detected of the pretreatment signal corresponding to the current signal to be detected.

In the embodiment, the target content modal component of the optimal hierarchy is determined according to a large amount of experimental data, the energy entropy of the current signal to be detected in the hierarchy is calculated to serve as the characteristic quantity of the energy entropy to be detected, the difference of the energy entropy characteristics of the photovoltaic power station direct current system before and after the arc fault can be more accurately reflected, and the problem that the wavelet decomposition algorithm with more applied areas in the current field is not adaptive is solved by using the EEMD algorithm.

In one embodiment, performing empirical mode decomposition on the corresponding noise processing signal to obtain content modal components corresponding to each hierarchy, respectively, includes: obtaining a signal to be decomposed corresponding to the current iteration; the signal to be decomposed in the first iteration is a corresponding noise processing signal; acquiring all local maximum points and local minimum points in a signal to be decomposed, and determining an upper envelope curve and a lower envelope curve of a noise processing signal based on the local maximum points and the local minimum points; calculating the average value of the upper envelope curve and the lower envelope curve to obtain an average value curve, and determining a response sequence according to the noise processing signal and the average value curve; decomposing based on the response sequence to obtain an connotation modal component of the current level; separating the connotation modal component of the current level from the noise processing signal to obtain a residual processing signal; if more than two extreme values exist in the residual processing signals, the residual processing signals are used as signals to be decomposed of the next iteration, and the step of returning to the step of obtaining all local maximum values and minimum value points in the signals to be decomposed is continuously executed until stopping when stopping conditions are reached, and connotation modal components corresponding to all layers respectively are obtained.

Specifically, the method for the computer device to perform empirical mode decomposition on each group of noise processing signals specifically includes:

1) The method comprises the steps that computer equipment obtains signals to be decomposed corresponding to current iteration, wherein the signals to be decomposed of the first iteration are a first group of noise processing signals;

2) The computer equipment determines all local maximum value points and local minimum value points of the first group of noise processing signals, and connects all local maximum value points and local minimum value points of the first group of noise processing signals by using a cubic spline interpolation method to form an upper envelope curve and a lower envelope curve of the first group of noise processing signals;

3) The computer equipment calculates the average value of the upper envelope curve and the lower envelope curve of the first group of noise processing signals to obtain a first group of average value curves, and makes a difference between the first group of noise processing signals and the first group of average value curves to obtain a first group of response sequences;

4) The computer equipment judges whether the first group of response sequences meet IMF conditions, wherein the IMF conditions are that the number of local extreme points and zero crossing points of the corresponding sequences is equal or at most one difference in the whole time range, and the average value of the envelope determined by the local maximum points and the envelope determined by the local minimum points is zero;

5) If the first group of response sequences meet the IMF conditions, corresponding connotation modal components can be obtained according to the IMF conditions, and the computer equipment separates the corresponding connotation modal components from the corresponding noise processing signals to obtain residual processing signals;

6) If more than two extreme values exist in the residual processing signals, the computer equipment takes the residual processing signals as signals to be decomposed of the next iteration, returns to the step 1) to continue execution until the number of the extreme values existing in the residual processing signals is not more than 2, and stops decomposition to obtain connotation modal components of a hierarchy corresponding to each decomposition;

7) And if the first group of response sequences do not meet the IMF condition, the computer equipment returns the first group of response sequences as noise processing signals to the step 1) to continue execution until the stopping condition is reached, and the content modal components corresponding to all the layers respectively are obtained.

In this embodiment, the computer device performs empirical mode decomposition on multiple groups of noise processing signals to obtain connotation mode components of different levels of each group of noise processing signals, so as to obtain a target connotation mode component, so that the energy entropy feature quantity to be detected has universality, and the accuracy of serial arc fault diagnosis in the energy entropy feature dimension is improved.

It should be understood that, although the steps in the flowcharts related to the embodiments described above are sequentially shown as indicated by arrows, these steps are not necessarily sequentially performed in the order indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in the flowcharts described in the above embodiments may include a plurality of steps or a plurality of stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of the steps or stages is not necessarily performed sequentially, but may be performed alternately or alternately with at least some of the other steps or stages.

Based on the same inventive concept, the embodiment of the application also provides a series arc fault diagnosis device for realizing the series arc fault diagnosis method. The implementation of the solution provided by the device is similar to that described in the above method, so specific limitations in one or more embodiments of the series arc fault diagnosis device provided below may be referred to above for limitations of the series arc fault diagnosis method, and will not be repeated here.

In one embodiment, as shown in fig. 7, there is provided a series arc fault diagnosis apparatus comprising: an acquisition module 701, a processing module 702, an extraction module 703 and a determination module 704, wherein:

The acquisition module 701 is configured to acquire a current signal to be detected in a direct current system of the photovoltaic power station, and map the current signal to be detected to a digital twin model of the photovoltaic power station;

the processing module 702 is configured to perform wavelet preprocessing on a current signal to be detected to obtain a preprocessed signal;

The extracting module 703 is configured to extract a time domain feature to be detected of the pre-processed signal, extract a frequency domain feature to be detected of the pre-processed signal, and extract an energy entropy feature to be detected of the pre-processed signal;

The acquisition module 701 is further configured to acquire a three-dimensional feature space previously constructed in the digital twin model of the photovoltaic power station;

the determining module 704 is configured to trigger an alarm operation if a fault area where the current signal to be detected falls in the three-dimensional feature space is determined based on the time domain feature to be detected, the frequency domain feature to be detected, and the energy entropy feature to be detected.

The respective modules in the above-described series arc fault diagnosis apparatus may be implemented in whole or in part by software, hardware, and combinations thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.

In one embodiment, a computer device is provided, which may be a server or a terminal, and the internal structure of which may be as shown in fig. 8. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database of the computer device is used for storing current signal data to be detected. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a series arc fault diagnosis method.

It will be appreciated by those skilled in the art that the structure shown in FIG. 8 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.

In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the method embodiments described above when the computer program is executed.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored which, when executed by a processor, carries out the steps of the method embodiments described above.

In an embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, implements the steps of the method embodiments described above.

The user information (including but not limited to user equipment information, user personal information, etc.) and the data (including but not limited to data for analysis, stored data, presented data, etc.) related to the present application are information and data authorized by the user or sufficiently authorized by each party.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of the embodiments of the methods described above. Any reference to memory, database, or other medium used in embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high density embedded nonvolatile Memory, resistive random access Memory (ReRAM), magneto-resistive random access Memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric Memory (Ferroelectric Random Access Memory, FRAM), phase change Memory (PHASE CHANGE Memory, PCM), graphene Memory, and the like. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, and the like. By way of illustration, and not limitation, RAM can be in various forms such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The databases referred to in the embodiments provided herein may include at least one of a relational database and a non-relational database. The non-relational database may include, but is not limited to, a blockchain-based distributed database, and the like. The processor referred to in the embodiments provided in the present application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, or the like, but is not limited thereto.

The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.

The foregoing examples illustrate only a few embodiments of the application and are described in detail herein without thereby limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of the application should be assessed as that of the appended claims.

Claims

1. A method of series arc fault diagnosis, the method comprising:

performing wavelet pretreatment on the current signal to be detected to obtain a pretreated signal, wherein the wavelet pretreatment is used for filtering direct current interference signals in the current signal to be detected;

Extracting a time domain feature quantity to be detected of the preprocessing signal;

Calculating frequency band energy corresponding to a preset target frequency band, and taking the calculated frequency band energy as a frequency domain feature quantity to be detected of the preprocessing signal; the preset target frequency band is a frequency band of which the frequency band energy comparison value of the frequency domain current corresponding to each of the normal series current data and the fault series current data meets the difference condition;

extracting the energy entropy characteristic quantity to be detected of the preprocessing signal;

2. The method according to claim 1, wherein the method further comprises:

Respectively inputting the normal series current data and the fault series current data into the digital twin model of the photovoltaic power station;

Extracting a normal time domain characteristic quantity of the pretreatment normal current and a fault time domain characteristic quantity of the pretreatment fault current respectively;

Extracting a normal frequency domain characteristic quantity of the pretreatment normal current and a fault frequency domain characteristic quantity of the pretreatment fault current respectively;

Extracting a normal energy entropy characteristic quantity of the pretreatment normal current and a fault energy entropy characteristic quantity of the pretreatment fault current respectively;

Constructing a three-dimensional feature space in the digital twin model of the photovoltaic power station according to the normal region and the fault region; the three-dimensional feature space comprises a normal area, a fault area and an interference area.

3. The method according to claim 1, wherein the extracting the time domain feature quantity to be detected of the pre-processed signal comprises:

4. The method according to claim 1, wherein the extracting the energy entropy feature quantity to be detected of the pre-processed signal comprises:

Acquiring a preset group of different random white noise sequences;

5. The method of claim 4, wherein performing empirical mode decomposition on the respective noise-processed signals to obtain content modal components corresponding to each hierarchy level, respectively, comprises:

6. A series arc fault diagnostic apparatus, the apparatus comprising:

the processing module is used for carrying out wavelet pretreatment on the current signal to be detected to obtain a pretreatment signal, and the wavelet pretreatment is used for filtering direct current interference signals in the current signal to be detected;

The extraction module is used for extracting the time domain feature quantity to be detected of the preprocessing signal;

The extraction module is used for carrying out Fourier transform on the preprocessing signal to obtain a frequency domain current to be detected; determining a frequency band interval according to the sampling frequency and the sampling time of the current signal to be detected, and carrying out frequency band division on the frequency domain current to be detected according to the frequency band interval; calculating frequency band energy corresponding to a preset target frequency band, and taking the calculated frequency band energy as a frequency domain feature quantity to be detected of the preprocessing signal; the preset target frequency band is a frequency band of which the frequency band energy comparison value of the frequency domain current corresponding to each of the normal series current data and the fault series current data meets the difference condition;

the extraction module is used for extracting the energy entropy characteristic quantity to be detected of the preprocessing signal;

7. The apparatus of claim 6, wherein the extraction module is configured to calculate a current signal mean change rate and a current period maximum difference of the pre-processed signal; and taking the difference between the current signal mean change rate and the current period maximum value as the time domain characteristic quantity to be detected of the preprocessing signal.

8. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 5 when the computer program is executed.

9. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 5.

10. A computer program product comprising a computer program, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 5.