CN113552447A - Series arc fault detection method based on random forest - Google Patents
Series arc fault detection method based on random forest Download PDFInfo
- Publication number
- CN113552447A CN113552447A CN202110847162.9A CN202110847162A CN113552447A CN 113552447 A CN113552447 A CN 113552447A CN 202110847162 A CN202110847162 A CN 202110847162A CN 113552447 A CN113552447 A CN 113552447A
- Authority
- CN
- China
- Prior art keywords
- frequency
- low
- series arc
- characteristic quantity
- point data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 47
- 238000007637 random forest analysis Methods 0.000 title claims abstract description 21
- 238000000034 method Methods 0.000 claims abstract description 38
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 26
- 239000013598 vector Substances 0.000 claims abstract description 26
- 238000012549 training Methods 0.000 claims abstract description 23
- 238000004458 analytical method Methods 0.000 claims abstract description 12
- 230000003190 augmentative effect Effects 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000005070 sampling Methods 0.000 claims description 10
- YBJHBAHKTGYVGT-ZKWXMUAHSA-N (+)-Biotin Chemical compound N1C(=O)N[C@@H]2[C@H](CCCCC(=O)O)SC[C@@H]21 YBJHBAHKTGYVGT-ZKWXMUAHSA-N 0.000 claims description 3
- 230000003247 decreasing effect Effects 0.000 claims description 3
- BTCSSZJGUNDROE-UHFFFAOYSA-N gamma-aminobutyric acid Chemical compound NCCCC(O)=O BTCSSZJGUNDROE-UHFFFAOYSA-N 0.000 claims description 3
- FEPMHVLSLDOMQC-UHFFFAOYSA-N virginiamycin-S1 Natural products CC1OC(=O)C(C=2C=CC=CC=2)NC(=O)C2CC(=O)CCN2C(=O)C(CC=2C=CC=CC=2)N(C)C(=O)C2CCCN2C(=O)C(CC)NC(=O)C1NC(=O)C1=NC=CC=C1O FEPMHVLSLDOMQC-UHFFFAOYSA-N 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 11
- 230000000694 effects Effects 0.000 description 10
- 238000000354 decomposition reaction Methods 0.000 description 9
- 238000003066 decision tree Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000009413 insulation Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 238000003745 diagnosis Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 206010000369 Accident Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 239000000428 dust Substances 0.000 description 1
- 238000010891 electric arc Methods 0.000 description 1
- 238000005485 electric heating Methods 0.000 description 1
- 230000005670 electromagnetic radiation Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001681 protective effect Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/08—Locating faults in cables, transmission lines, or networks
- G01R31/088—Aspects of digital computing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
- Y04S10/52—Outage or fault management, e.g. fault detection or location
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention relates to a series arc fault detection method based on a random forest, which comprises the following steps: acquiring a characteristic vector to be detected of the series arc current, and inputting the trained detection model based on the random forest algorithm to obtain a fault state of the series arc; wherein, the training process of the detection model comprises the following steps: acquiring sample characteristic vectors of a plurality of series arc currents and corresponding fault state labels to form a training set, and training a detection model by using the training set; the acquisition process of the feature vector comprises the following steps: collecting a high-frequency current component and a low-frequency current component of the series arc current, respectively obtaining a high-frequency characteristic quantity and a low-frequency characteristic quantity by performing waveform analysis on the high-frequency current component and the low-frequency current component, and combining the high-frequency characteristic quantity and the low-frequency characteristic quantity into a characteristic vector. Compared with the prior art, the method solves the problem that a single time domain characteristic cannot represent the series arc fault under different load conditions, and is suitable for arc fault detection of circuits with different loads mixed.
Description
Technical Field
The invention relates to the field of arc fault detection, in particular to a series arc fault detection method based on a random forest.
Background
Many serious fire accidents are caused by fault arcs in the line below the rated current or expected short circuit current, which may occur on improperly designed or aged supply lines, appliance plugs, and the power supply lines, internal wiring harnesses or component insulation of household appliances. When a fault arc occurs, protective devices such as leakage, overcurrent, and short circuit on the line may fail to detect the fault arc or act to cut off the power supply quickly.
Current research on arc fault detection can be broadly divided into three categories:
(1) establishing an arc simulation model, and identifying whether a fault occurs by detecting the change of corresponding parameters;
(2) detecting arc faults according to physical phenomena such as arc noise, radiation, temperature change and the like generated when arcs occur;
(3) the arc fault is detected from the electrical characteristic quantity at the time of occurrence of the arc fault.
The third type of research is the research direction of most researchers at present, and a method for detecting an arc fault according to an electrical characteristic quantity when the arc fault occurs has two problems at present:
firstly, the applicable scene has limitation. Based on characteristics such as light, heat, sound and electromagnetic radiation during arc discharge, the arc fault detection effect for switch cabinets and power distribution cabinets is good. However, for the arc fault of the building line, the installation position of the sensor cannot be determined because the position of the arc fault cannot be predicted, and meanwhile, because a large number of normal arc events such as switching arcs exist in the actual line, whether the normal arc events belong to the arc fault is difficult to distinguish by physical phenomena alone, so that the arc fault detection method based on the physical phenomena has limitation in detecting the arc fault in the building line;
secondly, the detection effect on the mixed load is not ideal. The characteristic quantity extracted in the time domain has good effect on the arc fault diagnosis of a single load line, but has poor effect on circuits with different loads mixed, and the energy entropy based on Empirical Mode Decomposition (EMD) has good effect on the characteristic quantity, but the waveform analyzed by the EMD algorithm has the problem of Mode aliasing, and the accuracy of arc fault detection is low.
Disclosure of Invention
The invention aims to overcome the defects in the prior art, provides a random forest-based series arc fault detection method, overcomes the problem that a single time domain characteristic cannot represent series arc faults under different loads, and is suitable for arc fault detection of circuits with different loads mixed.
The purpose of the invention can be realized by the following technical scheme:
a series arc fault detection method based on a random forest comprises the following steps:
acquiring a characteristic vector to be detected of the series arc current, and inputting the trained detection model based on the random forest algorithm to obtain a fault state of the series arc;
wherein, the training process of the detection model comprises the following steps:
acquiring sample characteristic vectors of a plurality of series arc currents and corresponding fault state labels to form a training set, and training a detection model by using the training set;
the acquisition process of the feature vector comprises the following steps:
collecting a high-frequency current component and a low-frequency current component of the series arc current, respectively obtaining a high-frequency characteristic quantity and a low-frequency characteristic quantity by performing waveform analysis on the high-frequency current component and the low-frequency current component, and combining the high-frequency characteristic quantity and the low-frequency characteristic quantity into a characteristic vector;
the problem that a single time domain feature cannot represent series arc faults under different load conditions is solved by respectively extracting the low-frequency feature quantity in the low-frequency current component and the high-frequency feature quantity in the high-frequency current component, and the accuracy of arc fault detection is improved by adopting a random forest algorithm based on an integrated learning technology to diagnose the series arc faults on the feature vectors.
Further, the waveform analysis process of the high-frequency characteristic quantity comprises:
21) decomposing the high-frequency current components through a VMD algorithm to obtain bandwidth and minimum K IMF components;
22) calculating the variance contribution rate and the energy entropy of each IMF component, and taking the energy entropy of the IMF component with the maximum variance contribution rate as a high-frequency characteristic quantity;
the VMD algorithm is adopted to carry out waveform decomposition on the high-frequency current component, the problems of mode mixing and end effect of EMD algorithm decomposition are solved, and the method is suitable for arc fault detection of circuits with different loads mixed.
Further, step 21) comprises:
decomposing the high frequency current component into K IMF components, the IMF component ukThe expression of (a) is:
uk(t)=Ak(t)cos(φk(t))
The following constraint variation problem is constructed:
wherein u iskAnd ωkRespectively IMF component and its center frequency, uk={u1,u2,u3,···,uK},ωk={ω1,ω2,ω3,···,ωK};
Solving a constraint variation problem by introducing augmented Lagrange to obtain a bandwidth and minimum K IMF components;
the expression of the augmented Lagrange L is as follows:
the process for solving the constraint variation problem by introducing the augmented Lagrange comprises the following steps:
iterative computation of u in augmented Lagrangian Lk、ωkAnd λkU is as describedk、ωkAnd λkThe iterative calculation formula of (a) is:
and when the relative error e is smaller than a set value epsilon, stopping iterative computation to obtain the bandwidth and the minimum K IMF components, wherein epsilon is larger than 0.
Further, the variance contribution rate SvarAnd energy entropy HEmThe calculation formula of (2) is as follows:
wherein E ismIs the energy of the IMF component, varmIs the variance of the IMF component.
Further, the waveform analysis process of the low-frequency characteristic quantity includes:
carrying out data sampling on the low-frequency current component waveform to obtain a plurality of sampling point data, and calculating low-frequency characteristic quantity according to the sampling point data;
the low-frequency characteristic quantity comprises kurtosis, a form factor, a crest factor, a pulse factor and a margin factor.
Further, the kurtosis ku is calculated by the following formula:
where x is the sample point data, μ is the mean of x, σ is the standard deviation of x, and E represents expectation.
Further, the formula for calculating the form factor S is:
where N is the total amount of sample point data.
Further, the formula for calculating the crest factor C is:
where N is the total amount of sample point data, xmaxAnd xminRespectively the maximum and minimum values in all sample point data.
Further, the formula for calculating the pulse factor I is as follows:
where N is the total amount of sample point data, xmaxAnd xminRespectively the maximum and minimum values in all sample point data.
Further, the calculation formula of the margin factor L is as follows:
where N is the total amount of sample point data, xmaxAnd xminRespectively the maximum and minimum values in all sample point data.
Compared with the prior art, the invention has the following beneficial effects:
(1) the invention collects the high-frequency current component and the low-frequency current component of the series arc current, respectively obtains the high-frequency characteristic quantity and the low-frequency characteristic quantity by performing waveform analysis on the high-frequency current component and the low-frequency current component, combines the high-frequency characteristic quantity and the low-frequency characteristic quantity into a characteristic vector, obtains a plurality of sample characteristic vectors of the series arc current and corresponding fault state labels to form a training set, utilizes the training set to train a detection model, obtains the characteristic vector to be detected of the series arc current, inputs the trained detection model based on the random forest algorithm to obtain the fault state of the series arc, overcomes the problem that the single time domain characteristic cannot represent the series arc fault under different load conditions by respectively extracting the low-frequency characteristic quantity in the low-frequency current component and the high-frequency characteristic quantity in the high-frequency current component, and diagnoses the arc series fault by adopting the random forest algorithm based on the integrated learning technology, the accuracy rate of arc fault detection is improved;
(2) the method comprises the steps of decomposing high-frequency current components through a VMD algorithm to obtain K IMF components with the smallest bandwidth and the smallest bandwidth, calculating the variance contribution rate and the energy entropy of each IMF component, taking the energy entropy of the IMF component with the largest variance contribution rate as a high-frequency characteristic quantity, and performing waveform decomposition on the high-frequency current components by adopting the VMD algorithm, so that the problems of modal mixing and end point effect of EMD algorithm decomposition are solved, and the method is suitable for arc fault detection of circuits with different loads mixed.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a circuit diagram of a simulation experiment of a series arc generator;
FIG. 3 is a schematic diagram of a series arc generator configuration;
FIG. 4 is a waveform of a low frequency current component when the electric drill load series arc is normal;
FIG. 5 is a waveform of a high frequency current component when the electric drill load series arc is normal;
FIG. 6 is a waveform of a low frequency current component when a power drill load series arc fails;
FIG. 7 is a waveform diagram of high frequency current components when a power drill load series arc fails;
the reference numbers in the figures illustrate:
1. the device comprises a fixed electrode, 2 a movable electrode, 3 a first insulating clamping element, 4 a sliding block, 5 a transverse adjusting knob, 6 a base and 7 a second insulating clamping element.
Detailed Description
The invention is described in detail below with reference to the figures and specific embodiments. The present embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation manner and a specific operation process are given, but the scope of the present invention is not limited to the following embodiments.
A method for detecting series arc faults based on a random forest, as shown in fig. 1, includes:
1) acquiring sample characteristic vectors of a plurality of series arc currents and corresponding fault state labels to form a training set;
2) training a detection model by using a training set;
3) acquiring a characteristic vector to be detected of the series arc current, and inputting the trained detection model based on the random forest algorithm to obtain a fault state of the series arc;
wherein, the training process of the detection model comprises the following steps:
acquiring sample characteristic vectors of a plurality of series arc currents and corresponding fault state labels to form a training set, and training a detection model by using the training set;
the acquisition process of the feature vector comprises the following steps:
collecting a high-frequency current component and a low-frequency current component of the series arc current, respectively obtaining a high-frequency characteristic quantity and a low-frequency characteristic quantity by performing waveform analysis on the high-frequency current component and the low-frequency current component, and combining the high-frequency characteristic quantity and the low-frequency characteristic quantity into a characteristic vector;
the problem that a single time domain feature cannot represent series arc faults under different load conditions is solved by respectively extracting the low-frequency feature quantity in the low-frequency current component and the high-frequency feature quantity in the high-frequency current component, and the accuracy of arc fault detection is improved by adopting a random forest algorithm based on an integrated learning technology to diagnose the series arc faults on the feature vectors.
When the series arc is judged to have a fault, the arc fault protection electric appliance acts.
The waveform analysis process of the high-frequency characteristic quantity comprises the following steps:
21) decomposing the high-frequency current components through a VMD algorithm to obtain bandwidth and minimum K IMF components;
22) calculating the variance contribution rate and the energy entropy of each IMF component, and taking the energy entropy of the IMF component with the maximum variance contribution rate as a high-frequency characteristic quantity;
the VMD algorithm is adopted to carry out waveform decomposition on the high-frequency current component, the problems of mode mixing and end effect of EMD algorithm decomposition are solved, and the method is suitable for arc fault detection of circuits with different loads mixed.
Step 21) includes:
decomposing the high frequency current component into K IMF components, IMF component ukThe expression of (a) is:
uk(t)=Ak(t)cos(φk(t))
The following constraint variation problem is constructed:
wherein u iskAnd ωkRespectively IMF component and its center frequency, uk={u1,u2,u3,···,uK},ωk={ω1,ω2,ω3,···,ωK};
Solving a constraint variation problem by introducing augmented Lagrange to obtain a bandwidth and minimum K IMF components;
the expression of the augmented lagrange L is:
the process of solving the constraint variation problem by introducing augmented Lagrange comprises the following steps:
iterative computation of u in augmented Lagrangian Lk、ωkAnd λk,uk、ωkAnd λkThe iterative calculation formula of (a) is:
and when the relative error e is smaller than a set value epsilon, stopping iterative computation to obtain the bandwidth and the minimum K IMF components, wherein epsilon is larger than 0.
The energy entropy value can measure the regularity degree of a time sequence, the energy characteristics of signals in different frequency bands, the current can be suddenly changed when a series arc fault occurs, the energy can be changed, the variance contribution rate refers to the proportion of the variation caused by a single factor to the total variation, and represents the influence of the single factor on a dependent variable, and the variance contribution rate SvarAnd energy entropy HEmThe calculation formula of (2) is as follows:
wherein E ismIs the energy of the IMF component, varmIs the variance of the IMF component.
The waveform analysis process of the low-frequency characteristic quantity comprises the following steps:
selecting low-frequency current component waveforms of two periods, namely 40ms, carrying out data sampling on the low-frequency current component waveforms to obtain 2500 sampling point data, and calculating low-frequency characteristic quantity according to the sampling point data;
the low frequency feature quantities include kurtosis, form factor, crest factor, pulse factor, and margin factor.
The kurtosis ku is a numerical statistic reflecting the distribution characteristics of random variables, is normalized 4-order central moment, and has the calculation formula as follows:
where x is the sample point data, μ is the mean of x, σ is the standard deviation of x, and E represents expectation.
The form factor S is the ratio of the effective value (RMS) to the rectified mean value, and is calculated as:
where N is the total amount of sample point data.
The crest factor C is the ratio of the peak value to the effective value of the signal, and the calculation formula is as follows:
where N is the total amount of sample point data, xmaxAnd xminRespectively the maximum and minimum values in all sample point data.
The ratio of the pulse factor I signal peak value to the rectified mean value (mean value of absolute values) is calculated by the formula:
where N is the total amount of sample point data, xmaxAnd xminRespectively the maximum and minimum values in all sample point data.
The ratio of the margin factor L signal peak value to the square root amplitude value is calculated by the following formula:
where N is the total amount of sample point data, xmaxAnd xminRespectively the maximum and minimum values in all sample point data.
The method comprises the steps that 6 feature values in total form a feature vector, a fault state label is set to be 0 or 1, 0 and 1 respectively represent a normal state and a fault state, a plurality of feature vectors are extracted and trained in a training set randomly through a decision tree algorithm to obtain a decision tree, the decision tree algorithm is C4.5, all decision trees obtained through repeated training form a detection model based on a random forest algorithm, the feature vector of series arc current to be detected is collected and input into the detection model, and whether series arc fails or not is judged.
Simulating series arc fault detection through a series arc generator, wherein the structural schematic diagram and the simulation experiment circuit diagram of the series arc generator are respectively shown in fig. 3 and fig. 2, a fixed electrode 1 is fixed on a base 6 through a first insulation clamping piece 3, a movable electrode 2 is fixed on a second insulation clamping piece 7, the second insulation clamping piece 7 is movably arranged on the base 6 through a sliding block 4, the position of the sliding block 4 is adjusted through a transverse adjusting knob 5, so as to adjust the distance between the fixed electrode 1 and the movable electrode 2, the fixed electrode 1 and the movable electrode 2 are respectively connected with a power supply and a load, and an arc is generated between the fixed electrode 1 and the movable electrode 2.
The series arc generator simulates the occurrence of series arc faults under different loads, the low-frequency current transformer and the high-frequency current transformer respectively collect low-frequency current components and high-frequency current components in the circuit under different load conditions and convert the low-frequency current components and the high-frequency current components into voltage output, filtering low-frequency and high-frequency currents respectively by a low-pass filter and a high-pass filter, setting the cut-off frequency to be 1kHz, acquiring current data by an oscilloscope, setting the sampling frequency to be 62.5kHz, taking an electric drill as an example, fig. 4 and 5 are respectively a low-frequency current component waveform diagram and a high-frequency current component waveform diagram when the electric drill load series arc is normal, figures 6 and 7 are waveform diagrams of low frequency current components and high frequency current components respectively when a power drill load series arc fails, the waveform diagrams are normalized current waveform diagrams, and the calculation results of each characteristic value of the series arc in the fault state and the normal state are shown in table 1:
TABLE 1 characteristic value calculation result Table
Characteristic value | Normal state | Series arc fault |
IMF component energy entropy with maximum variance contribution rate | 0.1458 | 0.1595 |
Kurtosis | 2.8555 | 6.2499 |
Form factor | 1.1903 | 1.4039 |
Crest factor | 5.5329 | 8.2308 |
Pulse factor | 6.5858 | 11.5554 |
Margin factor | 7.3481 | 14.7023 |
The electric heating kettle, the electric hair drier, the electric drill, the switching power supply, the resistor, the dust collector and other loads and mixed load conditions are analyzed, and a statistical table of the accuracy of the series arc fault detection of a single load is shown in table 2:
TABLE 3 Serial connection arc fault detection accuracy statistical table for single load
The statistical table of the serial arc fault detection accuracy of the hybrid load is shown in table 4:
TABLE 4 Serial arc fault detection accuracy statistical table for mixed load
As can be seen from tables 3 and 4, the detection effect of the method for detecting the series arc fault based on the random forest provided by the embodiment is good, the diagnosis rate for a single load and a mixed load is more than 97%, and meanwhile, the number of times that the series arc is diagnosed by test sample data under normal work is very small, and the misdiagnosis rate is low.
The embodiment provides a series arc fault detection method based on a random forest, which overcomes the problem that a single time domain feature cannot represent series arc faults under different load conditions by respectively extracting a low-frequency feature quantity in a low-frequency current component and a high-frequency feature quantity in a high-frequency current component, and improves the accuracy of arc fault detection by adopting a random forest algorithm based on an integrated learning technology to diagnose the series arc faults on the feature vectors; the high-frequency current components are decomposed through the VMD algorithm to obtain the bandwidth and the minimum K IMF components, the variance contribution rate and the energy entropy of each IMF component are calculated, the energy entropy of the IMF component with the maximum variance contribution rate is used as the high-frequency characteristic quantity, the VMD algorithm is adopted to carry out waveform decomposition on the high-frequency current components, the problems of mode mixing and end point effect of EMD algorithm decomposition are solved, and the method is suitable for arc fault detection of circuits with different loads mixed.
The foregoing detailed description of the preferred embodiments of the invention has been presented. It should be understood that numerous modifications and variations could be devised by those skilled in the art in light of the present teachings without departing from the inventive concepts. Therefore, the technical solutions available to those skilled in the art through logic analysis, reasoning and limited experiments based on the prior art according to the concept of the present invention should be within the scope of protection defined by the claims.
Claims (10)
1. A series arc fault detection method based on a random forest is characterized by comprising the following steps:
acquiring a characteristic vector to be detected of the series arc current, and inputting the trained detection model based on the random forest algorithm to obtain a fault state of the series arc;
wherein, the training process of the detection model comprises the following steps:
acquiring sample characteristic vectors of a plurality of series arc currents and corresponding fault state labels to form a training set, and training a detection model by using the training set;
the acquisition process of the feature vector comprises the following steps:
collecting a high-frequency current component and a low-frequency current component of the series arc current, respectively obtaining a high-frequency characteristic quantity and a low-frequency characteristic quantity by performing waveform analysis on the high-frequency current component and the low-frequency current component, and combining the high-frequency characteristic quantity and the low-frequency characteristic quantity into a characteristic vector.
2. A method as claimed in claim 1, wherein the waveform analysis process of the high frequency characteristic quantity comprises:
21) decomposing the high-frequency current components through a VMD algorithm to obtain bandwidth and minimum K IMF components;
22) and calculating the variance contribution rate and the energy entropy of each IMF component, and taking the energy entropy of the IMF component with the maximum variance contribution rate as the high-frequency characteristic quantity.
3. A method of detecting series arc faults based on a random forest as claimed in claim 2, wherein step 21) comprises:
decomposing the high frequency current component into K IMF components, the IMF component ukThe expression of (a) is:
uk(t)=Ak(t)cos(φk(t))
The following constraint variation problem is constructed:
wherein u iskAnd ωkRespectively IMF component and its center frequency, uk={u1,u2,u3,···,uK},ωk={ω1,ω2,ω3,···,ωK};
Solving a constraint variation problem by introducing augmented Lagrange to obtain a bandwidth and minimum K IMF components;
the expression of the augmented Lagrange L is as follows:
the process for solving the constraint variation problem by introducing the augmented Lagrange comprises the following steps:
iterative computation of u in augmented Lagrangian Lk、ωkAnd λkU is as describedk、ωkAnd λkThe iterative calculation formula of (a) is:
and when the relative error e is smaller than a set value epsilon, stopping iterative computation to obtain the bandwidth and the minimum K IMF components, wherein epsilon is larger than 0.
4. A method as claimed in claim 2, wherein the variance contribution ratio S is a function of the number of arc faults detected in the seriesvarAnd energy entropy HEmThe calculation formula of (2) is as follows:
wherein E ismIs the energy of the IMF component, varmIs the variance of the IMF component.
5. A method as claimed in claim 1, wherein the waveform analysis process of the low frequency characteristic quantity comprises:
carrying out data sampling on the low-frequency current component waveform to obtain a plurality of sampling point data, and calculating low-frequency characteristic quantity according to the sampling point data;
the low-frequency characteristic quantity comprises kurtosis, a form factor, a crest factor, a pulse factor and a margin factor.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110847162.9A CN113552447A (en) | 2021-07-27 | 2021-07-27 | Series arc fault detection method based on random forest |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110847162.9A CN113552447A (en) | 2021-07-27 | 2021-07-27 | Series arc fault detection method based on random forest |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113552447A true CN113552447A (en) | 2021-10-26 |
Family
ID=78104460
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110847162.9A Pending CN113552447A (en) | 2021-07-27 | 2021-07-27 | Series arc fault detection method based on random forest |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113552447A (en) |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114169398A (en) * | 2022-01-28 | 2022-03-11 | 国网天津市电力公司 | Photovoltaic direct-current arc fault identification method and device based on random forest algorithm |
CN114707550A (en) * | 2022-04-06 | 2022-07-05 | 天津科技大学 | Method for identifying direct current series arc fault at inverter end of photovoltaic system |
CN116298735A (en) * | 2023-05-10 | 2023-06-23 | 广东电网有限责任公司湛江供电局 | AC arc fault detection method and related device for low-voltage distribution network |
CN117761453A (en) * | 2023-11-28 | 2024-03-26 | 中国电力科学研究院有限公司 | Low-voltage fault arc identification method, device, medium and equipment |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107315996A (en) * | 2017-05-19 | 2017-11-03 | 燕山大学 | A kind of noise characteristic extracting method of ships under water based on IMF Energy-Entropies and PCA |
CN110488161A (en) * | 2019-07-23 | 2019-11-22 | 南京航空航天大学 | A kind of detection of multi-load series arc faults and localization method |
CN111257684A (en) * | 2020-04-01 | 2020-06-09 | 浙江人民电器有限公司 | Arc fault detection method and system |
CN112904157A (en) * | 2021-01-19 | 2021-06-04 | 重庆邮电大学 | Fault arc detection method based on integrated machine learning |
CN112904156A (en) * | 2021-01-19 | 2021-06-04 | 重庆邮电大学 | Fault arc detection method based on frequency domain classification |
CN112986712A (en) * | 2019-12-16 | 2021-06-18 | 施耐德电器工业公司 | Apparatus for detecting arc faults and associated arc fault protection unit |
CN113009271A (en) * | 2021-04-22 | 2021-06-22 | 威胜集团有限公司 | Fault arc detection and positioning method, device and computer readable storage medium |
CN113050594A (en) * | 2021-03-11 | 2021-06-29 | 北京航空航天大学 | Multi-information fusion fault diagnosis method and system for vectoring nozzle control system |
CN113076986A (en) * | 2021-03-29 | 2021-07-06 | 西安交通大学 | Photovoltaic fault arc characteristic selection method combining filtering type and packaging type evaluation strategies |
CN113094983A (en) * | 2021-03-29 | 2021-07-09 | 西安交通大学 | Online simulation method for multi-dimensional time-varying characteristics of direct-current fault electric arc of photovoltaic system |
WO2021139331A1 (en) * | 2020-01-08 | 2021-07-15 | 重庆交通大学 | Bearing fault diagnosis method based on instantaneous frequency optimization vmd |
-
2021
- 2021-07-27 CN CN202110847162.9A patent/CN113552447A/en active Pending
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107315996A (en) * | 2017-05-19 | 2017-11-03 | 燕山大学 | A kind of noise characteristic extracting method of ships under water based on IMF Energy-Entropies and PCA |
CN110488161A (en) * | 2019-07-23 | 2019-11-22 | 南京航空航天大学 | A kind of detection of multi-load series arc faults and localization method |
CN112986712A (en) * | 2019-12-16 | 2021-06-18 | 施耐德电器工业公司 | Apparatus for detecting arc faults and associated arc fault protection unit |
WO2021139331A1 (en) * | 2020-01-08 | 2021-07-15 | 重庆交通大学 | Bearing fault diagnosis method based on instantaneous frequency optimization vmd |
CN111257684A (en) * | 2020-04-01 | 2020-06-09 | 浙江人民电器有限公司 | Arc fault detection method and system |
CN112904157A (en) * | 2021-01-19 | 2021-06-04 | 重庆邮电大学 | Fault arc detection method based on integrated machine learning |
CN112904156A (en) * | 2021-01-19 | 2021-06-04 | 重庆邮电大学 | Fault arc detection method based on frequency domain classification |
CN113050594A (en) * | 2021-03-11 | 2021-06-29 | 北京航空航天大学 | Multi-information fusion fault diagnosis method and system for vectoring nozzle control system |
CN113076986A (en) * | 2021-03-29 | 2021-07-06 | 西安交通大学 | Photovoltaic fault arc characteristic selection method combining filtering type and packaging type evaluation strategies |
CN113094983A (en) * | 2021-03-29 | 2021-07-09 | 西安交通大学 | Online simulation method for multi-dimensional time-varying characteristics of direct-current fault electric arc of photovoltaic system |
CN113009271A (en) * | 2021-04-22 | 2021-06-22 | 威胜集团有限公司 | Fault arc detection and positioning method, device and computer readable storage medium |
Non-Patent Citations (3)
Title |
---|
KONSTANTIN DRAGOMIRETSKIY AND DOMINIQUE ZOSSO: "Variational Mode Decomposition", 《IEEE TRANSACTIONS ON SIGNAL PROCESSING》 * |
刘晓明,王丽君,侯春光,赵洋,刘湘宁: "基于小波包能量熵的低压串联故障电弧诊断", 《沈阳工业大学学报》 * |
崔芮华,王传宇,王洋: "VMD-ApEn在航空交流串联型电弧故障检测中的应用", 《电机与控制学报》 * |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114169398A (en) * | 2022-01-28 | 2022-03-11 | 国网天津市电力公司 | Photovoltaic direct-current arc fault identification method and device based on random forest algorithm |
CN114707550A (en) * | 2022-04-06 | 2022-07-05 | 天津科技大学 | Method for identifying direct current series arc fault at inverter end of photovoltaic system |
CN116298735A (en) * | 2023-05-10 | 2023-06-23 | 广东电网有限责任公司湛江供电局 | AC arc fault detection method and related device for low-voltage distribution network |
CN117761453A (en) * | 2023-11-28 | 2024-03-26 | 中国电力科学研究院有限公司 | Low-voltage fault arc identification method, device, medium and equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113552447A (en) | Series arc fault detection method based on random forest | |
Jiang et al. | Series arc fault detection based on random forest and deep neural network | |
Yao et al. | Characteristic study and time-domain discrete-wavelet-transform based hybrid detection of series DC arc faults | |
Liu et al. | Application of the variational mode decomposition-based time and time–frequency domain analysis on series DC arc fault detection of photovoltaic arrays | |
Routray et al. | High impedance fault detection in radial distribution system using S-transform and neural network | |
Chen et al. | Time–frequency distribution characteristic and model simulation of photovoltaic series arc fault with power electronic equipment | |
CN106324406B (en) | A kind of transformer DC magnetic bias method for diagnosing faults and device | |
CN112505511A (en) | Non-invasive low-voltage fault arc detection and positioning method and system | |
CN114252682B (en) | Large building abnormal load control method based on non-invasive load monitoring | |
CN112816838B (en) | GIS equipment defect diagnosis device and method based on FFT, VMD and LS-SVM | |
Wu et al. | Smart detection technology of serial arc fault on low-voltage indoor power lines | |
CN108169540A (en) | A kind of measuring method of wind power generating set voltage flicker | |
Gu et al. | Identification method for low-voltage arc fault based on the loose combination of wavelet transformation and neural network | |
Lin et al. | A novel non-intrusive arc fault detection method for low-voltage customers | |
Wang et al. | DC series arc fault detection method in photovoltaic system based on multiple frequency selections for common-mode conductive voltage | |
Wang et al. | A novel series arc fault detection method based on mel-frequency cepstral coefficients and fully connected neural network | |
CN112949563B (en) | Non-invasive load identification method based on variable point detection and improved KNN algorithm | |
Chu et al. | TDV-LSTM: A new methodology for series arc fault detection in low power AC systems | |
Wu et al. | Intelligent detection of serial arc fault on low voltage power lines | |
Strobl et al. | Refined fault detection in LVDC-grids with signal processing, system identification and machine learning methods | |
CN111025143A (en) | Bus conversion current switching test circuit device and parameter determination method thereof | |
Zhao et al. | Series arc fault diagnosis based on variational mode decomposition and random forest | |
Gupta et al. | Wavelet Based Enhanced Fault Detection Scheme for A Distribution System Embedded with Electric Vehicle Charging Station | |
Ala et al. | A smart sensing method for real-time monitoring of low voltage series-arc-fault | |
Liu et al. | Arc fault diagnosis and analysis based on wavelet neural network |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20211026 |