CN112255500A - Power distribution network weak characteristic fault identification method based on transfer learning - Google Patents
Power distribution network weak characteristic fault identification method based on transfer learning Download PDFInfo
- Publication number
- CN112255500A CN112255500A CN202011084220.9A CN202011084220A CN112255500A CN 112255500 A CN112255500 A CN 112255500A CN 202011084220 A CN202011084220 A CN 202011084220A CN 112255500 A CN112255500 A CN 112255500A
- Authority
- CN
- China
- Prior art keywords
- fault
- distribution network
- power distribution
- model
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- 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
- G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
- G01R31/52—Testing for short-circuits, leakage current or ground faults
-
- 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
- G01R31/081—Locating faults in cables, transmission lines, or networks according to type of conductors
- G01R31/086—Locating faults in cables, transmission lines, or networks according to type of conductors in power transmission or distribution networks, i.e. with interconnected conductors
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/18—Network design, e.g. design based on topological or interconnect aspects of utility systems, piping, heating ventilation air conditioning [HVAC] or cabling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2111/00—Details relating to CAD techniques
- G06F2111/02—CAD in a network environment, e.g. collaborative CAD or distributed simulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2113/00—Details relating to the application field
- G06F2113/04—Power grid distribution networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Computer Networks & Wireless Communication (AREA)
- Computational Mathematics (AREA)
- Computing Systems (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention provides a power distribution network weak characteristic fault identification method based on transfer learning, and relates to the technical field of power grid detection. The identification method comprises the following steps: firstly, a 20kV neutral point ungrounded alternating current distribution network model is built, two data acquisition points are respectively arranged at outgoing lines, and a training sample set and a transfer learning sample set are built; the sparse self-encoder is trained by using the training sample set, high accuracy of fault identification is realized, and finally, a small amount of migration sample sets are used for migration learning of the network model, so that the accuracy of the algorithm model can reach 98% under a new topological structure. The method can realize the type identification and fault line selection of the high-resistance fault in the power distribution networks with different topology types, and distinguish the interference signals of capacitor switching and load switching; the method is simple in principle, high in reliability, small in training sample number and strong in generalization capability, and weak characteristic fault identification can be realized in different power distribution network topologies.
Description
Technical Field
The invention belongs to the technical field of power grid detection, and particularly relates to a power distribution network weak characteristic fault identification method based on transfer learning.
Background
The power distribution network has the characteristics of complex topological structure, flexible and changeable operation mode and high fault occurrence probability, wherein the single-phase earth fault occurrence probability is highest. In China, after a distribution network has a single-phase earth fault, the energy band fault continues to operate for 1-2 hours, but the earth leakage capacitive current of the distribution network is obviously increased along with the enlargement of the distribution network scale. If the system is allowed to operate in a fault for a long time, the fault is further developed probably due to the excessive fault current, and the safe operation of the system and the personal safety of residents are threatened. Therefore, when the distribution network has a fault, corresponding measures should be taken quickly to remove the fault and find out the cause of the fault in time.
Distribution network fault identification is beneficial to quickly finding out the system fault reason and quickly taking corresponding measures to remove the fault. Meanwhile, fault identification is also a prerequisite for fault line selection and fault location. Therefore, the effective and accurate identification of the system fault is the guarantee of the safe operation of the distribution network. In a power distribution network, the fault types mainly include three-phase faults, two-phase faults, single-phase earth faults and the like. Meanwhile, disturbance signals of system operation, including nonlinear load switching, compensation capacitor switching and low-frequency and asymmetric characteristics of transformer excitation inrush current, exist in the power distribution network, and misjudgment of the fault detection device is easily caused. The single-phase high-resistance earth fault has the highest fault occurrence probability and high misdiagnosis rate due to weak fault characteristics, and becomes a research hotspot.
The existing fault identification method mainly adopts index threshold value judgment methods such as input impedance change values, wavelet energy moments, wavelet correlation coefficient energy, three-phase voltage absolute values, voltage and current correlation coefficients and the like to identify faults, but the method is adopted to extract high-frequency signals generated by fault arcs, and the noise resistance is poor. There are documents describing fault characteristics from the perspective of waveform distortion forms, such as the concavity and convexity of zero sequence current and the change characteristics of volt-ampere characteristic curve, but the reliability of these two methods in the case of distortion deviation is reduced.
Aiming at the problem that the conventional identification method is poor in sensitivity and reliability, domestic and foreign scholars apply a deep learning technology to power distribution network fault identification, automatically extract features of different analysis domains, realize direct mapping from original data to identification results, identify fault types through a classifier, can quickly discriminate interference signals and improve accuracy, but the method has the defects of data starvation and poor environment interactivity. The fault identification method based on deep learning needs a large number of training samples, and the actual power system fault data is less, so that the number of samples needed by applying an artificial intelligence technology cannot be achieved. The migration learning can complete the learning of similar networks through a small number of samples, and the probability that the original network based on a simulation layer is suitable for an actual power distribution network is increased through fine adjustment of the existing network topology, so that the application value of a fault identification model is improved.
Disclosure of Invention
The invention aims to provide a power distribution network weak characteristic fault identification method based on transfer learning, and solves the problems of data starvation and poor environment interactivity of the conventional power distribution network fault identification method.
In order to solve the technical problems, the invention is realized by the following technical scheme:
the invention relates to a power distribution network weak characteristic fault identification method based on transfer learning, which mainly comprises the following steps:
the method comprises the following steps: building a distribution network model and a fault model; building an 220/20kV power distribution network simulation model fed by double power supplies by referring to the European power distribution network standard, adjusting according to the actual situation of the Chinese power distribution network, setting a neutral point to be a non-grounding mode, and adopting an pi-shaped equivalent circuit with lumped parameters for a circuit; the topological structure of the line is changed by changing the opening and closing state of the line switch; in order to simulate the fault type of a power distribution network, a high-resistance grounding model, a low-resistance single-phase grounding model and an electric power system operation model with disturbance influence, namely a capacitance switching model and a load switching model, are built, and two data acquisition points are respectively arranged on two sections of buses connected with a 220kV system and are used for acquiring fault information of three-phase current; three-phase fault current information of 11 working conditions is obtained in total by changing fault types and fault positions, wherein the fault types comprise grounding types and disturbance signal types, the grounding types comprise A-phase high-resistance arc grounding, B-phase high-resistance arc grounding, C-phase high-resistance arc grounding, A-phase low-resistance grounding, B-phase low-resistance grounding and C-phase low-resistance grounding, and the disturbance signals comprise capacitance switching and load switching;
step two: data processing and sample set construction; dividing a power distribution network model into two radiation networks, namely a topology I and a topology II, acquiring three-phase fault current data of the topology I as a fault data set of a training and verification deep learning model, acquiring three-phase fault current information of the topology II as a fault data set of a training and verification migration network, wherein the sampling time is 0.1s, and the sampling frequency is 10 Hz; 15 fault points are set in total, each fault point simulates 11 working conditions, and finally 165 groups of samples are formed;
step three: training a network model; taking 80% of a sample set acquired by the topology I as a training sample set, training a sparse self-encoder neural network (SAE) by adopting an unsupervised learning mode to realize the sparseness of data characteristics, and then sending the sparse data characteristics into a classifier to train so as to realize 11 classifications of fault type identification; then, taking 20% of the topology-sample set as a verification set to verify the final network identification capability and accuracy;
step four: training a transfer learning model, and improving the generalization capability of the algorithm; and taking 80% of a small amount of sample data of the topology II as a training set, training a transfer learning model through fine tuning, automatically adjusting various parameters of the network, realizing fault type classification under a new topology structure, and finally taking 20% of data as a verification set to detect the network performance.
Preferably: in the first step, when the fault type of the power distribution network is identified, the weak characteristic fault of the power distribution network is effectively identified under the influence of the interference signal according to the influence of the disturbance signal on the identification reliability.
Preferably: the unsupervised learning mode in the third step is an unsupervised sparse autoencoder network learning method, and the learning method comprises the following steps: after normalization processing is carried out on the collected original data, the data are input into a sparse self-encoder network for dimension reduction processing, the most representative features are automatically extracted, and the identification precision of the classifier is improved.
Preferably: the learning method for training the transfer learning model in the fourth step comprises the following steps: the SAE model trained by a large amount of sample data is migrated to a topology II with only a small amount of sample information, and network parameters are automatically updated through fine tuning, so that the network is suitable for a new network topology in a short time, the generalization capability of the algorithm is improved, and the method is suitable for a complex and changeable actual power distribution network.
The invention has the following beneficial effects:
according to the method for identifying the weak characteristic fault of the power distribution network based on the transfer learning, the function of automatic characteristic extraction of the artificial neural network is utilized, and the high-resistance grounding fault is identified under the complex operation environment and various fault characteristics; by the transfer learning method, fault identification under different topological conditions can be realized under the condition of a small sample amount; the method is simple in training, convenient and fast to operate, high in sensitivity and reliability, capable of improving generalization capability of the algorithm and capable of adapting to power distribution network weak characteristic fault identification with flexible operation and complex and changeable topological structure.
Of course, it is not necessary for any product in which the invention is practiced to achieve all of the above-described advantages at the same time.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a network topology diagram of a power distribution network of the present invention;
FIG. 2 is a high resistance fault model diagram of the present invention;
FIG. 3 is a schematic diagram of sparse autoencoder data processing of the present invention;
FIG. 4 is a schematic diagram of the neural network structure of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, the present invention is a method for identifying weak characteristic faults of a power distribution network based on transfer learning, which mainly includes the following steps:
the method comprises the following steps: building a distribution network model and a fault model; building an 220/20kV power distribution network simulation model fed by double power supplies by referring to the European power distribution network standard, adjusting according to the actual situation of the Chinese power distribution network, setting a neutral point to be a non-grounding mode, and adopting an pi-shaped equivalent circuit with lumped parameters for a circuit; the topological structure of the line is changed by changing the opening and closing state of the line switch; in order to simulate the fault type of a power distribution network, a high-resistance grounding model, a low-resistance single-phase grounding model and an electric power system operation model with disturbance influence, namely a capacitance switching model and a load switching model, are built, and two data acquisition points are respectively arranged on two sections of buses connected with a 220kV system and are used for acquiring fault information of three-phase current; three-phase fault current information of 11 working conditions is obtained in total by changing fault types and fault positions, wherein the fault types comprise grounding types and disturbance signal types, the grounding types comprise A-phase high-resistance arc grounding, B-phase high-resistance arc grounding, C-phase high-resistance arc grounding, A-phase low-resistance grounding, B-phase low-resistance grounding and C-phase low-resistance grounding, and the disturbance signals comprise capacitance switching and load switching;
the model parameters are shown in the following tables 1-3, and with reference to fig. 1, the invention researches the situation when the switch 1 is disconnected, and two radiation nets are led out from the 220/20kV bus;
meter 120 kV power distribution network line parameters
TABLE 2 specific parameters of the lines
TABLE 3 Transformer parameters
Wherein, the capacitance is set to 240 muF, the load is an asymmetric three-phase load model with constant power, and the active power and the reactive power (unit is kW and kVar respectively) of the three phases are 10+ j0.1, 90+ j0.9 and 110+ j1.1 respectively; the intermittent arc high-resistance grounding fault model is shown in FIG. 3, specific parameters are shown in Table 4, and different types of arc high-resistance grounding faults can be simulated by changing the parameters of the elements;
TABLE 4 high resistance Fault model parameters
| Rp | Rn | Vp | Vn |
| 500 | 550 | 1500 | 750 |
| 700 | 750 | 2000 | 1500 |
| 600 | 650 | 3000 | 4000 |
Step two: data processing and sample set construction; dividing a power distribution network model into two radiation networks, namely a topology I and a topology II, collecting three-phase fault current data of the topology I as a fault data set of a training and verification deep learning model, collecting three-phase fault current information of the topology II as a fault data set of a training and verification migration network, as shown in figure 2, respectively setting two three-phase fault current sampling points on a bus 1 and a bus 12, wherein the sampling time is 0.1s, the sampling frequency is 10Hz, and controlling the added fault and disturbance signals by the on-off of a circuit breaker; totally setting 15 fault points, specifically setting positions as shown in fig. 1, traversing 11 faults and disturbances at each fault point, collecting corresponding three-phase fault current information, and finally forming 165 groups of samples;
the specific data acquisition method comprises the steps of numbering each fault point, wherein the numbering sequence is 1, 2, 3 and … …, the numbers 1 to 11 are training sample acquisition points with a topology one, and each group of faults acquire fault information of random time for 50 times, so that a large amount of (11 x 50 x 11 groups) fault simulation data are obtained and are used for training a self-encoder network; the serial number of the transfer learning sample acquisition points is 12-15, the transfer learning sample acquisition points are topological two, each fault point simulates 11 working conditions, each group of faults acquires 4 groups of fault information at random time, and a small amount of fault simulation data (11 x 4 x 11 groups) is obtained and is used for fine tuning training and verification of transfer learning; the specific fault types and corresponding labels are shown in table 5 below; transversely splicing the collected three-dimensional current vectors into arrays with the shapes of [1,303], and carrying out normalization processing on each array to finally form 6534 groups of samples, wherein 1-6050 groups are a sample set 1 and are used for training and verifying the model; the 6050-6534 group is a sample set 2 and is used for training and verifying the transfer learning. Data of two sample sets are randomly extracted, and 80% of the data are taken as a training set and 20% are taken as a verification set.
TABLE 5 Fault types and Classification tags
Step three: training a network model; taking 80% of a sample set acquired by the topology I as a training sample set, training a sparse self-encoder neural network (SAE) by adopting an unsupervised learning mode, realizing the sparseness of data characteristics, automatically learning the characteristics from label-free data by the SAE to obtain the characteristic description better than the original data, replacing the original data with the characteristics, enhancing the characteristics of the data sample, improving the accuracy, then sending the sparse data characteristics into a classifier for training, and realizing 11 classifications of fault type identification; then, taking 20% of the topology-sample set as a verification set to verify the final network identification capability and accuracy;
as shown in fig. 4, the process of data processing is implemented by introducing a penalty mechanism and a BP algorithm to solve the problem of minimum information loss, setting the characteristic number of a research sample to 303, setting the neuron number of an input layer to 303, setting a hidden layer to be one layer, setting the neuron numbers to 64, and finally stacking a softmax output layer with the neuron number of 11 to realize a multi-input and multi-output classification network structure; the hyper-parameters of the network are freely adjusted according to the rule of thumb, the processed data are input into the neural network for training and testing, and the trained model is stored.
Step four: training a transfer learning model, and improving the generalization capability of the algorithm; and finally, taking 20% of data as a verification set to detect the network performance, wherein the accuracy can reach 98.8%.
Wherein: in the first step, when the fault type of the power distribution network is identified, the weak characteristic fault of the power distribution network is effectively identified under the influence of the interference signal according to the influence of the disturbance signal on the identification reliability.
Wherein: the unsupervised learning mode in the third step is an unsupervised sparse autoencoder network learning method, and the learning method comprises the following steps: after normalization processing is carried out on the collected original data, the data are input into a sparse self-encoder network for dimension reduction processing, the most representative features are automatically extracted, and the identification precision of the classifier is improved.
Wherein: the learning method for training the transfer learning model in the fourth step comprises the following steps: the SAE model trained by a large amount of sample data is migrated to a topology II with only a small amount of sample information, and network parameters are automatically updated through fine tuning, so that the network is suitable for a new network topology in a short time, the generalization capability of the algorithm is improved, and the method is suitable for a complex and changeable actual power distribution network.
In the description herein, references to the description of "one embodiment," "an example," "a specific example" or the like are intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the invention. In this specification, the schematic representations of the terms used above do not necessarily refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
In addition, it can be understood by those skilled in the art that all or part of the steps in the method for implementing the embodiments described above can be implemented by instructing the relevant hardware through a program, and the corresponding program can be stored in a computer-readable storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, or the like.
The preferred embodiments of the invention disclosed above are intended to be illustrative only. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise embodiments disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best utilize the invention. The invention is limited only by the claims and their full scope and equivalents.
Claims (4)
1. A method for identifying weak characteristic faults of a power distribution network based on transfer learning is characterized by mainly comprising the following steps:
the method comprises the following steps: building a distribution network model and a fault model: building a power distribution network simulation model of 220/20kV fed by double power supplies, setting a neutral point as a non-grounding mode, and adopting an n-shaped equivalent circuit with lumped parameters for a circuit; the topological structure of the line is changed by changing the opening and closing state of the line switch; in order to simulate the fault type of a power distribution network, a high-resistance grounding model, a low-resistance single-phase grounding model and an electric power system operation model with disturbance influence, namely a capacitance switching model and a load switching model, are built, and two data acquisition points are respectively arranged on two sections of buses connected with a 220kV system and are used for acquiring fault information of three-phase current; three-phase fault current information of 11 working conditions is obtained in total by changing fault types and fault positions, wherein the fault types comprise grounding types and disturbance signal types, the grounding types comprise A-phase high-resistance arc grounding, B-phase high-resistance arc grounding, C-phase high-resistance arc grounding, A-phase low-resistance grounding, B-phase low-resistance grounding and C-phase low-resistance grounding, and the disturbance signals comprise capacitance switching and load switching;
step two: data processing and sample set construction: dividing a power distribution network model into two radiation networks, namely a topology I and a topology II, acquiring three-phase fault current data of the topology I as a fault data set of a training and verification deep learning model, acquiring three-phase fault current information of the topology II as a fault data set of a training and verification migration network, wherein the sampling time is 0.1s, and the sampling frequency is 10 Hz; 15 fault points are set in total, each fault point simulates 11 working conditions, and finally 165 groups of samples are formed;
step three: training of a network model: taking 80% of a sample set acquired by the topology I as a training sample set, training a sparse self-encoder neural network by adopting an unsupervised learning mode, realizing the sparseness of data characteristics, and then sending the sparse data characteristics into a classifier for training to realize 11 classifications of fault type identification; then, taking 20% of the topology-sample set as a verification set to verify the final network identification capability and accuracy;
step four: training a transfer learning model: and taking 80% of a small amount of sample data of the topology II as a training set, training a transfer learning model through fine tuning, automatically adjusting various parameters of the network, realizing fault type classification under a new topology structure, and finally taking 20% of data as a verification set to detect the network performance.
2. The method for identifying the weak characteristic fault of the power distribution network based on the transfer learning, according to claim 1, is characterized in that: in the first step, when the fault type of the power distribution network is identified, the weak characteristic fault of the power distribution network is effectively identified under the influence of the interference signal according to the influence of the disturbance signal on the identification reliability.
3. The method for identifying the weak characteristic fault of the power distribution network based on the transfer learning, according to claim 1, is characterized in that: the unsupervised learning mode in the third step is an unsupervised sparse autoencoder network learning method, and the learning method comprises the following steps: after normalization processing is carried out on the collected original data, the data are input into a sparse self-encoder network for dimension reduction processing, the most representative features are automatically extracted, and the identification precision of the classifier is improved.
4. The method for identifying the weak characteristic fault of the power distribution network based on the transfer learning, according to claim 1, is characterized in that: the learning method for training the transfer learning model in the fourth step comprises the following steps: the SAE model trained by a large amount of sample data is migrated to a topology II with only a small amount of sample information, and network parameters are automatically updated through fine tuning, so that the network is suitable for a new network topology in a short time, the generalization capability of the algorithm is improved, and the method is suitable for a complex and changeable actual power distribution network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011084220.9A CN112255500A (en) | 2020-10-12 | 2020-10-12 | Power distribution network weak characteristic fault identification method based on transfer learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011084220.9A CN112255500A (en) | 2020-10-12 | 2020-10-12 | Power distribution network weak characteristic fault identification method based on transfer learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112255500A true CN112255500A (en) | 2021-01-22 |
Family
ID=74242878
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011084220.9A Pending CN112255500A (en) | 2020-10-12 | 2020-10-12 | Power distribution network weak characteristic fault identification method based on transfer learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112255500A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113533904A (en) * | 2021-07-21 | 2021-10-22 | 南方电网科学研究院有限责任公司 | Method, device, equipment and medium for detecting high-resistance grounding fault of power distribution network |
| CN113642525A (en) * | 2021-09-02 | 2021-11-12 | 浙江大学 | Infant neural development assessment method and system based on skeletal points |
| CN114169249A (en) * | 2021-12-16 | 2022-03-11 | 福州大学 | Artificial intelligence identification method for high-resistance grounding fault of power distribution network |
| CN115184726A (en) * | 2022-07-12 | 2022-10-14 | 安徽省万企天成科技有限公司 | Intelligent power grid fault real-time monitoring and positioning system and method |
| CN115356596A (en) * | 2022-10-19 | 2022-11-18 | 广东电网有限责任公司佛山供电局 | Overhead line fault diagnosis method and system based on space vector conversion |
| CN116008731A (en) * | 2023-02-15 | 2023-04-25 | 重庆大学 | Power distribution network high-resistance fault identification method and device and electronic equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107491792A (en) * | 2017-08-29 | 2017-12-19 | 东北大学 | Feature based maps the electric network fault sorting technique of transfer learning |
| CN108510005A (en) * | 2018-04-08 | 2018-09-07 | 福州大学 | A kind of power distribution network high resistance earthing fault recognition methods based on convolutional neural networks |
| CN110161370A (en) * | 2019-04-25 | 2019-08-23 | 国网辽宁省电力有限公司 | A kind of electric network fault detection method based on deep learning |
| CN110780146A (en) * | 2019-12-10 | 2020-02-11 | 武汉大学 | Transformer fault identification and positioning diagnosis method based on multi-stage transfer learning |
| CN110879917A (en) * | 2019-11-08 | 2020-03-13 | 北京交通大学 | Electric power system transient stability self-adaptive evaluation method based on transfer learning |
| CN110943453A (en) * | 2019-12-23 | 2020-03-31 | 北京交通大学 | Power system fault sample generation and model construction method facing transfer learning |
| US20200150622A1 (en) * | 2018-11-13 | 2020-05-14 | Guangdong University Of Technology | Method for detecting abnormity in unsupervised industrial system based on deep transfer learning |
-
2020
- 2020-10-12 CN CN202011084220.9A patent/CN112255500A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107491792A (en) * | 2017-08-29 | 2017-12-19 | 东北大学 | Feature based maps the electric network fault sorting technique of transfer learning |
| CN108510005A (en) * | 2018-04-08 | 2018-09-07 | 福州大学 | A kind of power distribution network high resistance earthing fault recognition methods based on convolutional neural networks |
| US20200150622A1 (en) * | 2018-11-13 | 2020-05-14 | Guangdong University Of Technology | Method for detecting abnormity in unsupervised industrial system based on deep transfer learning |
| CN110161370A (en) * | 2019-04-25 | 2019-08-23 | 国网辽宁省电力有限公司 | A kind of electric network fault detection method based on deep learning |
| CN110879917A (en) * | 2019-11-08 | 2020-03-13 | 北京交通大学 | Electric power system transient stability self-adaptive evaluation method based on transfer learning |
| CN110780146A (en) * | 2019-12-10 | 2020-02-11 | 武汉大学 | Transformer fault identification and positioning diagnosis method based on multi-stage transfer learning |
| CN110943453A (en) * | 2019-12-23 | 2020-03-31 | 北京交通大学 | Power system fault sample generation and model construction method facing transfer learning |
Non-Patent Citations (5)
| Title |
|---|
| 佚名: "Tensorflow实现稀疏自动编码（SAE）", 《CSDN平台》 * |
| 佚名: "深度学习手记（六）之实现稀疏自编码算法（SAE）的优化过程", 《CSDN平台》 * |
| 佚名: "自编码算法（SAE）", 《CSDN平台》 * |
| 和敬涵 等: "新一代人工智能在电力系统故障分析及定位中的研究综述", 《中国电机工程学报》 * |
| 杨毅 等: "基于深度-迁移学习的输电线路故障选相模型及其可迁移性研究", 《电力自动化设备》 * |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113533904A (en) * | 2021-07-21 | 2021-10-22 | 南方电网科学研究院有限责任公司 | Method, device, equipment and medium for detecting high-resistance grounding fault of power distribution network |
| CN113642525A (en) * | 2021-09-02 | 2021-11-12 | 浙江大学 | Infant neural development assessment method and system based on skeletal points |
| CN114169249A (en) * | 2021-12-16 | 2022-03-11 | 福州大学 | Artificial intelligence identification method for high-resistance grounding fault of power distribution network |
| CN114169249B (en) * | 2021-12-16 | 2024-06-07 | 福州大学 | Artificial intelligent identification method for high-resistance ground fault of power distribution network |
| CN115184726A (en) * | 2022-07-12 | 2022-10-14 | 安徽省万企天成科技有限公司 | Intelligent power grid fault real-time monitoring and positioning system and method |
| CN115184726B (en) * | 2022-07-12 | 2023-11-28 | 广东中曼新能源科技有限公司 | Smart power grid fault real-time monitoring and positioning system and method |
| CN115356596A (en) * | 2022-10-19 | 2022-11-18 | 广东电网有限责任公司佛山供电局 | Overhead line fault diagnosis method and system based on space vector conversion |
| CN116008731A (en) * | 2023-02-15 | 2023-04-25 | 重庆大学 | Power distribution network high-resistance fault identification method and device and electronic equipment |
| CN116008731B (en) * | 2023-02-15 | 2023-08-25 | 重庆大学 | Power distribution network high-resistance fault identification method and device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112255500A (en) | Power distribution network weak characteristic fault identification method based on transfer learning | |
| CN110398663B (en) | Flexible direct current power grid fault identification method based on convolutional neural network | |
| Biswas et al. | State‐of‐the‐art on the protection of FACTS compensated high‐voltage transmission lines: a review | |
| Bo et al. | A new approach to phase selection using fault generated high frequency noise and neural networks | |
| CN108279364A (en) | Wire selection method for power distribution network single phase earthing failure based on convolutional neural networks | |
| CN108663602A (en) | Flexible direct current power distribution network monopole failure line selection and Section Location and system | |
| CN103728535A (en) | Extra-high-voltage direct-current transmission line fault location method based on wavelet transformation transient state energy spectrum | |
| CN103941163A (en) | Resonant earthed system fault line selection method utilizing fuzzy K-means clustering | |
| Sedighi et al. | Simulation of high impedance ground fault in electrical power distribution systems | |
| Ayyagari | Artificial neural network based fault location for transmission lines | |
| CN110247420B (en) | Intelligent fault identification method for HVDC transmission line | |
| CN112016473A (en) | Power distribution network high-resistance grounding fault diagnosis method based on semi-supervised learning and attention mechanism | |
| Rui et al. | Fault location for power grid based on transient travelling wave data fusion via asynchronous voltage measurements | |
| CN112787591B (en) | Photovoltaic array fault diagnosis method based on fine-tuning dense connection convolutional neural network | |
| Mishra et al. | A universal high impedance fault detection technique for distribution system using S-transform and pattern recognition | |
| Gao et al. | Fault line detection using waveform fusion and one-dimensional convolutional neural network in resonant grounding distribution systems | |
| Li et al. | A fault pattern and convolutional neural network based single-phase earth fault identification method for distribution network | |
| Nikoofekr et al. | Detection and classification of high impedance faults in power distribution networks using ART neural networks | |
| Narasimhulu et al. | LWT based ANN with ant lion optimizer for detection and classification of high impedance faults in distribution system | |
| Yuan et al. | Faulty feeder detection based on image recognition of voltage-current waveforms in non-effectively grounded distribution networks | |
| Yuan et al. | Faulty-Feeder Detection for Single Phase-to-Ground Faults in Distribution Networks Based on Waveform Encoding and Waveform Segmentation | |
| Yang et al. | Multi-Frequency bands based Pole-to-Ground fault detection method for MMC-Based radial DC distribution systems | |
| CN107462810A (en) | A kind of fault section location method suitable for active power distribution network | |
| Kanwal et al. | Artificial intelligence based faults identification, classification, and localization techniques in transmission lines-a review | |
| CN112710923A (en) | Data-driven single-phase earth fault line selection method based on post-fault steady-state information |
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: 20210122 |