CN115308532A - Power distribution network fault accurate positioning method and system based on random forest - Google Patents
Power distribution network fault accurate positioning method and system based on random forest Download PDFInfo
- Publication number
- CN115308532A CN115308532A CN202211005095.7A CN202211005095A CN115308532A CN 115308532 A CN115308532 A CN 115308532A CN 202211005095 A CN202211005095 A CN 202211005095A CN 115308532 A CN115308532 A CN 115308532A
- Authority
- CN
- China
- Prior art keywords
- difference value
- line
- positioning
- positioning result
- fault
- 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
- G01R31/08—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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- 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
Abstract
The invention discloses a method and a system for accurately positioning power distribution network faults based on random forests, which fully combine field operation practice, consider the influence of ubiquitous measurement errors and line parameter errors on positioning errors of a double-end synchronous algorithm, simulate the measurement errors and the line parameter errors by using an additive Gaussian white noise function, simultaneously select four basic algorithms with differences to respectively obtain four positioning results, then process the four positioning results of the basic algorithms to form training sample data, train a random forest model, and accurately position the power distribution network faults by using the trained random forest fault positioning model. The technical problem that in the prior art, a plurality of double-end fault positioning algorithms based on synchronous information are used for positioning the faults of the power distribution network, measurement errors and line parameter errors are not fully considered, and the fault positioning precision of the power distribution network is influenced is solved.
Description
Technical Field
The invention relates to the technical field of power distribution network fault monitoring, in particular to a power distribution network fault accurate positioning method and system based on a random forest.
Background
With the access of a Distributed Generation (DG) to a power distribution network, the scale of the power distribution network is continuously enlarged, the network topology is more and more complex, and the probability of power distribution network failure is more and more increased. The fault location after the fault occurs in the power distribution network is a basic link for fault processing of the power distribution network, and after the fault occurs, the fault location of the power distribution line is quickly and accurately completed, so that the fault isolation and power supply recovery speed can be improved, and the power supply reliability can be enhanced.
The existing power distribution network fault positioning method comprises a traveling wave method and a fault analysis method. The traveling wave method is used for accurately positioning the fault by measuring the traveling wave generated by the fault. The traveling wave method has high requirements on the sampling rate, more data transmission and high requirements on a communication system, and a large amount of special equipment needs to be installed when the traveling wave method is used, so that the economic cost is high, and therefore, in practical application scenes, the fault analysis method is more prone to be used for positioning the faults of the power distribution network. The fault analysis method establishes a loop equation by using system parameters and measurement data, obtains a fault distance by direct or iterative calculation, and can be mainly divided into a single-end quantity method and a double-end quantity method. The single-ended method is affected by the transition resistance and is slightly less reliable. In the trend of intellectualization and informatization, more and more devices with synchronous measurement functions are configured in a power distribution network. For example, the synchronous phasor measurement device can synchronously acquire voltage and current signals and transmit the signals to the dispatching center. The acquired synchronous data are substituted into a double-end fault positioning algorithm based on synchronous information, and then a positioning result can be obtained. Although the addition of the synchronization information greatly reduces the principle error of the positioning algorithm based on the double-end quantity, the influence of the measurement error and the line parameter error on the positioning result is still not negligible.
The invention provides a method and a system for accurately positioning power distribution network faults based on random forests, and aims to solve the problem that the measurement errors and line parameter errors of multiple double-end fault positioning algorithms based on synchronous information influence the positioning accuracy of the power distribution network faults in the conventional fault positioning method.
Disclosure of Invention
The invention provides a method and a system for accurately positioning a power distribution network fault based on a random forest, which are used for solving the technical problem that the power distribution network fault positioning accuracy is influenced by insufficient consideration of measurement errors and line parameter errors when a plurality of double-end fault positioning algorithms based on synchronous information are used for positioning the power distribution network fault in the prior art.
In view of this, the first aspect of the present invention provides a method for accurately positioning a power distribution network fault based on a random forest, including the following steps:
s1, constructing a power distribution network model, and performing fault simulation on the power distribution network model to obtain three-phase voltage and current at two ends of a fault line;
s2, performing measurement error simulation on the three-phase voltage and the three-phase current by using an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current containing the measurement error, and performing error simulation on the line parameters to obtain the line parameters containing the error;
s3, performing full-cycle Fourier transform and phase sequence transform on the three-phase voltage and current containing the measurement error to obtain sequence electric quantities at two ends of the line;
s4, respectively substituting sequence electrical quantities at two ends of the line and line parameters containing errors into a positive sequence voltage current method based on a lumped parameter line model, a positive and negative sequence impedance equality method based on the lumped parameter line model, a modulus solution based on the distributed parameter line model and a phase solution based on the distributed parameter line model to obtain a corresponding first positioning result, a corresponding second positioning result, a corresponding third positioning result and a corresponding fourth positioning result;
s5, calculating difference values of every two of the first positioning result, the second positioning result, the third positioning result and the fourth positioning result respectively to obtain a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value;
s6, taking the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value as input parameters, taking the actual fault position as a label, and constructing a training sample;
s7, training a random forest model by using the training sample to obtain a trained random forest model, and taking the trained random forest model as a power distribution network fault positioning model;
s8, three-phase voltage and current at two ends of a fault line, which are acquired when the power distribution network has a fault, are acquired, the steps S3 to S5 are executed, a first positioning result, a second positioning result, a third positioning result, a fourth positioning result, a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value are acquired, and the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value are input into a power distribution network fault positioning model to acquire a power distribution network fault positioning result.
Optionally, step S2 specifically includes:
respectively multiplying the three-phase voltage and the three-phase current by a first random number generated by an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current with measurement errors;
and multiplying the line parameter by a second random number generated by an additive white Gaussian noise function to obtain the line parameter containing errors.
Optionally, step S3 specifically includes:
after three-phase voltage and current containing measurement errors are obtained, carrying out full-cycle Fourier transform on three-phase voltage and current instantaneous values in the second period after the fault occurs to obtain phasor values of the three-phase voltage and the current at two ends of a fault line;
and (3) carrying out phase sequence transformation on the phasor values of the three-phase voltage and current by adopting a symmetrical component method to obtain the sequence electric quantities at two ends of the line.
Optionally, the number of CART trees of the random forest model is 100, the maximum depth of the CART trees is 20, and the minimum number of leaf node samples is 5.
Optionally, the first positioning result is:
wherein x is 1 For the first positioning result, l is the total length of the faulty line MN, z 1 Is the positive sequence impedance of the line unit,for the positive sequence voltage at the N-terminal of the faulty line,for the positive sequence current of the N-terminal of the fault line,for the positive sequence voltage at the end of the fault line M,is the positive sequence current of the fault line M terminal.
Optionally, the second positioning result is:
wherein x is 2 As a result of the second positioning, the positioning information,for the negative sequence voltage at the end of the fault line M,for the negative-sequence current at the end of the fault line M,for the negative sequence voltage at the N-terminal of the faulty line,is the negative sequence current at the N end of the fault line.
Optionally, the third positioning result is:
x 3 =ln(A 2 +B 2 )/4α 1
wherein x is 3 As a result of the third positioning, gamma 1 For faulty line MN positive sequence propagation coefficient, Z c1 As the characteristic impedance of the positive sequence of the line,y 1 for faulty line MN unit length admittance, α 1 As decay constant, beta 1 Is the phase constant, A is the real part and B is the imaginary part.
Optionally, the fourth positioning result is:
x 4 =arctan(B/A)/2β 1
wherein x is 4 Is the fourth positioning result.
The invention provides a power distribution network fault accurate positioning system based on a random forest, which comprises the following modules:
the modeling module is used for constructing a power distribution network model, carrying out fault simulation on the power distribution network model and acquiring three-phase voltage and current at two ends of a fault line;
the error simulation module is used for performing measurement error simulation on the three-phase voltage and the three-phase current by using an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current containing the measurement errors, and performing error simulation on the line parameters to obtain the line parameters containing the errors;
the sequence electric quantity module is used for carrying out full-cycle Fourier transformation and phase sequence transformation on three-phase voltage and current containing measurement errors to obtain sequence electric quantities at two ends of the line;
the initial positioning module is used for respectively substituting the sequence electrical quantities at two ends of the line and the line parameters containing errors into a positive sequence voltage current method based on a lumped parameter line model, a positive and negative sequence impedance equality method based on the lumped parameter line model, a modular value solution method based on a distributed parameter line model and a phase solution method based on the distributed parameter line model to obtain a corresponding first positioning result, a second positioning result, a third positioning result and a fourth positioning result;
the difference value calculation module is used for calculating pairwise difference values of the first positioning result, the second positioning result, the third positioning result and the fourth positioning result respectively to obtain a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value;
the training sample construction module is used for constructing a training sample by taking the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value as input parameters and taking the actual fault position as a label;
the model training module is used for training the random forest model by using the training sample to obtain a trained random forest model, and the trained random forest model is used as a power distribution network fault positioning model;
the fault positioning module is used for acquiring three-phase voltages and currents at two ends of a fault line acquired when the power distribution network has a fault, obtaining a first positioning result, a second positioning result, a third positioning result, a fourth positioning result, a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value after the three-phase voltages and the currents are processed by the sequence electric quantity module, the initial positioning module and the difference value calculation module, and inputting the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value into the power distribution network fault positioning model to obtain a power distribution network fault positioning result.
Optionally, the error simulation module is specifically configured to:
respectively multiplying the three-phase voltage and the three-phase current by a first random number generated by an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current with measurement errors;
and multiplying the line parameter by a second random number generated by an additive white Gaussian noise function to obtain the line parameter containing the error.
According to the technical scheme, the method and the system for accurately positioning the power distribution network fault based on the random forest have the following advantages:
the method for accurately positioning the faults of the power distribution network based on the random forest fully combines the field operation reality, considers the influence of ubiquitous measurement errors and line parameter errors on the positioning errors of the double-end synchronous algorithm, uses an additive white Gaussian noise function to perform measurement error simulation on three-phase voltage and current to obtain three-phase voltage and current containing the measurement errors, performs error simulation on the line parameters to obtain line parameters containing the errors, selects four basic algorithms with differences to calculate to respectively obtain four positioning results by taking the measurement results and the line parameters, which consider the errors, as fault analysis factors, and then processes the four positioning results of the basic algorithms to form training sample data and trains a random forest model. And finally, after the power distribution network fails, finishing accurate positioning by using the trained random forest fault positioning model. According to the method, the fault positioning model is constructed by utilizing the random forest in the integrated learning, the model positioning precision is high, the maintenance time is favorably shortened, the fault recovery speed is accelerated, the operation reliability of the system is improved, the influence of ubiquitous measurement errors and line parameter errors on double-end synchronous algorithm positioning errors is fully considered, the positioning effect is not influenced by factors such as transition resistance, fault initial angles and line lengths, the method is suitable for a novel power distribution network with high-permeability distributed power supplies, and the technical problems that the fault positioning of the power distribution network is carried out by using a plurality of double-end fault positioning algorithms based on synchronous information, the measurement errors and the line parameter errors are not fully considered, and the fault positioning precision of the power distribution network is influenced in the prior art are solved.
The principle and the obtained technical effect of the system for accurately positioning the faults of the power distribution network based on the random forest are the same as those of the method for accurately positioning the faults of the power distribution network based on the random forest, and the system is not repeated herein.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other related drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow diagram of a method for accurately positioning a fault of a power distribution network based on a random forest according to the present invention;
FIG. 2 is a schematic diagram of a power distribution network model built in the invention;
FIG. 3 is a lumped parameter line model sequence circuit provided in the present invention;
FIG. 4 is a distributed parametric line model positive sequence circuit provided in the present invention;
FIG. 5 is a schematic diagram of the overall construction process of the random forest model for fault accurate positioning provided in the present invention;
FIG. 6 is a schematic diagram of a relationship between the mean absolute error of the random forest model and the number of CART trees provided in the present invention;
fig. 7 is a schematic structural diagram of a power distribution network fault accurate positioning system based on a random forest provided in the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood, 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 obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
For convenience of understanding, referring to fig. 1, an embodiment of a method for accurately positioning a fault of a power distribution network based on a random forest is provided in the present invention, and includes the following steps:
It should be noted that, in the embodiment of the present invention, a power distribution network model is first built, a fault is set in the power distribution network model, and three-phase voltages and currents at two ends of a fault line are obtained, as shown in fig. 2, a Bergeron model is used for a power distribution network model line, an ideal voltage source line voltage is 35kV, a transformer transformation ratio is 35kV/10.5kV, a load is fixed by using a three-phase line to ground, and an interphase fault is set on a line B1M6 (line length 6460 meters). A large number of fault simulations are carried out by setting different line parameters, line types, fault positions, transition resistances, fault starting angles and the like. The line parameters are shown in table 1. The fault position takes 5% -95% (5% apart, 19 values altogether) from the head end of the line; the transition resistance is 0-55 omega (the interval is 5 omega, and the total number is 12); the failure initial angle is 0-324 degrees (36 degrees apart, and 12 values in total). After the fault happens, the measuring device arranged at the head end and the tail end of the circuit can synchronously measure the three-phase voltage and current at the installation position and send the three-phase voltage and current to the control center.
Table 1 line parameter set-up
And 102, performing measurement error simulation on the three-phase voltage and current by using an additive white Gaussian noise function to obtain the three-phase voltage and current with the measurement error, and performing error simulation on the line parameters to obtain the line parameters with the errors.
It should be noted that, for a single fault scenario generated by simulation, the instantaneous values of the three-phase voltages and currents at the two ends of the fault line and the line parameters (such as line length, impedance, etc.) are multiplied by random numbers, and the measurement error and the error simulation of the line parameters are performed. Considering that the measurement error and the line parameter error are randomly distributed, the random number for simulating the error in the embodiment of the invention is generated by an additive white gaussian noise function, namely, the three-phase voltage and the three-phase current are respectively multiplied by a first random number generated by the additive white gaussian noise function to obtain the three-phase voltage and the three-phase current containing the measurement error, and the line parameter is multiplied by a second random number generated by the additive white gaussian noise function to obtain the line parameter containing the error. The amplitude of the additive white Gaussian noise function follows Gaussian distribution, and the power spectral density is uniformly distributed. When the signal-to-noise ratio is 35dB, the average absolute error of the generated random numbers is 1.4%, and the actual conditions of measurement errors and line parameter errors are met. To expand the data volume, the simulation of the error is repeated 10 times for each fault scenario.
And 103, performing full-cycle Fourier transform and phase sequence transform on the three-phase voltage and current containing the measurement error to obtain sequence electric quantities at two ends of the line.
It should be noted that, after the measurement error and the line parameter error are simulated, the three-phase voltage and current with the measurement error are subjected to full-cycle fourier transform and phase-sequence transform to obtain the sequence electrical quantities at the two ends of the line. Specifically, the three-phase voltage and current instantaneous values in the second period after the fault occurs are subjected to full-period Fourier transform to obtain the phasor values of the three-phase voltage and current at the two ends of the fault line, and the phase sequence transformation is performed on the phasor values of the three-phase voltage and current by adopting a symmetrical component method to obtain the sequence electrical quantities at the two ends of the line. Thus, the electric quantity containing errors is obtained(positive sequence voltage at the faulty line M),(faulty line N)A terminal positive sequence voltage),(the positive sequence current at the end of the fault line M),(the positive sequence current at the N-terminal of the fault line),(negative sequence voltage at the end of the fault line M),(negative sequence voltage at the N-terminal of the faulty line),(negative sequence current at M end of fault line) and(negative sequence current at N end of faulty line), and line length l containing error, positive sequence resistance r per unit length 1 Positive sequence inductive reactance x per unit length L1 And positive sequence capacitive reactance x to ground C1 。
And 104, respectively substituting the sequence electrical quantities at two ends of the line and the line parameters containing errors into a positive sequence voltage current method based on a lumped parameter line model, a positive and negative sequence impedance equality method based on the lumped parameter line model, a modulus solution method based on the distributed parameter line model and a phase solution method based on the distributed parameter line model to obtain a corresponding first positioning result, a corresponding second positioning result, a corresponding third positioning result and a corresponding fourth positioning result.
It should be noted that, in combination with the development trend of the novel power distribution network and the fault characteristics of the distribution line, four basic algorithms are selected in the embodiment of the present invention, and are respectively: the method comprises a positive sequence voltage current method based on a lumped parameter line model, a positive and negative sequence impedance equality method based on the lumped parameter line model, a modulus solution based on a distributed parameter line model and a phase solution based on the distributed parameter line model. And respectively substituting the sequence electrical quantities at two ends of the line and the line parameters containing errors into a positive sequence voltage current method based on a lumped parameter line model, a positive and negative sequence impedance equality method based on the lumped parameter line model, a modulus solution based on a distributed parameter line model and a phase solution based on the distributed parameter line model to obtain a corresponding first positioning result, a corresponding second positioning result, a corresponding third positioning result and a corresponding fourth positioning result.
Specifically, the principle of implementing the first positioning result obtained by using the positive sequence voltage-current method based on the lumped parameter line model is as follows:
as shown in fig. 3, l is the total length of the fault line MN, x is the distance of the fault from the end M,is the sequence voltage at the location of the fault,andrespectively, the sequence voltage and the sequence current of the fault line M,andrespectively, the sequence voltage and sequence current, z, of the fault line N terminal i The impedance is a unit length sequence impedance of a fault line, wherein i =1 represents a positive sequence, and i =2 represents a negative sequence. After a fault occurs, the positive sequence voltage at the fault can be calculated by using the positive sequence voltage and the positive sequence current at the M end of the fault line MN, and can also be calculated by using the positive sequence voltage and the positive sequence current at the N end of the fault line MN. In conjunction with the uniqueness of the positive sequence voltage at the fault, the following equation can be written:
where l is the total length of the faulty line MN, z 1 Is a line unit positive sequence impedance, z 1 =r 1 +jx L1 ,r 1 Is a positive sequence resistance, x, per unit length of line L1 Is a positive sequence inductive reactance per unit length of the line,for the positive sequence voltage at the N-terminal of the faulty line,for the positive sequence current of the N-terminal of the fault line,for the positive sequence voltage at the end of the faulty line M,is the positive sequence current of the fault line M terminal.
Solving the formula (1) can obtain the fault distance, and for the convenience of distinguishing, the positioning result of the positive sequence voltage current method based on the lumped parameter line model is recorded as a first positioning result x 1 :
The principle of implementing the second positioning result obtained by using the positive-negative sequence impedance equivalence method based on the lumped parameter line model is as follows:
considering the distribution line as a stationary three-phase symmetrical element, the positive sequence impedance z 1 And negative sequence impedance z 2 Equal, i.e.:
z 1 =z 2 (3)
similar to the positive sequence voltage-current method, the negative sequence voltage at fault can be represented by the negative sequence voltage and the negative sequence current across the line, respectively, and the following equations are listed:
simultaneous equations (1), (3) and (4) are solvedObtaining a positioning result of a positive-negative sequence impedance equality method based on the lumped parameter line model, and recording the positioning result as a second positioning result x 2 :
Wherein the content of the first and second substances,for the negative sequence voltage at the end of the fault line M,for the negative-sequence current at the end of the fault line M,for the negative sequence voltage at the N-terminal of the faulty line,is the negative sequence current at the N end of the fault line.
The third positioning result and the fourth positioning result obtained by using the module value solution based on the distributed parameter line model and the phase solution based on the distributed parameter line model are realized by the following principle:
by using a distributed parameter line model, as shown in fig. 4, according to the transmission line theory, under the condition of considering the distribution characteristics of the line parameters, the voltage and current distribution at each point along the line and the relation between the voltage and current vectors at two ends of the line can be obtained by solving an equation, and by using this theory, the positive sequence voltage at the fault can be represented by the positive sequence voltage and current at the M end, or the positive sequence voltage and current at the N end, as shown in formula (6):
wherein ch is hyperbolic cosine function, sh is hyperbolic sine function, gamma 1 For faulty line MN positive sequence propagation coefficient, Z c1 For positive sequence characteristics of the lineThe impedance of the light source is measured,y 1 for faulty line MN unit length admittance, α 1 Is a damping constant, beta 1 Is a phase constant.
Equation (7) for the fault distance can be derived from equation (6):
wherein A is a real part and B is an imaginary part. Bound gamma 1 =α 1 +jβ 1 Formula (8) can be obtained:
thus, the third positioning result x is obtained based on a modulus solver of the distributed parameter line model 3 Comprises the following steps:
x 3 =ln(A 2 +B 2 )/4α 1 (9)
the phase solution method based on the distributed parameter line model utilizes the phase constant beta 1 And phase angle information of formula (8) to obtain tan (2 beta) 1 x) = B/a. Thus, the fourth positioning result x obtained by the phase solution based on the distributed parameter line model 4 Comprises the following steps:
x 4 =arctan(B/A)/2β 1
due to alpha 1 <β 1 Thus, the phase solution is more stable than the modulus solution.
The four basic algorithms have the following differences in aspects such as line models, solution ideas and solution parameters:
(1) a line model aspect. The positive sequence voltage current method and the positive and negative sequence impedance equality method adopt a lumped parameter model, and the calculated amount is small. The module value solution and the phase solution adopt a distributed parameter line model, and the theoretical positioning precision is higher.
(2) Solving the idea. The positive and negative sequence impedance equality method solves the fault distance by combining the positive and negative sequence network equations and the information of equal positive and negative sequence impedance of the line, and is suitable for the extreme scene of the missing line parameters. The distributed parameter modulus solution uses the attenuation constant and the modulus information to solve, and the distributed parameter phase solution uses the phase constant and the phase information to solve.
(3) And solving parameter aspects. The positive-sequence impedance equivalence method uses negative-sequence voltage and current information but does not use line impedance parameters, as compared to the positive-sequence voltage-current method. The distribution parameter modulus solution and the phase solution take the parameter of the positive sequence capacitance into consideration more finely, and although more parameter information needs to be acquired, the theoretical positioning accuracy is higher.
And 105, respectively calculating pairwise differences of the first positioning result, the second positioning result, the third positioning result and the fourth positioning result to obtain a first difference, a second difference, a third difference, a fourth difference, a fifth difference and a sixth difference.
It should be noted that, after step 104, positioning results corresponding to four basic algorithms are obtained: first positioning result x 1 The second positioning result x 2 And the third positioning result x 3 And a fourth positioning result x 4 . Compared with the method that only the positioning results of the four basic algorithms are used as input features, if the difference value of every two basic algorithms is also used as the input features, the model obtained by training can obtain better effect on the test sample. Therefore, in the embodiment of the present invention, pairwise differences between the first positioning result, the second positioning result, the third positioning result, and the fourth positioning result are respectively calculated (the number of the differences is:one).
And 106, taking the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value as input parameters, taking the actual fault position as a label, and constructing a training sample.
It should be noted that, for the ith fault sample, the positioning results of the above four basic algorithms are recorded as x i,1 ,x i,2 ,x i,3 And x i,4 The actual fault location is y i . After step 105, the ith training sample may be recordedWherein the content of the first and second substances, to input, y i Is the output. In order to simplify the model calculation process, the 10 input quantities of the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value may be normalized, and the normalized data is used as the training sample. The normalization method may be dividing the line length, that is, dividing the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference, the second difference, the third difference, the fourth difference, the fifth difference, and the sixth difference by the line length for normalization.
And 107, training a random forest model by using the training samples to obtain a trained random forest model, and taking the trained random forest model as a power distribution network fault positioning model.
It should be noted that, after the training samples are constructed, the training samples may constitute a training sample set. Training the random forest model by using the training sample. Randomly taking a part of samples from the training sample set as a verification set, and determining the hyper-parameters by observing the performance of the model on the verification set. And then, training the random forest model for one time by using all training sample set data. And finally, storing the trained model.
The random forest is an integrated learning algorithm based on Classification and Regression trees (CART), has the advantages of insensitivity to abnormal values, high operation speed, difficulty in overfitting and the like, and can complete Classification and Regression tasks. For regression problems, training sample setsConstituting an input space, CART recursively divides each region into two sub-regions and determines output values on each sub-region. The method comprises the following specific steps:
1. construction of stochastic solitaire model
(1) And traversing the feature j and the partition variable s, and selecting the variable pair (j, s) with the minimum equation (11).
Wherein R is 1 And R 2 Two sub-regions, c 1 And c 2 Is the output value of both sub-regions.
(2) Using the selected variable pair (j, s), the data is divided into two sub-regions and the output value of each sub-region is calculated.
Wherein the content of the first and second substances,as a result of the prediction of the regression tree, N m Is a region R m The number of samples in the sample.
(3) Repeating steps (1) and (2) for both sub-regions until a stop condition is met.
(4) Dividing the input space into M regions, R 1 ,R 2 ,…,R M And then obtaining a constructed CART regression tree:
wherein I is a function of 1 when the condition is satisfied and 0 when the condition is not satisfied.
The random forest algorithm capable of integrating a plurality of CART trees is better in performance. When the fusion of the basic algorithm positioning results is completed by utilizing the random forest, the general flow is as follows: first, random play-back sampling is employed to generate a number of subdata sets from a training data set. A plurality of CART regression trees are then generated utilizing the subdata sets independently and in parallel. It should be noted that, when dividing the left and right subtrees, the random forest randomly selects a part of features from all the features on the nodes, and then finds the optimal features and the dividing variables in the selected part of features. And finally, inputting the test samples into a plurality of regression trees, and obtaining the positioning result of the random forest after arithmetic mean of the regression trees. The overall construction flow of the fault location model based on the random forest is shown in fig. 5.
2. Adjusting parameters to determine hyper-parameters of random forest model
And randomly taking 95% of data in the training sample set for model training, taking the remaining 5% as a verification set for model verification, and adjusting and determining the hyper-parameters by observing the expression of the model on the verification set. Because the sample characteristic number is less, the internal nodes are subdivided into the minimum sample number, the maximum characteristic number, the maximum leaf node number and other hyperparameters, and default values are adopted. As such, there are only 3 hyper-parameters that need to be adjusted: number of CART trees, maximum depth of CART trees, and minimum number of samples of leaf nodes.
Taking the determination of the number of CART trees as an example, a random forest model is obtained by training 95% of training set data, and the change of the average absolute error of the random forest model on the verification set along with the number of CART trees is recorded, as shown in fig. 6. From fig. 6, three points of information can be obtained: (1) compared with 4 input features, the average absolute error of the model obtained by training 10 input features is smaller, and the positioning precision is higher; (2) with the increase of the number of CART trees, the average absolute error is gradually reduced; (3) when the number of trees reaches a certain number, the positioning precision can not be obviously improved. In order to consider the model effect and the training expenditure, the number of CART trees is taken as 100. By using the same method, the positioning precision and the model complexity are considered, and finally the maximum depth of the CART tree is determined to be 20 and the minimum sample number of the leaf nodes is determined to be 5.
3. After the hyper-parameters are determined, the data of the whole training sample set is used for training again
And (4) retraining the random forest model once again by using the whole training samples to finally obtain a power distribution network fault positioning model, and storing the power distribution network fault positioning model.
And 108, acquiring three-phase voltages and currents at two ends of a fault line acquired when the power distribution network has a fault, executing steps 103 to 105 to obtain a first positioning result, a second positioning result, a third positioning result, a fourth positioning result, a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value, and inputting the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value into a power distribution network fault positioning model to obtain a power distribution network fault positioning result.
It should be noted that the power distribution network fault location model obtained in step 107 is applied to a fault location scenario of practical application, after a fault occurs in a real power distribution line, three-phase voltages and currents at two ends of the fault line are collected, then corresponding first, second, third, fourth, fifth and sixth positioning results are obtained according to the processing procedures from step 103 to step 105, and then the first, second, third, fourth, fifth and sixth positioning results are input to the power distribution network fault location model (whether normalization processing is needed or not is determined according to actual needs to obtain normalized data), so as to obtain a power distribution network fault location result.
Aiming at the power distribution network fault location model obtained in the embodiment of the invention, the generalization capability of the power distribution network fault location model in the aspects of fault position, line parameter, fault starting angle, transition resistance, load, line length and the like is provided and the application effect display of the power distribution network fault location model in the aspect of high-permeability distributed power supply containing applicability is provided as follows:
generalization capability at fault location:
as shown in Table 2, the fault locations in the test set were different from the training set by 807.5m,1453.5m,2099.5m,2745.5m,3391.5m,4037.5m,4683.5m,5329.5m,5975.5m, respectively. And calculating the average absolute error of the four basic algorithms and the random forest model on the corresponding test set.
Due to the existence of measurement errors and line parameter errors, the positioning errors of the basic algorithm are large, and the requirements of the power distribution network are not met. Comparing the positioning errors of the algorithms in the table 2, the positioning error of the random forest model obtained by training is obviously reduced compared with that of the basic algorithm, and is less than 200 meters, so that the requirement of field operation is met. The constructed model has better generalization capability on fault positions.
TABLE 2 generalization Performance of random forest models at Fault location
In table 2, the numbers are the average absolute error of the different algorithms on the corresponding test set, in units: and (4) rice.
Generalization capability over line parameters:
different distribution networks and different lines often have differences in line parameters. And the generalization capability of the random forest model on line parameters is verified. Four sets of line parameters were randomly generated within a reasonable range, differing from the line parameters in the training set, as shown in table 3.
Similarly, the mean absolute error of the random forest model and the base algorithm over the test set was calculated for four sets of line parameters, as shown in table 4. The positioning error of the model obtained by training is still minimal and is less than 200m, and the requirement of field operation is also met.
TABLE 3 training set line parameter settings
TABLE 4 generalization Performance of random forest models on line parameters
In table 4, the numbers are the average absolute error of the different algorithms on the corresponding test set, in units: and (4) rice.
Generalization capability in terms of fault onset angle, transition resistance, load, line length, etc.:
the factors such as the initial fault angle, the transition resistance, the line load and the line length are changed, and the generalization performance of the random forest fault location model is tested and is respectively shown in tables 5, 6 and 7.
The result shows that the positioning error is obviously reduced and is less than 200m by fusing the random forest models of the positioning results of the four basic algorithms. The random forest positioning model has higher positioning accuracy and has better generalization capability in the aspects of fault initial angles, transition resistance and the like.
TABLE 5 generalization Performance of the random forest model on Fault Start Angle and transition resistance
TABLE 6 generalization Performance of random forest models on line load
TABLE 7 generalization Performance of random forest models over line Length
In tables 5-7, the numbers are the average absolute error in units for the different algorithms over the corresponding test set: and (4) rice.
Adaptability of distributed power supply during access:
against the background of the "dual carbon" strategy and the construction of new power systems, power distribution networks will have access to more and more distributed power sources. After the distributed power supply is connected, the fault accurate positioning algorithm in the power distribution network still has high positioning accuracy and good adaptability. Corresponding tests are carried out for verifying the adaptability of the constructed random forest model when a large number of distributed power supplies are accessed.
The method comprises the steps of respectively connecting a rotary type distributed power source and an inverter type distributed power source to the tail end of a distribution line, setting an interphase fault on the line, collecting and processing electric quantities at two ends of the line, and finally forming a test set sample. The average absolute error of the positioning results of the four basic algorithms and the random forest algorithm on the corresponding test set is calculated, as shown in table 8. The rotary distributed power supply is simulated by connecting an ideal voltage source and an internal resistance in series. The inverter type distributed power supply is photovoltaic. The low penetration means that the power of the fault feeder distributed power supply/the line load is approximately equal to 30%, and the high penetration means that the power of the fault feeder distributed power supply/the line load is approximately equal to 60%.
TABLE 8 adaptability of random forest model to distribution network with distributed power supplies
In table 8, the numbers are the average absolute error of the different algorithms on the corresponding test set, in units: and (4) rice.
It can be found that when different types of distributed power supplies are accessed, compared with a basic algorithm, the constructed random forest fault positioning model can still obtain the minimum positioning error, and the minimum positioning error is still less than 200m. The fault steady-state power frequency electrical quantity is used in a basic algorithm adopted in the process of constructing the random forest positioning model, and a complex transient process of a distributed power supply after a fault occurs is avoided. Therefore, the constructed random forest fault location model is still applicable to a novel power distribution network accessed by a high-permeability distributed power supply.
The method for accurately positioning the faults of the power distribution network based on the random forest fully combines the field operation reality, considers the influence of ubiquitous measurement errors and line parameter errors on the positioning errors of the double-end synchronous algorithm, uses an additive white Gaussian noise function to simulate the measurement errors of three-phase voltage and current to obtain the three-phase voltage and current containing the measurement errors, carries out error simulation on the line parameters to obtain the line parameters containing the errors, simultaneously selects four basic algorithms with differences to calculate by taking the measurement results and the line parameters which consider the errors as fault analysis factors to respectively obtain four positioning results, then processes the four positioning results of the basic algorithms to form training sample data, and trains a random forest model. And finally, after the power distribution network fails, finishing accurate positioning by using the trained random forest fault positioning model. According to the method, a fault positioning model is constructed by utilizing random forests in integrated learning, the model positioning precision is high, the overhaul time is favorably shortened, the fault recovery speed is accelerated, the system operation reliability is improved, the influence of ubiquitous measurement errors and line parameter errors on double-end synchronous algorithm positioning errors is fully considered, the positioning effect is not influenced by factors such as transition resistance, fault initial angles and line lengths, the method is suitable for a novel power distribution network containing a high-permeability distributed power supply, and the technical problems that the power distribution network fault positioning is carried out by using a plurality of double-end fault positioning algorithms based on synchronous information, the measurement errors and the line parameter errors are not fully considered, and the power distribution network fault positioning precision is influenced in the prior art are solved.
For easy understanding, please refer to fig. 7, an embodiment of a power distribution network fault accurate positioning system based on a random forest according to the present invention includes the following modules:
the modeling module is used for constructing a power distribution network model, carrying out fault simulation on the power distribution network model and acquiring three-phase voltage and current at two ends of a fault line;
the error simulation module is used for performing measurement error simulation on the three-phase voltage and the three-phase current by using an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current containing the measurement errors, and performing error simulation on the line parameters to obtain the line parameters containing the errors;
the sequence electrical quantity module is used for carrying out full-cycle Fourier transform and phase sequence transform on the three-phase voltage and current containing the measurement error to obtain the sequence electrical quantities at two ends of the line;
the initial positioning module is used for respectively substituting the sequence electrical quantities at two ends of the line and the line parameters containing errors into a positive sequence voltage current method based on a lumped parameter line model, a positive and negative sequence impedance equality method based on the lumped parameter line model, a modular value solution method based on a distributed parameter line model and a phase solution method based on the distributed parameter line model to obtain a corresponding first positioning result, a second positioning result, a third positioning result and a fourth positioning result;
the difference value calculation module is used for calculating pairwise difference values of the first positioning result, the second positioning result, the third positioning result and the fourth positioning result respectively to obtain a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value;
the training sample construction module is used for constructing a training sample by taking the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value as input parameters and taking the actual fault position as a label;
the model training module is used for training the random forest model by using the training sample to obtain a trained random forest model, and the trained random forest model is used as a power distribution network fault positioning model;
the fault positioning module is used for acquiring three-phase voltage and current at two ends of a fault line acquired when a power distribution network fails, obtaining a first positioning result, a second positioning result, a third positioning result, a fourth positioning result, a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value after the three-phase voltage and the current are processed by the sequence electric quantity module, the initial positioning module and the difference value calculation module, and inputting the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value into the power distribution network fault positioning model to obtain a power distribution network fault positioning result.
The error simulation module is specifically configured to:
respectively multiplying the three-phase voltage and the three-phase current by a first random number generated by an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current with measurement errors;
and multiplying the line parameter by a second random number generated by an additive white Gaussian noise function to obtain the line parameter containing the error.
The electrical sequence quantity module is specifically used for:
after three-phase voltage and current containing measurement errors are obtained, carrying out full-cycle Fourier transform on three-phase voltage and current instantaneous values in the second period after the fault occurs to obtain phasor values of the three-phase voltage and the current at two ends of a fault line;
and (3) carrying out phase sequence transformation on phasor values of three-phase voltage and current by adopting a symmetrical component method to obtain sequence electric quantities at two ends of the line.
The number of CART trees of the random forest model is 100, the maximum depth of the CART trees is 20, and the minimum sample number of leaf nodes is 5.
The first positioning result is:
wherein x is 1 For the first positioning result, l is the total length of the faulty line MN, z 1 Is a positive sequence impedance of a unit of line,for the positive sequence voltage at the N-terminal of the faulty line,for the positive sequence current of the N-terminal of the fault line,for the positive sequence voltage at the end of the faulty line M,is the positive sequence current of the fault line M terminal.
The second positioning result is:
wherein x is 2 As a result of the second positioning,for the negative sequence voltage at the end of the fault line M,for the negative-sequence current at the end of the fault line M,for the negative sequence voltage at the N-terminal of the faulty line,is the negative sequence current at the N end of the fault line.
The third positioning result is:
x 3 =ln(A 2 +B 2 )/4α 1
wherein x is 3 As a result of the third positioning, gamma 1 For faulty line MN positive sequence propagation coefficient, Z c1 For the line positive sequence characteristic impedance,y 1 for faulty line MN unit length admittance, α 1 As decay constant, beta 1 Is the phase constant, A is the real part and B is the imaginary part.
The fourth positioning result is:
x 4 =arctan(B/A)/2β 1
wherein x is 4 Is the fourth positioning result.
The system for accurately positioning the faults of the power distribution network based on the random forest fully combines the field operation reality, considers the influence of ubiquitous measurement errors and line parameter errors on the positioning errors of the double-end synchronous algorithm, uses an additive white Gaussian noise function to simulate the measurement errors of three-phase voltage and current to obtain the three-phase voltage and current containing the measurement errors, carries out error simulation on the line parameters to obtain the line parameters containing the errors, simultaneously selects four basic algorithms with differences to calculate by taking the measurement results and the line parameters which consider the errors as fault analysis factors to respectively obtain four positioning results, then processes the four positioning results of the basic algorithms to form training sample data, and trains a random forest model. And finally, after the power distribution network fails, finishing accurate positioning by using the trained random forest fault positioning model. According to the method, a fault positioning model is constructed by utilizing random forests in integrated learning, the model positioning precision is high, the overhaul time is favorably shortened, the fault recovery speed is accelerated, the system operation reliability is improved, the influence of ubiquitous measurement errors and line parameter errors on double-end synchronous algorithm positioning errors is fully considered, the positioning effect is not influenced by factors such as transition resistance, fault initial angles and line lengths, the method is suitable for a novel power distribution network containing a high-permeability distributed power supply, and the technical problems that the power distribution network fault positioning is carried out by using a plurality of double-end fault positioning algorithms based on synchronous information, the measurement errors and the line parameter errors are not fully considered, and the power distribution network fault positioning precision is influenced in the prior art are solved.
The principle and the obtained technical effect of the system for accurately positioning the faults of the power distribution network based on the random forest are the same as those of the method for accurately positioning the faults of the power distribution network based on the random forest, and the system is not repeated herein.
The above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A power distribution network fault accurate positioning method based on random forests is characterized by comprising the following steps:
s1, constructing a power distribution network model, and performing fault simulation on the power distribution network model to obtain three-phase voltage and current at two ends of a fault line;
s2, measuring error simulation is carried out on the three-phase voltage and the three-phase current by using an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current containing the measuring error, and error simulation is carried out on line parameters to obtain the line parameters containing the errors;
s3, performing full-cycle Fourier transform and phase sequence transform on the three-phase voltage and current containing the measurement error to obtain sequence electric quantities at two ends of the line;
s4, respectively substituting the sequence electrical quantities at two ends of the line and the line parameters containing errors into a positive sequence voltage current method based on a lumped parameter line model, a positive and negative sequence impedance equality method based on the lumped parameter line model, a modulus solution method based on a distributed parameter line model and a phase solution method based on the distributed parameter line model to obtain a corresponding first positioning result, a corresponding second positioning result, a corresponding third positioning result and a corresponding fourth positioning result;
s5, calculating difference values of every two of the first positioning result, the second positioning result, the third positioning result and the fourth positioning result respectively to obtain a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value;
s6, taking the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value as input parameters, taking the actual fault position as a label, and constructing a training sample;
s7, training a random forest model by using the training sample to obtain a trained random forest model, and taking the trained random forest model as a power distribution network fault positioning model;
s8, three-phase voltage and current at two ends of a fault line acquired when the power distribution network fails are acquired, the step S3 to the step S5 are executed to acquire a first positioning result, a second positioning result, a third positioning result, a fourth positioning result, a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value, and the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value are input into a power distribution network fault positioning model to acquire a power distribution network fault positioning result.
2. The method for accurately positioning the faults of the power distribution network based on the random forest as claimed in claim 1, wherein the step S2 specifically comprises:
respectively multiplying the three-phase voltage and the three-phase current by a first random number generated by an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current with measurement errors;
and multiplying the line parameter by a second random number generated by an additive white Gaussian noise function to obtain the line parameter containing errors.
3. The method for accurately positioning the faults of the power distribution network based on the random forest as claimed in claim 1, wherein the step S3 specifically comprises:
after three-phase voltage and current containing measurement errors are obtained, carrying out full-cycle Fourier transform on three-phase voltage and current instantaneous values in the second period after the fault occurs to obtain phasor values of the three-phase voltage and the current at two ends of a fault line;
and (3) carrying out phase sequence transformation on the phasor values of the three-phase voltage and current by adopting a symmetrical component method to obtain the sequence electric quantities at two ends of the line.
4. The method for accurately positioning the faults of the power distribution network based on the random forest as claimed in claim 1, wherein the number of CART trees of the random forest model is 100, the maximum depth of the CART trees is 20, and the minimum sample number of leaf nodes is 5.
5. The method for accurately positioning the faults of the power distribution network based on the random forest as claimed in claim 1, wherein the first positioning result is as follows:
wherein x is 1 For the first positioning result, l is the total length of the faulty line MN, z 1 Is a positive sequence impedance of a unit of line,for the positive sequence voltage at the N-terminal of the faulty line,for the positive sequence current of the N-terminal of the fault line,for the positive sequence voltage at the end of the faulty line M,is the positive sequence current of the fault line M terminal.
6. The method for accurately positioning the faults of the power distribution network based on the random forest as claimed in claim 5, wherein the second positioning result is as follows:
wherein x is 2 As a result of the second positioning,for the negative sequence voltage at the end of the fault line M,for the negative-sequence current at the end of the fault line M,for the negative sequence voltage at the N-terminal of the faulty line,is the negative sequence current at the N end of the fault line.
7. The method for accurately positioning the faults of the power distribution network based on the random forest as claimed in claim 6, wherein the third positioning result is as follows:
wherein x is 3 As a result of the third positioning, gamma 1 For faulty line MN positive sequence propagation coefficient, Z c1 For the line positive sequence characteristic impedance,y 1 admittance per unit length, α, for faulty line MN 1 Is a damping constant, beta 1 Is the phase constant, A is the real part and B is the imaginary part.
8. The method for accurately positioning the faults of the power distribution network based on the random forest as claimed in claim 7, wherein the fourth positioning result is that:
x 4 =arctan(B/A)/2β 1
wherein x is 4 Is the fourth positioning result.
9. The utility model provides a distribution network fault accurate positioning system based on random forest which characterized in that includes following module:
the modeling module is used for constructing a power distribution network model, carrying out fault simulation on the power distribution network model and acquiring three-phase voltage and current at two ends of a fault line;
the error simulation module is used for performing measurement error simulation on the three-phase voltage and the three-phase current by using an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current containing the measurement errors, and performing error simulation on the line parameters to obtain the line parameters containing the errors;
the sequence electrical quantity module is used for carrying out full-cycle Fourier transform and phase sequence transform on the three-phase voltage and current containing the measurement error to obtain the sequence electrical quantities at two ends of the line;
the initial positioning module is used for respectively substituting the sequence electrical quantities at two ends of the line and the line parameters containing errors into a positive sequence voltage current method based on a lumped parameter line model, a positive and negative sequence impedance equality method based on the lumped parameter line model, a modular value solution method based on a distributed parameter line model and a phase solution method based on the distributed parameter line model to obtain a corresponding first positioning result, a second positioning result, a third positioning result and a fourth positioning result;
the difference value calculation module is used for calculating pairwise difference values of the first positioning result, the second positioning result, the third positioning result and the fourth positioning result respectively to obtain a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value;
the training sample construction module is used for constructing a training sample by taking the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value as input parameters and taking the actual fault position as a label;
the model training module is used for training the random forest model by using the training sample to obtain a trained random forest model, and the trained random forest model is used as a power distribution network fault positioning model;
the fault positioning module is used for acquiring three-phase voltage and current at two ends of a fault line acquired when a power distribution network fails, obtaining a first positioning result, a second positioning result, a third positioning result, a fourth positioning result, a first difference value, a second difference value, a third difference value, a fourth difference value, a fifth difference value and a sixth difference value after the three-phase voltage and the current are processed by the sequence electric quantity module, the initial positioning module and the difference value calculation module, and inputting the first positioning result, the second positioning result, the third positioning result, the fourth positioning result, the first difference value, the second difference value, the third difference value, the fourth difference value, the fifth difference value and the sixth difference value into the power distribution network fault positioning model to obtain a power distribution network fault positioning result.
10. The system for accurately positioning the fault of the distribution network based on the random forest as recited in claim 9, wherein the error simulation module is specifically configured to:
respectively multiplying the three-phase voltage and the three-phase current by a first random number generated by an additive white Gaussian noise function to obtain the three-phase voltage and the three-phase current with measurement errors;
and multiplying the line parameter by a second random number generated by an additive white Gaussian noise function to obtain the line parameter containing the error.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211005095.7A CN115308532A (en) | 2022-08-22 | 2022-08-22 | Power distribution network fault accurate positioning method and system based on random forest |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202211005095.7A CN115308532A (en) | 2022-08-22 | 2022-08-22 | Power distribution network fault accurate positioning method and system based on random forest |
Publications (1)
Publication Number | Publication Date |
---|---|
CN115308532A true CN115308532A (en) | 2022-11-08 |
Family
ID=83862667
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202211005095.7A Pending CN115308532A (en) | 2022-08-22 | 2022-08-22 | Power distribution network fault accurate positioning method and system based on random forest |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115308532A (en) |
-
2022
- 2022-08-22 CN CN202211005095.7A patent/CN115308532A/en active Pending
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Mirshekali et al. | A novel fault location methodology for smart distribution networks | |
Shabestary et al. | Ladder network parameters determination considering nondominant resonances of the transformer winding | |
CN108051700B (en) | The phase component fault distance-finding method of distribution line parameter identification based on μ PMU | |
Jiang et al. | PMU-based fault location using voltage measurements in large transmission networks | |
Davoudi et al. | Transient-based fault location on three-terminal and tapped transmission lines not requiring line parameters | |
CN106054023B (en) | The method of two sides system impedance is estimated in a kind of transmission line of electricity single end distance measurement | |
CN101807795A (en) | Method for forming electric energy metering simulation system and device thereof | |
Estebsari et al. | An improved fault location method for distribution networks exploiting emerging LV smart meters | |
CN110348114B (en) | Non-precise fault identification method for power grid completeness state information reconstruction | |
CN111157845B (en) | Time domain inversion technology-based fault positioning method suitable for power distribution network | |
Jamali et al. | A Fast and accurate fault location method for distribution networks with DG using genetic algorithms | |
Wu et al. | Synchrophasor measurement‐based fault location technique for multi‐terminal multi‐section non‐homogeneous transmission lines | |
CN112415273A (en) | Method for accurately measuring zero sequence parameters of double-circuit non-full-line parallel transmission line | |
Wang et al. | An edge computing method using a novel mode component for power transmission line fault diagnosis in distribution network | |
Zhang et al. | Fault location method for three-core cable using amplitude ratio of shield-grounding wire currents | |
CN107462810A (en) | A kind of fault section location method suitable for active power distribution network | |
Han et al. | A new assessment for the total harmonic contributions at the point of common coupling | |
Abasi et al. | Location of double-circuit grounded cross-country faults in GUPFC-compensated transmission lines based on current and voltage phasors analysis | |
CN107132500B (en) | A kind of synchronous phasor measurement unit on-line calibration method and apparatus | |
Dağ et al. | Harmonic source location and meter placement optimization by impedance network approach | |
CN115685046A (en) | Mutual inductor metering abnormity identification method, device, equipment and storage medium | |
CN115308532A (en) | Power distribution network fault accurate positioning method and system based on random forest | |
Zhang et al. | Kalman filter approach for line parameter estimation for long transmission lines | |
Poudineh-Ebrahimi et al. | Accurate fault location algorithm for three-terminal lines without time synchronization considering unknown and different parameters of sections | |
Li et al. | Two-stage fault section location for distribution networks based on compressed sensing with estimated voltage measurements |
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 |