CN117520869B

CN117520869B - Method and system for identifying parameter interval of medium-voltage distribution network based on dynamic Bayesian network

Info

Publication number: CN117520869B
Application number: CN202410011378.5A
Authority: CN
Inventors: 姜学宝; 潘琪; 王亮; 陈康; 吴博文; 付柳笛; 徐洋; 周陈斌
Original assignee: Suzhou Power Supply Co of State Grid Jiangsu Electric Power Co Ltd
Current assignee: Suzhou Power Supply Co of State Grid Jiangsu Electric Power Co Ltd
Priority date: 2024-01-04
Filing date: 2024-01-04
Publication date: 2024-03-29
Anticipated expiration: 2044-01-04
Also published as: CN117520869A

Abstract

A medium-voltage distribution network parameter interval identification method and system based on a dynamic Bayesian network. The method comprises the steps of collecting historical physical quantity and historical meteorological data measured by a measuring device, and calculating historical distribution parameters to form historical measuring data, wherein the historical measuring data comprises a feature vector set and a label vector set; performing discretization and interval division on the historical data, and calculating a clustering center and an interval boundary value; constructing a static Bayesian network according to the influence relation between each characteristic and the distribution parameters, and learning the conditional probability of the static Bayesian network; expanding a static Bayesian network into a dynamic Bayesian network, and learning the time transition probability of the historical power distribution parameters; and collecting real-time measurement data of the medium-voltage distribution network, and inputting the real-time measurement data into a dynamic Bayesian network model to estimate interval boundary values of distribution parameters. The method and the device not only consider the physical information of the power distribution network, but also consider the influence of environmental factors, and improve the reliability of the parameter identification of the power distribution network.

Description

Method and system for identifying parameter interval of medium-voltage distribution network based on dynamic Bayesian network

Technical Field

The invention belongs to the technical field of medium-voltage distribution network parameter estimation, and particularly relates to a medium-voltage distribution network parameter interval identification method and system based on a dynamic Bayesian network.

Background

The medium-voltage distribution network is one of the most important energy infrastructures in the modern society, depends on accurate real-time distribution network line physical parameters, and is further widely applied to the fields of distribution network reactive power optimization, tide calculation, state estimation, fault diagnosis, topology identification and the like. In practical situations, the circuit parameters will change due to environmental and circuit aging, so as to indirectly influence the application. The current distribution network line parameters cannot be acquired or measured in real time, and the current distribution network line parameters are stored in the SCADA system and cannot be updated in real time. The number of installed measurement devices and the quality of measurement information often fail to meet the requirements of parameter identification. Therefore, in order to better meet the safe, stable and economic operation of the medium-voltage distribution network and reduce the situation of overlarge application deviation caused by inaccurate line parameter information parameters, how to accurately, quickly and reliably acquire and identify the parameter information of the distribution network by utilizing the data measured by the measuring device has important significance in ensuring the safe operation and economy of the distribution network. In order to further ensure that the line parameters can meet the quality requirements, the identification of the line parameter interval of the power distribution network plays a more important role.

The traditional medium-voltage distribution network line parameter identification method is mainly divided into two main types. The first type of method is a generalized single time section state estimation method. For example, a weighted least square method is used to solve the line parameters of the power distribution network, specifically, the line parameters are used as a linear regression problem, and the specified parameter variables are obtained through solving a power flow calculation model. The above method generally needs the intersection difference of two ends of the line when obtaining accurate line parameters, and the phasor measurement device in the power distribution network is generally expensive, which is difficult to meet the application in the practical system.

The second type of power distribution network topology identification method is mainly a dynamic state estimation algorithm. For example, a Kalman filtering algorithm is used to perform state estimation on the power distribution network line parameters, and an optimal parameter estimation result is obtained by establishing a measurement equation and an observation equation and updating a related covariance matrix in a time dimension. The method can perform online parameter identification through the collected measurement data, but generally needs to obey Gaussian distribution, has large calculation amount because the equation of the power system is nonlinear, is difficult to meet the application requirements in an actual system, and does not consider the variation range of the parameters, so that the reliability is reduced.

Disclosure of Invention

In order to solve the defects in the prior art, the invention provides a method and a system for identifying the parameter interval of a medium-voltage distribution network based on a dynamic Bayesian network, which acquire more accurate line parameter information of the medium-voltage distribution network by using data acquired by a current measuring device, so that the parameter identification of the medium-voltage distribution network is more effective and reliable, and the problems of lower parameter identification efficiency, poor tolerance and the like of the medium-voltage distribution network in the prior art are solved.

In order to solve the technical problems, the invention adopts the following technical scheme.

The invention firstly discloses a medium-voltage distribution network parameter interval identification method based on a dynamic Bayesian network, which comprises the following steps:

step 1: according to the historical physical quantity and the historical meteorological data measured by a measuring device installed on a bus node in a medium-voltage distribution network, calculating historical distribution parameters, and forming the historical physical quantity, the historical meteorological data and the historical distribution parameters into historical measuring data of a continuous time sequence, wherein the historical measuring data comprises a feature vector set and a label vector set;

step 2: discretizing historical measurement data of a continuous time sequence, dividing intervals of the historical measurement data, and calculating cluster centers of the feature vector set and the tag vector set and corresponding interval upper and lower boundary values according to the average difference degree of the feature vector set and the average difference degree of the tag vector set;

Step 3: constructing a static Bayesian network according to the influence relation between each characteristic of the historical physical quantity and the historical meteorological data and the historical distribution parameters, and learning the conditional probability of the static Bayesian network based on the clustering centers of the characteristic vector set and the label vector set;

step 4: expanding a static Bayesian network with a single time section into a time-varying dynamic Bayesian network, and learning the time transition probability of the historical power distribution parameters according to the clustering center of the label vector set to obtain a dynamic Bayesian network model;

step 5: and acquiring real-time measurement data of the medium-voltage distribution network, wherein the real-time measurement data comprises real-time physical quantities and real-time meteorological data measured by the measurement device, inputting the real-time measurement data into the dynamic Bayesian network model to estimate posterior probability of distribution parameters of the medium-voltage distribution network, and determining interval upper and lower boundary values of the distribution parameters according to the posterior probability.

The invention further comprises the following preferable schemes:

the historical physical quantity comprises branch active power of a medium-voltage distribution networkBranch reactive power->Lateral component of branch voltage drop- >The meteorological data comprises temperature->Moisture->Wind speed->Intensity of illumination->The power distribution parameters include a series branch resistance parameter and a series branch reactance parameter, and the forming the historical physical quantity, the historical meteorological data and the historical power distribution parameters into a continuous time series of historical measurement data further includes:

according to the corresponding branch voltage drop transverse components of different nodesBranch active power->Branch reactive power->Temperature of the position of the branch>Moisture->Wind speed->Intensity of illumination->Constructing a feature vector set->：

；

Constructing a label vector set according to the series branch resistance parameter and the series branch reactance parameter：

；

Representation->Time history measurement feature vector subset:

；

wherein the method comprises the steps ofRepresentation->Moment medium voltage distribution network node->And node->Average temperature between, wherein node->And nodeAny two nodes with entity connection relation in the medium-voltage distribution network; />Representation->Moment medium voltage distribution network node->And node->Average humidity between>Representation->Moment medium voltage distribution network node->And node->Average wind speed between>Representation ofMoment medium voltage distribution network node->And node->Average illumination intensity between- >Representation->Moment medium voltage distribution network node->And node->Branch active power between, +.>Representation->Moment medium voltage distribution network node->And node->Reactive power of the branch between>Representation->Moment medium voltage distribution network node->And node->The lateral component of the drop in branch voltage between, +.>Representing the length of time for collecting data;

the representation is according to->Tag vector subsets obtained from time historic measurement data:

；

wherein the method comprises the steps ofRepresentation->Moment medium voltage distribution network node->And node->Series branch resistance parameter between->Representation ofMoment medium voltage distribution network node->And node->A series leg reactance parameter therebetween.

In the step 2, the average difference degree of the feature vector set and the average difference degree of the label vector set are calculated according to the following procedures:

calculating any two subsets of vectors according to the historical measurement feature vector subsets and the label vector subsetsAnd (3) with、/>And->Euclidean distance between->And->：

；

Calculating any sample by using an average difference degree method according to the Euclidean distance between any two subsets obtained by calculationAnd->Corresponding average degree of difference->And->：

；

Calculating a feature vector set by using the average difference of all the subsetsAnd tag vector set->Mean degree of difference >And->：

；

。

In the step 2, calculating the cluster centers of the feature vector set and the label vector set and the corresponding upper and lower boundary values of the interval, further includes:

according to the feature vector setAnd tag vector set->Mean degree of difference>And->Determining an initial cluster center +.>And->Wherein->The number of cluster centers;and->Respectively the ith clustering center;

according to the initial clustering centerAnd->Number of cluster centers->For feature vector set->And tag vector set->K-means clustering of all subsets of (1) for each subset vector +.>And->Respectively calculating the distance between the clustering centers>And->And assigning it to the cluster centers closest to it to form cluster center sets +.>And->：

；

Wherein,representing subset vector +.>Subset vector corresponding to minimum distance from cluster center, +.>Representing subset vector +.>Subset vector corresponding to minimum distance from cluster center, +.>A subset vector representing a subset corresponding to a minimum value of the euclidean distance calculated by the cluster center;

according to the cluster center setAnd->Obtaining the cluster centers corresponding to the cluster centers and the upper and lower boundary values of the corresponding intervals +.>And->Wherein->And- >Respectively represent the lower boundary of the interval corresponding to the cluster center, < + >>And->Representing the upper boundary of the interval corresponding to the cluster center.

The step 3 further comprises:

step 3.1: temperature in the historical measurement dataWith moisture->Wind speed->Is +.>As 2 groups of 4 observations and 4 observations as hidden variable branch resistances +.>And branch reactance->Is a parent node of (a);

step 3.2: determining hidden variable branch resistance according to the branch voltage drop transverse component expressionBranch reactance->And branch active power->Branch reactive power->Lateral component of branch voltage drop->The relation between, i.e. the lateral component of the drop of the branch voltage +.>As branch active power +.>Branch reactive power->Child nodes of hidden variables:

；

wherein,representing branch active power, +.>Representing branch reactive power, +.>Representing the lateral component of the drop of the branch voltage, < >>Representing the nodes at the two ends of the branch +.>Or node->Node voltage amplitude, ">Representing hidden variable branch resistance, < ->Representing the branch reactance;

step 3.3: according to the cluster center setAnd->Calculating hidden variable +.>Temperature->Humidity ofWind speed->Intensity of illumination->Branch active power->Branch reactive power- >And the lateral component of the drop of the branch voltageConditional probability between->：

；

Wherein,representing a frequency statistics function, +.>Including temperature->Moisture->Wind speed->Intensity of illumination->Branch power->And combining the obtained plurality of conditional probabilities into the static Bayesian network.

In the step 4, the single-time-section static bayesian network is expanded into a time-varying dynamic bayesian network, and the method further includes:

and copying the static Bayesian network of the single time section at the current moment to obtain the static Bayesian network of the next time section, and determining the current moment only by the last moment according to the Markov assumption, thereby forming the dynamic Bayesian network of multiple time sections.

In the step 4, learning the time transition probability of the historical power distribution parameter to obtain a dynamic bayesian network model, and further comprising:

learning hidden variables on the basis of the dynamic Bayesian networkProbability of time transfer：

；

Wherein,representing a frequency count function;

conditional probability of completion of derivation in the static Bayesian networkAnd the time transition probability of the deduction in the dynamic Bayesian network +.>And outputting and storing the dynamic Bayesian network model structure.

In the step 5, inputting the real-time measurement data into the dynamic bayesian network model to estimate a posterior probability of a distribution parameter of the medium voltage distribution network, and determining a section upper and lower boundary value of the distribution parameter according to the posterior probability, further including:

step 5.1: real-time measurement data of the medium-voltage distribution network, which are acquired in real time, including temperature, humidity, wind speed, illumination intensity, branch active power, branch reactive power and branch voltage drop transverse components, are used as observation variables and are input into the dynamic Bayesian network model to obtain hidden variable branch resistance parameters and branch reactance parameters;

step 5.2: calculating the corresponding interval boundary according to the obtained clustering center of the hidden variable branch resistance parameter and branch reactance parameter and the corresponding probability value,/>]And outputs:

；

wherein,representing in clustersHeart is->The corresponding calculation probability, q is the number of clustering centers,/>Representing the calculated hidden variable +.>，/>Representing the calculated lower boundary of the values of the hidden variable branch resistance parameter and branch reactance parameter, +.>Representing the calculated upper boundary of the cryptovariable branch resistance parameter and branch reactance parameter values, Representing the lower boundary of the belonging cluster center +.>Or->，/>Upper boundary representing the belonging cluster center +.>Or->。

The invention also discloses a medium-voltage distribution network topology identification system using the medium-voltage distribution network parameter interval identification method based on the dynamic Bayesian network, which comprises a history measurement data acquisition and preprocessing module, a measurement data discretization and interval division module, a static Bayesian network construction module, a dynamic Bayesian network construction module and a medium-voltage distribution network real-time parameter interval identification module, and is characterized in that:

the historical measurement data acquisition and preprocessing module is used for calculating historical distribution parameters according to the historical physical quantity and the historical meteorological data measured by the measurement device installed on the bus node in the medium-voltage distribution network, and forming the historical physical quantity, the historical meteorological data and the historical distribution parameters into historical measurement data of a continuous time sequence, wherein the historical measurement data comprises a feature vector set and a label vector set;

the measurement data discretization and interval division module is used for discretizing the historical measurement data of the continuous time sequence, dividing the interval of the historical measurement data, and calculating the clustering centers of the feature vector set and the label vector set and the corresponding upper and lower boundary values of the interval according to the average difference degree of the feature vector set and the average difference degree of the label vector set;

The static Bayesian network construction module is used for constructing a static Bayesian network according to the influence relation between each characteristic of the historical physical quantity and the historical meteorological data and the historical distribution parameters, and learning the conditional probability of the static Bayesian network based on the characteristic vector set and the clustering center of the label vector set;

the dynamic Bayesian network construction module is used for expanding a static Bayesian network with a single time section into a time-varying dynamic Bayesian network, and learning the time transition probability of the historical power distribution parameters according to the clustering center of the label vector set to obtain a dynamic Bayesian network model;

the medium voltage distribution network real-time parameter interval identification module is used for acquiring real-time measurement data of the medium voltage distribution network, wherein the real-time measurement data comprise real-time physical quantities and real-time meteorological data measured by the measurement device, inputting the real-time measurement data into the dynamic Bayesian network model so as to estimate posterior probability of distribution parameters of the medium voltage distribution network, and determining interval upper and lower boundary values of the distribution parameters according to the posterior probability.

The application also discloses a terminal, which comprises a processor and a storage medium;

The storage medium is used for storing instructions;

the processor is used for operating according to the instruction to execute the steps of the medium-voltage distribution network parameter interval identification method based on the dynamic Bayesian network.

The application also discloses a computer readable storage medium, on which a computer program is stored, which when being executed by a processor, implements the steps of the medium voltage distribution network parameter interval identification method based on the dynamic Bayesian network.

Compared with the prior art, the method has the advantages that the data acquired by the measuring device are fully considered, and the rapid and accurate identification of the parameter interval of the medium-voltage distribution network is realized through the data. In addition, the invention fully considers the interval of the power distribution network line parameter estimation, enhances the credibility of the estimation result, and can inhibit the influence of measurement noise on the line parameter identification to a certain extent, so that the invention can improve the accuracy, the robustness and the practicability of the power distribution network line parameter identification.

Drawings

Fig. 1 is a flowchart of a method for identifying a medium-voltage distribution network parameter interval based on a dynamic bayesian network in the invention.

Fig. 2 is a topology diagram of a medium voltage distribution network in a preferred embodiment of the present invention.

Fig. 3 is a schematic diagram of a medium-voltage distribution network parameter interval identification model based on a dynamic bayesian network in the invention.

Fig. 4 is a block diagram of a medium voltage distribution network parameter interval identification system based on a dynamic bayesian network in the invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

The embodiments described herein are merely some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art without making any inventive effort, are within the scope of the present invention.

Aiming at the defects of the prior art, the invention provides a medium voltage distribution network parameter interval identification method and system based on a dynamic Bayesian network, which are used for collecting historical physical parameters measured by a measuring device installed on bus nodes in the medium voltage distribution network and weather data at related nodes, discretizing a continuous time sequence, dividing the data into interval centers of characteristic values and corresponding interval upper and lower boundary values on the basis of the discrete data, constructing a static Bayesian network according to the influence relation between the characteristics, learning related transition probability parameters on the basis of the constructed static Bayesian network, expanding the static Bayesian network with a single time section into a time-varying dynamic Bayesian network, learning time transition probabilities among specified characteristics, finally outputting a learned dynamic Bayesian network model, collecting real-time measurement data of the medium voltage distribution network, completing corresponding distribution network parameter estimation by using the obtained dynamic Bayesian network model, and calculating the upper and lower boundaries of a line parameter interval according to the posterior probability obtained by estimation. The method fully considers the data acquired by the measuring device, realizes rapid and accurate identification of the medium-voltage distribution network parameter interval through the data, and fully considers the interval of the distribution network line parameter estimation.

Referring to fig. 1, the method for identifying the parameter interval of the medium-voltage distribution network based on the dynamic bayesian network disclosed by the invention comprises the following steps:

step 1: according to the historical physical quantity and the historical meteorological data measured by a measuring device installed on a bus node in the medium-voltage distribution network, calculating historical distribution parameters, and forming the historical physical quantity, the historical meteorological data and the historical distribution parameters into historical measuring data of a continuous time sequence, wherein the historical measuring data comprises a feature vector set and a label vector set.

Specifically, step 1 further includes:

step 1.1: according to an original design scheme, a corresponding topological structure diagram of the medium-voltage distribution network is obtained, a bus in the medium-voltage distribution network is used as a node in a topological graph model, node numbering is carried out on the medium-voltage distribution network according to the topological graph model and power flow direction on the basis, and a low-voltage side node of a head-end transformer of the medium-voltage distribution network is set as a datum reference node.

Step 1.2: analyzing each physical quantity measured by a measuring device installed on a bus node, including branch active power of a medium-voltage distribution networkBranch reactive power->Lateral component of branch voltage drop- >Collecting node-related meteorological data including temperature +.>Moisture->Wind speed->Intensity of illumination->And collecting related physical parameters of branches among nodes in the medium-voltage distribution network, including branch series resistance +.>Branch series reactance->Relationship between them.

Step 1.3: according to the collected corresponding physical measurement value of each node in the medium-voltage distribution network and the related meteorological measurement data, calculating corresponding parameter information among all branches by combining a historical topological structure to form historical measurement data comprising different nodesCorresponding branch voltage drop transverse component of (2)Branch active power->Branch reactive power->Temperature of the position of the branch>Humidity ofWind speed->Intensity of illumination->The feature vector set is formed>And a tag vector set comprising corresponding series branch resistance parameter and series branch reactance parameter>The method comprises the following steps of:

；

wherein,representation->The time history measurement feature vector subset specifically comprises:

；

wherein the method comprises the steps ofRepresentation->Moment medium voltage distribution network node->And node->Average temperature between, wherein node->And node->Is any two nodes with entity connection relation in the medium voltage distribution network. />Representation->Moment medium voltage distribution network node->And node->Average humidity between >Representation->Moment medium voltage distribution network node->And node->Average wind speed between>Representation->Moment medium voltage distribution network node->And node->Average illumination intensity between->Representation->Moment medium voltage distribution network node->And node->Branch active power between, +.>Representation->Moment medium voltage distribution network node->And node->Reactive power of the branch between>Representation->Moment medium voltage distribution network node->And node->The lateral component of the drop in branch voltage between, +.>Representing the length of time the data was acquired.

Wherein,the representation is according to->The label vector subset obtained by the time history measurement data specifically comprises the following components:

；

wherein the method comprises the steps ofRepresentation->Moment medium voltage distribution network node->And node->Series branch resistance parameter between->Representation->Moment medium voltage distribution network node->And node->A series leg reactance parameter therebetween.

Step 2: discretizing historical measurement data of a continuous time sequence, dividing intervals of the historical measurement data, and calculating cluster centers of the feature vector set and the label vector set and corresponding interval upper and lower boundary values according to the average difference degree of the feature vector set and the average difference degree of the label vector set.

Specifically, the step 2 further includes:

step 2.1: and discretizing the sample set data aiming at the collected historical measurement data samples of the continuous time sequence, and calculating Euclidean distance between any two discrete sample subsets. Calculating any two subsets of vectors according to the historical measurement feature vector subsets and the label vector subsets obtained in the step 1And->、/>And->Euclidean distance between->And->The method is characterized by comprising the following steps:

；

step 2.2: based on the Euclidean distance between any two subsets obtained by calculationCalculating arbitrary samples by average degree of difference methodAnd->Corresponding average degree of difference->And->The method is characterized by comprising the following steps:

；

step 2.3: calculating a feature vector set by using the average difference degree of all the subsets calculated in the step 2.2And tag vector set->Mean degree of difference>And->The method is characterized by comprising the following steps:

；

。

step 2.4: according to the feature vector set obtained by calculationAnd tag vector set->Mean degree of difference>Andan initial cluster center is determined. Specifically, first the subsets are averaged for difference +.>And->The feature vector subset and the label vector subset corresponding to the maximum value in the cluster are respectively used as 1 st cluster center +.>And->And reject in the corresponding vector set And->A new set of vectors is formed. Then searching the largest subset average difference corresponding subset in the new vector set, if the difference between the subset and the selected initial cluster center is larger than the vector set average difference +.>And->The corresponding subset is taken as the 2 nd cluster center +.>And->Otherwise find the sub-largest average difference subset to judge, circulate until all the cluster centers satisfying the condition +.>And->Screening out, and adding->And->Is the initial cluster center, wherein->The number of cluster centers;

step 2.5: according to the initial cluster center determined in step 2.4And->Number of cluster centers->For feature vector set->And tag vector set->K-means clustering of all subsets of (1) for each subset vector +.>And (3) withRespectively calculating the distance between the clustering centers>And->And assigning it to the cluster centers closest to it to form cluster center sets +.>And->The method is characterized by comprising the following steps:

；

Step 2.6: according to the cluster center setAnd->Obtaining the cluster centers corresponding to the cluster centers and the upper and lower boundary values of the corresponding intervals +.>And->Wherein->And->Respectively represent the lower boundary of the interval corresponding to the cluster center, < + >>And->Representing the upper boundary of the interval corresponding to the cluster center.

Step 3: and constructing a static Bayesian network according to the historical physical quantity and the influence relation between each characteristic of the historical meteorological data and the historical distribution parameters, and learning the conditional probability of the static Bayesian network based on the clustering centers of the characteristic vector set and the label vector set.

Specifically, the step 3 further includes:

step 3.1: temperature in historical measurement dataWith moisture->Wind speed->Is +.>As 2 groups of 4 observations and 4 observations as hidden variable branch resistances +.>And branch reactance->Is a parent node of (a);

step 3.2: determining hidden variable branch resistance according to the branch voltage drop transverse component expressionBranch reactance->And branch active power->Branch reactive power->Lateral component of branch voltage drop->The relation between, i.e. the lateral component of the drop of the branch voltage +.>As branch active power +. >Branch reactive power->And child nodes of hidden variables, the specific expression is as follows:

；

step 3.3: according to the cluster center setAnd->Calculating hidden variable +.>Temperature->Humidity ofWind speed->Intensity of illumination->Branch active power->Branch reactive power->And the lateral component of the drop of the branch voltageConditional probability between->Z () represents a parameter set, and the specific expression is:

；

wherein,representing a frequency statistics function, +.>Including temperature->Moisture->Wind speed->Intensity of illumination->Branch power->S () represents a parameter set. I.e. temperature +.>Moisture->Wind speed->Intensity of illumination->Branch power->Each of the features of (a) is substituted by +.>Obtaining a plurality of conditional probabilities P, forming the static Bayesian network for characterizing each feature and hidden variable ++>Influence relation between the two.

Step 4: and expanding the static Bayesian network with the single time section into a time-varying dynamic Bayesian network, and learning the time transition probability of the historical power distribution parameters according to the clustering center of the label vector set to obtain a dynamic Bayesian network model.

Specifically, the step 4 further includes:

step 4.1: copying to obtain a static Bayesian network of a single time section at the next time section according to the static Bayesian network of the single time section at the current time, and determining the current time only by the last time according to the Markov assumption, thereby forming a dynamic Bayesian network of multiple time sections;

step 4.2: learning hidden variables on the basis of a dynamic Bayesian networkProbability of time transferThe specific expression is as follows:

；

wherein,representing a frequency count function;

step 4.3: the static Bayesian network is processedConditional probability of completion of a derivation in a complexAnd the time transition probability of the deduction in the dynamic Bayesian network +.>And outputting and storing the dynamic Bayesian network model structure.

Specifically, the step 5 further includes:

step 5.2: calculating the corresponding interval boundary according to the obtained clustering center of the hidden variable branch resistance parameter and branch reactance parameter and the corresponding probability value,/>]And output, specifically as follows:

；

wherein,representing the cluster center as +.>The corresponding calculation probability, q is the number of clustering centers,/>Representing the calculated hidden variable +.>，/>Representing the calculated lower boundary of the values of the hidden variable branch resistance parameter and branch reactance parameter, +.>Representing the calculated upper boundary of the cryptovariable branch resistance parameter and branch reactance parameter values,representing the lower boundary of the belonging cluster center +.>Or->，/>Upper boundary representing the belonging cluster center +.>Or->。

Compared with the prior art, the method has the beneficial effects that firstly, the historical physical parameters measured by the measuring device installed on the bus node in the medium-voltage distribution network and the meteorological data at the relevant node are collected; then discretizing the continuous time sequence, and dividing the data into intervals on the basis of discrete data to obtain interval centers of the characteristic values and corresponding interval upper and lower boundary values; secondly, constructing a static Bayesian network according to the influence relation among all the features, learning related transition probability parameters on the basis of the constructed static Bayesian network, expanding the static Bayesian network with a single time section into a time-varying dynamic Bayesian network, and learning time transition probabilities among the specified features; and finally, outputting the learned dynamic Bayesian network model, simultaneously collecting real-time measurement data of the medium-voltage distribution network, completing corresponding parameter estimation of the distribution network by using the obtained dynamic Bayesian network model, and calculating the upper and lower boundaries of a line parameter interval according to the posterior probability obtained by estimation. The invention fully considers the data acquired by the measuring device, realizes the rapid and accurate identification of the medium-voltage distribution network parameter interval through the data, fully considers the interval of the distribution network line parameter estimation, enhances the credibility of the estimation result, and can inhibit the influence of the measuring noise on the line parameter identification to a certain extent, so that the invention can improve the accuracy, the robustness and the practicability of the distribution network line parameter identification.

The present invention may be a system, method, and/or computer program product. Referring to fig. 4, the invention also discloses a medium-voltage distribution network topology identification system based on the medium-voltage distribution network parameter interval identification method based on the dynamic bayesian network, which comprises a history measurement data acquisition and preprocessing module 1, a measurement data discretization and interval division module 2, a static bayesian network construction module 3, a dynamic bayesian network construction module 4 and a medium-voltage distribution network real-time parameter interval identification module 5.

The historical measurement data acquisition and preprocessing module 1 is used for calculating historical distribution parameters according to the historical physical quantity and the historical meteorological data measured by a measurement device installed on a bus node in the medium-voltage distribution network, and forming the historical physical quantity, the historical meteorological data and the historical distribution parameters into historical measurement data of a continuous time sequence, wherein the historical measurement data comprises a feature vector set and a label vector set;

the measurement data discretization and interval division module 2 is used for discretizing the historical measurement data of the continuous time sequence, dividing the interval of the historical measurement data, and calculating the cluster centers of the feature vector set and the label vector set and the corresponding upper and lower boundary values of the interval according to the average difference of the feature vector set and the average difference of the label vector set;

The static Bayesian network construction module 3 is used for constructing a static Bayesian network according to the influence relation between each characteristic of the historical physical quantity and the historical meteorological data and the historical distribution parameters, and learning the conditional probability of the static Bayesian network based on the clustering centers of the characteristic vector set and the label vector set;

the dynamic Bayesian network construction module 4 is used for expanding a static Bayesian network with a single time section into a time-varying dynamic Bayesian network, and learning the time transition probability of the historical power distribution parameters according to the clustering center of the label vector set to obtain a dynamic Bayesian network model;

the medium voltage distribution network real-time parameter interval identification module 5 is configured to collect real-time measurement data of the medium voltage distribution network, where the real-time measurement data includes real-time physical quantities measured by the measurement device and real-time meteorological data, input the real-time measurement data into the dynamic bayesian network model, estimate posterior probability of a distribution parameter of the medium voltage distribution network, and determine interval upper and lower boundary values of the distribution parameter according to the posterior probability.

Based on the spirit of the present invention, those skilled in the art can easily think that a computer program product can be obtained based on the aforementioned medium voltage distribution network parameter interval identification method based on a dynamic bayesian network. The computer program product may include a computer readable storage medium having computer readable program instructions embodied thereon for causing a processor to implement aspects of the present disclosure. The application also comprises a terminal, which comprises a processor and a storage medium; the storage medium is used for storing instructions; the processor is used for operating according to the instruction to execute the steps of the medium-voltage distribution network parameter interval identification method based on the dynamic Bayesian network.

The computer readable storage medium may be a tangible device that can hold and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium would include the following: portable computer disks, hard disks, random Access Memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static Random Access Memory (SRAM), portable compact disk read-only memory (CD-ROM), digital Versatile Disks (DVD), memory sticks, floppy disks, mechanical coding devices, punch cards or in-groove structures such as punch cards or grooves having instructions stored thereon, and any suitable combination of the foregoing. Computer-readable storage media, as used herein, are not to be construed as transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (e.g., optical pulses through fiber optic cables), or electrical signals transmitted through wires.

The computer readable program instructions described herein may be downloaded from a computer readable storage medium to a respective computing/processing device or to an external computer or external storage device over a network, such as the internet, a local area network, a wide area network, and/or a wireless network. The network may include copper transmission cables, fiber optic transmissions, wireless transmissions, routers, firewalls, switches, gateway computers and/or edge servers. The network interface card or network interface in each computing/processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium in the respective computing/processing device.

Computer program instructions for performing the operations of the present disclosure can be assembly instructions, instruction Set Architecture (ISA) instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, c++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions may be executed entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of remote computers, the remote computer may be connected to the user computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (e.g., connected through the internet using an internet service provider). In some embodiments, aspects of the present disclosure are implemented by personalizing electronic circuitry, such as programmable logic circuitry, field Programmable Gate Arrays (FPGAs), or Programmable Logic Arrays (PLAs), with state information of computer readable program instructions, which can execute the computer readable program instructions.

Finally, it should be noted that the above embodiments are only for illustrating the technical solution of the present invention and not for limiting the same, and although the present invention has been described in detail with reference to the above embodiments, it should be understood by those skilled in the art that: modifications and equivalents may be made to the specific embodiments of the invention without departing from the spirit and scope of the invention, which is intended to be covered by the claims.

Claims

1. The medium-voltage distribution network parameter interval identification method based on the dynamic Bayesian network is characterized by comprising the following steps of:

step 5: collecting real-time measurement data of a medium-voltage distribution network, wherein the real-time measurement data comprises real-time physical quantities and real-time meteorological data measured by the measurement device, inputting the real-time measurement data into the dynamic Bayesian network model to estimate posterior probability of distribution parameters of the medium-voltage distribution network, and determining interval upper and lower boundary values of the distribution parameters according to the posterior probability;

the method for calculating the historical distribution parameters according to the historical physical quantity and the historical meteorological data measured by the measuring device installed on the bus node in the medium-voltage distribution network further comprises the following steps:

Obtaining a corresponding topological structure diagram of the medium-voltage distribution network according to an original design scheme, taking a bus in the medium-voltage distribution network as a node in a topological graph model, numbering the node of the medium-voltage distribution network according to the topological graph model and a power flow direction, and setting a low-voltage side node of a head-end transformer of the medium-voltage distribution network as a datum reference node;

analyzing each physical quantity measured by a measuring device installed on a bus node, including branch active power of a medium-voltage distribution networkBranch reactive power->Lateral component of branch voltage drop->Collecting node-related meteorological data including temperature +.>Moisture->Wind speed->Intensity of illumination->And collect related physical parameters of branch among nodes in medium voltage distribution network, including branch series resistanceBranch series reactance->A relationship between;

the composing the historical physical quantity, the historical meteorological data and the historical power distribution parameter into a continuous time series of historical measurement data further comprises:

according to the corresponding branch voltage drop transverse components of different nodesBranch active power->Branch reactive power->Temperature of the position of the branch>Moisture->Wind speed->Intensity of illumination- >Constructing a feature vector set->：

；

Wherein the method comprises the steps ofRepresentation->Time history measurement feature vector subset:

；

representation->Moment medium voltage distribution network node->And node->Average temperature between, wherein node->And node->Any two nodes with entity connection relation in the medium-voltage distribution network; />Representation->Moment medium voltage distribution network node->And node->Average humidity between>Watch->Moment medium voltage distribution network node->And node->Average wind speed between>Representation->Moment medium voltage distribution network node->And node->Average illumination intensity between->Representation->Moment medium voltage distribution network node->And node->Branch active power between, +.>Representation->Moment medium voltage distribution network node->And node->Reactive power of the branch between>Representation->Moment medium voltage distribution network node->And node->The lateral component of the drop in branch voltage between, +.>Representing the length of time for collecting data;

wherein the method comprises the steps ofRepresentation->Moment medium voltage distribution network node->And node->Between which are locatedSeries branch resistance parameter, < >>Representation->Moment medium voltage distribution network node- >And node->A series branch reactance parameter therebetween;

calculating any two subsets of vectors according to the historical measurement feature vector subsets and the label vector subsetsAnd->、/>And->Euclidean distance between->And->：

Calculating a feature vector set by using the average difference of all the subsetsAnd tag vector set->Mean degree of difference>And->：

；

according to the feature vector setAnd tag vector set->Mean degree of difference>And->Determining an initial cluster centerAnd->Wherein->The number of cluster centers; />And->Respectively the ith clustering center;

according to the initial clustering centerAnd->Number of cluster centers->For feature vector set->And tag vector set->K-means clustering of all subsets of (1) for each subset vector +. >And->Respectively calculating the distance between the clustering centers>Andand assigning it to the cluster centers closest to it to form cluster center sets +.>And->：

according to the cluster center setAnd->Obtaining the cluster centers corresponding to the cluster centers and the upper and lower boundary values of the corresponding intervals +.>And->Wherein->And->Respectively represent the lower boundary of the interval corresponding to the cluster center, < + >>And->Representing the upper boundary of the interval corresponding to the clustering center;

the step 3 further comprises:

step 3.2: determining hidden variable branch resistance according to the branch voltage drop transverse component expressionBranch reactance->And branch active power->Branch reactive power- >Lateral component of branch voltage drop->The relationship between, i.e. the lateral component of the drop of the branch voltageAs branch active power +.>Branch reactive power->Child nodes of hidden variables:

step 3.3: according to the cluster center setAnd->Calculating hidden variable +.>Temperature->Moisture->Wind speed->Intensity of illumination->Branch active power->Branch reactive power->And branch voltage drop transverse component->Conditional probability between->：

Wherein,representing a frequency statistics function, +.>Including temperature->Moisture->Wind speed->Intensity of illumination->Branch power->The obtained conditional probabilities are combined into the static Bayesian network;

copying to obtain a static Bayesian network of a single time section at the next time section according to the static Bayesian network of the single time section at the current time, and determining the current time only by the last time according to the Markov assumption, thereby forming a dynamic Bayesian network of multiple time sections;

Wherein,representing a frequency count function;

2. The method for identifying a parameter interval of a medium voltage distribution network based on a dynamic bayesian network according to claim 1, wherein in the step 5, the real-time measurement data is input into the dynamic bayesian network model to estimate a posterior probability of a distribution parameter of the medium voltage distribution network, and the determining the interval upper and lower boundary values of the distribution parameter according to the posterior probability further comprises:

wherein,representing the cluster center as +.>The corresponding calculation probability, q is the number of clustering centers,/>Representing the calculated hidden variable +.>，/>Representing the calculated lower boundary of the values of the hidden variable branch resistance parameter and branch reactance parameter, +.>Representing the calculated upper boundary of the values of the hidden variable branch resistance parameter and branch reactance parameter, +.>Representing the lower boundary of the belonging cluster center +.>Or->，/>Upper boundary representing the belonging cluster center +.>Or->。

3. The utility model provides a medium voltage distribution network parameter interval identification system, includes history measurement data acquisition and preprocessing module, measurement data discretization and interval division module, static Bayesian network construction module, dynamic Bayesian network construction module and medium voltage distribution network real-time parameter interval identification module, its characterized in that:

the medium voltage distribution network real-time parameter interval identification module is used for acquiring real-time measurement data of the medium voltage distribution network, wherein the real-time measurement data comprise real-time physical quantities and real-time meteorological data measured by the measurement device, inputting the real-time measurement data into the dynamic Bayesian network model so as to estimate posterior probability of distribution parameters of the medium voltage distribution network, and determining interval upper and lower boundary values of the distribution parameters according to the posterior probability;

The history measurement data acquisition and preprocessing module is further used for:

；

Constructing a label vector set according to the series branch resistance parameter and the series branch reactance parameter ：

；

Representation->Time history measurement feature vector subset:

；

wherein the method comprises the steps ofRepresentation->Moment medium voltage distribution network node->And node->Average temperature between, wherein node->And node->Any two nodes with entity connection relation in the medium-voltage distribution network; />Representation->Moment medium voltage distribution network node->And node->Average humidity between>Representation->Moment medium voltage distribution network node->And node->Average wind speed between>Representation->Moment medium voltage distribution network node->And node->Average illumination intensity between->Representation->Moment medium voltage distribution network node->And node->Branch active power between, +.>Representation->Moment medium voltage distribution network node->And node->Reactive power of the branch between>Representation->Moment medium voltage distribution network node->And node->The lateral component of the drop in branch voltage between, +.>Representing the length of time for collecting data;

wherein the method comprises the steps ofRepresentation->Moment medium voltage distribution network node->And node->Series branch resistance parameter between->Representation->Moment medium voltage distribution network node->And node->A series branch reactance parameter therebetween;

the measurement data discretization and interval division module is further configured to calculate an average difference between the feature vector set and the label vector set according to the following process:

From the history measurement featuresVector subsets and tag vector subsets, computing any two subset vectorsAnd->、/>And->Euclidean distance between->And->：

；

The measurement data discretization and interval division module is further used for:

according to the initial clustering centerAnd->Number of cluster centers->For feature vector set->And tag vector set->K-means clustering of all subsets of (1) for each subset vector +.>And->Respectively calculating the distance between each clustering center/>Andand assigning it to the cluster centers closest to it to form cluster center sets +.>And->：

Wherein,representing subset vector +.>Subset vector corresponding to minimum distance from cluster center, +.>Representing subset vector +. >Subset vector corresponding to minimum distance from cluster center, +.>A subset vector representing a subset corresponding to a minimum value of the euclidean distance calculated by the cluster center;

the static bayesian network construction module is further configured to:

temperature in the historical measurement dataWith moisture->Wind speed->Is +.>As 2 sets of 4 observed variables, and 4 were takenObservation variable as hidden variable branch resistance +.>And branch reactance->Is a parent node of (a);

determining hidden variable branch resistance according to the branch voltage drop transverse component expressionBranch reactance->And branch active powerBranch reactive power->Lateral component of branch voltage drop->The relation between, i.e. the lateral component of the drop of the branch voltage +.>As branch active power +.>Branch reactive power->Child nodes of hidden variables:

wherein,representing branch active power, +.>Representing branch reactive power, +.>Representing the lateral component of the drop of the branch voltage, < > >Representing the nodes at the two ends of the branch +.>Or node->Node voltage amplitude, ">Representing hidden variable branch resistance, < ->Representing the branch reactance;

according to the cluster center setAnd->Calculating hidden variable +.>Temperature->Moisture->Wind speed->Intensity of illumination->Branch active power->Branch reactive power->And branch voltage drop transverse component->Conditional probability between->：

the dynamic bayesian network construction module is further configured to:

Wherein,representing a frequency count function;

conditional probability of completion of derivation in the static Bayesian networkAnd the time transition probability of the deduction in the dynamic Bayesian network +. >And outputting and storing the dynamic Bayesian network model structure.

4. The medium voltage distribution network parameter interval identification system of claim 3, wherein the medium voltage distribution network real-time parameter interval identification module is further configured to:

real-time measurement data of the medium-voltage distribution network, which are acquired in real time, including temperature, humidity, wind speed, illumination intensity, branch active power, branch reactive power and branch voltage drop transverse components, are used as observation variables and are input into the dynamic Bayesian network model to obtain hidden variable branch resistance parameters and branch reactance parameters;

calculating the corresponding interval boundary according to the obtained clustering center of the hidden variable branch resistance parameter and branch reactance parameter and the corresponding probability value,/>]And outputs:

wherein,representing the cluster center as +.>The corresponding calculation probability, q is the number of clustering centers,/>Representing the calculated hidden variable +.>，/>Representing hidden variable branch resistance parameters and branch reactance parametersLower boundary calculated by the value, +.>Representing the calculated upper boundary of the values of the hidden variable branch resistance parameter and branch reactance parameter, +.>Representing the lower boundary of the belonging cluster center +.>Or->，/>Upper boundary representing the belonging cluster center +. >Or->。

5. A terminal comprising a processor and a storage medium; the method is characterized in that:

the storage medium is used for storing instructions;

the processor is configured to operate according to the instructions to perform the steps of the method for identifying a medium voltage distribution network parameter interval based on a dynamic bayesian network according to any of claims 1-2.

6. A computer readable storage medium having stored thereon a computer program, which when executed by a processor performs the steps of the medium voltage distribution network parameter interval identification method based on a dynamic bayesian network according to any of claims 1-2.