CN111860617A - Comprehensive optimization operation method for power distribution network - Google Patents
Comprehensive optimization operation method for power distribution network Download PDFInfo
- Publication number
- CN111860617A CN111860617A CN202010623750.XA CN202010623750A CN111860617A CN 111860617 A CN111860617 A CN 111860617A CN 202010623750 A CN202010623750 A CN 202010623750A CN 111860617 A CN111860617 A CN 111860617A
- Authority
- CN
- China
- Prior art keywords
- distribution network
- power distribution
- support vector
- vector machine
- stage
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2458—Special types of queries, e.g. statistical queries, fuzzy queries or distributed queries
- G06F16/2474—Sequence data queries, e.g. querying versioned data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/70—Smart grids as climate change mitigation technology in the energy generation sector
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
-
- 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
Abstract
The invention discloses a comprehensive optimization operation method of a power distribution network, which comprises the steps of constructing a high-dimensional random matrix, extracting system characteristics from operation data of the power distribution network as input of a model, taking a switch state as output, and identifying a topological structure of the current power distribution network by a support vector machine in a training stage 1; the extracted system characteristics and the switch state are used as input together, the reactive power control strategy is used as output, a support vector machine in the stage 2 is trained, the nonlinear mapping relation between input and output is learned, and a power distribution network comprehensive optimization model based on the two-stage support vector machine combination decision is established. The method provided by the invention does not depend on the model and parameters of the power distribution network, has high online decision speed and strong adaptability, can effectively deal with the inaccuracy of the model and parameters of the power distribution network and the uncertainty caused by DG large-scale grid connection, and provides a new way for the optimized operation of the complex power distribution network.
Description
Technical Field
The invention relates to a comprehensive optimization operation method for a power distribution network, and belongs to the technical field of power grid operation.
Background
The power distribution network reconstruction and the power distribution network reactive power optimization are important technical means for optimizing the operation of the power distribution network, the power distribution network reconstruction is to obtain a network topology structure under an optimal optimization target value by changing the closing of a network switch, and the aim of minimum active loss is achieved on the premise of ensuring higher voltage level.
The reactive power optimization of the power distribution network refers to that under the condition of given structural parameters and loads of the power distribution network, on the premise of meeting all constraint conditions, certain performance indexes of the power distribution network are optimized by means of adjusting reactive power output of a power supply, transformer transformation ratio, reactive power compensation and the like, the reconstruction of the power distribution network and the reactive power optimization of the power distribution network are important measures for improving the operation level of the power distribution network, but the maximum optimization of the power distribution network cannot be realized by a single measure, and therefore the two measures need to be comprehensively considered.
With the access of large-scale distributed power sources, electric vehicle random loads and the like, the intermittent supply and utilization and randomness of the distributed power sources and the electric vehicles bring great uncertainty to the power distribution network, and the difficulty of reactive power optimization and power distribution network reconstruction is greatly increased. The existing research improves the calculation efficiency and the convergence to a certain extent, but still does not get rid of the limitations of model simplification and iterative optimization of the traditional method.
The power distribution network reconstruction is a nonlinear combination optimization problem, the power distribution network reactive optimization is a nonlinear integer programming problem, the problem that the power distribution network reactive optimization and the nonlinear integer programming are comprehensively optimized can cause the problem that the solution is more difficult and complex, and the traditional mathematical method has the defects of long calculation time, difficulty in convergence, insufficient precision and the like, so that relevant scholars apply a heuristic method and an artificial intelligence method to the comprehensive optimization of the power distribution network.
The comprehensive optimization of the power distribution network is a complex nonlinear problem, the support vector machine technology can mine hidden effective information from historical data and directly analyze the nonlinear relation between input and output, and therefore in order to guarantee the power supply quality and stable operation of the power distribution network, a comprehensive optimization operation method of the power distribution network needs to be found.
Disclosure of Invention
The invention aims to solve the technical problem of providing a comprehensive optimization operation method for a power distribution network, which has the characteristics of no dependence on a model structure, rapidness and strong universality.
In order to solve the problems, the technical scheme adopted by the invention is as follows:
a comprehensive optimization operation method for a power distribution network comprises the following steps:
step S1, extracting system characteristics from the operation data of the power distribution network as the input of a model by constructing a high-dimensional random matrix, taking the switch state as the output, training a stage 1 support vector machine, and identifying the topological structure of the current power distribution network;
and step S2, the extracted system characteristics and the switch state are used as input, the reactive power control strategy is used as output, the stage 2 support vector machine is trained, the nonlinear mapping relation between the input and the output is learned, and the comprehensive optimization model of the power distribution network based on the two-stage support vector machine combination decision is established.
As a further improvement of the invention, the system characteristics include scene characteristics of load data, photovoltaic power generation, fan power generation, electric vehicle charging data, and local real-time environment data;
the local real-time environmental data includes temperature, wind speed, and light intensity.
As a further improvement of the method, original data of 7 types of scene characteristics are obtained by sampling according to a time sequence from a historical database of the power distribution network, and N multiplied by M dimension random matrixes are respectively constructed, wherein N is the number of state variables, and M is the length of the time sequence; because the number of state variables of DG, electric vehicle random load, environmental factors and the like at a certain sampling moment is less, high-dimensional random matrixes are respectively constructed by adopting a matrix expansion method.
As a further improvement of the invention, the average spectrum radius, the second-order center distance, the maximum spectrum radius, the minimum spectrum radius, the distribution proportion of the feature roots outside/on/in the ring and the matrix mode and the variance of a random matrix constructed by 7 types of original data per hour are calculated according to a single-ring theorem, and then, in addition to the total load, 64 feature variables are constructed per hour and used as a sample feature set input by a support vector machine, and the sample feature set represents the running state of the power distribution network.
As a further improvement of the invention, the two-stage support vector machine combined decision construction process is as follows:
taking the extracted system characteristics and the corresponding switch states as training samples, training the stage 1 support vector machine, and identifying the topological structure of the current power distribution network;
and (3) taking the system characteristics and the corresponding switch states as input, taking the corresponding reactive power control strategy as output, training the stage 2 support vector machine, and learning the mapping relation among the system characteristics, the topological structure and the reactive power optimization control strategy.
As further improvement of the method, the flow of the comprehensive optimization method of the power distribution network based on the two-stage support vector machine combination decision is divided into two parts of off-line training and on-line application;
during off-line training, firstly extracting system characteristics and corresponding switch states from historical data to serve as training samples, and off-line training the stage 1 support vector machine; secondly, taking the system characteristics and the corresponding switch states as input, taking the corresponding reactive power control strategy as output, and training the stage 2 support vector machine off line;
when the device is in online use, statistical characteristics are extracted from measured data and input into a model of the stage 1 support vector machine, and the on-off state combination at the current moment is given; and then inputting the statistical characteristics and the predicted switch state into a trained model of the stage 2 support vector machine together, namely, giving a current reactive power control strategy.
As a further improvement of the method, the offline modeling step of the comprehensive optimization model of the power distribution network based on the two-stage support vector machine combined decision is as follows:
step Q1, sampling and preprocessing of the sample data set;
step Q2, selecting parameters of a combined decision model of the two-stage support vector machine;
step Q3, model performance is evaluated.
As a further improvement of the present invention, the sampling and preprocessing of the sample data set are performed as follows: acquiring original data from a power distribution network historical database, constructing an input characteristic data set through a high-dimensional random matrix, and simultaneously acquiring a reactive power control strategy and a switching state at a corresponding moment to form an output strategy data set, and jointly forming a sample data set; the input data is mapped into a range of [ 0,1 ] by adopting a linear mapping method, as shown in formula 1:
in the formula (I), the compound is shown in the specification,
、
respectively are numerical values before and after the normalization of the ith characteristic quantity;
、
the maximum value and the minimum value of the feature quantity in the sample data set are respectively.
As a further improvement of the invention, the selection process of the parameters of the combined decision model of the two-stage support vector machine is as follows: and selecting a radial basis kernel function, searching optimal parameters c and g by adopting a cross validation method, and training and regression predicting the model by using the obtained optimal parameters.
As a further improvement of the present invention, the process of evaluating the performance of the model is as follows: applying a control strategy generated based on a two-stage support vector machine combined decision model to a test set to test whether the control effect of reducing the active power loss of the system and reducing the node voltage offset can be achieved;
selecting a loss reduction rate
And deviation of voltage
To measure the effect of the reactive power optimization,
the larger the size of the tube is,
the smaller the value, the better the optimization effect, and the formula is defined as:
in the formula (I), the compound is shown in the specification,
line loss before system optimization; a
Line loss after system optimization;
is the actual voltage value of the ith node;
is the rated voltage value of the node; n is the total number of system nodes.
Adopt the produced beneficial effect of above-mentioned technical scheme to lie in:
the invention simultaneously considers the problems of reactive power optimization and reconstruction of the power distribution network, realizes the optimized operation of the power distribution network from the data driving angle, and has the following characteristics:
1) a random matrix is used for carrying out statistical modeling and feature extraction on a large amount of electric state quantity and non-electric environment data generated in the operation of the power distribution network, and the time-space characteristics of the operation of the power distribution network can be effectively reflected.
2) The comprehensive optimization model of the power distribution network based on the two-stage support vector machine combined decision directly excavates the nonlinear mapping relation between the system characteristics and the control strategy, realizes the coordination control between the reactive power optimization control and the power distribution network reconstruction, and obviously improves the optimization effect of the power distribution network.
3) The method provided by the invention does not depend on the model and parameters of the power distribution network, has high online decision speed and strong adaptability, can effectively deal with the inaccuracy of the model and parameters of the power distribution network and the uncertainty caused by DG large-scale grid connection, and provides a new way for the optimized operation of the complex power distribution network.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed for the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without creative efforts.
FIG. 1 is a schematic diagram of a two-stage SVM decision making process of the present invention;
FIG. 2 is a schematic flow chart of a comprehensive optimization method for the power distribution network according to the present invention;
fig. 3 is a diagram of an improved IEEE-37 node topology of the present invention.
Detailed Description
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the application, its application, or uses. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments according to the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, and it should be understood that when the terms "comprises" and/or "comprising" are used in this specification, they specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof, unless the context clearly indicates otherwise.
The relative arrangement of the components and steps, the numerical expressions, and numerical values set forth in these embodiments do not limit the scope of the present application unless specifically stated otherwise. Meanwhile, it should be understood that the sizes of the respective portions shown in the drawings are not drawn in an actual proportional relationship for the convenience of description. Techniques, methods, and apparatus known to those of ordinary skill in the relevant art may not be discussed in detail but are intended to be part of the specification where appropriate. In all examples shown and discussed herein, any particular value should be construed as merely illustrative, and not limiting.
Thus, other examples of the exemplary embodiments may have different values. It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, further discussion thereof is not required in subsequent figures.
In the description of the present application, it is to be understood that the orientation or positional relationship indicated by the directional terms such as "front, rear, upper, lower, left, right", "lateral, vertical, horizontal" and "top, bottom", etc., are generally based on the orientation or positional relationship shown in the drawings, and are used for convenience of description and simplicity of description only, and in the case of not making a reverse description, these directional terms do not indicate and imply that the device or element being referred to must have a particular orientation or be constructed and operated in a particular orientation, and therefore, should not be considered as limiting the scope of the present application; the terms "inner and outer" refer to the inner and outer relative to the profile of the respective component itself.
The main uncertain factor influencing the reactive power optimization and reconstruction results in the traditional power distribution network is load, and after random loads such as DG grid connection and electric vehicles are connected, the uncertain factors in the power distribution network are further increased due to the randomness of power supply and power consumption and the influence of various weather factors, and the running data of the power distribution network also presents random distribution characteristics while periodically changing.
The invention provides a comprehensive optimization operation method for a power distribution network, which comprises the following steps:
step S1, extracting system characteristics from the operation data of the power distribution network as the input of a model by constructing a high-dimensional random matrix, taking the switch state as the output, training a stage 1 support vector machine, and identifying the topological structure of the current power distribution network;
and step S2, the extracted system characteristics and the switch state are used as input, the reactive power control strategy is used as output, the stage 2 support vector machine is trained, the nonlinear mapping relation between the input and the output is learned, and the comprehensive optimization model of the power distribution network based on the two-stage support vector machine combination decision is established.
Further, the system characteristics comprise scene characteristics of load data, photovoltaic power generation, fan power generation, electric vehicle charging data and local real-time environment data;
the local real-time environmental data includes temperature, wind speed, and light intensity.
Further, sampling is carried out on the power distribution network historical database according to a time sequence to obtain original data of 7 types of scene characteristics, N × M dimension random matrixes are respectively constructed, N is the number of state variables, and M is the length of the time sequence; because the number of state variables of DG, electric vehicle random load, environmental factors and the like at a certain sampling moment is less, high-dimensional random matrixes are respectively constructed by adopting a matrix expansion method.
Further, the average spectrum radius, the second-order center distance, the maximum spectrum radius, the minimum spectrum radius, the distribution proportion of the characteristic roots outside/on/in the circular ring and 9 statistical characteristics of the matrix mode and the variance of a random matrix constructed by 7 types of original data per hour are calculated according to a single-ring theorem, and 64 characteristic variables are constructed per hour and serve as a sample characteristic set input by a support vector machine, and the sample characteristic set represents the running state of the power distribution network.
As shown in fig. 1, the two-stage support vector machine combined decision construction process is as follows:
taking the extracted system characteristics and the corresponding switch states as training samples, training the stage 1 support vector machine, and identifying the topological structure of the current power distribution network;
and (3) taking the system characteristics and the corresponding switch states as input, taking the corresponding reactive power control strategy as output, training the stage 2 support vector machine, and learning the mapping relation among the system characteristics, the topological structure and the reactive power optimization control strategy.
As shown in fig. 2, the process of the comprehensive optimization method for the power distribution network based on the two-stage support vector machine combination decision is divided into two parts, namely offline training and online application;
during off-line training, firstly extracting system characteristics and corresponding switch states from historical data to serve as training samples, and off-line training the stage 1 support vector machine; secondly, taking the system characteristics and the corresponding switch states as input, taking the corresponding reactive power control strategy as output, and training the stage 2 support vector machine off line;
When the device is in online use, statistical characteristics are extracted from measured data and input into a model of the stage 1 support vector machine, and the on-off state combination at the current moment is given; and then inputting the statistical characteristics and the predicted switch state into a trained model of the stage 2 support vector machine together, namely, giving a current reactive power control strategy.
Further, the offline modeling step of the comprehensive optimization model of the power distribution network based on the two-stage support vector machine combined decision is as follows:
step Q1, sampling and preprocessing of the sample data set;
step Q2, selecting parameters of a combined decision model of the two-stage support vector machine;
step Q3, model performance is evaluated.
Further, the sampling and preprocessing process of the sample data set is as follows: acquiring original data from a power distribution network historical database, constructing an input characteristic data set through a high-dimensional random matrix, and simultaneously acquiring a reactive power control strategy and a switching state at a corresponding moment to form an output strategy data set, and jointly forming a sample data set; the input data is mapped into a range of [ 0,1 ] by adopting a linear mapping method, as shown in formula 1:
in the formula (I), the compound is shown in the specification,
、
respectively are numerical values before and after the normalization of the ith characteristic quantity; 
、
Respectively, the maximum value of the feature quantity in the sample data setAnd a minimum value.
Further, the selection process of the two-stage support vector machine combined decision model parameters is as follows: and selecting a radial basis kernel function, searching optimal parameters c and g by adopting a cross validation method, and training and regression predicting the model by using the obtained optimal parameters.
Further, the process of evaluating the performance of the model is as follows: applying a control strategy generated based on a two-stage support vector machine combined decision model to a test set to test whether the control effect of reducing the active power loss of the system and reducing the node voltage offset can be achieved;
selecting a loss reduction rate
And deviation of voltage
To measure the effect of the reactive power optimization,
the larger the size of the tube is,
the smaller the value, the better the optimization effect, and the formula is defined as:
in the formula (I), the compound is shown in the specification,
line loss before system optimization; a
Line loss after system optimization;
is the actual voltage value of the ith node;
is the rated voltage value of the node; n is the total number of system nodes.
In order to verify the correctness and the effectiveness of the method, the embodiment improves a classic IEEE-37 node system, adds random loads such as a distributed power supply DG for photovoltaic power generation, fan power generation and the like and an electric vehicle charging station and the like, and accesses reactive power optimization equipment such as a voltage regulator, a capacitor bank, a static reactive power compensator and the like to form an active power distribution network. The system has 37 nodes, 41 branches, 5 tie switches from S1 to S5, and the improved IEEE-37 node topology is shown in fig. 3 assuming that each branch is provided with a sectionalizer.
In order to simulate historical data and control strategies of a power distribution network under traditional reactive power optimization control, OpenDSS and Matlab are used for simulation generation. Firstly, system simulation is carried out in OpenDSS by combining historical load data and environment data of a certain distribution network for one year (8760 h), and relevant data such as load, distributed power output, electric vehicle charging and the like collected on site are simulated.
Then, a high-dimensional random matrix is constructed to extract input features, and an input vector of 8760 historical samples is formed by taking one hour as a sample. According to the hourly load condition of the distribution network in the region, the minimum active network loss and the minimum node voltage deviation of the system are taken as objective functions, a reactive power control strategy and a switching state in each hour are obtained by adopting a particle swarm algorithm, and a corresponding strategy of 8760 historical samples is formed. Partial parameters of the particle swarm algorithm are set as follows: the population specification modulus is 30, the maximum iteration number is 50, the value range of the inertia factor is more than or equal to 0.4 and less than or equal to 0.9, and the value of the learning factor is c1= c2= 2. When power distribution network reconstruction is researched, a decimal coding mode based on loops is adopted, and a control variable is a switch number selected to be disconnected in each loop; when the reactive power optimization problem is researched, the tap position of the transformer, the capacitor bank 1, the capacitor bank 2 and the SVC are used as control variables.
And finally, sorting the input features and the output strategy into a sample database to form a sample data set containing 8760 effective historical samples, wherein the first 75% of samples are used as a training set, and the last 25% of samples are used as a testing set.
And optimizing and selecting core parameters c and g of the model support vector machine at different stages by adopting a cross verification method. In stage 1, the optimal parameters of the support vector machine selected by a cross validation method are c =0.015625 and g = 0.125; similarly, the best parameters of the support vector machine in the phase 2 are c =2 and g = 0.5. The performance of the comprehensive optimization decision model established by the method is evaluated, and the evaluation index results of the test set are shown in table 1.
As can be seen from Table 1, by adopting the method, the system loss can be obviously reduced, the node voltage offset can be reduced, the line loss is averagely reduced by 31.44% when the line loss ratio is not optimized, the voltage deviation is averagely reduced by 5.96%, the running performance of the power distribution network is greatly improved, and the correctness of the established model is verified.
The invention provides a reconstruction-considered two-stage power distribution network comprehensive optimization method based on a random matrix and a support vector machine, simultaneously considers the problems of reactive power optimization and reconstruction of a power distribution network, and realizes the optimized operation of the power distribution network from the data driving angle. Through theoretical analysis and example verification, the main conclusion is as follows:
1) A random matrix is used for carrying out statistical modeling and feature extraction on a large amount of electric state quantity and non-electric environment data generated in the operation of the power distribution network, and the time-space characteristics of the operation of the power distribution network can be effectively reflected.
2) The comprehensive optimization model of the power distribution network based on the two-stage support vector machine combined decision directly excavates the nonlinear mapping relation between the system characteristics and the control strategy, realizes the coordination control between the reactive power optimization control and the power distribution network reconstruction,
the optimization effect of the power distribution network is obviously improved.
3) The method provided by the invention does not depend on the model and parameters of the power distribution network, has high online decision speed and strong adaptability, can effectively deal with the inaccuracy of the model and parameters of the power distribution network and the uncertainty caused by DG large-scale grid connection, and provides a new way for the optimized operation of the complex power distribution network.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; 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; it is obvious as a person skilled in the art to combine several aspects of the invention. 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 comprehensive optimization operation method for a power distribution network is characterized by comprising the following steps:
step S1, extracting system characteristics from the operation data of the power distribution network as the input of a model by constructing a high-dimensional random matrix, taking the switch state as the output, training a stage 1 support vector machine, and identifying the topological structure of the current power distribution network;
and step S2, the extracted system characteristics and the switch state are used as input, the reactive power control strategy is used as output, the stage 2 support vector machine is trained, the nonlinear mapping relation between the input and the output is learned, and the comprehensive optimization model of the power distribution network based on the two-stage support vector machine combination decision is established.
2. The comprehensive optimization operation method for the power distribution network is characterized in that the system characteristics comprise scene characteristics of load data, photovoltaic power generation, fan power generation, electric vehicle charging data and local real-time environment data;
the local real-time environmental data includes temperature, wind speed, and light intensity.
3. The comprehensive optimization operation method of the power distribution network according to claim 2, characterized in that 7 types of original data of the scene features are obtained by sampling according to a time sequence from a historical database of the power distribution network, and N x M-dimensional random matrices are respectively constructed, wherein N is the number of state variables, and M is the length of the time sequence; because the DG, the electric automobile random load and the environmental factors have fewer state variables at a certain sampling moment, a high-dimensional random matrix is respectively constructed by adopting a method of an extended matrix.
4. The comprehensive optimization operation method of the power distribution network according to claim 3, characterized in that an average spectrum radius, a second-order center distance, a maximum spectrum radius, a minimum spectrum radius, a distribution proportion of feature roots outside/on/in a ring, and 9 statistical features of a matrix model and a variance of a random matrix constructed by 7 types of original data per hour are calculated according to a single-ring theorem, and 64 feature variables are constructed per hour in addition to a total load to serve as a sample feature set input by a support vector machine, wherein the sample feature set represents the operation state of the power distribution network.
5. The comprehensive optimization operation method for the power distribution network, according to claim 4, is characterized in that a two-stage support vector machine combined decision construction process is as follows:
taking the extracted system characteristics and the corresponding switch states as training samples, training the stage 1 support vector machine, and identifying the topological structure of the current power distribution network;
and (3) taking the system characteristics and the corresponding switch states as input, taking the corresponding reactive power control strategy as output, training the stage 2 support vector machine, and learning the mapping relation among the system characteristics, the topological structure and the reactive power optimization control strategy.
6. The comprehensive optimization operation method for the power distribution network, according to claim 5, is characterized in that the flow of the comprehensive optimization method for the power distribution network based on the two-stage support vector machine combination decision is divided into two parts, namely offline training and online application;
During off-line training, firstly extracting system characteristics and corresponding switch states from historical data to serve as training samples, and off-line training the stage 1 support vector machine; secondly, taking the system characteristics and the corresponding switch states as input, taking the corresponding reactive power control strategy as output, and training the stage 2 support vector machine off line;
when the device is in online use, statistical characteristics are extracted from measured data and input into a model of the stage 1 support vector machine, and the on-off state combination at the current moment is given; and then inputting the statistical characteristics and the predicted switch state into a trained model of the stage 2 support vector machine together, namely, giving a current reactive power control strategy.
7. The comprehensive optimization operation method for the power distribution network according to claim 6, wherein the offline modeling step of the comprehensive optimization model for the power distribution network based on the two-stage support vector machine combination decision is as follows:
step Q1, sampling and preprocessing of the sample data set;
step Q2, selecting parameters of a combined decision model of the two-stage support vector machine;
step Q3, model performance is evaluated.
8. The method according to claim 7, wherein the sampling and preprocessing of the sample data set are performed as follows: acquiring original data from a power distribution network historical database, constructing an input characteristic data set through a high-dimensional random matrix, and simultaneously acquiring a reactive power control strategy and a switching state at a corresponding moment to form an output strategy data set, and jointly forming a sample data set; the input data is mapped into a range of [ 0, 1 ] by adopting a linear mapping method, as shown in formula 1:
In the formula (I), the compound is shown in the specification,
、
numbers before and after normalization for ith characteristic quantity respectivelyA value;
、
the maximum value and the minimum value of the feature quantity in the sample data set are respectively.
9. The comprehensive optimization operation method for the power distribution network according to claim 7, wherein the selection process of the parameters of the two-stage support vector machine combined decision model is as follows: and selecting a radial basis kernel function, searching optimal parameters c and g by adopting a cross validation method, and training and regression predicting the model by using the obtained optimal parameters.
10. The comprehensive optimization operation method for the power distribution network according to claim 7, wherein the process of evaluating the performance of the model is as follows: applying a control strategy generated based on a two-stage support vector machine combined decision model to a test set to test whether the control effect of reducing the active power loss of the system and reducing the node voltage offset can be achieved;
selecting a loss reduction rate
And deviation of voltage
To measure the effect of the reactive power optimization,
the larger the size of the tube is,
the smaller the value, the better the optimization effect, and the formula is defined as:
in the formula (I), the compound is shown in the specification,
line loss before system optimization; a
Line loss after system optimization;
is the actual voltage value of the ith node;
is the rated voltage value of the node; n is the total number of system nodes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010623750.XA CN111860617A (en) | 2020-07-01 | 2020-07-01 | Comprehensive optimization operation method for power distribution network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010623750.XA CN111860617A (en) | 2020-07-01 | 2020-07-01 | Comprehensive optimization operation method for power distribution network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111860617A true CN111860617A (en) | 2020-10-30 |
Family
ID=72988947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010623750.XA Pending CN111860617A (en) | 2020-07-01 | 2020-07-01 | Comprehensive optimization operation method for power distribution network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111860617A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112487991A (en) * | 2020-12-02 | 2021-03-12 | 国网湖南综合能源服务有限公司 | High-precision load identification method and system based on feature self-learning |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107134790A (en) * | 2016-02-26 | 2017-09-05 | 中国电力科学研究院 | A kind of GA for reactive power optimization control sequence based on big data determines method |
| CN109214708A (en) * | 2018-10-08 | 2019-01-15 | 国网甘肃省电力公司电力科学研究院 | Study of Risk Evaluation Analysis for Power System method based on cross entropy theoretical optimization support vector machines |
| CN109546659A (en) * | 2018-10-19 | 2019-03-29 | 北京交通大学 | Based on random matrix and the matched var Optimization Method in Network Distribution of intelligent scene |
| CN109598386A (en) * | 2018-12-12 | 2019-04-09 | 国网山东省电力公司临沂供电公司 | A kind of accurate analysis method of distribution optimization and system based on deep learning |
| CN110336298A (en) * | 2019-08-17 | 2019-10-15 | 广东博慎智库能源科技发展有限公司 | A kind of idle planing method of the distribution containing distributed generation resource based on integrated intelligent algorithm |
-
2020
- 2020-07-01 CN CN202010623750.XA patent/CN111860617A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107134790A (en) * | 2016-02-26 | 2017-09-05 | 中国电力科学研究院 | A kind of GA for reactive power optimization control sequence based on big data determines method |
| CN109214708A (en) * | 2018-10-08 | 2019-01-15 | 国网甘肃省电力公司电力科学研究院 | Study of Risk Evaluation Analysis for Power System method based on cross entropy theoretical optimization support vector machines |
| CN109546659A (en) * | 2018-10-19 | 2019-03-29 | 北京交通大学 | Based on random matrix and the matched var Optimization Method in Network Distribution of intelligent scene |
| CN109598386A (en) * | 2018-12-12 | 2019-04-09 | 国网山东省电力公司临沂供电公司 | A kind of accurate analysis method of distribution optimization and system based on deep learning |
| CN110336298A (en) * | 2019-08-17 | 2019-10-15 | 广东博慎智库能源科技发展有限公司 | A kind of idle planing method of the distribution containing distributed generation resource based on integrated intelligent algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 安然 等: "基于随机矩阵和支持向量机的两阶段配电网综合优化", 《供用电》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112487991A (en) * | 2020-12-02 | 2021-03-12 | 国网湖南综合能源服务有限公司 | High-precision load identification method and system based on feature self-learning |
| CN112487991B (en) * | 2020-12-02 | 2022-12-16 | 国网湖南综合能源服务有限公司 | High-precision load identification method and system based on characteristic self-learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106655207A (en) | Power distribution network reactive power optimization system and method based on multi-data analysis | |
| CN104037776B (en) | The electric network reactive-load capacity collocation method of random inertial factor particle swarm optimization algorithm | |
| CN107800140A (en) | A kind of large user for considering load characteristic, which powers, accesses decision-making technique | |
| CN104463357A (en) | Method for evaluating random intermittent DG optimized integration based on random optimal power flow | |
| CN106803130B (en) | Planning method for distributed power supply to be connected into power distribution network | |
| Wang et al. | Automatic adjustment method of power flow calculation convergence for large-scale power grid based on knowledge experience and deep reinforcement learning | |
| CN112994017A (en) | Distributed photovoltaic power supply site selection optimization method based on power distribution network probability load flow calculation | |
| CN109412161A (en) | A kind of Probabilistic Load calculation method and system | |
| CN109066819A (en) | A kind of idle work optimization method of the power distribution network based on case reasoning | |
| CN109242136A (en) | A kind of micro-capacitance sensor wind power Chaos-Genetic-BP neural network prediction technique | |
| CN105529703A (en) | Urban power net reconstruction planning method based on power supply capability bottleneck analysis | |
| Abass et al. | Methods comparison for optimal capacitor placement in distribution system | |
| CN112446175B (en) | Distributed power supply site selection optimization method based on power distribution network probability load flow calculation | |
| CN113437756B (en) | Micro-grid optimization configuration method considering static voltage stability of power distribution network | |
| Ni et al. | A review of line loss analysis of the low-voltage distribution system | |
| CN113872213A (en) | Power distribution network voltage autonomous optimization control method and device | |
| CN111860617A (en) | Comprehensive optimization operation method for power distribution network | |
| Wang et al. | Analysis of network loss energy measurement based on machine learning | |
| CN114169117A (en) | Power transmission network extension planning method based on improved particle swarm optimization | |
| CN113591224A (en) | Urban power grid cascading failure risk assessment method and system | |
| CN109698516A (en) | The maximum capacity computing system and method for renewable energy access power distribution network | |
| CN104793107B (en) | A kind of power grid cascading fault determination method based on improvement OPA models | |
| CN113673065A (en) | Loss reduction method for automatic reconstruction of power distribution network | |
| Chen et al. | Research on Real-time Carbon Emission Factor Prediction Method for Load Nodes of Power System based on Gated Recurrent Cell Network | |
| Bai et al. | Correlation analysis and prediction of power network loss based on mutual information and artificial neural network |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20201030 |