WO2020140743A1 - 配电台区源-荷协同接入方法、终端、存储介质 - Google Patents

配电台区源-荷协同接入方法、终端、存储介质 Download PDF

Info

Publication number
WO2020140743A1
WO2020140743A1 PCT/CN2019/125811 CN2019125811W WO2020140743A1 WO 2020140743 A1 WO2020140743 A1 WO 2020140743A1 CN 2019125811 W CN2019125811 W CN 2019125811W WO 2020140743 A1 WO2020140743 A1 WO 2020140743A1
Authority
WO
WIPO (PCT)
Prior art keywords
load
wind speed
distribution
light intensity
access
Prior art date
Application number
PCT/CN2019/125811
Other languages
English (en)
French (fr)
Inventor
戚艳
王旭东
穆云飞
尚学军
骆柏锋
丁一
霍现旭
张志君
吴磊
Original Assignee
国网天津市电力公司电力科学研究院
国网天津市电力公司
国家电网有限公司
天津大学
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 国网天津市电力公司电力科学研究院, 国网天津市电力公司, 国家电网有限公司, 天津大学 filed Critical 国网天津市电力公司电力科学研究院
Priority to US16/759,456 priority Critical patent/US11500341B2/en
Publication of WO2020140743A1 publication Critical patent/WO2020140743A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
    • G05B13/047—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators the criterion being a time optimal performance criterion
    • 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
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • 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
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02S—GENERATION OF ELECTRIC POWER BY CONVERSION OF INFRARED RADIATION, VISIBLE LIGHT OR ULTRAVIOLET LIGHT, e.g. USING PHOTOVOLTAIC [PV] MODULES
    • H02S10/00—PV power plants; Combinations of PV energy systems with other systems for the generation of electric power
    • H02S10/10—PV power plants; Combinations of PV energy systems with other systems for the generation of electric power including a supplementary source of electric power, e.g. hybrid diesel-PV energy systems
    • H02S10/12—Hybrid wind-PV energy 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
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
    • Y02B10/00—Integration of renewable energy sources in buildings
    • Y02B10/10—Photovoltaic [PV]
    • 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
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/50—Photovoltaic [PV] energy
    • 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
    • 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

Definitions

  • the present application belongs to the technical field of access to a power supply system containing electric energy instead of a load, and for example, relates to a source-charge cooperative access method, terminal, and storage medium in a distribution station area.
  • Planners still use more extensive regulations or standards for source-charge access to the station area, which is likely to cause long-term high-load operation in some station areas and light load in some station areas for long periods. Therefore, for the source-charge cooperative access, it is urgent to solve the problem of improving the feeder power supply capacity while ensuring the full absorption of new energy.
  • This application proposes a source-charge cooperative access method, terminal, and storage medium for a power distribution station area, which can reduce the construction cost of the station area, improve the utilization rate of the feeder area of the station area, and fully enhance the ability of the station area to consume clean energy .
  • This application provides a source-charge cooperative access method for a distribution station area, which includes the following steps:
  • Establish the timing characteristic model of the distributed power supply and the timing characteristic model of the load and use the maximum likelihood estimation and the timing characteristic model of the distributed power supply to obtain the distributed power supply of the access station and use the classification regression tree and the timing characteristic model of the load to obtain The timing characteristics of the load of the access station area;
  • establishing a timing characteristic model of a distributed power supply includes:
  • P s is the output power of photovoltaic power generation
  • E is the light intensity
  • D is the photovoltaic panel area
  • ⁇ is the light energy conversion efficiency
  • P w is the output power of the fan
  • v is the natural wind speed
  • V r is the rated wind speed of the fan
  • V ci is the cut-in wind speed of the fan
  • V co is the cut-out wind speed of the fan
  • Pr is the rated output power of the fan
  • a , B and C are the fitting coefficients of the nonlinear part of the fan output.
  • v takes 14m/s
  • V ci takes 4m/s
  • V co takes 25m/s.
  • Using the maximum likelihood estimation and the distributed power supply timing feature model to obtain the distributed power supply timing features of the access station area includes: In the case where the light intensity follows the Beta distribution, according to the access station area The historical data of the measured light intensity, using the maximum likelihood method, to solve the shape parameters ⁇ and ⁇ of the light intensity Beta distribution; based on the shape parameters ⁇ , ⁇ of the light intensity Beta distribution and the Monte Carlo method to simulate the Beta distribution Light intensity, and calculate the time series characteristics of the photovoltaic unit of the access station according to the time series characteristic model of photovoltaic power generation and the simulated light intensity; in the case that the natural wind speed follows Weibull distribution, according to the The historical data of the natural wind speed measured in the station area, using the maximum likelihood method, to solve the shape parameter s and scale parameter l of the Weibull distribution of the natural wind speed; the shape parameter s, scale parameter l and the Weibull distribution based on the natural wind speed The Monte Carlo method simulates the natural wind speed satisfying the Weibull distribution, and calculates the
  • the establishment of load time-series feature model includes: using fuzzy C-Means (FCM) algorithm to classify multiple pieces of historical load data collected by the negative control system to obtain multiple types of load data, thereby extracting the load data of each type Class center as the time series characteristic of each type of load
  • Using the classification regression tree and the time series feature model of the load to obtain the time series characteristics of the load of the access station area includes: constructing a classifier through a classification regression tree (Classification And Regression Trees, CART), and setting the current node of the classifier to t, the current
  • the sample data set of the node is S, which is composed of M samples.
  • the time series feature model of the load it is determined that the M samples belong to L time series features; if the number of samples belonging to the kth class center X k in t is s k , the Gini index (GINI) characterizing the impurity degree of the current node is: In the formula, p(X k
  • t) s k /M represents the proportion of samples belonging to X k in S;
  • using the maximum likelihood method to solve the shape parameters ⁇ and ⁇ of the Beta distribution of the light intensity includes: constructing a log-likelihood Function L( ⁇ , ⁇ ):
  • E i is the light intensity at time i
  • using the maximum likelihood method to solve the shape parameter s and scale parameter l of the Weibull distribution of natural wind speed includes:
  • v i is the natural wind speed at time i
  • the initial value is selected, and iterative iterations are performed until the convergence criterion is met; after convergence, the shape parameter s and scale parameter l of the Weibull distribution are obtained.
  • the shape parameters ⁇ , ⁇ based on the Beta distribution of the light intensity and the Monte Carlo simulation of the light intensity satisfying the Beta distribution include:
  • (.) ⁇ is the gamma function, E m is the maximum light intensity; shape parameter ⁇ and ⁇ is the Beta distribution;
  • the shape parameter s and scale parameter l based on the Weibull distribution of the natural wind speed and the Monte Carlo simulation of the natural wind speed satisfying the Weibull distribution include:
  • s is the shape parameter of Weibull distribution
  • l is the scale parameter of Weibull distribution
  • the fuzzy C-means clustering FCM algorithm is used to classify multiple pieces of historical load data collected by the negative control system, and obtaining multiple types of load data includes:
  • the multiple pieces of historical load data are divided into L categories through the FCM algorithm, where the normalized ith piece of historical load data D i is expressed as:
  • d i,j represents the load active power in the jth period of the ith historical load data
  • m is the number of sampling periods in a day
  • i is an integer greater than 0;
  • centroid-like matrix X is expressed as:
  • N is the number of samples of historical load data
  • the FCM iteratively updates the centroid matrix X and the membership matrix V, and calculates the F of the FCM objective function until F no longer changes, and determines the attribution time series characteristics of the multiple historical load data according to the principle of maximum membership;
  • D i is classified as X k , V i,k satisfies:
  • V i,k max ⁇ V i,1 ,V i,2 ,...,V i,k ,...,V i,L ⁇
  • max ⁇ . ⁇ represents the maximum function
  • the timing characteristic of the new access load i is A i
  • a i can be expressed as:
  • a i [A i,1 ,A i,2 ,...,A i,k ,...,A i,24 ]
  • the combined optimization model is used to determine the feeder to which a certain load is connected when n feeders are available for access in the access station area, and the decision variable of the combined optimization model can be expressed by a 0-1 variable :
  • the load before the kth feeder is not connected is l k , then after the new load is connected, the feeder load L k is updated to:
  • the combined optimization model aims at the maximum power supply capacity of the feeder.
  • the objective function of the combined optimization model is:
  • n represents the number of optional feeders
  • P kN is the upper limit of the k- th feeder's active transmission capacity
  • P k.max represents the maximum load of the k- th feeder
  • Each feeder in the access station area satisfies the following constraint: P k.max ⁇ P kN .
  • An embodiment of the present application also provides a computer-readable storage medium that stores a computer program, and when the computer program is executed by a processor, the method as provided in any embodiment of the present application is implemented.
  • An embodiment of the present application further provides a terminal, which includes a processor and a memory, and a program is stored on the memory, and when the program is executed by the processor, a method as provided in any embodiment of the present application is implemented.
  • FIG. 1 is a flowchart of a source-charge cooperative access method for a distribution station area provided by an embodiment of this application;
  • FIG. 2 is a schematic diagram of source-charge cooperation in a station area provided by an embodiment of the present application
  • FIG. 3a is a schematic diagram of the timing characteristics of photovoltaic power generation provided by an embodiment of the present application.
  • FIG. 3b is a schematic diagram of timing characteristics of wind power generation provided by an embodiment of the present application.
  • 3c is a schematic diagram of load timing feature 1 provided by an embodiment of the present application.
  • 3d is a schematic diagram of load timing feature 2 provided by an embodiment of the present application.
  • 3e is a schematic diagram of load timing feature 3 provided by an embodiment of the present application.
  • 3f is a schematic diagram of load timing feature 4 provided by an embodiment of the present application.
  • 3g is a schematic diagram of load timing feature 5 provided by an embodiment of the present application.
  • FIG. 3h is a schematic diagram of load timing feature 6 provided by an embodiment of the present application.
  • FIG. 4 is a schematic diagram of load data for accessing a feeder in a foreground area provided by an embodiment of the present application
  • 5a is a comparison diagram of daily load curves of feeder 1 before and after optimization provided by an embodiment of the present application
  • FIG. 5b is a comparison diagram of daily load curves of feeder 2 before and after optimization provided by an embodiment of the present application.
  • FIG. 1 is a flowchart of a source-charge cooperative access method for a distribution station area provided by an embodiment of the present application. Referring to FIG. 1, the method provided by the embodiment of the present application includes the following steps:
  • Step 110 Establish a timing characteristic model of the distributed power supply and a timing characteristic model of the load respectively; and obtain the timing characteristics of the distributed power supply of the access station area by using the maximum likelihood estimation and the timing characteristic model of the distributed power supply and use classification
  • the regression tree and the time-series feature model of the load obtain the time-series feature of the load of the access station area.
  • Step 120 Input the timing characteristics of the distributed power source of the access station area and the timing characteristics of the load of the access station area into a combined optimization model, and determine the feeder of the source-charge access through optimization.
  • This application proposes a source-charge cooperative access method in the distribution station area.
  • the timing characteristics of distributed power sources and loads are analyzed.
  • the maximum likelihood estimation and classification regression tree are used to obtain the timing characteristics of distributed power sources and users accessing the station area respectively;
  • the optimization model aims at the maximum power supply capacity of the feeder in the station area, and uses the source-charge combination of the access station as the decision variable to give the power supply access strategy.
  • the effectiveness of the proposed method is verified by an example.
  • the result of the calculation example shows that the source-dutch cooperative access method based on time sequence characteristics proposed in this application can reduce the construction cost of the station area and improve the utilization rate of the feeder in the station area by accessing the optimized combination of the source and charge in the station area. At the same time, it has fully enhanced Taiwan's ability to consume clean energy.
  • This application provides a source-charge cooperative access method in a distribution station area, as shown in FIG. 2, which includes the following steps:
  • Step 10 Analyze the timing characteristics of the distributed power and load, and establish the timing characteristics model of the distributed power and load respectively; and use the maximum likelihood estimation and classification regression tree to obtain the timing characteristics of the distributed power and user access to the station area, respectively .
  • step 10 The steps of step 10 include:
  • the photovoltaic power output power P s is shown in (1):
  • E is the local light intensity
  • D is the photovoltaic panel area
  • ⁇ is the light energy conversion efficiency
  • the light intensity E approximately conforms to the Beta distribution, and its probability density function f(E) and probability distribution function F(E) are shown as (2) and (3) respectively:
  • E m is the maximum light intensity
  • ⁇ is the shape parameter of the Beta distribution
  • the output power of the fan depends on the natural wind speed, and its output power P w is shown in (5):
  • v is the natural wind speed
  • V r is the rated wind speed of the fan
  • V ci and V co are the cut-in wind speed and cut-out wind speed of the fan
  • A, B, and C are the fitting coefficients of the nonlinear part of the fan output.
  • v takes 14m/s
  • V ci takes 4m/s
  • V co takes 25m/s.
  • wind speed v approximates the Weibull distribution, and its probability density function f(v) and probability distribution function F(v) are:
  • the maximum likelihood method is used to solve the parameters of the light intensity Beta distribution and the wind speed Weibull distribution; further, Based on the Monte Carlo method, the solar radiation irradiance and the Weibull distribution wind speed satisfying the Beta distribution are simulated, and the timing characteristics of the photovoltaic units and wind turbines in the station area are calculated according to the photovoltaic power generation and wind power generation model formulas (1) and (5).
  • FCM Fuzzy C-Means
  • d i,j represents the load active power in the j-th period in the i-th historical load data; a total of 96 sampling periods in a day.
  • the historical load data is divided into L categories by FCM algorithm, and the centroid matrix X is expressed as:
  • N is the number of samples of historical load data.
  • the objective function of FCM is that the minimum Euclidean distance between D i and the corresponding centroid X k can be expressed as:
  • FCM updates X and membership matrix V through iterative equations (20) and (21), and calculates F in equation (19) until it no longer changes. According to the principle of maximum membership, determine the time series characteristics of the sample data. If D i is classified as Xk , then the membership must meet:
  • V i,k max ⁇ V i,1 ,V i,2 ,...,V i,k ,...,V i,L ⁇ (22)
  • Accurate identification of the load timing characteristics of the access station area is the basis of the source-load optimized access of the station area.
  • this application uses the decision tree model to realize the identification of the load timing characteristics of the station area.
  • the decision tree model is a widely used non-parametric classifier. First, a sample data set is formed, and each sample has a predetermined set of attributes and corresponding categories. Then, a classifier can be obtained through learning. Finally, through this classification The device can correctly classify the newly appeared objects.
  • the core of the decision tree model is the construction of the classifier.
  • the classification and regression tree (Classification And Regression Trees, CART) is used to construct the classifier.
  • the current node of the classifier is t
  • the current sample data set is S, which is composed of M samples and belongs to L time series features. If the number of samples belonging to X k in t is sk , the Gini index (GINI), which represents the impurity degree of the current node, is:
  • p L GINI(t L ) is the probability that the sample in t is divided into t L
  • p R GINI(t R ) is the probability that the sample in t is divided into t R.
  • the load characteristic index with the largest amount of impurity reduction can be selected as the classification attribute.
  • the classifier can guide the identification of station load time series features.
  • the load characteristic index includes a load rate, a daily minimum load rate, a daily peak-valley difference rate, a peak period load rate, a normal period load rate, and a valley period load rate.
  • the meaning of the user and the load are the same, and the meaning of the installation parameter and the load characteristic index are the same.
  • Step 20 Input the distributed power source of the station area and the timing characteristics of the user into the combined optimization model, and determine the feeder connected to the source-charge by optimizing to obtain the power supply scheme.
  • step 20 is:
  • the application After effectively identifying the source-charge timing characteristics of the access station area, the application proposes an optimization model of source-charge cooperative access based on the timing characteristics, and enters the distributed power supply of the area and the timing characteristics of users into the combined optimization model. Through heuristic algorithm optimization, the feeder connected with source-charge is determined, and the power supply scheme is obtained.
  • a i the distributed power source as a user whose useful power is negative
  • a i [A i,1 ,A i,2 ,...,A i,k ,...,A i,24 ] (25)
  • the core of the combined optimization model is to determine the feeder that a user accesses.
  • the decision variables of the optimization model can be expressed by 0-1 variables:
  • the load before the k-th feeder is not connected is l k , then after connecting a new user, the feeder load L k can be updated to:
  • the combined optimization model aims at the maximum power supply capacity of the feeder:
  • n represents the number of optional feeders
  • P kN is the upper limit of the active transmission capacity of the k- th feeder
  • P k.max represents the maximum load of the k- th feeder.
  • Equations (25)-(29) form the source-charge cooperative optimization access model mentioned in this application, and the particle swarm algorithm is used to solve the optimization results, and the feeder access scheme for the project's upper stage area can be obtained.
  • two-shift and three-shift loads are matched with photovoltaic output, and wind power is matched with night shift output.
  • Wind power and photovoltaic power generation have complementary characteristics in time series. Wind power generation output is small during the day, while photovoltaic power generation output is large, and wind power generation output is large at night, and photovoltaic power generation output is small.
  • the timing characteristics of the load and DG are comprehensively considered, and the complementarity between the DG and the load sequential power is used to ensure the load balance of the station feeder at the planning stage through the optimal combination of DG and load to realize the station area. Economic operation.
  • a source-charge cooperative access method provided in the distribution station area fully considers the complementary characteristics of the distributed power source and the user load timing characteristics as much as possible.
  • the types of DG mentioned in this application are such photovoltaic power generation and wind power generation.
  • the key is to effectively estimate the timing characteristics of the distributed power and load connected to the station area. Considering that the main factor affecting the output power of photovoltaic power generation and wind power generation is the climate environment.
  • this application establishes the timing characteristics models of distributed power sources and loads, respectively, and then uses maximum likelihood estimation and classification regression trees to obtain the timing characteristics of distributed power sources and users accessing the station area, respectively.
  • the electric power enterprise enters the distributed power source of the station area and the sequence characteristics of users into the combined optimization model, and determines the feeder connected to the source-charge through optimization to obtain the power supply scheme.
  • a company's actual station power supply access plan design is used as an example for verification.
  • a station area there are 12 users, 2 photovoltaics with the same specifications, 2 fans with the same specifications to be connected, and two feeders available for access in the station area.
  • the original power supply plan formulated by the power company is shown in Table 1.
  • Power companies combine the user's reporting parameters and use the classification regression tree (lassification And Regression Trees, CART) to mine the classifier to identify the connected user load in the existing load timing characteristics.
  • the user load identification is shown in Table 1.
  • the load of the feeder that is not connected to the reception area can be obtained through the negative control system.
  • the feeder 1 and 2 models are YJV22-8.7/15kV-3*25, and the upper limit of the line active transmission capacity is 1500kW.
  • the load data of feeders that are not connected to the foreground area is shown in Figure 4.
  • One strategy given by solving the power supply access combination optimization model through optimization is shown in Table 1.
  • formula (27) calculate the daily load curve of the feeder under the traditional method and optimization method as shown in Figure 5a and Figure 5b. It can be seen from Figures 5a and 5b that the difference in source-charge timing characteristics has an impact on the formulation of the access scheme. The reason is that, by synthesizing the probability distribution characteristics of 24 periods, the difference complementarity between the distribution characteristics of DG and load time sequence in different periods can be taken into account, so the power supply capacity of the station area has been significantly improved. Taking the original access scheme of feeder 1 as an example, the wind turbine is planned to connect to feeder 1 at this time, and the initial load of feeder 1 is at the peak of electricity consumption from 10:00 to 15:00. In the proposal of this application, the complementarity of source and load characteristics is used. Re-connecting the PV with the peak output at noon time to feeder 1 can realize the peak cut of feeder 1.
  • the daily load rate, daily minimum load rate, and daily peak-valley difference rate indicators are used to further investigate the feeder load before and after optimization.
  • the indexes before and after optimization of the feeder 1 and 2 are calculated as shown in Table 2.
  • feeder 1 As an example, after using the optimization method, the daily load rate of the feeder is increased by 18%, the daily minimum load rate is increased by 22%, and the daily peak-valley difference rate is reduced by 33% compared with the traditional method.

Abstract

本申请涉及一种配电台区源-荷协同接入方法，所述方法包括：分别建立分布式电源的时序特征模型和负荷的时序特征模型；并利用最大似然估计获取的时序特征接入台区的分布式电源的时序特征以及利用分类回归树获取接入台区的用户的时序特征；将接入台区的分布式电源和用户的时序特征输入组合优化模型，通过寻优确定源-荷接入的馈线。

Description

配电台区源-荷协同接入方法、终端、存储介质
本申请要求在2018年12月30日提交中国专利局、申请号为201811648669.6的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本申请属于含电能替代负荷供能系统的接入技术领域，例如涉及一种配电台区源-荷协同接入方法、终端、存储介质。
背景技术
随着国民经济不断发展，配电台区(以下简称台区)负荷特性正发生深刻变化。在荷端，台区下属负荷呈现出多元、互补、关联复杂形态，台区变压器(以下简称台变)峰谷差不断加大，用电高峰期供需矛盾日益突出，电网调峰困难；在源端，以光伏、风电为代表的分布式电源(Distributed generator，DG)大量接入配电台区。然而，DG大量接入除了对配电台区的安全运行提出了更高要求，也相应改变了台区的“源-荷”组合后所表达出的综合负荷特性，甚至出现峰谷差加大，功率倒送等问题。
针对同时接有DG和负荷(以下简称源-荷)的台区，已有大量研究展开，并提出了不同的控制策略，集中在通过合理的调控手段与电价引导机制，有序引导可控负荷与分布式电源的协同优化运行。有的建立了居民用电概率模型，基于负荷状态概率矩阵制定了以分布式光伏消纳最大化为目标的主动负荷需求响应方案；有的文献提出的电动汽车互动充电控制策略，能够使电动汽车匹配光伏出力，促进光伏消纳；有的文献提出了可控负荷，DG和储能装置的协同调度模型，支撑可再生能源的全额消纳；有的文献提出了一种基于改进Colored power算法的加权系数排队算法，可用于直接控制家居环境常见的空调、热泵等温控负荷参与需求侧响应，促进分布式电源优化利用；有的文献通过对电采暖负荷的优化调度，体现了电采暖负荷在消纳风电方面的协同潜力；有的文献建立了计及空调负荷群控制的源-荷协同优化调度模型，降低了发电成本和负荷控制成本。然而，已有文献多侧重于台区内源-荷间的协同控制。
规划人员依然采用较为粗放的规程或标准对台区进行源-荷接入，易造成某些台区长期高负载运行，而一些台区则长期轻载。因此对于源-荷协同接入，在保证充分消纳新能源的同时提高馈线供电能力，是亟待解决的问题。
发明内容
本申请提出一种配电台区源-荷协同接入方法、终端、存储介质，能够降低台区建设成本，提升台区馈线的利用率的同时，充分提升台区对清洁能源的消纳能力。
本申请提供的一种配电台区源-荷协同接入方法，包括以下步骤：
建立分布式电源的时序特征模型和负荷的时序特征模型；并利用最大似然估计和分布式电源的时序特征模型获取接入台区的分布式电源以及利用分类回归树和负荷的时序特征模型获取所述接入台区的负荷的时序特征；
将接入台区的分布式电源的时序特征输入和负荷的时序特征输入组合优化模型，通过寻优确定源-荷接入的馈线。
在一实施例中，建立分布式电源的时序特征模型包括：
建立立光伏发电的时序特征模型：
P s＝EDη
上式中，P s为光伏发电的输出功率，E为光照强度；D为光伏板面积；η为光能转换效率；
建立风机的时序特征模型：
Figure PCTCN2019125811-appb-000001
其中，
Figure PCTCN2019125811-appb-000002
其中，P w为风机的输出功率，v为自然风风速；V r为风机额定风速；V ci为风机的切入风速，V co为风机的切出风速；P r为风机的额定输出功率；A、B、C分别为风机出力非线性部分的拟合系数。本申请v取14m/s；V ci取4m/s；V co取25m/s。
(2)利用最大似然估计和分布式电源的时序特征模型获取接入台区的分布式电源的时序特征包括：在所述光照强度服从Beta分布的情况下，根据在所述接入台区测得的光照强度的历史数据，利用最大似然法，求解光照强度Beta分布的形状参数λ和μ；基于所述光照强度Beta分布的形状参数λ、μ以及蒙特卡洛方法模拟满足Beta分布的光照强度，并根据光伏发电的时序特征模型和模拟的所述光照强度计算所述接入台区光伏机组的时序特征；在所述自然风风速服从Weibull分布的情况下，根据在所述接入台区测得的自然风风速的历史数据，利用最大似然法，求解自然风风速Weibull分布的形状参数s和尺度参数l；基于所述自然风风速Weibull分布的形状参数s、尺度参数l以及所述蒙特卡洛方 法模拟满足Weibull分布的自然风风速，并根据所述风机的时序特征模型和模拟的所述自然风风速计算所述接入台区的风机的时序特征
建立负荷的时序特征模型包括：利用模糊C均值聚类(Fuzzy C-Means，FCM)算法将负控系统采集的多条历史负荷数据进行分类，获得多类负荷数据，从而提取每类负荷数据的类心作为所述每类负荷的时序特征
利用分类回归树和负荷的时序特征模型获取所述接入台区的负荷的时序特征包括：通过分类回归树(Classification And Regression Trees，CART)构建分类器，设分类器的当前节点为t，当前节点的样本数据集为S，由M个样本组成，通过负荷的时序特征模型确定所述M个样本分属于L个时序特征；若t中属于第k个类心X k的样本个数为s k，则表征当前节点的杂质度的基尼指数(Gini index,GINI)为：
Figure PCTCN2019125811-appb-000003
式中，p(X k|t)＝s k/M表示归属X k的样本在S中的比重；
采用负荷特性指标将t分裂，产生两子节点t R和t L，子节点GINI为GINI(t R)和GINI(t L)，则本次分裂的杂质度削减量为：Φ(t)＝GINI(t)-p RGINI(t R)-p LGINI(t L)式中，p LGINI(t L)为t中样本被分到t L的概率；p RGINI(t R)为t中样本被分到t R的概率；杂质削减量最大的负荷特性指标选为分类属性。重复上述过程，不断选择分类属性，直到满足最小GINI即可得到最终的分类器，结合新接入负荷的报装参数以及所述分类器识别所述接入台区的负荷时序特征。
在一实施例中，所述根据在所述接入台区测得的光照强度的历史数据，利用极大似然法，求解光照强度Beta分布的形状参数λ和μ包括：构造对数似然函数L(λ，μ)：
Figure PCTCN2019125811-appb-000004
其中，E i为时刻i的光照强度；
令：
Figure PCTCN2019125811-appb-000005
Figure PCTCN2019125811-appb-000006
使用牛顿拉夫逊法迭代求解F 1和F 2，得修正方程式：
Figure PCTCN2019125811-appb-000007
根据所述修正方程，选取初始值，经过反复迭代直到满足收敛判据；收敛后得出Beta分布的形状参数λ和μ；
所述根据根据在所述接入台区测得的自然风风速的历史数据，利用最大似然法，求解自然风风速Weibull分布的形状参数s和尺度参数l包括：
Figure PCTCN2019125811-appb-000008
其中，v i为时刻i的自然风风速；
令：
Figure PCTCN2019125811-appb-000009
Figure PCTCN2019125811-appb-000010
使用牛顿拉夫逊法迭代求解F 3和F 4，得修正方程式：
Figure PCTCN2019125811-appb-000011
根据所述修正方程，选取初始值，经过反复迭代直到满足收敛判据；收敛后得出Weibull分布的形状参数s和尺度参数l。
在一实施例中，所述基于所述光照强度Beta分布的形状参数λ、μ以及蒙特卡洛方法模拟满足Beta分布的光照强度包括：
基于所述光照强度Beta分布的形状参数λ和μ获取所述光照强度E的概率分布函数F(E)；根据所述光照强度E的概率分布函数F(E)和蒙特卡洛方法模拟满足Beta分布的光照强度；
其中，
Figure PCTCN2019125811-appb-000012
Γ(.)是伽玛函数，E m为光照强度最大值；λ和μ为Beta分布的形状参数；
所述基于所述自然风风速Weibull分布的形状参数s和尺度参数l以及所述蒙特卡洛方法模拟满足Weibull分布的自然风风速包括：
基于所述自然风风速Weibull分布的形状参数s和尺度参数l以获取所述自然风风速v的概率分布函数F(v)；根据所述自然风风速v的概率分布函数F(v)和所述蒙特卡洛方法模拟满足Weibull分布的自然风风速；
其中，所述自然风风速v的概率分布函数F(v)为：
Figure PCTCN2019125811-appb-000013
s为Weibull分布的形状参数；l为Weibull分布的尺度参数。
在一实施例中，利用模糊C均值聚类FCM算法将负控系统采集的多条历史 负荷数据进行分类，获得多类负荷数据包括：
对多条历史负荷数据，通过FCM算法将所述多条历史负荷数据分为L类，其中，归一化的第i条历史负荷数据D i表示为：
D i＝[d i,1,d i,2,…,d i,j,…,d i,m]
式中d i,j表示第i条历史负荷数据中第j时段负荷有功功率；m为一天采样时段的个数，i为大于0的整数；
类心矩阵X表示为：
X＝[X 1,X 2,...,X k,...,X L] T
D i对第k个聚类中心X k的隶属度V i,k满足如下条件：
Figure PCTCN2019125811-appb-000014
其中，N为历史负荷数据的样本数；
FCM的目标函数为：
Figure PCTCN2019125811-appb-000015
通过拉格朗日变换求解上式得：
Figure PCTCN2019125811-appb-000016
Figure PCTCN2019125811-appb-000017
FCM通过迭代更新类心矩阵X和隶属度矩阵V，并计算FCM的目标函数的F直到F不再变化，并按照最大隶属度原则确定所述多条历史负荷数据的归属时序特征；其中，在D i归类于X k的情况下，V i,k满足：
V i,k＝max{V i,1,V i,2,…,V i,k,…,V i,L}
其中，max{.}表示最大值函数。
在一实施例中，所述新接入负荷i的时序特征为A i，A i可表达为：
A i＝[A i,1,A i,2,...,A i,k,...,A i,24]
所述组合优化模型用于在所述接入台区有n条馈线可供接入的情况下，确定某一负荷接入的馈线，所述组合优化模型的决策变量可以用0-1变量表示：
Figure PCTCN2019125811-appb-000018
第k条馈线未接入前的负荷为l k，则接入新负荷后，馈线负载L k更新为：
Figure PCTCN2019125811-appb-000019
组合优化模型为以馈线最大供电能力为目标，所述组合优化模型的目标函数为：
Figure PCTCN2019125811-appb-000020
式中，n表示可选馈线的数量；P k.N为第k条馈线有功传输容量上限，P k.max表示第k条馈线的最大负载；
所述接入台区内的每条馈线满足如下约束：P k.max≤P k.N。
本申请实施例还提供一种计算机可读存储介质，存储有计算机程序，所述计算机程序被处理器执行时实现如本申请任意实施例提供的方法。
本申请实施例还提供一种终端，包括处理器和存储器，所述存储器上存储有程序，所述程序被所述处理器执行时实现如本申请任意实施例提供的方法。
附图说明
图1为本申请实施例提供的配电台区源-荷协同接入方法的流程图；
图2为本申请实施例提供的台区的源-荷协同示意图；
图3a为本申请实施例提供的光伏发电时序特征示意图；
图3b为本申请实施例提供的风力发电时序特征示意图；
图3c为本申请实施例提供的负荷时序特征1示意图；
图3d为本申请实施例提供的负荷时序特征2示意图；
图3e为本申请实施例提供的负荷时序特征3示意图；
图3f为本申请实施例提供的负荷时序特征4示意图；
图3g为本申请实施例提供的负荷时序特征5示意图；
图3h为本申请实施例提供的负荷时序特征6示意图；
图4为本申请实施例提供的接入前台区馈线的负荷数据示意图；
图5a为本申请实施例提供的优化前后馈线1的日负荷曲线对比图；
图5b为本申请实施例提供的优化前后馈线2的日负荷曲线对比图。
具体实施方式
以下结合附图对本申请实施例作进一步描述：
图1为本申请实施例提供的配电台区源-荷协同接入方法的流程图。参见图1，本申请实施例提供的方法包括如下步骤：
步骤110、分别建立分布式电源的时序特征模型和负荷的时序特征模型；并 利用最大似然估计和所述分布式电源的时序特征模型获取接入台区的分布式电源的时序特征以及利用分类回归树和所述负荷的时序特征模型获取所述接入台区的负荷的时序特征。
步骤120、将所述接入台区的分布式电源的时序特征和所述接入台区的负荷的时序特征输入组合优化模型，通过寻优确定源-荷接入的馈线。
本申请提出了一种配电台区源-荷协同接入方法。首先，分析了分布式电源以及负荷的时序特征，其次利用最大似然估计以及分类回归树分别获取接入台区的分布式电源和用户的时序特征；进而，提出台区源-荷接入组合优化模型，以台区馈线最大供电能力为目标，以接入台区的源-荷组合为决策变量，给出供电接入策略；最后，通过算例验证所提方法的有效性。算例结果表明，本申请提出的基于时序特征的台区源-荷协同接入方法，可通过接入台区源-荷的优化组合，在降低台区建设成本，提升台区馈线的利用率的同时，充分提升了台区对清洁能源的消纳能力。
本申请提供的一种配电台区源-荷协同接入方法，如图2所示，包括以下步骤：
步骤10、分析分布式电源以及负荷的时序特征，分别建立分布式电源和负荷的时序特征模型；并利用最大似然估计以及分类回归树分别获取接入台区的分布式电源和用户的时序特征。
所述步骤10的步骤包括：
(1)分析分布式电源时序特征：
光伏发电输出功率P s如(1)所示：
P s＝EDη         (1)
上式中，E为地区光照强度；D为光伏板面积；η为光能转换效率；
其中，光照强度E近似符合Beta分布，其概率密度函数f(E)与概率分布函数F(E)分别如(2)和(3)所示：
Figure PCTCN2019125811-appb-000021
Figure PCTCN2019125811-appb-000022
上式中，E m为光照强度最大值；λ、μ为Beta分布的形状参数；
根据式(2)、(3)经过等价变换得到模拟光照强度如公式(4)所示：
E＝E mF -1(E)       (4)
风机输出功率取决于自然风风速，其输出功率P w如(5)所示：
Figure PCTCN2019125811-appb-000023
Figure PCTCN2019125811-appb-000024
Figure PCTCN2019125811-appb-000025
Figure PCTCN2019125811-appb-000026
其中，v为自然风风速；V r为风机额定风速；V ci、V co分别为风机的切入风速、切出风速；A、B、C分别为风机出力非线性部分的拟合系数。本申请v取14m/s；V ci取4m/s；V co取25m/s。
风速v近似符合Weibull分布，其概率密度函数f(v)和概率分布函数F(v)分别为：
Figure PCTCN2019125811-appb-000027
Figure PCTCN2019125811-appb-000028
式中，s、l分别为Weibull分布的形状参数、尺度参数。根据式(9)、(10)经过等价变换得到模拟风速为：
v＝-l(lnF) 1/s      (11)
(2)辨识台区分布式电源时序特征；
在光照强度和风速分别服从Beta分布和Weibull分布前提下，根据台区测得的光照强度和风速的历史数据，利用极大似然法，求解光照强度Beta分布以及风速Weibull分布的参数；进而，基于蒙特卡洛方法模拟满足Beta分布的太阳光辐照度和Weibull分布的风速，并根据光伏发电和风力发电的出力模型式(1)、(5)计算台区光伏机组和风机的时序特征。
最大似然估计根据地区风速和光照强度历史数据样本出现的概率最大的原则，来求Beta分布和Weibull分布参数的估计值，以求解Beta分布参数为例，构造对数似然函数L(λ，μ)：
Figure PCTCN2019125811-appb-000029
令：
Figure PCTCN2019125811-appb-000030
Figure PCTCN2019125811-appb-000031
使用牛顿拉夫逊法迭代求解式(13)、(14)，可得相应的修正方程式：
Figure PCTCN2019125811-appb-000032
根据(15)，选取合适的初始值，经过反复迭代直到满足收敛判据。收敛后就可得出求得Beta分布的参数λ、μ。同理，亦可求得Weibull分布的参数s、l。
(3)分析负荷时序时序特征
接入台区的负荷种类丰富，不同类别的电力负荷特点不尽相同。为最终实现台区供电接入的优化，有必要分析多种类型负荷特点及用电习惯。本申请利用模糊C均值聚类(Fuzzy C-Means，FCM)算法将负控系统采集的用户历史负荷数据进行分类，从而提取多类类心分别作为多类负荷的时序特征。归一化的历史负荷数据D i可表示为：
D i＝[d i,1,d i,2,…,d i,j,…,d i,96]           (16)
式中d i,j表示第i条历史负荷数据中第j时段负荷有功功率；一天共96个采样时段。对历史负荷数据，通过FCM算法将其分为L类，其中类心矩阵X表示为：
X＝[X 1,X 2,...,X k,...,X L] T        (17)
历史负荷数据的任一样本D i对第k个聚类中心X k的隶属度V i,k应有：
Figure PCTCN2019125811-appb-000033
式中，N为历史负荷数据的样本数。
FCM的目标函数为D i与对应类心X k的欧式距离最小，可表达为：
Figure PCTCN2019125811-appb-000034
通过拉格朗日变换求解式(19)的可得：
Figure PCTCN2019125811-appb-000035
Figure PCTCN2019125811-appb-000036
FCM通过迭代式(20)、(21)更新X和隶属度矩阵V，并计算式(19)的F直到不再变化。按照最大隶属度原则确定样本数据的归属时序特征，若D i归类于 Xk，则其隶属度需满足：
V i,k＝max{V i,1,V i,2,…,V i,k,…,V i,L}        (22)
其中max{.}表示最大值函数；
(4)辨识台区负荷时序特征辨识
准确识别接入台区的负荷时序特征，是台区源-荷优化接入的基础，对此本申请通过决策树模型实现台区负荷时序特征识别。决策树模型是一种广泛使用的非参数分类器，首先形成样本数据集，且每个样本都有事先确定的一组属性和对应类别，接着通过学习可得到一个分类器，最后，通过这个分类器可对新 出现的对象给出正确的分类。
决策树模型的核心是分类器的构建，本申请通过分类回归树(Classification And Regression Trees，CART)构建分类器。设分类器的当前节点为t，当前的样本数据集为S，由M个样本组成，分属于L个时序特征。若t中属于X k的样本个数为s k，则表征当前节点的杂质度的基尼指数(Gini index,GINI)为：
Figure PCTCN2019125811-appb-000037
式中p(X k|t)＝s k/M表示归属X k的样本在S中的比重。采用的负荷特性指标将t分裂产生两子节点t R和t L，子节点GINI为GINI(t R)和GINI(t L)，则本次分裂的杂质度削减量为：
Φ(t)＝GINI(t)-p RGINI(t R)-p LGINI(t L)       (24)
式中p LGINI(t L)为t中样本被分到t L的概率；p RGINI(t R)为t中样本被分到t R的概率。杂质削减量最大的负荷特性指标即可选为分类属性。
重复上述过程，不断选择分类属性，直到满足最小GINI即可得到最终的分类器。结合新接入用户的报装参数，分类器可指导台区负荷时序特征识别。
一实施例中，负荷特性指标包括负荷率，日最小负荷率，日峰谷差率，峰期负载率，平期负载率，谷期负载率。
本申请实施例中，用户与负荷的含义相同，且报装参数与负荷特性指标的含义相同。
步骤20、将台区的分布式电源和用户的时序特征输入组合优化模型，通过寻优确定源-荷接入的馈线，得出供电方案。
所述步骤20的方法是：
在有效识别接入台区的源-荷时序特征后，本申请提出基于时序特征的台区源-荷协同接入优化模型，将台区的分布式电源和用户的时序特征输入组合优化模型，通过启发式算法寻优确定源-荷接入的馈线，得出供电方案。
设接入前识别出用户i(分布式电源作为有用功率为负的用户)时序特征为A i，可表达为：
A i＝[A i,1,A i,2,...,A i,k,...,A i,24]          (25)
如图2所示，若当前台区有n条馈线可供接入，则组合优化模型的核心就是确定某一用户接入的馈线，此时优化模型的决策变量可以用0-1变量表示：
Figure PCTCN2019125811-appb-000038
第k条馈线未接入前的负荷为l k，则接入新用户后，馈线负载L k可以更新为：
Figure PCTCN2019125811-appb-000039
组合优化模型为以馈线最大供电能力为目标：
Figure PCTCN2019125811-appb-000040
式中，n表示可选馈线的数量；P k.N为第k条馈线有功传输容量上限，P k.max表 示第k条馈线的最大负载。
出于台区安全性的考虑，还要求台区内每条馈线日峰值负荷小于该条馈线的额定传输容量，则有如下约束：
P k.max≤P k.N      (29)
式(25)-(29)组成了本申请所提源-荷协同优化接入模型，通过粒子群算法求解优化结果，即可得到工程上台区馈线接入方案。
本申请的工作原理是：
传统供电接入方案制定仅考虑负荷或者分布式电源的报装值，通过报装值的加和得到接入方案，与实际情况存在较大差异。例如，负荷的报装值加和高于DG出力和，当负荷处于峰值而DG出力低谷时，台区内潮流不反向，但是某些时段，当DG出力处于高峰而此时台区内负荷较小时，有可能出现潮流反向。考虑到负荷需求与DG的出力存在时间分布上的差异，这给二者的协同接入提供了优化空间。例如两班型和三班型负荷与光伏出力匹配，风力发电与夜班型出力匹配。风力和光伏发电在时序上呈现出互补的特性。在白天风力发电出力较小，而光伏发电出力较大，在晚上风力发电出力较大，而光伏发电出力较小。综上所述，只有综合考虑负荷与DG的时序特征，利用DG同负荷时序功率之间的互补性，通过DG与负荷的优化组合，在规划阶段上保证台区馈线负载平衡，从而实现台区的经济运行。
为此本申请提供的一种配电台区源-荷协同接入方法，尽可能充分考虑到分布式电源和用户负荷时序特征的互补特性。光伏发电和风力发电作为分布式清洁能源的代表，常在低压台区分散接入、就地消纳。因此本申请所提的DG类型即为该类光伏发电和风力发电。对于图2所示的台区，为充分挖掘源-荷时序互补性，关键在于有效估计接入台区的分布式电源和负荷的时序特征。考虑到影响光伏发电和风力发电输出功率的主要因素为气候环境。因此明晰台区风速和光照的时序分布，对于模拟分布式电源的出力至关重要。对此，本申请分别建立分布式电源以及负荷的时序特征模型，其次利用最大似然估计以及分类回归树分别获取接入台区的分布式电源和用户的时序特征。电力企业将台区的分布式电源和用户的时序特征输入组合优化模型，通过寻优确定源-荷接入的馈线，得出供电方案。
为验证本申请所提方法的有效性，以某公司实际台区供电接入方案设计为例进行验证。在某台区，有12户用户，2台规格相同的光伏，2台规格相同的风机待接入，台区内可供接入的馈线有两条。电力公司制定的原供电方案如表1所示。
首先，根据台区测得的光照强度和风速的历史数据，利用极大似然法求得Beta分布的参数λ＝0.61、μ＝2.63，Weibull分布的参数s＝2.43、l＝8.46。进而，通过蒙特卡洛方法模拟光辐照度和风速，并根据式(1)，(5)计算台区光伏机组和 风机的时序特征。另一方面，选取456户用户通过FCM算法分为L＝6类获取负荷时序特征，由此获取如图3a～图3h所示的台区源-荷时序特征：
电力企业结合用户的报装参数，利用分类回归树(lassification And Regression Trees，CART)挖掘出的分类器，在既有负荷时序特征中识别接入的用户负荷，用户负荷识别情况见表1。未接入前台区馈线的负荷可通过负控系统获得，馈线1、2型号为YJV22-8.7/15kV-3*25，线路有功传输容量上限都为1500kW。未接入前台区馈线的负荷数据见图4，通过优化求解供电接入组合优化模型给出的一个策略如表1所示。
表1接入方案对比
Table 1Power supply access strategy
Figure PCTCN2019125811-appb-000041
按式(27)分别计算传统方法和优化方法下馈线日负荷曲线如图5a和图5b所示。由图5a和图5b可知，源-荷时序特征差异性对接入方案的制定产生影响。原因为，综合24个时段概率分布特性，能计及不同时段DG与负荷时序分布特征的差异互补性，因此台区的供电能力达到显著提升。以馈线1原有的接入方案中为例，风机此时规划接入馈线1，而10时到15时馈线1初始负荷处于用电高峰，本申请方案中，利用源荷特征的互补性，将中午时段出力处于峰值的光伏改接到馈线1，即可实现馈线1的削峰。
在本实施例中，通过日负荷率，日最小负荷率，日峰谷差率指标进一步考察优化前后的馈线负载情况，计算馈线1、2优化前后的指标如表2所示。
表2馈线负荷特性指标计算结果
Table 2Load characteristics index for large consumers
Figure PCTCN2019125811-appb-000042
Figure PCTCN2019125811-appb-000043
考虑源-荷特征的优化接入，多项指标均优于传统方法。以馈线1为例，采用优化方法后，与传统方法相比，馈线日负荷率提高了18％，日最小负荷率提高了22％，日峰谷差率降低了33％。

Claims (8)

  1. 一种配电台区源-荷协同接入方法，包括：
    分别建立分布式电源的时序特征模型和负荷的时序特征模型；并利用最大似然估计和所述分布式电源的时序特征模型获取接入台区的分布式电源的时序特征以及利用分类回归树和所述负荷的时序特征模型获取所述接入台区的负荷的时序特征；
    将所述接入台区的分布式电源的时序特征和所述接入台区的负荷的时序特征输入组合优化模型，通过寻优确定源-荷接入的馈线。
  2. 根据权利要求1所述的方法，其中，
    所述建立分布式电源的时序特征模型包括：
    建立光伏发电的时序特征模型：
    P s＝EDη
    其中，P s为光伏发电的输出功率，E为光照强度；D为光伏板面积；η为光能转换效率；
    建立风机的时序特征模型：
    Figure PCTCN2019125811-appb-100001
    其中，
    Figure PCTCN2019125811-appb-100002
    P w为风机的输出功率，v为自然风风速；V r为风机额定风速；V ci为风机的切入风速；V co为风机的切出风速；P r为风机的额定输出功率；A、B、C分别为风机出力非线性部分的拟合系数；V ci取4m/s；V co取25m/s；
    所述利用最大似然估计和所述分布式电源的时序特征模型获取接入台区的分布式电源的时序特征包括：在所述光照强度服从Beta分布的情况下，根据在所述接入台区测得的光照强度的历史数据，利用最大似然法，求解光照强度Beta分布的形状参数λ和μ；基于所述光照强度Beta分布的形状参数λ、μ以及蒙特卡洛方法模拟满足Beta分布的光照强度，并根据光伏发电的时序特征模型和模拟的所述光照强度计算所述接入台区的光伏机组的时序特征；在所述自然风风速服从Weibull分布的情况下，根据在所述接入台区测得的自然风风速的历史数据，利用最大似然法，求解自然风风速Weibull分布的形状参数s和尺度参数l；基于所述自然风风速Weibull分布的形状参数s、尺度参数l以及所述蒙特卡洛 方法模拟满足Weibull分布的自然风风速，并根据所述风机的时序特征模型和模拟的所述自然风风速计算所述接入台区的风机的时序特征；
    所述建立负荷的时序特征模型包括：利用模糊C均值聚类FCM算法将负控系统采集的多条历史负荷数据进行分类，获得多类负荷数据，提取每类负荷数据的类心作为所述每类负荷数据的时序特征。
    所述利用分类回归树和所述负荷的时序特征模型获取所述接入台区的负荷的时序特征包括：
    通过分类回归树CART构建分类器，设分类器的当前节点为t，当前节点的样本数据集为S，样本数据集由M个样本组成，通过所述负荷的时序特征模型确定所述M个样本分属于L个时序特征，L为大于0的整数；在t中属于第k个类心X k的样本个数为s k的情况下，表征当前节点t的杂质度的基尼指数GINI为：
    Figure PCTCN2019125811-appb-100003
    其中，p(X k|t)＝s k/M表示归属X k的样本在S中的比重；
    采用负荷特性指标将t分裂，产生两子节点t R和t L，子节点GINI为GINI(t R)和GINI(t L)，则本次分裂的杂质度削减量为：Φ(t)＝GINI(t)-p RGINI(t R)-p LGINI(t L)；
    其中，p LGINI(t L)为t中样本被分到t L的概率；p RGINI(t R)为t中样本被分到t R的概率；杂质削减量最大的负荷特性指标选为分类属性；
    重复上述过程，不断选择分类属性，直到满足最小GINI，得到最终的分类器；根据新接入负荷的报装参数以及所述分类器识别所述接入台区的所述新接入负荷的负荷时序特征。
  3. 根据权利要求2所述的方法，其中，所述根据在所述接入台区测得的光照强度的历史数据，利用极大似然法，求解光照强度Beta分布的形状参数λ和μ包括：构造对数似然函数L(λ，μ)：
    Figure PCTCN2019125811-appb-100004
    其中，E i为时刻i的光照强度；
    令：
    Figure PCTCN2019125811-appb-100005
    Figure PCTCN2019125811-appb-100006
    使用牛顿拉夫逊法迭代求解F 1和F 2，得修正方程式：
    Figure PCTCN2019125811-appb-100007
    根据所述修正方程，选取初始值，经过反复迭代直到满足收敛判据；收敛后得出Beta分布的形状参数λ和μ；
    所述根据根据在所述接入台区测得的自然风风速的历史数据，利用最大似然法，求解自然风风速Weibull分布的形状参数s和尺度参数l包括：
    Figure PCTCN2019125811-appb-100008
    其中，v i为时刻i的自然风风速；
    令：
    Figure PCTCN2019125811-appb-100009
    Figure PCTCN2019125811-appb-100010
    使用牛顿拉夫逊法迭代求解F 3和F 4，得修正方程式：
    Figure PCTCN2019125811-appb-100011
    根据所述修正方程，选取初始值，经过反复迭代直到满足收敛判据；收敛后得出Weibull分布的形状参数s和尺度参数l。
  4. 根据权利要求3所述的方法，其中，所述基于所述光照强度Beta分布的形状参数λ、μ以及蒙特卡洛方法模拟满足Beta分布的光照强度包括：
    基于所述光照强度Beta分布的形状参数λ和μ获取所述光照强度E的概率分布函数F(E)；根据所述光照强度E的概率分布函数F(E)和蒙特卡洛方法模拟满足Beta分布的光照强度；
    其中，
    Figure PCTCN2019125811-appb-100012
    Γ(.)是伽玛函数，E m为光照强度最大值；λ和μ为Beta分布的形状参数；
    所述基于所述自然风风速Weibull分布的形状参数s、尺度参数l以及所述蒙特卡洛方法模拟满足Weibull分布的自然风风速包括：
    基于所述自然风风速Weibull分布的形状参数s和尺度参数l获取所述自然 风风速v的概率分布函数F(v)；根据所述自然风风速v的概率分布函数F(v)和所述蒙特卡洛方法模拟满足Weibull分布的自然风风速；
    其中，所述自然风风速v的概率分布函数F(v)为：
    Figure PCTCN2019125811-appb-100013
    s为Weibull分布的形状参数；l为Weibull分布的尺度参数。
  5. 根据权利要求2-4任一项所述的方法，其中，所述利用模糊C均值聚类FCM算法将负控系统采集的多条历史负荷数据进行分类，获得多类负荷数据包括：
    对多条历史负荷数据，通过FCM算法将所述多条历史负荷数据分为L类，其中，归一化的第i条历史负荷数据D i表示为：
    D i＝[d i,1,d i,2,···,d i,j,···,d i,m]
    式中d i,j表示第i条历史负荷数据中第j时段负荷有功功率；m为一天采样时段的个数，i为大于0的整数；
    类心矩阵X表示为：
    X＝[X 1,X 2,...,X k,...,X L] T
    D i对第k个类心X k的隶属度V i,k满足如下条件：
    Figure PCTCN2019125811-appb-100014
    其中，N为历史负荷数据的样本数；
    FCM的目标函数为：
    Figure PCTCN2019125811-appb-100015
    通过拉格朗日变换求解上式得：
    Figure PCTCN2019125811-appb-100016
    Figure PCTCN2019125811-appb-100017
    FCM通过迭代更新类心矩阵X和隶属度矩阵V，并计算FCM的目标函数的F直到F不再变化，并按照最大隶属度原则确定所述多条历史负荷数据的归属时序特征；其中，在D i归类于X k的情况下，V i,k满足：
    V i,k＝max{V i,1,V i,2,···,V i,k,···,V i,L}
    其中，max{.}表示最大值函数。
  6. 根据权利要求1所述的方法，其中，
    新接入负荷i的时序特征为A i，A i表达为：
    A i＝[A i,1,A i,2,...,A i,k,...,A i,h]
    其中，h为用户数量；
    所述接入优化模型用于在所述接入台区有n条馈线可供接入的情况下，确定某一负荷接入的馈线，所述组合优化模型的决策变量可以用0-1变量表示：
    Figure PCTCN2019125811-appb-100018
    第k条馈线未接入前的负荷为l k，接入新负荷后，馈线负载L k更新为：
    Figure PCTCN2019125811-appb-100019
    所述组合优化模型为以馈线最大供电能力为目标，所述组合优化模型的目标函数为：
    Figure PCTCN2019125811-appb-100020
    其中，n表示可选馈线的数量；P k.N为第k条馈线有功传输容量上限，P k.max表示第k条馈线的最大负载；
    所述接入台区内的每条馈线满足如下约束：P k.max≤P k.N。
  7. 一种终端，包括处理器和存储器，所述存储器上存储有程序，所述程序被所述处理器执行时实现如权利要求1-6任一项所述的方法。
  8. 一种计算机可读存储介质，存储有计算机程序，所述计算机程序被处理器执行时实现如权利要求1-6任一项所述的方法。
PCT/CN2019/125811 2018-12-30 2019-12-17 配电台区源-荷协同接入方法、终端、存储介质 WO2020140743A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US16/759,456 US11500341B2 (en) 2018-12-30 2019-12-17 Source-load cooperation access method for a power distribution region, terminal and storage medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201811648669.6 2018-12-30
CN201811648669.6A CN110009123A (zh) 2018-12-30 2018-12-30 一种考虑时序互补特性的配电台区源-荷协同接入方法

Publications (1)

Publication Number Publication Date
WO2020140743A1 true WO2020140743A1 (zh) 2020-07-09

Family

ID=67165264

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/125811 WO2020140743A1 (zh) 2018-12-30 2019-12-17 配电台区源-荷协同接入方法、终端、存储介质

Country Status (3)

Country Link
US (1) US11500341B2 (zh)
CN (1) CN110009123A (zh)
WO (1) WO2020140743A1 (zh)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112084652A (zh) * 2020-09-07 2020-12-15 中国电力工程顾问集团西北电力设计院有限公司 一种风光储互补发电系统容量配置计算方法及系统
CN112365076A (zh) * 2020-11-19 2021-02-12 西安建筑科技大学 一种并联冷机负荷分配优化方法、存储介质及计算设备
CN113452018A (zh) * 2021-06-29 2021-09-28 湖南大学 一种电力系统备用不足风险场景辨识方法
CN113516195A (zh) * 2021-07-20 2021-10-19 南方电网科学研究院有限责任公司 一种配电网极限线损率计算方法、装置、设备和介质
CN114881439A (zh) * 2022-04-24 2022-08-09 中国华能集团清洁能源技术研究院有限公司 基于模糊综合评价计算储能变压器运行得分的方法及系统
CN115358640A (zh) * 2022-10-20 2022-11-18 四川中电启明星信息技术有限公司 一种基于阶梯式激励的居民台区需求响应方法

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110009123A (zh) * 2018-12-30 2019-07-12 国网天津市电力公司电力科学研究院 一种考虑时序互补特性的配电台区源-荷协同接入方法
CN111612248B (zh) * 2020-05-20 2023-09-08 云南电网有限责任公司电力科学研究院 一种配电网侧源-荷协调方法及系统
CN112686308A (zh) * 2020-12-28 2021-04-20 青岛鼎信通讯股份有限公司 一种基于kh与fcm的低压台区归属识别方法及系统
CN114202011A (zh) * 2021-10-30 2022-03-18 湖南江军科技有限责任公司 一种用电安全感知方法
CN114462725B (zh) * 2022-04-13 2022-09-02 国网浙江省电力有限公司营销服务中心 一种基于动态资源池的非直控型需求侧响应优化调度方法
CN115207921B (zh) * 2022-09-14 2022-12-30 国网江西省电力有限公司电力科学研究院 一种光伏接入配电网的位置优化方法及系统
CN116911806B (zh) * 2023-09-11 2023-11-28 湖北华中电力科技开发有限责任公司 基于互联网+的电力企业能源信息管理系统

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106230026A (zh) * 2016-08-30 2016-12-14 华北电力大学（保定） 基于时序特性分析的含分布式电源的配电网双层协调规划方法
CN107688879A (zh) * 2017-10-20 2018-02-13 云南电网有限责任公司 一种考虑源荷匹配度的主动配电网分布式电源规划方法
CN110009123A (zh) * 2018-12-30 2019-07-12 国网天津市电力公司电力科学研究院 一种考虑时序互补特性的配电台区源-荷协同接入方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107147152B (zh) * 2017-06-15 2019-12-10 广东工业大学 新能源配电网多类型有功无功源协同优化配置方法及系统
CN107800140B (zh) * 2017-10-18 2021-08-13 天津大学 一种考虑负荷特征的大用户供电接入决策方法
CN108288231B (zh) * 2018-01-19 2019-12-13 广东电网有限责任公司河源供电局 一种分布式光伏接入对配电台区负荷特性影响的评估方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106230026A (zh) * 2016-08-30 2016-12-14 华北电力大学（保定） 基于时序特性分析的含分布式电源的配电网双层协调规划方法
CN107688879A (zh) * 2017-10-20 2018-02-13 云南电网有限责任公司 一种考虑源荷匹配度的主动配电网分布式电源规划方法
CN110009123A (zh) * 2018-12-30 2019-07-12 国网天津市电力公司电力科学研究院 一种考虑时序互补特性的配电台区源-荷协同接入方法

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112084652A (zh) * 2020-09-07 2020-12-15 中国电力工程顾问集团西北电力设计院有限公司 一种风光储互补发电系统容量配置计算方法及系统
CN112084652B (zh) * 2020-09-07 2023-06-20 中国电力工程顾问集团西北电力设计院有限公司 一种风光储互补发电系统容量配置计算方法及系统
CN112365076A (zh) * 2020-11-19 2021-02-12 西安建筑科技大学 一种并联冷机负荷分配优化方法、存储介质及计算设备
CN112365076B (zh) * 2020-11-19 2023-06-20 西安建筑科技大学 一种并联冷机负荷分配优化方法、存储介质及计算设备
CN113452018A (zh) * 2021-06-29 2021-09-28 湖南大学 一种电力系统备用不足风险场景辨识方法
CN113516195A (zh) * 2021-07-20 2021-10-19 南方电网科学研究院有限责任公司 一种配电网极限线损率计算方法、装置、设备和介质
CN114881439A (zh) * 2022-04-24 2022-08-09 中国华能集团清洁能源技术研究院有限公司 基于模糊综合评价计算储能变压器运行得分的方法及系统
CN114881439B (zh) * 2022-04-24 2023-04-07 中国华能集团清洁能源技术研究院有限公司 基于模糊综合评价计算储能变压器运行得分的方法及系统
CN115358640A (zh) * 2022-10-20 2022-11-18 四川中电启明星信息技术有限公司 一种基于阶梯式激励的居民台区需求响应方法

Also Published As

Publication number Publication date
US11500341B2 (en) 2022-11-15
CN110009123A (zh) 2019-07-12
US20210208548A1 (en) 2021-07-08

Similar Documents

Publication Publication Date Title
WO2020140743A1 (zh) 配电台区源-荷协同接入方法、终端、存储介质
CN103151797B (zh) 基于多目标调度模型的并网运行方式下微网能量控制方法
CN107910863B (zh) 综合考虑光伏出力和负荷需求预测区间的配电网调度方法
Li et al. A novel bi-level robust game model to optimize a regionally integrated energy system with large-scale centralized renewable-energy sources in Western China
CN105591406B (zh) 一种基于非合作博弈的微电网能量管理系统的优化算法
CN105071389B (zh) 计及源网荷互动的交直流混合微电网优化运行方法及装置
Li et al. Hybrid time-scale energy optimal scheduling strategy for integrated energy system with bilateral interaction with supply and demand
Aghamolaei et al. Feasibility analysis of community-based PV systems for residential districts: A comparison of on-site centralized and distributed PV installations
CN111754037A (zh) 区域终端集成供能系统长期负荷混合预测方法
Huang et al. Optimal configuration planning of multi-energy microgird based on deep joint generation of source-load-temperature scenarios
Chupradit et al. Modeling and Optimizing the Charge of Electric Vehicles with Genetic‎ Algorithm in the Presence of Renewable Energy Sources
Buechler et al. Optimal energy supply scheduling for a single household: Integrating machine learning for power forecasting
Wu et al. Multi-objective optimization of integrated energy systems considering renewable energy uncertainty and electric vehicles
CN111082446B (zh) 一种考虑电池自消耗的储能优化配置方法
CN105514986A (zh) 一种基于虚拟电厂技术的der用户竞价并网方法
CN112670982A (zh) 一种基于奖励机制的微电网有功调度控制方法及系统
CN109066823B (zh) 一种适于三端口电力电子变压器的交直流混合微网两层优化方法
Wang et al. Research on short‐term and mid‐long term optimal dispatch of multi‐energy complementary power generation system
CN114240203A (zh) 一种包含热泵的配电网与分布式风电协同规划方法和装置
Fan et al. Energy flexibility of space-heating or cooling in Spain based on Developed Wildebeest Herd Optimization algorithm
CN109149644B (zh) 一种基于大数据分析的光储一体在线策略匹配和协同优化方法
Wang et al. Optimal Planning of Distributed Generation and Loads in Active Distribution Network: A Review
Lv et al. Research on energy storage allocation strategy considering smoothing the fluctuation of renewable energy
Zhu et al. DSM and optimization of multihop smart grid based on genetic algorithm
Tang et al. Benefit Evaluation Indicators and Methods for Regional Energy Internet

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19906879

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 19906879

Country of ref document: EP

Kind code of ref document: A1