CN114513004A - New energy station equivalence method based on improved k-means algorithm and application - Google Patents

New energy station equivalence method based on improved k-means algorithm and application Download PDF

Info

Publication number
CN114513004A
CN114513004A CN202210126799.3A CN202210126799A CN114513004A CN 114513004 A CN114513004 A CN 114513004A CN 202210126799 A CN202210126799 A CN 202210126799A CN 114513004 A CN114513004 A CN 114513004A
Authority
CN
China
Prior art keywords
new energy
voltage
station
equivalent
clustering
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
Application number
CN202210126799.3A
Other languages
Chinese (zh)
Inventor
贾科
孔繁哲
张旸
温志文
余磊
刘海涛
吴建云
牛健
栗磊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
North China Electric Power University
Electric Power Research Institute of State Grid Ningxia Electric Power Co Ltd
Original Assignee
North China Electric Power University
Electric Power Research Institute of State Grid Ningxia Electric Power Co Ltd
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 North China Electric Power University, Electric Power Research Institute of State Grid Ningxia Electric Power Co Ltd filed Critical North China Electric Power University
Priority to CN202210126799.3A priority Critical patent/CN114513004A/en
Publication of CN114513004A publication Critical patent/CN114513004A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • 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
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/906—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • H—ELECTRICITY
    • H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
    • H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
    • H02J3/00—Circuit arrangements for ac mains or ac distribution networks
    • H02J3/38—Arrangements for parallely feeding a single network by two or more generators, converters or transformers
    • H02J3/46—Controlling of the sharing of output between the generators, converters, or transformers
    • H02J3/466—Scheduling the operation of the generators, e.g. connecting or disconnecting generators to meet a given demand
    • 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
    • Y02E10/70—Wind energy
    • Y02E10/76—Power conversion electric or electronic aspects

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Databases & Information Systems (AREA)
  • Power Engineering (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Automation & Control Theory (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The new energy station equivalence method based on the improved k-means algorithm and the application thereof comprise the steps of deducing a new energy power supply multi-machine short circuit current analytical expression and carrying out phase conversion; establishing an analytic relation between the equivalent error and the station electric quantity; based on the characteristics of the voltage-controlled current source of the new energy power supply, equivalent voltage and impedance of the unit divided into the same type are calculated by adopting an optimization algorithm, the obtained equivalent error is used as the distance between each particle, the equivalent voltage is used as a clustering center, the algorithm is adopted to perform clustering equivalence on the new energy unit, the final clustering quantity is determined according to the contour coefficient, and the unit clustering equivalence is completed. According to the method, the equivalent error is used as the particle spacing of the clustering algorithm, the improved k-means algorithm is adopted for clustering equivalence, the model complexity of the new energy power supply is simplified, certain precision is guaranteed, the calculation amount and the simulation duration are reduced, the number of nodes is reduced in the process of calculating the current of the new energy power supply network, and the calculation speed and the iterative calculation efficiency are improved.

Description

New energy station equivalence method based on improved k-means algorithm and application
Technical Field
The invention relates to a method and application thereof, in particular to a new energy station equivalence method based on an improved k-means algorithm and application thereof.
Background
With the large-scale wind power and photovoltaic centralized grid connection, the difference between the fault characteristics of the new energy power supply and the synchronous power supply is large, the protection action performance is greatly influenced, and an accurate new energy station fault equivalence method needs to be researched urgently.
At present, a single-machine equivalence method or a multi-machine equivalence method is mainly adopted for the new energy station fault equivalence method, the model simplification degree of the single-machine equivalence method is large, the calculated amount is small, but when equivalence modeling is carried out on new energy power supplies in large running states, the accuracy of a single-machine equivalence model is poor, and the station fault characteristics cannot be accurately represented.
The existing multi-machine equivalence method of the new energy power supply mostly adopts a clustering algorithm, and a large number of relevant factors need to be considered in the existing research on the equivalence of the new energy power supply: the method is characterized in that the influence of the factors such as low voltage ride through characteristics, output, collection lines and fault occurrence conditions has certain ambiguity, and the clustering problem related to dynamic values and the like also has certain ambiguity, so that the key point for establishing a wind power plant dynamic voltage equivalent model is how to describe the factors and the interaction of the factors to form a reasonable clustering result.
In the traditional clustering algorithm, the vector distance of the particle coordinates is used as the distance between the particles and the clustering center, an Euclidean distance equidistance formula is usually adopted for description, and the average value of the particle coordinates which are classified into the same type is used as the clustering center. The method takes the equivalence problem as an optimization problem, adopts a particle swarm optimization algorithm to optimize the equivalence result to obtain equivalent voltage, takes the obtained equivalent error as the particle distance, and takes the particle distance as the small equivalent error and the optimal objective function as the clustering center to carry the equivalent voltage into the algorithm for iteration.
Disclosure of Invention
In order to solve the defects in the prior art, the technical scheme of the new energy station equivalence method based on the improved k-means algorithm comprises the following steps:
step 1, deducing a multi-machine short-circuit current analytical expression of the new energy power supply according to national standard requirements and terminal voltage of each station, carrying out phase conversion, and establishing an analytical relation between equivalent errors and station electric quantity;
and 2, based on the characteristics of the voltage-controlled current source of the new energy power supply, calculating equivalent voltage and impedance of the units divided into the same type by adopting an optimization algorithm, taking the obtained equivalent error as the distance between each particle, taking the equivalent voltage as a clustering center, performing clustering equivalence on the new energy units by adopting an improved k-means algorithm, and determining the final clustering quantity according to the contour coefficient to finish the unit clustering equivalence.
The invention also discloses a new energy station equivalence method based on the improved k-means algorithm, which is applied to the new energy station fault analysis system.
Advantageous effects
According to the method, the equivalent error is used as the particle spacing of the clustering algorithm, the improved k-means algorithm is adopted for clustering equivalence, the model complexity of the new energy power supply is simplified, certain precision is guaranteed, the calculated amount and the simulation duration are reduced, the number of nodes is reduced in the process of calculating the network current of the new energy power supply, and the calculating speed and the iterative calculating efficiency are improved.
Drawings
FIG. 1 is a detailed model topology diagram of a research site of the present invention;
FIG. 2 is a topological diagram of an equivalent model of a research station according to the present invention;
FIG. 3 is a flow chart of the new energy station equivalence method based on the improved k-means algorithm.
Detailed Description
The invention discloses a new energy field station equivalence method based on an improved k-means algorithm, which is described in detail below with reference to the accompanying drawings.
Step 1, deducing a multi-machine short-circuit current analytical expression of the new energy power supply according to national standard requirements and terminal voltage of each station, carrying out phase conversion, and establishing an analytical relation between equivalent errors and station electric quantity:
at present, the grid-connected new energy power supply in China is mainly divided into a full-power inverter power supply represented by a double-fed fan and a partial-power inverter power supply represented by a photovoltaic direct-drive fan and a permanent-magnet direct-drive fan.
According to technical provisions that photovoltaic power stations are connected to an electric power system, and a d-axis voltage orientation strategy is generally adopted by an inverter, the short-circuit current output by the photovoltaic station after the fault is obtained
Figure BDA0003500289840000031
Figure BDA0003500289840000032
Wherein id and iq are dq axis current output by the full-power inverter type energy power supply, Us station terminal voltage, IN is station rated current, Imax is the maximum value of current allowed to be output by the inverter, K1 and K2 are reactive current support coefficients of low-pass control stages 1 and 2 respectively, UN is the rated value of station terminal voltage, and P0 is active power of the new energy power supply;
different from a photovoltaic voltage ride-through strategy, according to technical provisions for connecting wind power plant to power system, when the voltage of wind power drops below 0.2p.u., the fan is disconnected, and the short-circuit current output by the wind turbine generator set after the fault is obtained
Figure BDA0003500289840000041
Figure BDA0003500289840000042
The id and the iq are dq-axis current output by the permanent magnet wind field, the per unit value of the station terminal voltage of the Us, UN is a rated value of the station terminal voltage, IN is station rated current, Imax is the maximum value of current allowed to be output by the inverter, K1 is a reactive current support coefficient of low-voltage penetration control, and P0 is active power of the new energy power supply.
For the double-fed fan, unlike the permanent-magnet direct-drive fan, the short-circuit current provided in the fault process of the double-fed fan mainly consists of the stator short-circuit current and the short-circuit current of the grid-side converter, and the capacity of the grid-side converter is small, so that the stator current output by the double-fed fan is approximately equal to the short-circuit current provided by the double-fed fan. The double-fed fan adopts different control modes according to different voltage drop degrees, the fault characteristics of the double-fed fan are different, and the double-fed fan is divided into three types according to the control modes: (1) when the voltage drop delta U is less than 0.2p.u., the doubly-fed fan does not enter a low penetration interval at the moment, RSC keeps outer loop control, and the fault characteristic of the doubly-fed fan is consistent with the fault characteristic of the full-power inverter power supply at the moment, so that repeated description is omitted; (2) when the voltage drops by 0.2p.u. < delta U < alpha p.u., the low-penetration control is started, and the RSC single-ring control is carried out according to the national standard when the outer ring is disconnected; (3) when the voltage drop degree alpha p.u. < delta U <0.8p.u., the rotor current is over-current, and the rotor converter is cut off and crowbar resistor is used for controlling in order to prevent the rotor side converter from being damaged due to over-current.
Crowbar resistance on-phase, 0.2p.u. < U < α p.u.
Under a two-phase rotating coordinate system, the stator and rotor voltage flux linkage equation of the doubly-fed wind turbine is as follows:
stator and rotor voltage equations:
Figure BDA0003500289840000043
stator and rotor flux linkage equation:
Figure BDA0003500289840000051
in the formula: u. ofsd、usq、urd、urqDq-axis voltages, i, of stator and rotor, respectivelysd、isq、ird、 irqDq-axis currents, psi, of stator and rotor, respectivelysd、ψsq、ψrd、ψrqDq-axis flux linkage, L, of stator and rotor, respectivelym、Ls、LrRespectively a stator and rotor coaxial equivalent winding mutual inductance, a stator self-inductance, a rotor self-inductance and R in dq coordinatess、RrRespectively, the stator and rotor resistances, ω is synchronous speed, and s is ω - ω1， ω1Is the rotor speed.
When the terminal voltage falls, the stator flux linkage can be expressed as:
Figure BDA0003500289840000052
after a fault, a crowbar is instantly put into use, and the crowbar resistor bypasses the rotor converter and the rotor voltage urdqWhen the rotor resistance is equal to 0, the rotor resistance is changed into the sum R of the original rotor resistance and the crowbar resistancere＝Rr+Rc,RcIs a crowbar resistor.
From the stator-rotor flux linkage equation, a relational expression between the stator-rotor current and the flux linkage can be obtained:
Figure BDA0003500289840000053
simultaneous equations (11) - (14) can obtain that the steady-state short-circuit current output by the doubly-fed wind turbine after the crowbar resistor is put into use is as follows:
Figure BDA0003500289840000054
in the formula: σ ═ 1-Lm 2/(LrLs)，LmThe stator and the rotor are mutually inducted.
Rsc control phase α p.u. < U <0.8p.u.
According to technical provisions for accessing a wind power plant to a power system, the reactive current output by the double-fed fan in the RSC control stage can be obtained as follows:
isq＝K(0.9-Us)IN α＜Us＜0.8 (16)
wherein: i.e. isqFor stator q-axis current, UsIs the stator voltage per unit value, INIs the rated current.
The doubly-fed wind turbine is considered to adopt stator voltage directional control, and the rotor q-axis current of the doubly-fed wind turbine can be obtained according to a stator flux linkage equation of the wind turbine:
Figure BDA0003500289840000061
in order to prevent the rotor current from overcurrent damage and the rotor side converter in the fault process, the rotor current needs to be limited, and the active power limit value of the fan is considered, so that the d-axis current of the rotor can be obtained as follows:
Figure BDA0003500289840000062
wherein IrmaxLimiting the amplitude, i, of the rotor converterrdmaxIn order to consider the rotor current of the active power limit value of the fan, the expression is as follows:
Figure BDA0003500289840000063
wherein P is0The maximum active power which can be output by the fan.
In summary, in the RSC control stage of the voltage α p.u. < U <0.8p.u., the short-circuit current output by the dual-feed fan is:
Figure BDA0003500289840000064
the short-circuit current of the double-fed fan in different control stages is as follows:
Figure BDA0003500289840000065
Figure BDA0003500289840000066
because the new energy power supply adopts voltage orientation control, the dq-axis current output by each station takes the output voltage of each phase-locked loop as a reference system, so the dq-axis short-circuit current output by each station is reduced to the phase of the system synchronous power supply:
Figure BDA0003500289840000067
wherein idΣ、iqΣCalculating the dq axis current of each station to the phase of the reference system; i.e. idi、iqiDq-axis currents, θ, respectively, output by stations iiIs the difference between the terminal voltage of the station i and the reference phase.
For the equivalent model there are:
Figure BDA0003500289840000071
wherein idΣeq、iqΣeqRespectively calculating the dq axis current of the equivalent station; ideq、iqeqRespectively, the dq-axis current, theta, before the return of the equivalent stationeqAnd obtaining the difference value of the voltage at the terminal i of the equivalent unit and the phase of the reference system.
According to the analysis, the new energy power supply presents the characteristics of a voltage-controlled current source after the fault, the steady-state fault current of the new energy power supply is determined by different control modes, voltage drop degrees, the phase of the power generated during the fault and the phase of the phase-locked loop output, namely the phase of the fan outlet, so that the relevant electric quantity influencing the fault steady-state characteristic of the new energy power supply is selected as a grouping index, and the station topologies before and after the equivalence are shown in fig. 1 and 2.
Step 2, calculating equivalent voltage and impedance by adopting an optimization algorithm on the basis of the characteristics of the voltage-controlled current source of the new energy power supply; and taking the obtained equivalent error as the distance between each particle, taking the equivalent voltage as a clustering center, performing clustering equivalence on the new energy unit by adopting an improved k-means algorithm, and determining the final clustering quantity according to the contour coefficient to finish unit clustering equivalence.
In the traditional clustering algorithm, the vector distance of the particle coordinates is used as the distance between the particles and the clustering center, an Euclidean distance equidistance formula is usually adopted for description, and the average value of the particle coordinates which are classified into the same type is used as the clustering center. The method takes the equivalence problem as an optimization problem, adopts a particle swarm optimization algorithm to optimize the equivalence result to obtain equivalent voltage, takes the obtained equivalent error as the particle distance, and takes the particle distance as the small equivalent error and the optimal objective function as the clustering center to carry the equivalent voltage into the algorithm for iteration.
1. Distance establishment of particles from cluster centers
The clustering algorithm firstly initializes the clustering center, the initial center is randomly assigned to a certain particle, the equivalent error is taken as the particle distance, and the particles are distributed to the center with the minimum equivalent error as a cluster. And converting the clustering problem into an optimization problem by taking an error generated by equivalence as an objective function.
The equivalent resulting error is given by:
Q＝(idΣ-idΣeq)2+(iqΣ-iqΣeq)2 (25)
calculating equivalent errors by adopting an optimization algorithm, firstly randomly generating particles, taking equivalent voltage as the positions of the particles, and for a certain determined equivalent voltage, calculating the equivalent errors according to the topology of an equivalent model:
Figure BDA0003500289840000081
in the formula: delta U and delta U are the longitudinal component and the transverse component of the voltage drop, UeqFor randomly generated equivalent voltage, U, of the optimization algorithmfTo the fault point voltage, Peq、QeqThe active power and the reactive power output by the equivalent unit are respectively, r and x are respectively the resistance and the reactance of the line in unit length, and l is the length of the equivalent line.
For a determined voltage drop, the equivalent model is solvable, a determined line impedance length and a voltage phase are solved according to the equation (26), and the reduced dq axis current output by the equivalent station can be calculated according to the equation (24). According to the error calculation in (25), the objective function Q is optimized to obtain leq、UeqAnd the error E is taken as the distance d between the particles of the clustering algorithm and the clustering centerijAnd finishing the establishment of the algorithm space distance.
2. Determination of cluster centers in an iterative process
The clustering algorithm first initializes the cluster center and assigns particles to the closest cluster according to the inter-particle distance. The invention randomly generates equivalent voltage for optimizing the fans classified into the same type according to the grouping result obtained in the last step, the target function is kept unchanged, and the equivalent voltage is randomly generated for optimizing again,at this time, an equivalent voltage U is obtainedeqAnd using the equivalent voltage as a clustering center, recalculating the distance between the particles and the clustering center, redistributing the particles based on the principle of closest distance, and performing the next iteration until the algorithm converges and outputs a clustering result, wherein the specific flow is shown in fig. 3.
3. Determining the number of clusters from the contour coefficients
The profile coefficient formula is as follows:
Figure BDA0003500289840000082
wherein a (i) represents the sample point cohesion degree, and the calculation formula is as follows
Figure BDA0003500289840000083
Where j represents other sample points within the same class as sample i and distance represents the distance between samples ij. Smaller a (i) indicates a tighter sample.
b (i) is calculated in a similar manner to a (i). Except that it is necessary to traverse other clusters to get multiple values { b (1), b (2), … …, b (n) } from which the smallest value is selected as the final result. The new energy power supply is divided into more types, errors and evaluation indexes are not obviously improved, but the required simulation time length is obviously increased, so that the grouping number can be set within a reasonable range, and the equivalent flow is shown in figure 3.
Example 1
Taking actual topologies and parameters of a plurality of wind farms in a certain region of Jilin as an example, a wind farm detailed model (including a permanent magnet wind farm of 10 × 100MW and a double-fed wind farm) shown in FIG. 1 is built in PSCAD. The wind turbine generator set is connected to a grid-connected point through a box transformer substation (0.69kV/35kV), and is connected with an external power grid through an overhead line through a main transformer substation (35kV/220 kV). The short-circuit impedance of the box transformer is 6.39%, and the short-circuit impedance of the main transformer is 13.54%. The distance between stations is 5 km.
1. When the voltage of the grid-connected point is about 90%, because the permanent magnet and double-fed steady state characteristics of the grid-connected point do not enter the low-penetration interval are consistent, only the simulation verification is carried out on the permanent magnet wind field, the current after the normalization of each wind field can be obtained according to the step one and the voltage information of each station, the equivalent voltage is generated by adopting an optimization algorithm according to the equivalent method in the step two, the optimal solution is found, the equivalent error among samples is obtained, and the primary classification is carried out; and (4) obtaining the equivalent voltage by the stations classified into the same type through an optimization algorithm, calculating the distance from the equivalent voltage to each sample again, finishing classification again, and continuously iterating until the algorithm converges.
TABLE 1 wind field clustering results
Figure BDA0003500289840000091
The current errors generated before and after the equivalence of the equivalent model provided by the invention in the wind power plant are shown in table 2, the fault current output by the equivalent model and the detailed model can be well fitted, the current error of the equivalent model is 1.1%, the precision is high, and the model can be simplified on the premise of ensuring the model precision.
Comparison of equivalence method and single-machine equivalence error provided in Table 2
Figure BDA0003500289840000101
2 the voltage of the grid-connected point drops by 60 percent
When the voltage of the grid-connected point falls to 60%, the grid-connected point is divided into a permanent magnet type and a double-fed type according to the type of the wind field, and then the grid-connected point is classified according to the voltage information of each station.
Table 3 wind field clustering results
Figure BDA0003500289840000102
The current errors generated before and after the equivalence of the equivalent model provided by the invention is adopted in the wind power plant are shown in table 4, the fault current output by the equivalent model and the detailed model can be well fitted, the current error of the equivalent model is 1.34%, the precision is high, and the model can be simplified on the premise of ensuring the model precision.
Comparison of the equivalence method presented in Table 4 with the single-machine equivalence error
Figure BDA0003500289840000103
The voltage of the grid-connected point drops by 30 percent
When the voltage of a grid-connected point falls to 30%, the RSC of the double-fed machine set is cut off and crowbar is put into use, the permanent magnetic wind field enters low penetration control, classification is carried out according to machine types, and grouping is carried out according to terminal voltage information.
TABLE 5 wind field clustering results
Figure BDA0003500289840000111
The current errors generated before and after the equivalence of the equivalent model provided by the invention is adopted in the wind power plant are shown in table 6, the fault current output by the equivalent model and the detailed model can be well fitted, the current error of the equivalent model is 2.7%, the precision is high, and the model can be simplified on the premise of ensuring the model precision.
Comparison of the equivalence method presented in Table 6 with the single-machine equivalence error
Figure BDA0003500289840000112
Compared with a detailed model by adopting the equivalent method in combination with tables 2, 4 and 6, the fault current error is less than 3%. The finally obtained new energy power source equivalence method suitable for engineering practicability is high in simplification degree, ensures certain precision, reduces the calculation amount and simulation time, reduces the number of nodes in the process of calculating the network current containing the new energy power source, and improves the calculation speed and the iterative calculation efficiency.
Theoretically, grouping of multiple units can be realized by adopting a clustering algorithm through the fault characteristics and the electric quantity characteristics of the new energy power supply. The traditional clustering algorithm takes the vector distance of the particle coordinates as the distance between the particles and the clustering center, but because the clustering problem is actually a nonlinear problem, the objective function is not uniform and monotonous, and the partial derivatives at each coordinate point are different, the particles and the clustering center are close to each other in a space coordinate system and are not equivalent to the objective function after clustering, so that the objective function is more optimized. The method takes an equivalence problem as an optimization problem, adopts a particle swarm optimization algorithm to optimize an equivalence result to obtain equivalent voltage, takes the obtained equivalent error as a particle distance, and takes the particle distance as a small equivalent error and a more optimal target function when the particle distance is close, and takes the equivalent voltage as a clustering center to be introduced into the algorithm for iteration. The method takes the equivalent error as the inter-particle distance of the clustering algorithm, adopts the improved k-means algorithm to perform clustering equivalence, simplifies the model complexity of the new energy power supply, ensures certain precision, reduces the calculated amount and the simulation time, reduces the number of nodes in the process of calculating the network current containing the new energy power supply, and improves the calculation speed and the iterative calculation efficiency.
The foregoing shows and describes the general principles, essential features, and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are merely illustrative of the principles of the invention, but that various changes and modifications may be made without departing from the spirit and scope of the invention, which fall within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.

Claims (9)

1. A new energy field station equivalence method based on an improved k-means algorithm is characterized by comprising the following steps: the method comprises the following steps:
step 1: deducing a multi-machine short circuit current analytical expression of the new energy power supply according to national standard requirements and terminal voltage of each station, carrying out phase conversion, and establishing an analytical relation between equivalent errors and station electric quantity;
step 2: based on the characteristics of the voltage-controlled current source of the new energy power supply, equivalent voltage and impedance of the unit divided into the same type are calculated by adopting an optimization algorithm, the obtained equivalent error is used as the distance between each particle, the equivalent voltage is used as a clustering center, the new energy unit is subjected to clustering equivalence by adopting an improved k-means algorithm, the final clustering quantity is determined according to the contour coefficient, and the unit clustering equivalence is completed.
2. The new energy site equivalence method for improving k-means algorithm according to claim 1, characterized by comprising the following steps: the step 1 further comprises the following steps: according to technical provisions for connecting photovoltaic power stations to an electric power system, and considering that inverters generally adopt a d-axis voltage orientation strategy, short-circuit current output by the photovoltaic stations after faults is obtained as follows:
Figure RE-551296DEST_PATH_IMAGE001
the method comprises the steps that id and iq are dq axis currents output by a photovoltaic power station respectively, a per unit value of a Us station terminal voltage, UN is a rated value of the station terminal voltage, IN is a station rated current, Imax is a maximum value of current allowed to be output by an inverter, K1 and K2 are reactive current support coefficients of low-voltage-through control stages 1 and 2 respectively, and P0 is active power of a new energy power supply.
3. The new energy site equivalence method for improving k-means algorithm according to claim 2, characterized by comprising the following steps: according to technical provisions for accessing wind power plants to a power system, when the voltage of wind power drops to be below 0.2p.u., a fan is disconnected from a network, short-circuit current output by a wind turbine generator after a fault can be obtained:
Figure RE-736946DEST_PATH_IMAGE002
the id and the iq are dq-axis current output by the permanent magnet wind field, the per unit value of the station terminal voltage of the Us, UN is a rated value of the station terminal voltage, IN is station rated current, Imax is the maximum value of current allowed to be output by the inverter, K1 is a reactive current support coefficient of low-voltage penetration control, and P0 is active power of the new energy power supply.
4. The new energy site equivalence method for improving k-means algorithm according to claim 3, characterized by comprising the following steps: the doubly-fed fan adopts different control modes according to the different degrees of voltage drop, and its fault characteristics are also different, divide into three types according to control mode: (1) when the voltage drop delta U is less than 0.2p.u., the doubly-fed fan does not enter a low penetration interval at the moment, the RSC keeps outer loop control, and the fault characteristic of the doubly-fed fan is consistent with the fault characteristic of the full-power inverter type power supply at the moment; (2) when the voltage drops by 0.2p.u. < delta U < alpha p.u., the low-penetration control is started, and the RSC single-ring control is carried out according to the national standard when the outer ring is disconnected; (3) when the voltage drop degree alpha p.u. < delta U <0.8p.u., the rotor current is overcurrent, and in order to prevent the rotor side converter from being damaged due to overcurrent, the rotor converter is cut off and crowbar resistance is put into control; α is the voltage boundary between crowbar resistance control and low-pass control.
5. The new energy site equivalence method for improving k-means algorithm according to claim 4, characterized by: the crowbar resistance is controlled by the following steps:
a. in a crowbar resistor input stage, when the voltage drops by 0.2p.u. < delta U < alpha p.u., the steady-state short-circuit current output by the double-fed fan after the crowbar resistor is input is as follows:
Figure RE-844579DEST_PATH_IMAGE003
in the formula: σ =1-Lm 2/(LrLs)，LmThe mutual inductance Ls and Lr of the stator and the rotor are respectively the self inductance of the stator and the rotor, Rre = Rcb + Rr, Rcb is a crowbar resistance value, Rr is a rotor resistance, s is wn-ws, wn is a synchronous rotating speed, ws is a stator rotating speed, and U is a terminal voltage;
rsc control phase α p.u. < U <0.8p.u.
The short-circuit current output by the double-fed fan is as follows:
Figure RE-770947DEST_PATH_IMAGE004
the system comprises a double-fed fan, a rotor converter, a rotor q-axis current, a reactive current support coefficient and a new energy power supply, wherein isd and isq are respectively a stator dq-axis current output by the double-fed fan, a Us station terminal voltage, IN is a station rated current, Irmax is a maximum value of a current allowed to be output by the rotor converter, irq is a rotor q-axis current, K is a reactive current support coefficient of low-penetration control, and P0 is active power of the new energy power supply;
the short-circuit current of the doubly-fed wind turbine at different control stages can be obtained as follows:
Figure RE-DEST_PATH_IMAGE005
。
6. the new energy site equivalence method for improving k-means algorithm according to claim 4, characterized by comprising the following steps: because the new energy power supply adopts voltage orientation control, the dq-axis current output by each station takes the output voltage of each phase-locked loop as a reference system, so the dq-axis short-circuit current output by each station is reduced to the phase of the system synchronous power supply:
Figure RE-382057DEST_PATH_IMAGE006
wherein idΣ、iqΣCalculating the dq axis current of each station to the phase of the reference system; i.e. idi、iqiD q-axis currents, θ, respectively output by the stations iiThe difference value of the terminal voltage of the station i and the reference phase is obtained;
for an equivalent model there are:
Figure RE-735678DEST_PATH_IMAGE007
wherein idΣeq、iqΣeqRespectively the dq-axis current after the normalization of the equivalent station；ideq、iqeqRespectively, the dq-axis current, theta, before the return of the equivalent stationeqAnd the difference value of the voltage at the i-machine end of the equivalent unit and the phase of the reference system is obtained.
7. The new energy site equivalence method for improving k-means algorithm according to claim 1, characterized by comprising the following steps: the step 2 further comprises the following steps:
(1) establishing the distance between the particles and the clustering center;
(2) determining a clustering center in an iterative process;
(3) and determining the grouping number according to the contour coefficient.
8. The new energy site equivalence method for improving k-means algorithm according to claim 7, characterized by comprising the following steps:
the distance establishment of the particles from the cluster center further comprises the following steps: taking the equivalent error as the particle distance, distributing the particles to the center with the minimum equivalent error as a cluster, and taking the error generated by the equivalent as a target function to convert the clustering problem into an optimization problem;
the determining the number of clusters according to the contour coefficients further comprises the following steps: for fans classified into the same type, the target function is kept unchanged, equivalent voltage is randomly generated and optimized again, and equivalent voltage U is obtained at the momenteqThe equivalent voltage at the moment is used as a clustering center, the distance between the particles and the clustering center at the moment is recalculated, the particles are redistributed based on the principle of closest distance, and the next iteration is carried out until the algorithm converges and outputs a clustering result;
the profile coefficient formula is as follows:
Figure RE-533869DEST_PATH_IMAGE008
wherein a (i) represents the sample point cohesion, and the calculation formula is as follows:
Figure RE-263928DEST_PATH_IMAGE009
where j represents other sample points within the same class as sample i and distance represents the distance between samples ij. Smaller a (i) indicates a tighter sample.
9. The new energy station equivalence method based on the improved k-means algorithm according to any one of claims 1-8 is applied to a new energy station fault analysis system.
CN202210126799.3A 2022-02-10 2022-02-10 New energy station equivalence method based on improved k-means algorithm and application Pending CN114513004A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210126799.3A CN114513004A (en) 2022-02-10 2022-02-10 New energy station equivalence method based on improved k-means algorithm and application

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210126799.3A CN114513004A (en) 2022-02-10 2022-02-10 New energy station equivalence method based on improved k-means algorithm and application

Publications (1)

Publication Number Publication Date
CN114513004A true CN114513004A (en) 2022-05-17

Family

ID=81551353

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210126799.3A Pending CN114513004A (en) 2022-02-10 2022-02-10 New energy station equivalence method based on improved k-means algorithm and application

Country Status (1)

Country Link
CN (1) CN114513004A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114865703A (en) * 2022-06-08 2022-08-05 合肥工业大学 Method for identifying high-penetration characteristic parameters of direct-drive fan inverter
CN116432541A (en) * 2023-06-08 2023-07-14 国网江西省电力有限公司电力科学研究院 New energy station modeling method and system based on optimization clustering algorithm

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114865703A (en) * 2022-06-08 2022-08-05 合肥工业大学 Method for identifying high-penetration characteristic parameters of direct-drive fan inverter
CN114865703B (en) * 2022-06-08 2024-03-08 合肥工业大学 High-pass characteristic parameter identification method for direct-drive fan inverter
CN116432541A (en) * 2023-06-08 2023-07-14 国网江西省电力有限公司电力科学研究院 New energy station modeling method and system based on optimization clustering algorithm
CN116432541B (en) * 2023-06-08 2023-10-20 国网江西省电力有限公司电力科学研究院 New energy station modeling method and system based on optimization clustering algorithm

Similar Documents

Publication Publication Date Title
Liu et al. A dynamic equivalent model for DFIG-based wind farms
CN114513004A (en) New energy station equivalence method based on improved k-means algorithm and application
CN106446458B (en) Consider the weak looped distribution network tidal current computing method of distributed power source
CN109412203B (en) Grid-connected capacity limit calculation method for wind power multi-feed-in power system
CN106451539B (en) It is a kind of meter and permanent magnet direct-drive wind turbine group dynamic characteristic wind farm grid-connected Method of Stability Analysis
CN106168994B (en) A kind of subsynchronous frequency equivalent modeling method of wind power plant
Xue et al. Equivalent modeling of DFIG based wind farm using equivalent maximum power curve
CN111027179B (en) Equivalent modeling method for double-fed wind power plant considering auxiliary frequency modulation service
CN110165644B (en) New energy station pilot protection method based on transient current time-frequency characteristics
CN114491886A (en) General modeling method and device for active power distribution network containing multi-type distributed new energy
CN108460228B (en) Wind power plant equivalence method based on multi-objective optimization algorithm
CN108054753B (en) Direct-drive wind power plant cluster division method considering low-voltage ride through characteristics
CN108258725B (en) Doubly-fed wind turbine dynamic equivalence method based on equivalent power angle coherence
Shao et al. Aggregation modeling algorithm of PMSM wind farm considering submarine cable power loss
CN110571794B (en) Transient model equivalent calculation method suitable for doubly-fed wind power plant
Wang et al. Evaluation method of wind turbine group classification based on calinski harabasz
Ju et al. Three-phase DFIG steady model and fast three-phase load flow algorithm for distribution power systems
Qi et al. Dynamic equivalent modeling of direct-drive PMSG wind farms based on the transient active power response characteristics
Li et al. Structure preserving aggregation method for doubly-fed induction generators in wind power conversion
CN110212570B (en) Wind power plant equivalent model based on MMSE mining and construction method and application thereof
Ye et al. A New Method of Wind Farm Clustering Based on Improved Fuzzy C-Means Clustering and MPSO
CN112260225A (en) Double-fed wind turbine generator three-phase short-circuit current calculation method and system based on combined model
Jiang et al. Rapid Generation of DFIG Farm Equivalent Model Evaluation Scenario Based on Deep Belief Network
Sun et al. A practical clustering method of DFIG wind farms based on dynamic current error
CN113139259B (en) Wind power plant clustering modeling method for dynamic equivalence of power grid

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