CN113162028A - New energy station AGC comprehensive performance evaluation method based on RankBoost - Google Patents

New energy station AGC comprehensive performance evaluation method based on RankBoost Download PDF

Info

Publication number
CN113162028A
CN113162028A CN202110334224.6A CN202110334224A CN113162028A CN 113162028 A CN113162028 A CN 113162028A CN 202110334224 A CN202110334224 A CN 202110334224A CN 113162028 A CN113162028 A CN 113162028A
Authority
CN
China
Prior art keywords
new energy
index
energy station
agc
namely
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.)
Granted
Application number
CN202110334224.6A
Other languages
Chinese (zh)
Other versions
CN113162028B (en
Inventor
任洲洋
杨志学
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.)
Chongqing University
Original Assignee
Chongqing University
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 Chongqing University filed Critical Chongqing University
Priority to CN202110334224.6A priority Critical patent/CN113162028B/en
Publication of CN113162028A publication Critical patent/CN113162028A/en
Application granted granted Critical
Publication of CN113162028B publication Critical patent/CN113162028B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • 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
    • H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
    • H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]

Landscapes

  • Engineering & Computer Science (AREA)
  • Power Engineering (AREA)
  • Feedback Control In General (AREA)
  • Supply And Distribution Of Alternating Current (AREA)

Abstract

The invention discloses a new energy station AGC comprehensive performance evaluation method based on RankBoost, which comprises the following steps: 1) acquiring historical output data of the new energy station, and establishing an effective data set V; 2) calculating a frequency modulation performance index of the new energy station in each control period based on the effective data set V; 3) calculating NdThe method comprises the steps that an index set I of each new energy station in M control periods is processed to obtain a new energy AGC effective index set S; 4) and evaluating the comprehensive performance of the new energy AGC new energy station by utilizing RankBoost based on the new energy AGC effective index set S. The method can be widely applied to the evaluation of the new energy AGC new energy station with different installed capacities and different regions and different frequency modulation performances, and can provide beneficial reference for the operation control of the secondary frequency modulation of the power system and the income distribution of the auxiliary frequency modulation market by the new energy.

Description

New energy station AGC comprehensive performance evaluation method based on RankBoost
Technical Field
The invention relates to the field of new energy AGC scheduling, in particular to a new energy station AGC comprehensive performance evaluation method based on RankBoost.
Background
In recent years, the installed capacity of new energy plants represented by photovoltaic power and wind power has been increasing year by year. Due to the randomness and the low controllability of the new energy, the large-scale centralized operation of the new energy will affect many aspects such as peak load regulation and frequency modulation of a power grid, tie line control, system transient stability and the like, and new challenges are brought to the safe and stable operation of a power system.
Therefore, many new energy stations are beginning to gradually assemble AGC control systems, which have active regulation capability and can participate in grid frequency modulation. Due to the fact that the physical performance difference of the new energy stations is large, and the operation management levels of all power plants are different, all AGC new energy stations can hardly keep good adjusting performance after being put into operation. Therefore, a set of comprehensive performance evaluation methods for the AGC new energy station is needed to formulate a corresponding examination and reward punishment method to evaluate the performance of the AGC new energy station executing the AGC instruction, so as to optimize the performance of the AGC new energy station and realize the safe and stable control of the power grid.
Many scholars have developed a series of researches on AGC optimization control at present, but unfortunately, the series of researches developed at present mainly focus on frequency modulation performance evaluation of conventional thermal power new energy stations, and the comprehensive performance evaluation of the new energy stations is discussed only rarely. And the new energy participating in the grid frequency modulation has become an unblocked trend. Therefore, the research on the comprehensive performance evaluation strategy of the new energy station is extremely necessary and has practical engineering value. However, no method for evaluating the comprehensive performance of the new energy station based on RankBoost is considered in research.
Disclosure of Invention
The invention aims to provide a new energy station AGC comprehensive performance evaluation method based on RankBoost, which comprises the following steps:
1) acquiring historical output data of the new energy station, and establishing an effective data set V;
2) calculating a frequency modulation performance index of the new energy station in each control period based on the effective data set V;
the frequency modulation performance indexes comprise a commissioning performance index, a tracking performance index, an adjusting speed index, a generating capacity index, a plan completion rate index and a utilization hour index.
The commissioning performance index comprises commissioning rate K1Namely:
Figure BDA0002996671240000011
in the formula, ToAnd TgThe AGC operation time and the grid-connected operation of the new energy station are respectivelyTime;
the tracking performance index comprises an excess power generation amount K for representing the response capability of the new energy station to the control command2And a tracking deviation K for representing the tracking capability of the new energy station on the control command3Namely:
Figure BDA0002996671240000021
Figure BDA0002996671240000022
in the formula, N represents the number of sampling points in a control period; i is an arbitrary sampling point; delta PtA value indicating that the AGC actual output is higher than the control command during the adjustment period; piA sampling value representing the power of the new energy station; p1A target output representing a control command;
the regulation speed index comprises an upper speed regulation degree K of the new energy station4And the down-regulation speed K5Namely:
Figure BDA0002996671240000023
in the formula, P0And P1The target output, t, of the initial value of regulation and the control command of the new energy station0And t1The new energy station is used for controlling the output of the new energy station;
the power generation capacity index comprises actual accumulated power generation capacity K used for representing power generation capacity of the new energy station6And theoretical generated energy index K7Namely:
Figure BDA0002996671240000024
wherein N represents the number of sampling points in the control period, PiIs a sampling value of the power of the new energy station,
Figure BDA0002996671240000025
predicting a force value for a theory;
the plan completion rate index comprises a plan completion rate K of the new energy station8Namely:
Figure BDA0002996671240000026
in the formula, PGIs the actual generated energy of the new energy station, PplanThe planned generating capacity of the new energy station is obtained;
the utilization hour number index comprises the utilization hour number K of the new energy station9Namely:
Figure BDA0002996671240000031
in the formula, PGIs the actual generated energy, P, of the AGC new energy stationICThe installed capacity of the new energy station.
3) Calculating NdThe method comprises the steps that an index set I of each new energy station in M control periods is processed to obtain a new energy AGC effective index set S;
calculating NdThe method for setting the index set I of the new energy station in the M control periods comprises the following steps:
3.1) let the new energy station serial number d equal to 1, IdSetting the maximum iteration number N as 0d；
3.2) calculating the AGC frequency modulation performance index of the d-th new energy station in each control period, thereby obtaining an index set I under M control periodsd＝{Id,1,Id,2,…,Id,MF, wherein, index set I under each control periodd,i＝{Kd,i,1,Kd,i,2,…,Kd,i,9}；
3.3) calculating the average index set of the d-th new energy station
Figure BDA0002996671240000032
Namely:
Figure BDA0002996671240000033
3.4) determination of d>NdIf not, making d equal to d +1, and returning to the step 3.2); if yes, ending iteration and outputting an average index set
Figure BDA0002996671240000034
3.5) to the set of average indicators
Figure BDA0002996671240000035
Carrying out normalization treatment to obtain NdAnd (4) setting effective indexes S of the new energy station in M control periods.
4) And evaluating the comprehensive performance of the new energy AGC new energy station by utilizing RankBoost based on the new energy AGC effective index set S.
Based on the new energy AGC effective index set S, the step of evaluating the comprehensive performance of the new energy AGC new energy station by using RankBoost comprises the following steps:
4.1) initialization iteration number r is 1, weight D r＝10, objective function H0Setting the maximum iteration number as RT when the value is 0;
4.2) calculating the initial weight Dr(Sk(i),Sk(j) Namely:
Dr(Sk(i),Sk(j))＝c·max{0,φ(Sk(i),Sk(j))} (9)
in the formula, Sk(i) Calculating a k index in the ith new energy station; c is an order
Figure BDA0002996671240000036
Generalization factor of (1), NdThe number of new energy stations to be evaluated is determined;
wherein the sign function phi (S)k(i),Sk(j) As follows)：
Figure BDA0002996671240000041
4.3) obtaining the coefficient factor a by utilizing a sequencing learning methodrOptimal characteristic index hrAnd a loss function Zr；
Calculating a coefficient factor arOptimal characteristic index hrAnd a loss function ZrThe steps of (1):
4.3.1) number of initialization iterations q equals 1, index number k equals 1, parameter RMK is the number of indices;
4.3.2) order thetaq0.01 × q, Sk(i) Establishing a symbolic function m for the calculation result of the kth index in the ith new energy stationkNamely:
Figure BDA0002996671240000042
in the formula, thetaqIs a preset threshold value;
4.3.3) calculating the cumulative sign function R of the ith new energy stationk(i) Namely:
Rk(i)＝Rk(i-1)+mk(i) (12)
wherein R isk(0)＝0；
4.3.4) making i equal to i +1, and returning to the step 4.3.2) until the accumulated symbolic functions of all the new energy stations are calculated;
4.3.5) determining the maximum cumulative sign function RkmaxNamely:
Rkmax＝max{|Rk(i)|,i＝1,2...Nd} (13)
4.3.6) determination of Rkmax>RMIf true, proceed to step 4.3.7), if false, proceed to step 4.3.8);
4.3.7) let parameter RM＝RkmaxCharacteristic index hq＝mkAnd based on the index feature hqCalculating a corresponding ordering loss function ZqNamely:
Figure BDA0002996671240000043
wherein q is the number of iterations, aqIs a coefficient factor; h isq(i) The index characteristics of the ith new energy station are obtained;
coefficient factor aqAs follows:
Figure BDA0002996671240000044
wherein, the weight s of the optimal characteristic index is as follows:
Figure BDA0002996671240000045
in the formula, Sk(i) Calculating a k index in the ith new energy station;
4.3.7) judging whether k is less than or equal to M, if so, making k equal to k +1, and returning to the step 4.3.2); if not, ending the iteration and outputting the optimal characteristic index hq；
4.3.8) judging that q is less than or equal to qmaxIf true, let q be q +1 and return to step 4.3.2), if false, go to step 4.3.9); q. q.smaxIs the maximum iteration number;
4.3.9) comparison of qmaxAll loss function Z under sub-iterationqBy taking the loss function ZqAnd taking the t characteristic index and the coefficient factor under the minimum as the optimal characteristic value and the coefficient factor which are finally output.
4.4) calculating to obtain the sequencing index H of each AGC new energy stationr(i) Namely:
Hr(i)＝Hr-1(i)+ar*hr(i) (17)
4.5) update the weight Dr+1Namely:
Figure BDA0002996671240000051
4.6) determining the number of iterations r>If RT is not true, let r be r +1, and return to step 4.3); if yes, ending the iteration and outputting a new energy station sequencing objective function value Hr。
The technical effects of the method are undoubted, the comprehensive, scientific and visual new energy station AGC frequency modulation performance evaluation method is established, the RankBoost method is adopted for performance evaluation, and compared with the traditional weight value analysis method, the subjectivity of weight value coefficient selection is avoided; according to the invention, a series of indexes such as commissioning, tracking, adjusting speed and generating capacity of the new energy AGC are fully considered, and various factors influencing the frequency modulation performance of the new energy station are scientifically and comprehensively considered; meanwhile, the comprehensive performance evaluation is carried out by adopting an unsupervised machine learning method Rank boost in combination with the characteristics of an index system, and a unique strong sequencing result is obtained through the learning of a plurality of weak sequencing results, so that the defects of the traditional evaluation method are avoided. The method can be widely applied to the evaluation of the new energy AGC new energy station with different installed capacities and different regions and different frequency modulation performances, and can provide beneficial reference for the operation control of the secondary frequency modulation of the power system and the income distribution of the auxiliary frequency modulation market by the new energy.
Drawings
FIG. 1 is a flow chart of the present invention;
fig. 2 is a comprehensive performance ranking diagram of each new energy station.
Detailed Description
The present invention is further illustrated by the following examples, but it should not be construed that the scope of the above-described subject matter is limited to the following examples. Various substitutions and alterations can be made without departing from the technical idea of the invention and the scope of the invention is covered by the present invention according to the common technical knowledge and the conventional means in the field.
Example 1:
referring to fig. 1 to 2, a new energy station AGC comprehensive performance evaluation method based on rankbost includes the following steps:
1) acquiring historical output data of the new energy station, and establishing an effective data set V;
2) calculating a frequency modulation performance index of the new energy station in each control period based on the effective data set V;
the frequency modulation performance indexes comprise a commissioning performance index, a tracking performance index, an adjusting speed index, a generating capacity index, a plan completion rate index and a utilization hour index.
The commissioning performance index comprises commissioning rate K1Namely:
Figure BDA0002996671240000061
in the formula, ToAnd TgThe AGC operation time and the grid-connected operation time of the new energy station are respectively set;
the tracking performance index comprises an excess power generation amount K for representing the response capability of the new energy station to the control command2And a tracking deviation K for representing the tracking capability of the new energy station on the control command3Namely:
Figure BDA0002996671240000062
Figure BDA0002996671240000063
in the formula, N represents the number of sampling points in a control period; i is an arbitrary sampling point; delta PtA value indicating that the AGC actual output is higher than the control command during the adjustment period; piA sampling value representing the power of the new energy station; p1A target output representing a control command;
the regulation speed index comprises an upper speed regulation degree K of the new energy station4And the down-regulation speed K5Namely:
Figure BDA0002996671240000064
in the formula, P0And P1The target output, t, of the initial value of regulation and the control command of the new energy station0And t1The new energy station is used for controlling the output of the new energy station;
the power generation capacity index comprises actual accumulated power generation capacity K used for representing power generation capacity of the new energy station6And theoretical generated energy index K7Namely:
Figure BDA0002996671240000071
wherein N represents the number of sampling points in the control period, PiIs a sampling value of the power of the new energy station,
Figure BDA0002996671240000072
predicting a force value for a theory;
the plan completion rate index comprises a plan completion rate K of the new energy station8Namely:
Figure BDA0002996671240000073
in the formula, PGIs the actual generated energy of the new energy station, PplanThe planned generating capacity of the new energy station is obtained;
the utilization hour number index comprises the utilization hour number K of the new energy station9Namely:
Figure BDA0002996671240000074
in the formula, PGIs the actual generated energy, P, of the AGC new energy stationICThe installed capacity of the new energy station.
3) Calculating NdThe method comprises the steps that an index set I of each new energy station in M control periods is processed to obtain a new energy AGC effective index set S;
calculating NdThe method for setting the index set I of the new energy station in the M control periods comprises the following steps:
3.1) let the new energy station serial number d equal to 1, IdSetting the maximum iteration number N as 0d；
3.2) calculating the AGC frequency modulation performance index of the d-th new energy station in each control period, thereby obtaining an index set I under M control periodsd＝{Id,1,Id,2,…,Id,MF, wherein, index set I under each control periodd,i＝{Kd,i,1,Kd,i,2,…,Kd,i,9}；
3.3) calculating the average index set of the d-th new energy station
Figure BDA0002996671240000075
Namely:
Figure BDA0002996671240000076
3.4) determination of d>NdIf not, making d equal to d +1, and returning to the step 3.2); if yes, ending iteration and outputting an average index set
Figure BDA0002996671240000077
3.5) to the set of average indicators
Figure BDA0002996671240000078
Carrying out normalization treatment to obtain NdAnd (4) setting effective indexes S of the new energy station in M control periods.
4) And evaluating the comprehensive performance of the new energy AGC new energy station by utilizing RankBoost based on the new energy AGC effective index set S.
Based on the new energy AGC effective index set S, the step of evaluating the comprehensive performance of the new energy AGC new energy station by using RankBoost comprises the following steps:
4.1) initialization iteration number r is 1, weight D r＝10, objective function H0Setting the maximum iteration number as RT when the value is 0;
4.2) calculating the initial weight Dr(Sk(i),Sk(j) Namely:
Dr(Sk(i),Sk(j))＝c·max{0,φ(Sk(i),Sk(j))} (9)
in the formula, Sk(i) Calculating a k index in the ith new energy station; c is an order
Figure BDA0002996671240000081
Generalization factor of (1), NdThe number of new energy stations to be evaluated is determined;
wherein the sign function phi (S)k(i),Sk(j) As follows):
Figure BDA0002996671240000082
4.3) obtaining the coefficient factor a by utilizing a sequencing learning methodrOptimal characteristic index hrAnd a loss function Zr；
Calculating a coefficient factor arOptimal characteristic index hrAnd a loss function ZrThe steps of (1):
4.3.1) number of initialization iterations q equals 1, index number k equals 1, parameter RMK is the number of indices;
4.3.2) order thetaq0.01 × q, Sk(i) Establishing a symbolic function m for the calculation result of the kth index in the ith new energy stationkNamely:
Figure BDA0002996671240000083
in the formula, thetaqIs a preset threshold value;
4.3.3) calculating the cumulative sign function R of the ith new energy stationk(i) Namely:
Rk(i)＝Rk(i-1)+mk(i) (12)
wherein R isk(0)＝0；
4.3.4) making i equal to i +1, and returning to the step 4.3.2) until the accumulated symbolic functions of all the new energy stations are calculated;
4.3.5) determining the maximum cumulative sign function RkmaxNamely:
Rkmax＝max{|Rk(i)|,i＝1,2...Nd} (13)
4.3.6) determination of Rkmax>RMIf true, proceed to step 4.3.7), if false, proceed to step 4.3.8);
4.3.7) let parameter RM＝RkmaxCharacteristic index hq＝mkAnd based on the index feature hqCalculating a corresponding ordering loss function ZqNamely:
Figure BDA0002996671240000091
wherein q is the number of iterations, aqIs a coefficient factor; h isq(i) The index characteristics of the ith new energy station are obtained; h isq(j) The index characteristics of the jth new energy station are obtained;
coefficient factor aqAs follows:
Figure BDA0002996671240000092
wherein, the weight s of the optimal characteristic index is as follows:
Figure BDA0002996671240000093
in the formula, Sk(i) For the ith new energy stationThe calculation result of the k index;
4.3.7) judging whether k is less than or equal to M, if so, making k equal to k +1, and returning to the step 4.3.2); if not, ending the iteration and outputting the optimal characteristic index hq；
4.3.8) judging that q is less than or equal to qmaxIf true, let q be q +1 and return to step 4.3.2), if false, go to step 4.3.9); q. q.smaxIs the maximum iteration number;
4.3.9) comparison of qmaxAll loss function Z under sub-iterationqBy taking the loss function ZqAnd taking the t characteristic index and the coefficient factor under the minimum as the optimal characteristic value and the coefficient factor which are finally output.
4.4) calculating to obtain the sequencing index H of each AGC new energy stationr(i) Namely:
Hr(i)＝Hr-1(i)+ar*hr(i) (17)
4.5) update the weight Dr+1Namely:
Figure BDA0002996671240000094
4.6) determining the number of iterations r>If RT is not true, let r be r +1, and return to step 4.3); if yes, ending the iteration and outputting a new energy station sequencing objective function value Hr。
Example 2:
referring to fig. 1, a verification test of an AGC comprehensive performance evaluation method for a new energy station based on RankBoost includes the following steps:
1) selecting eight new energy AGC in a certain inland area in China, acquiring historical statistical data of one year, and establishing an effective data set V;
2) calculating frequency modulation performance indexes of the new energy station in each control period based on an effective data set V of AGC historical data, wherein the frequency modulation performance indexes comprise a commissioning performance index, a tracking performance index, an adjusting speed index, a generating capacity index, a plan completion rate index and a utilization hour index;
the main steps of the frequency modulation index calculation are as follows:
2.1) calculating the commissioning rate K according to the following formula based on AGC statistical data of the new energy station1And representing the commissioning performance of the new energy station by the index.
Figure BDA0002996671240000101
In the formula, ToAnd TgThe AGC operation time and the grid-connected operation time of the new energy station are respectively set.
2.2) introduction of excess power K2Deviation from tracking K3Two different indexes representing the response and tracking capability of the new energy AGC field station to the control command, wherein the over-generation amount K2And a tracking offset K3The calculation formula of (a) is as follows:
Figure BDA0002996671240000102
Figure BDA0002996671240000103
wherein N represents the number of sampling points in the control period, Δ PtIndicating that the actual AGC output is above the value of the control command, P, during the settling periodiSampled values representing the power of the unit, P1Representing the target output of the control command.
2.3) calculating the Up-regulating speed K of the AGC field station according to the following formula4And the down-regulation speed K5:
Figure BDA0002996671240000104
In the formula, P0And P1Respectively, the initial value and the target of the regulation of the unit, t0And t1The time when the unit starts to adjust and the time when the unit reaches the adjustment target are respectively.
2.4) introduction of the actual cumulative energy production K6And theoretical generated energy index K7The power generation capacity of the new energy station is represented, and the calculation formula is as follows:
Figure BDA0002996671240000105
wherein N represents the number of sampling points in the control period, PiIs the sampling power of the unit and is,
Figure BDA0002996671240000106
the force values are predicted for the theory.
2.5) calculating the planned completion rate K of the new energy AGC field station based on the following formula8：
Figure BDA0002996671240000111
In the formula, PGIs the actual power generation of the unit, PplanIs the planned generating capacity of the unit.
2.6) calculating the number of utilization hours K of the set of the new energy AGC field station based on the following formula9:
Figure BDA0002996671240000112
In the formula, PGIs the actual power generation capacity of the AGC unit, PICIs the installed capacity of the unit.
3) Calculating an index set I of 8 new energy stations in 365 control periods, and averaging and normalizing the index set I to obtain an effective index set S;
3.1) let d be 1, IdSetting the maximum iteration number N as 0d＝8；
3.2) calculating the AGC frequency modulation performance index of the d new energy station in each control period, thereby obtaining an index set I under 365 control periodsd＝{Id,1,Id,2,…,Id,365F, wherein, index set I under each control periodd,i＝{Kd,i,1,Kd,i,2,…,Kd,i,9}；
3.3) calculating the average index set of the d unit
Figure BDA0002996671240000113
The calculation formula is as follows:
Figure BDA0002996671240000114
3.4) determination of d>If the result is not true, making d equal to d +1, and entering the step 2; if yes, ending iteration and outputting an average index set
Figure BDA0002996671240000115
3.5) to the set of average indicators
Figure BDA0002996671240000116
And carrying out normalization processing to obtain effective index sets S of 8 new energy stations in 365 control periods, wherein the effective index sets S are shown in table 1:
TABLE 1 normalized AGC index data
Figure BDA0002996671240000117
4) Based on a new energy AGC effective index set S, RankBoost is adopted to carry out comprehensive performance evaluation on a new energy AGC unit, and the steps are as follows:
4.1) initialization iteration number r is 1, weight D r＝10, objective function H0Setting the maximum iteration number as RT to be 50;
4.2) calculating the initial weight Dr(Sk(i),Sk(j) Namely:
Dr(Sk(i),Sk(j))＝c·max{0,φ(Sk(i),Sk(j))} (9)
in the formula, Sk(i) Is the calculation result of the k index in the ith new energy AGC field station, and c is order
Figure BDA0002996671240000121
The generalization factor of (1). Wherein the sign function phi (S)k(i),Sk(j) As follows):
Figure BDA0002996671240000122
4.3) obtaining the coefficient factor a through a sequencing learning processrOptimum characteristic index hrAnd a loss function ZrThe method comprises the following steps:
4.3.1) number of initialization iterations q ═ 1, k ═ 1, RMThe index number K is 9;
4.3.2) order thetaq0.01 × q, Sk(i) Establishing a symbol function m for the calculation result of the kth index in the ith new energy AGC field stationkNamely:
Figure BDA0002996671240000123
in the formula, thetaqIs a given threshold.
4.3.3) calculating the cumulative sign function RkNamely:
Rk(i)＝Rk(i-1)+mk(i) (12)
wherein R isk(0)＝0；
4.3.4) determining RkmaxNamely:
Rkmax＝max{Rk(i)|,i＝1,2...8} (13)
4.3.5) determination of Rkmax>RMIf yes, the step 6 is carried out, and if not, the step 7 is carried out;
4.3.6) reaction of RM＝RkmaxCharacteristic index hq＝mkAnd based on the index feature hqAccording to the following formulaCalculating corresponding sorting loss function Zq：
Figure BDA0002996671240000124
Wherein q is the number of iterations, aqThe calculation method is as follows:
Figure BDA0002996671240000125
wherein, the weight s of the optimal characteristic index is as follows:
Figure BDA0002996671240000126
in the formula, Sk(i) Calculating a k index calculation result in the ith new energy AGC station;
4.3.7), judging whether k is less than or equal to 9, if so, making k equal to k +1, and returning to the step 2); if not, ending the iteration and outputting the optimal characteristic index hq。
4.3.8) determining whether r is less than or equal to 100, if so, making r equal to r +1 and returning to step 2), and if not, performing step 9).
4.3.9) comparison of the loss function Z for the above-mentioned 100 iterationsqBy taking the loss function ZqMinimum hqAnd aqAnd the optimal characteristic value and the coefficient factor are finally output.
4.4) calculating the sequencing index of each AGC unit according to the following formula:
Hr(i)＝Hr-1(i)+ar*hr(i) (17)
4.5) update the weight Dr+1:
Figure BDA0002996671240000131
4.6) determining the number of iterations r>If the RT is not true, let r be r +1, and go to step 3; if yes, ending the iteration and outputting a new energy station sequencing objective function value Hr。
After the Rank-boot machine learning algorithm is adopted, the value of the ranking function H of each station is as shown in FIG. 2, and the larger the ranking function H is, the better the frequency modulation comprehensive performance of the station is. The result shows that the unit 8 is the optimal unit, the unit 2 is the worst unit, and by combining the AGC index data in table 1, it can be easily seen that all indexes of the unit 8 are far greater than the average value of the indexes, and the indexes 1, 3, 6, 8, and 9 are all ranked in the front. The indexes of the inverted unit 2 are basically far lower than the average value, and the indexes 1, 5, 6, 8 and 9 are arranged at the tail end. Therefore, the selected sequencing result of the invention is consistent with the performance of the unit, and the reasonability of the model provided by the invention is verified from the side. Compared with conventional methods such as an analytic hierarchy process and the like, the RankBoost method generates a unique strong sorting result by combining a plurality of weak sorting results, and avoids the difficulty of selecting index weight coefficients when comprehensively evaluating the performance of a plurality of indexes.

Claims (6)

1. A new energy station AGC comprehensive performance evaluation method based on RankBoost is characterized by comprising the following steps:
1) and acquiring historical output data of the new energy station, and establishing the effective data set V.
2) Calculating a frequency modulation performance index of the new energy station in each control period based on the effective data set V;
3) calculating NdThe method comprises the steps that an index set I of each new energy station in M control periods is processed to obtain a new energy AGC effective index set S;
4) and evaluating the comprehensive performance of the new energy AGC new energy station by utilizing RankBoost based on the new energy AGC effective index set S.
2. The method for evaluating the comprehensive performance of the new energy station AGC based on the RankBoost as claimed in claim 1, wherein the method comprises the following steps: the frequency modulation performance indexes comprise a commissioning performance index, a tracking performance index, an adjusting speed index, a generating capacity index, a plan completion rate index and a utilization hour index.
3. The method for evaluating AGC (automatic Generation control) comprehensive performance of the new energy station based on the RankBoost as claimed in claim 2, wherein the commissioning performance index comprises commissioning rate K1Namely:
Figure FDA0002996671230000011
in the formula, ToAnd TgThe AGC operation time and the grid-connected operation time of the new energy station are respectively set;
the tracking performance index comprises an excess power generation amount K for representing the response capability of the new energy station to the control command2And a tracking deviation K for representing the tracking capability of the new energy station on the control command3Namely:
Figure FDA0002996671230000012
Figure FDA0002996671230000013
in the formula, N represents the number of sampling points in a control period; i is an arbitrary sampling point; delta PiA value indicating that the AGC actual output is higher than the control command during the adjustment period; piA sampling value representing the power of the new energy station; p1A target output representing a control command;
the regulation speed index comprises an upper speed regulation degree K of the new energy station4And the down-regulation speed K5Namely:
Figure FDA0002996671230000021
in the formula, P0And P1The target output, t, of the initial value of regulation and the control command of the new energy station0And t1The new energy station is used for controlling the output of the new energy station;
the power generation capacity index comprises actual accumulated power generation capacity K used for representing power generation capacity of the new energy station6And theoretical generated energy index K7Namely:
Figure FDA0002996671230000022
wherein N represents the number of sampling points in the control period, PiIs a sampling value of the power of the new energy station,
Figure FDA0002996671230000023
predicting a force value for a theory;
the plan completion rate index comprises a plan completion rate K of the new energy station8Namely:
Figure FDA0002996671230000024
in the formula, PGIs the actual generated energy of the new energy station, PplanThe planned generating capacity of the new energy station is obtained;
the utilization hour number index comprises the utilization hour number K of the new energy station9Namely:
Figure FDA0002996671230000025
in the formula, PGIs the actual generated energy, P, of the AGC new energy stationICThe installed capacity of the new energy station.
4. A substrate according to claim 1The AGC comprehensive performance evaluation method for the new energy station of RankBoost is characterized by calculating NdThe method for setting the index set I of the new energy station in the M control periods comprises the following steps:
1) let the serial number d of the new energy station be 1, IdSetting the maximum iteration number N as 0d；
2) Calculating AGC frequency modulation performance indexes of the d-th new energy station in each control period, thereby obtaining an index set I under M control periodsd＝{Id,1,Id,2,…,Id,MF, wherein, index set I under each control periodd,i＝{Kd,i,1,Kd,i,2,…,Kd,i,9}；
3) Calculating an average index set of the d-th new energy station
Figure FDA0002996671230000031
Namely:
Figure FDA0002996671230000032
4) judgment of d>NdIf the result is not true, making d equal to d +1, and returning to the step 2); if yes, ending iteration and outputting an average index set
Figure FDA0002996671230000033
5) For average index set
Figure FDA0002996671230000034
Carrying out normalization treatment to obtain NdAnd (4) setting effective indexes S of the new energy station in M control periods.
5. The method for evaluating the comprehensive performance of the new energy station AGC based on the RankBoost according to claim 1, wherein the step of evaluating the comprehensive performance of the new energy station AGC by using the RankBoost based on the new energy AGC effective index set S comprises the following steps:
1) the number of initialization iterations r is 1, and the weight Dr＝10, objective function H0Setting the maximum iteration number as RT when the value is 0;
2) calculating an initial weight Dr(Sk(i),Sk(j) Namely:
Dr(Sk(i),Sk(j))＝c·max{0,φ(Sk(i),Sk(j))} (9)
in the formula, Sk(i) Calculating a k index in the ith new energy station; c is an order
Figure FDA0002996671230000035
Generalization factor of (1), NdThe number of new energy stations to be evaluated is determined;
wherein the sign function phi (S)k(i),Sk(j) As follows):
Figure FDA0002996671230000036
3) obtaining coefficient factor a by using sequencing learning methodrOptimal characteristic index hrAnd a loss function Zr；
4) Calculating to obtain a sequencing index H of each AGC new energy stationr(i) Namely:
Hr(i)＝Hr-1(i)+ar*hr(i) (11)
5) update the weight value to Dr+1Namely:
Figure FDA0002996671230000037
6) judging the number of iterations r>If the RT is not established, making r be r +1, and returning to the step 3); if yes, ending the iteration and outputting a new energy station sequencing objective function value Hr。
6. According to the rightThe AGC comprehensive performance evaluation method for the new energy station based on the RankBoost, as recited in claim 5, characterized in that a coefficient factor a is calculatedrOptimal characteristic index hrAnd a loss function ZrThe steps of (1):
1) the number of initialization iterations q is 1, the index number k is 1, and the parameter RMK is the number of indices;
2) let thetaq0.01 × q, Sk(i) Establishing a symbolic function m for the calculation result of the kth index in the ith new energy stationkNamely:
Figure FDA0002996671230000041
in the formula, thetaqIs a preset threshold value;
3) calculating the accumulative sign function R of the ith new energy stationk(i) Namely:
Rk(i)＝Rk(i-1)+mk(i) (14)
wherein R isk(0)＝0；
4) Making i equal to i +1, and returning to the step 2) until the accumulative sign functions of all the new energy stations are calculated;
5) determining a maximum cumulative sign function RkmaxNamely:
Rkmax＝max{|Rk(i)|,i＝1,2...Nd} (15)
6) judgment of Rkmax>RMWhether the determination is true, if true, the routine proceeds to step 7), and if false, the routine proceeds to step 8);
7) let parameter RM＝RkmaxCharacteristic index hq＝mkAnd based on the index feature hqCalculating a corresponding ordering loss function ZqNamely:
Figure FDA0002996671230000042
in which q is an overlapGeneration number, aqIs a coefficient factor; h isq(i) The index characteristics of the ith new energy station are obtained;
coefficient factor aqAs follows:
Figure FDA0002996671230000043
wherein, the weight s of the optimal characteristic index is as follows:
Figure FDA0002996671230000044
in the formula, Sk(i) Calculating a k index in the ith new energy station;
7) judging whether k is less than or equal to M, if so, making k equal to k +1, and returning to the step 2); if not, ending the iteration and outputting the optimal characteristic index hq；
8) Judging q is less than or equal to qmaxIf true, making q equal to q +1, and returning to step 2), if false, proceeding to step 9); q. q.smaxIs the maximum iteration number;
9) comparison qmaxAll loss function Z under sub-iterationqBy taking the loss function ZqAnd taking the t characteristic index and the coefficient factor under the minimum as the optimal characteristic value and the coefficient factor which are finally output.
CN202110334224.6A 2021-03-29 2021-03-29 New energy station AGC comprehensive performance evaluation method based on RankBoost Active CN113162028B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110334224.6A CN113162028B (en) 2021-03-29 2021-03-29 New energy station AGC comprehensive performance evaluation method based on RankBoost

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110334224.6A CN113162028B (en) 2021-03-29 2021-03-29 New energy station AGC comprehensive performance evaluation method based on RankBoost

Publications (2)

Publication Number Publication Date
CN113162028A true CN113162028A (en) 2021-07-23
CN113162028B CN113162028B (en) 2023-05-30

Family

ID=76885562

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110334224.6A Active CN113162028B (en) 2021-03-29 2021-03-29 New energy station AGC comprehensive performance evaluation method based on RankBoost

Country Status (1)

Country Link
CN (1) CN113162028B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113743724A (en) * 2021-08-02 2021-12-03 南方电网科学研究院有限责任公司 New energy station configuration energy storage evaluation method, system, medium and power terminal
CN115508650A (en) * 2022-10-13 2022-12-23 西安德纳检验检测有限公司 New energy station frequency modulation detection method and system based on multipoint synchronous measurement

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101441239A (en) * 2008-12-09 2009-05-27 张家港三得利新能源科技有限公司 Verification method of parallel networking type photovoltaic power station power generation performance
CN102290825A (en) * 2011-08-05 2011-12-21 辽宁省电力有限公司 Real-time measuring and evaluation-querying system for regulating performance of power generating set
CN103543403A (en) * 2013-09-02 2014-01-29 国家电网公司 Method for detecting primary frequency modulation capacity of power system units
CN104517199A (en) * 2015-01-16 2015-04-15 国家电网公司 New energy power generation online monitoring method based on real time data
CN104834479A (en) * 2015-04-24 2015-08-12 清华大学 Method and system for automatically optimizing configuration of storage system facing cloud platform
CN111242451A (en) * 2020-01-07 2020-06-05 中国南方电网有限责任公司 Power frequency modulation auxiliary service clearing method, system, device and storage medium
CN111293681A (en) * 2020-01-22 2020-06-16 重庆大学 Photovoltaic station output fluctuation quantitative evaluation method based on RankBoost

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101441239A (en) * 2008-12-09 2009-05-27 张家港三得利新能源科技有限公司 Verification method of parallel networking type photovoltaic power station power generation performance
CN102290825A (en) * 2011-08-05 2011-12-21 辽宁省电力有限公司 Real-time measuring and evaluation-querying system for regulating performance of power generating set
CN103543403A (en) * 2013-09-02 2014-01-29 国家电网公司 Method for detecting primary frequency modulation capacity of power system units
CN104517199A (en) * 2015-01-16 2015-04-15 国家电网公司 New energy power generation online monitoring method based on real time data
CN104834479A (en) * 2015-04-24 2015-08-12 清华大学 Method and system for automatically optimizing configuration of storage system facing cloud platform
CN111242451A (en) * 2020-01-07 2020-06-05 中国南方电网有限责任公司 Power frequency modulation auxiliary service clearing method, system, device and storage medium
CN111293681A (en) * 2020-01-22 2020-06-16 重庆大学 Photovoltaic station output fluctuation quantitative evaluation method based on RankBoost

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
冯署能: "辅助服务中心AGC性能改善方法及竞争模型研究", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ专辑（月刊）》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113743724A (en) * 2021-08-02 2021-12-03 南方电网科学研究院有限责任公司 New energy station configuration energy storage evaluation method, system, medium and power terminal
CN113743724B (en) * 2021-08-02 2024-05-28 南方电网科学研究院有限责任公司 New energy station configuration energy storage evaluation method, system, medium and power terminal
CN115508650A (en) * 2022-10-13 2022-12-23 西安德纳检验检测有限公司 New energy station frequency modulation detection method and system based on multipoint synchronous measurement

Also Published As

Publication number Publication date
CN113162028B (en) 2023-05-30

Similar Documents

Publication Publication Date Title
CN107482692B (en) Active control method, device and system for wind power plant
CN113162028A (en) New energy station AGC comprehensive performance evaluation method based on RankBoost
CN111431215B (en) Double-layer control method for participating in frequency modulation of power distribution network side through large-scale energy storage
CN112186761B (en) Wind power scene generation method and system based on probability distribution
CN113555881A (en) Wind power plant frequency modulation fan sequencing method
CN105573123A (en) Thermal power generating unit mechanical furnace coordination control method based on improved T-S fuzzy prediction modeling
CN111245032B (en) Voltage prediction control method considering loss reduction optimization of wind power plant collector line
Malmir et al. Controlling megawatt class WECS by ANFIS network trained with modified genetic algorithm
CN110397553B (en) Model-free wind power plant wake flow management method and system
CN116470513A (en) Multi-type photo-thermal power station coordinated scheduling operation method responding to power grid requirements
CN110336285B (en) Optimal economic power flow calculation method for power system
CN114336592A (en) Wind power plant AGC control method based on model predictive control
CN108270244B (en) Method for scheduling new energy power system in multiple control domain operation modes
CN115758193A (en) Distributed energy storage aggregation control method and device
CN115764875A (en) Method for determining inertia and frequency modulation reserve capacity of power system
CN111082442B (en) Energy storage capacity optimal configuration method based on improved FPA
CN111178601B (en) Wind turbine generator power prediction method based on meteorological data post-processing
CN113410873A (en) Regional comprehensive energy system configuration method considering source load multiple uncertainty
CN110932334A (en) Wind power plant power control method with constraint multi-objective optimization
CN112968480B (en) Wind-thermal power combined optimization scheduling method and system based on unit load response capability
CN110717694A (en) Energy storage configuration random decision method and device based on new energy consumption expected value
CN117439090B (en) Flexible resource allocation or scheduling method taking flexible adjustment coefficient as index
CN115377963B (en) Novel temperature set value control method for heterogeneous cluster air conditioner load
CN114944659B (en) Automatic control method and system for wind-solar-energy-storage combined power station
CN116388299B (en) Wind-solar energy storage station group power tracking optimization control method, system and equipment

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
GR01 Patent grant
GR01 Patent grant