CN105406521A - Method for computing new index for evaluating unit AGC regulation performance - Google Patents

Abstract

The invention discloses a method for computing a new index for evaluating unit AGC regulation performance, comprising the following steps that actual outputs and target outputs of all units participating into AGC regulation are recorded per 5 seconds; and a computing unit tracks the correlation index C, delay index D, mileage index M, precision index P and overall regulation performance index K of an AGC control command. The method provided by the invention completely adopts the unit to track the historical data of the actual output and target output of the AGC control command without being influenced by set unit evaluation parameters, so that the time frame of the unit with poor tracking effect can be effectively recognized, the regulation performance of the AGC unit for tracking the AGC control command can be comprehensively and accurately evaluated, and the regulation performance of the unit under different running states can be highlighted through modifying different performance index weights.

Description

A kind of computational methods evaluating unit AGC adjusting function New Set

Technical field

The invention belongs to electric power system real power control field, the present invention relates to more precisely and a kind ofly utilize that unit is actual exerts oneself and target historical data statistical computation of exerting oneself obtains evaluating the method for unit AGC adjusting function New Set.

Background technology

Frequency is one of important indicator of the quality of power supply, is also the main control parameters of power system operation.Modern power systems ensures that the important tool of frequency quality is automatic generation control (AutomaticGenerationControl, AGC) apply, the amount of unbalance of load and generating in AGC Real-Time Monitoring electrical network, and by this amount of unbalance by ACE (AreaControlArea, district control deviation) mode distribute in electrical network the unit participating in AGC and control, the balance of control command guarantee electrical network hair electricity of unit by following the tracks of AGC and issuing.Therefore, the ability that unit follows the tracks of AGC control objectives is the key factor ensureing frequency quality of power grid, grid company is assessed the quality of unit AGC adjusting function with regard to needing, and the corresponding examination of formulation and way of rewards and punishments promote and optimize the AGC adjusting function of unit.

Document one " AGC unit regulation efficiency qualitative assessment and compensation way are studied " (electric power network technique, calendar year 2001,25 volume 8 phases: 15-19 page) disclose a kind of method that unit to dropping into automatic generation control operation carries out rational adjusting function qualitative assessment and examination, and propose the regulation efficiency of AGC unit to be divided into 3 Performance Coefficient: pondage, regulations speed and adjusting deviation, mention these indexs in literary composition and comparatively objectively can reflect the actual contribution of AGC unit to system.

Document two " the North China AGC unit regulation efficiency on-line determination software development based on CC-2000 system " (electric power network technique, 29th volume 18 phase 22-25 page in 2005) similar with document one, disclose the pondage of the real-time regulation efficiency of unit, regulations speed and degree of regulation 3 indexs, the basis taking into account thermal power unit operation feature analyzes the method for these 3 indexs being carried out to quantitatively calculating, and gives the algorithm flow based on CC-2000 system establishment AGC unit regulation efficiency on-line determination software.

The method that document three " evaluating unit AGC adjusting function new method " (electrical network and clean energy resource, 25 volumes the 2nd phase 21-27 page in 2009) discloses one " loudspeaker " curve evaluation unit AGC adjusting function is simply and intuitively evaluated the quality that unit performs AGC instruction.Unit practical adjustments process and standard adjustment process graphically compare by " loudspeaker " curve that this article discloses, subregion is carried out to unit tracing process, for subregion calculation of performance indicators respectively, then two subregion index values are added the adjusting function overall target K obtaining AGC unit.

Make a general survey of the sign AGC unit adjusting function index mentioned in current document, follow the tracks of the actual tracing process proposition of single control command process based on unit substantially, mainly comprise the response time of unit, regulations speed, degree of regulation and pondage etc., these parameters are physically clearer and more definite, computational process is also fairly simple, but also there are some problems, the result of calculation of the AGC unit adjusting function index such as proposed at present is larger by the impact of unit examination parameter, the instruction less with actual output deviation of exerting oneself of part of generating units target is then difficult to calculate the adjusting function index of unit simultaneously, the adjustment process that the computational methods of the adjusting function index of current practical application do not follow the tracks of even reverse regulation for unit in addition also cannot effectively calculate.

Summary of the invention

In order to the AGC adjusting function of more accurate thoroughly evaluating unit, the effect of comprehensive reflection unit trace command, the historical data regulated by participating in AGC to unit carries out mathematical analysis, propose correlation metric C, retardance index D, inner level index M and accuracy index P, and by the weighted analysis to these indexs, realize rational evaluation unit being followed the tracks of to AGC control command adjusting function.

For achieving the above object, the present invention takes following technical scheme to realize:

Evaluate computational methods for unit AGC adjusting function New Set, comprise the following steps:

1) to the unit that all participation AGC regulate, that records a unit every 5s actually to exert oneself and target is exerted oneself, and will record result writing in files;

2) calculate the correlation metric C that unit follows the tracks of AGC control command, computational methods are as follows:

21) definition 5 minutes is a fixing computing cycle, and each fixing computing cycle comprises 60 sampled points;

22) at each fixing computing cycle initial time, the unit target calculated in 5 minutes is thereafter exerted oneself the relative coefficient of data sequence and the actual data sequence of exerting oneself of unit;

23) δ time of delay is set, δ ∈ [5 seconds, 10 seconds, 15 seconds, 20 seconds ... 300 seconds], keep unit target sequence of exerting oneself constant, the actual sequence of exerting oneself of unit postpones a sampled point successively backward, calculates each actual exert oneself sequence and unit target postponed and to exert oneself the relative coefficient of data sequence;

24) get step 23) in maximum in all relative coefficients take advantage of the correlation metric Cmin being this fixed cycle in 100%;

25) calculate the correlation metric of each whole 5 minutes initial time in a hour, averaged is as the correlation metric Chour of this hour of unit;

3) calculate the retardance index D that unit follows the tracks of AGC control command, computational methods are as follows:

31) in step 2) calculate in the process of correlation metric, the time of delay that each fixed cycle initial time unit actual exert oneself sequence and target are exerted oneself corresponding to serial correlation coefficient maximum can be obtained, equal decay time this time of delay and take advantage of in sampling time interval;

32) according to the retardance index Dmin calculating this fixed cycle time of delay again, computational methods are:

$D=(1-\frac{t}{T})\×100\%---\left(1\right)$

In formula: t is time of delay, T is fixed cycle duration, is 300 seconds here;

33) calculate the retardance index Dmin of in a hour each whole 5 minutes, averaged is the retardance index Dhour of this hour of unit;

4) calculate unit and follow the tracks of level index M in AGC control command, computational methods are as follows:

41) calculate unit target in each whole 5 minutes to exert oneself change absolute value summation, as the theoretical mileage of unit, be designated as M _expected;

42) calculate the actual tracking target of exerting oneself of unit in each whole 5 minutes exert oneself change summation, be designated as M _actualthis value is actual actual poor the adding up of exerting oneself with a upper moment of exerting oneself when being the instruction of unit tracing control, exert oneself if actual and exceed target and exert oneself, be not counted in, comprise actual the exerting oneself of timing and be greater than target and exert oneself and actual the exerting oneself of lower timing is less than target and exerts oneself two kinds of situations;

43) calculate level index M in unit these whole 5 minutes, method is as follows:

$M=\frac{M_{actual}}{M_{\exp ected}}\×100\%---\left(2\right)$

44) to calculate in one hour level index Mmin in each whole 5 minutes, averaged is level index Mhour in this hour of unit;

5) calculate the accuracy index P of unit, computational methods are as follows:

51) first obtain the specified regulations speed of unit, unit is converted into MW/s by MW/min;

52) analyze the target of unit to exert oneself sequence, build a virtual target and to exert oneself sequence; Construction method is as follows: the real goal for each unit is exerted oneself, if the next sampled point that unit realistic objective is exerted oneself equals current sampling point, then unit virtual target is exerted oneself and equaled real goal and exert oneself; Exert oneself if next sampled point is not equal to current target, illustrate that AGC has issued new control command to unit, exert oneself from lower beginning unit virtual target and equal to exert oneself from current goal that basis goes out force according to specified regulations speed slope to next target is close, until virtual target is exerted oneself, the target equaling next point is exerted oneself;

53) fixing any sampling instant in 5 minute cycle, the accuracy FACTOR P s adopting formula (3) to calculate unit to follow the tracks of, then the Ps of all sampling instants in 5 minutes is averaging, obtain each accuracy index P fixing 5 minute cycle unit;

$P=(1-|\frac{P_{actualgen}-P_{v-t\arg et}}{Avg{\left(P_{v-t\arg et}\right)}_{\min }}\left|\right)\×100\%---\left(3\right)$

In formula: P _actualgenfor actual the exerting oneself of each sampling instant of unit, P _v-targetfor the virtual target of each sampling instant is exerted oneself, Avg (P _v-target) _minfor the mean value that virtual target in current 5 minutes sections is exerted oneself;

54) calculate the accuracy index Pmin of in a hour each whole 5 minutes, averaged is the accuracy index Phour of this hour of unit;

6) calculate unit and follow the tracks of AGC control command adjusting function overall performane K, computational methods are as follows:

K _hour＝αC _hour+βD _hour+γM _hour+ηP _hour(4)

C in formula _hour, D _hour, M _hour, P _hourbe respectively the correlation metric of each whole hour unit, retardance index, inner level index and accuracy index; Wherein:

α+β+γ+η＝1(5)

After calculating the adjusting function overall performane K of each hour, then calculate the average adjusted performance index K of 24 hours every days.

Further, in described step 2) in, in each fixing computing cycle, utilize that unit is actual exerts oneself and target is exerted oneself the correlation of sequence and the correlation metric of the percentage by postponing the correlation maximum that the actual sequence of exerting oneself of unit obtains as unit.

Further, in described step 3) in, within each fixed cycle, unit is actual exerts oneself and target exerts oneself time of delay of correlation maximum as the degree of delay index of unit.

Further, in described step 4) in, within each fixed cycle, add up unit target respectively and to exert oneself the actual variable quantity of exerting oneself of absolute change amount and the actual tracking of unit, the percentage of both business is as level index in unit.

Further, in described step 5) in, within each fixed cycle, when discovery unit target exerts oneself change, according to the specified regulations speed of unit, fictionalize the virtual controlling target of the desirable aircraft pursuit course of unit as unit, then to exert oneself the percentage of business of mean value with virtual target output deviation and fixed cycle internal object as the accuracy index of unit with actual exerting oneself.

Further, by step 2), 3), 4), 5) each independent performance index of obtaining are weighted and obtain unit comprehensive adjustment performance index, the weight coefficient summation of each individual index is 1.

Further, by revising the value of weight coefficient, the significance level of amendment individual index in overall target.

Beneficial effect of the present invention:

(1) what the evaluation unit AGC adjusting function index that the present invention proposes utilized unit to follow the tracks of AGC control command completely actually to exert oneself and target is exerted oneself historical data, the impact of the unit examination parameter do not arranged, effectively can identify the unit tracking effect poor period, the adjusting function that AGC unit follows the tracks of AGC control command can be evaluated more comprehensively and accurately, and can by amendment different performance index weights, the adjusting function of unit under outstanding different running status;

(2) what the present invention utilized unit to follow the tracks of AGC control command actually to exert oneself and target is exerted oneself historical data, adopt set time section to actually exerting oneself, target sequence of exerting oneself carries out mathematical analysis, draw and comprise correlation metric C, retardance index D, inner level index M and accuracy index P tetra-single indexs, again by being weighted analysis to these indexs, draw a unit adjusting function overall target K, unit follows the tracks of the quality of AGC control command to utilize these indexs accurately to reflect, facilitates grid company to examine compensation accordingly to AGC unit.

Be described further below with reference to the technique effect of accompanying drawing to design of the present invention, concrete structure and generation, to understand object of the present invention, characteristic sum effect fully.

Accompanying drawing explanation

Fig. 1 is correlation metric and the degree of delay index calculate Principle of Process figure of the embodiment of the present invention;

Fig. 2 is actual mileage and desirable mileage schematic diagram in the mileage degree index calculate process of the embodiment of the present invention;

Fig. 3 is that the accuracy index virtual target of embodiment of the present invention sequence of exerting oneself generates schematic diagram.

Embodiment

Be described below in detail embodiments of the invention, the example of described embodiment is shown in the drawings.Be exemplary below by the embodiment be described with reference to the drawings, be intended to for explaining the present invention, and can not limitation of the present invention be interpreted as.

Evaluate computational methods for unit AGC adjusting function New Set, comprise the following steps:

1) to the unit that all participation AGC regulate, an every 5 seconds records unit actual is exerted oneself and target is exerted oneself, and will record result writing in files, every day a file; As shown in Figure 1;

2) calculate the correlation metric C that unit follows the tracks of AGC control command, computational methods are as follows:

21) definition 5 minutes is a fixing computing cycle, and unit is actual exerts oneself and target is exerted oneself sampled point data break 5 seconds, and namely every 60 sampled points are a fixing computing cycle;

22) at each fixing computing cycle initial time, the unit target calculated in 5 minutes is thereafter exerted oneself the relative coefficient of data sequence and the actual data sequence of exerting oneself of unit, as shown in Figure 1;

23) δ time of delay is set, δ ∈ [5,10,15,20 ..., 300] and (unit is second, 5 seconds, interval, totally 60 values), keep unit target sequence of exerting oneself constant, the actual sequence of exerting oneself of unit postpones a sampled point successively backward, calculate each relative coefficient postponed, as shown in Figure 1;

24) maximum of getting in all 60 relative coefficients is taken advantage of in 100% is again the correlation metric Cmin of this fixed cycle;

25) calculate the correlation metric of each whole 5 minutes initial time in a hour, averaged is as the correlation metric Chour of this hour of unit;

Particularly, as shown in Figure 1, to a fixing computing cycle i.e. 5 minutes (00:00:00-00:04:55) from 00:00:00, this 5 minutes internal objects are exerted oneself sequence Y, in these 5 minutes, actual sequence of exerting oneself is X1 (numbering from 1 to 60), calculate target exert oneself sequence Y and actual sequence X 1 of exerting oneself relative coefficient C1 and record; Then, form actual sequence X 2 (numbering from 2 to 61) of exerting oneself by moving a sampled point after actual sequence of exerting oneself, then calculate target exert oneself sequence Y and actual sequence X 2 of exerting oneself relative coefficient C2 and record; By that analogy, calculate target respectively and to exert oneself 60 relative coefficient Ci of sequence Y and the actual sequence X i that exerts oneself, get maximum relative coefficient in these 60 relative coefficients and be multiplied by 100% as the correlation metric Cmin of this 5 minute period; Move afterwards again 5 minutes (namely from 00:05:00-00:09:55) calculate this after move the correlation metric Cmin of 5 minute period, by that analogy, calculate the correlation metric of each whole 5 minutes initial time in a hour, averaged is as the correlation metric Chour of this hour of unit.

3) calculate the retardance index D that unit follows the tracks of AGC control command, computational methods are as follows:

31) in step 2) calculate in the process of correlation metric, the time of delay that each fixed cycle initial time unit actual exert oneself sequence and target are exerted oneself corresponding to serial correlation coefficient maximum can be obtained, equal decay time this time of delay and take advantage of in sampling time interval;

32) according to the retardance index Dmin calculating this fixed cycle time of delay again, computational methods are:

$D=(1-\frac{t}{T})\×100\%---\left(1\right)$

In formula: t is time of delay, T is fixed cycle duration, is 300 seconds here.

34) calculate the retardance index Dmin of in a hour each whole 5 minutes, averaged is the retardance index Dhour of this hour of unit;

4) calculate unit and follow the tracks of level index M in AGC control command, computational methods are as follows:

41) calculate unit target in each whole 5 minutes to exert oneself change absolute value summation, as the theoretical mileage of unit, be designated as M _expected;

42) calculate the actual tracking target of exerting oneself of unit in each whole 5 minutes exert oneself change summation, be designated as M _actualthis value is actual actual poor the adding up of exerting oneself with a upper moment of exerting oneself when being the instruction of unit tracing control, exert oneself if actual and exceed target and exert oneself, be not counted in (upper timing is actual exert oneself be greater than that target is exerted oneself, actual the exerting oneself of lower timing be less than target and exert oneself), as shown in Figure 2;

43) calculate level index M in unit these whole 5 minutes, method is as follows:

$M=\frac{M_{actual}}{M_{\exp ected}}\×100\%---\left(2\right)$

45) to calculate in one hour level index Mmin in each whole 5 minutes, averaged is level index Mhour in this hour of unit;

5) calculate the accuracy index P of unit, computational methods are as follows:

51) first obtain the specified regulations speed of unit, unit is converted into MW/s by MW/min;

52) analyze the target of unit to exert oneself sequence, build a virtual target and to exert oneself sequence.Real goal for each unit is exerted oneself, if the next sampled point that unit realistic objective is exerted oneself equals current sampling point, then unit virtual target is exerted oneself and equaled real goal and exert oneself; Exert oneself if next sampled point is not equal to current target, illustrate that AGC has issued new control command to unit, exert oneself from lower beginning unit virtual target and equal to exert oneself from current goal that basis goes out force according to specified regulations speed slope to next target is close, until virtual target is exerted oneself, the target equaling next point is exerted oneself, as is shown in phantom in fig. 3;

53) fixing any sampling instant in 5 minute cycle, the accuracy FACTOR P s adopting formula (3) to calculate unit to follow the tracks of, then the Ps of all sampling instants in 5 minutes is averaging, obtain each accuracy index P fixing 5 minute cycle unit;

$P=(1-|\frac{P_{actualgen}-P_{v-t\arg et}}{Avg{\left(P_{v-t\arg et}\right)}_{\min }}\left|\right)\×100\%---\left(3\right)$

In formula: P _actualgenfor actual the exerting oneself of each sampling instant of unit, P _v-targetfor the virtual target of each sampling instant is exerted oneself, Avg (P _v-target) _minfor the mean value that virtual target in current 5 minutes sections is exerted oneself.

55) calculate the accuracy index Pmin of in a hour each whole 5 minutes, averaged is the accuracy index Phour of this hour of unit;

6) calculate unit and follow the tracks of AGC control command adjusting function overall performane K, computational methods are as follows:

K _hour＝αC _hour+βD _hour+γM _hour+ηP _hour(4)

C in formula _hour, D _hour, M _hour, P _hourbe respectively the correlation metric of each whole hour unit, retardance index, inner level index and accuracy index; Wherein:

α+β+γ+η＝1(5)

After calculating the adjusting function overall performane K of each hour, then calculate the average adjusted performance index K of 24 hours every days.

More than describe preferred embodiment of the present invention in detail.Should be appreciated that those of ordinary skill in the art just design according to the present invention can make many modifications and variations without the need to creative work.Therefore, all technical staff in the art, all should by the determined protection range of claims under this invention's idea on the basis of existing technology by the available technical scheme of logical analysis, reasoning, or a limited experiment.

Claims (7)

1. evaluate computational methods for unit AGC adjusting function New Set, it is characterized in that: comprise the following steps:

1) to the unit that all participation AGC regulate, that records a unit every 5s actually to exert oneself and target is exerted oneself, and will record result writing in files;

2) calculate the correlation metric C that unit follows the tracks of AGC control command, computational methods are as follows:

21) definition 5 minutes is a fixing computing cycle, and each fixing computing cycle comprises 60 sampled points;

22) at each fixing computing cycle initial time, the unit target calculated in 5 minutes is thereafter exerted oneself the relative coefficient of data sequence and the actual data sequence of exerting oneself of unit;

23) δ time of delay is set, δ ∈ [5 seconds, 10 seconds, 15 seconds, 20 seconds ... 300 seconds], keep unit target sequence of exerting oneself constant, the actual sequence of exerting oneself of unit postpones a sampled point successively backward, calculates each actual exert oneself sequence and unit target postponed and to exert oneself the relative coefficient of data sequence;

24) get step 23) in maximum in all relative coefficients take advantage of the correlation metric Cmin being this fixed cycle in 100%;

25) calculate the correlation metric of each whole 5 minutes initial time in a hour, averaged is as the correlation metric Chour of this hour of unit;

3) calculate the retardance index D that unit follows the tracks of AGC control command, computational methods are as follows:

31) in step 2) calculate in the process of correlation metric, the time of delay that each fixed cycle initial time unit actual exert oneself sequence and target are exerted oneself corresponding to serial correlation coefficient maximum can be obtained, equal decay time this time of delay and take advantage of in sampling time interval;

32) according to the retardance index Dmin calculating this fixed cycle time of delay again, computational methods are:

$D=(1-\frac{t}{T})\×100\%---\left(1\right)$

In formula: t is time of delay, T is fixed cycle duration, is 300 seconds here;

33) calculate the retardance index Dmin of in a hour each whole 5 minutes, averaged is the retardance index Dhour of this hour of unit;

4) calculate unit and follow the tracks of level index M in AGC control command, computational methods are as follows:

41) calculate unit target in each whole 5 minutes to exert oneself change absolute value summation, as the theoretical mileage of unit, be designated as M _expected;

42) calculate the actual tracking target of exerting oneself of unit in each whole 5 minutes exert oneself change summation, be designated as M _actualthis value is actual actual poor the adding up of exerting oneself with a upper moment of exerting oneself when being the instruction of unit tracing control, exert oneself if actual and exceed target and exert oneself, be not counted in, comprise actual the exerting oneself of timing and be greater than target and exert oneself and actual the exerting oneself of lower timing is less than target and exerts oneself two kinds of situations;

43) calculate level index M in unit these whole 5 minutes, method is as follows:

$M=\frac{M_{actual}}{M_{\exp ected}}\×100\%---\left(2\right)$

44) to calculate in one hour level index Mmin in each whole 5 minutes, averaged is level index Mhour in this hour of unit;

5) calculate the accuracy index P of unit, computational methods are as follows:

51) first obtain the specified regulations speed of unit, unit is converted into MW/s by MW/min;

52) analyze the target of unit to exert oneself sequence, build a virtual target and to exert oneself sequence; Construction method is as follows: the real goal for each unit is exerted oneself, if the next sampled point that unit realistic objective is exerted oneself equals current sampling point, then unit virtual target is exerted oneself and equaled real goal and exert oneself; Exert oneself if next sampled point is not equal to current target, illustrate that AGC has issued new control command to unit, exert oneself from lower beginning unit virtual target and equal to exert oneself from current goal that basis goes out force according to specified regulations speed slope to next target is close, until virtual target is exerted oneself, the target equaling next point is exerted oneself;

53) fixing any sampling instant in 5 minute cycle, the accuracy FACTOR P s adopting formula (3) to calculate unit to follow the tracks of, then the Ps of all sampling instants in 5 minutes is averaging, obtain each accuracy index P fixing 5 minute cycle unit;

$P=(1-|\frac{P_{actua\lg en}-P_{v-t\arg et}}{Avg{\left(P_{v-t\arg et}\right)}_{min}}\left|\right)\×100\%---\left(3\right)$

In formula: P _actualgenfor actual the exerting oneself of each sampling instant of unit, P _v-targetfor the virtual target of each sampling instant is exerted oneself, Avg (P _v-target) _minfor the mean value that virtual target in current 5 minutes sections is exerted oneself;

54) calculate the accuracy index Pmin of in a hour each whole 5 minutes, averaged is the accuracy index Phour of this hour of unit;

6) calculate unit and follow the tracks of AGC control command adjusting function overall performane K, computational methods are as follows:

K _hour＝αC _hour+βD _hour+γM _hour+ηP _hour(4)

C in formula _hour, D _hour, M _hour, P _hourbe respectively the correlation metric of each whole hour unit, retardance index, inner level index and accuracy index; Wherein:

α+β+γ+η＝1(5)

After calculating the adjusting function overall performane K of each hour, then calculate the average adjusted performance index K of 24 hours every days.

2. a kind of New Set evaluating unit AGC adjusting function according to claim 1, it is characterized in that: in described step 2) in, in each fixing computing cycle, utilize that unit is actual exerts oneself and target is exerted oneself the correlation of sequence and the correlation metric of the percentage by postponing the correlation maximum that the actual sequence of exerting oneself of unit obtains as unit.

3. a kind of New Set evaluating unit AGC adjusting function according to claim 1, it is characterized in that: in described step 3) in, within each fixed cycle, unit is actual exerts oneself and target exerts oneself time of delay of correlation maximum as the degree of delay index of unit.

4. a kind of New Set evaluating unit AGC adjusting function according to claim 1, it is characterized in that: in described step 4) in, within each fixed cycle, add up unit target respectively to exert oneself the actual variable quantity of exerting oneself of absolute change amount and the actual tracking of unit, the percentage of both business is as level index in unit.

5. a kind of New Set evaluating unit AGC adjusting function according to claim 1, it is characterized in that: in described step 5) in, within each fixed cycle, when discovery unit target exerts oneself change, according to the specified regulations speed of unit, fictionalize the virtual controlling target of the desirable aircraft pursuit course of unit as unit, then to exert oneself the percentage of business of mean value with virtual target output deviation and fixed cycle internal object as the accuracy index of unit with actual exerting oneself.

6. a kind of New Set evaluating unit AGC adjusting function according to any one of claim 1 to 5, it is characterized in that: by step 2), 3), 4), 5) each independent performance index of obtaining are weighted and obtain unit comprehensive adjustment performance index, the weight coefficient summation of each individual index is 1.

7. a kind of New Set evaluating unit AGC adjusting function according to claim 6, is characterized in that: by revising the value of weight coefficient, the significance level of amendment individual index in overall target.