CN103345546B - The governor parameter discrimination method that frequency locus combines with particle cluster algorithm - Google Patents

Abstract

The present invention relates to the governor parameter discrimination method that a kind of frequency locus combines with particle cluster algorithm, can be applicable to machine unit speed regulating device parameter is carried out parameter identification.The present invention is to set up the input/output model of speed regulator system, and demonstrates the identifiability of relevant parameter；The electromechanical transient simulation computing module of PSASP software is directly invoked by interface, and obtain governor parameter, including obtain power system disturbed after simulation frequency curve, and then based on particle swarm optimization algorithm governor parameter is adjusted so that simulation frequency curve and practical frequency curve as close possible to；Optimum governor parameter is obtained by parameter optimization repeatedly.The present invention only need to utilize the failure wave-recording curve of frequency and the parameter of the other parts of system, gets final product the parameter identification of the governing system of completion system, can avoid the most loaded down with trivial details on-the-spot test, facilitate dispatcher to obtain speed regulator relevant parameter in time.

Description

The governor parameter discrimination method that frequency locus combines with particle cluster algorithm

Technical field

The present invention relates to the governor parameter discrimination method that a kind of frequency locus combines with particle cluster algorithm, can be applicable to Machine unit speed regulating device parameter is carried out parameter identification.

Background technology

Governing system directly affects electricity as the important component part of generator control system, the correctness of its parameter The stability of network operation and security.Refer to " modern power systems analysis " that Wang Xifan writes. Beijing: Science Press, 2007.Due to system operation mode conversion frequently, the governing system of each unit for the conversion of the method for operation of adaptive system, Often it is also required to frequently adjust.For control centre, obtain system all set speed adjustments system the most in time Parameter, is one of the strong guarantee of system safety in operation and stability.

About the parameter identification problem of governing system, research and engineers and technicians propose many interesting ways, Substantially can be divided into two class methods at present.

First kind method is that test combines identification technique and obtains the parameter of governing system." the steamer write by Wen Libin Machine and electrohydraulic governor system model important parameter method of testing thereof ", Guangxi electric power .2012,35 (1): 9-12,16, propose to adopt Obtaining generator power measurement links parameter by load dump test, by introducing test command signal, and servo card PID signal obtains Take the parameter of PID loop joint.

" research of turbine regulating system parameter identification method " write by Wu Haicheng, Cai Bing, Chen Shengli etc., Anhui electricity Power, 2010,27 (1): 21-25, according to differential evolution algorithm and combine steam turbine load disturbance experiments and propose steam turbine speed control The method of Parameter identification.

" Parameter Recognition of Steam Turbine Governor System Elements and the application thereof " write by Sun Ping, Yuan Baiying, China Power, 1995,11:7-12, propose to be obtained by off-line testing input and the curve of output of steam turbine electro-hydraulic governing system, then use Least square method carries out identification to the parameter of governing system links and obtains corresponding governing system parameter.Such method Advantage is that obtained model and parameter are relatively more accurate, and shortcoming is just must to carry out in the case of unit maintenance.

Equations of The Second Kind method is based on measured curve, carries out the identification of corresponding set speed adjustment systematic parameter.

" difficulty when actual measurement track extracts knowledge and the prospect " write by xuwei, Xue Yusheng, Chen Shi etc., power system Automation, 2009,33 (15): 1-7, it is proposed that utilize system failure practical frequency fluctuation path implementation set speed adjustment systematic parameter The conception of identification, but specific algorithm is not proposed.

By Li Yulong, Wang Hao, appoint to continue and red etc. write " power system frequency Dynamic Simulation Results degree of accuracy factor analysis ", black Longjiang electric power, 2012,34 (1): 50-52,61；And " made merit frequency process imitate by what Liu Hongbo, Mu Gang, Xu Xingwei etc. write True track approaches the model parameter research of actual measurement track ".Electric power network technique, 2006,30 (18): 20-24.These two articles analyze There is with frequency fluctuation track the coefficient type of the governing system of strong correlation relation, and obtain as drawn a conclusion: difference system adjusted by speed regulator Number affects minimum point and the rise frequency that frequency declines, and governor dead time affects the minimum point that frequency declines.

" complicated electric power system dynamic process of frequency based on the track analysis " write by Xu Xingwei, Beijing, North China electric power University, 2010.Propose so-called system frequency Stability Assessment index, including initial frequency fall, frequently fall time, frequently drop angle rate, return Rise frequency etc.；But the index that this article is carried its essence is the evaluation index of the fluctuation size of system frequency track, and non-frequency is steady Determine index.

Summary of the invention

Governing system parameter based on practical frequency track is realized for currently available technology does not has concrete effective ways Identification, the invention provides the governor parameter discrimination method that a kind of frequency locus combines with particle cluster algorithm.The present invention is Traffic department obtains the problem of governor parameter in time and provides new method, and the present invention devises when obtaining simulation curve simultaneously The interface of PSASP simulation algorithm model can be directly invoked, realize Transient calculation provide new approaches for calling PSASP, it is desirable to should Invent and play a significant role in the middle of these problems of solution.

The technical solution adopted for the present invention to solve the technical problems is:

The governor parameter discrimination method that frequency locus combines with particle cluster algorithm is: set up the input of speed regulator system Output model, and demonstrate the identifiability of relevant parameter；The electromechanical transient simulation of PSASP software is directly invoked by interface Computing module, and obtain governor parameter, including acquisition power system simulation frequency curve after disturbed, and then based on particle Colony optimization algorithm governor parameter is adjusted so that simulation frequency curve and practical frequency curve as close possible to；By instead Multiple parameter optimization and obtain optimum governor parameter.

The described input/output model setting up speed regulator system, when setting up the input/output model of speed regulator system, Require that the electrohydraudic servomechanism of speed regulator, steam turbine model, and generator system, excitation system and network etc. are known System.

Described interface, by ODBC interface, Putcollect and Getcollect statement, SetCursorPos function and Mouse_event functions etc. realize reading in from file, read governor parameter and directly invoke PSASP software transient stability meter Calculate module.

During described acquisition governor parameter, analyze for a plurality of measured curve simultaneously, the function of structure be based on A plurality of measured curve obtains with PSASP simulation curve.

Described concrete calculation process, is first to give iterations and convergence and governor parameter to be identified Initial value and constant interval, be then read into parameter the concrete module of PSASP and carry out multilayer output feedback network, by simulation frequency number According to reading into corresponding document, and calculate acquisition exemplary frequency deviation values, it is judged that whether iteration convergence criterion meets, and utilize population to calculate Method is optimized.

Described basic identification principle, is first to obtain reality outputExport with model, obtaining reality outputWith Model exportsRear calculating obtains deviation delta z, uses optimized algorithm to be adjusted, when inclined difference DELTA z reaches deviation delta z obtained Parameter to the corresponding speed regulator electro-hydraulic control system also just obtained time minimum.

The present invention proposes the new method of the governor parameter identification that a kind of frequency locus combines with particle cluster algorithm.First First set up the input/output model of speed regulator system, and demonstrate the identifiability of relevant parameter；Then by design interface Directly invoke PSASP software electromechanical transient simulation computing module and obtain power system disturbed after simulation frequency curve, And then based on particle swarm optimization algorithm, governor parameter is adjusted so that simulation frequency curve and practical frequency curve to the greatest extent may be used Can be close；Obtained optimum governor parameter by parameter optimization repeatedly, send out it is achieved thereby that based on practical frequency curve The identification of group of motors governing system parameter.

The invention has the beneficial effects as follows: only need to utilize the failure wave-recording curve of frequency and the parameter of the other parts of system, Get final product the parameter identification of the governing system of completion system, the most loaded down with trivial details on-the-spot test can be avoided, facilitate dispatcher timely Obtain speed regulator relevant parameter.

The present invention is further detailed explanation below in conjunction with the accompanying drawings.

Accompanying drawing explanation

Fig. 1 is that total system of the present invention transmits function block diagram；

Fig. 2 is that module 1 transmits function；

Fig. 3 is that module 2 transmits function；

Fig. 4 is that module 3 transmits function；

Fig. 5 is that module 4 transmits function；

Fig. 6 is that module 5 transmits function；

Fig. 7 is that module 6 transmits function；

Fig. 8 is the schematic diagram of the present invention；

Fig. 9 is the calculation flow chart of the present invention；

Figure 10 is 20090731 failure-frequency tracks/simulation frequency track；

Figure 11 is 20090906 failure-frequency tracks/simulation frequency track；

Figure 12 is 200909062 failure-frequency tracks/simulation frequency track；

Figure 13 is 20091107 failure-frequency tracks/simulation frequency track；

Figure 14 is 20091111 failure-frequency tracks/simulation frequency track.

Detailed description of the invention

The present invention is the governor parameter discrimination method that a kind of frequency locus combines with particle cluster algorithm.Mainly complete Governing system modeling analysis；The Identification of parameter of speed regulator electro-hydraulic control system based on optimisation technique realizes；Soft with PSASP Part calls relevant operation realization and waits the work of three aspects.

The governor parameter discrimination method that frequency locus of the present invention combines with particle cluster algorithm is: set up speed regulator system Input/output model, and demonstrate the identifiability of relevant parameter；PSASP software is directly invoked by the interface of design Electromechanical transient simulation computing module, and obtain governor parameter, including acquisition power system simulation frequency curve after disturbed, And then based on particle swarm optimization algorithm, governor parameter is adjusted so that simulation frequency curve and practical frequency curve to the greatest extent may be used Can be close；Optimum governor parameter is obtained by parameter optimization repeatedly.

This governor parameter discrimination method is when setting up the input/output model of speed regulator system, it is desirable to the electricity of speed regulator Liquid servo control mechanism, steam turbine model, and generator system, excitation system and network etc. are known system.

This governor parameter discrimination method is when design interface, by ODBC interface, Putcollect and Getcollect statement, SetCursorPos function and mouse_event function etc. realize reading in from file, read speed regulator Parameter also directly invokes PSASP software multilayer output feedback network module.

This governor parameter discrimination method, when obtaining governor parameter, is to analyze for a plurality of measured curve simultaneously , the function of structure obtains with PSASP simulation curve based on a plurality of measured curve.

Concrete operation step is as follows:

1, governing system modeling.Speed regulator electro-hydraulic control system is modeled in detail, the electrohydraudic servomechanism of speed regulator, Steam turbine model, and generator system, excitation system and network etc. are known system.In speed regulator electro-hydraulic control system The transmission function block diagram of the whole system including excitation, generator and network under Zuo Yong, as shown in Figure 1.In FIG, G₁ (s) represent electrohydraudic servomechanism, steam turbine model etc. four the internal remainder of type speed regulator comprehensively transmit function；G₂(s) generation Generator system included by table mechanical output-frequency translation process, excitation system and comprehensively transmit letter corresponding to network Number.Y (s) represents system frequency excursion, and U (s) represents system input, in this system, U (s)=Y (s).

Due to the internal remainder of the speed regulator such as electrohydraudic servomechanism, steam turbine model, generator system during whole All parameters of system, excitation system and network are known and constant, therefore G₁(s), G₂S () is changeless.Have only to Transmission function to each link in dotted line frame in Fig. 1, sets up the input-output function relation of corresponding each link, and then Set up the functional relation between the input of whole speed regulator electro-hydraulic control system and output.

Separately below each module in Fig. 1 dotted line frame is modeled respectively, and then obtains its whole mathematical model.

Transmitting function for the module 1 shown in Fig. 2, wherein U (s) represents the system frequency excursion Δ f of input.x₁Representative is treated Identified parameters T₁；Y₁S () represents output；U, y₁It is respectively input and output time-domain functional form (as follows).Then this transmission function The corresponding differential equation is:

$u=y_1+\frac{{\mathrm{dy}}_1}{\mathrm{dt}}{T}_1$

Abbreviation obtains:

y₁=f (T₁, u)=f₁(x₁,u) (1)；

Function, Y are transmitted for module 2 as shown in Figure 3₁S () represents input, Y₂S () represents output, x₂Represent to be identified Parameter ε, K；Transmission function is turned to algebraic equation obtain:

Make f₂₁(y₁,x₂)=K(y₁+ε/2),f₂₂(y₁,x₂)=K(y₁-ε/2),

f₂₃(y₁,x₂)=0。

Then $y_2=\left\{\begin{array}{cc}K(y_1+\mathrm{\&epsiv;}/2)& y_1<-\mathrm{\&epsiv;}/2\\ K(y_1-\mathrm{\&epsiv;}/2)& y_1>\mathrm{\&epsiv;}/2\\ 0& -\mathrm{\&epsiv;}/2\&le;y_1\&le;\mathrm{\&epsiv;}/2\end{array}\right.=\left\{\begin{array}{cc}{f}_{21}(y_1,x_2)& y_1<{-x}_2/2\\ f_{22}(y_1,x_2)& y_1>x_2/2\\ f_{23}(y_1,x_2)& -x_2/2\&le;y_1\&le;x_2/2\end{array}\right.---\left(2\right)$ ；

Function, Y are transmitted for module 3 as shown in Figure 4₃S () represents input, Y₄S () represents output, x₃Represent to be identified Parameter K_I.As seen from Figure 1, input Y₃(s)=P_REF-Y₂(s)-P_E, wherein P_REFRepresent given load, P_ERepresent feedback negative Lotus, therefore, with the equation corresponding to the transmission function represented by Fig. 1-9 be:

Make f₄₁(x₃,y₃)=P_up,f₄₂(x₃,y₃)=P_down,

f₄₃(x₃,y₃)=∫ K_Iy₃Dt=∫ x₃ y₃ dt

Then $y_4=f_4(x_3,y_3)$

$=\left\{\begin{array}{cc}{P}_{\mathrm{up}}& f_{43}(x_3,y_3)>P_{\mathrm{up}}\\ P_{\mathrm{down}}& f_{43}(x_3,y_3)<P_{\mathrm{down}}\\ f_{43}(x_3,y_3)& P_{\mathrm{down}}\&le;f_{43}(x_3,y_3)\&le;P_{\mathrm{up}}\end{array}\right.=\left\{\begin{array}{cc}{f}_{41}(x_3,y_3)& f_{43}(x_3,y_3)>P_{\mathrm{up}}\\ f_{42}(x_3,y_3)& f_{43}(x_3,y_3)<P_{\mathrm{down}}\\ f_{43}(x_3,y_3)& P_{\mathrm{down}}\&le;f_{43}(x_3,y_3)\&le;P_{\mathrm{up}}\end{array}\right.---\left(3\right)$ ；

Function, Y are transmitted for module 4 as shown in Figure 5₄S () represents input, Y₅S () is output, x₄Represent ginseng to be identified Number K_P,K_D.The equation corresponding with Fig. 5 is:

$y_5=y_4+K_P{y}_3+K_D\frac{{\mathrm{dy}}_3}{\mathrm{dt}}{f}_5(y_3,y_4,x_4)---\left(4\right);$

Function, Y are transmitted for module 5 as shown in Figure 6₅S () represents input；x₅Represent parameter to be identified, including P_max、 P_min；Y₆S () represents output.The equation corresponding with Fig. 6 is:

Make f₆₁(x₅,y₅)=P_max,f₆₂(x₅,y₅)=P_min,

f₆₃(x₅,y₅)=ky₅

Then $y_6=f_6(x_5,y_5)$

$=\left\{\begin{array}{cc}{P}_{\max }& f_{63}(x_5,y_5)>P_{\max }\\ P_{\min }& f_{63}(x_5,y_5)<P_{\min }\\ f_{63}(x_5,y_5)& P_{\min }\&le;f_{63}(x_5,y_5)\&le;P_{\max }\end{array}\right.=\left\{\begin{array}{cc}{f}_{61}(x_5,y_5)& f_{63}(x_5,y_5)>P_{\max }\\ f_{62}(x_5,y_5)& f_{63}(x_5,y_5)<P_{\min }\\ f_{63}(x_5,y_5)& P_{\min }\&le;f_{63}(x_5,y_5)\&le;P_{\max }\end{array}\right.---\left(5\right)$ ；

Function, Y are transmitted for module 6 as shown in Figure 7₆(s), Y₇S () represents input；As shown in Figure 1, Y₇=K₂(P_REF- Y₂), P_CVRepresent output, represent the pitch instruction exported by electro-hydraulic control system.The equation corresponding with Fig. 7 is:

P_CV=y₆+y₇=f₇(y₆,y₇) (6)；

U is substituted into formula (1) and obtains y₁, by y₁Substitution formula (2) obtains y₂, by y₂Represent y₃And y₇, and by y₃Substitution formula (3) Obtain y₄, by y₃And y₄Substitution formula (4) obtains y₅, by y₅Substitution formula (5) obtains y₆, by y₆And y₇Substitution formula (6) obtains in Fig. 1 empty Input represented in line frame graph and output function expression formula:

P_CV=f (u, x) (7)；

Wherein u represents is that the system frequency excursion Δ f, x of input represents the to be identified of each link in electro-hydraulic control system Parameter, including T₁,K,ε,K_I,K_D,K_P,K₂,P_up,P_down,P_max,P_min；P_CVRepresent output mechanical power.May certify that, formula (7) In parameter x be cognizable.

2, the Identification of parameter of speed regulator electro-hydraulic control system based on optimisation technique.

If the actual value vector of parameter to be identified in Fig. 1Represent, actually enter variable vectorRepresent, push away Output function express formula formula (7) and G₁(s),G₂The corresponding time-domain expression ff (P of (s)_CV) be multiplied, when obtaining Fig. 1 overall Territory expression formula；If the actual output vector of Fig. 1Represent, then the mathematic(al) representation between input and the output of Fig. 1 is:

$\stackrel{\&RightArrow;}{z}=f(\stackrel{\&RightArrow;}{x},\stackrel{\&RightArrow;}{u})\&CenterDot;\mathrm{ff}\left(f(\stackrel{\&RightArrow;}{x},\stackrel{\&RightArrow;}{u})\right)---\left(8\right);$

If the variable estimate to be identified of speed regulator electro-hydraulic control system is designated asEmulation input isTime corresponding model defeated Go out forAccording to the identification principle of the present invention shown in Fig. 8, first obtain reality outputExport with modelReal obtaining Border exportsExport with modelRear calculating obtains deviation, uses optimized algorithm to be adjusted deviation delta z obtained, works as deviation The parameter of the corresponding speed regulator electro-hydraulic control system also just obtained when value Δ z minimizes.Then speed regulator electro-hydraulic control system Parameter identification problem be ask the output of the output so that analogue system (i.e. model) and real system closest in the case of Optimal parameter problem, available mathematic(al) representation is expressed as:

$\underset{\hat{x}}{\mathrm{Min}}{(\hat{z}-\stackrel{\&RightArrow;}{z})}^2=\underset{\hat{x}}{\mathrm{Min}}{(f(\hat{x},\hat{u})\&CenterDot;\mathrm{ff}\left(f(\hat{x},\hat{u})\right)-\stackrel{\&RightArrow;}{z})}^2---\left(9\right);$

The present invention is practical frequency track and PSASP software systems parameter identification process being combined real system, proposes The parameter identification new method of speed regulator electro-hydraulic control system.Basic ideas are: for all systems in addition to speed regulator, bag Include excitation system, generator system, network all known s bar frequency record wave profile, one group of speed regulator electro-hydraulic control system of primary election Initial parameter, utilizes PSASP software to obtain the s bar frequency simulation track corresponding to this group speed regulator electro-hydraulic control system parameter, And calculate the s bar frequency simulation track deviation relative to actual measurement track respectively, choose emulation and corresponding practical frequency curve centering 1 pair of curve that deviation value is maximum, is modified existing speed regulator electro-hydraulic control system parameter according to optimized algorithm, and obtains new Speed regulator electro-hydraulic control system parameter；And carry out new round calculating and correction, such loop iteration as stated above, until stop Only calculate criterion and meet requirement.First its concrete calculation flow chart as it is shown in figure 9, give iterations and convergence and treat The initial value of identification governor parameter and constant interval, be then read into parameter the concrete module of PSASP and carry out transient stability Calculate, by simulation frequency data read-out to corresponding document, and calculate acquisition exemplary frequency deviation values, it is judged that iteration convergence criterion is the fullest Foot, and utilize particle cluster algorithm to be optimized.Details are as follows:

(1) initialize: k=0；Given iterations N, convergence criterion ε；The initial value of given governor parameter to be identified and Its constant interval.

(2) estimate based on parameter to be identified(k represent carry out kth subparameter estimation), is read into PSASP by parameter Concrete module also carries out multilayer output feedback network, it is thus achieved that organize the s bar frequency simulation track of governor parameter (corresponding to being corresponding to this The s kind fault of system), and utilize formula (10) to calculate the s bar frequency simulation track deviation relative to actual measurement track.

$g_i\left({\hat{x}}_k\right)\underset{t=0}{\overset{m}{\mathrm{\&Sigma;}}}{[{\hat{z}}_i({\hat{x}}_k,\hat{u},t)-{\stackrel{\&RightArrow;}{z}}_i(\stackrel{\&RightArrow;}{x},\stackrel{\&RightArrow;}{u},t)]}^2---\left(10\right);$

Wherein, i represents i-th frequency locus, m represent frequency curve discretization after hits, t represents frequency locus The t sampled point,Represent the kth time estimate of parameter to be identified,Represent the true value of identified parameters,Table T sampled point is corresponded to after showing the i-th article of frequency simulation track discretization utilizing kth time estimates of parameters to be identified to obtain Value,The value of t sampled point is corresponded to after representing i-th article of frequency actual measurement track discretization,Represent to treat I-th frequency simulation track of the kth of identified parameters time estimate acquisition and the quadratic sum of the deviation of actual measurement track.

(3) curve pair that emulation is maximum with practical frequency curve alignment deviation is chosen, it may be assumed that

$g\left({\hat{x}}_k\right)=\underset{i}{\max }\left[g_i\left({\hat{x}}_k\right)\right]---\left(11\right);$

IfIllustrate that the simulation curve of deviation maximum meets requirement with the deviation value of measured curve, obtain The best estimate of parameter to be identified, iteration terminates；Otherwise use Particle Swarm Optimization, based on the emulation chosen and reality Border frequency curve pair, for current estimate to be measuredIt is modified, it is thus achieved that its new value ${\hat{x}}_{k+1}$ 。

(4) k=k+1, turns to step (1).

The realization of 3 operations relevant to PSASP software transfer.

For the simulation calculation of real system, being an important step in inventive algorithm, the present invention is connect by design Mouth directly invokes the computing module of Based on Power System Analysis Software Package (PSASP), and has conveniently realized simulation calculation.

PSASP software provides user-defined modeling (UD) and user program interface (UPI) environment, but PSASP is also Do not provide computing module and the interface of output data, it is impossible to realize directly invoking and place to result of calculation of its computing module Reason.The present invention devises PSASP software computing module and user program interface interchange method, it is achieved that its computing module and calculating The calling and processing of result.In the present invention, in governor parameter identification algorithm, the data relevant to PSASP software are read automatically The concrete methods of realizing of the 3 aspect problems such as call taking, the most automatically carrying out computing module and result of calculation is as follows:

(1) from file, governor parameter is read in.

PSASP provides the parameter reading in all control systems from interactive interface, but does not provide reading phase from file The parameter function answered.Owing to after every suboptimization, the parameter of speed regulator is stored in file, therefore to simulation calculation is anti- Carrying out again, the governor parameter that must realize being stored in file reads in the operation in the computing module of PSASP.

The commpar.dbc database being stored under lib file due to the governor parameter of PSASP software systems In gov.dbf form, i.e. its computing module can read the relevant parameter being stored in gov.dbf form automatically.Therefore, this Bright employing ODBC is as interface, it is achieved file links with gov.dbf form, and by Putcollect statement by corresponding document In governor parameter be read in the gov.dbf form of commpar.dbc database, it is achieved thereby that governor parameter is from literary composition Part directly reads in PSASP calculation procedure.Corresponding logical code is:

(2) the most automatically the calling of PSASP multilayer output feedback network module；

In order to the emulation of many class frequencys record wave profile is approached, need to carry out repeatedly calling of PSASP computing module.And PSASP system provides only and once calculates, and the manual operation triggering the start button that must pass through interactive interface calculated just can be entered OK, do not provide calculate automatically to carry out.The method of the present invention is by batch processing module hstbatcal.exe, SetCursorPos function and mouse_event function realize the repeated multiple times of PSASP computing module and automatically call.Specifically Wpsasp.exe program interface is opened, then by SetCursorPos function and mouse_ by window function findwindow Event function call batch processing module hstbatcal.exe, obtains start button position point in PSASP system interaction interface Hit, thus real the most automatically calling now for computing module.Corresponding logical code is as follows:

While (" not clicking on multilayer output feedback network button ")

Sleep();

WM-close();

End。

(3) the frequency locus curve data obtained by simulation calculation reads in corresponding document；

After PSASP has calculated, its result is automatically credited database stresult.dbc under result file In Output table, the present invention is by ODBC interface, and Getcollect statement is by obtained by the simulation calculation in Output table Frequency locus curve data read and be stored in corresponding txt file.Due to this process with from file governor parameter The process in corresponding gov.dbf form of reading in is similar to, and its corresponding logical code just repeats no more at this.

The present invention is by carrying out application experiment in the Speed-adjustable system parameter identification of said method electrical network northeastward, to verify the party The validity of method.The following unit of Northeast China Power Grid 100MW is to use mechanical governing system, and existing field measurement parameter, but 100MW and above unit, be respectively mounted electric gas/liquid pressure pressure type governing system, and Northeast China Power Grid has carried out the governor model ginseng of typical case's unit Number actual measurement work, wherein: Suizhong factory #2 machine uses PSASP3 type governor model and actual measurement parameter；Bai Shan factory #5 machine uses PSASP7 type governor model and actual measurement parameter；Yi Min factory #1 machine, #4 machine uses PSASP4 type governor model and actual measurement parameter, And the governing system for other unit does not has measured data.

The governing system of other unit surveying unit place power plant is directly applied mechanically this factory and surveys parameter, to unit The PSASP4 type governor model of the governing system equal Northeast China Power Grid scheduling suggestion that more than capacity 100MW does not surveys, i.e. before this paper State model, and according to unit capacity, the parameter of corresponding governing system be divided into 3 classes: unit capacity at 100MW to 300MW, its Governor parameter is the 1st class；Unit capacity is at 300MW to 500MW, and its governor parameter is the 2nd class；Unit capacity exists More than 500MW's, its governor parameter is the 3rd class.The method described in the present invention is utilized to pick out above three groups of tune respectively The parameter of speed device, and contrast with practical frequency curve, to prove its validity.

Northeast China Power Grid has frequency record ripple geometric locus its corresponding network structure known and unit for 2009 except governing system Outside the failure condition of all control system parameters be shown in Table 1.

Owing to Northeast China Power Grid has detailed recorder data to front four groups of faults, sample once every 0.02s, so according to table In 1, front four groups of faults each systematic parameter corresponding and frequency record ripple geometric locus, use method proposed by the invention to east The parameter of the 3 class set speed adjustment systems that NORTEL net is not surveyed carries out repeatedly identification, selects the most best one group as finally distinguishing The result known, the 3 class governor parameter values obtained are as shown in table 2.After in the 3 class governor parameters picked out and table 1 Other parameter of system that five groups of fault phases are corresponding carries out the contrast emulating obtained system frequency curve with frequency record ripple track Figure is all listed in Figure 10 to Figure 14.

It can be seen that utilize the governor parameter of institute of the present invention identification to combine corresponding failure and system from Figure 10 to Figure 14 The parameter of other parts can obtain with frequency record ripple geometric locus very close to frequency curve.Also demonstrate institute of the present invention accordingly The validity of extracting method, thus avoid the most loaded down with trivial details on-the-spot test, only need to utilize the failure wave-recording curve of frequency and system The parameter of other parts, gets final product the parameter identification of the governing system of completion system.

Table 1 Northeast China Power Grid failure condition in 2009.

Table 2 Northeast China Power Grid speed regulator system 3 class identified parameters.

Claims (1)

1. the governor parameter discrimination method that frequency locus combines with particle cluster algorithm, including the input setting up speed regulator system Output model, and demonstrate the identifiability of relevant parameter；Obtain governor parameter, including obtain power system disturbed after Simulation frequency curve, is characterized in that: described acquisition governor parameter, is the machine being directly invoked PSASP software by interface Electricity transient emulation computing module, and then based on particle swarm optimization algorithm, governor parameter is adjusted so that simulation frequency curve With practical frequency curve as close possible to；Optimum governor parameter is obtained by parameter optimization repeatedly；

The described input/output model setting up speed regulator system, when setting up the input/output model of speed regulator system, it is desirable to The electrohydraudic servomechanism of speed regulator, steam turbine model, and generator system, excitation system and network are known system；

Described interface, by ODBC interface, Putcollect and Getcollect statement, SetCursorPos function and Mouse_event function realizes reading in from file, reads governor parameter and directly invokes PSASP software multilayer output feedback network Module；

During described acquisition governor parameter, analyzing for a plurality of measured curve, the function of structure is based on a plurality of simultaneously Measured curve obtains with PSASP simulation curve；Its concrete calculation process, be first give iterations and convergence and The initial value of governor parameter to be identified and constant interval, be then read into parameter the concrete module of PSASP and to carry out transient state steady Devise a stratagem is calculated, and by simulation frequency data read-out to corresponding document, and calculates acquisition exemplary frequency deviation values, it is judged that iteration convergence criterion whether Meet, and utilize particle cluster algorithm to be optimized；

The identifiability of described relevant parameter, its identification principle, is first to obtain reality outputExport with modelObtaining Actual outputExport with modelRear calculating obtains deviation delta z, uses optimized algorithm to be adjusted deviation delta z obtained, when The parameter of the corresponding speed regulator electro-hydraulic control system also just obtained when difference DELTA z minimizes partially.