CN103812122A - Method for PSS parameter optimization setting based on constraint mutative scale - Google Patents

Abstract

A method for PSS parameter optimization setting based on a constraint mutative scale comprises a first step of obtaining phase-frequency characteristic data of a generator excitation system through a test or theoretical method; a second step of establishing an optimization model for solving a PSS lead-lag link time constant, adopting a constraint mutative scale method to solve the model, obtaining an optimum value of the PSS lead-lag link time constant of a machine set, and enabling the value to be an optimum setting value; and a third step of calculating an amplification coefficient of a PSS according to signal damping in the PSS, the excitation system and a generator link. A practical setting value needs to undergo test verification and is generally very close to a calculated value.

Description

Based on constrained variable metric PSS parameter optimization setting method

Technical field

The invention belongs to power system stability control technology field.

Background technology

Along with putting into operation and electrical network interconnected of a large amount of units, the dynamic antivibration of long-distance and large-capacity power transmission declines, and causes regional power grid generation low-frequency oscillation when serious, threatens power network safety operation.At present, the main PSS(of employing power system stabilizer, PSS) improve dynamic antivibration or suppress low-frequency oscillation, PSS is an optional feature of excitation system, it produces additional signal by extracting the signal relevant with vibration, make generator produce additional damping moment, the dynamic antivibration of raising system and the conveying capacity of each section, suppress low-frequency oscillation and occur.

If there is no suitable PSS parameter, PSS just can not provide additional damping to suppress low-frequency oscillation effectively, even likely helps increasing effect to having vibrated.PSS parameter tuning is generally completed by power test unit at present, and method is varied.As: the phase-frequency characteristic test data of unit is input to Matlab program, adopts evolutionary programming algorithm to draw PSS lead-lag time constant after need to processing.After the artificial processing of data, cause the PSS parameter of adjusting inaccurate, unreasonable, the dynamic antivibration deficiency that causes this unit to provide system, and affect conveying capacity.For example Xiaowan Hydroelectric Power Station #1 unit PSS test in the time doing BPA program stability check, is found the positive damping deficiency of this unit to Grid oscillation mode after adjusting, afterwards in #2 unit test time update setting parameter.

Summary of the invention

Object of the present invention provide in order to overcome the defect that above-mentioned prior art exists just a kind of can effectively provide additional damping suppress low-frequency oscillation based on constrained variable metric PSS parameter optimization setting method.

The present invention adopts and sets up Optimized model, adopts constrained variable metric method to solve PSS lead-lag time constant; PSS amplification coefficient is calculated in the damping producing in PSS, excitation system, generator link according to input signal; The inventive method convergence is good, highly versatile, have high practical value.

The present invention is achieved by the following technical solution.

Summary of the invention is divided into two parts:

(1) solve the Optimized model of PSS lead-lag time constant

Its model is:

$\begin{array}{c}\min F({\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="HDA0000417501590000011.png,HDA0000417501590000012.png"\; state="\{\{state\}\}">T</figure-callout>}}_1,T_2,T_3,T_4,T_5,T_6)=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{({\mathrm{\&theta;}}_{\mathrm{PSS},i}-{\mathrm{\&theta;}}_{\mathrm{GE},i})}^2\\ s.t.\left\{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="HDA0000417501590000011.png,HDA0000417501590000012.png"\; state="\{\{state\}\}">T</figure-callout>}}_1\&le;15T_2,T_3\&le;15T_4,T_5\&le;15T_6\\ T_{k,\min }\&le;T_k\&le;T_{k,\min },k=1,...,6\end{array}\right.\end{array}---\left(1\right)$

θ _PSS,i=arctg(ω _iT ₁)+arctg(ω _iT ₃)+arctg(ω _iT ₅)-arctg(ω _iT ₂)-arctg(ω _iT ₄)-arctg(ω _iT ₆)（2）

Wherein: θ _{pSS, i}for PSS model is in angular frequency _ithe offset angle at place, θ _{gE, i}for generator excited system phase-frequency characteristic is at ω _ithe phase angle at place, N is test data group number, T _kfor the lead-lag time constant of PSS model, T _{k, min}, T _{k, max}for minimum value and the maximum of PSS model lead-lag time constant regulation.

This model adopts constrained variable metric Optimized model to solve.Constrained variable metric optimization method and program, have that convergence is fast, efficiency is high, good reliability, be suitable for the advantages such as ability is strong, is the part achievement (country identified in 1986) of " six or five " national science and technology research project;

(2) solve the model of PSS amplification coefficient

The amplification coefficient K of PSS model _pSSmodel be

$K_{\mathrm{PSS}}=\frac{2\mathrm{\&xi;}{\mathrm{\&omega;}}_d{T}_J-D-D_e}{K_2{\left|G_{\mathrm{PSS}}\left(s\right)G_E\left(s\right)\right|}_{s={\mathrm{j\&omega;}}_d}}---\left(3\right)$

Wherein: || represent plural mould; Damping ratio ξ=0.1～0.3;

ω _dfor hunting angle frequency; T _jfor the inertia time constant of generating set;

D is the mechanical damping coefficient of generator amature;

D _ethe electrical damping moment coefficient of generator, computing formula is

$D_e=\frac{K_2{K}_5{K}_A{T}_{d0}^{\&prime;}{\mathrm{\&omega;}}_0}{{\left(K_6{K}_A\right)}^2+{\left(T_{d0}^{\&prime;}\mathrm{\&omega;}\right)}^2}---\left(4\right)$

K ₂, K ₅, K ₆it is the parameter of the Mathematical Modeling of low-frequency oscillation;

Power frequency angular velocity omega ₀=2 π f ₀; K _afor voltage regulator gain.

Beneficial effect of the present invention is, can effectively provide additional damping to suppress low-frequency oscillation, and convergence is good, highly versatile, practical value are high, and the present invention has general applicability, in the test of Yunnan unit PSS parameter tuning, is applied.

Accompanying drawing explanation

Fig. 1 is PSS model provided by the invention (PSS2A model);

Fig. 2 is low-frequency oscillation model provided by the invention;

Fig. 3 is voe #1 machine PSS parameter optimization of the present invention interface.

Embodiment

Below according to specification, accompanying drawing, and in conjunction with concrete embodiment, technical scheme of the present invention is further explained in detail.

Based on constrained variable metric PSS parameter optimization setting method, it is characterized in that, set up Optimized model, with constraint

Variable metric method solves PSS parameter, as the following step is:

Step 1: in the time that PSS parameter tuning is tested, obtain unit phase-frequency characteristic data, these data are an Excel file (for example: voe 1# machine .xls, data content is shown in Fig. 2);

Following table is unit phase-frequency characteristic test data provided by the invention:

Wherein: the 1st classifies frequency (Hz) as, the 2nd row lag angle (degree), "-" represents to lag behind.For example: frequency=2.01416Hz produces the lag angle of 141.789108 degree at excitation system, generator system.When phase-frequency characteristic test, measuring equipment is directly to generate Excel file.

Step 2: according to phase-frequency characteristic data, optimizing application model, adopts constrained variable metric method to calculate the time constant of PSS lead-lag link; Be formula (1):

$\min F({\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="HDA0000417501590000011.png,HDA0000417501590000012.png"\; state="\{\{state\}\}">T</figure-callout>}}_1,T_2,T_3,T_4,T_5,T_6)=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{({\mathrm{\&theta;}}_{\mathrm{PSS},i}-{\mathrm{\&theta;}}_{\mathrm{GE},i})}^2$

$s.t.\left\{\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="T"\; filenames="HDA0000417501590000011.png,HDA0000417501590000012.png"\; state="\{\{state\}\}">T</figure-callout>}}_1\&le;15T_2,T_3\&le;15T_4,T_5\&le;15T_6\\ T_{k,\min }\&le;T_k\&le;T_{k,\max },k=1,...,6\end{array}\right.$

θ _PSS,i=arctg(ω _iT ₁)+arctg(ω _iT ₃)+arctg(ω _iT ₅)-arctg(ω _iT ₂)-arctg(ω _iT ₄)-arctg(ω _iT ₆)

Wherein: θ _{pSS, i}for PSS model is in angular frequency _ithe offset angle at place, θ _{gE, i}for generator excited system phase-frequency characteristic is at ω _ithe phase angle at place, N is test data group number, T _kfor the lead-lag time constant of PSS model, T _{k, min}, T _{k, max}for minimum value and the maximum of PSS model lead-lag time constant regulation;

Application formula (1) is calculated the time constant T of PSS lead-lag link ₁～T ₆.Optimized model and constraint variable metric method solve T ₁～T ₆for a part of content of the present invention;

Step 3: by PSS model and parameter, unit parameter, system parameters numerical value, calculate PSS amplification coefficient; Foundation solves PSS amplification coefficient model, that is:

$K_{\mathrm{PSS}}=\frac{2\mathrm{\&xi;}{\mathrm{\&omega;}}_d{T}_J-D-D_e}{K_2{\left|G_{\mathrm{PSS}}\left(s\right)G_E\left(s\right)\right|}_{s={\mathrm{j\&omega;}}_d}}$

Wherein: || represent plural mould; Damping ratio ξ=0.1～0.3; ω _dfor hunting angle frequency; T _jfor the inertia time constant of generating set; D is the mechanical damping coefficient of generator amature; D _ethe electrical damping moment coefficient of generator, computing formula is

k ₂, K ₅, K ₆it is the parameter of the Mathematical Modeling of low-frequency oscillation; Power frequency angular velocity omega ₀=2 π f ₀; K _afor voltage regulator gain;

Calculate PSS amplification coefficient K _pSS.K _pSScomputational methods are the second parts of the present invention; The COEFFICIENT K of first calculating chart 3 ₁～K ₆, then according to formula (3) calculating K _pSS.

The PSS optimization of embodiment 1---voe 1# machine is adjusted

(1) phase-frequency characteristic of voe 1# machine test.Data as shown in Figure 2, are named as voe 1# machine .xls.

(2) application formula (1) Optimized model, adopts constrained variable metric method for solving, calculates the time constant T1～T6 of voe 1# machine PSS lead-lag link.Calculate:

T1=0.3501，T3=0.0584，T5=0.0930

T2=0.0377，T4=0.0050，T6=0.0062

(3) according to PSS model and parameter, unit parameter, system parameters, the amplification coefficient K of calculating parameter PSS _pSS.Through calculating, the K of voe 1# machine _pSScalculated value is:

K _PSS=12.49

The PSS optimization of embodiment 2---the 1# machine of making a pilgrimage to a temple on a famous mountain is greatly adjusted

(1) make a pilgrimage to a temple on a famous mountain the greatly phase-frequency characteristic test of 1# machine.As shown in the table, these data are named as the 1# machine .xls that makes a pilgrimage to a temple on a famous mountain greatly.

(2) application formula (1) Optimized model, adopt constrained variable metric method for solving, calculate the time constant T1～T6 of the PSS lead-lag link of the 1# machine of making a pilgrimage to a temple on a famous mountain greatly.Calculate:

T1=0.3113，T3=0.3105，T5=0.5005

T2=0.0208，T4=0.0207，T6=2.0000

(3) according to PSS parameter, unit parameter, system parameters, the amplification coefficient K of calculating parameter PSS _pSS.Through calculating, the K of the 1# machine of making a pilgrimage to a temple on a famous mountain greatly _pSScalculated value is:

K _PSS=14.30。

PSS2A model parameter T in Fig. 1 ₁～T ₆for the time constant of lead-lag link, it optimizes setting method is Part I content of the present invention; Parameter K _pSSfor the amplification coefficient of PSS, its setting method is the second part of the present invention.

Other parameter tuning Comparision is simple, and setting method is as follows:

(1) measurement links parameter

The time constant of measurement links is very little, general T _rW=T _rP=0.02～0.04 second, K _r=1.

(2) calculation of parameter of PSS isolation link

Adopt electromagnetic power signal and angular frequency signal, the isolated DC in PSS transfer function, the time constant of isolation high frequency filter device are generally:

T _1W=4 seconds, T _2W=4 seconds

T _p=4 seconds, T ₇=4 seconds

(3) calculation of parameter of PSS comparing element

K _s, T ₇, K forms comparing element, its parameter is generally set as:

T ₇=4 seconds, K _s=T ₇/ T _j(T _jfor unit set inertia time constant), K=1.

(4) the filtering link of PSS to shafting torsional oscillation

Parameter T ₈, T ₉, M, N selection to meet the angle that this link produces and be approximately-3600, desirable:

T ₈=0.1 second, M=5, N=1, T ₉=0.2 second.

(5) the amplitude limit link of PSS outlet

During for fear of large disturbance, PSS plays ill-effect, adds an amplitude limit link in PSS outlet, and amplitude limit value is: ± (0.05～0.1) pu.

In Fig. 2:

(1) K ₁～K ₆for the parameter of low-frequency oscillation model.Come according to the incremental computations of electromagnetic torque, excitation winding voltage, set end voltage formula, can consult in general teaching material data.

(2) Δ T _efor electromagnetic torque increment; Δ T _e1with Δ T _e2for Δ T _etwo components; Δ T _mfor machine torque increment; Δ δ is generator's power and angle increment; Δ ω is generator angular speed increment; Δ E ' _qfor transient potential; Δ E _ffor exciter output voltage; Δ U _reffor excitation system reference voltage increment; Δ U _pSSfor PSS output voltage increment.

Claims (1)

1. based on constrained variable metric PSS parameter optimization setting method, it is characterized in that, set up Optimized model, solve PSS parameter by constrained variable metric method, as the following step is:

Step 1: in the time that PSS parameter tuning is tested, obtain unit phase-frequency characteristic data, i.e. Excel file;

Step 2: according to phase-frequency characteristic data, optimizing application model, adopts constrained variable metric method to calculate the time constant of PSS lead-lag link; That is:

$\min F(T_1,T_2,T_3,T_4,T_5,T_6)=\underset{i=1}{\overset{N}{\mathrm{\&Sigma;}}}{({\mathrm{\&theta;}}_{\mathrm{PSS},i}-{\mathrm{\&theta;}}_{\mathrm{GE},i})}^2$

$s.t.\left\{\begin{array}{c}{T}_1\&le;15T_2,T_3\&le;15T_4,T_5\&le;15T_6\\ T_{k,\min }\&le;T_k\&le;T_{k,\max },k=1,...,6\end{array}\right.$

θ _PSS,i=arctg(ω _iT ₁)+arctg(ω _iT ₃)+arctg(ω _iT ₅)-arctg(ω _iT ₂)-arctg(ω _iT ₄)-arctg(ω _iT ₆)

Wherein: θ _{pSS, i}for PSS model is in angular frequency _ithe offset angle at place, θ _{gE, i}for generator excited system phase-frequency characteristic is at ω _ithe phase angle at place, N is test data group number, T _kfor the lead-lag time constant of PSS model, T _{k, min}, T _{k, max}for minimum value and the maximum of PSS model lead-lag time constant regulation;

Step 3: by PSS model and parameter, unit parameter, system parameters numerical value, calculate PSS amplification coefficient; Foundation solves PSS amplification coefficient model, that is:

$K_{\mathrm{PSS}}=\frac{2\mathrm{\&xi;}{\mathrm{\&omega;}}_d{T}_J-D-D_e}{K_2{\left|G_{\mathrm{PSS}}\left(s\right)G_E\left(s\right)\right|}_{s={\mathrm{j\&omega;}}_d}}$

Wherein: || represent plural mould; Damping ratio ξ=0.1～0.3; ω _dfor hunting angle frequency; T _jfor the inertia time constant of generating set; D is the mechanical damping coefficient of generator amature; D _ethe electrical damping moment coefficient of generator, computing formula is

; K ₂, K ₅, K ₆it is the parameter of the Mathematical Modeling of low-frequency oscillation; Power frequency angular velocity omega ₀=2 π f ₀; K _afor voltage regulator gain.

Images