Professional Documents
Culture Documents
05159763-IM Load
05159763-IM Load
org
ISSN 1751-8687
Abstract: Load modelling plays an important role in power system dynamic stability assessment. One of the
widely used methods in assessing load model impact on system dynamic response is parametric sensitivity
analysis. A composite load model-based load sensitivity analysis framework is proposed. It enables
comprehensive investigation into load modelling impacts on system stability considering the dynamic
interactions between load and system dynamics. The effect of the location of individual as well as patches of
composite loads in the vicinity on the sensitivity of the oscillatory modes is investigated. The impact of load
composition on the overall sensitivity of the load is also investigated.
690 IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690– 700
& The Institution of Engineering and Technology 2009 doi: 10.1049/iet-gtd.2008.0108
www.ietdl.org
reported in [11, 12]. It is established that the numerical composite static and dynamic load model gives the best
approach can compute all the sensitivity factors in a one-time simulation results in comparison with the field results. The
run whereas the analytical method has to compute this by static model of the IM is presented using a curve-fitting
each parameter; hence the former method is computationally technique in [17]. Recent advances in identification of IM
efficient for the large number of parameters in a big system. loads from power system composite load models are
The sensitivity analysis is used for ranking loads in a power presented in [18, 19]. A comprehensive discussion on
system. various issues of load modelling can be found in [20],
which focuses on the development of physically based load
The load that is highly sensitive with respect to its parameters models based on field test data and natural disturbances
should be modelled carefully in the system [13]. Load sensitivity data. The fitted model is simulated from the actual
provides a useful insight to proper system operation and control dynamic response of the motor and is expected to be quite
to prevent instability. Proper load sensitivity analysis provides realistic. In this paper, however, a detailed model for the
computational efficiency in measuring selected important IM is used to investigate the effect of the composite load
loads for sensitivity and stability analysis. So far most of the on the system stability.
load sensitivity works are mainly concerned with the generic
load model, which is essentially an exponential recovery The load model considered is the composition of a ZIP
model. However, modelling all the loads as static or dynamic load and an IM load as shown in Fig. 1 and is expressed as
loads may not truly reflect the load ranking. Moreover, it is
widely recognised that a power system can be modelled with Stot ¼ (1 a)SZIP þ aSind (1)
static and dynamic components. This is usually described as a
ZIP (constant impedance, constant current and constant where a is the percentage of IM load in the composite load
power) load and IM loads [14]. Different load structures may model; SZIP and Sind are the complex powers of the static
have different impacts on system dynamic characteristics. It is (ZIP) and IM loads, respectively
necessary to evaluate the combination of load, which is most
" 2 #
sensitive to system stability. This paper presents a V V o
comprehensive load ranking analysis framework, which can SZIP ¼ PZIP þ jQZIP ¼ AP þ BP þ CP PZIP
Vo Vo
handle both static and dynamic loads subject to different load
" #
structures. V 2 V
þ AQ þ BQ þ CQ QoZIP (2)
Vo Vo
The contributions of this paper are as follows: (i) effect of
the location of composite load on the overall sensitivity and
the eigenvalue movement of the system, (ii) finding the where AP, BP, CP, AQ , BQ and CQ are coefficients that
most and least sensitive parameters to the critical respectively represent the constant impedance, constant
eigenvalues of the system, (iii) effect of the parametric current and constant power portion of the real and reactive
variation on the mobility of the critical modes of the power of the ZIP load. For the purpose of study only the
system, (iv) effect of the increase in the number of IMs in constant power portion of the ZIP load is considered.
the system on the overall sensitivity of the load and (v)
effect of the variation of the composition of load on overall The IM model used is in compliance with the IEEE task
sensitivity. force [3]. The IM part of the load can be ‘small’ or ‘large’
motor depending on the specific load characteristic to be
The paper is structured as follows. Section 2 introduces modelled. As a general rule, the ‘small’ motor runs at a
composite load modelling with ZIP load and IM load. much higher slip than most ‘large’ motors above about
Section 3 describes system modelling and parameter 30 kW [21, 22]. A third-order model is usually adequate
sensitivity as well as the concept of overall sensitivity of the for aggregated motors in the bulk system dynamic simulation.
loads with respect to load parameters. The proposed
methodology is tested using the 10 machine 39 bus New Stator transients are neglected and electrical transients in
England system in Section 4. This is followed by conclusions one rotor per axis are represented. The transient-state
in Section 5.
IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690 – 700 691
doi: 10.1049/iet-gtd.2008.0108 & The Institution of Engineering and Technology 2009
www.ietdl.org
Xo ¼ Xs þ Xm
where the symbols are of the typical synchronous generators
0 Xr Xm and have the usual meaning as in the literature [23].
X ¼ Xs þ
Xr þ Xm (4)
Xr þ Xm The IEEE type-1 exciter [23, 24] is used in this paper –
To0 ¼ see Fig. 3.
vo Rr
Tm ¼ (am þ bm sm þ cm s2m )Tmo (5) The first step for the sensitivity and stability analysis is to
model the power system. The differential and algebraic
0 0
Te ¼ (Emx Imx þ Emy Imy )=vo (6) equation (DAE) form of system modelling is used in this
paper
0
Imy ¼ [Rs (Emx Vmx ) þ X 0 (Emy
0
Vmy )]=(Rs2 þ Xm2 )
Ẋ ¼ F(X , Y , u)
Imy ¼ [X 0 (Emx
0 0
Vmx ) Rs (Emy Vmy )]=(Rs2 þ Xm2 ) (9)
0 ¼ G(X , Y , u)
(7)
where Xs , Xm , Xr , Rr , Rs and H are stator impedance, where X is a vector of state variables, Y is a vector of algebraic
magnetising reactance, rotor impedance, rotor resistance, variables and system parameters, and u is the control input.
stator resistance and inertia of the IM, respectively [21]. The DAEs can be linearised and rearranged at operating
Vmx and Vmy are the real and imaginary axis terminal
voltages. am , bm and cm are the torque coefficients of the
torque slip characteristics.
692 IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690– 700
& The Institution of Engineering and Technology 2009 doi: 10.1049/iet-gtd.2008.0108
www.ietdl.org
IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690 – 700 693
doi: 10.1049/iet-gtd.2008.0108 & The Institution of Engineering and Technology 2009
www.ietdl.org
694 IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690– 700
& The Institution of Engineering and Technology 2009 doi: 10.1049/iet-gtd.2008.0108
www.ietdl.org
the least damped local mode and the inter-area mode of Again, the load at bus 3 is modelled as a composite load
oscillation is noticeable when the load at bus 29 is (50% IM þ 50% ZIP), keeping all the other loads as 100%
modelled as composite load. There is eigenmode shift in ZIP loads as in the above section. The change in the
both the modes. particular eigenmode with a small change in IM load
IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690 – 700 695
doi: 10.1049/iet-gtd.2008.0108 & The Institution of Engineering and Technology 2009
www.ietdl.org
Table 1 Parametric sensitivity of the critical eigenvalues for composite load at bus 3 (50% IM)
S. no. Critical eigenvalue Frequency Damp @li @l i @l i @l i @l i @l i X @l
i
(Hz) @Rs @Xs @Rr @Xr @Xm @H K
@Kj
j
Table 2 Parametric sensitivity of the critical eigenvalues for composite load at bus 4 (70% IM)
S. no. Critical eigenvalue Frequency Damp @li @l i @l i @l i @l i @l i X @l
i
(Hz) @Rs @Xs @Rr @Xr @Xm @H K
@Kj
j
1 20.1587 + j6.9999 1.11 0.022 4096 4276 2302 4010 208 0 35 610
2 20.2222 + j7.0131 1.11 0.031 1420 1472 8256 1382 74 0 12 604
3 20.2426 + j6.8956 1.09 0.035 854 86 4792 808 44 0 7358
4 20.3154 + j8.4851 1.35 0.037 168 146 662 136 6 0 1118
5 20.2897 + j7.4850 1.19 0.038 0.558 0.448 1.467 0.416 0.014 0 2.905
6 20.1105 + j2.7362 0.43 0.040 323 2988 17 704 2818 164 4 26 904
7 20.3678 + j8.7524 1.39 0.042 4 2 42 2 0 0 50
8 20.2785 + j5.9814 0.95 0.045 18 58 1618 6 14 0 1768
9 20.4202 + j8.7011 1.38 0.048 556 59 4736 562 42 0 6488
parameter is calculated using (17-18) for each parameter. Xm , Xr , Rr , Rs and H ) of the IM are changed one by one.
The result is presented in Table 1. It is found that Rr is Every time, only one parameter is increased while keeping
the most sensitive parameter as (@li =@Rr ) is maximum for the rest constant. Each load parameter is increased by 2%
all the critical modes whereas H is the least. A similar for up to three steps and the eigenmodes are plotted. The
pattern is observed when the loads at other buses are effect of increasing each IM load parameter on the mobility
modelled as composite loads. of the eigenmode is depicted in Fig. 7. Increasing the
parameters changes the damping of the least damped local
mode. For example, if parameter Xr is changed to
4.3 Parametric changes and critical ‘1.06 Xr’, the eigenvalue would change from
20.1655 þ j6.9755 (top right corner) to 20.1675 þ j6.967
modes mobility (bottom left corner) as shown in Fig. 7. A similar effect is
The effect of the increase in a particular parameter of the IM observed in the inter-area mode of oscillation, which is
load is observed in the mobility of the eigenmode. This is shown in Fig. 8. The effect on the eigenmode movement is
achieved by modelling the load at bus 3 as a composite significant because of the increase in parameters Xs and Xr ,
load (10% IM þ 90% ZIP) and then the parameters (Xs , which are stator and rotor reactances. However, it is
696 IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690– 700
& The Institution of Engineering and Technology 2009 doi: 10.1049/iet-gtd.2008.0108
www.ietdl.org
IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690 – 700 697
doi: 10.1049/iet-gtd.2008.0108 & The Institution of Engineering and Technology 2009
www.ietdl.org
Table 3 OS of the composite loads with the variable number of IMs modelled in the system
Load bus Number of IMs in the system S.D.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
3 0.76 0.11 0.61 0.35 1.52 0.27 1.05 0.58 2.02 0.30 0.72 0.32 1.75 0.17 0.92 1.25 1.95 0.631
4 0 1.06 0.42 4.30 2.88 2.05 2.08 6.33 2.28 3.55 1.16 3.93 2.66 1.97 3.14 9.32 4.54 2.271
7 0 0 0.05 0.14 0.99 0.22 1.08 0.19 0.68 0.19 0.26 0.18 0.80 0.68 0.49 0.51 0.73 0.348
8 0 0 0 3.99 3.95 2.67 2.49 6.07 2.33 3.61 2.61 3.66 10.32 7.11 2.76 9.66 6.28 3.059
12 0 0 0 0 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.001
15 0 0 0 0 0 0.39 0.98 2.33 1.25 1.76 1.72 3.06 2.57 1.99 1.81 4.10 1.70 1.228
16 0 0 0 0 0 0 1.07 0.56 1.53 0.96 1.72 0.74 2.20 0.94 0.95 0.82 1.34 0.692
18 0 0 0 0 0 0 0 0.57 0.45 0.56 0.13 0.58 0.76 0.56 0.27 0.74 0.31 0.293
20 0 0 0 0 0 0 0 0 17.33 1.32 2.47 6.00 15.68 3.07 3.74 3.62 6.96 5.373
21 0 0 0 0 0 0 0 0 0 0.71 0.56 2.13 1.55 1.29 1.06 2.52 1.36 0.844
23 0 0 0 0 0 0 0 0 0 0 0.26 0.21 1.62 0.21 0.76 0.10 0.87 0.446
24 0 0 0 0 0 0 0 0 0 0 0 2.03 1.00 1.38 0.38 1.71 1.30 0.716
25 0 0 0 0 0 0 0 0 0 0 0 0 0.60 0.09 0.32 0.10 0.78 0.234
26 0 0 0 0 0 0 0 0 0 0 0 0 0 0.12 0.08 0.38 0.56 0.159
27 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.50 0.06 1.93 0.476
28 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.25 1.08 0.265
29 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.00 1.57
Figure 9 S.D. of the motor loads with a variable number of Figure 10 OS of the load with 16 and 17 motors in the
IMs in the system system
698 IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690– 700
& The Institution of Engineering and Technology 2009 doi: 10.1049/iet-gtd.2008.0108
www.ietdl.org
6 Acknowledgment
The first author is supported by a UQIRS scholarship.
7 References
[1] HILL D.J., HISKENS I.A.: ‘Dynamic analysis of voltage
collapse in power systems’. Proc. 31st IEEE Conf. Decision
and Control, 1992, vol. 3, pp. 2904 – 2909
Figure 11 OS of the load with 11 and 12 motors in the
system at bus 23 and bus 24
[2] KARLSSON D., HILL D.J.: ‘Modeling and identification of
nonlinear dynamic loads in power system’, IEEE Trans.
Power Syst., 1994, 9, (1), pp. 157 – 166
[4] PEREIRA L., KOSTEREV D., MACKIN P., ET AL .: ‘An interim dynamic
induction motor model for stability studies in the WSCC’, IEEE
Trans. Power Syst., 2002, 17, (4), pp. 1108 – 1115
[5] MA J., DONG Z.Y., HE R., HILL D.J.: ‘System energy analysis
incorporating comprehensive load characteristics’, IET
Figure 12 Overall sensitivity of the load with a different Gener. Transm. & Distrib, 2007, 1, (6), pp. 855 – 863
percentage of dynamic load component
[6] WANG K.W., CHUNG C.Y. , TSE C.T., TSANG K.M.: ‘Improved
probabilistic method for power system dynamic stability
5 Conclusion studies’, IEE Proc. Gener. Transm. Distrib., 2000, 147, (1),
This paper restates the importance of composite load pp. 37– 43
modelling, especially dynamic load models with IMs, in
power system stability assessment. Misrepresentation of the [7] WANG K.W. , TSE C.T., BIAN X.Y., DAVID A.K. : ‘Probabilistic
actual load could lead to quite different stability assessment eigenvalue sensitivity analysis and PSS design in
results. Based on detailed system dynamic analysis, a multimachine systems’, IEEE Trans. Power Syst., 2003, 18,
comprehensive study is performed on load sensitivity (1), pp. 1439 – 1445
considering IM loads in a composite load model. Various
scenarios are observed to understand the effect of [8] ANDERS G.J. : ‘Probability concepts in electric power
composite load modelling on the system small-signal systems’ (Wiley, 1990)
stability characteristics. The main findings of this paper and
their implication in power system small-signal stability [9] HOCKENBERRY J., LESIEUTRE B.: ‘Evaluation of uncertainty in
analysis are also given following the detailed comprehensive dynamic simulation of power system models: the
case studies. Using the presented approach, the most probabilistic collocation method’, IEEE Trans. Power Syst.,
sensitive composite load location can be found for further 2000, 15, (1), pp. 299– 306
stability analysis. Significant parameters of the IM load are
identified based on sensitivity analysis. It is also observed [10] LIMA E.E.S., FERNANDES L.F.J. : ‘Assessing eigenvalue
that the vicinity of motors has an increasing effect on the sensitivities’, IEEE Trans. Power Syst., 2000, 15, (1),
overall sensitivity, and the increasing dynamic component pp. 299– 306
IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690 – 700 699
doi: 10.1049/iet-gtd.2008.0108 & The Institution of Engineering and Technology 2009
www.ietdl.org
[11] DONG Z.Y., PANG C.K., ZHANG P.: ‘Power system sensitivity [19] MA J., HE R., HILL D.J.: ‘Load modeling by finding support
analysis for probabilistic small signal stability assessment vectors of load data from field measurements’, IEEE
in a deregulated environment’, Int. J. Control Autom. Trans. Power Syst., 2006, 21, (2), pp. 726– 735
Syst., 2005, 3, (2), pp. 355 – 362
[20] Advanced load modelling, EPRI, Palo Alto, CA and
[12] MAKAROV Y.V., DONG Z.Y.: ‘Eigenvalues and eigenfunctions, Public Service Company of New Mexico, Albuquerque,
Vol. Computational science & engineering, encyclopedia NM, 2002, 1007318
of electrical and electronics engineering’ (Wiley, 1998)
[21] TAYLOR C.W.: ‘Power system voltage stability’ (McGraw-
[13] MILANOVIC J.V. , HISKENS I.A., MASLENNIKOV V.A.: ‘Ranking Hill, 1994)
loads in power systems – comparison of different
approaches’, IEEE Trans. Power Syst., 1999, 14, (2), [22] MISHRA Y., DONG Z.Y., MAJUMDER R., ET AL .: ‘Investigation of
pp. 614– 619 sensitivity based load ranking with composite load ranking
in power system’. Proc. Eighth Int. Power Engineering Conf.
[14] LI S.H., CHIANG H.D. , LIU S.: ‘Analysis of composite load (IPEC 2007), December 2007, pp. 572 – 577
models on load margin of voltage stability’. Int. Conf.
Power System, 2006, pp. 1 – 7 [23] KUNDUR P.: ‘Power system stability & control’ (McGraw-
Hill, 1994)
[15] UNDRILL J.M., LASKOWSKI T.F.: ‘Model selection and data
assembly for power system simulation’, IEEE Trans. Power [24] SAUER P.W. , PAI M.A.: ‘Power system dynamics and
Appl. Syst., 1982, PAS- 01, pp. 3333– 3341 stability’ (Pearson Education, 2003)
[16] KAO W.A.: ‘The effect of load models on unstable low- [25] DONG Z.Y., HILL D.J. , GUO Y. : ‘A power system
frequency oscillation damping in tai power system control scheme based on security visualisation in
experience w/wo power system stabilizers’, IEEE Trans. parameter space’, Int. J. Electr. Power Energy, 2005, 27, (7),
Power Syst., 2001, 14, pp. 463 – 473 pp. 488– 495
[17] MORISON K., HAMADANI H., WANG L.: ‘Practical issues of load [26] HILL D.J., DONG Z.Y.: ‘Nonlinear computation and control
modeling for voltage stability studies’. Proc. IEEE PES for small disturbance stability’. Invited Paper, IEEE/PES SM
General Meeting 2003, Toronto, Canada, July 2003 2000 Panel Session on Recent Applications of Linear
Analysis Techniques, 2000
[18] MA J., HAN D., HE R.M., DONG Z.Y., HILL D.J.: ‘Research on
identifiability of equivalent motor in composite load [27] MA J., HAN D., HE R.M., DONG Z.Y., HILL D.J.: ‘Reducing identified
model’. Proc. Power Tech 2007, Lausanne, Switzerland, parameters of measurement based composite load model’,
1 – 5 July 2007 IEEE Trans. Power Syst., 2008, 23, (1), pp. 76– 83
700 IET Gener. Transm. Distrib., 2009, Vol. 3, Iss. 7, pp. 690– 700
& The Institution of Engineering and Technology 2009 doi: 10.1049/iet-gtd.2008.0108