This action might not be possible to undo. Are you sure you want to continue?

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

Contents lists available at ScienceDirect

**Energy Conversion and Management
**

journal homepage: www.elsevier.com/locate/enconman

**Small disturbance voltage stability assessment of power systems by modal analysis and dynamic simulation
**

Nima Amjady *, Mohammad Reza Ansari

Department of Electrical Engineering, Semnan University, Semnan, Iran

a r t i c l e

i n f o

a b s t r a c t

The introduction of liberalized electricity markets in many countries has resulted in more highly stressed power systems. On the other hand, operating points of a power system are acceptable in the feasible region, which is surrounded by the borders of different stabilities. Power system instability is critical for all participants of the electricity market. Determination of different stability margins can result in the optimum utilization of power system with minimum risk. This paper focuses on the small disturbance voltage stability, which is an important subset of the power system global stability. This kind of voltage stability is usually evaluated by static analysis tools such as continuation power ﬂow, while it essentially has dynamic nature. Besides, a combination of linear and nonlinear analysis tools is required to correctly analyze it. In this paper, a hybrid evaluation method composed of static, dynamic, linear, and nonlinear analysis tools is proposed for this purpose. Effect of load scenario, generation pattern, branch and generator contingency on the small disturbance voltage stability are evaluated by the hybrid method. The test results are given for New England and IEEE68 bus test systems. Ó 2008 Elsevier Ltd. All rights reserved.

Article history: Received 9 September 2007 Accepted 13 April 2008 Available online 5 June 2008 Keywords: Voltage stability Voltage collapse Dynamic simulation Eigenvalue analysis

1. Introduction Nowadays, in many countries, the introduction of competitive supply and corresponding organizational separation of supply, transmission, and system operation has resulted in more highly stressed and unpredictable operating conditions and more vulnerable networks [1]. These conditions, brought on by natural load growth with a signiﬁcant increase in long-distance transmission usage, often result in heavy transmission circuit loadings, depressed bus voltage magnitudes, and closer proximity to voltage instability. Many power system blackouts all over the world have been reported where the reason for the blackout has been voltage instability [2,3]. Power system stability is an essential requirement for all participants of an electricity market. So, voltage stability can appear as a main limiting factor for transmission system loading. Thus, it is required to have a clear understanding of the problem. Power system stability is the ability of an electric power system, for a given initial operating condition, to regain a state of operating equilibrium after being subjected to a physical disturbance, with most system variables bounded so that practically the entire system remains intact [4]. Analysis of stability, including identifying key factors that contribute to instability and devising methods of improving stable operation, is greatly facilitated by classiﬁcation of stability into appropriate categories. Power system stability

* Corresponding author. Tel.: +98 021 88797174; fax: +98 021 88880098. E-mail address: amjady@tavanir.org.ir (N. Amjady). 0196-8904/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.enconman.2008.04.010

can be categorized to rotor angle, frequency and voltage stabilities [4]. Voltage stability refers to the ability of a power system to maintain steady voltages at all buses in the system after being subjected to a disturbance from a given initial operating condition [4]. Voltage instability stems from the attempt of load dynamics to restore power consumption beyond the capability of the combined transmission and generation system [5]. A possible outcome of voltage instability is loss of load in an area, or tripping of transmission lines and other elements by their protective systems leading to cascading outages. Voltage collapse is the catastrophic result of a sequence of events leading to a low voltage proﬁle suddenly in a major part of the power system [2]. Voltage stability can be further classiﬁed to small disturbance, large disturbance, short term, and long term categories [2]. Small disturbance voltage stability considers the power system’s ability to control voltages after small disturbances, e.g. changes in load [6]. Large-disturbance voltage stability refers to the system’s ability to maintain steady voltages following large disturbances such as system faults, loss of generation, or circuit contingencies. Besides, the time frame of interest for voltage stability problems may vary from a few seconds to tens of minutes. Therefore, voltage stability may be either a short-term or a long-term phenomenon. Short term voltage stability is characterized by components such as induction motors, excitation of synchronous generators, and electronically controlled devices such as HVDC and static VAR compensator [5]. The study period of interest is in the order of several seconds, and analysis requires solution of appropriate

Fig. and time delays) [4]. Considering the above explanations. dynamic. we found more nonlinear behaviors of the stability than those mentioned in the previous works.2630 N. when the voltage instability occurs in the transient period of disturbance. In the case of saddle-node bifurcations. Response of the system to this perturbation is found by the dynamic simulation. generator reactive power limits) [11]. the stability of the power system cannot be fully guaranteed with steady state studies. Then. Besides. For instance. Effect of load scenario. generation scenario. the reader is referred to the web version of this article.7]. Amjady. power system has EPs obtained by the steady state analysis. which is typically correlated with the lack of power ﬂow solutions. Therefore. discrete tap steps. cannot account for nonlinear effects such as tap changer controls (deadbands. how the system voltages will respond to small system changes. The long term dynamics such as response of power plant controls.g. However. emergence of oscillatory instability or Hopf bifurcation in response to the perturbation is detected. Especially. our analysis method proceeds by one step ahead. In Section 3 or the main section of the paper. which can determine actual dynamic status of small disturbance voltage stability. On the other hand. Hopf bifurcation [11]. SLM. controllers and loads. The dynamics of the long term time scale last for several minutes. A brief review of the paper and the future research are in Section 4. generator contingency and branch contingency on the status of this stability is evaluated and its illcondition and well-condition behaviors are introduced. Steady state voltage stability studies investigate long term voltage stability [2]. With appropriate assumptions.3. SLM is the maximum loading level beyond which steady state solutions cannot be obtained for the system [12]. The long term voltage stability is characterized by scenarios such as load recovery by the action of on-load tap changer or through load self restoration. in this work. However. 68-bus test system. Between DLM and SLM on the green curve. detailed numerical results for New England and IEEE68 test systems are presented and discussed. 68-bus test system is shown in Fig. 1. Important conclusions based on the results for this kind of stability are derived. In steady state voltage stability studies based on the load ﬂow equations it is assumed that all dynamics are died out and all controllers have done their duty. i. is at the end of green curve. the proposed method.e. At each point of the continuation curve (e. 1. system equations can be linearized for analysis thereby allowing computation of valuable sensitivity information useful in identifying factors inﬂuencing stability.g. If in a point of the continuation curve. This procedure is successively repeated for the points of the continuation curve until both dynamic load margin (DLM) and static load margin (SLM) are determined. and is quite often associated with the voltage collapse in power systems.R. previous works proposed static analysis tools (like power ﬂow based methods) to solve small disturbance voltage stability [2. at ﬁrst. delayed corrective control actions such as shunt compensation switching or load shedding. 2. However. This paper focuses on the small disturbance voltage stability. In dynamic analysis the transition itself is of interest and it checks that the transition will lead to an acceptable operating condition. characteristics of small disturbance voltage stability problem are described. linear and nonlinear analysis tools. we consider a combination of static and dynamic analysis tools in our evaluation. a combination of linear and nonlinear analyzes is used in a complementary manner. a singularity of a system Jacobian and/or state matrix results in the disappearance of such static solutions. The remaining parts of the paper are organized as follows. however. Then a small perturbation in the form of load increment is applied to the equilibrium point (EP) of the power system. each of these 1 For interpretation of color in Fig. i. Then. PV curve). a combination of static and dynamic analyzes based on the voltage continuation curve is proposed for small disturbance voltage stability evaluation. the lack of steady state solutions are due to system controls reaching limits (e. the steady state conditions of the power system are analyzed by the load ﬂow equations.8–10]. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 system differential equations.7]. In Section 2. this kind of stability is also inﬂuenced by the dynamic characteristics of the system such as those of generators. which is a major concern of today power systems [2. whereas in the case of limit-induced bifurcations.e. M. this is similar to analysis of rotor angle stability [4]. The main contribution of this paper is analysis of various aspects of small disturbance voltage stability. Saddlenode and limit-induced bifurcations basically consist of loss of system equilibrium. This concept is useful in determining. at any instant. but the use of steady state analysis tools can provide useful information about it. Especially. Besides. 1. The advantage of using algebraic equations compared to differential equations of dynamic studies is the computation speed. indicating DLM. boiler dynamics and automatic generation control also affect long term voltage stability [5]. voltage stability is dynamically lost at that point (DLM). In this ﬁgure Hopf bifurcation point (at the end of blue1 curve). composed of static. the voltage stability margin obtained by the dynamic method can be considerably less than that of the static method. The small disturbance voltage stability is a dynamic phenomenon by nature. is shown. an execution of the proposed method for IEEE 16machine. So. This approach provides the most accurate response of the actual dynamics of voltage instability when appropriate modeling is included [2]. Even we adopted this procedure in our some previous works [3. Sample execution of the proposed method in the form of PV curve for the IEEE 16-machine. the post disturbance equilibrium point of the power system (with incremented load) is determined by the load ﬂow equations. The proposed method Limitations of steady state power system studies (algebraic equations) are associated with stability of nonlinear dynamic systems. . The time domain simulations capture the events and chronology leading to voltage instability. Saddle node bifurcation for constant power loads usually appears on the nose of the PV curves. drivers of small disturbance voltage stability and their sensitivities with respect to various system variables are presented. which is another contribution of this paper. to evaluate this problem is introduced. This linearization. associated with saddle node bifurcation.

voltage setting points. this transition of eigenvalues does not impact the stability properties of a power system.y. we know that the inverse of gy will become inﬁnity. which determine the base case conditions or initial point of the continuation curve. Besides. the system will lose its stability [5. which is called ‘‘singularity induced inﬁnity”. but is usually a prelude to other bifurcations. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 2631 EPs will be lost by a small perturbation. which indicates DLM. there are three different kinds of bifurcation points. Amjady. Assuming gy is nonsingular. If no solution can be found. indicating where the parameter p deﬁnes speciﬁc system conﬁgurations and operation conditions. x 2 X Rn . the above model can be represented as " # _ fx fy Dx Dx . In obtaining this solution. (7) Determine post-disturbance EP (with changed load and generation patterns) as described in step 2. which is deﬁned by a scheduled system operating strategy. as discussed in the previous paragraphs). Bifurcation theory assumes that power system parameters vary slowly and predicts how a power system becomes unstable.N. the continuation curve. is changed at once (which is also the case considered here). The dynamic state variables x (slow modes). such as exciter control systems. Through tracing the eigenvalues of matrix Jr. two complex conjugate eigenvalues of reduced Jacobian cross the imaginary axis. load demand. where it is not easy to compute and analyze the stability of the system. The instantaneous variables y (fast modes) satisﬁes algebraic constraints (2). saddle node or limit induced bifurcation. For a structural stability problem. a better viewpoint for this matter can be presented based on the eigenvalue analysis. At singularity induced bifurcation (SIB) point. EPs before Hopf bifurcation are stable equilibrium points (SEP).14]. There are two steps involved to identify the dynamic stability of the power system as the parameter p slowly changes. an EP is a solution of the system: ( fðx. controller limits (such as reactive power limits of the generators) and on-load tap changers (OLTC) are considered.16]. are the most interesting points in voltage stability studies. (5) Determine dynamic response of the power system to the step change by time domain simulation. Through (6). from stationary to oscillatory. At this point the reduced Jacobian has a zero eigenvalue. When one EP passes through the boundary. Evaluate occurrence of HB or SNB based on the behavior of eigenvalues. e. representing unstable equilibrium point (UEP). However.17]. form reduced Jacobian matrix and analyze the eigenvalues at each EP [16. the other components of Ju are partial Jacobians.17]. (3) Form reduced Jacobian at the EP and calculate its eigenvalues.g. describe the generation dynamics of power systems. 0 ¼ g ðx. Similarly. etc. (4) Change the load and generation patterns of the power system by one step (apply the small perturbation) according to the scheduled system operating strategy or the predeﬁned security assessment scenario. In [5] it has been mentioned that usually one parameter. More details about the above DAE model can be found in [1. gy becomes singular. Saddle-node bifurcation (SNB) point is where two EPs coalesce and then disappear. p 2 P Rq ð1Þ ð2Þ ð3Þ an EP of the DAE model for a given p depends on the eigenvalues of the reduced Jacobian matrix Jr [17]. The stability of . Ju ¼ ¼ Ju ð5Þ gx gy Dy 0 where Ju is called the unreduced Jacobian. For a better illustration of the performance of the method we begin with the dynamic model of the voltage stability problem. augmented Jacobian. f : Rnþmþq ! Rn g : Rnþmþq ! Rm y 2 Y Rm .g. or augmented system state matrix [14–16]. Now the proposed analysis method can be summarized as the following step by step algorithm: (1) Deﬁne the set of parameters p including power system conﬁguration. such as power ﬂow equations.pÞ ¼ 0 _ ¼0) x ð4Þ g ðx. At this point. load pattern (consist of load level and distribution). we use from bifurcation theory to study the actual behavior of the system. changing the load and generation patterns may be simply in the form of linearly increasing consumption of a load bus and accordingly increasing the generation of slack bus. For ﬁxed values of p parameters. (5) can be reduced to ordinary state space equations by eliminating Dy: _ Dx ¼ ½fx À fy g À1 g x Dx ¼ J r Dx y ð6Þ where Jr is called reduced Jacobian or reduced system matrix.e. First. solve and trace the EP along the path (i. or from order to chaos. These three bifurcation sets usually are the boundary of the feasible region of the DAE model (1) and (2) [13.pÞ . (6) If emergence of oscillatory instability or Hopf bifurcation in the dynamic response is detected. the partial derivative of vector f with respect to vector x.y. Dynamic models of the power system components considered in our time domain simulation are described in Appendix and more details can be found in [19]. which can be presented in the form of parameter dependent differential algebraic equations (DAE) [14]: _ x ¼ fðx. The change of parameters moves the system slowly from one equilibrium to another until it reaches the collapse point [5].pÞ . Then.16]. which is implicitly assumed to have an instantaneously converging transient. by means of linearization. more signiﬁcant ones described above. Hopf bifurcation determined by the eigenvalue analysis (step 3) and time domain simulation should be the same so that we can judge about its occurrence. Fig. Matrix fx is the partial Jacobian. we can study the local dynamic stability of the power system [6. We discussed about the bifurcation points from the viewpoint of static and dynamic analyzes. The obtained Jacobian matrix describes the linear system which best approximates the nonlinear equations close to the equilibrium.pÞ ¼ 0 To determine the stability margin of the EP. M. valuable sensitivity information useful in identifying factors inﬂuencing small disturbance voltage stability can be obtained by linearization of the system at the equilibrium point [4]. (2) Determine EP by the load ﬂow solution. e. voltage stability is dynamically lost at that point. Bifurcation points where change from stable to unstable occurs.y. Now. However. There are also other types of bifurcations in the DAE models such as focus node transition where two complex eigenvalues coalesce and change to real ones [18]. generation. For instance. Hopf bifurcation (HB) point is where there is an emergence of oscillatory instability. 1. In that case the stability of the nonlinear system can be studied like the stability of linear systems in the neighborhood of operating equilibrium [2]. Voltage stability is a nonlinear phenomenon. in which case there is a possibility to achieve either saddle node or Hopf bifurcation. such as loads. respectively [18]. deﬁned by (1). The proposed analysis method at each equilibrium point along the continuation curve calculates eigenvalues of the Jacobian matrix and evaluates possibility of occurrence of different bifurcations. and generation pattern. Bifurcation analysis deals with the problem of loss of stability of a nonlinear dynamic system under changing parameter values [13].R. gy is called algebraic Jacobian or static Jacobian [14–16].y.

M. sample results for the other load models. New England and IEEE68 test systems are presented. 3. different load models have been evaluated and it is concluded that from the voltage stability viewpoint. in the unreduced Jacobian matrix Ju. and the stability of the system will change [16]. 2 and 3. At SIB. respectively.g. 1). We evaluate effect of load and generation scenarios. New England and IEEE68 test systems. Single line diagram of the New England test system with 10 machines and 39 buses and IEEE68 test system with 16 machines and 68 bus are shown in Figs. SIB occurs in a DAE system when the equilibrium point is placed at a singularity manifold of the algebraic subsystem. Amjady. . Single line diagram of the New England test system.2632 N.22]. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 SLM. determinant of Jr also goes to inﬁnity. has been occurred.g. Single line diagram of IEEE68 bus test system. As seen.e. 3. IEEE30. at least one eigenvalue of reduced Jacobian matrix Jr will change from negative inﬁnity to positive inﬁnity.20]. the constant power load usually create the most stress in the system. where their data can be found in [21.1. Dynamic models of these test systems are as described in Appendix. Fig. At the SIB point. Fig. Similar results have been obtained for the other ones. frequently considered in the voltage stability studies [9–12.p) = 0 is divided into the voltage causal regions where gy is not singular separated by the singularity manifold at which gy is singular [18]. So. However. (8) The procedure described from step 2 to step 7 is repeated until both DLM and SLM be found (e. i.y. are also presented for comparison. In [12]. respectively. group of buses and all buses. Sample Fig.R. For the sake of conciseness only the results of the larger ones. the above algorithm mainly focus on the HB and SNB. At singular point. e. constant current and constant impedance. The constraint manifold g(x. 2. 3.18. this kind of load model is considered for the load buses in most examinations of this paper to better illustrate the bifurcation points (constant power factor is also assumed). Numerical results We examined the proposed method (the step by step algorithm) on the IEEE14. only one eigenvalue changes the sign. and generator and branch contingencies on different aspects of small disturbance voltage stability by the proposed method. Load scenario We examined many load scenarios including increasing load of a single bus. Occurrence of saddle node bifurcation should be also detected by the eigenvalue analysis.

The critical eigenvalues are the same in Figs. 4. in the range [À1. However. 5. The nonlinear behavior of the dominant eigenvalues of the algebraic Jacobian gy with respect to the load variation has been discussed in the previous works [1. In Fig. DLM and SLM of the power system are sensitive to the load scenario. and all buses are increased. 4. where the dominant eigenvalues in the current operating point may be different from the critical ones. Critical eigenvalues are those which ﬁnally cross the imaginary axis and cause HB. In Figs. By increasing load(18) or L. i. PV curve for two different load scenarios are shown in Fig. Dominant or critical eigenvalue(s) can be a single real one or a pair of complex conjugate eigenvalues.0]. 4. another aspect of nonlinearity can be seen in the behavior of critical eigenvalues. the critical eigenvalues (indicated by the arrows) appear after three other ones. but the velocity by which these eigenvalues approach to the imaginary axis are dependent on the load scenario. Here. A similar situation is seen in Fig. Indeed the same critical eigenvalues were observed in all various examined load scenarios. which are the best candidate(s) for HB in that operating point.R. some interesting results can be also seen. are shown. where load of a single bus 18. cross the imaginary axis and cause HB. 4. For instance. 4 and 5. By further increasing load(18). in the ﬁrst two parts of Fig. It is noted that from the dominant eigenvalue(s) in the current operating point. locus of eigenvalues of the reduced Jacobian Jr close to imaginary axis. Eigenvalue locus for increasing load of bus 18 in the New England test system. 4 and 5 owning completely different load scenarios. . the critical eigenvalues proceed the ﬁrst one. 6. indicated by load(18). Amjady. 4 and 5. Fig. 5. indicated in the ﬁfth and sixth parts of Fig. M. instead of monotonically approaching the imaginary axis.23]. a similar situation is seen for the reduced Jacobian Jr where the closest eigenvalues to the imaginary axis are changed by the load variation. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 2633 results for the New England test system are shown in Figs.e. Although the critical eigenvalues are ﬁxed with respect to load scenario. where load of a single bus 8 and a single bus 23 has been separately increased. they goes farther at ﬁrst and then rapidly return and cross the imaginary axis. By increasing load(18). the critical eigenvalues appear in the second place after only one eigenvalue in the third and fourth parts of Fig. respectively. For a better illustration of this matter. respectively. L is the load growth factor of all buses.N. In spite of these nonlinear behaviors. resulting in different DLMs and SLMs. we mean the closest one(s) to the imaginary axis in the complex plane. Besides.

The participation factors represent contribution of each state of the power system in the critical modes [15. HB point Factor 1 0. For instance.717 0. respectively.603 0.518 0.R. 4 and 5 (HB point) Load(18) = 1540 + 308i.473 0.559 0. participation factors for examinations of Figs.24]. PV curve for two load scenarios in the New England test system. 6.467 State K+3 K+3 K+3 K K+3 K K K+2 K K+2 Model ESDC1A ESDC1A ESDC1A GENROU ESDC1A GENROU GENROU GENROU GENROU GENROU Bus 33 36 37 36 32 37 33 33 32 36 L = 1. Eigenvalue locus for increasing load of all buses in the New England test system. Amjady.475 0. M.806 0.347 State K+3 K+3 K K+3 K K+3 K+2 K K K+3 Model ESDC1A ESDC1A GENROU ESDC1A GENROU ESDC1A GENROU GENROU GENROU ESDC1A Bus 33 37 33 32 37 36 33 36 32 38 Fig. 4 and 5 are shown in columns 1 and 5. 10 states with the largest participation factors (normalized with respect to their maximum) at HB point are sorted.627 0. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 Fig. HB point Factor 1 0. In Table 1.678 0.25233.2634 N. Parameters of this table are described in Appendix.608 0.636 0.917 0. state K + 3 of the model ESDC1A of bus 33 (which according to Table 7 of Appendix is exciter output EFD) has the largest participation factor in the critical modes .559 0.556 0. 5.757 0. Table 1 Participation factors for load scenarios of Figs.717 0. In this Table.

439 K+2 0. we also changed the framework. However. 7. the new framework can be also seen in practice. As seen. In supplying reactive load in the load ﬂow solution. Although in this examination. To obtain dynamic response of the system to the small perturbation (step 5 of the step by step algorithm) all generators participate according to their dynamic characteristics [19]. In spite of Figs. 8 and 9 common states are seen. Iran’s power network feeds large iron foundries owning Fig. when dynamic response to a step change approximately reaches to EP.N. In the new framework. 4 and 5. M. where results of Table 1 are presented for the initial point of the load scenarios. load level and load scenario (direction of loading) can change participation factors in the critical eigenvalues. Another interesting point can be seen from Table 2. So. State variables owning the highest participation factors in the critical eigenvalues at HB point are shown in Table 3. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 2635 in the both load scenarios. but critical eigenvalues are the same. respectively. 3. these eigenvalues show a different behavior here. In the previous works. DLM and SLM also change with variation of generation scenario as shown in Fig.301 K+3 0. 7. 7.7–12].190 K+3 0. Amjady. In other words. . in spite of different load scenarios considered.627 K 0. by increasing load in Fig.R. although both load and generation scenarios are changed. where the PV curve is constructed by tracing EPs of the successive dynamic responses.142 K+3 0. 7 and 8 out of 10 states of Tables 1 and 2 are the same for the ﬁrst and second load scenarios. It is seen that the set of the most participating states in the critical modes have low sensitivity with respect to both load and generation scenarios. So. For instance. all generators participate based on their AVR (automatic voltage regulator) setting and reactive power limits. 8. but with different ranks and participation factors (especially the ﬁrst one is the same). a set of step load increases are applied to the initial EP and response of the power system to the step changes is dynamically found by the time domain simulation without midway load ﬂow solutions. In this framework. the critical eigenvalues monotonically approaches the imaginary axis. initial point 1 K+3 0. generation of the slack bus (bus 39 in the New England test system) is increased accordingly to supply the required active load to obtain EP of the Table 2 Participation factors for the initial point of the load scenarios Factor State Model ESDC1A GENROU GENROU ESDC1A GENROU ESDC1A GENROU ESDC1A GENROU GENROU Bus 33 33 33 37 37 32 32 38 38 38 Load(18) = 150 + 30i or L = 1.259 K 0. the next step is applied and this cycle is repeated until the HB point be found. the selected states except the last one are the same and only their rank and participation factor are somewhat different. Eigenvalue locus when bus 32 is slack bus. a different load scenario and more importantly a different generation scenario have been applied.089 K+3 power system (step 2 of the step by step algorithm). We examined other generation scenarios. Eigenvalue locus when slack bus is changed to bus 32 and load of bus 8 is increased are shown in Fig.115 K 0. but the main participating states are relatively constant. In addition to changing the slack bus in the step by step algorithm. but their rank and especially participation factor may be changed. Between results of Table 3 and two examinations of Table 1. respectively. another dimension of nonlinearity is observed in the behavior of drivers of critical eigenvalues. As seen.2. However.152 K 0. Generation scenario In all load scenarios of the previous subsection. PV curve is usually constructed by successive load ﬂow solutions [3.

Obtained results for load scenario load(18) within this framework are shown in Fig. M. Between Tables 3 and 4 (with different load and generation scenarios) 9 out of 10 states are common but with different ranks and participation factors.574 K+2 Fig. It is noted that the same critical eigenvalues are seen in Fig. Similar results (e.690 K 0.859 K 0.151 K 0.353 K 0. and participation factors) were obtained in all examined generation scenarios.186 K+3 0.781 K 0.609 K 0. Eigenvalue locus in the new framework.959 K+3 0. 3. Generator contingency Fig.3. PV curve for two different generation scenarios in response to load(8).734 K 0. 9.946 K+3 0. On the other hand. HB point 1 K+3 0. like Fig. These furnaces are successively ﬁred with a small time difference among them imposing the set of step load changes to the power system. although both have the same load scenario. 7 and Table 3. HB point 1 K+3 0.8i. . 8.421 K+3 0.g.861 K+3 0.878 K+3 0.144 K Load(18) = 854 + 170. Eigenvalue locus for increasing load of bus 1 Table 3 Participation factors when bus 32 is slack bus (HB point) Factor State Model ESDC1A GENROU GENROU ESDC1A GENROU ESDC1A ESDC1A GENROU ESDC1A GENROU Bus 33 33 33 37 37 38 36 38 32 36 Table 4 Participation factors in the new framework (HB point) Factor State Model ESDC1A ESDC1A ESDC1A ESDC1A ESDC1A GENROU GENROU GENROU GENROU GENROU Bus 36 37 33 32 38 36 37 38 32 33 Load(8) = 900 + 300i.g. relatively ﬁxed main participating states and variable DLM. We discussed about different kinds of generator and branch contingencies in [1]. Amjady. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 several furnaces (e. 9. ﬁxed critical eigenvalues.2636 N.155 K+3 0. 9 and Table 4. 10 Â 65 MW). SLM.459 K+2 0.R. 9 is completely different from DLM in Fig. DLM in Fig. respectively. 4.148 K+3 0.

10 and 11 are shown in Table 5. 11. although load scenario is the same. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 2637 of IEEE68 test system is shown in Fig.e. In other words. Fig. the DAE model of the system is changed. 10. Participation factors at HB point of Figs. 11 shows this locus with the same load scenario when generator of bus 67. however the generator contingency changes critical modes (not seen in the previous examinations) due to variation of dynamic characteristics of the power system. Eigenvalue locus for the IEEE68 test system without contingency. . Amjady. i. Eigenvalue locus for the IEEE68 test system when G15 is disconnected. 10. M.R. G15. is disconnected.N. As seen. Only 4 out of 10 states are common between the Fig. Fig.

R. but in the branch contingency a main transmission line with signiﬁcant effect on the stability margin is disconnected [1].054 State K+5 K+4 K+5 K+4 K K+3 K K+3 K+5 K+4 Model GENROU GENROU GENROU GENROU GENROU ESDC1A GENROU ESDC1A GENROU GENROU Bus 66 66 65 65 65 65 63 63 64 64 two cases (compare it with 9 common states between two cases of Table 1). 96). 10 and 11 Without contingency (base case) Factor 1 0. . Fig. the critical eigenvalues do not change.997 0. By further comparing Tables 1 and 5. respectively.177 0.e.178 State K+5 K+4 K+5 K+4 K+5 K+4 K+5 K+4 K K+3 Model GENROU GENROU GENROU GENROU GENROU GENROU GENROU GENROU GENROU ESDC1A Bus 66 66 65 65 67 67 68 68 68 68 With generator contingency Factor 1 0. In other words. Fig. it is observed that the main participating states of the IEEE68 are mostly from the mechanical part equations of the generators (e. However this is not the case for all contingencies. 10. i. i. are (87. 11. Only 5 out of 10 states are common between Table 6 and the base case of Table 5. results of the time domain simulation were in accordance with the modal analysis as indicated in the step 6 of the step by step algorithm. if in this test system transmission line 1–47 be disconnected.150 0.087 0.094 0.758 0.g. respectively. In the mentioned examinations.286 0. Participation factors for the critical modes of Fig.289 0. 3. with the generator contingency. For instance.g.2638 N. 10 and 12 it can be seen that with the same load scenario. 88).e. Amjady. exciter output EFD) or electrical part equations of the generators (e. severe generator or branch contingencies can change the critical modes.249 0. Comparing Figs. It is noted that severe generator and branch contingencies are intentionally selected in the previous and current examinations. Branch contingency Fig. Large differences among the participation factors of these two cases are also seen. rotor angle and speed) while those of the New England are mostly from the exciters (e. and with the branch contingency. 12 shows Eigenvalue locus for IEEE68 when transmission line 41–42 is disconnected.936 0. i. and (95.g. larger differences among the participation factors of the two cases of Table 5 are also seen. 12 at HB point are shown in Table 6.290 0. Pair of critical eigenvalues of IEEE68 in the base case.285 0. Besides. 77).4. branch contingency changes the critical eigenvalues. (76.055 0. 10. Eigenvalue locus for the IEEE68 test system when transmission line 41–42 is disconnected.761 0. a small perturbation at load(1) = Fig. Fig. 10–12 reveals that the generator contingency increases DLM but the branch contingency greatly decreases it. 12. M. For instance.e. to illustrate effect of these contingencies on critical eigenvalues. In the generator contingency a weak generator owning high participation factors in the critical modes (base case of Table 5) is eliminated. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 Table 5 Participation factors at HB point of Figs. Considering Figs. in the load scenario shown in Fig.937 0. 12.995 0. Eq) [19] indicating different dynamic characteristics of these two test systems.

the oscillations are well damped here. variation of voltage of bus 1 in response to the same perturbation at load(1) = 2440 + 1220i (considerably larger load level) is shown in Fig. we placed four power system stabilizers (PSS) on the generator buses 65. The same load and generation scenarios of Fig.e. mostly because of too low bus voltage for operation. or (ii) rotor oscillations of increasing amplitude due to lack of sufﬁcient damping torque.037 K 0. Response to a small perturbation in load of bus 1 of the IEEE68 test system without PSS at load(1) = 2000 + 1000i (step change is applied at 5 s). Although some eigenvalues approach to the imaginary axis. Instability that may result can be of two forms: (i) increase in rotor angle through a nonoscillatory or aperiodic mode due to lack of synchronizing torque. PV curve of the New England test system with constant impedance load model. Eigenvalue locus in Fig.029 K+5 Fig. Amjady. However. slightly after the HB). With the installed stabilizers. In spite of Fig. To enhance DLM of the power system. 13. The applied stabilizers are speed sensitive units with ESTAB1 model. 13. By comparing Figs. indicating emergence of the oscillatory instability. On the other hand. 13 and 14. The PSS can shift the critical eigenvalues toward the left hand side of the complex plane and so can increase DLM. 10) to 2470 + 1235i. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 Table 6 Participation factors at HB point of Fig. which is another consequence of this paper. 13. 4 are considered here. M. with the mentioned stabilizers. but do not cross it. 14. A system Fig. Response to a small perturbation in load of bus 1 of the IEEE68 test system with PSS at load(1) = 2440 + 1220i.043 K 0. in today’s power systems. small-disturbance rotor angle stability problem is usually associated with insufﬁcient damping of oscillations [4]. However. which its block diagram is presented in the Appendix and more details can be found in [24]. we can even increase DLM of the power system up to its SLM (HB is eliminated).034 K+5 0. 67.N. effectiveness of PSS for improving small disturbance voltage stability is also seen. 15 and 16.030 K+3 0. there is no dominant eigenvalue here. HB point 1 K+5 0. Fig. As a comparison.25]. 66. 14 shows that the SLM does not change with addition of PSS. 2000 + 1000i (i.065 K 0. 15. Besides. PV curve and eigenvalue locus of the New England test system with constant impedance model for the load buses are shown in Figs.R. However. and 68 of the IEEE68 test system. DLM of the power system was increased from 1966 + 983i (HB in Fig. . Indeed. 12 Factor State Model GENROU GENROU GENROU ESDC1A GENROU GENROU GENROU GENROU ESDC1A GENROU Bus 66 66 53 53 62 66 68 68 66 65 2639 Load(1) = 500 + 250i. Variation of voltage of bus 1 in response to this perturbation is shown in Fig. Fig. 16 conﬁrms this result.052 K+3 0. no HB or oscillatory instability is observed in this case. In spite of Fig. However. In some cases. by adjustment of PSS parameters.986 K+4 0. EPs on the lower curve are typically not viable. it can be seen that the trace of SEPs continues after the nose point to the lower part of the PV curve until no EP can be found. 4. In the previously mentioned examinations. is applied to load of bus 1. small-disturbance voltage stability refers to the system’s ability to maintain steady voltages when subjected to small perturbations such as incremental changes in system load [4]. Small-disturbance (or small-signal) rotor angle stability is concerned with the ability of the power system to maintain synchronism under small disturbances. the load buses have constant power load model.033 K+4 0. respectively. 14. Some researchers work on the PSS and its parameter tuning to enhance the angle stability [20.

however we frequently observed occurrence of HB and SNB in practice. an important subset of power system global stability. . Effect of load and generation scenarios. Effect of PSS to enhance DLM of this stability is also shown. voltage swings and even occurrence of oscillatory instability is the major concern in the southern part of Iran’s power system network [14]. Besides. and generator and branch contingencies on the critical eigenvalues and dynamic and static stability margins are evaluated. we found more nonlinear behaviors of the stability than previous works. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 Fig. M. Besides. linear and nonlinear analysis tools. The research work is under way in order to provide a more comprehensive evaluation of the stability boundaries with more dynamic models (e. dynamic. Although results of this paper are presented for the IEEE test systems. drivers of the small disturbance voltage stability and their sensitivities with respect to power system states under various conditions are presented and discussed. The ill-condition and well-condition behaviors of this stability are introduced. 17. Important conclusions are drawn giving a better insight to both operators and planners of the power system about small disturbance voltage stability. product of PTI company. Numerical results of this paper have been obtained by the software package PSS/E 25. Conclusion In this paper a more deeper evaluation of the small disturbance voltage stability. Fig.2640 N. Amjady. Similar results have been obtained for constant current load model.4.g. Table 7 State variables of the dynamic models GENROU and ESDC1A GENROU model for generators States K K+1 K+2 K+3 K+4 K+5 Description E0q E0d WKd Wkq D Speed Angle ESDC1A model for excitation system States K K+1 K+2 K+3 K+4 Description Sensed VT Lead lag Regulator output VR Exciter output EFD Rate feedback integrator 4. Eigenvalue locus of the New England test system with constant impedance load model. assessment of small disturbance voltage and angle stabilities in a uniﬁed framework can be considered as a matter of the future research.R. is presented by a combination of static. For instance. dynamic load). 16. Especially. Block diagram of the ESTAB1 model for PSS. break up by selective protection will follow.

[20] Hongesombut K. Technical Report 2003 – 3. NJ/USA: PrenticeHall. IEEE Trans Power Syst 2004. [10] Yorino N. Lee KY. vol. In: 14th PSCC. 1998. Reeve J. Deﬁnition and classiﬁcation of power system stability. Ramos ER. [5] Cutsem TV. 2003. A new bifurcation analysis for power system dynamic voltage stability studies. [15] Nam HK. USA. Canizares C. Anderson G. Block diagram of the ESTAB1 model for PSS is shown in Fig.edu/research/pstca/. 2001.21:19–27. 1998. Vournas C.148:201–8.ee. Esmaili M.25:705–15.4 Program application guide – vol.20:973–83. The EPRI power system engineering series. Identiﬁcation of electromechanical modes and placement of PSSs using relative gain array. [16] Huang GM. Zhao L.19: 1387–401. Amjady. Fast computation of voltage stability security margins using nonlinear programming techniques. Paserba J. Power system stabilizer tuning in multimachine power system based on a minimum phase control loop method and genetic algorithm. Model of the excitation systems is IEEE type DC1 (ESDC1A). Indices to detect Hopf bifurcations in power systems. II. 2000. Electric Power Syst Res 2003. 1. Power system stability and control. Boston/ USA: Kluwer Academic Publishers. Boston/ USA: Kluwer. Song X. [6] Kundur P.washington. IEEE Trans Power Syst 2000. [8] Conejo AJ. Canizares CA. Kim YK.R. [18] Marszalek W. Serrano Duque AC. Proc Inst Elect Eng Gen Trans Distrib 2001. Tampere University of Technology Publications 344. [13] Chen H. M.19:410–7. Esmaili M. 15–23.148: 263–7. Esmaili M. Li HQ. Transient and voltage stability enhancement via coordinated excitation and UPFC control. Dynamic voltage security assessment by a neural network based method. Voltage security assessment and vulnerable bus ranking of power systems. p. p. Identifying a vanishing eigenvalue in voltage collapse analysis with consideration of limits. [11] Mithulananthan N. [14] Amjady N. [25] Milanovic JV. 1994. 2002. Canada. IEEE Trans Power Syst 2004. IEEE Trans Power Syst 2004.15:363–9. Power Technologies.. [21] Pai MA. soc winter meeting. A method of voltage stability evaluation for branch and generator outage contingencies.64:227–37. [9] Zarate LAL. References [1] Amjady N. 17.19:252–9. Electric Power Syst Res 2003. 2. Castro CA. [23] Zambroni AC. Inc. Ramos JLM.19:1005–14. 1. Milano F.N. Improving voltage security assessment and ranking vulnerable bus with consideration of power system limits. 2002. [24] PSS/E 25. More details can be found in [24]. [2] Repo S. 882–7. 1989. [12] Kodsi SKM. A new Eigen-sensitivity theory of augmented matrix and its applications to power system stability analysis. Wang Y. [19] Amjady N. [7] Amjady N. In: Proc IEEE power eng. Pai MA. p. Congestion management ensuring voltage stability. Bose A. IEE Proc Gener Transm Distrib 2001. Canizares CA. Trzaska ZW. vol. 1997. . [17] Sauer PW. Ohta A. Shim KS. Energy function analysis for power system stability. Int J Elect Power Energy Syst 2003. Inc. [3] Amjady N. A framework of reliability assessment with consideration effect of transient and voltage stabilities. IEEE Trans Power Syst 2005. [22] http://www. Harada S. Zhou R. Application of a new sensitivity analysis framework for voltage contingency ranking. [4] Kundur P.66:215–26.. Voltage stability of electric power systems. McGraw-Hill. ON. Mitani Y Tsuji K. Singularity-induced bifurcations in electrical power systems. IEEE Trans Power Syst 2004. IEEE Trans Power Syst 2006. Modeling and simulation of IEEE 14bus system with facts controllers. Sasaki H. Ansari / Energy Conversion and Management 49 (2008) 2629–2641 2641 List of symbols p parameters of system conﬁguration and operation condition x dynamic state variables (slow modes) y instantaneous variables (fast modes) Ju unreduced Jacobian partial derivative of vector f with respect to vector x fx Jr reduced Jacobian gy algebraic Jacobian Appendix In our time domain simulation. Sevilla. IEEE Trans Power Syst 2006. University of Waterloo. Waterloo. On-linevoltage stability assessment ofpower systems – an approach of black box modeling. In: Proc of NAPS. State variables of the dynamic models GENROU and ESDC1A are represented in Table 7. Ajjarapu V. IEEE Trans Power Syst 2006. the generator is modeled by 6th order differential equations (GENROU model). Bertrand RG. Power system dynamics and stability. vol.20:312–20.21:357–64. 1–7.

voltage stability modal analysis

voltage stability modal analysis

- Condition Based Maintenance (CBM)
- New Chalkboard
- RealScan-10_MTBF
- Teixeira Et Al 08
- High-Level Power Analysis for Intellectual Property-Based Digital Systems
- MTBF
- IBP1365_12
- (eBook) Transient Pipe Flow in Pipelines and Networks - The Newest Simulation Methods
- Face recognition based on composite correlation filters
- Hazop Method
- Lean Tools
- Reduce DOE Costs
- 99-1980 - copia
- A physics-based simulation approach for cooperative erection activities.pdf
- Important
- Veronesi 2016 - Paper statistical learning
- Dynamic Simulation of Grinding Circuits
- Input Design
- Calibration
- Seismic 3sadghd
- push 30 2012.pdf
- I08TPC-293Juke
- PowerTech - Gentesting_Brochure
- Analysis of the Adsorption Process and of Desiccant Cooling Systems
- IPC2012-90210
- Uso Su Tre Casi Dei Modelli Cfd Per Prevedere l'Andamento Dei Fumi in Ambienti Chiusi Complessi
- งานวิจัย 11
- Anand Parmar Paper
- Evaluation of Observer Structures With Application to Fault Detection
- The Importance of First-principles, Model-based Steady-state Gain Calculations in MPC-A Refinery Case Study

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue listening from where you left off, or restart the preview.

scribd