This action might not be possible to undo. Are you sure you want to continue?
system network by optimally scheduling the parameters of the available controllers. A set of real-time stability indicators, including a transient stability index, transient voltage stability index and damping factor are used to reflect the stabilities of power systems. Then, the relative significances of a group of controllers to the power system are used as references for the optimiser to schedule new controller parameters. This coordinated control scheme effectively allocates power system resources to handle different stability issues in different time frames. The results suggest that power system security can be further enhanced by means of coordinated control.
Power systems are becoming more complex with the increasingly stressed stability margins resulting from the electricity market. Traditional vertically integrated power systems are being replaced by market-based operation after deregulation. Consequently, it becomes more difficult to predict the behavior of power systems. Furthermore, the introduction of independent power producers and power electronic loads, which are nonlinear in nature, have also complicated the operation of the grid. Although the operation and the interconnection of power systems are evolving, two classes of approaches appear in the mainstream control of power systems. The first approach is a decentralised controller design where classical linear controllers as well as modern nonlinear controllers are designed for handling predictable local problems [1,2]. Although remote interactions can be considered in terms of local control variables, decentralised controllers rarely coordinate together to resolve any particular problems from the viewpoints of the network, not to mention problems of an unusual nature. The other approach is a hierarchical control design based on a wide-area measurement [3–5].
Accelerated by fast communication facilities, the upper level controllers can obtain remote information from different local controllers, and if local controllers have conflicting effects with upper level control objectives, they will tune these local control systems in a coordinated way according to various global optimisation procedures. These hierarchical optimal controllers usually can achieve good performance for a single control objective or two, such as transient stability, voltage regulation or oscillation damping, but may neglect the other potential stability issues in power systems. Power system stability issues are not mutually exclusive. There have been qualitative studies [1, 6] showing that some controller interactions could degrade the power system stability under certain circumstances for which they are not designed. The concept of coordinated different control evolved from previous works [7–13]. A novel global control strategy was introduced for coordinated transient stability and voltage regulation by excitation in [7, 8], developing a framework for more general schemes involving the coordination of many controls with diverse control goals. The framework was further refined in [9, 10], applied to a benchmark test system and covered more different network scenarios. In addition, preventive and emergency control actions were coordinated to enhance transient stability in [11, 12], and an adaptive optimal control scheme has been applied against cascading blackouts in China . This paper extends the framework  to investigate coordinated security control. First, various system trajectories are utilised to analyse the instability ‘symptoms’ of transient stability, voltage security assessment and small signal stability. The formulation of remedial actions has to be based on the due consideration of these ‘symptoms’. Then, the ‘relative effectiveness’ of local controllers is evaluated for the capabilities to resolve transient instability, voltage instability and small signal oscillatory problems.
Power system signals such as rotor angles of the generators, bus voltage profiles and oscillating frequencies are sensed and used to calculate the transient voltage and small signal stability costs. Later, a ‘sensitivity function’ is formed, which is used by the optimizer later as a reference for control parameter scheduling. The severities of the stability issues are compared based on their weighted costs, and a new set of controller parameters are optimally calculated with the aim of reducing the total cost of system stability. It is to be emphasised that it is not a centralised controller design but rather the maximal exploitation of available control resources. The simulation results for a benchmark system show that this coordinated control scheme is effective, robust and provides significant stability improvement for many kinds of system disturbances. The paper is structured as follows. The coordinated control scheme is described in Section 2. Section 3 presents the power system models and the formulation of the coordinate control system. The example power system, its coordinated control system and simulation results are provided in Section 4, followed by discussions in Section 5 and conclusions in Section 6.
2. Coordinated control scheme
The goal of the proposed adaptive coordinated control is to maintain system stability under all operating conditions, especially unforeseeable ones. It is a framework within which a variety of methods and different controllers can be coordinated to handle complex system problems to the extent allowed by the available information and communication technology. Hill et al. and Guo et al. [7, 8] offered the possibility of coordinating control actions across the whole system geographically and for all operating situations in an adaptive way. Hill et al.  referred to the collection of available control elements ‘swarming’ onto the problems as they arise.
The coordinated control framework as shown in Fig. 1 has a two-tier structure, which is local control and some degree of central coordinated control. The local controllers perform their normal operations with the original local settings until an abnormal situation occurs where the coordination control takes over. Based on the global control objectives for the power system, a set of high-level coordination controllers (CCs) is built as the supervisory controllers for the existing local controllers. This system-wide coordinated control has four main tasks, namely: (1) indicator processing, (2) scheduling, (3) optimisation and (4) actuation. The function of each task is described as follows: 2.1. Indicator processing : The signal-processing layer of the CC collects and processes raw network information from local controllers in real-time such as power flow, system frequency, bus voltages and alarm signals. The signal is then transformed into meaningful signals such as stability indicators. The CC will then determine the severity of each kind of instability, that is to decide whether the system is in the proximity of transient, voltage or frequency instability. 2.2. Scheduling : Depending on the severity of the particular type of instability, a group of control actions will be scheduled (typically of the switching type) which have the capacity, with further tuning, to overcome the instability. 2.3. Optimisation : The optimisation layer measures the degree of instability of the power system and determines the necessary amount of control action to be provided from each selected local controller. The stabilising effects of each local controller towards different types of instabilities are calculated beforehand. The results can be recorded in the CC either in the form of a lookup table or a set of linear/nonlinear equations.
The information enables the optimisation algorithms to fine tune the action plan established by the scheduler in step 2 for optimal performance.
Fig.1 Adaptive coordination control framework
Remark 1: The upper level CCs can handle either one single issue (such as voltage regulation, oscillation damping and so on.) across a wide area or regional tasks, spanning multiple stability issues from a small regional area to a wider interconnected power grid. The designs of these upper level controllers are based on the global control objective of the power system.
These coordinated controllers can be installed in different areas of the power system and linked together via a communication network, which will necessarily be limited in structure and speed. Note that if a communication link to the controller is broken down, local controllers can still provide adequate performance although it may not maximise their
contributions to the optimisation of the whole power system. Remark 2: In real-time utilisation, the availability of Phase Measurement Units (PMUs) and the locations of local controllers have to be considered such that the optimization procedure can be simplified and the global control objective(s) guaranteed. In addition, the two-tier structure of the coordinated control framework enables information processing and decision making to be made at the local control systems by the use of distributed computation technology. This can significantly reduce the computational burden on the CC and, thus, reaction to the event in real time is possible. The proposed control scheme should be able to reduce the overall processing time in comparison with the centralised control scheme.
3. Power system models and controller design
3.1 Power system models : The dynamics of interconnected power systems can be represented by a set of highorder nonlinear differential–algebraic formulas as follows
x˙ = f (x, y, u, m) 0 = g(x, y, u, m) (1)
where the differential equation describes the dynamics of the power system components including generators, dynamic load and control devices; the algebraic equation expresses the voltage, current and network power flow equations in the system; x € Rn is the dynamic state vector of the power system; y € Rl is the algebraic constraint vector ; u € Rp is disturbances,
which make the power system move from one equilibrium point to another; and m € Rq is the control vector, which consists of adjustable system parameters. The operating point (x, y) of the power system (1) varies according to the disturbance µ and the control vector m. When the specific disturbance µ0 occurs, control vector µ=µ0+µ+ where µ0 is the control vector of the original steady state and µ+ is the adjusted control vector. The aim of the proposed control scheme in Section 2 is to obtain the optimised adjustable control vector µ . At the equilibrium point (x0, y0), power system dynamic state vector x˙ and disturbances u˙0 are equal to zero. 3.2. Stability assessment and cost Three categories of stability analysis  are mainly involved in the systematical controller design of power system, including transient stability, voltage stability and small signal stability analysis. In order to effectively allocate power system resources to handle different stability issues in different time frames and situations, stability assessments and indices are evaluated through various system trajectories in real time. Despite different methods to study transient stability, the responses of power angle of the generators are always regarded as the key trajectories to analyze the synchronization of power system. Thus, it is obviously proper to employ power angle excursions to calculate transient stability cost Cts , which is an important transient stability indicator for the optimisation of the CCs. For voltage stability assessment, a cost Ctvs which is based on the trajectories of bus voltages, is used to measure how ‘voltage unstable’ the system is.
In the case of small signal oscillation, it is common that many different modes of oscillation are presented in a multimachine power system. Each mode is characterised by its natural frequency and damping factor. The normalized deviation of the damping factor of the targeted mode is used to evaluate the small signal stability cost Css . The costs Cts, Ctvs and Css can be estimated in real-time and are dependent on the existing power system state variable x0, algebraic variables y0, disturbances u0 , controller parameters µ0 and adjusted control variables µ+ . Thus, the estimated costs can be written as
Where Cts ,
Css represent the changes of transient stability cost Cts , voltage
stability cost Ctvs and small signal stability cost Css when control variables µ are adjusted. The changes of stability cost Cts , Ctvs and Css are the indications of the capabilities of
the power system local controllers for different stabilities. In this paper, the ‘sensitivity matrix’ is proposed to measure the changes of stability cost at a nominal operating point. As a first-order approximation, the relationship between cost changes and controller parameters can be represented as follows
Where C is the sum of changes of different stability costs, A is the gradient between the change of stability costs and the adjusted controller parameters K, that is the sensitivity matrix.
(Note that K can be any controller parameters, such as for load shed. Generally, K is the parameters of the controllers and affects the output of the controllers µ.) The details of calculating different stability costs will be provided later. Certainly, higher order nonlinear controller function can be designed, but it is not the focus and is therefore not covered in this paper. 3.3. Optimisation The determination of the local controller parameter actuation is formulated as a constrained optimisation program. A quadratic objective function is proposed
min J = αtsC2ts+ αtvsC2tvs+ αssC2ss
αts + αtvs + αss = 1
Where the coefficients αts, αss and αtvs are the weights for the corresponding stability indices. They are determined with the aim to balance the contribution of each kind of stability in different time intervals. Therefore J is a function of current power system conditions and predicted optimal control parameters, µ
J = J (x0, y0, u0, µ)
This objective function is subject to two inequality constraints (a) Upper and lower bounds
(b) Limited rate of change in controller parameters (8)
The inequalities can be converted into the following general form Gi ( ) 0 (9)
3.4 Coordinated control design The proposed coordinated controller design provides a supervisory feedback for the adjustments of local controller parameters according to different network structures, distinct operating conditions and various control objectives. The CC design involves the following procedure: 1. design the scopes of upper level CCs based on the global control objectives and local controllers in the existing state of the power system prior to any contingencies; 2. Choose the corresponding methods for upper level CCs to assess transient, voltage and small signal stability; 3. Calculate different stability costs of local controllers and create a set of local controller schedules for overcoming potential instabilities of power systems; 4. Analyse the relationship between cost changes and controller parameters and form the Amatrix in (3); 5. solve the objective optimisation (4) and set up the relative significances of a group of control actions according to the sensitivity matrix A and 6. schedule the actuation of the local controller parameters.
4. Case study :
To illustrate the proposed control framework, control of a typical two-area power system will be investigated. 4.1 Two-area power system Fig. 2 shows the diagram of a modified version of the two area system described in . The modifications are as follows: Line 6–7 and line 9–10 are split into double circuits
Fig. 2 A modified 2-area system Line 7–8 and line 8–9 are turned into four-line circuits Two static VAr compensator’s (SVC) are installed at Bus 7 and Bus 9
The 0.5–0.6 Hz inter-area mode of oscillation, which has the lowest damping factor compared with other local modes of oscillation, is chosen as the target for damping in this paper [1, 14]. Other modes are not adversely affected to an extent that they need special attention in this system. All the dynamic data of the generators and static data of the system remain unchanged from that given in .
Throughout this paper, a number of contingencies are referred to. The descriptions of them are as follows: C1: Three-phase fault at Bus 7. Fault is cleared by tripping line 7–8 at 0.1 s later; C2: Three-phase fault at Bus 8. Fault is cleared by tripping line 7–8 and line 8–9 at 0.1 s later; C3: Three-phase fault at Bus 9. Fault is cleared by tripping line 8–9 at 0.1 s later; C4: 10% load increase at Bus 9.
Fig. 3. Block diagrams of simple exciter system, PSS and SVC
In the two-area power system, each of the four generators is equipped with a simple exciter and power system stabilizer (PSS). The control blocks of the simple exciter, PSS and the SVC are shown in Fig. 3. The parameters are listed in the Appendix. 4.2. Controller sensitivity tests Power system controllers are mostly designed for some kind of primary objectives. For example, PSS is designed for damping and SVC is for voltage support. However, their contributions to other types of stability controls are also valuable. Information like the relative strength of each controller is required for the coordinated control system to formulate suitable switching actions. In this paper, the stability assessments were calculated by using Powertech’s Dynamic Security Assessment Software Tools (DSA ToolsTM ).
4.2.1 Transient stability assessment:
The transient stability
index of a multi-machine system can be computed by using the extended (EEAC) equal . area This criteria method
separates the system generators into two groups, that is the critical cluster and the rest. The cluster of critical machines is identified as the group of ‘candidate’ critical machines which give the smallest critical clearing time (CCT). Fig. 4. Power angle diagram
The stability index is obtained by calculating the difference between the ‘acceleration area’ and ‘deceleration area’ in the corresponding EEAC assessment. Fig. 4 identifies the quantities referred to. Then the index
Where Ainc is equal to kinetic energy increasing area (during fault) and Adec is equal to kinetic energy decreasing area (post fault). The initial gain values of the exciter, PSS and SVC are set to be 200, 30 and 200, respectively. By varying the gain of one kind of controller at a time, the transient stability indices are found (see Fig. 5). The following points are observed: Small load change (in this case, 10%) does not affect transient stability – see Case C4; There are discontinuities in the stability indices. This is because the EEAC algorithm separates different machines into different clusters according to the degrees they are affected. Different combinations of critical generators will change the inertia base of the calculation, resulting in the discontinuity of the plots; It shows that increased gain in the excitation system does not always increase the stability index, and increase in PSS gain does not necessarily worsen the transient stability. These two figures show that the actions of the excitation system and PSS in transient stability tend to be oppositional; Increase in SVC gain does not improve the transient stability
Fig. 5. Transient stability index with adjustable excitation, PSS and SVC gains
4.2.2 Transient voltage stability assessment: 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 . By estimating the ‘distance to the proximity of voltage collapse’, many stability indices have been proposed . In this paper, voltage security (or transient voltage Stability), which describes the deviation of transmission system voltage from its nominal value is used . Based on signal energy analysis , a transient voltage instability index, which measures the normalised distance of the transient voltage from the steady-state value, is chosen. The equation is stated below
where ɳtvs is the transient voltage instability index, ncl is the time step at the clearing time; and the subscript i denotes buses and i denotes the time step. Note that unlike the transient stability index, ɳts (that the larger the index value, the more stable is the system), bigger ɳtv implies that the system is more voltage unstable. Inspection on Fig. 6 reveals that: the transient voltage instability index increases monotonically in a close-to-linear fashion as the excitation gains change from 100 to 300; change in PSS gains has relatively much less effect on transient voltage stability; the SVC is the most effective device for transient voltage stability improvement; load variation has minimal effect on transient voltage stability.
It is therefore concluded that a reduction in excitation gain and an increase in SVC gain can improve transient voltage stability.
Fig. 6. Transient voltage stability index with adjustable excitation, PSS and SVC gains 4.2.3 Small signal stability assessment: Different modes of oscillation are present in a multi-machine power system. Each mode is characterised by its natural frequency and damping factor.
Prony’s method  is a common way to extract the frequency information in real-time. It considers the sampled data as a linear combination of exponentials.
N, where T is the sample interval, Ak is the amplitude, αk is the damping factor,
f k is the sinusoidal frequency,
is the sinusoidal initial phase.
The p-exponent discrete-time function of (12) can be expressed in the form
(13) The damping factor ( ), corresponding frequency ( f ) and damping index (ɳss) can be calculated by
The diagrams in Fig. 7 show the damping variations with different controller gains. It is noted that the damping ratio of the inter-area mode increases almost linearly with the increase of PSS gains. On the other hand, although the damping factor increases in the excitation gain ranges from 0 to about 60, it begins to decrease monotonically after the peak. The increase of SVC gains also decreases the damping ratio in this case.
Fig.7 Small signal stability index with adjustable excitation, PSS and SVC gains (Frequency = 0.55 Hz)
In general, the procedure of sensitivity measurements demonstrated above can be adopted for any kinds of dynamic problems, indices and control actions. Those chosen here are for illustration. 4.2.4 Stability costs: Although the transient stability index from EEAC is a well-accepted indicator for transient stability, it cannot provide a continuous assessment of the overall status of the generator angles. A modified method to measure the overall rotor angle deviation is, thus, proposed
where Mi is the moment of inertia of the ith machine in the S-cluster, Mj the moment of inertia of the jth machine in the A-cluster,
the rotor angle of the ith machine in the S-cluster,
rotor angle of the jth machine in the A-cluster To represent the angle excursion as a cost to the power system transient stability, the following function is used
(18) Where dc is the critical angle of the system. The value of Cts is between 0 and 1. The transient stability ‘limit’ is defined by Cts = 1. Transient voltage instability index can be directly used as the cost to be minimised for transient voltage stability, that is
Note that the value of the transient voltage stability cost is kept below l. During phase-to-earth faults, voltage would momentarily drop to zero. The voltage deviation (thus voltage stability cost) from the nominal value would be large when compared with other stability costs. However, transient stability cost is considered to be the most important during the fault period. For illustration here l is taken as 0.3, which covered the voltage fluctuations encountered. The small signal stability cost is the normalised deviation of the damping factor of the frequency from the targeted one
where ɳss0 is the target damping factor. The maximum value of Css is 1 and the minimum value is limited to 0, even if ɳss 4.2.5 Sensitivity matrix: From the analysis of above three sections, the ‘relative effectiveness’ of three controllers, namely an excitation controller, PSS and SVC have been evaluated for the capabilities to resolve transient instability, voltage instability and small signal oscillatory problems. Thus, the sensitivity matrix A in (3) can be summarised at a nominal operating point. Due to the relatively simple relationship between cost changes and controller parameters actuation, a linearised sensitivity matrix is formed as follows
The value of this A-matrix is provided in the Appendix.
4.2.6 Optimisation function: In order to realise the coordinated control and actuate various local controller parameters, the quadratic objective function (4) must be chosen in advance. Since the twoarea power system is a comparatively small system and system response after disturbances with limited time interval, the coefficients αts, αss and αtvs are considered as constants. The values are taken to be [0.6, 0.3, 0.1], respectively, in the simulation. 4.3. Simulation results In this section, a network contingency scenario is demonstrated. The effects of the global coordinated control scheme are compared with a conventional method where no global control is used, that is parameters are fixed at values determined in the Appendix. The control scheme is tested by simulations in MATLAB Power System Toolbox . The load at Bus 9 is increased by 10% at t = 2.0 s. Then at t = 6.0 s, a fault occurs at Bus 8. The fault is cleared by tripping one line 8–9. The results are shown in Figs. 8a–8d. Fig. 8a shows that the bus voltage under global control settles at a new equilibrium much faster than the conventional scheme. The cost comparisons in Figs. 8b and 8c confirm that transient stability, transient voltage stability and damping are improved by the proposed control scheme. The controller gains in Fig. 8d highlight the point that the controllers can automatically reschedule their parameters for successive events.
Fig. 8 Figure 8 Responses of the global coordinated control with successive events a Voltage magnitude at Bus 8 b Stability cost of the power system with global control c Stability cost of the power system without global control d Actual controller gain scheduling
The simulations have shown some significant improvements over a wide range of power system stability problems by the proposed coordination scheme. There are some points worth noting: 1. The coordinated control scheme used in Section 4 only coordinates controllers of similar types. In general, the global control will be capable of handling a mixture (hybrid) of continuous and discrete, dynamic and static control actions. Furthermore, advanced controllers like FACTS devices are also suitable for this scheme . 2. This coordinated control scheme utilises a basic feedback control algorithm to improve power system stability. Other advanced control algorithms such as model predictive control  can also be adopted to predict the power system behaviour and to evaluate parameters in advance. 3. The voltage cost in the objective function only considers’transient voltage stability’ Long-term voltage stability can also be taken into account as an additional cost to the objective function. Indicators like ‘loading margin’ can be used and remedial actions such as load shedding and capacitor switching can also be coordinated for voltage regulation.
This paper presented a generic adaptive coordinated control scheme for power systems as a further step towards universal type controllers which flexibly respond to all stability problems. The controller uses measurable network information to determine the locality and type of the stability nproblems. By formulating and coordinating the necessary control actions from all available local control actions, we can overcome system-wide power system problems. This framework is flexible to adopt new elements such as stability indicators and optimisation algorithms. The development of tests and models is further needed for evaluating the capabilities of different controllers under different instabilities.
7. References  KUNDUR P.: ‘Power system stability and control’ (McGraw-Hill, 1994)  CHAPMAN J.W., ILIC M.D., KING C.A., ENG L., KAUFMAN H.: ‘Stabilizing a multimachine power system via decentralized feedback linearizing excitation control’, IEEE Trans. Power Syst., 1993, 8, (3), pp. 830–839  TAYLOR C.W., ERICKSON D.C., MARTIN K.E., WILSON R.E., VENKATASUBRAMANIAN V.: ‘WACS-Wide-Area stability and voltage control system: R&D and online demonstration’, Proc. IEEE Energy Infrastruct. Defense Syst., 2005, 93, (5), pp. 892–906  OKOU F., DESSAINT L.A., AKHRIF O.: ‘Power systems stability enhancement using a wide-area signals based hierarchical controller’, IEEE Trans. Power Syst., 2005, 20, (3), pp. 1465–1477  MARQUES A.B., TARANTO G.N., FALCAO D.M.: ‘A knowledge-based system for supervision and control of regional voltage profile and security’, IEEE Trans. Power Syst., 2005, 20, (1), pp. 400–407  KUNDUR P., PASERBA J., AJJARAPU V., ANDERSSON G., BOSE A., CANIZARES C., ET AL.: ‘Definition and classification of power system stability’, IEEE Trans. Power Syst., 2004, 19, (3), pp. 1387–1401 [IEEE/CIGRE Joint Task Force on Stability Terms & Definitions]  HILL D.J., GUO Y., LARSSON M., WANG Y.: ‘Global hybrid control of power systems’. Preprints of Bulk Power System Dynamics and Control V, 26–31 August 2001
 GUO Y., HILL D.J., WANG Y.: ‘Global transient stability and voltage regulation for power systems’, IEEE Trans. Power Syst., 2001, 16, (4), pp. 678– 688  LEUNG J., HILL D.J., NI Y., HUI R.: ‘Global power system control using a unified power flow controller’. Proc. of the 14th Power System Computation Conf., PSCC, Sevilla, Spain, June 2002  LEUNG J.S.K., HILL D.J.: ‘Coordinated power system security control’. IREP Symp. 2004, VI Bulk Power Systems Dynamics and Control Symp. Managing complexity in power systems: from micro-grids to mega-
interconnections, Cortina D’Ampezzo, Italy, August 2004  XUE Y., LI W., HILL D.J.: ‘Optimization of transient stability control part-I: For cases with identical unstable modes’, Int. J. Contr., Auto. Syst., 2005, 3, (2), pp. 334–340  XUE Y., LI W., HILL D.J.: ‘Optimization of transient stability control part-II: For cases with different unstable modes’, Int. J. Contr., Auto. Syst., 2005, 3, (2), pp. 341–345  XUE Y.: ‘Defense schemes against power system blackouts in china with high load growth’. Proc. 15th Power System Computation Conf., PSCC, liege, Belgium, August 2005  KLEIN M., ROGERS G.J., KUNDUR P.: ‘A fundamental study of inter-area oscillations in power systems’, IEEE Trans. Power Syst., 1991, 6, (3), pp. 914–921
 XUE Y., CUTSEM T.V., RIBBENS-PAVELLA M.: ‘Extended equal area criterion justifications, generalizations, applications’, IEEETrans. Power Syst., 1989, 4, (1), pp. 44–52  Voltage Stability Assessment: Concepts, Practices and Tools: ‘IEEE PES, Power System Stability Subcommittee Special Publication’, August 2002  HISKENS I.A., HILL D.J.: ‘Energy functions, transient stability and voltage behaviour in power systems with nonlinear loads’, IEEE Trans. Power Syst., 1989, 4, (4), pp. 1525–1533  TUGLIE E.D., DICORATO M., SCALA M.L., SCARPELLINI P.: ‘A corrective control for angle and voltage stability enhancement on the transient time-scale’, IEEE Trans. Power Syst., 2000, 15, (4), pp. 1345–1353  MARPLE S.L. JR., CAREY W.W.: ‘Digital spectral analysis with applications’ (Prentice-Hall, 1987)  CHOW J.H.: ‘Power system toolbox, a set of coordinated M-files for use with MATLAB’ (Cherry Tree Scientific Software, 1996)  LARSSON M., HILL D.J., OLSSON G.: ‘Emergency voltage control using search and predictive control’, Int. J. Electr. Power Energy Syst., 2002, 24, pp. 121–130
8. Appendix The simple exciters have the following parameters:
TR = 0.01,TA = 0.2, KEXC = 200.0,VRMAX = 7.0,VRMIN = 27.0, KEXC(MAX) = 300.0, KEXC(MIN) = 0.0 The time constants (s) of the power system stabiliser are: T1 = .05, T2 = 0.02, T3 = 3.0, T4 = 5.4, Tw = 10, KPSS = 30. The SVC parameters are specified as follows: TR = 0.05, KSVC = 200.0, BMAX = 3.0, BMIN = 23.0, KSVC(MAX) = 300.0, KSVC(MIN) = 0.0 The upper and lower bounds of PSS are 150 and 0. The sensitivity matrix used is: [ -0.2, 0.01, 0.05; 0.2, - 0.02, - 25 ×10-6; 0.01, 0.005, -1.5]