Recent Philosophies of Automatic Generation Control Strategies in Power Systems
Ibraheem,Prabhat Kumar,and Dwarka P.Kothari,Senior Member,IEEE
Abstract—An attempt is made in this paper to prent critical lit-erature review and an up-to-date and exhaustive bibliography on the AGC of power systems.Various control aspects concerning the AGC problem have been highlighted.AGC schemes bad on pa-rameters,such as linear and nonlinear power system models,clas-sical and optimal control,and centralized,decentralized,and mul-tilevel control,are discusd.AGC strategies bad on digital,lf-tuning control,adaptive,VSS systems,and intelligent/soft com-puting control have been included.Finally,the investigations on AGC systems incorporating BES/SMES,wind turbines,FACTS devices,and PV systems have also been discusd.
Index Terms—AC/DC transmission links,automatic generation control,deregulated power systems,load frequency control,mul-tilevel control.
N OMENCLATURE
ACE Area Control Error
AFC Automatic Frequency Control
AFRC Automatic Frequency Ratio Control
AGC Automatic Generation Control
雅思培训的费用AI Artificial Intelligence
ANN Artificial Neural Network
A VR Automatic V oltage Regulator
BES Battery Energy Storage
CES Capacitive Energy Storage
CPS Control Performance Standard
DCS Disturbance Control Standard
DPM DISCO Participation Matrix
DTS Dispatcher Training Simulator
EACC Error Adaptive Control Computer
EDC Economic Dispatch Controller
FACTS Flexible Alternating Current Transmission System GAs Genetic Algorithms
GENCOs Generation Companies
GRC Generation Rate Constraint
IGBT Insulated Gate Bipolar Transistor
ISE Integral Square Error
LFC Load Frequency Control
renegadeLQI Linear Quadratic Integral
LQR Linear Quadratic Regulator
MES Magnetic Energy Storage
Manuscript received July10,2004.Paper no.TPWRS-00272-2003. Ibraheem is with Department of Electrical Engineering,Faculty of En-gineering and Technology,Jamia Millia Islamia,New Delhi110025,India (e-mail:ibraheem_).
P.Kumar is with Department of Electrical Engineering,Faculty of Engi-neering and Technology,Aligarh Muslim University,Aligarh202002,India.
D.P.Kothari is with Centre for Energy Studies,Indian Institute of Tech-nology,New Delhi110016,India.
Digital Object Identifier10.1109/TPWRS.2004.840438NERC North American Electric Reliability Control
NNs Neural Networks
在线中英翻译
NRPS Northern Region Power System
OPF Optimal Power Flow
PI Proportional plus Integral
PID Proportional,Integral,and Derivative
PLCC Power Line Carrier Communication
PV Photovoltaic
RF Redox Flow
RTOPF Real-Time Optimal Power Flow
SA Simulated Annealing
SMES Super conducting Magnetic Energy Storage
STC Self Tuning Control
SVC Static Var Compensator
生活大爆炸第二季全集下载
UHV AC Ultra-High-V oltage Alternating Current
VSC Variable Structure Controller
VSS Variable Structure System
I.I NTRODUCTION
T HE successful operation of interconnected power systems requires the matching of total generation with total load demand and associated system loss.With time,the operating point of a power system changes,and hence,the systems may experience deviations in nominal system frequency and sched-uled power exchanges to other areas,which may yield undesir-able effects[1].
There are two variables of interest,namely,frequency and tie-line power exchanges.Their variations are weighted together by a linear combination to a single variable called the ACE.The AGC problem has been augmented with the valuable rearch contributions from time to time,like AGC regulator designs in-corporating parameter variations/uncertainties,load character-istics,excitation control,and parallel ac/dc transmission links. The microprocessor-bad AGC regulator,lf-tuning regulator, and adaptive AGC regulator designs have also been prented. The most recent advancement in this area is the application of concepts like neural networks,fuzzy logic,and genetic algo-rithms to tackle the difficulties associated with the design of AGC regulators for the power systems with nonlinear models and/or insufficient knowledge about the system required for its accurate modeling.Apart from 教师节快乐 英文
advances in control concepts, there have been many changes during the last decade or more, such as deregulation of power industry and u of SMES,wind turbines,and PV cells as other sources of electrical energy to the system.Due to the,the control philosophies associated with AGC have changed to accommodate their dynamics and ef-fects on overall system dynamic performance.The prent study covers the critical review of a wide range of methodologies of
0885-8950/$20.00©2005IEEE
AGC regulator designs of power systems with their salient fea-tures.
II.O VERVIEW OF AGC S CHEMES
Thefirst attempt in the area of AGC has been to control the frequency of a power system via theflywheel governor of the synchronous machine.This technique was subquently found to be insufficient,and a supplementary control was included to the governor with the help of a signal directly proportional to the frequency deviation plus its integral.This scheme consti-tutes the classical approach to the AGC of power systems.Very early works in this important area of AGC have been by Cohn et al.[2]–[6].The works were bad on tie-line bias control strategy.Quazza[7]illustrated noninteractive control consid-ering i)noninteraction between frequency and tie-line powers controls a
care是什么意思nd ii)each control area taking care of its own load vari-ations.The investigations with large signal dynamics of LFC systems were reported by Aggarwal and Bergth[8].The rev-olutionary optimal control concept for AGC regulator designs of interconnected power systems was initiated by Elgerd[9].A technique bad on coordinated system-wide correction of time error and inadvertent interchange was incorporated in an AGC study by Cohn[10].Supplementary controllers were designed to regulate the ACEs to zero effectively.Later on,energy source dynamics were incorporated in AGC regulator design[11]. The standard definitions of the terms associated with the AGC of power systems werefinalized in[12].Following that,sugges-tions for dynamic modeling for LFC are discusd thoroughly in[13]–[15].Bad on the experiences with actual implemen-tation of AGC schemes,modifications to the definition of ACE are suggested from time to time to cope with the changed power system environment[16]–[19].Since many prently regulated markets are likely to evolve into a hybrid scheme,and some deregulated markets are already of this ,Norway),the effects of deregulation of the power industry on LFC have been addresd through[20].
A.Types of Power System Models
The AGC problem has been dealt with extensively for more than three decades.The major part of the
work reported so far has been performed by considering linearized models of two/multiarea power systems[3],[4],[7],[9],[21],[22].Later on,the effect of GRC was included in the types of studies, considering both continuous and discrete power system models [11],[22].Incorporating the dynamics of the energy source in AGC regulator design,Kwatny et al.[11]have propod an optimal tracking approach to AGC,considering load to be the output of the dynamic system.
The small signal analysis is justified for studying the system respon for small perturbations.However,the implementation of AGC strategy bad on a linearized model on an esntially nonlinear system does not necessarily ensure the stability of the system.Considerable attention has been paid by rearchers to consider the system nonlinearities[24]–[27].Tripathy[27] demonstrated the destabilizing effect of governor dead-band nonlinearity on conventional the AGC system.It is shown that governor dead-band nonlinearity tends to produce continuous oscillations in the area frequency and tie-line power transient respon.
B.Control Techniques
The pioneering work by a number of control engineers, namely Bode,Nyquist,and Black,has established links be-tween the frequency respon of a control system and its clod-loop transient
performance in the time domain.The investigations carried out using classical control approaches reveal that it will result in relatively large overshoots and tran-sient frequency deviation[9],[28],[29].Moreover,the ttling time of the system frequency deviation is comparatively long and is of the order of10–20s.
The AGC regulator design techniques using modern optimal control theory enable the power engineers to design an optimal control system with respect to given performance criterion. Fosha and Elgerd[30]were thefirst to prent their pioneering work on optimal AGC regulator design using this concept.A two-area interconnected power system consisting of two iden-tical power plants of nonreheat thermal turbines was considered for investigations.A new formulation for optimal AGC strategy has been witnesd in[31].
The feasibility of an optimal AGC scheme requires the avail-ability of all state variables for feedback.However,the efforts em unrealistic,since it is difficult to achieve this.Then,the problem is to reconstruct the unavailable states from the avail-able outputs and controls using an obrver.Considering state reconstruction,many significant contributions have been made [32]–[37].Bohn and Miniesy[32]have studied the optimum LFC of a two-area interconnected power system by making the u of i)differential approximation and ii)a Luenberger ob-rver and by introd
ucing an adaptive obrver for identification of unmeasured states and unknown deterministic demands,re-spectively.Exploiting the fact that the nonlinearity of the power system model,namely,the tie-line powerflow,is measurable, the obrver has been designed to give zero asymptotic error, even for the nonlinear model.
AGC schemes bad on an optimal obrver,which is a state estimator with decaying error at a desired speed,using a nonlinear transformation[33]and reduced-order models with a local obrver[34]have appeared in the literature.A simplified generating unit model oriented toward LFC and the method for its transfer function identification bad on a two-stage procedure indirectly reducing both noi effects and transfer function order is prented in[37].
Due to practical limitations in the implementation of regula-tors bad on feedback of all state variables,suboptimal AGC regulator designs were considered[38]–[40].A suboptimal and near-optimal LFC concept using modern control theory is pre-nted by Moorthy and Aggarwal[38].
Apart from optimal/suboptimal control concepts,modal con-trol theory has also been ud to design AGC regulators for power systems.The design method employing modal and sin-gular perturbation techniques to affect decoupling of the in-terconnection into its subsystem components has appeared
in [41].In the method,after achieving the decoupling,local con-trollers for each subsystem are designed individually to place the
clod-loop poles of each subsystem in some prespecified loca-tions in the complex plane,and then,the resulting controllers are ud to generate local control inputs,using local informa-tions only.The AGC regulator design using Lyapunov’s cond method and utilizing minimum ttling time theory has been propod by Shirai[42].The importance of the dominant time constant of the clod-loop systems in designing the regula-tors has been emphasized.The author has reported a bang-bang AGC policy bad on this method.
C.Control Strategies
In the early days,the AGC problem of power systems was dealt with using control strategies bad on centralized con-trol strategy[7],[9],[30],[40].Many control strategies have been propod on the basis of class of disturbances[7].Elgerd and Fosha[9]suggested a feedback and loop gain to eliminate the disturbance,and they also suggested a different feedback form to develop optimal controllers[30]for an electrical energy system.They assumed the load disturbances to be deterministic. They propod a proportional controller,disregarding the steady state requirements an
d compensation of load disturbances.The main limitation of the works prented on AGC considering centralized control strategy is the need to exchange information from control areas spread over distantly connected geographical territories along with their incread computational and storage complexities.
The decentralized AGC concept appeared in the power system control scenario to deal with such problems very ef-fectively,and conquently,many rearch papers using this concept with continuous and discrete time system models have appeared in the literature[43]–[50].In[46],the authors have examined the structural properties of obrvability and control-lability for a class of interconnected power system models.The propod scheme provides for the complete decentralization of a global state feedback control policy in the n that the area control feedback loops are completely decoupled.Again,a class of systematic distributed control design methods bad on i)distributed implementations of centralized control systems, ii)model reduction of dynamical systems,and iii)modeling of the interactions between the subsystems comprising the global control system is prented in[77].The beauty of the design is to achieve almost identical results as obtained with the centralized one.The design of decentralized load frequency controllers bad on structured singular values is discusd in [78].
Various AGC schemes bad on two-level[81]and multi-level[82]–[84]control concepts have been reported in the lit-erature.A two-level suboptimal controller has been suggested by Miniesy and Bohn[81].However,this approach does not en-sure zero steady state error,and hence,a multilevelfinite time optimal controller design ensuring zero steady-state error has been reported in[82].The advantage of hierarchical structure is reflected in the fact that even if one of the control levels fails, the system remains in operation.
A global controller,which also exploits the possible benefi-cial aspects of interconnections,has been applied for the LFC problem[84],and favorable results have been achieved.The re-duction of control efforts required in the AGC of interconnected power systems is sought with the help of a singular perturbation approach.This can be achieved by decomposing the system into slow and fast subsystems and designing controllers parately for each of the subsystems,and the controllers are combined to yield a composite controller.Using this approach,the investi-gations on the AGC of large power systems are available in the literature[87],[88].The parate controllers were designed for slow and fast subsystems and were combined in such a way that the slow subsystem always interacts with only one of the fast subsystems at a time[88].The study also involves the effect of parameter variation and GRC.
D.Excitation Control and Load Characteristics
In most of the AGC studies,it is assumed that there is no inter-action between the power/frequency and reactive-power/voltage control loops.It may be permissible only when the speed of the excitation systems is much faster than that of the LFC system, but in practical systems,during dynamic perturbations,there does exist some interaction between the two control chan-nels[27].Some papers consider this aspect[83],[89]–[93].A literature survey[90]shows that Durick[89]is probably the first to investigate the damping effects of voltage control in a two-area LFC system,assuming that i)reactive-power/voltage control loop has a much faster respon than power/frequency control loop and,thus,taking the area voltage perturbation to be directly available as a control variable,and ii)area voltage perturbation does not have any effect on the area load.Consid-ering the assumptions unrealistic,a realistic LFC model was developed by including the excitation control in one area and voltage-perturbation as the input in the other[90].The change in load demand due to voltage perturbation is considered in both areas.
Considerable rearch work has been carried out for the AGC of interconnected power systems incorporating the load charac-teristics[49],[94],[95],[98],[99].A method to obtain the re-spon of a large power system to cyclic load variations by mod-eling the power system by a t offirst-order,line
ar differential equations and the load variation as a Fourier ries pattern was demonstrated by Van Ness[94].The solution of the problem of optimum load frequency sampled data control with either un-known deterministic load or randomly varying system distur-bances is discusd in[49].Introducing an adaptive obrver treated the random load demands and random disturbances. The study has also been carried out with exciter and speed governor control loops for voltage-dependent load character-istics on stabilizing intersystem oscillations[62].The AGC problem has been investigated using disturbance-accommoda-tion control.It was shown that the optimal accommodation of load disturbances could lead to significantly better performance than that of conventional controllers.It was further shown that the complete cancellation of all disturbance effects in the class of power systems considered is impossible.Nevertheless, the disturbance effects in system frequency can be cancelled completely.
E.Digital AGC Schemes
Since digital control is more accurate and reliable,compact in size,less nsitive to noi and drift,and moreflexible,the re-
archers have focud their attention on proposing digital AGC control schemes[64]–[71].Ross[64]
was probably thefirst to prent a comprehensive direct digital LFC regulator for power systems.Later,the control philosophy and design techniques of a digital LFC incorporating dynamic control criteria for perfor-mance evaluation of digital control system bad onfield test results was outlined by Ross and Green in[65].As the ACE reprenting generation mismatch in an area can be derived in discrete mode by sampling the tie-line power and system fre-quency deviation and then transferring over the telemetering links and unlike in continuous-time system,the control vector in the discrete mode is constrained to remain constant between the sampling instants.Bad on this fact,Bohn and Miniesy [32]have analyzed the effect of the sampling period on the system’s dynamic behavior using a discrete model of a single area power system.An informative work on digital AGC mod-eling,including the criterion for evaluation of system dynamic performance with the help of indices that measure the effec-tiveness of control relative to control efforts,is discusd by Demello and Mills[66].
Kothari and coworkers[23],[70]have studied the AGC in dis-crete mode.The investigations were carried out with more real-istic modeling of AGC ,considering that the system is operating in continuous mode and the controller is operating in discrete mode[23].In[70],discrete mode AGC of an inter-connected power system with reheat thermal plants considering a new ACE is described.Th
e new ACE is derived from tie-line power deviation,frequency deviation,time error,and inadver-tent interchanges.Optimum integral and proportional integral controllers using the concept of stability margin and the ISE technique have been obtained with conventional and new ACEs, and their dynamic performance was compared for a step-load disturbance.
F.Sensitivity Features
An optimal AGC regulator design bad on nominal system parameter values may not really be optimal for the system with parametric variations/uncertainties due to various system oper-ating and environmental conditions,and therefore,the imple-mentation of the regulators on the system may be inadequate to provide the desired system functioning.This could result in a degraded system dynamic performance and sometimes also in the loss of system stability.Therefore,considerable work has also been prented on AGC that considers nsitivities of the system parameter variations[72]–[84].
In the late1960s,a nsitivity study was included in an opti-mization analysis to determine optimal parameter values of con-ventional AGC systems by Van Ness[72].The VSS controllers have an advantage over the controllers bad on the linear op-timal control theory in lecting the values of t
he parameters in many different ways of a VSS controller.Innsitivity to pa-rameter variation can be achieved by designing variable struc-ture AGC regulators.Erschler et al.[76]are probably thefirst to investigate the AGC of hydropower systems using the VSS technique.
It may be noted that the VSS controllers have improved tran-sient respon due to load disturbances in the power system. By properly lecting the parameters of the controller,the fre-quency deviations and tie-line powers effectively can be con-trolled.The rearch publications regarding the design of load frequency controllers for interconnected power systems incor-porating the system parametric uncertainties are reported in the literature[80]–[85].A control technique bad on the applica-tion of linear
好学教育
feedback infinity robust controllers in the power system model to control the frequency deviations was propod by Ismail[80].This approach suggests that the con-troller respon should be fast enough to offt the frequency errors due to load variations.
A robust controller bad on the Riccati equation approach has been propod for the power system by Wang and coworkers [81],[82].Later,bad on a combination of the robust control approach and an adaptive control technique,a design procedure of a new robust adaptive controller was propod for power system load-frequency control with system parametric uncer-tainties.The motivation of combining the robust control with an adaptive control was to u the robust control approach to deal with the small parametric uncertainties[82].The other re-arch contributions on decentralized robust LFC bad on the Riccati equation approach have appeared in[84].The design of decentralized robust LFC applying structured singular values is propod by Yang et al.[85].It has been demonstrated that when the frequency respon-bad diagonal dominance cannot be achieved,the structured singular values can be applied to de-sign decentralized LFC to achieve the desired system dynamic performance[85].
G.Adaptive and Self-Tuning AGC Schemes
Apart from various AGC schemes,adaptive control has been a topic of rearch for more than a quarter of a century.Ba-sically,the adaptive control systems can be classified into two categories,namely,the lf-tuning regulators and the model ref-erence control systems.The task of adaptive control is to make the process under control less nsitive to changes in process parameter
s and to unmodeled process dynamics.A number of articles have been reported on adaptive AGC schemes[86]–[91] and STC schemes[92]–[96]for AGC of power systems.
In1966,Ross[86]described control criteria in LFC and the related practical difficulties encountered in trying to achieve the criteria.The implementation and analysis of an adaptive LFC system on the Hungarian power system has been done by Vajk et al.[88].An adaptive controller using a proportional in-tegral adaptation to meet the hyperstability condition require-ments to take care of the parameter changes of the system was prented by Pan and Liaw[89].A multiarea adaptive LFC scheme for AGC of power systems[90]and a reduced-order adaptive LFC for interconnected hydrothermal power system [91]are reported in the literature.A multivariable lf-tuning controller has been derived by defining a cost function with a term reprenting the constraints on the control effort and then minimizing that with respect to the control vector.Later on,a lf-tuning algorithm for AGC of interconnected power systems was prented by Lee[95].
H.Concepts of AI Techniques:NN,FL,and GA
In practice,many nonlinear process are approximated by reduced-order models,possibly linear,that are clearly related to
the underlying process characteristics.However,the models may be valid only within certain specific operating ranges,and a different model may be required in the wake of changed op-erating conditions,or the control system should adopt the new system model parameters.The advent of AI techniques,such as neural networks,has solved this problem to a great extent. The neural technology offers many more benefits in the area of nonlinear control problems,particularly when the system is operating over the nonlinear operating range.The applications of neural networks in power system control are witnesd in
[97]–[102].
A new AGC scheme to incorporate the nonconforming load problem was prented by Douglas et al.[98],in which an ef-fort had been undertaken to develop algorithms capable of dis-criminating between noncontrollable short-term excursions and controllable long-term excursions.Out of the two techniques de-scribed,one was developed using a neural network algorithm for pattern recognition of controllable signals,and the other tech-nique was bad on the detection of the controllable signal in the prence of a noisy random load using a random signal proba-bility model.Test results reveal that neural network-bad AGC implementation had a significant improvement over the modern AGC implementation.LFC system performance was evaluated with a nonlinear neural network controller u
dotardsing a generalized neural structure to yield better system dynamic performance than the individual neurons[99].
Recently,a four-area interconnected power system model with reheat nonlinearity effect of the steam turbine and upper and lower constraints for generation rate nonlinearity of hydro turbine was considered for the investigation in[101].It has been shown in[102]that the AGC problem can be viewed as a stochastic multistage decision-making problem or a Markov Chain control problem and have prented algorithms for designing AGC bad on a reinforcement learning approach. The fuzzy logic control concept departs significantly from traditional control theory,which is esntially bad on math-ematical models of the controlled process.Instead of deriving a controller via modeling the controlled process quantitatively and mathematically,the fuzzy control methodology tries to es-tablish the controller directly from domain experts or operators who are controlling the process manually and successfully.Re-cently,many studies exploiting the fuzzy logic concept in AGC regulator design dealing with various system aspects have ap-peared in the literature[103]–[105].
More recent contributions considering the problem of de-composition of multivariable systems for the purpo of dis-tributed fuzzy control was reported by Gegov[104].The pro-pod decomposition method has reduced the number of inter-active fuzzy relations among subsystems.The concept and
de-velopment of AGC using ANN and fuzzy t theory to utilize the novel aspects of both in single hybrid AGC system design for power systems has also been mooted[106].
The days,GA is the most popular and widely ud algo-rithm of all the intelligent algorithms.GAs have been widely applied to solve complex nonlinear optimization problems in a number of engineering disciplines in general and in the area of AGC of power systems in particular[106]–[113].In[107],op-timum adjustment of the classical AGC parameters using GAs is investigated.An array of performance indices bad on var-ious functions of error and time is considered for the study.
A reinforced GA has been propod as an appropriate op-timization method to tune the membership functions and rule ts for fuzzy gain scheduling of load frequency controllers of multiarea power systems to improve the dynamic performance [108].The propod control scheme incorporates dead-band and generation-rate constraints also.Later,contrary to the trial-and-error lection of the variable structure feedback gains,a genetic algorithm-bad lection of feedback gains has been advocated for load frequency variable structure controller in[109].The -lection scheme provides an optimal feedback gains lection in the VSC,and the test results show that not only the dynamic performance has been improved,but also,the control effort is dramatically reduced.Karnavas et al.[10
6]have prented a comprehensive study on AGC of an autonomous power system using combined intelligent techniques.
A higher order robust dynamic performance is achieved with LFC designs bad on GA and linear matrix inequalities [112].The desired control parameters have been obtained by coordinating GA with linear matrix inequalitie control toolbox optimization.In[113],a new GA/GA-SA-bad fuzzy AGC scheme of a multiarea thermal generating system is developed. The scheme is capable of evaluating thefitness of GA/hybrid GA-SA optimization by lecting a function like“figure of merit,”which directly depends on transient performance char-acteristics like ttling times,undershoots,overshoots,and time derivative of frequency.The hybrid GA-SA technique yields more optimal gain values than the GA method.
成人英文I.Types of Inter-Ties
The HVDC transmission has emerged on a power scenario, due to its numerous technical and economic advantages,for a large chunk of power transfer over large distances.Besides other applications,the commissioning of an HVDC link in parallel with existing ac links has shown beneficial effects from the point of view of stabilization of the system.
李阳疯狂英语学习方法Considerable attention has been paid to consider the damping effect of the dc system as an area interconnection between ac systems.As far as the system frequency control of power sys-tems interconnected via a dc link is concerned,very few pub-lications have appeared on this topic[114]–[116].An AFRC system on an HVDC transmission utilizing the high-speed con-trol features of a dc system,cooperating with automatic fre-quency control on interconnected ac systems,is developed by Yoshida et al.[114].Later,the effects of an AFRC system on an HVDC transmission to the AFC on ac systems when AFRC is applied to a random load disturbance in a steady state.The frequency improving and reduction effects of the output power of regulating power stations by AFRC are analyzed by a digital computer[115].A new dc AFC system,which applies a mul-tivariable control to the dc system-bad frequency control and capable of controlling the frequencies of the two ac systems op-timally while maintaining their stability,is developed by Sanpei et al.[116].
Considerable rearch work on the LFC of interconnected power systems incorporating ac and dc links is contained in [117]–[121].Investigations on decentralized robust LFC of a
