Intelligent Robust Load-frequency Control Strategies For An Electric Power Interconnection

Load-Frequency Control(LFC)has been considered to be an important part of Automatic Generation Control(AGC)for a large-scale interconnected power system.It is the fact that loads in an electric power grid depending only on users are continually and randomly changeable,effecting strongly the system frequency and tie-line power of such an electric power interconnection.As a result of this phenomenon,it is essential to design an effective LFC strategy,whose objectives are mainly to maintain the network frequency at its nominal value and the tie-line power flow at a scheduled demand.It means that such effective LFC controllers must be able to quickly damp both dynamic fluctuations of the system frequency and tie-line power flow;thereby the stability and reliability of the network can be assured successfully.In order to seek a viable solution for the LFC problem,tie-line bias-based control strategy has been widely applied to design two typical types of LFC controllers,including conventional and intelligent regulators.The conventional controllers such as Integral,Proportional-Integral and Proportional-Integral-Derivative ones are adopted as initial LFC schemes.Resulting from these traditional LFC controllers,an electric power interconnection may only obtain highly poor performances such as large overshoots and long settling times;thus the control goals cannot be reached in this case.As several promising candidates for replacing the conventional LFC regulators,a number of intelligent controllers applying modern control techniques,e.g.,Artificial Neural Network(ANN),Genetic Algorithm(GA),Fuzzy Logic(FL),Particle Swarm Optimization(PSO),have been taken into account recently.The design rule of such intelligent controllers is basically dependent on the working mechanism of human being.This leads to their outstanding control features,such as high flexibility,efficiency and adaptability.Consequently,they have been commonly chosen as feasibly intelligent candidates to solve many control problems of nonlinear-complicated systems,especially for the LFC of a multi-area electric interconnection.Such intelligent controllers are built depending on the principle of inheritance and integration.Particularly,the basic intelligent control strategies which employ the basic FL load-frequency controllers are offered at the beginning.After that,the weaknesses of these fundamental FL architectures concerning the determination of the membership functions,rule base together with scaling factors will be overtaken by utilizing the optimization techniques(GA,PSO)to design adaptive FL load-frequency controllers.Another powerful way to further enhance the control characteristics of the basic FL load-frequency controllers is to coordinate with the ANN-based regulators.In this way,a hybrid controller will also be designed to solve effectively the LFC issue.Moreover,in dealing with the effect of nonlinearities and uncertainties to the LFC,Sliding Mode Control(SMC)strategy is investigated in this work.Specifically,some of the main contributions of this work are as follows.First,a mathematical model of electric power interconnections consisting of n control-areas(CAs)would be established,and it depends on the tie-line bias control methodology.This model can be considered to be a typical case study of a complex interconnected power system since it is practically characterized by various types of turbines as well as generation units.Although such a model might be linearized and does not contain any uncertainties,it is able to be adopted as an efficient candidate to demonstrate the accuracy and feasibility of most LFC schemes proposed in this study.Secondly,fundamental intelligent LFC strategies based on FL and ANN techniques will be proposed as the next contribution of this work.Two basic FL-based LFC controllers(for example,PI-and PD-type FL architectures)with the properly certain rule base are studied initially.Thereafter,two ANN-based LFC strategies including NARMA-L2 and MRAC are also presented.These LFC schemes are verified to obtain the better control performances in comparison with three conventional LFC regulators(I,PI and PID).However,the optimization of their membership functions,rule bases and scaling factors of the two FL-based LFC architectures still needs to be investigated to adapt to the complexity of a practical large-scale power system.Moreover,the integration of FL and ANN technique to design a hybrid intelligent LFC strategy may be of interest to further enhance the control features in dealing with the LFC problem.Thirdly,a novel hybrid LFC controller based on an effective integration of ANN and FL techniques is proposed as the following contribution.In particular,this novel control strategy includes two consecutive phases:MRAC-based model(ANN stage)and PD-type FL-based architecture(FL stage).Due to the two-stage operating mechanism,the proposed hybrid LFC controller is able to utilize the outstanding advantages of both the FL and ANN techniques in order to satisfy effectively the requirements of the LFC strategy.In addition,Superconducting Magnetic Energy Storage(SMES)devices which are working efficiently for the frequency stabilization will be employed to considerably enhance control quality of the hybrid LFC methodology;thereby a superior LFC architecture in dealing with practical power interconnections can be devised in this study.Together with the hybrid LFC controller,another solution to further improve the control features of the fundamental FL architecture is to design an optimal FL-based load-frequency controller.This new control architecture is an improved model of the basic PI-based FL inference system proposed in the second contribution in combination with the biology-inspired optimization mechanisms(i.e.,PSO and GA)and an online tuning method.In principle,with a reasonably executed mechanism,the PSO and GA algorithms and the online tuning method are employed to optimize a number of parameters of the basic PI-like FL model,including the membership functions,rule base and scaling factors.Under this integration,a novel adaptive PI-type FL controller is set up and it can afford to obtain much better control performances compared to the fundamental counterpart.Such a novel LFC strategy is also the fourth contribution of this work.The following contribution proposes a robust LFC architecture based on the SMC to tackle the nonlinearities and uncertainties of a practical power system.It is found that a practical power interconnection might be usually characterized by a number of the nonlinearities and uncertainties,making the design of control methodologies highly challenging.The SMC,which is typically adopted to deal with this problem,has been applied for a number of nonlinear and/or uncertain control systems,especially for a modern power network with the LFC.In this study,it will be employed particularly to significantly minimize the effect of some nonlinearities such as Governor Dead-band(GDB)and Generation Rate Constraint(GRC)to the frequency maintenance.The uncertainties are also taken into account in this context.Eventually,the feasibility,effectiveness and superiority of the proposed intelligent control methods will be validated through numerical simulations executed in various cases of multi-CA interconnected power system models and load conditions.It is verified clearly from the simulation results that such intelligent controllers could be employed effectively for the LFC strategy of a practical interconnected power system;thus they are promising control solutions to substitute for the conventional LFC regulators.