Look-ahead Voltage Stability Analysis And Control For Large-scale Power Systems

The fast development of interconnected power grids makes the voltage stability analy-sis and control of large-scale power systems more and more important. This dissertationinvestigates the development of the numerical analysis in power system studies and pre-sents a fast method for power flow calculations. Considering the varying load and the con-trol device, a new look-ahead voltage stability analysis and control method is proposed inthis dissertation. The main works and contributions are summarized as follows:(1) A Newton-FGMRES method is proposed for solving the power flow equations andthree accelerating schemes, including a hybrid scheme, a partial preconditioner updatescheme, and an adaptive tolerance control scheme, is developed to further improve the per-formance of the Newton-FGMRES method. Numerical results show the advantages of theproposed fast Newton-FGMRES method as opposed to the traditional Newton-GMRESmethod in terms of the convergence characteristics and computation time. For the robust-ness of the proposed fast Newton-FGMRES method, numerical results also show the supe-riority of the proposed method under different loading conditions.(2) A computational procedure to numerically construct local bifurcation boundaryand steady-state security boundary of large electric power system models. The local bifur-cation boundary and steady-state security boundary play important roles in power systemoperations and planning. Local bifurcation boundary defines the physical limitations ofpower systems while steady-state security boundary defines the operational limitations ofpower systems. Numerical results show that operational and engineering constraints areviolated before bifurcations occur, i.e. steady-state security boundary is inside its corre-sponding local bifurcation boundary. The impact of contingencies on the local bifurcationboundary and on the steady-state security boundary are presented. The convexity propertyof these boundaries is also observed. The convex property of these two boundaries is pro- nounced if the traditional treatment of Q limit; i.e. non-Q-limit or hard-Q-limit is made.However the effect of generator capability curve, capability-Q-limit, leads to the non-con-vexity of local bifurcation boundary and steady-state security boundary.(3) The validity of the traditional post-contingency PV curve is investigated. Based onthe look-ahead voltage stability analysis, a new method to evaluate the exact post-contin-gency PV curve is presented. Two strategies is applied to the proposed method to evaluatemore points on the curves. Numerical results show that the traditional post-contingency PVcurve can not reflect the change in look-ahead voltage stability analysis and its results isalways conservative.(4) Considering the local logical action of voltage control devices, a look-ahead volt-age stability analysis method is presented. The reason why the contingency ranking methodbased on sensitivities can not captures the most severe contingencies is studies and a newmulti-level contingencies ranking method is proposed. Numerical results show that the pro-posed method can be more effective in look-ahead voltage stability contingency analysis.(5) The voltage preventive control problem and voltage corrective control problem arestudied. Based on the varying load and the reactive power optimization problem, a newmodel of the optimization problem for the look-ahead voltage control is introduced. In thismodel, the constrains that ensure minimal load margin in each time interval is considered.The corrective-preventive interior point method is used to solve the proposed optimizationproblem. Numerical results show the effectiveness of the proposed method.(6)The look-ahead voltage stability with uncertainties is studied. With the develop-ment of renewable energy resources, there will be more renewable energy resources in thefuture and more uncertainties will be brought to power systems. Although the short-termload forecast can be very accurate, we still can not guarantee that loads would follow theresults of forecast. This paper considers uncertainties of both renewable energy resourcesand loads and present a method to evaluate load margins under these conditions.