Research On Novel Methodolagy Of Power System Preventive Control With Transient Stabiltiy Constraints

Currently stability-constrained preventive control and optimization have beenunder intensive investigation, with much progress being made especially on transientstability preventive control. As yet few (no) scheme has reached a practical level dueto the inherent complexity. In this thesis, several (three) preventive control modelsconstrained by transient stability are constructed step by step, and the methodologyfor solving the models is discussed in detail.Enlightened by the nearly linear relation between critical clearing time (CCT)and generation, a new dynamic security dispatching strategy based on trajectorysensitivity is proposed. It removes non-damaging faults from reliable contingency listand distinguishes the most severe fault and the most advanced machinecorrespondingly;then calculates the least active power generation to be regulated onthis machine according to the approximate linear relation, so as to make the CCT ofthe most severe fault longer than the actual CCT;finally, by simulating along thefault-on and post-fault trajectory, trajectory sensitivities of rotor angels versusgeneration are obtained, which guides the computation of active power output changefor all machines. In the above approach, however, the implication of initial values for trajectorysensitivity is still vague. To overcome this disadvantage, conjugate equation of thesystem representation is derived by applying calculus of variation in optimal control,and a further gradient of the performance index (PI) with respect to (w.r.t.) controlvariables is established-namely, inverse trajectory sensitivity, which measures theaction between the degree of stability and control variable. The kernel of thealgorithm first integrates the differential-algebraic equation to the end timenumerically, and treats the final value of the functional PI w.r.t. state variables as theinitial condition for the conjugate function;thus the gradient of PI w.r.t. controlvariables is computed after solving the conjugate function in reverse order on timeaxis. On this basis another preventive control models are constructed. The systematic model for optimal power flow with transient stability constraints(OTS) developed in this thesis is based on the above inverse trajectory sensitivity.Instead of tackling the difficult problem directly and integrally, the OTS isequivalently converted into two sub-problems of optimization-one is conventionalOPF and the other belongs to optimal control. The two sub-problems are solved in analternatively manner, where the one of optimal control assesses the active powergeneration limit based on the operating point produced by OPF. The outstandingadvantage of the treatment is that the computation burden can be reduced significantlywith better convergence.To fulfill the practical potential of OTS, this dissertation further convertstransient stability constraints into constraints of active power flow along critical lines,so that the OTS can be solved by reliable and ready tools used for OPF. In therealization of the algorithm some techniques are presented. 1. a criterion to cease timedomain simulation as soon as possible that speeds the analysis;2. a conception of"effective contingency" and the method for identifying it;3. a model to process a setof contingences, i.e. a dynamic security-constrained rescheduling method to ensuresystem stability for a set of credible contingencies and to satisfy the economic goal.Last but not the least, it can avoid the possible situation that stability improvement toone single contingency might worsen the dynamic behavior of the remaining oneseven in a never ending process. System modeling issue is not a limiting concern in themethodology. The estimated accuracy suggests that this method will be of practicalvalue in avoiding transient instability.