Reliability Optimization Research On Spinning Reserve Of Bulk Power System And Distributed Generation Of Distribution System

This dissertation is supported by National Natural Science Foundation of China(50977094) and Natural Science Foundation Project of CQ CSTC (2011BB6047).As power system is increasingly large and complex, many uncertainties pose athreat to system safe and reliable operation. Consequently, the security and stabilityproblems of power system are becoming more prominent. Allocation of power source isoptimized to get maximum benefit with limited conditions through the optimalcombination and distribution. This study research on the allocation optimization ofspinning reserve(SR) in bulk power systems and the allocation optimization ofdistributed generation(DG) in distribution systems，from the perspective of probabilisticrisk assessment. Both SR and DG can provide a certain amount of subsequent powersupply capacity and ensure the reliability when power system is disturbed.This paper is divided into two parts，the main research contents include thefollowing aspects：①To improve the operational reliability of bulk power systems, the operationalconstraints of SR are embedded into the conventional optimal load shedding modelincorporating unit types, ramp rates, component reliability parameters, networkconfiguration and transmission capacity limits. Moreover, a sensitivity formula of theexpected demand not supplied (EDNS) with respect to the capacity of SR is deducedbased on the actual physical meanings of Lagrange multipliers in the inequalityconstraints. Therefore utilizing the ranking of the sensitivity indices, a heuristic idea isproposed to determine the optimal allocation of SR.②The optimal capacity and allocation scheme of SR are identified to minimize thetotal cost (the sum of SR cost and customer interruption cost) combining with theoptimal allocation model of spinning reserve. Finally, the RBTS and IEEE-RTS79power systems are evaluated and the simulation results verify its validity.③To obtain the reliability indexes of load points and distribution system, a methodwhich combines the optimal island searching with the reliability evaluation ofdistribution system is proposed, according to the difference of importance of load. Anoptimized islanding scheme can be formed based on the actual running situation，failurelocation and DG output model while the fault occurs. A mathematical model of allocation optimization of distributed generation is established based on the powersupply reliability of distribution network.④To improve the computational efficiency，a new state variable coding methodaccording to the actual situation is proposed. Simultaneously，to remedy the prematureconvergence of particle swarm optimization (PSO) algorithm and the slow convergencespeed of artificial fish school algorithm (AFSA), the particle and artificial fish hybridalgorithm is improved with adaptability, which enhances the speed and precision ofcomputation for optimizing the siting and sizing of DG in distribution system.Simulation results of the main feeder F4via IEEE-RBTS Bus6system verify itsvalidity.