Studies On Generator Start-up Sequence After Power System Major Blackout

With the rapid development of social economy, the electric load and generation capacity is continuously growing. As a result, the network configuration and dynamic behavior of power systems are becoming more complicated than before. Moreover, the integration of large amount of wind power which is of variability and uncertainty would be not favorable for the stability of power systems. All these factors have increased the complexity of power system operations and maintenances significantly. Power systems are more frequently operated to approach their security limits. As a consequence, a great blackout might occur in the case of unpredictable disturbances and inappropriate operations.In contemporary society, the demand for reliable power supply is also becoming more and more intensive than before. The widespread blackout has a serious negative effect on the economy and society. In order to speed up the restoration process and reduce the duration of a blackout, the study of power system restoration and the determination of proper schemes for power system restoration following a blackout is quite necessary. The objective of power system restoration is to restore the load as soon as possible. And the successful start-up of generators is the precondition of load restoration and the basis of power system restoration. The optimal start-up plan of generators can facilitate power system restoration and reduce the duration as well as the cost of a blackout significantly.The optimization of generator start-up sequencing is intended to provide an optimal starting sequence of all non-black-start (NBS) units and optimal restoration paths. The generator start-up sequencing problem is related to many factors and is a complex multi-stage, multi-constraint and nonlinear combinational optimization problem. Based on current research, a detail research on the optimization of the generator start-up sequence after a blackout has been carried out in this dissertation. Many theories, such as multi-objective optimization, preference optimization and complex network theory are applied to solve the generator start-up sequencing problem in the research, and many specific aspects, such as mathematical models, solving algorithms and checking constraints are involved. Main contributions and innovations are described as following:(1) This paper studies the origin and influence of the inrush current in black start field test, and presents a novel method to select the optimal generator started during black start stage in a short time considering the influence of inrush current. During the black start test, the harmonic of inrush current caused distortion of the terminal voltage of the black start unit, and had a negative influence on the excitation system of black start unit with self-shunt excitation. Based on the analysis of inrush current, a transformer index is introduced to indicate the severity of the inrush current. The analysis of charging no-load transmission lines is also carried out and a line index is proposed to reflect the overvoltage level and the risk of self-excitation when charging transmission lines. Similarly, the start-up of auxiliaries is also analyzed and the under-frequency deviation and voltage dip are indicated by an auxiliary motor index when starting auxiliaries. Based on the indexes, the success rate of generator start-up is defined to evaluate the success possibility of starting different generators firstly. A mathematical model whose objective is to maximize the success rate of generator start-up is proposed, and the model is solved promptly by preselection of units and restoration paths. The simulation results of Shandong power grid demonstrate that the proposed method can select the optimal generator and restoration path in a short time. The comparison of checking results of different plans validates the effectiveness of the success rate of generator start-up.(2) In order to coordinate the problems of the generator start-up sequence and restoration paths interaction, a preference multi-objective model which can optimize the start-up sequence and restoration paths of NBS units simultaneously is proposed during network reconfiguration stage. The optimization model consists of three optimization objectives which are to maximize the total MWh output, the reliability of restoration paths and the importance of the restoration paths. The final goal of system restoration is to restore the load as soon as possible, and the first objective can facilitate load restoration. Therefore, the first objective is regarded to be more important than the other two objectives. The difference of the objective importance is also considered in this model. The preference information is expressed as a reference point in an a priori way. The r-dominance which takes its origin from hybridization between the Pareto dominance principle and the reference point method is introduced, and a variant of nondominated sorting genetic algorithm II, r-NSGA-II which incorporates the r-dominance concept is applied to search for the Pareto optimal solutions according to the reference point. Then, a decision-making method which combines the analytic hierarchy process and preference ranking organization method for enrichment evaluation is applied to sort the Pareto optimal solutions and to determine the final solution. The simulation results of Shandong power grid demonstrate that the proposed model can obtain the Pareto optimal start-up sequences and restoration paths of NBS units simultaneously. The r-NSGA-II is effective to express the preference information of the preference multi-objective model and select the preference optimal solutions from the Pareto optimal solution set. The scale of Pareto optimal solution set is also convenient to be controlled.(3) During the restoration process, the optimization objectives of generator start-up sequencing problem vary with the system conditions. In order to reduce the time cost to optimize generator start-up sequence and obtain practical restoration plans, a two stage optimal generator start-up method is proposed. At the beginning of system restoration process, the restored system is weak. Therefore, the generator with the maximum successful rate of start-up is selected as the first unit to be started for safe restoration in stage I. The main optimization objective of stage II is fast restoration, and the reliability of energizing the restoration path and the importance of the restoration path are also considered. Then, a preference multi-objective optimization model is built. Based on the idea of variable neighborhood search algorithm, a new method is proposed to transform the preference objective into an ordinary one by designing a solution space. A combination of differential evolution and estimation of distribution algorithm for the global continuous optimization problem, namely DE-EDA, is applied to search the optimal solutions in the solution space. The simulation results of typical system as well as a practical power grid demonstrate that the two-stage method can obtain generator start-up plans which conform to the actual requirements better than the plans which are obtained with an ordinary method. The time cost is also reduced. The solving algorithm is effective to solve the proposed optimization model and can reduce the calculation burden.