Transient Voltage Stability Research For Shipboard Power System Under Large Disturbance Of Impact Load

As the development and application of electric science technology, the capacity of Shipboard Power System (SPS) is continuing to be larger. All kinds of drive, monitor control and protect equipments have realized electrical automation, demanding more and more power. Meanwhile, the capacity of single load has also become larger. As a result, voltage drops of SPS caused by impact load connecting often occur. Thus, the research of transient voltage stability for SPS has become more and more important. Based on the summarization of previous work, this dissertation conducts systematical research work on several key aspects of transient voltage stability under large disturbance of SPS.Characters of scenarios on losing transient voltage stability of SPS under large disturbance are analyzed. Furthermore, it is concluded that the research on transient voltage stability of SPS is very special. Based on the characters and analyses, math model is built for SPS. Then, numerical simulation on transient voltage stability under large disturbance is executed by utilizing time domain analysis in transient stability analysis in electric power system. Finally, by using a SPS example, simulation demonstrates math model and numerical integration method correct. It is concluded that this method could be basic and validating methodology for further research of transient voltage stability.Time varying character analysis method for judging SPS on the transient voltage stable edge or not is proposed. This method reviews the main conclusions of character analysis in small disturbance analysis. Then, system status equations are expended according to Taylor series at each epoch. After that, liberalized Jacobian expression is deduced at the epoch. Meanwhile, small disturbance analysis for the linear equations is executed. By calculating all the eigen values and left and ri ght eigen vectors and dynamic participating factors of eigen vectors corresponding with all status variables, theory criterion for judging SPS on the transient voltage stable edge or not is concluded. Finally, by utilizing a SPS simulation example, the effectiveness of the methodology is proved.A methodology of calculating Controlling Unstable Equilibrium Point (CUEP) on SPS transient voltage stability boundary based on BP neural network is presented. First, high dimensions non-linear mapping relationship between status variables at different fault cut times and at stable state is built according to electric power system stability boundary theory. Second, transient stability numerical simulations under different large disturbance cut time are executed for times to record data of status variables at different fault cut times and at stable state. Data of status variables at different fault cut times is regarded as input sample and data of status variables at stable state is regarded as output sample, respectively. Furthermore, all the data is used to train BP neural network. Third, the algorithm results CUEP by inputting variables at critical cut time, which is calculated by the intersection point of fault trajectory and stable manifold into the trained BP neural network. Finally, the methodology is demonstrated correct and effective by IEEE39nodes system. Meanwhile, the novel method is more advanced and practical compared with traditional Newton Method.A fast reconfiguration method based on DNSPPSO algorithm after losing trainsient voltage stability for SPS is proposed. The methodology converts emergency control problem after large disturbance of SPS into a shipboard power system reconfiguration problem. Furthermore, a fast and novel particle swarm optimization algorithm is presented based on the security and real-time demand of SPS. This algorithm combines small population technology and dynamic neighborhood technology together. Small population technology could reduce population size greatly. Besides, the technique regenerate all the particles with retaining global and personal best finesses and positions within the search space every few iterations. At the same time, dynamic neighborhood technique can overcome the shortcomings of weight method effectively when handling multi-objectives optimization problem. Meanwhile, dynamic neighborhood technique could result the final solution directly according to specified mission without pick is out from Pareto front. Numerical simulation is run through an8-bus SPS example to demonstrate the effectiveness of this algorithm by comparing with SPPSO and DPSO algorithm.