梯形舰船电力系统网络重构模型及算法研究

发布时间：2018-05-05 10:26

本文选题：舰船电力系统 + 网络重构　；参考：《上海交通大学》2015年硕士论文

【摘要】：舰船电力系统网络重构是舰船能源管理中非常关键的部分,也是系统故障恢复及提高舰船生命力的重要途径。在舰船电力系统中,网络重构的作用是在故障发生并被自动隔离之后,找到需要恢复供电的负荷或区域,重新调整网络中各支路开关的开闭状态,在所有可能的重构策略中,快速的找出一套既能使目标函数最大化,又能满足网络各运行约束的重构方案。在网络重构策略生成过程中有两个关键子问题:重构策略的评估及重构优化算法。潮流计算是重构方案评估的主要内容及重要基础。随着舰船电力系统网络拓扑结构的日益复杂化及网络容量的日益增加,越来越多的国家采用多电站梯形船舶电网的结构,但其调度算法却非常复杂。故进行针对梯形舰船电力系统中适用的潮流计算及网络重构模型及其优化算法的研究是该领域的一项研究热点及难点。本文首先对舰船电力系统的网络重构问题进行了描述,分析了重构问题为多目标、多约束、非线性的整数规划问题的本质。随后介绍了梯形舰船电力系统的结构特点与优势及舰船电力系统中的各设备元件的简化数学模型。之后,本文建立了完整的梯形舰船系统网络重构的数学模型。其次,本文介绍了几种经典的潮流计算方法及其具体的算法逻辑。随后针对梯形船舶电网,提出了一种改进的并行潮流计算算法,将梯形舰船电力系统进行分层等效成供电主网与配电子网,并给不同的子网络分配不同的计算单元,针对供电网和配电网的网络结构特点分别采用节点电势法和前推回推法迭代计算。仿真实验验证了其有效性。随后,本文简要介绍了两种基本的重构优化算法,并尝试使用动态规划算法中的有背包约束的0-1背包问题对其求解。此外,本文还改进了粒子群算法在梯形船舶电网重构问题中的应用,针对系统中存在部分双路供电负荷的情况引入了多级编码方式,降低了粒子的维数,并将多级编码方式推广至多条供电路径同时存在的情况,并针对多级编码对粒子群算法的离散化过程进行了改进。本文通过典型的多电站梯形舰船电力系统仿真实例对的改进算法进行了仿真验证,仿真结果验证了该算法拥有较强的寻优能力和较快的收敛速度。最后,本文还给出了基于所提出的综合分布式并行潮流计算算法和离散粒子群网络重构算法的仿真平台设计框架和主要实现技术。本文结合其图形交互界面给出了详细的系统输入模块的数据定义,介绍了网络拓扑结构分析模块、潮流计算模块及网络重构模块具体的实现逻辑细节,并结合几个具体的算例分别验证了各功能模块的有效性。
[Abstract]:Network reconfiguration of naval power system is a very important part of ship energy management, and it is also an important way to recover system fault and improve ship vitality. In warship power system, the function of network reconfiguration is to find the load or area that needs to be restored after the fault occurs and is automatically isolated, and to readjust the open and close states of branch switches in the network, and in all possible reconfiguration strategies. Quickly find out a reconstruction scheme which can maximize the objective function and satisfy the network running constraints. There are two key sub-problems in network reconfiguration strategy generation: reconfiguration strategy evaluation and reconfiguration optimization algorithm. Power flow calculation is the main content and important foundation of reconfiguration scheme evaluation. With the increasing complexity of the network topology and the increasing capacity of ship power system, more and more countries adopt the structure of multi-power station trapezoidal ship network, but its scheduling algorithm is very complex. Therefore, it is a hot and difficult point to study the power flow calculation and network reconfiguration model and its optimization algorithm in trapezoidal ship power system. In this paper, the problem of network reconfiguration of ship power system is described, and the essence of integer programming problem with multi-objective, multi-constraint and nonlinear is analyzed. Then it introduces the structural characteristics and advantages of trapezoidal ship power system and the simplified mathematical model of each equipment component in ship power system. After that, a complete mathematical model of network reconstruction of trapezoidal ship system is established. Secondly, this paper introduces several classical power flow calculation methods and their specific algorithm logic. Then an improved parallel power flow calculation algorithm for trapezoidal ship power network is proposed. The trapezoidal ship power system is stratified as the main power supply network and the distribution electronic network, and different computing units are assigned to different sub-networks. According to the network structure characteristics of power supply network and distribution network, the node potential method and the forward push back method are used to calculate the network structure. Simulation results show its effectiveness. Then, this paper briefly introduces two basic reconstruction optimization algorithms, and tries to solve them by using the 0-1 knapsack problem with knapsack constraints in dynamic programming algorithm. In addition, this paper also improves the application of particle swarm optimization algorithm in the problem of grid reconfiguration of trapezoidal ships. A multilevel coding method is introduced to reduce the dimension of particles in the case of partial dual power supply loads in the system. The multilevel coding method is extended to the simultaneous existence of multiple power supply paths, and the discretization process of particle swarm optimization algorithm is improved for multilevel coding. The improved algorithm is verified by a typical simulation example of trapezoidal warship power system of multi-power station. The simulation results show that the algorithm has strong optimization ability and faster convergence speed. Finally, the design framework and main implementation techniques of the simulation platform based on the proposed distributed parallel power flow algorithm and discrete particle swarm optimization network reconstruction algorithm are presented. In this paper, the detailed data definition of the system input module is given based on the graphical interface, and the logic details of the network topology analysis module, the power flow calculation module and the network reconstruction module are introduced. The validity of each functional module is verified by several concrete examples.
【学位授予单位】：上海交通大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U674.703.3

【相似文献】