【作者】 王飞；

【导师】 章健；

【作者基本信息】 郑州大学 ， 电力系统及其自动化， 2015， 硕士

【摘要】 电力系统数字仿真具有经济性、便捷性等优点,已成为电力系统设计规划运行的重要工具,并且仿真的结果是电力系统相关决策的重要依据。电力系统负荷模型是数字仿真的基础,各种类型负荷元件数学模型的精确程度直接影响到仿真结果的准确程度,从而影响以仿真结果为依据的决策方案的制定。随着日益突出的电网动态稳定和电压稳定等问题,众多专家学者对负荷模型和参数对系统稳定性的影响进行了深入的研究。电力系统负荷建模及仿真技术越来越需要考虑实际的负荷类型。电力系统中负荷类型多种多样,如何根据实际电力网络的负荷情况选择最优的最能满足负荷特性的负荷模型组合方案将会影响到数字仿真的精度。本文以电力系统采集的负荷数据为研究对象,针对电力系统中负荷类型多种多样,系统综合负荷模型的组合方案不可能一一列出的问题,提出了一种通过建立负荷模型库,以代表负荷模型投入情况的0-1逻辑变量为决策变量的最优负荷模型自动选择的方法:根据实际电力系统的负荷结构组合出各种动态等值模型,并且在众多模型组合方案中自动选择出最优的模型组合,并将传统启发式算法引入模型组合选择和参数辨识的优化求解过程。同时,考虑到负荷模型参数辨识过程中优化计算方法选择的不同,引入了自动微分技术。由于在负荷模型自动选择的过程中,模型组合的种类很多,造成微分表达式的手工推导以及微分代码的编写工作非常繁琐,而利用自动微分解决这些问题会显得很高效,程序的开发效率也会显著提高。本文最后以实际电力网络的激励和响应数据为研究对象,通过对实际电网动态等值模型的自动选择,得到所选模型组合能够较好的描述负荷行为,表明电力系统负荷模型自动选择这一方法的可行性和有效性。更多还原

【Abstract】 Digital simulation of power system is economic and convenience, and has become a major for the design, planning and operation in power system. The digital simulation results are the main basis for corresponding decision of power system.Power system load model is the basis for digital simulation, the accuracy of the mathematical model of each component directly affects the simulation results, thus affecting the decision scheme which is based on simulation results. With the increasingly prominent problems of grid dynamic stability and voltage stability,many experts and scholars take in-depth study on the influence that load models and parameters impact on system stability. Actual load types are increasingly needed to consider in loading modeling and simulation technology of power system. The types of load model in power system are various. The establishment of load model in accordance with the characteristics of various load types can effectively improve simulation accuracy. Choosing the best load combination that satisfies the actual grid load characteristic will affect the accuracy of digital simulation.A comprehensive load model considering the distribution network branch is referenced in this paper. Order the load data collected in power system as the research object. The types of load model in power system are various, thus the combinations of power system aggregate load model can not be completely listed. In view of the above problems, through the establishment of load model library, a automatic selection method of the optimal load model is put forward in this paper,which regards the minimum fitting error that represents the proximity of fitting response and practical response as objective function, and put the 0-1 logic variables that represents the investment situation of load model as decision variable. A variety of dynamic equivalence model are composed according to the actual power system structure, and the optimal model combination is automatically selected. Traditional heuristic algorithm is introduced to optimize the solution process. At the same time,considering the different optimization calculation methods in the process of loadmodel parameter identification, the automatic differentiation technique is introduced in this paper. Because of the variety of model combination in the process of load model automatic selection, hand deriving of differential expression and preparation of differential code are very complicated. Using automatic differentiation to solve the problems will be very efficient and the development efficiency of program will be significantly increased.Finally, order the motivation and response data of the actual electricity network as simulation object. The effectiveness and feasibility of the proposed method is verified by the automatic selection of the actual power grid dynamic equivalent model, and the selected model combination can describe the load behavior well.更多还原