【作者】 王巍；

【导师】 毛弋；

【作者基本信息】 湖南大学 ， 电气工程， 2011， 硕士

【摘要】 电力工业是国民经济一个非常重要的部门,是现代社会发展的重要动力,随着国民经济的快速发展及人民生活水平的迅速提高,社会对电力的需求量越来越大,同时,电力工业的建设又涉及到大量一次能源消耗和巨额的投资,合理地进行电力系统规划可以获得很大的经济利益和社会效益;相反,电力系统规划的失误将会给国家建设带来无法挽回的损失。在这种背景下,如何提高电力系统规划的水平就成了一个紧迫的任务。输电网络规划是一个十分复杂的组合优化问题,求解非常复杂。将传统启发式方法和数学优化方法应用于求解电网规划问题时,虽然在实际工作中有一定的突破,但是仍存在着维数灾、局部最优和约束条件及目标函数不易处理等问题。蚁群算法是一种人工智能方法,这种方法的主要特征是正反馈、分布式计算以及富于建设性的贪婪启发式搜索的运用。蚁群算法特别适合于整型变量优化的求解,从而为输电网络规划问题的求解开创了一条新的途径。到目前为止,蚁群算法及其各种改进算法已成功应于输电网络扩展规划问题的求解,取得了很好的结果。并且显示出了一定的优越性,是一种很有发展前景的方法。针对传统蚁群算法易陷入停滞和局部最优的缺点,本文对蚁群算法进行了改进。通过构造新的信息素释放函数,加快了信息素在较优路径上的正反馈过程,提高了蚁群算法的搜索效率。在应用阶段,结合单阶段输电网络数学模型,采用直流潮流方程求解网络潮流,并针对输电网络规划的特点对线路编码方式进行了改进,有效改善了蚁群算法在寻优过程中易陷入停滞的现象,文章最后以Matlab为平台,对3个电网规划算例进行了仿真计算,测试结果验证了新型蚁群算法应用于输电网络规划的有效性。更多还原

【Abstract】 Power industry is a vital economic sector and a crucial driving force of modern social development, with the rapid development of the national economy and rising living standards, the electricity demand of society is increasing considerably. Meanwhile, the construction of power industry involves a large number of primary energy consumption and large investment and reasonable power system planning can gain a great economic and social benefits. In contrast, errors in power system planning can cost irreparable loss of national construction, which makes power system planning an urgent objection.Power distribution network planning is a very complex and combinatorial optimization problem, which is demanding in calculation. The traditional heuristic methods and mathematical optimization method is applied to solve the problem of network planning, although some breakthroughs have been made in practical work, the fact that contains disaster of dimensionality, locally optimization, objective function and constraints, and difficulty to deal with target function still exist. Artificial ant colony algorithm is an artificial intelligence, which is characterized by positive feedback, distributed computing, and full of constructive use of the greedy heuristic search. Ant colony optimization is particularly suitable for solving the integer variables, which leads a new solution of power distribution network planning. Until recently, ant colony algorithm and its various improved method have been applied in solving power distribution network expansion planning problem, which has achieved good results. And it also showed some advantages, which is a promising method.Considering its tendency of falling into stagnation and local optimum drawbacks when adopting traditional method, ant colony algorithm is improved in this paper. By constructing the new pheromone release function and speeding up the optimum path of pheromones in positive feedback process, search efficiency of ant colony algorithm is improved. In application stage, with single-stage distribution network model, calculating power flow trend by using DC load flow equations and improvement on line encoding regarding to its characteristics have prevented its tendency of falling into stagnation. In the end of this paper, with three simulation samples, the test results validate the new algorithm for its effectiveness on distribution network planning by using MATLAB as a platform.更多还原