高速铁路接触网系统维修决策优化

Optimization of maintenance decision for catenary system of high-speed railway

接触网系统作为电气化铁路的重要组成部分,其安全性和可靠性将直接影响铁路的正常运行。考虑到接触网系统结构复杂,就接触网的7个主要部件进行分析,并认为整个系统是由各部件串联而组成的,随后将电气化铁路中接触网系统的维修方式划分为4类,不同的维修方式下对应着不同的可靠性和维修费用,基于此建立不同维修方式下的可靠性模型和维修费用模型。为提升接触网系统的可靠性并尽可能地减少其维修费用,提出一种多目标布谷鸟-遗传算法来对接触网维修这一多目标优化模型进行求解。该算法采用Tent混沌映射生成多样性和分布性较好的初始种群。为了避免算法早熟,在基本的NSGA2算法基础上加入布谷鸟搜索算法中的levy-flights算子,提高其局部搜索能力。此外,引入优化选择因子Pr,动态调整NSGA2的全局性寻优机制与布谷鸟搜索算法的局部优化机制。为了测试该算法的性能,以实际案例进行分析。仿真结果表明,与基本的NSGA2相比,本文所提多目标布谷鸟-遗传算法的初代种群多样性更优,最优解的精度更高。使用多目标布谷鸟-遗传算法所求解出的最优维修方案可提升接触网系统的可靠性且保持较低的维修费用,达到一个较好的符合工程实际的效果,进一步验证了该算法的可行性和优越性。

As an important part of electrified railway, the safety and reliability of catenary system will directly affect the normal operation of railway. Considering the complex structure of the catenary system, we analyzed the seven main components of the catenary, and believed that the entire system was composed of various components connected in series. Then this paper divided the maintenance methods of the catenary system in the electrified railway into four categories, different maintenance methods correspond to different reliability and maintenance costs. Based on the case, reliability models and maintenance cost models under different maintenance methods were established. In order to improve the reliability of catenary system and reduce its maintenance cost as much as possible, a multi-objective cuckoo-genetic algorithm was proposed to solve the multi-objective optimization model of catenary maintenance. The tent chaotic map algorithm was used to generate initial population with better diversity and distribution. In order to avoid the prematurity of the algorithm, the levy-flights operator in the cuckoo search algorithm was added to the basic NSGA2 algorithm to improve its local search ability. In addition,an optimization selection factor was introduced to dynamically adjust the global optimization mechanism of NSGA2 and the local optimization mechanism of the cuckoo search algorithm. To test the performance of the algorithm, a real case was analyzed. The simulation results show that, compared with the basic NSGA2, the multiobjective cuckoo-genetic algorithm proposed in this paper has better population diversity and higher accuracy of the optimal solution. The optimal maintenance scheme solved by the multi-objective cuckoo-genetic algorithm can improve the reliability of the catenary system and keep the maintenance cost low, and achieve a good effect in line with the actual engineering, which further verifies the feasibility and superiority of the algorithm.