蜻蜓算法的改进及在甘蔗收获机中的应用

发布时间：2018-05-02 01:33

本文选题：蜻蜓算法 + 甘蔗收获机　；参考：《广西民族大学》2017年硕士论文

【摘要】：目前,甘蔗机械化收割过程中存在着堵塞严重、破头率高、切割质量差等问题,进而导致来年甘蔗宿根的发芽率偏低,极大影响了甘蔗的产量和甘蔗机械化的推广。其中,剥叶系统性能的好坏是影响堵塞的关键所在,而切割器是直接影响甘蔗宿根切割质量的关键部件,为实现对剥叶性能的优化,以及探究复杂因素对刀盘轴向振动及切割质量的影响规律并实现对切割质量的预测与控制,本文利用智能优化算法的自适应性、容错性以及强鲁棒性以提高甘蔗收获机剥叶系统的性能和对刀盘切割系统的切割质量预测。蜻蜓算法是一种新型群体智能优化算法,该算法源于自然界中蜻蜓捕食和迁徙的群体行为,通过模拟蜻蜓群体飞行、捕食、躲避外敌等过程进行全局搜索和局部搜索,从而实现优化的目的。该算法具有结构简单、易于实现、搜索性能优异且鲁棒性强等特点,然而在解决一些复杂的优化问题时存在收敛后期易陷入局部最优的缺陷,一定程度上影响了对甘蔗收获机的结构性能的优化。为提高甘蔗收获机剥叶系统的性能以及切割质量预测的精度,本文针对基本蜻蜓算法存在的不足进行分析和改进,并将改进后的算法应用于解决甘蔗收获机的优化问题。本文的主要工作如下:(1)引入精英反向学习策略,在保证种群多样性的同时,扩大了搜索区域的范围,同时,在迭代中对蜻蜓个体的位置更新利用指数函数步长来代替原始的线性步长,有效提高了算法的收敛速度,从而增强了算法的全局勘探能力和收敛速度。(2)利用上述改进后的算法实现对剥叶系统的PID控制器参数优化,解决了传统PID参数优化方法易出现费时、震荡且不能保证所调参数最优的问题,同时通过PID控制实现了剥叶和输送工序的速度匹配问题,有效缓解了阻塞问题。(3)为解决传统预测方法对刀盘振动预测精度低、参数选取盲目等问题,提出一种基于蜻蜓算法支持向量机的甘蔗收获机刀盘振动及性能的预测模型。该方法利用蜻蜓群体寻优的过程实现对支持向量机参数的优化,并将优化后的支持向量机对刀盘振动进行预测,实验数据表明,基于蜻蜓算法的支持向量机预测模型具有更高的预测精度和泛化能力,有效地实现了对甘蔗收获机刀盘振动的预测。(4)引入免疫选择算子,利用免疫选择操作对蜻蜓种群进行更新,能够有效抑制算法在收敛过程中易出现的早熟停滞现象,以提高其全局寻优能力和寻优精度。然后利用优化后的蜻蜓算法优化支持向量机的训练参数,并将优化后的支持向量机对甘蔗收获机的宿根切割质量进行预测,仿真结果表明利用改进后的蜻蜓算法优化的支持向量机具有更优的预测性能。
[Abstract]:At present, there are some problems in the process of sugarcane mechanized harvesting, such as serious blockage, high head breaking rate and poor cutting quality, which leads to the low germination rate of sugarcane roots in the coming year, which greatly affects the yield of sugarcane and the popularization of sugarcane mechanization. Among them, the performance of the leaf-stripping system is the key to affect the clogging, and the cutter is the key component that directly affects the cutting quality of sugarcane root. And to explore the influence of complex factors on the axial vibration and cutting quality of the cutter head, and to realize the prediction and control of the cutting quality, this paper makes use of the self-adaptability of the intelligent optimization algorithm. Fault tolerance and strong robustness are used to improve the performance of the leaf stripping system of sugarcane harvester and to predict the cutting quality of the cutter cutting system. Dragonfly algorithm is a new kind of swarm intelligence optimization algorithm, which originates from the swarm behavior of predation and migration of dragonflies in nature. The global search and local search are carried out by simulating the flight, predation and avoidance of foreign enemies of dragonflies. In order to achieve the purpose of optimization. The algorithm has the advantages of simple structure, easy to implement, excellent search performance and strong robustness. However, there are some defects in solving some complex optimization problems, such as the local optimum is easy to fall into in the later stage of convergence. To some extent, it affects the optimization of the structure and performance of sugarcane harvester. In order to improve the performance of leaf stripping system of sugarcane harvester and the precision of cutting quality prediction, this paper analyzes and improves the shortcomings of the basic dragonfly algorithm, and applies the improved algorithm to solve the optimization problem of sugarcane harvester. The main work of this paper is as follows: (1) the elite reverse learning strategy is introduced, which not only ensures the diversity of the population, but also expands the scope of the search area, at the same time, In the iteration, the exponential function step size is used to replace the original linear step to update the position of individual dragonfly, which effectively improves the convergence speed of the algorithm. Thus, the global exploration ability and convergence rate of the algorithm are enhanced. The improved algorithm is used to optimize the parameters of PID controller of the leaf-stripping system, and the traditional PID parameter optimization method is easy to take time. In order to solve the problem of vibration prediction of cutter head by traditional prediction method, the speed matching problem of blade stripping and conveying process is realized by PID control, which effectively alleviates the blockage problem. This paper presents a prediction model of vibration and performance of sugarcane harvester based on dragonfly support vector machine (SVM). In this method, the parameters of support vector machine are optimized by using dragonfly population optimization process, and the optimized support vector machine is used to predict the vibration of cutter head. The experimental data show that, The prediction model of support vector machine based on dragonfly algorithm has higher prediction precision and generalization ability, and the immune selection operator is introduced to predict the vibration of cutter head of sugarcane harvester effectively. Using immune selection to update the dragonfly population can effectively suppress the premature stagnation in the convergence process of the algorithm in order to improve its global optimization ability and optimization accuracy. Then the optimized dragonfly algorithm is used to optimize the training parameters of support vector machine, and the optimized support vector machine is used to predict the cutting quality of sugarcane harvester. The simulation results show that the support vector machine (SVM) optimized by the improved dragonfly algorithm has better prediction performance.
【学位授予单位】：广西民族大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S225.53;TP18

【参考文献】