基于PSO优化LSSVM参数的矿井巷道场强预测

发布时间：2018-02-13 20:23

本文关键词： 矿井巷道电波传输粒子群支持向量机　出处：《西安科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：井下开采过程中,无线矿井通信系统为井下安全生产、管理,抢救救灾,人员定位和追踪等工作的顺利进行提供重要保障。然而,地下作业不同于地面,巷道地质环境非常的复杂、无线电波的传输容易受到矿井环境诸多因素的干扰。这些因素,严重阻碍井下电磁波发展。针对传统的矿井巷道电磁波的传输受地质地质环境的影响严重,建模过程复杂、建模难度大的问题,提出利用最小二乘支持向量机建立矿井巷道场强预测模型,并利用粒子群算法优化最小二乘支持向量机的参数,提高了模型的精准度。首先,本课题研究最小二乘支持向量机理论、矩形矿井巷道无线电传播理论基础上,研究巷道中对电磁波传播的影响因素,包括载波的频率、巷道壁的电参数、矿井巷道横截面的大小、巷道壁倾斜度、巷道壁粗糙程度等,确定支持向量机网络的输入、输出。其次,对于矿井巷道复杂的环境分别建立大、小尺度预测模型。仿真建模初期,先利用MATLAB仿真来构造初期的样本数据,选取了模式衰减、巷道壁粗糙度引起的电磁波损耗、巷道倾斜度引起的电磁波损耗和多径衰落引起的电磁波损耗在内的模型的总和。在研究矿井巷道小尺度变化时,通过相空间重构理论构建衰落信道的变化序列,同时利用构建的序列样本训练支持向量机的网络预测模型并利用粒子群算法优化最小二乘支持向量机的参数。最后,结合我校实际矿井环境对模型进行测试验证。通过利用MATLAB软件完成仿真实验,得出LSSVM能够准确的学习矿井巷道无线信道的变化特性。通过传统的粒子群算法与自适应算法比较可以发现,PSO在优化LSSVM的参数时表现出了一定的优越性。实验结果证明:通过利用粒子群算法优化最小支持向量机的参数,可以提高模型的精准度。
[Abstract]:In the process of underground mining, the wireless mine communication system provides an important guarantee for the safe production, management, rescue and relief, personnel positioning and tracing. However, underground operation is different from the ground. The geological environment of the roadway is very complicated, and the transmission of radio waves is easily disturbed by many factors in the mine environment. The transmission of electromagnetic wave in traditional mine roadway is seriously affected by geological and geological environment, the modeling process is complex, and the modeling is difficult. The prediction model of mine tunnel field strength is established by using least square support vector machine (LS-SVM), and the parameters of LS-SVM are optimized by particle swarm optimization (PSO) to improve the accuracy of the model. Based on the theory of least square support vector machine (LS-SVM) and radio propagation theory of rectangular mine roadway, this paper studies the influencing factors of electromagnetic wave propagation in roadway, including the frequency of carrier wave, the electrical parameters of roadway wall. The input and output of the support vector machine network are determined by the size of the cross section of mine roadway, the slope of roadway wall, the roughness of roadway wall, etc. Secondly, the complex environment of mine roadway is established respectively. Small scale prediction model. In the early stage of simulation modeling, MATLAB simulation is used to construct the initial sample data, and the mode attenuation and electromagnetic wave loss caused by roadway wall roughness are selected. The sum of the models including electromagnetic wave loss caused by roadway inclination and electromagnetic wave loss caused by multipath fading. When studying the small-scale variation of mine roadway, the change sequence of fading channel is constructed by the theory of phase space reconstruction. At the same time, the network prediction model of support vector machine is trained by the sequence samples, and the parameters of least squares support vector machine are optimized by particle swarm optimization. Finally, Combined with the actual mine environment of our school, the model is tested and verified. The simulation experiment is completed by using MATLAB software. By comparing the traditional particle swarm optimization algorithm with the adaptive algorithm, it can be found that LSSVM has some advantages in optimizing the parameters of LSSVM. It is proved that the parameters of minimum support vector machine are optimized by particle swarm optimization. The accuracy of the model can be improved.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;TN011;TD65

【参考文献】