基于支持向量机和免疫遗传BP的瓦斯浓度预测研究

发布时间：2018-03-31 16:09

本文选题：瓦斯预测　切入点：支持向量机　出处：《西安科技大学》2017年硕士论文

【摘要】：煤炭占我国一次能源消费比例最重,是我国的主要能源,在我国有着重要的战略地位,因此煤矿安全生产是煤矿的一个重大问题。但是我国煤炭地质构造复杂,煤层瓦斯含量大,煤矿安全事故发生率远远高于世界主要产煤国家,其中瓦斯灾害事故发生频率最高,伤害最大。因此,瓦斯浓度的预测对于煤矿生产安全和职工人身安全意义重大。本文以煤矿安全生产为目的,以现有瓦斯监测技术为基础,结合煤矿井下实际,提出基于支持向量机的瓦斯数据去噪算法和基于免疫遗传BP神经网络的瓦斯浓度数据预测算法,对井下采集瓦斯浓度数据进行去噪和预测研究。本文主要研究工作如下:首先,分析煤矿实际生产中井下瓦斯数据的特点,得出其受井下复杂环境影响普遍含有噪声,因此提出基于最小二乘支持向量机的瓦斯数据去噪算法,对采集到的瓦斯数据进行处理,通过对瓦斯数据的仿真实验,验证所提算法的有效性。其次,针对实际井下瓦斯浓度预测不足的问题,提出基于免疫遗传BP神经网络对瓦斯浓度预测的算法。针对BP神经网络结构难以确定的问题,结合井下瓦斯浓度数据的特点,提出以相空间重构理论为依据的解决方法,通过求取最佳嵌入维数m,确定了BP神经网络的结构;针对BP神经网络存在收敛速度慢及易困入局部极值的缺陷,提出基于免疫遗传理论的优化算法,将BP神经网络权值和阈值作为待求解问题(抗原),产生初始抗体种群,通过引入免疫遗传机制,提高算法运行效率,来克服BP神经网络易困入局部极值的缺陷。通过对煤矿井下瓦斯浓度数据的仿真实验,验证所提算法的有效性。最后,将本文算法分别应用于搭建的采煤工作面瓦斯采集系统和本校的煤矿瓦斯监测系统,对算法的实用性进行验证。此外,还研究井下人员定位和无线通讯系统,以满足煤炭生产的需要。
[Abstract]:Coal accounts for the heaviest proportion of primary energy consumption in China and is the main energy in our country, and has an important strategic position in our country. Therefore, the safety of coal production is a major problem in coal mine. However, the geological structure of coal in China is complex. The coal seam gas content is large, the coal mine safety accident rate is far higher than the world main coal producing country, the gas disaster accident occurrence frequency is the highest, the harm is biggest. The prediction of gas concentration is of great significance to the safety of coal mine production and the personal safety of workers. This paper proposes a gas data denoising algorithm based on support vector machine and a gas concentration prediction algorithm based on immune genetic BP neural network. Based on the analysis of the characteristics of underground gas data in coal mine production, it is found that there is generally noise in the underground gas data under the influence of complex underground environment. Therefore, a gas data de-noising algorithm based on least square support vector machine is proposed to deal with the collected gas data. Through the simulation experiment of gas data, the validity of the proposed algorithm is verified. Secondly, aiming at the problem of insufficient gas concentration prediction in actual underground, This paper presents an algorithm for predicting gas concentration based on immune genetic BP neural network. In view of the problem that BP neural network structure is difficult to determine and considering the characteristics of underground gas concentration data, a solution based on phase space reconstruction theory is proposed. The structure of BP neural network is determined by obtaining the best embedding dimension m, and the optimization algorithm based on immune genetic theory is proposed to solve the problems of slow convergence speed and easy entrapment in local extremum of BP neural network. The weight and threshold of BP neural network are taken as the problem to be solved (antigen), and the immune genetic mechanism is introduced to improve the efficiency of the algorithm. To overcome the defect that BP neural network is easily trapped into the local extremum. Through the simulation experiment of gas concentration data in coal mine, the validity of the proposed algorithm is verified. Finally, The algorithm is applied to the coal face gas collection system and the coal mine gas monitoring system, and the practicability of the algorithm is verified. In addition, the underground personnel positioning and wireless communication system are also studied. To meet the needs of coal production.
【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;TD712

【参考文献】