基于GA-LSSVR的煤矿瓦斯数据去噪研究

发布时间：2018-04-08 09:50

  本文选题：瓦斯浓度　切入点：数据去噪　出处：《矿业安全与环保》2017年01期

【摘要】：针对煤矿瓦斯数据普遍含有噪声的问题,提出一种基于遗传算法优化的最小二乘支持向量回归机(GA-LSSVR)的数据去噪算法。LSSVR通过求解只含一个等式约束的二次规划问题来求得最优解,从而改进了小波去噪局部最优的缺点。但LSSVR也存在收敛速度慢的缺点,通过遗传算法(GA)优化LSSVR,以提高算法的收敛速度。首先,对某煤矿的瓦斯浓度时间序列进行异常数据和缺失数据的处理,然后用GA-LSSVR建模训练。仿真实验结果表明,与小波去噪方法相比,GA-LSSVR能有效去除噪声,并且能够避免数据失真,把有效信号分离出来,经过计算,GA-LSSVR能将输入输出均方根误差降低0.002 94,相对降低了34.59%,去噪效果较好;与LSSVR方法相比,GA-LSSVR能明显缩短程序运行时间,可提高运行效率。
[Abstract]:Aiming at the problem that coal mine gas data generally contain noise, a data denoising algorithm named GA-LSSVRbased on genetic algorithm optimization is proposed. LSSVR solves the quadratic programming problem with only one equality constraint to obtain the optimal solution.Thus, the shortcomings of local optimal wavelet denoising are improved.However, LSSVR also has the disadvantage of slow convergence speed. Genetic algorithm (GA) is used to optimize LSS VRR to improve the convergence speed of the algorithm.Firstly, the abnormal data and missing data are processed in the time series of gas concentration in a coal mine, and then the GA-LSSVR modeling training is used.The simulation results show that GA-LSSVR can effectively remove noise, avoid data distortion and separate the effective signal compared with wavelet denoising method.GA-LSSVR can reduce the root mean square error (RMS) of input and output by 0.002, and reduce the RMS error by 34.59. Compared with the LSSVR method, GA-LSSVR can significantly shorten the running time and improve the running efficiency.
【作者单位】： 西安科技大学电气与控制工程学院;河南科技学院信息工程学院;
【分类号】：TD712

【相似文献】

中国期刊全文数据库 前2条

1 赵清艳;张超;;基于GA-LSSVR的铣削加工变形预测[J];组合机床与自动化加工技术;2011年12期

2 ;[J];;年期

中国重要会议论文全文数据库 前1条

1 郭志明;赵春江;陈立平;黄文倩;;基于GA-LSSVR的烟草尼古丁含量的近红外光谱分析[A];中国农业工程学会电气信息与自动化专业委员会、中国电机工程学会农村电气化分会科技与教育专委会2010年学术年会论文摘要[C];2010年

，

本文编号：1721090