基于多重分形去趋势波动分析的高速列车运行状态识别方法  

Approach to recognizing high-speed trains running state using multifractal detrended fluctuation analysis

作　　者：张美兰[1] 金炜东[1] 孙永奎[1] 王江丽[1] 

机构地区：[1]西南交通大学电气工程学院,成都610031

出　　处：《计算机应用研究》2015年第10期2978-2980,共3页Application Research of Computers

基　　金：国家自然科学基金重点资助项目(61134002)

摘　　要：针对高速列车运行中的状态识别问题,提出基于多重分形去趋势波动分析的高速列车状态识别新方法。通过分析发现,高速列车在不同运行状态下的多重分形奇异谱和广义Hurst指数谱均有明显的区别,因此提取多重分形奇异谱参数和广义Hurst指数谱参数作为高速列车运行状态的特征,并使用支持向量机对其状态进行识别。实验结果表明,高速列车在运行速度200 km/h及以上时,状态识别率达到100%。多重分形奇异谱参数和广义Hurst指数谱参数能够有效地描述高速列车的运行状态,为高速列车运行状态的识别提供了一种有效的方法。In order to evaluate in-service performance of high-speed trains,this paper proposed a new approach to recognize the running state of high-speed train using multifractal detrended fluctuation analysis. The analysis revealed that the multifractal singularity spectrum and generalized hurst exponent spectrum of the high-speed train at different running states made a significant distinction. Therefore,it extracted the multifractal singularity spectrum parameters and generalized hurst exponent spectrum parameters as the characteristics of the high-speed train running states,and using support vector machines recognized the states. Experimental results prove that the high-speed train state recognition rate is 100% when the speed is more than 200 km / h,and using multifractal singularity spectrum parameters and generalized hurst exponent spectrum parameters can represent effectively the running state of high-speed train. The proposed method provides an effective method for the recognition of the high-speed train running states.

关 键 词：高速列车 状态识别 多重分形去趋势波动分析 多重分形奇异谱 广义Hurst指数谱 支持向量机 

分 类 号：U279[机械工程—车辆工程] TP391.4[交通运输工程—载运工具运用工程]