作者：杜峰, 唐岚, 丁峻强
作者单位：西华大学交通与汽车工程学院, 四川 成都 610039
关键词：信号处理; 功率谱密度; MATLAB软件; PSD函数; PWELCH函数; 工程随机振动
摘要：
针对MATLAB中两个内建功率谱密度计算函数psd()和pwelch()计算结果迥异的现象,在功率谱密度估计理论的基础上,根据经典的周期图理论和Welch平均周期图方法,通过详细分析源程序,解析计算方法,发现psd()计算的并不是工程单边功率谱密度,而是采样信号双边谱,故与pwelch()结果迥异,另外pwelch()不能对分段信号数据进行预处理。就上述不足提出了相应的改进措施,比较验证表明改进措施行之有效。
Improvement and application of PSD and PWELCH
DU Feng, TANG Lan, DING Jun-qiang
School of Transportation and Automotive Engineering, Xihua University, Chengdu 610039, China
Abstract: Because the results calculated by PSD are quit different from those by PWELCH, the source code is analyzed in detail for these two functions which are used to calculate the power spectrum density in Matlab. On the basis of the theory of power spectrum density estimation, it is found with classical cyclic periodogram and Welch average periodogram what function PSD calculates actually is not power spectrum density used on engineering but double-edge spectrum of collected signals. That is reason for the phenomena mentioned at the beginning of the article. On the other hand, segmental signal can not be pre-processed by function PWELCH. Some ways are proposed to improve those issues, and these actions are proved be effective.
Keywords: Signal process; Power spectrum density; Matlab; Function PSD; Function PWELCH; Engineering random vibration
2010, 36(1): 93-96 收稿日期: 2009-8-6;收到修改稿日期: 2009-9-28
基金项目: 四川省教育厅重点科研项目(2004A112);四川省重点学科建设项目(SZD0410-1)
作者简介: 杜峰(1980-),男,硕士研究生,河南邓州市人,专业方向为汽车电子技术。
参考文献
[1] 卢文祥,杜润生.机械工程测试·信息·信号分析[M].2版.武汉:华中科技大学出版社,2006.
[2] 李晓雷,俞德孚,孙逢春.机械振动基础[M].2版.北京:北京理工大学出版社,2007.
[3] 柏文春.随机振动中几个工程计算问题[J].电子工程,2008(1):13-16.
[4] 赵若红,傅继阳,吴玖荣,等.MATLAB内建psd函数在工程随机振动谱分析中的修正方法[J].暨南大学学报:自然科学版,2007,28(5):435-439.
[5] 王 济,胡 晓.MATLAB在振动信号处理中的应用[M].北京:中国水利水电出版社,2006.
[6] 黄治宇,刘保华,陈高平.随机信号的功率谱估计及MATLAB实现[J].现代电子技术,2002(3):21-23.
[7] 高西全.数字信号处理原理、实现及应用[M].北京:电子工业出版社,2006.
[8] 庄表中,梁启德,张佑起.结构随机振动[M].北京:国防工业大学出版社,1995.
[9] 刘会灯,朱 飞.MATLAB编程基础与典型应用[M].北京:人民邮电出版社,2008.
[10] 陈财政,邢动秋.基于MATLAB的加速度传感器振动信号处理方法研究[J].现代电子技术,2007,30(7):134-136.
