基于经验模态分解的井眼防碰特征信号识别

在采用钻头振动波方法进行丛式井防碰监测时,为了准确识别井下钻头振动信号,将Hil-bert-Huang变换中的经验模态分解(EMD)应用于钻头信号的特征提取。采用EMD方法可将复杂环境下的钻头振动加速度信号分解成固有模态分量,通过分析各分量的归一化峭度及能量分布,得到包含钻头冲击振动特征的4个IMF分量;对相应的特征IMF分量进行功率谱分析,得到钻头钻进时特征信号的频域范围。当钻头趋近邻井套管时,信号特征分量将发生显著变化,通过海上丛式井防碰的现场试验数据分析验证了该方法的有效性。

Study on Feature Extraction Method for Wellbore Anti-collision Based on EMD

A new method of vibration signal analysis of drill bit based on EMD of Hilbert-Huang transform is presented for the offshore cluster well anti-collision.To further extract useful information contained in response signals under complicated environment,empirical mode decomposition algorithm was used to decompose the original drill bit vibration signal into the intrinsic modes.By analyzing kurtosis and energy spectrum of the intrinsic mode function components,the first four IMFs were found that contains the feature of the drill bit.The specific frequency interval of the drill bit is found after power spectrum analysis.When the bit is approaching to the casing of the adjacent well,the analysis result of the specific IMFs changed significantly and agreed well with the conclusion.The proposed processing method was confirmed effective through real experiment data analysis on the offshore cluster well.