基于2D-VMD的偶极阵列声波测井反射波提取方法

Extracting reflected waves from dipole array acoustic logging data based on 2D-VMD

偶极远探测声波测井是利用偶极子阵列声波测井仪器接收到的反射波进行井旁构造成像的测井技术,广泛应用于井周远探测储层评价领域,该技术数据处理的关键是从接收到的声波全波列中分离出反射波并进行成像。目前的反射波分离算法虽然实现了波场分离,但存在直达波和反射波的模态混叠问题,分离后的反射波信号能量损失较大。针对这些问题,基于引入的二维变分模态分解(2D-VMD)建立反射波提取新算法,将原始信号分解为多个具有特定方向和振动特性的固有模态分量,然后选取特定的分量进行重构,实现反射波的分离。采用2D-VMD、中值滤波、F-K变换对模拟和实测数据进行反射波提取。结果表明,相比中值滤波、F-K变换,基于2D-VMD的方法能够在有效压制模式波和随机噪声的同时,准确地提取出完整的反射波信号,最大限度保留反射波的幅度。证明了该方法在偶极阵列声波测井反射波提取中的有效性和优越性。

Single borehole reflection imaging is an effective approach to the direct evaluation of geologic structures and reservoirs around a borehole using reflected shear waves extracted from array acoustic logging data.To enhance azimuthal resolution and radial depth of investigation,the key to data processing is complete separation of reflections from original data with the minimum loss of signal energy.Existing separation methods suffer from the problems of modes mixing and energy loss of extracted reflections and consequent small resolution and radial depth.We propose a new algorithm of 2D variational mode decomposition for wave separation without great time consumption of recursive iteration.Original signals are decomposed into several inherent modes with different directions and vibration characteristics,and appropriate modes are used for reconstruction and reflection separation.Reflected waves are extracted from synthetic data with different noise levels and field data using 2D variational mode decomposition,median filtering,and F-K transform;our method performs better than median filtering and F-K transform in effective suppression of mode waves and random noises and extraction of complete reflected signals.Synthetic and field data tests demonstrate the effectiveness and advantages of our method.