关键词: 感应测井, 测井数据, 系统传递函数, 谱估计法, 地层电导率, 校正曲线, 七侧向测井, 统计模型, 分辨率, 约束条件

Abstract: This paper demonstrates the application of AR spectral estimation to obtain the vertical transfer function of induction log, and to deconvolve the induction logging data with it for calibration. With a statistical model of formation conductivity distribution, this paper determines two constraints of the transfer function as: H(0)=1 and H(f)=h(-f). The vertical transfer functions of induction logging data in two wells located in different district are estimated by means of AR spectral estimation. Then, deconvolution as a calibration, are made by conjugate gradient method. Comparison of the data calibrated with core conductivity data from a well with oil base mud,7 late-rolog data and original induction data indicates that this calibration improves obviously the resolution and accurary of inducation logging data. In addition, scatter diagrams of induction logging data before and after the calibration shows that the calibration makes induction logging data closer to core data than the original ones.