Deconvolution of Wireline Logs Using Point-Spread Function

Most wireline logs used for the petrophysical evaluation of reservoirs have a vertical resolu tion in the order of one meter. This poses a problem when the typical layer thickness if less than one meter, since no correct reading will be obtained. High-resolution logs like the electromagnetic propagation tool, the dipmeters or acoustic and electrical imaging devices have a resolution well below one meter, sometimes as high as one centimeter, but their application to petrophysical reservoir evaluation is limited. We present an approach which uses information from high-resolution logs to deconvolve low-resolution logs. From the high-resolution log we first identify a sharp bed boundary, on both sides of which the petrophysical properties assume a locally constant but different value. From this control interval we determine the point-spread function (PSF) of the low-resolution log using the basic convolution theorem. The PSF is mathematically equivalent to the tool response function under actual borehole conditions. It can be obtained at various sharp bed boundaries in order to obtain a more representative tool response. The low-resolution log is then deconvolved using Fast Fourier Transforms over the entire interval of interest. We apply this method to a number of synthetic and field data sets. Noise filtering prior to deconvolution is found to be important, as is correct depth shifting and a proper choice of the control interval.