The inversion of NMR log data sets with different measurement errors

We present a composite-data processing method which simultaneously processes two or more data sets with different measurement errors. We examine the role of the noise level of the data in the singular value decomposition inversion process, the criteria for a proper cutoff, and its effect on the uncertainty of the solution. Examples of processed logs using the composite-data processing method are presented and discussed. The possible usefulness of the apparent T-1/T-2 ratio extracted from the logs is illustrated, (C) 1999 Academic Press.