A reduction in well log computing costs and an exchange of technique between professions is accelerating the use of computers in interpretation. A new computer program for Fast Fourier Transformation allows economic application of several processes previously too cumbersome. The digital approach to seismic analysis provides capability in that branch of the oil industry directly applicable to log analysis.
These techniques permit rapid and economic data screening and new opportunities for classification and correlation. Significant correlation of large amounts of data with reference such as core analysis can be made. These give important local guidance as to the contribution or lack of contribution of various logs to a desired end result. Many of the tools are simple and straightforward. They allow recalibration of raw input data, the elimination of extraneous data and the smoothing and enbancing of data. This permits easier recognition of trends and correlation. Correlation and classification techniques taken from such sciences as biomedicine and psychology can be applied directly to the determination of lithology. This is improved through the use of Linear Programming. Once estimates of lithology are available from logs, several geological mapping techniques can be exploited.
A major problem in processing log data is the initial preparation, calibration, normalization and editing. Once the data is in computer compatible form and exonerated, many convenient manipulations can be made. These are best presented in graphic form. Visual analysis of graphs allows interactive with the computer form improvement of this processing.
The paper covers general techniques which can be applied to western Canadian prospects, and have been applied to Devonian carbonate sequences. The complexity of Devonian carbonates requires aids such as Time Series Analysis and Multivariate Analysis. Statistical procedures now available help dramatically in the choice of log suites and derivation of porosities and lithologies. At the same time, the impersonal contact that the analyst then has with the raw data allows a variety of errors and distortions to prevail unless precautions are taken.
During 1968 a price breakthrough in well log digital processing started a renaissance of computer log analysis. Libraries of economical digitized logs are being organized. Large-scale computer facilities, with their economies and conveniences, are increasingly available. These factors are combining to encourage computer log analysis and fresh general techniques made attractive and necessary by wealth of data.
The manual approach to data analysis retains some advantages" This is still the subject of experimentation. Project "MAC" at the Massachusetts Institute of Technology is an example. "MAC" has advanced from "Multiple Access Computer to "Machine Aided Cognition". It has been quickly discovered that while a computer could be taught to perceive and recognize, the programming task is huge. The eye, together with the mind" can perform certain graphic analysis with relative ease. For example, hen the human v1ews a cross-plot of data points he can quickly appreciate: smoothness, averaging, circularity, cyclicity, clusters, and patterns. A promising combination of man and machine calls for an iterative exchange of graphic data.
