The Multiwell Faciolog (mark of Schlumberger) computer process is a data-reduction technique that uses multivariate analysis to integrate data from wireline logs, core data, and geologic knowledge into a geologically significant display of electrofacies. Electrofacies are defined by Serra and Abbott as the "set of log responses that characterizes a sediment and permits the sediment to be distinguished from others."

The computing chain begins with log normalization and correction for environmental effects. The corrected logs from a selected key well are then used to construct an n dimensional crossplot incorporating data from n number of input logs. The program selects n principal component axes through the resulting cloud of data points and automatically clusters the data into local modes, which are then displayed in two-dimensional principal component space. Each group of local modes represents intervals that have similar log response. Local modes are then clustered into a smaller number of terminal modes, which are identified by rock type, using core lithology data and geologic knowledge, and are manually clustered into a significant number of electrofacies. Once the electrofacies are defined n the key well, a data base is established, containing information on the average log values for each electrofacies, geologic descriptions, and display patterns. Each subsequent well with a similar logging suite is then compared to the data base. Those zones corresponding to electrofacies previously defined in the data base will be identified as the same electrofacies. Those zones with different average log values will be identified as different electrofacies and will be added to the data base.

The advantage of the Multiwell Faciolog technique is that voluminous data, including wireline, core, and geologic information,can be integrated into one or two key displays. This program is flexible and allows the interpreter to adjust and modify the data base, as necessary, with the addition of each new well. The data base does not require a unique data set and can be constructed from open-hole logs, cased-hole logs, or computed logs common to the study wells. The final cross-sectional displays can be an aid in tracing geologic characteristics from well to well.

End_of_Article - Last_Page 682------------