Quantitative Use of Calibration Data to Correct Miscalibrated Well Logs*
- E.A. Maciula (Mobil Oil Corp.) | J.E. Cochrane (Chevron Oil Field Research Co.)
- Document ID
- Society of Petroleum Engineers
- Journal of Petroleum Technology
- Publication Date
- July 1968
- Document Type
- Journal Paper
- 663 - 670
- 1968. Society of Petroleum Engineers
- 5.6.1 Open hole/cased hole log analysis
- 1 in the last 30 days
- 172 since 2007
- Show more detail
- View rights & permissions
A method of mathematically correcting miscalibrated well logs on the basis of calibration data is presented. The model is a logging system with a linear-response function calibrated to two standards. The model assumes that interaction between calibration error functions is negligible, and that these error functions remain essentially constant during the log recording.
The general model and the resulting correction equation were adapted to induction logging systems, and field Trials were performed to test the validity of the model. A computer program permitted induction log data to be generated continuously through the correction equation. Results indicate the assumptions are reasonable, and that adequate correction is accomplished for the range of conductivity values investigated.
The correction technique was applied to induction logs run by four service companies in the same wellbore in the Wilmington field, Calif. The corrected logs essentially are identical.
Well logs formerly were used primarily for qualitative formation evaluation and correlation. Log accuracy has become of prime importance as logging has begun to provide quantitative parameters for the determination of formation characteristics. Measurement accuracy is dependent upon precise logging-system calibration. However, little attention has been paid by oil companies to these calibration data, and use of these data has been confined to intercompany accuracy control by the logging companies. Recently oil companies have come to realize the dependence of competent formation evaluation on accurate logging system calibration. With the nascence of this interest in highly accurate logs, we foresee advantages to be gained from a method through which a miscalibrated log can be corrected adequately on the basis of calibration information. Such a method would enable the company witness to accept such a log if he were provided with a complete calibration record. This latitude would prove especially profitable where hole conditions are severe or when rig time is a critical consideration.
A mathematical technique is proposed for correcting erroneous logs that are produced by logging systems having a linear response characteristic and calibrated to two standards. To simplify the model, assumptions were made concerning interaction of high- and low-error components. The ideal model has been adapted to fit the induction logging system. Field tests, utilizing purposely miscalibrated induction logs, were conducted to test the validity of the revised model. Results of these tests indicate the equation is correct for usual conductivity ranges observed in California formations. A computer program has been written to generate a corrected induction log using error inputs calculated from the calibration records.
Logging System Model
Definition of Calibration Errors
The model used in this study is the ideal logging system that has a linear response function and that is calibrated to two known standards in the high and low sensitivity ranges. The response of such a system, correctly calibrated, is shown in Fig. 1A. We can describe three types of calibration errors that will directly affect the log accuracy. These are high calibration point error Delta, low calibration point error delta, and recorder offset error sigma. For the case of the simple two-point calibration method, these errors are determined conveniently at the high and low standard calibration points on the response function; that is, they are read from the calibration data as recorded deviations from the known value of the standards.
The imposition of high and low calibration point errors may cause a shift or a rotation of the response function as illustrated in Fig. 1B. The necessary assumption is that there is no interaction between these two errors. This seems reasonable because most present-day logging systems are provided with high and low sensitivity adjustments that are relatively independent.
|File Size||581 KB||Number of Pages||8|