Trying to calibrate the new key to help you log investigation, a simple non-linear regression dating try run-in Matlab 7
Table 4 summarizes the pertinent well data used for calculation of the gamma ray, potassium, and the thorium indexes. Their comparison with the actual measured clay content from the XRD analysis (Fig. 5) showed that the unmodified gamma ray, potassium, thorium indexes calculated from the linear Eq. (1) provided wide overestimation of the Shurijeh clay content, which leads eventually to the misestimating of the original hydrocarbon in place and reserves. According to the Shurijeh age (Early Cretaceous), Larinov calibration for highly consolidated formations was used and based on the data in Fig. 6, ? sh in Dewan equation which is corresponding to the highest gamma ray reading considered to be 2.75 g/cc units. The C factor in the Bhuyan–Passey modification was considered to be 60 for the Shurijeh Formation after measuring the clay contents of reference adjacent shale in many surface samples. The estimated clay content from modified gamma ray, potassium, and thorium indexes are given in Table 5.
The newest testing ranging from unmodified ray indexes and you may center-measured clay content material. Black community Gas producing better study activities, black right up-pointing triangle low-generating really investigation points
Occurrence against GR throughout the key examples of the new Shurijeh Development. Black network Gasoline producing better data factors, and black colored up-pointing triangle non creating better study issues
Therefore, the values of linear gamma beam, potassium, and you will thorium spiders was indeed altered, having fun with each one of the empirically derived non-linear transform equations delivered by Larinov , Clavier , Steiber , Dewan , otherwise Bhuyan and you will Passey (all of the relationship is listed in Table step 1) to track down a less incorrect quote out-of Shurijeh clay blogs
Because the problems of the early in the day variations produced a somewhat high variation to the overall performance, it is essential to derive an empirical matchmaking to the clay blogs estimations inside development. eleven.0 application. Found below is the gotten non-linear calibration matchmaking on the Shurijeh Creation when it comes to a rational function ranging from natural gamma ray directory since the separate variable in addition to lab-derived weight per cent clay based on the X-beam diffraction analysis as the established variable:
Contour 8 shows the partnership away from lbs % clay in the XRD measurements of each other wells and you will changed sheer gamma beam index playing with various other equations as well as extremely consolidated Larionov changes, Clavier et al
The natural gamma ray index was chosen for running the regression analysis due to the stronger correlation coefficient in compare to the potassium or thorium indexes, with the core clay contents in both wells. The unique feature of new equation is to calculate the clay content of less than muzmatch 100% with a given IGR of 1.0, while all other previous modifications give clay content of 100% for such IGR value. The assumption used in developing the non-linear relationships was based on the fact that the entire radioactivity is not due to the clay minerals only. The goodness of agreement and the reliability of the regression equation were then both verified by a correlation coefficient of 0.992 upon application on some other core samples from another wells drilled in the Shurijeh Formation. It is clear from the data in Fig. 7 that the core data, verify both the very low and the medium range of clay contents, estimated from the non-linear empirical relationship. The average percent relative error was also minimized to 11.4%. Due to the statistical bias of comparing data samples with very different sizes and variances (76 samples versus 11 samples), the error cannot be reduced further. , Steiber, Dewan, Bhuyan–Passey and the empirical transforms. The clay content was also estimated from the potassium and thorium indexes using the empirical non-linear calibration (Table 5) and a comparison of average percent relative errors for different equations has been shown in Fig. 9.