Designing and Implementing an Algorithm Based On an Autoregressive Kalman Filter to Estimate Well-Log Data

Traditional Kalman filters (KF) rely strongly on former knowledge of the possibility of unstable processes and noise consideration statistics. Inadequate priority filter statistics with noise instability decrease the accuracy and precision of the estimated modes and cause bias in the estimation. Adaptive KF according to the autoregressive (AR) predictor model, in this article, is proposed for well-logging analysis. One of the primary devices for well-logging is natural gamma-ray (NGT), which finds out the fluctuation in natural radioactive emitting from the concentration of three elements (Potassium (K), Thorium (Th), and Uranium (U)). The data are obtained from the implementation of the natural gamma ray sensor and the advanced signal processing on these data. The NGT tools in this research are shown in a combination with the energy level of the concentration of these substances in diverse depths. Conventional methods have uncertainties in estimating gamma rays received from the radioactive energy level to assess the reservoir. To cope with this problem, the adaptive KF base on AR is offered and proposed. However, the result is improved in comparison with KF. By applying the optimal AR method, the result is better and improved. Practical implementation and analysis confirm the validity and accuracy of our research.