Analysis of the cumulative effect of bit error rate in acoustic telemetry while drilling with multiple repeaters

Abstract

Acoustic telemetry enables data transmission from downhole to surface equipment in drilling operations. However, signal reliability is frequently jeopardized by the cumulative effect of bit error rate (BER) as acoustic signals pass through multiple repeaters in harsh environments. Prior studies have primarily focused on individual factors that influence BER, but there has been no comprehensive analysis of how multiple factors, such as repeater quality, environmental noise, and drilling depth, interact to influence cumulative BER. A structured approach is needed to model and predict the accumulation of BER during drilling. The purpose of this research is to develop a predictive framework for evaluating cumulative BER in acoustic telemetry while drilling with multiple repeaters. It proposes Cumulative BER in the Acoustic Telemetry Drilling Dataset (CBATD) and proposes the Cumulative BER Predictor algorithm for predicting the Final Cumulative BER based on key influencing factors. The CBATD dataset includes 10 input features: Initial BER, Number of Repeaters, Repeater Quality Score, Signal Power, Repeater Distance, Environmental Noise, Temperature, Mud Density, Drill Depth, and Pressure. The target attribute is the Final Cumulative BER (%). The proposed Cumulative BER Predictor algorithm loads and normalizes the dataset selects features depending on correlation, and makes predictions using Linear Regression and Random Forest Regression. Model performance is evaluated using MSE, R², MAE, RMSE, and MAPE, followed by an analysis of key features impacting bit error rate accumulation during drilling using acoustic telemetry. The analysis identified Initial BER, Repeater Quality Score, and Environmental Noise as significant influences on Final Cumulative BER. Linear regression yielded an MSE of 0.35, R² of 0.88, MAE of 0.41, RMSE of 0.59, and MAPE of 7.8%. Random Forest Regression showed improved performance with an MSE of 0.18, R² of 0.94, MAE of 0.28, RMSE of 0.42, and MAPE of 5.2%. The Cumulative BER Predictor algorithm outperformed both, attaining an MSE of 0.10, R² of 0.97, MAE of 0.19, RMSE of 0.32, and MAPE of 3.6% on the CBATD dataset. This confirms its superior accuracy and reliability in predicting cumulative bit error rates in acoustic telemetry drilling. The Cumulative BER Predictor algorithm models BER accumulation and emphasizes enhancing repeater quality and decreasing noise to improve drilling communication, laying the groundwork for better acoustic telemetry designs in harsh environments. Performance improvement is measured by lower prediction errors and a higher R² value compared to baseline models, indicating more accurate and reliable cumulative BER estimation.