Optimum Space Throughput for Rate of Penetration in Realtime

Abstract

Abstract Over the years, the industry has come across a variety of drilling optimization techniques. One of the advantages of optimization, its inherent mathematical accuracy, is also considered a disadvantage because the downhole conditions are unknown and uncertain. As a result, models based on traditional mathematics have inaccurate optimal drilling parameters and can become unusable, especially in real-time due to non-convergence deadlocks, making them impractical for real-time use. To address this issue, this paper proposes an opto-satisficed time-optimal batch methodology is proposed that brackets the feasible solution space using various engineering models. The various engineering and data science models are connected in the form of microservices and can be called as many times as needed during optimization. To curb conservatism, the model is enhanced with dynamic constraints to ensure optimality. Additional limits can be calculated for each discretized subinterval. Taking into account bit wear and vibration, the optimum penetration rate limit is calculated when drilling the well at the specified intervals. It is assumed that the bit wears continuously with each drilling interval and the amount of wear depends on uncontrolled and uncertain parameters such as formation, anisotropy and slope. Note that it is the combination of optimization, satisfaction, and sufficient conditions that makes the drilling operation parameters optimal. Analysis results show that convergence of optimum bound space is very quickly bracketed for optimal drilling parameters in real-time, with predictability in test wells showing best solution results under uncertainty. We also found that the results provided a reasonable threshold if more data were used in well drilling. Results show that as long as the rig is within the operating range, the operating parameters are satisfactory and sufficient to produce the desired results. In other words, almost normal technical solutions are achieved. In this methodology, we found that operating conditions can be adjusted by combining theoretical engineering models with real-time measurements. We use real results from multiple wells to demonstrate the efficiency of this method compared to conventional well optimization algorithms. The study presents comprehensive results including various examples using operating parameters derived while drilling a well in real time to the satisfaction of drilling engineers.