Abstract
The primary objective of the proposed method is to evaluate the current conditions of the drilling system and suggest modifying values of main drilling control parameters to optimize the efficiency of the drilling in whole, while reducing the probability of premature wear of the drill bit.
The fundamental theory behind the proposed approach is based on some elements of fractal analysis as well as artificial neural networks (NN). Many different complicated multi-parameter systems, including drilling systems, frequently discover scale invariance, or "fractality" of spatial-time properties. On this basis, some common engineering approaches to modeling and diagnostic methods have been developed and used.
In our case, the "system" consists of the drilling components (e.g., surface equipment, drillstring, drill bit, mud, etc.), as well as the formation being drilled. Drilling control parameters include both the parameters adjustable in real-time, such as hook load, RPM, or mud flow rate, as well as those that can be modified with some delay (e.g., mud properties or configuration of the bottom-hole assembly).
We consider specific examples of operational situations accompanying the drilling process, in which the fractal performances’ application allows receiving important practical information during a standard operation, i.e., without carrying out a special controlled field experiment.
We present some field examples that we used for gathering required data, training our model, and making an assessment of the current state of the system (specifically, the drill bit conditions) by using the proposed methodology. We also present the results of monitoring the system's response to specific adjustments of control parameters’ values for these cases.
We believe that the proposed methodology opens new opportunities for real-time drilling optimization that can be efficiently implemented within the scope of the existing drilling practice.
