Cutting concentration prediction in horizontal and deviated wells using artificial intelligence techniques

Abstract

Improper hole cleaning or drilled-cutting transportation impacts drilling operations. To illustrate, inefficient cleaning of the wellbore can lead to many drilling problems such as low drilling rate (i.e. low ROP), early bit wear and, in the severe cases, a complete loss of the well due to stuck pipe. To understand efficiency in cutting transport in drilling and to provide solutions for the problem, many studies have been conducted. In all cases, they provide empirical models based on experimental data. In this study, different artificial intelligence (AI) techniques are employed to estimate the concentration of cuttings present in the wellbore. The purpose of this study is to indirectly measure the hole-cleaning efficiency in order to predict the cutting concentration from drilling parameters using artificial intelligence techniques. The study is based on 116 experimental data records from the studies. Two AI techniques were selected, namely artificial neural network (ANN) and support vector machine (SVM), to estimate the cutting concentration in the wellbore. The input parameters comprise mud density and mud rheological properties (yield point and plastic viscosity) in addition to drilling parameters including the hole inclination angle, pipe eccentricity (i.e. location of the drill pipe from the axis of the well), the rate of penetration (ROP), flow rate (GPM), drill pipe rotary speed (RPM) and temperature. The results obtained show the ability of the two employed techniques to accurately predict the cutting concentration in the wellbore with average absolute errors (AAE) less than 5% and correlation coefficients (R) higher than 0.9. Comparison of these results with a literature model showed that the AI techniques provide better predictions of cutting concentration and higher accuracy than that model. Applying the developed AI technique will help the drilling engineers to assess the hole cleaning in a real time.

Optimization of Flow Rate and Pipe Rotation Speed Considering Effective Cuttings Transport Using Data-Driven Models

“…Al-Azani et al [14] used SVM to estimate the cuttings concentration in the wellbore by correlating it with the drilling fluid properties and the drilling parameters such as the pump rate, rotation speed, etc. In a later study [15], they extended the initial work by incorporating ANN models. They trained the models using the data published by Yu et al [16].…”

Application of Artificial Intelligence Techniques in Predicting the Lost Circulation Zones Using Drilling Sensors

“…ANN is an information-processing system, which attempts to imitate the performance features of the biological nervous system. The network is adapted as a computer model, which can advance transformations, associations, or mappings between data [17]. The feature of ANN is that it does not require any physical phenomenon that explains the system under study [18].…”

A review of technological advances and open challenges for oil and gas drilling systems engineering

“…They compared the performance of their ANN model with the results of a mechanistic cuttings transport model and realised an absolute average deviation of less than 6%. A similar approach by Al-Azani et al 108 was adopted for hole cleaning; this was however based on 116 experimental data records.…”