Predicting stuck pipe problems during oil and gas drilling operation is one of the most complex problems in the drilling business. The complexity of the problem is driven not only by the complexity of the natural factors, but it extends to the nature of the drilling operation itself. The drilling operation is continuously influenced by a dynamic smart system. The dynamic part of the system is impacted by natural forces like formation related characteristics, and also is impacted by human activities during the operation such as drilling, tripping and hole cleaning. The smartness of this system is driven by the fact that the operation is controlled by a number of experts, i.e. drilling engineers, trying to run the best sequence of operations using best operation parameters to achieve operation objective. At the top of that, the engineers can change their operation plan whenever they find it necessary to address any operational condition, including a potential stuck pipe problem. In this paper we prove the stuck pipe prediction problem is not a binary classification problem. Instead, we define the stuck pipe prediction problem as a multi-class problem which takes into consideration the dynamic nature of the drilling operation. A reinforcement learning based algorithm is proposed to solve the redefined problem, and its performance and evaluation results is shared in details. The accuracy of the developed algorithm in terms of detecting true stuck pipe events is shown. The results will compare the performance of different machine learning algorithms, which is then used to justify the selection of the best performing method. In addition, we show the accuracy performance improvement through time by employing the feedback channel to retrain the model. The presented method is using a reinforcement logic, in which the solution is connected to the operation reporting to label the solution prediction for false and true predictions. This information is then used to return the neural networks to learn new operational patterns to enhance accuracy.

No abstract

In recent years, the Drilling and Workover (D&WO) operations are growing significantly. The growth of active operations required and produced more data from D&WO operations. With very large number of rig activities daily transmitting more than 60,000 real-time data points every second, it became necessary to understand and utilize this Big Data in order to predict drilling troubles and discover hidden knowledge. The adaption of the industrial Revolution (IR) 4.0 contributed to the use of advanced and novel approaches such as Artificial intelligence (AI) and Machine learning (ML) models. However, those models require continues improvement as drilling data change. When using the industrial standard and adapted Wellsite Information Transfer Specification Markup Language (WITSML) based Big Data environment, the task to monitor the performance of a model at a large scale becomes challenging due to common reasons such as a large number of wells, different models being deployed and different data stored in different systems. In this paper, a new approach is introduced using WITSML based Big Data environment. The methods employed utilize an advanced engine to monitor and evaluate active AI/ML models at a large scale. The engine utilizes anomaly detection methods to monitor abnormal behaviors of the models such as sudden high rate of alerts per day/well or a sudden drop in true event detection. The paper will also demonstrate how such technology can help in early detection of model's decay signs or sudden changes in real-time data quality. The solution improved and automated the process of monitoring and maintaining of AI/ML models in the Drilling domain. It also made the decay detection of models possible and showed how models improve when iterative enhancements are deployed.

As stuck pipe continues to be a major contributor to non-productive time (NPT) in drilling for oil and gas operations, efforts to mitigate its incidence cannot be over emphasized. A machine learning approach is presented in this paper to identify warning signals and give early indications for an impending stuck pipe possibility during drilling activities so as to take proactive measures to mitigate its occurrence. The model uses a moving window-based approach to capture key drilling parameters trends and apply an unsupervised machine learning algorithm to predict abnormalities in the parameters’ rate of change. It utilizes most commonly available drilling real-time data and is therefore deployable in all type of wells. No pre-drill model is essentially required as the model utilizes a self-learning and self-adjusting model. The methodology involves the use of change point detection in identifying rig activity and the associated drilling parameters so as to capture relevant parametric trend for analysis. Inherent in the parameter trend are the different factors that affects their readings; such as wellbore geometry, bottom-hole assembly (BHA), dogleg severity (DLS), formation characteristics, pump flow rate and pipe rotations. The algorithm has been tested on historical wells data in which stuck pipe incidence, near-miss stuck pipe occurred, and incidence-free wells to prove the concept. The results of the model performance is hereby presented along with an accuracy measure.

The Drilling Performance Curve (DPC) is a simple yet powerful tool to assess the drilling performance in any given area where a consecutive series of similar wells have been drilled. Have shown the Drilling Performance Curve as a valid measure of drilling performance, it provides an indicator of relative improvement over a campaign, and position on learning curve can provide a measure of process maturity. Brett and Millheim has proven that Learning Curve Theory (LCT) mathematically describes the ability of organizations to improve their performance over time. It has been used under many different names - the progress curve, the experience curve, etc. All the information that is needed to perform the analysis is the sequence numbers of the well and the time it takes to reach a given depth. The expression relating the sequence in which a well was drilled and the total activity duration is given by: T = C 1 * e c 2 ( 1 − n ) + C 3 Where: T: Time required to drill the nth well, n: Well number in an area of uniform geology, C1: Constant reflecting how much longer the initial well takes to drill than the idealized final well, C2: Constant reflecting the speed with which the drilling organization reaches the minimum drilling time for an area, C3: Constant reflecting the idealized minimum drilling time for the area. Defining the value of the three constants; C1, C2, and C3, is the most critical step in the process as it affects the accuracy of the total duration predications for future wells. In this paper, Nonlinear Least Squares Method is used to estimate the parameters in the nonlinear equation shown above. Based on the operation classification and coding reported in the daily morning report, well operation durations were collected at many different levels of granularities such as: well, hole sections, major operations, and activity levels. LCT equation was applied at these predefined levels of operation time breakdown and Nonlinear Least Squares Method was utilized to calculate the value of the equation three constants that will be used to predicate future wells duration. It was found that, LCT is a good practical tool for performance monitoring and predication with highly accurate results up to 90 % where the experience gaining by the involved teams is the key fact consideration in this technique in addition to the technology technical limitations. The degree of detail collected in our drilling data allows more variety of drilling performance evaluation. Learning curve calculation is performed by the system on well level and can be decomposed into different levels of analysis covering from total well duration down to action duration level. The proposed technique allows any codes combinations as well. LCT was found to be an objective, neutral tool to measure the performance of the organization and different teams at many different levels with a very high accuracy while using Nonlinear Least Squares Method to estimate the LCT equations parameters.