Measurement-while-drilling Data Research Articles

Profiling of weathered argillaceous limestone rock with MWD data from advanced drilling for tunnelling along Wu-Kai expressway in Chongqing, China

Advanced drilling was employed for the ongoing excavation of a tunnel on Wu-Kai Expressway in Chongqing, aiming to characterize the ground condition ahead of the tunnel face and avoid potential geohazards. Measurement While Drilling (MWD) system and the hydraulic rotary crawler drilling machine are used for the advanced drillhole in the weathered argillaceous limestone rock mass. However, the MWD data exhibits random variations in penetration rate and cannot be correlated with ground conditions. In addition, there is little MWD data on the weathered argillaceous limestone with time-series algorithm in relevant open literature. This paper presents further analysis of the MWD data, resolves the random variation issue in penetration rate, and re-interprets the MWD data along three advanced drillholes in this tunnel project. The re-calculated penetration rate with the time-series algorithm for each drillhole is a stepped function of the drillhole depth. The constant penetration rates for the different homogeneous zones can be further correlated with the weathering degrees of the rock and compared with those from other five ground sites with different ground conditions in China and London. The findings of the paper can enhance the general applicability of MWD data for accurate profiling of weathered argillaceous limestone rock mass for safe tunnelling.

Normalizing Large Scale Sensor-Based MWD Data: An Automated Method toward A Unified Database.

In the context of geo-infrastructures and specifically tunneling projects, analyzing the large-scale sensor-based measurement-while-drilling (MWD) data plays a pivotal role in assessing rock engineering conditions. However, handling the big MWD data due to multiform stacking is a time-consuming and challenging task. Extracting valuable insights and improving the accuracy of geoengineering interpretations from MWD data necessitates a combination of domain expertise and data science skills in an iterative process. To address these challenges and efficiently normalize and filter out noisy data, an automated processing approach integrating the stepwise technique, mode, and percentile gate bands for both single and peer group-based holes was developed. Subsequently, the mathematical concept of a novel normalizing index for classifying such big datasets was also presented. The visualized results from different geo-infrastructure datasets in Sweden indicated that outliers and noisy data can more efficiently be eliminated using single hole-based normalizing. Additionally, a relational unified PostgreSQL database was created to store and automatically transfer the processed and raw MWD as well as real time grouting data that offers a cost effective and efficient data extraction tool. The generated database is expected to facilitate in-depth investigations and enable application of the artificial intelligence (AI) techniques to predict rock quality conditions and design appropriate support systems based on MWD data.

Bottom hole pressure prediction based on hybrid neural networks and Bayesian optimization

Many scholars have focused on applying machine learning models in bottom hole pressure (BHP) prediction. However, the complex and uncertain conditions in deep wells make it difficult to capture spatial and temporal correlations of measurement while drilling (MWD) data with traditional intelligent models. In this work, we develop a novel hybrid neural network, which integrates the Convolution Neural Network (CNN) and the Gate Recurrent Unit (GRU) for predicting BHP fluctuations more accurately. The CNN structure is used to analyze spatial local dependency patterns and the GRU structure is used to discover depth variation trends of MWD data. To further improve the prediction accuracy, we explore two types of GRU-based structure: skip-GRU and attention-GRU, which can capture more long-term potential periodic correlation in drilling data. Then, the different model structures tuned by the Bayesian optimization (BO) algorithm are compared and analyzed. Results indicate that the hybrid models can extract spatial-temporal information of data effectively and predict more accurately than random forests, extreme gradient boosting, back propagation neural network, CNN and GRU. The CNN-attention-GRU model with BO algorithm shows great superiority in prediction accuracy and robustness due to the hybrid network structure and attention mechanism, having the lowest mean absolute percentage error of 0.025%. This study provides a reference for solving the problem of extracting spatial and temporal characteristics and guidance for managed pressure drilling in complex formations.

Formation lithology predictions based on measurement while drilling (MWD) using gradient boosting algorithms

Many subsurface engineering applications require petrophysical and geomechanical properties as inputs for estimating the recoverable hydrocarbon volume and the reservoir deformation and properties' change during the life cycle of the production. Well log data plays a key role in estimating these reservoir properties, including volumes of minerals, Young's Modulus, and Poisson's Ratio. A conventional workflow using a multi-mineral petrophysical inversion model based on core-log integration requires an a priori rock-fluid model; therefore, it is often unavailable in reservoirs of complex minerals, such as carbonate or unconventional fields. In addition, most unconventional wells do not acquire Logging While Drilling (LWD) and/or wireline logs due to economic constraints or borehole instability reasons. Available measurements for these wells are typically Measurement While Drilling (MWD) Gamma Ray (GRMA) logs along with drilling dynamics measurements including Surface Weight-On-Bit (SWOB), Rate Of Penetration (ROP), Torque (TQA), Revolutions Per Minute (RPM), and the differential pressure. Thus, in this study, we demonstrate a machine learning approach, which does not require an a priori model, to predict formation lithologies based on conventional MWD and wireline well logs.We applied the Extreme Gradient Boosting Regression (XGBR) method to simultaneously predict the weight fractions of 5 minerals: QFM (Quartz-Feldspar-Mica), carbonate, clays, pyrite, and siderite from drilling dynamics measurements. The XGBR algorithm was tested with datasets from four adjacent wells in the Volve Field, located in the North Sea. A total of 10,575 data points were used to train and validate the model. The trained model was then applied to the blind test dataset of 10,446 data points from the same field. Lithology predictions from the trained model yield excellent results of weight concentrations with Root-Mean-Square Errors (RMSE) of 6.01 and 7.44% for training and test datasets, respectively. The prediction results agree with the lithology obtained from the Elemental Capture Spectroscopy (ECS) data as well as the dominant lithology record from rock cuttings. With the developed model, lithologic information and sonic logs of the formation could be predicted with MWD data, which could ultimately help optimize real-time drilling and completion operations in the field.

BILSTM-Based Deep Neural Network for Rock-Mass Classification Prediction Using Depth-Sequence MWD Data: A Case Study of a Tunnel in Yunnan, China

Measurement while drilling (MWD) data reflect the drilling rig–rock mass interaction; they are crucial for accurately classifying the rock mass ahead of the tunnel face. Although machine-learning methods can learn the relationship between MWD data and rock mechanics parameters to support rock classification, most current models do not consider the impact of the continuous drilling-sequence process, thereby leading to rock-classification errors, while small and unbalanced field datasets result in poor model performance. We propose a novel deep neural network model based on Bi-directional Long Short-Term Memory (BILSTM) to extract information-related sequences in MWD data and improve the accuracy of the rock-mass classification. Two optimization modules were designed to improve the model’s generalization performance. Stratified K-fold cross-validation was used for model optimization in small and unbalanced datasets. Model validation is based on the MWD dataset of a highway tunnel in Yunnan, China. Multiple metrics show that the prediction ability of the network is significantly better than those of a multilayer perceptron (MLP) and a support-vector machine (SVM), while the model exhibits an improved generalization performance. The accuracy of the network can reach 90%, which is 13% and 15% higher than the MLP and SVM, respectively.

MWD Data-Based Marble Quality Class Prediction Models Using ML Algorithms

Brønnøy Kalk AS operates an open pit mine in Norway producing marble, mainly used by the paper industry. The final product is used as filler and pigment for paper production. Therefore, the quality of the product has utmost importance. In the mine, the primary quality indicator, called TAPPI, is quantified through a laborious sampling process and laboratory experiments. As a part of digital transformation, measurement while drilling (MWD) data have been collected in the mine. The purpose of this paper is to use the recorded MWD data for the prediction of marble quality to facilitate quality blending in the pit. For this purpose, two supervised classification modelling algorithms such as conventional logistic regression and random forest have been employed. The results show that the random forest classification model presents significantly higher statistical performance, and it can be used as a tool for fast and efficient marble quality assessment.

Deep Learning for Intelligent Prediction of Rock Strength by Adopting Measurement While Drilling Data

Precise, rapid, and reliable prediction of rock strength parameters is of great significance for underground engineering. This paper presents a method for predicting rock strength parameters including the Poisson’s ratio (P), elastic modulus (E), and uniaxial compressive strength (UCS) based on computer drilling jumbo measurement while drilling (MWD) data. First, the distribution characteristics and correlation of MWD data are studied; second, a filtering method of MWD data is proposed, which reduces the influence of operational and mechanical factors; finally, an intelligent prediction model of rock mechanics parameters was established, 30 groups of test data were used for application, and the mean absolute percentage error (MAPE) of prediction results for P, E and UCS are 2.11%, 3.11%, and 2.9%, the determination coefficients (R2) are 0.4346, 0.8241, and 0.6616. Compared with the data before optimization, the accuracy of prediction results is improved significantly, it shows that the deep neural network model can accurately predict rock mass parameters.

Rock mass structural recognition from drill monitoring technology in underground mining using discontinuity index and machine learning techniques

A procedure to recognize individual discontinuities in rock mass from measurement while drilling (MWD) technology is developed, using the binary pattern of structural rock characteristics obtained from in-hole images for calibration. Data from two underground operations with different drilling technology and different rock mass characteristics are considered, which generalizes the application of the methodology to different sites and ensures the full operational integration of MWD data analysis. Two approaches are followed for site-specific structural model building: a discontinuity index (DI) built from variations in MWD parameters, and a machine learning (ML) classifier as function of the drilling parameters and their variability. The prediction ability of the models is quantitatively assessed as the rate of recognition of discontinuities observed in borehole logs. Differences between the parameters involved in the models for each site, and differences in their weights, highlight the site-dependence of the resulting models. The ML approach offers better performance than the classical DI, with recognition rates in the range 89% to 96%. However, the simpler DI still yields fairly accurate results, with recognition rates 70% to 90%. These results validate the adaptive MWD-based methodology as an engineering solution to predict rock structural condition in underground mining operations.

Improving face decisions in tunnelling by machine learning‐based MWD analysis

AbstractIn Norwegian drill and blast tunnelling, contracts stipulate collecting Measurement While Drilling (MWD) data from all drillholes. The MWD approach is an objective way of collecting, processing and visualising advance drilling data that have been successfully used in making face decisions for many years in Norwegian tunnelling. MWD data collection consists of equipping drill rigs with sensors recording different drilling parameters, with subsequent near real‐time data processing for access by on‐site personnel and face engineers in decision‐making process. A deficiency in the MWD approach is still the subjective data interpretation necessary to translate visualised data into actual face decisions. Digital scepticism and a lack of digital knowledge are other obstacles in automating the process from data interpretation to decisions. Thus, MWD data are sometimes only used for nice visualisations and as‐built documentation. This study proposes machine learning (ML)‐based methods to characterise the rock mass from sensor data using data from five twin‐tube tunnels in a Norwegian highway project. Results show that the deep learning‐based method – convolutional neural network – is capable of translating complex patterns in MWD data to functional rock mass characterisation, thereby aiding face engineers with data analysis. The study is, to our knowledge, the first known attempt to use deep learning‐based computer vision techniques to interpret MWD data framed as images.

Deep Convolutional Network Improves Completion Design

This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper SPE 201545, “Completion Design Improvement With a Deep Convolutional Network,” by Liwei Li, Nasser Nasrabadi, and Timothy Carr, SPE, West Virginia University. The paper has not been peer reviewed. Maximizing stimulated natural and hydraulic-fracture networks is a primary concern for economic production from a horizontal shale gas well. Geomechanical facies and preexisting fractures in each stage are identified based on similarities in formation characteristics. This often requires analyzing large volumes of drilling, logging-while-drilling (LWD), and measurement-while-drilling (MWD) data. In the complete paper, the authors develop a methodology that calculates mechanical specific energy (MSE) using real-time drillstring acceleration signals. Introduction The Marcellus Shale Energy and Environmental Laboratory (MSEEL) aimed to improve stimulation efficiency by collecting, integrating, and analyzing real-time LWD and MWD data. In addition, engineered completions were combined with fiber-optic distributed-temperature-sensor and distributed-acoustic-sensor data. To reduce issues with data quality measured using surface sensors, advanced MWD and LWD tools are proposed to collect data close to the bit. In the analysis of nonstationary data such as drillstring vibration measurements, the conventional Fourier transform is limited to frequency spectrum without time information. To address this problem, new signal-processing methods have been developed to provide frequency features localized in time. Deep convolutional neural networks (CNNs) are a promising approach for image classification. Feature extractions can be performed by adding convolution and pooling filters, reducing spatial resolutions, and increasing depth to build a deep learning network. The authors propose a decision chart that labels rock laterals using calculated MSE values and interpreted natural fracture intensities to improve stage-by-stage perforation efficiency. The borehole-imaging tool is used to achieve a cost-effective solution. This tool collects both high-resolution acoustic amplitudes and triaxial accelerations.

Predicting rock fragmentation based on drill monitoring: A case study from Malmberget mine, Sweden

Fragmentation analysis is an essential part of the optimization process in any mining operation. The costs of loading, hauling, and crushing the rock are strongly influenced by the size distribution of the blasted rock. Several direct and indirect methods are used to analyse or predict fragmentation, but none is entirely applicable to fragmentation assessment in sublevel caving mines, mainly because of the limitations imposed by the underground environment and the lack of all the required data to adequately describe the rock mass. Over the past few years, measurement while drilling (MWD) data has emerged as a potential tool to provide more information about the in-situ rock mass. This research investigated if MWD can be used to predict rock fragmentation in sublevel caving. The MWD data obtained from a sublevel caving mine in northern Sweden were used to find the relationship between rock fragmentation and the nature of the rock mass. The loading operation of the mine was filmed for more than 12 months to capture images of loaded load-haul-dump (LHD) buckets. The blasted material in those buckets was classified into four categories based on the median particle size (X50). The results showed a stronger correlation for fine and medium fragmented material with rock type (MWD data) than coarser material. The paper presents a model for prediction of fragmentation, which concludes that it is possible to use MWD data for fragmentation prediction.

Formation Pressure Inversion Method Based on Multisource Information

SummaryThe exploration and development of offshore oil and gas have greatly alleviated the tension of global oil and gas resources. However, the abnormal pressure of offshore reservoirs is more common compared with terrestrial oil and gas reservoirs, and the marine geological structure is complex, with the development of faults, fractures, and high and steep structures, which leads to the strong anisotropy of formation pore pressure distribution and uncertainty of pressure system change. In this paper, considering the corresponding characteristics of the randomness of the formation pressure prediction results in the Eaton equation for their respective variables, a formation pressure inversion method based on multisource information, such as predrilling data, bottomhole while drilling data, seabed measured data, and surface measured data, is established. On this basis, combined with the data of a well in the South China Sea, the variation law of the uncertainty of formation pressure prediction results under the conditions of predrilling data, measurement while drilling (MWD) data, and their mutual coupling is analyzed. The simulation results show that the uncertainty distribution of formation pressure prediction based solely on predrilling data shows linear accumulation trend with well depth, and the formation pressure inversion method based on multisource information can significantly curb the increasing trend of uncertainty when MWD data are introduced. Therefore, through the analysis of typical change patterns of monitoring parameters under normal/abnormal conditions during drilling, combined with the method of multisource information, the abnormal pressure information can be accurately predicted and inversed, which provides important support for wellbore pressure regulation under complex formation conditions.

Specifics of MWD Data Collection and Verification during Formation of Training Datasets

This paper presents a structured analysis in the area of measurement while drilling (MWD) data processing and verification methods, as well as describes the main nuances and certain specifics of “clean” data selection in order to build a “parent” training database for subsequent use in machine learning algorithms. The main purpose of the authors is to create a trainable machine learning algorithm, which, based on the available “clean” input data associated with specific conditions, could correlate, process and select parameters obtained from the drilling rig and use them for further estimation of various rock characteristics, prediction of optimal drilling and blasting parameters, and blasting results. The paper is a continuation of a series of publications devoted to the prospects of using MWD technology for the quality management of drilling and blasting operations at mining enterprises.

Significant Surge and Swab Offshore Brazil Induced by Rig Heave during Drillpipe Connections

SummaryIn this paper, we describe measured and simulated downhole pressure variations (“surge and swab”) during drillpipe connections when drilling an ultradeepwater well offshore Brazil on Bacalhau (former Carcará) Field. Floating rig motion caused by waves and swell (“rig heave”) induces surge and swab when the drillstring is suspended in slips to make up or break a drillpipe connection and topside heave compensation is temporarily deactivated. This is a known issue in regions with harsh weather, such as the North Sea, where pressure oscillations of up to 20 bar have been reported during connections. Recorded downhole drilling data from Bacalhau Field reveals significant pressure oscillations downhole (in the same order of magnitude as in the North Sea) each time the drillstring was suspended in slips to make a connection in the subsalt 8½-in. section of the well. Mud losses were experienced around the same well depth, and they might have been caused by surge and swab.Measured surge and swab pressure variations have been reproduced in an advanced proprietary surge and swab simulator that considers rig heave, drillpipe elasticity, well friction, non-Newtonian drilling mud, well trajectory, and geometry. Moreover, findings in this paper suggest that surge and swab was in fact significantly higher than recorded by the measurement while drilling (MWD) tool. The true magnitude of surge and swab is not captured in the recorded MWD data due to low sampling frequency of the downhole pressure recording (one measurement every 6 seconds, a standard downhole pressure sampling rate used on many operations today).This work shows that surge and swab during drillpipe connections on floaters may challenge the available pressure window for some wells, even in regions with calm weather such as Brazil. Managed pressure drilling (MPD) is a technique that improves control of the downhole pressure. It is, however, not possible to compensate fast downhole pressure transients, such as heave-induced surge and swab, using MPD choke topside. This is due to the long distance between the choke and the bit, which translates into a time delay in the same order of magnitude as typical wave and heave periods. A downhole choke combined with continuous circulation is one of the potential solutions.Surge and swab during drillpipe connections can result in a loss or an influx and should be considered in the well planning phase when mud weight, section lengths, etc. are selected.

An event-triggered approach to torsional vibration control of drill-string system using measurement-while-drilling data

Suppressing torsional vibration of drill-string has great safety and economic importance to drilling process. Note that most existing control methods only measure surface information, their performance are limited, especially for a long drill-string. In this paper, an advanced control system is developed to mitigate torsional vibration of drill-string, using a measurement-while-drilling (MWD) communication tool to collect downhole information. The drill-string is described by a lumped parameter model to include full torsional flexibility modes with less computational complexity. A state observer and a filter are combined to estimate an equivalent control torque to cope with uncertain bit–rock interaction. Considering that the communication and power resources are quite precious for MWD, the downhole state is periodically sampled and an event-triggered mechanism is devised to pick out only some key sampled data for transmission to the surface. Using both the surface continuous measurement and the discrete MWD data with transmission delays, the closed loop of the drill-string and the control system is modeled as a nonlinear time-delay system, based on which sufficient conditions are derived for stability analysis and calculation of gain matrices. A numerical case study is carried out to verify the effectiveness of the approach.

Drill Bit Deformations in Rotary Drilling Systems under Large-Amplitude Stick-Slip Vibrations

In oil and gas industry, rotary drilling systems are used for energy exploration and productions. These types of systems are composed of two main parts: mechanical and electrical parts. The electrical part is represented by rotating motor called top drive; however, the mechanical part of the system is composed of tool string with many pipes, at the bottom end of these pipes the bit is attached to cut the rock during their contact. Since the bit is in a direct contact with rock characteristic variations, it can be under risk for heavy damage. The latter is principally caused by the fact that the rock–bit interaction term is highly nonlinear and unpredictable. In literature, many mathematical models have been proposed for rock–bit interaction, but they do not reflect the dynamic of the bit under vibrations since torsional and axial vibrations are strongly coupled and synchronized with it. In industrial development, the design of drill bit has faced many improvements in order to overcome these vibrations and mitigate unpredictable phenomena. Even though, the practical use of these drill bits confirmed that there are still many failures and damages for the new designs; moreover, bits’ virtual life become shorter than before. The objective of this study is to analyze the drill bit deformations caused by the stick-slip vibration phenomenon which is characterized by high-frequency high-amplitude in rotary drilling systems. The obtained results were validated through a case study of MWD (measurement while drilling) data of well located in a Southern Algerian oil field.

Estimating the blast sill thickness using changepoint analysis of MWD data

Measurement while drilling (MWD) data are collected during production drilling in the open pit operation in question. On every bench in the open pit, a fractured marble zone (sill) is left on top of the hard and un-fractured marble to facilitate subsequent loading and transportation. The boreholes are drilled through this sill and then continue into the solid marble. The varying sill thicknesses cause challenges in the charging and blasting processes since charging the sill can cause fly-rocks and failing to charge enough can result in big boulders that need to be reduced in size through a cost- and time-consuming hammering process. If the sill thickness in each borehole is known, borehole specific charging depths can be developed. This paper shows how a step-wise changepoint analysis on MWD data can be used to estimate sill thicknesses. The estimated sill thicknesses are compared with optical televiewer (OTV) images to validate the predictions. Principal component analysis is used to create variables which gradually explain a decreasing amount of the variances in the data. The study concludes that a mean and variance-based changepoint analysis on the first principal component (PC1) can detect the sill thicknesses with a correlation coefficient equal to 0.98 and sum of squared deviation 0.12 from actual sill thickness. The PC1 is mainly (46% of the variance) described by rotation pressure, feed pressure and dampening pressure, indicating that the variations in these three variables are most important when studying high contrast physical and mechanical differences in rocks.

Rock support prediction based on measurement while drilling technology

In tunnelling, rock mass support is designed based on the observed rock mass conditions. These conditions are determined by rock mass characterization. Measurement while drilling (MWD), one such characterization method, acquires drill parameter data during drilling. These data can be used to predict rock mass conditions ahead of the face. In this study, grout-hole MWD data are collected at one of the main tunnels in the large infrastructural project Stockholm bypass tunnels. The normalized and filtered MWD data are correlated to the rock mass characteristics using multilinear regression and the Levenberg-Marquardt method. Although a strong numerical correlation cannot be obtained, the results are promising. As a result, a holistic visual approach, which linked the MWD parameters with the rock mass classification and rock support requirements, was developed.

Improved filtering and normalizing of Measurement-While-Drilling (MWD) data in tunnel excavation

Measurement-While-Drilling (MWD) data are complex because of the drilling process and drilling system. Swift and accurate normalization and filtration processes are required if MWD data are to be used for rock mass characterization. The current method for MWD data filtration does not consider all the effects of the drilling process, e.g., collaring or extension rod coupling. This paper presents a new method for MWD data filtration and normalization. The new method filters the operational parameters and normalizes the data on a stepwise algorithm, including production factors (collaring drill hole and rod extension coupling) and rock drill settings. It shows reliable results for the normalization and filtration of MWD data gathered for drill hole collaring and coupled extension drill rods at grout holes and for different rock drills.

Establishing relationships between structural data from close-range terrestrial digital photogrammetry and measurement while drilling data

Geologists, mine planners, geotechnical, and mining engineers always strive for maximum information to get a better insight of the rock mass before interacting with it. Over the recent decades, close-range terrestrial digital photogrammetry (CRTDP) has been increasingly used for data acquisition and to support the conventional methods for rock mass characterization. It provides a safe, time-saving and contact-free way to gather enough data to minimize user dependent biases. However, it requires an expensive camera, fieldwork and some software to extract the information from images. In addition, it can over-estimate the rock fracturing sometimes due to weathering of the rock face or poor blasting practices. Measurement while drilling (MWD) data include the responses of different drilling parameters to the variations in the rock mass. MWD data are produced in large quantity, as they come from every hole drilled. These data correspond to the inside variations of rock rather than the surface ones counted in photogrammetry.In this paper, structural data are obtained from different bench faces of an open pit mine using a commercial software package, ShapeMetriX3D (by 3GSM). These data are compared to the MWD data of the boreholes that were blasted to produce these bench faces to establish certain relationships between drilling parameters and rock mass structures. Half casts of the boreholes with MWD data were visible on the bench faces of the pre-split wall that allowed a better correlation. The results show abrupt changes in MWD parameters for open joints or cavities with some infilling material and overall increases or decreases in parameters for closely spaced bedding planes, fractures or foliations. The results are promising and suggest the method can be used to characterize the rock mass, modify the charging of explosives in blasting operations and facilitate the geological modeling of the rock mass.