The early detection of a stuck pipe during drilling operations is challenging and crucial. Some of the studies on stuck detection have adopted supervised machine learning approaches with ordinal support vector machines or neural networks using datasets for “stuck” and “normal”. However, for early detection before stuck occurs, the application of ordinal supervised machine learning has several concerns, such as limited stuck data, lack of an exact “stuck sign” before it occurs, and the various mechanisms involved in pipe sticking. This study acquires surface drilling data from various wells belonging to several agencies, examines the effectiveness of multiple learning models, and discusses the possibility of the early detection of pipe sticking before it occurs. Unsupervised machine learning using data on the normal activities is a possible advanced method for early stuck detection, which is adopted in this study. In addition, as a countermeasure to another concern that even normal activities involve various operations, we apply unsupervised learning with multiple learning models.

Japan Agency for Marine-Earth Science and Technology (JAMSTEC) operates the scientific drillship Chikyu for scientific research. However, Chikyu has experienced problems with respect to drill pipe failure. This may be due to the limitation of indication of anomalies such as drill pipe failure in conventional drilling data monitoring. As one of primary aims of scientific drilling is to recover core samples from sediment layers under the seabed, improving core recovery rate is very important because it can enhance the operation efficiency. Obtaining the information of lithology of drilling layer is also helpful for both scientific and operational aspects. However, there is no direct information regarding the core recovery rate and lithology. The recovery rate and lithology can be determined after retrieving a coring tool. Therefore, this study applies a machine learning technique to identify the drilling states, which includes anomaly detection of the drilling torque assuming the drill pipe failure, the prediction of core recovery rate as well as lithology. This study aims to achieve real-time drilling state identification. Accordingly, a drilling data acquisition and distribution system was developed. The drilling data distributed from the system is read by data analysis systems in several languages for real-time analysis. The drilling state identification models created in Python include the anomaly detection model and the prediction models of the core recovery and lithology. The models were installed in the real-time drilling data analyzing system, and real-time drilling state identification was attempted during the operation to confirm the health of the real-time drilling data analyzing system and to demonstrate identification with machine learning.

In recent years, many disasters due to earthquakes and tsunamis occurred in various countries, including Japan. Two triple junctions exist below the Japanese archipelago and around Japan, where three plates meet to form a complicated crustal structure. To conduct research on the Tohoku earthquake, which accompanied the tsunami that devastated northern Japan in March 2011, Japan Agency for Marine-Earth Science and Technology (JAMSTEC) constructed the scientific deep-sea drilling vessel Chikyu. To understand the mechanism of the earthquakes, a direct access to the fault zone is necessary. During drilling operations, the drill pipes are stressed by different forces and the occurrence of drill-string failures is high, which causes significant loss of money and time. This study focused on the definition of a practical fatigue assessment procedure useful to avoid failures during drilling operations. In this paper the proposed procedure is outlined, and the sensitivity analysis carried out to highlight most influencing parameters is presented.

The Stick-Slip is one of troublesome problems which happens in scientific drilling. In order to clarify the mechanism of Stick-Slip phenomenon, it is necessary to develop a reliable numerical model [1]-[4] to estimate drill bit motion for various drilling conditions. We have investigated on an NDDE (Neutral Delay Differential Equation) [5] to solve Stick-Slip of drill pipe [6]-[8]. In order to solve this NDDE, it is necessary to determine the frictional torque acting on the drill bit. In the previous study, we have employed the Balanov’s [4] frictional torque model which is the function of drill bit rotation with four unknown model parameters. This numerical model calculates the Stick-Slip phenomena of the drill pipe model in the laboratory experiments correctly under the appropriate four model parameter settings. On the other hand, it is quite difficult to calculate the actual drill pipe Stick-Slip [9], because the calculated results change drastically depending on the four unknown parameter settings. In the present study, the frictional torque acting on the drill bit is defined as a periodic function and solved the NDDE to calculate the stick-slip of the actual drill pipe. The average, variation and period of the frictional torque on the drill bit were set under the consideration of measured top drive torques. The calculated numerical solutions of the drill bit motion and the top drive torque agreed qualitatively with the measured ones.

During riserless drilling operations conducted in some scientific drillings and the initial stages of all oil and gas drilling operations, drill pipe motions such as vortex induced vibration, whirl motion, and motion due to the Magnus effect are generated. The last motion represents an interesting and important phenomenon that generates a lift force in addition to a drag force due to the ocean current and the rotation of the drill pipe. Accordingly, this study focuses on the drill pipe motions owing to the Magnus effect. An analytical model of a drill pipe was established by applying an absolute nodal coordinate formulation (ANCF) that can capture the behavior of a relatively flexible and long pipe, such as a drill pipe. The lifting and drag forces are calculated using computational fluid dynamics (CFD), and the lift and drag coefficients are calculated for several different drill pipe rotational velocities and ocean current velocities. A series of model experiments were conducted in a towing tank, with changing water flow velocities and rotational speed of the drill pipe model to observe the corresponding changes in the Magnus effect and to measure the resulting drill pipe motions. Additionally, the resulting drag and lift forces were measured. It was observed from the experiments that the motions in the cross-flow direction increased as the rotational speed of the drill pipe model increased, and that the lifting force increased as the rotational speed increased. The drill pipe motions were then simulated using a previously established analytical model and the results of the CFD simulations. The results of the simulations were evaluated against the results of the experiments, and reasons for observed discrepancies are discussed.

The scientific drilling vessel Chikyu is performing Nankai Trough Seismogenic Zone Experiment (NanTroSEIZE), a challenging deep drilling activity, for scientific purpose. We faced difficulty to drill deep during past NanTroSEIZE operations due to unstable sediments and insufficient cutting removal. Non-Stop Driller concept is, therefore, applied for the operation of NanTroSEIZE scheduled to start Oct. 2018 to enable continuous circulation of drilling fluid circulation. The Non-Stop Driller concept requires an additional, specially-designed sub called an “NSD sub” with a ball valve for drilling fluid inlet. Generally, the fatigue strength of a drill pipe is a critical factor governing the performance of challenging deep drilling. This study, therefore, focused on the fatigue failure of the NSD sub due to the bending stress caused by interference with risers including flex joints, ship structure, or drilling equipment resulting from ship motions. The bending stress leads to cyclic stress caused by rotation of the drill pipe. This is especially the case at the Nankai Trough where ocean currents are very strong reaching or sometimes exceeding 4 knots, a high bending stress is assumed to be exerted on the NSD sub. Full-scale fatigue tests of the NSD sub were first conducted to acquire the actual fatigue curve. Further, the bending stress distribution of a drill pipe, which refers to the locus of the bending stress during the drilling operation, was analyzed by considering interference of the drill string with the structure, drilling equipment, and risers that are deformed by the ocean current. Time-series ship motions is prepared using the response amplitude operators of the Chikyu for the sea states at Nankai Trough area, and then time-series stress response is obtained by considering the operational conditions such as rate of penetration and rotational velocity of drill pipe. The numbers of occurrence of each stress amplitude can be counted from the time-series stress response. Consequently, the cumulative damage ratio is calculated for evaluating the fatigue of the NSD sub. The results confirmed that the cumulative fatigue is within a safe range. This study focused on the evaluation of the fatigue strength of the specially designed NSD sub for the challenging scientific drilling operation at Nankai Trough, a harsh environment because of the presence of strong ocean currents. This paper presents the overview of NanTroSEIZE including the Non-Stop Driller concept, and the results of fatigue evaluations.

The scientific drillship Chikyu, operated by JAMSTEC has faced the problems with drill pipe damage and failure. After a failure, we examined the surface-measured drilling data, including Fourier transform analysis and spectrogram, and conducted an outlier investigation with an autoregressive model. We found some possible indications of top drive torque anomalies among the surface-measured drilling data. The purpose of this study is to attempt anomaly detection of top drive torque by applying machine learning approaches to such pre-examinations. Data acquired during the drilling operations of the Chikyu are first analyzed to observe the operations and determine the characteristics of the drilling data. Several hundred sets of surface drilling data are extracted and prepared as labeled learning data. Machine learning is then performed using several algorithms to build predictive models for anomaly detection of top drive torque. Simulations using these predictive models are then conducted for other sets of drilling data, including the anomaly conditions acquired during the operations of the Chikyu. The results demonstrate the validity of the predictive models. This study applies machine learning technique to the anomaly detection of the top drive torque with aim of preventing the drill pipe damage and failure. This paper describes our approaches in detail and discuss the results.

Offshore drilling with drill string over 10,000m long has many technical challenges. Among them, the challenge to control the weight on bit (WOB) between a certain range is inevitable for the integrity of drill pipes and the efficiency of the drilling operation. Since WOB cannot be monitored directly during drilling, the tension at the top of the drill string is used as an indicator of the WOB. However, WOB and the surface measured tension are known to show different features. The deviation among the two is due to the dynamic longitudinal behavior of the drill string, which becomes stronger as the drill string gets longer and more elastic. One feature of the difference is related to the occurrence of high-frequency oscillation. We have analyzed the longitudinal behavior of drill string with lumped-mass model and captured the descriptive behavior of such phenomena. However, such physics-based models are not sufficient for real-time operation. There are many unknown parameters that need to be tuned to fit the actual operating conditions. In addition, the huge and complex drilling system will have non-linear behavior, especially near the drilling annulus. These features will only be captured in the data obtained during operation. The proposed hybrid model is a combination of physics-based models and data-driven models. The basic idea is to utilize data-driven techniques to integrate the obtained data during operation into the physics-based model. There are many options on how far we integrate the data-driven techniques to the physics-based model. For example, we have been successful in estimating the WOB from the surface measured tension and the displacement of the drill string top with only recurrent neural networks (RNNs), provided we have enough data of WOB. Lack of WOB measurement cannot be avoided, so the amount of data needs to be increased by utilizing results from physics-based numerical models. The aim of the research is to find a good combination of the two models. In this paper, we will discuss several hybrid model configurations and its performance.

Stick-slip is one of the typical phenomenon which is observed in offshore drilling and considered as a critical problem for the drilling operation. The stick-slip makes a large fluctuation of drill bit rotation, even though the top of the drill pipe is rotating at a steady velocity and sometimes causes the damage of the drill bit. Additionally, it leads a crushing of the sediment layer which is a big problem especially for the scientific drilling [1][2][3]. The main purpose of the scientific drilling is to correct high quality core samples of sediment layers under the seabed. However, once the stick-slip occurs, it makes difficult to recover a high-quality sediment layer core sample. Therefore, it is necessary to detect the occurrence of stick-slip and its fundamental characteristics such as oscillation periods and amplitudes by simulation with the aid of surface drilling data, which can be monitored during the drilling operation to mitigate or prevent stick-slip. It would be advantageous to identify the characteristics of the stick-slip from the surface drilling data. The past study [4][5][6] investigated a numerical method to analyze the stick-slip by solving the NDDE (Neutral Delay Differential Equation) which is derived from torsional vibration equation. A small-scale model experiment was conducted in a water tank to observe the stick-slip phenomenon, and the result from the analytical model is evaluated with that obtained from the experiments. In this study, the numerical model is applied for the stick-slip analysis not only of the drill pipe model but also the actual drill pipe in operation. The solutions of the NDDE is depend on not the initial value but the initial history of the solution, because NDDE contains a delayed function term. Especially, the initial history settings have much effect on the numerical solution of NDDE in case of the actual drill pipe. Additionally, to solve the NDDE for stick-slip analysis, we must set some model parameters concerned with the frictional torque on drill bit. The present study investigated the effects of the initial history and the model parameters settings on numerical solutions in detail and presented an procedure to determine the appropriate settings of the initial history and the model parameters by reference to the measured top drive torque.

During riserless drilling operations, which are carried out in some scientific drillings and in the initial stages of all drilling operations in oil and gas exploration, a lifting force is generated in addition to a drag forces in ocean current environment owing to the ocean current and the rotation of the drill pipe. This is called the Magnus effect, and it is a critical phenomenon during such operations. First, the lifting and drag forces are calculated using the computational fluid dynamic (CFD), and the lift and drag coefficients are calculated for several rotational velocities of the drill pipe and the velocities of the ocean current. It can be observed through the calculations that the lifting force increases as the rotational velocity of the drill pipe increases, and it reaches a level of approximately several times that of the drag force. The force reaches such a considerably high magnitude that it can induce the motions of the drill pipe, resulting in the generation of a high bending moment. An analytical model of a drill pipe has been established by applying an absolute nodal coordinate formulation (ANCF), which can express a relatively flexible and long pipe, such as a drill pipe. ANCF is a finite element method, and was basically developed to analyze deformable linear objects such as the cable. With ANCF, the absolute slopes of elements are defined based on the absolute nodal coordinate. Finally, the drill pipe motions are simulated using the established model by applying the results of CFD simulations for sample cases and referencing the operation of the Chikyu.

In both of scientific and industrial drilling, it is not desirable to any have accompanying vibrations, large fluctuations, or other unexpected dynamics disturbing a steady and stable drilling process. One of the origins of such disturbances can be the flexibility of the drill pile; drill pipe is easily deformed by external perturbations because it is very long and slender. To specify the stable range of parameters of drill pipe, we theoretically and numerically analyzed a Cosserat-type mathematical model of drill pile. The values of the parameters of this model were set to appropriate values representing actual riserless drilling situations adopted in the operations of the scientific deep-sea drilling vessel Chikyu owned by the Japan Agency for Marine-Earth Science and Technology (JAMSTEC). We concluded that the buckling of pipe should be treated as a possible accident, but buckling does not happen in the normal operations of the Chikyu.

Stick-slip, a term for the spontaneous torsional vibrations occurring on a drill bit, is a critical problem in offshore drilling because of the damage it causes to the drill bits. Moreover, it leads to crushing or grinding of the sediment layer, which is a critical problem especially in scientific drilling where the purpose is to recover core samples from the layers. To detect the occurrence of stick-slip and its fundamental characteristics onboard in real-time, we developed an analytical model is developed to express the torsional motions at the drill bit, leading to the stick-slip phenomenon. The analytical model includes the unknown frictional torque at the drill bit, which depends on the properties of the drilling layers and is a dominant factor for demonstrating the stick-slip phenomenon. This study proposes a method to determine the frictional torque by integrating the analytical model with the surface drilling data. It would be advantageous to identify the characteristics of stick-slip from the surface drilling data. The method is validated with the results of small-scale model experiments in a water tank to observe the stick-slip phenomenon. Furthermore, we acquired the actual drilling data, which are surface drilling data and downhole data, during the drilling operation of the scientific drilling vessel Chikyu. This study focuses on attempts to estimate the stick-slip characteristics by the simulation based on the established analytical model with actual surface drilling data acquired during the drilling operation of the scientific drilling vessel Chikyu, and a comparison with the downhole data including the bit motions.

Japan Agency for Marine-Earth Science and Technology (JAMSTEC) operates the scientific deep-sea drilling vessel Chikyu, which undertook the drilling project under Japan Trench Fast Drilling Program (JFAST) to determine the mechanism of the catastrophic Tohoku Earthquake and the resulting huge tsunami. As targeted drilling sites were located about 1,000 m below the seafloor, at water depths of 7,000 m, JFAST was a challenging deep riserless drilling. Drilling operation can also provide information on the properties of the drilling layer. This is important as a main aim of the scientific drilling is to obtain core samples from sediment layers under the seabed for evaluating sediment properties. Even a rough estimate of sediment properties could potentially provide an important notice for conducting coring operations. Thus, we attempted to estimate properties of sediment layers using surface drilling data, such as drilling torque acquired from JFAST, and compare with properties of the core samples. As shear stress of sediment is a key parameter, it was estimated from surface drilling data. It was then compared with shear stress data determined from laboratory tests using core samples, and data from temperature sensors set in borehole.

The stick-slip is one of the critical problems for the scientific drilling, because it causes a crushing of the sampled layer. The present study investigates the characteristics of stick-slip phenomena of the drill pipe with the model experiments and numerical methods. The model experiments are carried out using a 1m length drill pipe model made with the Teflon. The angular velocity at the top and the bottom of the pipe are measured with the gyro sensor on some conditions of rotating speed at the pipe top and the weight on bit (load at the pipe bottom). The numerical simulations are also carried out to reproduce the stick-slip phenomena of the model experiments. The stick-slip is a kind of torsional vibration which is governed by the convection equation. By considering the boundary condition at the top and bottom of the pipe, we can obtain a neutral delayed differential equation (NDDE). The solutions of the NDDE is depend on not the initial value but the initial history of the solution, because NDDE contains a delayed function term. Therefore, it should be solved carefully to avoid the numerical error. The NDDE is solved with the 4th order Runge-Kutta scheme with very small time increment until the truncation error could be neglected. And also, we have found out that the effect of the initial history on the solution become to be very small after a certain period of time. The experimental results are compared with the numerical results under the same rotating condition. The experimental results of the stick-slip suggest that the period of the slip is mainly depend on the rotation speed at the pipe top and the magnitude of the slip is mainly depend on the weight on bit. Those characteristics of the stick-slip such as the period or the magnitude of slip are also obtained with the numerical calculations. However, in order to obtain an acceptable numerical results of NDDE, we have to adjust the frictional torque acting on the drill bit. Though, the frictional torque model was determined by reference to the measured torque at the top of the drill pipe model in the present study, it is desired to be improved. Therefore, the physical model of the frictional torque on the drill bit should be evaluated much carefully for the precise estimation of the stick slip in the future.

A crawler system is an advanced function that can potentially allow remotely operated vehicles (ROVs) to operate on the seafloor. A quasi-static theoretical motion model for a crawler ROV was established to facilitate evaluation and design that incorporate the characteristic parameters in water, including hydrodynamic forces and the effect of an underwater cable, that are expected to affect mobility. The results confirmed that these parameters considerably affect and decrease the mobility of crawler ROVs. The established motion model was applied to examine the mobility of a sample small-size crawler ROV deployed from a launching station.

Stick-slip is a major problem in offshore drilling because it may cause damage to the drill bit as well as crushing or grinding the sediment layer, which is crucial problem in scientific drilling because the purpose of the scientific drilling is to recover core samples from the layers. To mitigate stick-slip, first of all it is necessary to establish a model of the torsional motion of the drill bit and express the stick-slip phenomenon. Toward this end, the present study proposes a model of torsional waves propagating in a drillstring. An analytical model is developed and used to derive a neutral delay differential equation (NDDE), a special type of equation that requires time history, and an analytical model of stick-slip is derived for friction models between the drill bit and the layer as well as the rotation speed applied to the uppermost part of the drill string. In this study, the stick-slip model is numerically analyzed for several conditions and a time series of the bit motions is obtained. Based on the analytical results, the appearance of stick-slip and its severity are discussed. A small-scale model experiment was conducted in a water tank to observe the stick-slip phenomenon, and the result is discussed with numerical analysis. In addition, utilizing surface drilling data acquired from the actual drilling operations of the scientific drillship Chikyu , occurrence of stick-slip phenomenon is discussed.

A ROV (Remotely Operated Vehicle) which has a crawler based driving system is considered to be one of the appropriate underwater vehicles for seafloor exploration or seabed resources development [1][2][3][4][5][6][7]. The crawler driven ROV is able to move on sea floor, stay on a fixed sea bottom location and is capable to do heavy works such as digging the seafloor. In order to utilize those capabilities, it is important to know the fundamental moving capability of crawler driven ROV. According to the previous investigations [8][9], the crawler driven ROVs are easy to run in bow-up attitude in some running conditions due to the buoyancy and the hydrodynamic forces acting on the ROV. This irregular running sometimes causes a turning over. Therefore, we have to know the restrictions on the design parameters of the ROV not to run in bow-up attitude to design the ROV. The authors have been investigating the moving capability of crawler driven ROV and showed a method to estimate the restrictions of design parameters to avoid the bow-up running, which is called normal running condition [10][11][12][13]. This method is based on a simple dynamic model which considers the forces acting on ROV as concentrated loads; those are gravity, buoyancy, reaction from the ground, thrust and hydrodynamic drag. The loading position of ground reaction in steady running is obtained from the balance condition of forces. We consider that the loading point of ground reaction should be inside between the fore and rear wheels for the normal running. This constrained condition indicates the relation between gravity and buoyancy center locations for the normal run of ROV under the given body geometry, weight, displacement and running speed. This method estimates the ROV’s running capability in acceptable accuracy compared with the model experiments. However, this method does not consider the tension of the cable which is connected to the ROV. As you can easily imagine, the cable tension has a big influence on the movable area of the ROV. If the ROV keeps going forward, it will turn over due to the tension of the cable at a certain point. We must know the movable area of the crawler driven ROV for the operation planning. The present study shows a method to estimate the movable area of the crawler driven ROV under the restriction of the cable by extending the previous method to estimate the normal running condition.

During drilling operations, the scientific drilling vessel Chikyu encounters vortex-induced vibration (VIV) of the riser during riser drilling as well as that of the drill pipe during riserless drilling. The characteristics of the VIV of the drill pipe will differ from those of the riser because the drill pipe rotates during operations. Therefore, investigating the characteristics of the VIV during drill pipe rotation is important. In this paper, the results of model experiments that show the characteristics of the VIV of a rotating drill pipe and drill pipe whirl are described. The authors conducted model experiments on the VIV response of a rotating cylinder in a water tank to observe the motion of the rotating cylinder. The results confirmed that rotating affects the characteristics of the VIV. When the rotating frequency of the drill pipe was low, the VIV resonated with the rotation; as the rotating frequency increased, the frequency of the VIV decreased. In addition, a self-excited whirl motion was also observed during the experiments. This “drill pipe whirl,” which rotated in the direction opposite to the rotation of the drill pipe at a low frequency, was larger in amplitude than the VIV. The results demonstrate that rotation of the drill pipe affects the characteristics of the VIV and generates the whirl motion.

The Remotely Operated Vehicle, so called “ROV” which has crawler based moving system is considered as one of the appropriate underwater vehicles for seafloor exploration or seabed resources development [1][2][3][4][5][6][7]. The advantages of crawler driven ROV are to be able to stay on a fixed sea bottom location and to be capable to do heavy works such as digging the seafloor. However, the ROV moving on the sea bottom with crawler based driving system easily turn over due to the buoyancy and hydrodynamic forces [8][9][10][11][12]. Therefore, it is important to know the moving capability of the ROV on the sea bottom for the design point of view. The authors have shown the condition for the normal running of the ROV which moves on horizontal and inclined flat sea bottom by means of a simple dynamic model [11]. Normal running means that the ROV runs without bow-up or stern-up situations and the crawlers touch the ground normally. The normal running condition of ROV indicates the constrained condition of the relation between gravity and buoyancy center locations for any given design parameters such as geometry, weight, displacement and running speed of the ROV. Though this method estimates the ROVs’ moving capability with acceptable accuracy, the hydrodynamic forces on the ROV and its application point are required for accurate estimation. In the previous research, those quantities are roughly estimated from the past experimental investigations. The present study investigated the flow around the crawler driven ROV which runs on seafloor with CFD (Computational Fluid Dynamics) analysis to evaluate the characteristics of hydrodynamic forces acting on the ROV. The open source CFD code, OpenFOAM [13] was applied for flow calculation and the results were validated with model experiments. By using the calculated hydrodynamic forces on ROV, the moving capability of ROV was evaluated with a method the authors had shown. The estimates of the running capability of the ROV by using the CFD calculations are quite different from past estimations in some running conditions.

The scientific drilling vessel Chikyu was designed to have the deep drilling capability to reach the deep earthquake zone. To realize such deep drilling, a drill pipe with higher than ever strength and reliability is of necessity. A strength evaluation of such high strength drill pipe is also necessary. Around Japan, the earthquake zones are widely located under the seabed in deep water. For example, the Nankai Trough located beneath the ocean off the southwest coast of Japan is one of the most active earthquake zones where large-scale earthquakes have occurred repeatedly throughout history. Thus, the Chikyu has started the first scientific drilling at the Nankai Trough as the Nankai Trough Seismogenic Zone Experiment (NanTroSEIZE). NanTroSEIZE targets the megasplay fault zone at 3500 m below the seafloor and finally 6000m deep drilling into the seismogenic zone and across the plate interface into the subducting crust at water depths of around 2500m. In addition, a huge earthquake zone is expected to be located about 1000m below the seafloor at water depths of around 7000m. For this drilling task, the riserless drilling technique should be applied. To realize such deep drilling with both riser and riserless techniques, a S150 drill pipe was developed during the construction phase and has been used in the past scientific drillings of the Chikyu. For deeper drillings in the future and drilling operations in harsh environments, we are developing superior high strength drill pipes, S155 and S160, possessing high reliability including corrosion resistance to achieve high toughness and reduction of stress concentration. An evaluation of the maximum possible stress was conducted. In the maximum strength evaluation, we considered dynamic stress and bending stress due to the current and the vessel inclination. This paper describes the development of superior high strength drill pipes and the strength evaluation of such drill pipes for deep earthquake zone drilling.