|Theme||Visible||Selectable||Appearance||Zoom Range (now: 0)|
Abstract Uniformity of proppant distribution among multiple perforation clusters affects treatment efficiency in multistage fractured wells stimulated using the plug-and-perf technique. Multiple physical phenomena taking place in the well and perforation tunnels can cause uneven proppant distribution among multiple clusters. The problem has been studied in the recent years with experimental and computational fluid dynamics (CFD) methods, which provide useful insights but are impractical for routine designs. Simplified models that incorporated the proppant transport efficiency (PTE) correlation derived from the CFD results in a hydraulic fracture model have been also presented in literature. In this paper, we present a numerical model that simulates the transient proppant slurry flow in the wellbore, considering proppant transport and settling including bed formation, rate- and concentration-dependent pressure drop, PTE, and dynamic pressure coupling with the hydraulic fractures. The model is efficient and is designed to be an independent wellbore transport model so it can be integrated with any fracture models, including fully 3D and/or complex fracture network models, for practical design optimization. The model predictions are compared and found to agree with previously published studies. Parametric studies demonstrate sensitivity of proppant distribution to grain size, fluid viscosity, and pumping rate for fixed perforation designs. Analysis of the simulation results shows that the dominant cause of uneven proppant distribution is proppant inertia. Possible slurry stratification is less important, except for the cases with relatively low flow rates and near toe clusters. Accordingly, proppant distribution is less sensitive to perforation phasing than to the number of perforations in clusters. Alterations of the number of perforations per cluster within a stage enable achieving more even proppant distribution.
Thiessen, Scott (Hunting Energy Services - Titan Division) | Han, Oliver (Hunting Energy Services - Titan Division) | Ahmed, Ramadan (University of Oklahoma) | Elgaddafi, Rida (University of Oklahoma)
ABSTRACT In hydraulic fracturing, determining the perforation pressure loss is a critical step in the design strategy, on-site troubleshooting diagnostics and post-fracture analysis. Historically, the most widely assumed and thus unknown components in the perforation friction equationare the coefficient of discharge and the holistic perforation diameter. The perforation coefficient of discharge has long been assumed as a dynamic variable dependent on the amount of fluid and proppant pumped through the perforations. This variable becomes increasingly important when clusters are spaced closer together and fewer perforations are shot such as in a limited entry design. Limited entry is a perforating technique used to generate uniform fractures along the wellbore by creating appropriate pressure differentials from cluster to cluster. With the adoption of consistent hole perforating shaped charges, the perforating diameters are more consistent and predictable. While not all consistent hole shaped charges have low diameter variability, the perforating diameters downhole are no longer an unknown, particularly after the introduction of downhole cameras. Therefore, the coefficient of discharge is the only unknown variable remaining. This paper presents an experimental methodology to accurately define the true coefficient of discharge in common completions perforated by a known consistent hole shaped charge. The test setup is illustrated, detailed test steps are discussed, and experimental data with correlations of rate per perforation and discharge coefficient is presented. Completions tested included 4-1/2", 5", and 5-1/2" casings in common weights and grades. Various perforating strategies were examined such as single shot and angled shot. Critical parameters such as entry hole diameters were made by the actual shaped charges and measured before and after the test. Freshwater and slickwater were used as hydraulic fluid and circulated at real-world pump rates through each perforation to simulate the actual field flow conditions. Based on the study, several correlations for the coefficient of discharge of flow through a perforation are created considering casing thickness, entry hole diameter and rate per perforation for the given consistent hole shaped charges. These correlations can improve perforation and fracturing designs where perforation friction are important variables.
Gudala, Manojkumar (Indian Institute of Technology, Madras) | Naiya, Tarun Kumar (Indian Institute of Technology (Indian School of Mines) Dhanbad) | Govindarajan, Suresh Kumar (Indian Institute of Technology, Madras)
Summary The present work focuses on the improvement of flow properties during the transportation of heavy oil via 0.0254-, 0.0381-, and 0.0508-m-diameter pipelines. The effect of temperature, water cut, natural extract Madhuca Longifolia (ML), and potato starch (PS) on pressure drop, shear viscosity, and flow behavior index (n) was experimentally investigated. Minimum pressure drop occurred in the 0.0508-m-inner-diameter (ID) pipeline because of the combined consequence of temperature and 2,000 ppm ML during the transportation of 85% heavy oil þ 15% water. A new correlation was developed to predict the friction factor for the heavy oil/emulsions during its transportation in a 0.0254-m-ID pipeline using the linear regression method for friction factor. Flow behavior index inclined toward Newtonian from shear-thinning behavior (i.e., n ¼ 0.2181 to 0.9834) after the addition of 2,000 ppm ML at 50 A new hybrid artificial intelligence (AI) technique was developed and used to optimize flow-influencing parameters to minimize the pressure drop and shear viscosity and improve flow behavior index. Minimum pressure drop (58,659.72 Pa), shear viscosity (1.56 Pas), and maximum flow behavior index (0.71) were achieved during the heavy oil flow in the 0.0508-m-ID pipeline after addition of 15% water, 1,320 ppm ML for 12.33-m However, from the studies, it was concluded that ML shows better performance compared with PS. Because both ML and PS are biodegradable and nontoxic, the petroleum industry may use both as a cost-effective alternative to decrease pour point and improve flowability for heavy crude oil. Introduction Because of a decrease in conventional light crude oil reserves, heavy oil has gained importance in the global hydrocarbon market in the last decade (Hein 2017). This fact leads to an increased economic interest in the exploration and production of heavy oil. The high viscosity of oil creates primitive problems in production and transportation (Speight 2013; Nadirah et al. 2014). It leads to the requirement of high pumping pressure, power, and consequently damage to the equipment, ultimately increasing the cost of heavy oil production and transportation. Heavy oil viscosity is decreased either by heating or emulsification or the addition of water with drag-reducing additives (DRAs) (Omer and Pal 2010; Martínez-Palou et al. 2011; Hart 2014).
Summary There is a great deal of interest in the oil and gas industry (OGI) in seeking ways to implement machine learning (ML) to provide valuable insights for increased profitability. With buzzwords such as data analytics, ML, artificial intelligence (AI), and so forth, the curiosity of typical drilling practitioners and researchers is piqued. While a few review papers summarize the application of ML in the OGI, such as Noshi and Schubert (2018), they only provide simple summaries of ML applications without detailed and practical steps that benefit OGI practitioners interested in incorporating ML into their workflow. This paper addresses this gap by systematically reviewing a variety of recent publications to identify the problems posed by oil and gas practitioners and researchers in drilling operations. Analyses are also performed to determine which algorithms are most widely used and in which area of oilwell-drilling operations these algorithms are being used. Deep dives are performed into representative case studies that use ML techniques to address the challenges of oilwell drilling. This study summarizes what ML techniques are used to resolve the challenges faced, and what input parameters are needed for these ML algorithms. The optimal size of the data set necessary is included, and in some cases where to obtain the data set for efficient implementation is also included. Thus, we break down the ML workflow into the three phases commonly used in the input/process/output model. Simplifying the ML applications into this model is expected to help define the appropriate tools to be used for different problems. In this work, data on the required input, appropriate ML method, and the desired output are extracted from representative case studies in the literature of the last decade. The results show that artificial neural networks (ANNs), support vector machines (SVMs), and regression are the most used ML algorithms in drilling, accounting for 18, 17, and 13%, respectively, of all the cases analyzed in this paper. Of the representative case studies, 60% implemented these and other ML techniques to predict the rate of penetration (ROP), differential pipe sticking (DPS), drillstring vibration, or other drilling events. Prediction of rheological properties of drilling fluids and estimation of the formation properties was performed in 22% of the publications reviewed. Some other aspects of drilling in which ML was applied were well planning (5%), pressure management (3%), and well placement (3%). From the results, the top ML algorithms used in the drilling industry are versatile algorithms that are easily applicable in almost any situation. The presentation of the ML workflow in different aspects of drilling is expected to help both drilling practitioners and researchers. Several step-by-step guidelines available in the publications reviewed here will guide the implementation of these algorithms in the resolution of drilling challenges.
Summary Decline curve analysis (DCA) has been the mainstay in unconventional reservoir evaluation. Because of the extremely low matrix permeability, each well is evaluated economically for ultimate recovery as if it were its own reservoir. Classification and normalization of well potential is difficult because of ever-changing stimulation total contact area and a hyperbolic curve fit parameter that is disconnected from any traditional reservoir characterization descriptor. A new discrete fracture model approach allows direct modeling of inflow performance in terms of fracture geometry, drainage volume shape, and matrix permeability. Running such a model with variable geometrical input to match the data in lieu of standard regression techniques allows extraction of a meaningful parameter set for reservoir characterization, an expected outcome from all conventional well testing. Because the entirety of unconventional well operation is in transient mode, the discrete fractured well solution to the diffusivity equation is used to model temporal well performance. The analytical solution to the diffusivity equation for a line source or a 2D fracture operating under constrained bottomhole pressure consists of a sum of terms, each with exponential damping with time. Each of these terms has a relationship with the constant rate, semisteady-state solution for inflow, although the well is not operated with constant rate, nor will this flow regime ever be realized. The new model is compared with known literature models, and sensitivity analyses are presented for variable geometry to illustrate the depiction of different time regimes naturally falling out of the unified diffusivity equation solution for discrete fractures. We demonstrate that apparent hyperbolic character transitioning to exponential decline can be modeled directly with this new methodology without the need to define any crossover point. The mathematical solution to the physical problem captures the rate transient functionality and any and all transitions. Each exponential term in the model is related to the various possible interferences that may develop, each occurring at a different time, thus yielding geometrical information about the drainage pattern or development of fracture interference within the context of ultralow matrix permeability. Previous results analyzed by traditional DCA can be reinterpreted with this model to yield an alternate set of descriptors. The approach can be used to characterize the efficacy of evolving stimulation practices in terms of geometry within the same field and thus contribute to the current type curve analyses subject to binning. It enables the possibility of intermixing of vertical and horizontal well performance information as simply gathering systems of different geometry operating in the same reservoir. The new method will assist in reservoir characterization and evaluation of evolving stimulation technologies in the same field and allow classification of new type curves.
Dooply, Mohammed (Schlumberger) | Schupbach, Michael (Murphy Exploration & Production Co) | Hampshire, Kenneth (Murphy Exploration & Production Co) | Contreras, Jose (Schlumberger) | Flamant, Nicolas (Schlumberger)
Summary Two of the most important parameters to monitor during a primary cementing job are the flow rate in and return flow rate measurements. To achieve optimum job results of a primary cementing job, measuring annular return rates and comparing them with simulated data in real time will provide a better understanding of job signatures and result in the best possible top of cement (TOC) estimation prior to running any cement evaluation log or making a decision to continue drilling the next section of the well. The return rate job signature along with the wellhead pressure is essential to understanding the behavior and discrepancies between simulated and acquired surface data. Therefore, to assess the risk of job issues, such as unsuspected washout and lost circulation among others, accurate measurements of the return rate are critical. Historically, the cement job evaluation has been limited by the fact that most drilling rigs do not have an accurate flowmeter installed on the annulus return line, and a simple verification of mud tanks volume vs. pumped volume, as reported by drillers or mud loggers, more often than not results in an unreliable assessment of the volume lost downhole, due to the unfamiliarity with the U-tubing effect and lack of data consolidation from the cement unit (flow rate in) and the rig (flow rate in and flow rate out). In this paper, we will review a solution developed to mitigate the lack of a direct flow-rate measurement by computing and displaying the return rate using either a paddle meter measurement or the derivative over time of the volume observed in the rig tanks.
Elgaddafi, Rida (University of Oklahoma) | Ahmed, Ramadan (University of Oklahoma (Corresponding author) | Karami, Hamidreza (email: firstname.lastname@example.org)) | Nasser, Mustafa (University of Oklahoma) | Hussein, Ibnelwaleed (Qatar University)
Summary The accumulation of rock cuttings, proppant, and other solid debris in the wellbore caused by inadequate cleanout remarkably impedes field operations. The cuttings removal process becomes a more challenging task as the coiled-tubing techniques are used during drilling and fracturing operations. This article presents a new hole cleaning model, which calculates the critical transport velocity (CTV) in conventional and fibrous water-based fluids. The study is aimed to establish an accurate mechanistic model for optimizing wellbore cleanout in horizontal and inclined wells. The new CTV model is established to predict the initiation of bed particle movement during cleanout operations. The model is formulated considering the impact of fiber using a special drag coefficient (i.e., fiber drag coefficient), which represents the mechanical and hydrodynamic actions of suspended fiber particles and their network. The dominant forces acting on a single bed particle are considered to develop the model. Furthermore, to enhance the precision of the model, recently developed hydraulic correlations are used to compute the average bed shear stress, which is required to determine the CTV. In horizontal and highly deviated wells, the wellbore geometry is often eccentric, resulting in the formation of flow stagnant zones that are difficult to clean. The bed shear stress in these zones is sensitive to the bed thickness. The existing wellbore cleanout models do not account for the variation in bed shear stress. Thus, their accuracy is limited when stagnant zones are formed. The new model addresses this problem by incorporating hydraulic correlations to account for bed shear stress variation with bed height. The accuracy of the new model is validated with published measurements and compared with the precision of an existing model. The use of fiber drag and bed shear stress correlations has improved model accuracy and aided in capturing the contribution of fiber in improving wellbore cleanout. As a result, for fibrous and conventional water-based fluids, the predictions of the new model have demonstrated good agreement with experimental measurements and provided better predictions than the existing model. Model predictions show a noticeable reduction in fluid circulation rate caused by the addition of a small quantity of fiber (0.04% w/w) in the fluid. In addition, results show that the existing model overpredicts the cleaning performance of both conventional and fibrous water-basedmuds.
Md Yusof, Muhammad Aslam (Universiti Teknologi PETRONAS) | Ibrahim, Mohamad Arif (Universiti Teknologi Malaysia) | Mohamed, Muhammad Azfar (Universiti Teknologi PETRONAS) | Md Akhir, Nur Asyraf (Universiti Teknologi PETRONAS) | M Saaid, Ismail (Universiti Teknologi PETRONAS) | Ziaudin Ahamed, Muhammad Nabil (Universiti Teknologi PETRONAS) | Idris, Ahmad Kamal (Universiti Teknologi Malaysia) | Awangku Matali, Awangku Alizul (Vestigo Petroleum)
Abstract Recent studies indicated that reactive interactions between carbon dioxide (CO2), brine, and rock during CO2 sequestration can cause salt precipitation and fines migration. These mechanisms can severely impair the permeability of sandstone which directly affect the injectivity of supercritical CO2 (scCO2). Previous CO2 injectivity change models are ascribed by porosity change due to salt precipitation without considering the alteration contributed by the migration of particles. Therefore, this paper presents the application of response surface methodology to predict the CO2 injectivity change resulting from the combination of salt precipitation and fines migration. The impacts of independent and combined interactions between CO2, brine, and rock parameters were also evaluated by injecting scCO2 into brine saturated sandstone. The core samples were saturated with NaCl brine with salinity between 6,000 ppm to 100,000 ppm. The 0.1, 0.3, and 0.5 wt.% of different-sized hydrophilic silicon dioxide particles (0.005, 0.015, and 0.060 μm) were added to evaluate the effect of fines migration on CO2 injectivity alteration. The pressure drop profiles were recorded throughout the injection process and the CO2 injectivity alteration was represented by the ratio between the initial and final injectivity. The experimental results showed that brine salinity has a greater individual influence on permeability reduction as compared to the influence of particles (jamming ratio and particle concentration) and scCO2 injection flow rate. Moreover, the presence of both fines migration and salt precipitation during CO2 injection was also found to intensify the permeability reduction by 10%, and reaching up to threefold with increasing brine salinity and particle size. The most significant reductions in permeability were observed at higher brine salinities, as more salts are being precipitated out which, in turn, reduces the available pore spaces and leads to a higher jamming ratio. Thus, more particles were blocked and plugged especially at the slimmer pore throats. Based on comprehensive 45 core flooding experimental data, the newly developed model was able to capture a precise correlation between four input variables (brine salinity, injection flow rate, jamming ratio, and particle concentration) and CO2 injectivity changes. The relationship was also statistically validated with reported data from five case studies.
Abstract To increase the production and recovery of marginal, mature, and challenging oil reservoirs, developing new inflow control technologies is of great importance. In cases where production of surrounding reservoir fluids such as gas and water can cause negative effects on both the total oil recovery and the amounts of energy required to drain the reservoir, the multiphase flow performances of these technologies are of particular significance. In typical cases, a Long Horizontal Well (LHW) will eventually start producing increasing amounts of these fluids. This will cause the Water Cut (WC) and/or Gas Oil Ratio (GOR) to rise, ultimately forcing the well to be shut down even though there still are considerable amounts of oil left in the reservoir. In earlier cases, Inflow Control Devices (ICD) and Autonomous Inflow Control Devices (AICD) have proven to limit these challenges and increase the total recovery by balancing the influx along the well and delaying the breakthrough of gas and/or water. The Autonomous Inflow Control Valve (AICV) builds on these same principles, and in addition has the ability to autonomously close when breakthrough of unwanted gas and/or water occurs. This will even out the total drawdown in the well, allowing it to continue producing without the WC and/or GOR reaching inacceptable limits. As part of the qualification program of the light-oil AICV, extensive flow performance tests have been carried out in a multiphase flow loop test rig. The tests have been performed under realistic reservoir conditions with respect to variables such as pressure and temperature, with model oil, water, and gas at different WC's and GOR's. Conducting these multiphase experiments has been valuable in the process of establishing the AICV's multiphase flow behavior, and the results are presented and discussed in this paper. Single phase performance and a comparison with a conventional ICD are also presented. The results display that the AICV shows significantly better performance than the ICD, both for single and multiphase flow. A static reservoir modelling method have been used to evaluate the AICV performance in a light-oil reservoir. When compared to a screen-only completion and an ICD completion, the simulation shows that a completion with AICV's will outperform the above-mentioned completions with respect to WC and GOR behavior. A discussion on how this novel AICV can be utilized in marginal, mature, and other challenging reservoirs will be provided in the paper.
Wang, Fuyong (China University of Petroleum, Beijing) | Zai, Yun (China University of Petroleum, Beijing) | Zhao, Jiuyu (China University of Petroleum, Beijing) | Fang, Siyi (China University of Petroleum, Beijing)
Abstract Well real-time flow rate is one of the most important production parameters in oilfield and accurate flow rate information is crucial for production monitoring and optimization. With the wide application of permanent downhole gauge (PDG), the high-frequency and large volume of downhole temperature and pressure make applying of deep learning technique to predict flow rate possible. Flow rate of production well is predicted with long short-term memory (LSTM) network using downhole temperature and pressure production data. The specific parameters of LSTM neural network are given, as well as the methods of data preprocessing and neural network training. The developed model has been validated with two production wells in the Volve Oilfield, North Sea. The field application demonstrates that the deep learning is applicable for flow rate prediction in oilfields. LSTM has the better performance of flow rate prediction than other five machine learning methods, including support vector machine (SVM), linear regression, tree, and Gaussian process regression. The LSTM with a dropout layer has a better performance than a standard LSTM network. The optimal numbers of LSTM layers and hidden units can be adjusted to obtain the best prediction results, but more LSTM layers and hidden units lead to more time of training and prediction, and LSTM model might be unstable and cannot converge. Compared with only downhole pressure or temperature data used as input parameters, flow rate prediction with both of downhole pressure and temperature used as input parameters has the higher prediction accuracy.