The last decade has spotted a tremendous upsurge in casing failures. The aftermaths of casing failure can include the possibility of blowouts, environmental pollution, injuries/fatalities, and loss of the entire well to name a few. The motivation behind this work is to present findings from a predictive analytics investigation of casing failure data using supervised and unsupervised data mining algorithms. Scientists and researchers have speculated the underlying causes but to date this type of work remains unpublished and unavailable in the public domain literature.
This study assembled comprehensive data from eighty land-based wells during drilling, fracturing, workover jobs, and production. Twenty wells suffered from casing failure while the remaining sixty offset wells were compiled from well reports, fracturing treatment data, drilling records, and recovered casing data. The failures were unsystemic but included fatigue failure, bending stresses from excessive dogleg, buckling, high hoop stress on connections, and split coupling. The failures were detected at various depths, both in cemented and uncemented hole sections. Failures were spotted at the upper and lower production casing.
Using a predictive analytics software from SAS, twenty-four to twenty-six variables were evaluated through the application of various data mining techniques on the failed casing data sets. The missing data was accounted for using multivariate normal imputation. The study outcome addressed common casing sizes and couplings involved with each failure, failure location, hydraulic fracturing stages, cement impairment, dogleg severity, thermal and tensile loads, production-induced shearing, and DLS. The predictive algorithms used in this study included Logistic Regression, Hierarchal Clustering, and Decision Trees. While the descriptive analytics manifested in visual representations included Scatterplot Matrix, and PivotTables. Failure causes were identified. A total of five statistical techniques using the aforementioned algorithms were developed to evaluate the concurrent effect of the interplay of these variables. Nineteen variables were believed to possess the highest contribution to failure. Scatterplot matrix suggested a complex correlation between the total base water used in fracturing simulation and casing thickness. Logistic Regression suggested nine variables were significant including: TVD, operator, frac start month, MD of most severe DL, heel TVD, hole size, BHT, total proppant mass, cumulative DLS in lateral and build sections variables as significant failure contributors. PivotTables showed that the rate of casing failure was highest during the winter season.
This investigation is aimed to develop a thorough understanding of casing failures and the myriad of contributing factors to develop comprehensive predictive models for future failure prediction via the application of data mining algorithms. These models intend to provide a theoretical and statistical basis for cost-effective, safe, and better drilling practices.
Ouled Ameur, Zied (Cenovus Energy Inc) | Kudrashou, Viacheslau (Texas A&M Engineering) | Nasr-El-Din, Hisham A. (Texas A&M University) | Forsyth, Jeffrey (nFluids Inc) | Mahoney, John (Mahoney Geochemical Consulting) | Daigle, Barney (AkzoNobel)
The acidizing of sour, heavy-oil, weakly consolidated sandstone formations under steam injection is challenging because of fines migration, sand production, inorganic-scale formation, corrosion issues, and damage caused by asphaltene precipitation associated with these sandstone formations. These and other similar problems cause decline in the productivity of the wells, and there is a recurring need to stimulate them to restore productivity. The complexity of sandstone ormations requires a mixture of acids and several additives, especially at temperatures up to 360°F, to accomplish successful stimulation. Three treatments were tested on a horizontal well in the field: hydrochloric acid (HCl); Chelating Agent B, a high-pH chelant; and Chelating Agent A, or glutamic acid N,N-diacetic acid (GLDA). The first two treatments with 15 wt% HCl and high-pH (pH=10) Chelating Agent B produced results below expectations. The third treatment using GLDA was successful, and the well productivity increased significantly. The field treatment with GLDA included pumping the treatment fluid, which was foamed to create proper rheological characteristics and a better-controlled pumping process. The treatment fluids were displaced into the formation by pumping produced water and were allowed to soak for 6 hours. In this paper, we evaluate the field applications of GLDA using geochemical modeling, production data, and analysis of well-flowback fluids after the field treatments.
Group of microorganisms Example Metabolic Function Influence on corrosion Oxidizes Sulfide or other reduced sulfide pH may decrease locally.
Enhanced Oil Recovery (EOR) has been utilized in Trinidad and Tobago for over 50 years. Most projects so far have focused on thermal as well as gas injection along with the more conventional waterfloods. In spite of that, recovery factors are still relatively low and the country's oil production has been declining for some time. Surprisingly, given the progress in chemical EOR and in particular polymer flooding in the last 10 years, these processes have not been used in Trinidad and we suggest that it might be time to consider their application. Similarly, foam has been used extensively worldwide to improve performances of gas and steam injection but has not yet been used in the country.
The situation of EOR in Trinidad will be first reviewed along with the characteristics of the main reservoirs. Then the potential for the application of chemical-based EOR methods such as polymer, surfactant and foams will be studied by comparing the characteristics of Trinidad's reservoirs to others worldwide which have seen the applications of chemical-based EOR methods.
This review and screening suggests that there is no technical barrier to the application of all these EOR methods in Trinidad. Most reservoirs produce heavy oil and are heavily faulted, but polymer injection has been widely applied in heavy oil reservoirs as well as in faulted reservoirs before, and suitable examples will be provided in the paper. Similarly, these characteristics do not present any specific difficulty for foam-enhanced gas or steam injection. The main issue appears to be the identification of suitable water sources for the projects.
This paper proposes a new look at EOR opportunities in Trinidad using conventional methods which have not been used in the country. This will help reservoir engineers who are considering such applications in the country and hopefully will eventually result in an increase in the oil production in the future.
ABSTRACT: Wellbore instability and formation sand production pose potential risks for wellbore drilling, completion and production operations. In many sandstone reservoirs worldwide, sand production has been observed to accompany oil and gas production. In this study, we aim to estimate, predict and quantify wellbore instability and sand production potentials in the Hajdúszoboszló field, Pannonian Basin, Hungary, using the Mechanical Earth Model (MEM). Our study relies on petrophysical log data obtained from an onshore gas well within the field as input data. Our 1-D MEM utilizes a workflow that develops wellbore and sand failure mechanisms, first creating the mechanical stratigraphy of the reservoir rock; followed by estimating the pore pressure, overburden stress, rock strength, rock elastic properties, and horizontal stresses of the reservoir rock with reference to the depth of stratigraphic column, from compressional slowness, shear slowness, density, porosity and shale volume. Lastly, we conduct a wellbore stability and sand management analysis. Our results show the mechanical stratigraphy of unconsolidated sandstone and shale distribution in the reservoir, wellbore shear and tensile failures, wellbore breakout and breakdown pressures, wellbore sensitivity analysis, sanding interval analysis, critical drawdown pressure (CDDP) profile and sand failure zones. Based on careful observation of our results, we predict the wellbore intervals with high sand production potentials and wellbore instability within the reservoir formations. Therefore, we suggest significant wellbore failure during drilling process and also a high possibility of sand production into the wellbore during well completion at a formation interval of 550-937 m. Although there is need for data from additional wells in the field to be incorporated into our model prediction, we suggest that our preliminary model can be useful for critical decision making during drilling and completion operations across the Hajdúszoboszló field, Pannonian Basin, Hungary. In addition, our study provides a platform for further investigation into wellbore stability and sanding analysis in other parts of the Pannonian Basin where available well data can also be incorporated in our model.
Roostaei, M. (University of Alberta) | Guo, Y. (University of Alberta) | Velayati, A. (University of Alberta) | Nouri, A. (University of Alberta) | Fattahpour, V. (RGL Reservoir Management) | Mahmoudi, M. (RGL Reservoir Management)
ABSTRACT: Unconsolidated sand was packed on a slotted-liner coupon in large-scale sand retention tests (SRT) and was subjected to several stress conditions, corresponding to the evolving stress conditions during the life cycle of a SAGD producer. Cumulative produced sand at the end of testing was measured as the indicator for sand control performance. Retained permeability was calculated by measuring pressure drops near the liner and was considered as the quantification of the flow performance of the liner. Experimental results indicate the liner performance is significantly affected by the stress induced compaction of the oil sand. The stress results in the sand compaction, leading to a denser sand, hence, a lower porosity and permeability. The lower porosity results in a higher pore-scale flow velocity, which can trigger more fines mobilization, hence, a higher skin buildup. With respect to sanding, the higher stress can stabilize the sand bridges: Increased normal forces between near-slot sand particles result in a higher inter-particle friction, hence, more stable sand bridges and less produced sand. The lower and upper bounds of slot window are governed by plugging and sand production, respectively. Experimental results indicate an upward shift in both the lower and upper bounds at elevated stress conditions
Steam Assisted Gravity Drainage (SAGD) is a thermal recovery technology currently employed to extract heavy oil and high viscosity bitumen from Alberta oil sands.
Due to the unconsolidated nature of oil sands, SAGD wells are prone to producing sand, hence, requiring sand control devices to prevent sanding during oil production. Slotted liners are a prominent sand control technique, which have been extensively used in Alberta's SAGD wells to avoid sand production problems. The design of the slots must allow a free flow of fines and clays through the slots and the porous medium around the well, with minimal plugging.
In SAGD recovery method, a large volume of high-pressure steam is injected by the injector well to reduce the bitumen viscosity and facilitate the production. Continuous injection of the high-pressure steam leads to a complex alteration of the in-situ stresses and the associated geomechanical properties within the reservoir and even the neighboring strata. Porosity and permeability of the reservoir sand are influenced in this process.
The importance of uncertainty quantification and risks assessment in the petroleum industry cannot be overstated. Uncertainty will always be present in production forecasts and reserves estimates. Underestimation of uncertainty when estimating reserves and profitability of projects can lead to poor decision making and disappointment.
Water Displacement Curve (WDC) models are alternative to Decline Curve Analysis (DCS) models in waterflooded oil fields. These curves allow engineers to estimate reserves and forecast production performance in waterflooded oil reservoirs taking into account either liquid or water production. Compared with DCA, WDC models are expected to perform better in forecasting oil production in waterflooded oil fields.
In this study I applied WDC models probabilistically using Bayesian methodology and Markov Chain Monte Carlo (MCMC) stochastic simulation. I also developed a Multimodel approach based on eleven WDC models to quantify uncertainty in production forecasts by assessing differences in matches and forecasts provided by each model.
Both Multimodel and MCMC with WDC models were calibrated and compared to MCMC with DCA method which was recently developed and introduced by Gong et al. (2011). Reliability of the developed methods was assessed using production history of 100 wells from actual waterflooded oil fields. I performed hindcast studies in which I assumed that some fraction of the actual historical production data is known (6, 12, 24 and 36 months) and the rest of the actual production is unknown (5 – 7 years). I then matched the assumed known production fraction of the history and forecasted production to the end of the actual historical period. The cumulative production at the end of the hindcast is compared to the actual cumulative production at this time to test the probabilistic reliability of the methodology when production history is limited. The study showed that both developed methods are well-calibrated probabilistic methods. Also, computer software was developed during this research to make the process of calculations more convenient.
The use of acid is an important well maintenancetool in removing near wellbore damage to restore a reservoir's natural permeability and represents one of the most economic options in managing base decline. The selection of acid maintenance candidates however can be a complex process, particularly in wells completed across multiple sands, involving many factors both on the surface and subsurface. As a consequence, individual acid maintenance jobs have had a mixed success rate historically, with certain jobs resulting in a lack of response, or worse, higher water production rates and equipment failures.
This paper uses the Wilmington Oil Field, located in southern California, as a case study to examine the typical characteristics associated with low volume acid maintenance success and provides a novel approach using machine learning (ML) algorithms to aid in the screening and selection of future candidates. The developed algorithms, which make use of the open-source statistical software R, is trained based on results from over 500 producer and 3900 injector acid maintenance jobs that were executed at the field and incorporates predictors from the following groupings determined from literature and subject matter experts (SMEs): Production/injection history, Reservoir properties, Acid type and volume, Delivery mechanism, Formation damage, Well completion design, and Surface facility properties. Over 100+ predictor variables were compiled and screened using supervised feature selection to identify those variables providing the greatest explanatory power. A series of machine learning models: Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), and Gradient Boosting Machines (GBM)) were then used in a classification task to successfully predict whether a producer acid maintenance job would be economic.
The logistic regression model ultimately yielded the best classification results (71% prediction accuracy for the producer jobs and 77% for the injector jobs) and proved to be the ML algorithm with the best balance of accuracy, interpretability, and ease of implementation in the field. The model robustness is examined by applying the algorithm outside of the training and test datasets, to acid maintenance jobs executed in 2016-17 and shows similar predictive accuracy. As a result the model is being actively used to automatically screen for treatment candidates among all 700+ producers and 400+ injectors in the Wilmington Field, which are then validated by SMEs before being executed. The overall process has resulted in significant cost savings by both improving the performance of the acid maintenance program and greatly reducing the amount of time spent by technical staff in selecting candidates. These results indicate that ML algorithms can be effective analytical tools not only for ‘big data’ problems (i.e. largen, time series datasets) which are featured heavily in industry literature, but also for smaller datasets thus opening up a variety of potential applications that can be deployed by surveillance teams alongside traditional approaches.
The use of data-driven cognitive solutions represents a major advancement in the management of oil and gas operations. Tools that integrate concepts from disciplines such as informatics, machine learning and predictive analytics can offer powerful solutions to allow improvements in the safety, efficiency and integrity of oil and gas operations. We discuss data systems that enhance traditional, human-based monitoring systems, with the aim of approaching risk-free operations.
Ethical decision-making is critical in the management of oil and gas operations. Digital data solutions effectively compensate for the limitations of human-based decision processes when confronted with data overload and multi-dimensional data systems. Increased adoption of data gathering, automation, data analytics and advanced computer-aided process control has already made its mark in the industry. Examples have emerged as for how in areas such as artificial lift, pipeline transportation and offshore operations, these data analytics techniques have helped in failure detection and prediction and smart management of such operations. The incorporation of data-intensive decision-making and smart risk management solutions have resulted in a step change in the improvement of the ethical foundation and the base underlying the industry. Moreover, these digital tools and machine-based cognitive processes for risk-avoidance solutions can help to build and restore the public's faith and trust in the industry. We also discuss how relying on digital solutions alone can have its limitations when it comes to professional ethics and responsibilities.
A rapid, user-friendly & easily deployable workflow solution for managing steam flood operations & reservoir management is presented that combines day to day surveillance techniques with performance prediction that uses an analytical model for proactive management of operations. The proposed workflow has been implemented in multiple projects (using public data and data provided by operators) to answer multiple critical questions like pattern response date, steam breakthrough time, expected peak rate from a pattern, arrival of the peak rate, underperforming or over-performing regions in a field, redistribution of the steam volumes, where to minimize steam oil ratio, classify patterns based on efficiency and so on. It is built up on two core components. A set of diagnostic plots and data analytics that enable quick historical performance assessment of patterns as well as the surveillance of the current state of operations. Secondly, an analytical performance model by Jeff Jones is applied to model the oil production from the patterns (early time, mid-life and late pattern life) that serves as a fast-computational tool for practising engineers. This paper discusses the results from multiple fields on the application of proposed workflow and how it has assisted in field management and optimization. These methods are efficient and fast and more importantly user friendly for reservoir management of steamfloods.