Data-driven modeling of heavy oil viscosity in the reservoir from geophysical well logs

Viscosity is the most crucial fluid property on recovery and productivity of hydrocarbon reservoirs, more particularly heavy oil reservoirs. In heavy and extra heavy oil reservoirs e.g. bitumen and tar sands more energy is required to be injected into the system in order to decrease the viscosity to make the flow easier. Therefore, attempt to develop a reliable and rapid method for accurate estimation of heavy oil viscosity is inevitable. In this study, a predictive model for estimating of heavy oil viscosity is proposed, utilizing geophysical well logs data including gamma ray, neutron porosity, density porosity, resistivity logs, spontaneous potential as well as P-wave velocity and S-wave velocity and their ratio (Vp/Vs). To this end, a supervised machine learning algorithm, namely least square support vector machine (LSSVM), has been employed for modeling, and a dataset was provided from well logs data in a Canadian heavy oil reservoir, the Athabasca North area. The results indicate that the predicted viscosity values are in agreement with the actual data with correlation coefficient (R²) of 0.84. Furthermore, the outlier detection analysis conducted shows that only one data point is out of the applicability of domain of the develop model.     

Modeling of the density of mixtures of Athabasca bitumen and a high boiling n-alkane

Recent studies revealed the more availability of heavy oil resources, such as bitumen than other types. So, the injection of solvents such as tetradecane with the aim of diluting bitumen is applied as an enhanced oil recovery (EOR) method for such reservoirs. This study has investigated the prediction of density for Athabasca bitumen–tetradecane mixture, under different temperature, pressure, and solvent's weight percent conditions, using a radial basis function neural network (RBF-NN) technique. Results were then compared with experimental values and values reported based on the previous correlation. MSE and R² values were 0.10496 and 1.00, respectively. Thus, this proposed model has been introduced as a very appropriate model for density prediction of bitumen–tetradecane mixture.     

Evolving ANFIS model to estimate density of bitumen-tetradecane mixtures

Recent investigations have proved more worldwide availability of heavy crude oil resources such as bitumen than those with conventional crude oil. Diluting the bitumen through injection of solvents including tetradecane into such reservoirs to decrease the density and viscosity of bitumen has been found to be an efficient enhanced oil recovery approach. This study focuses on introducing an effective and robust density predictive method for Athabasca bitumen-tetradecane mixtures against pressure, temperature and solvent weight percent through implementation of adaptive neuro-fuzzy interference system technique. The emerged results of proposed model were compared to experimentally reported and correlation-based density values in different conditions. Values of 0.003805 and 1.00 were achieved for mean square error and R², respectively. The developed model is therefore regarded as a highly appropriate tool for the purpose of bitumen-tetradecane mixture density estimation.     

Applying FCM-ANFIS algorithm as a novel computational method for prediction of viscosity of bitumen and heavy alkane mixture

Recently the studies expressed that the noticeable number of oil reservoirs in all over the world are heavy oil and bitumen reservoirs. So the importance of enhancement of oil recovery (EOR) processes for heavy oil and bitumen reservoirs is highlighted. The Dilution of the reservoir fluid by solvents such as tetradecane is one of well-known methods for these types of reservoirs which effects oil recovery by decreasing viscosity. In the present study, Fuzzy c-means (FCM) algorithm was coupled with Adaptive neuro-fuzzy inference system (ANFIS) to predict viscosity of bitumen and tetradecane in terms of temperature, pressure and weight percent of tetradecane. The coefficients of determination for training and testing steps were calculated such as 0.9914 and 0.9613. The comparison of results and experimental data expressed that FCM-ANFIS algorithm has great potential for estimation of viscosity of bitumen and tetradecane.     

PSO-ANFIS modeling of viscosity for mixtures of Athabasca bitumen and a high-boiling n-alkane

The bitumen and heavy oil reservoirs are more in number than light crude oil reservoirs in the world. To increase the empty space between molecules and decrease viscosity, the bitumen was diluted with a liquid solvent such as tetradecane. Due to the sensitivity of enhanced oil recovery process, the accurate approximation for the viscosity of mentioned mixture is important. The purpose of this study was to develop an effective relation between the viscosity of Athabasca bitumen and heavy n-alkane mixtures based on pressure, temperature, and the weight percentage of n-tetradecane using the adaptive neuro-fuzzy inference system method. For this model, the value of MRE and R² was obtained as 0.34% and 1.00, respectively; so this model can be applied as an accurate approximation for any mixture of heavy oil with a liquid solvent.     

Application of fuzzy c-means algorithm as a novel method to predict density of mixtures of Athabasca bitumen and heavy n-alkane

The significant number of oil reservoir are bitumen and heavy oil. One of the approaches to enhance oil recovery of these types of reservoir is dilution of reservoir oil by injection of a solvent such as tetradecane into the reservoirs to modify viscosity and density of reservoir fluids. In this investigation, an effective and robust estimating algorithm based on fuzzy c-means (FCM) algorithm was developed to predict density of mixtures of Athabasca bitumen and heavy n-alkane as function of temperature, pressure and weight percent of the solvent. The model outputs were compared to experimental data from literature in different conditions. The coefficients of determination for training and testing datasets are 0.9989 and 0.9988. The comparisons showed that the proposed model can be an applicable tool for predicting density of mixtures of bitumen and heavy n-alkane.     

Utilization of LSSVM algorithm for estimating synthetic natural gas density

In the gas engineering the accurate calculation for pipeline and gas reservoirs requires great accuracy in estimation of gas properties. The gas density is one of major properties which are dependent to pressure, temperature and composition of gas. In this work, the Least squares support vector machine (LSSVM) algorithm was utilized as novel predictive tool to predict natural gas density as function of temperature, pressure and molecular weight of gas. A total number of 1240 experimental densities were gathered from the literature for training and validation of LSSVM algorithm. The statistical indexes, Root mean square error (RMSE), coefficient of determination (R²) and average absolute relative deviation (AARD) were determined for total dataset as 0.033466, 1 and 0.0025686 respectively. The graphical comparisons and calculated indexes showed that LSSVM can be considered as a powerful and accurate tool for prediction of gas density.     

Modeling of viscosity for mixtures of Athabasca bitumen and heavy n-alkane with LSSVM algorithm

The resources of heavy oil and bitumen are more than those of conventional light crude oil in the world. Diluting the bitumen with liquid solvent can decrease viscosity and increase the empty space between molecules. Tetradecane is a candidate as liquid solvent to dilute the bitumen. Owning to the sensitivity of enhanced oil recovery process, the accurate approximation for the viscosity of aforementioned mixture is important to decrease uncertainty. The aim of this study was to develop an effective relation between the viscosity of Athabasca bitumen and heavy n-alkane mixtures based on temperature, pressure, and weight percentage of n-tetradecane using the least square support vector machine. This computational model was compared with the previous developed correlation and its accuracy was confirmed. The value of R² and MSE obtained 1.00 and 1.02 for this model, respectively. This developed predictive tool can be applied as an accurate estimation for any mixture of heavy oil with liquid solvent.     

Applying LSSVM algorithm as a novel and accurate method for estimation of interfacial tension of brine and hydrocarbons

One of the critical parameters in petroleum and chemical engineering is the interfacial tension between brine and hydrocarbon which has major effects on trapping and residual oil in reservoir pore throat so it becomes one of the interesting topics in enhancement of oil recovery in this work Least squares support vector machine (LSSVM) algorithm was applied as a novel predicting machine for prediction of interfacial tension of brine and hydrocarbons in terms of hydrocarbon carbon number, temperature, pressure and ionic strength of brine. A total number of 175 interfacial tensions were collected from literature in the purpose of training and testing of the model. The root mean squared error (RMSE), average absolute relative deviation (AARD) and the coefficient of determination (R²) were calculated overall datasets as 0.23964, 0.27444 and 0.98509 respectively. The results of study showed that predicting LSSVM machine can be applicable for estimation of interfacial tension and EOR processes.     

Prediction of Oil Production Rate Using Vapor-extraction Technique in Heavy Oil Recovery Operations

Heavy oil and bitumen are major parts of the petroleum reserves in north of America. Owning to this fact and produce this type of oils various methods could be considered. Vapor extraction (VAPEX) method is one of the promising methods that have been executed successfully through North America, specifically in Canada, and is a solvent-based approach. The authors present the implication of the new type of network approach with low parameters called least square support vector machine (LSSVM) in prediction of the oil production rate via VAPEX method. To evaluate and examine the accuracy and effectiveness of both developed models in estimation oil production rate via VAPEX method, extensive experimental VAPEX data were faced to the two addressed models. Moreover, statistical analysis of the output results of the LSSVM was conducted. Based on the determined statistical parameters, the outcomes of the LSSVM model has lower deviation from relevant actual value. Knowledge about oil production via enhanced oil recovery (EOR) methods could help to select and design more proper EOR approach for production purposes. Outcomes of this research communication could improve precision of the commercial reservoir simulators for heavy oil recovery specifically in thermal techniques.     

Application of Grid partitioning based Fuzzy inference system and ANFIS as novel approach for modeling of Athabasca bitumen and tetradecane mixture viscosity

The heavy oil and bitumen reservoirs have effective role on supplying energy due to their availability in the world. The bitumen has extremely high viscosity so this type of reservoirs has numerous problems in production and trans- portation.one of the common approach for reduction of viscosity is injection of solvents such as tetradecane. In the present study the Grid partitioning based Fuzzy inference system was coupled with ANFIS to propose a novel algorithm for prediction of bitumen and tetradecane mixture viscosity in terms of pressure, temperature and weight fraction of the tetradecane. In the present study, the coefficients of determination for training and testing phases are determined as 0.9819 and 0.9525 respectively and the models are visualized and compared with experimental data in literature. According to the results the predicting method has acceptable accuracy for prediction of bitumen and tetradecane mixture viscosity.     

On prediction of asphaltene precipitation in different operational conditions utilization of LSSVM algorithm

Asphaltene which is known as one of the fractions of oil, can cause the important problems during production of crude oil in reservoir, tubing and surface facilities so these problems can influence the production cost and time. In order to predicting and solving asphaltene problems, a powerful Least squares support vector machine (LSSVM) algorithm were developed for asphaltene precipitation estimation as function of dilution ratio, temperature, precipitant carbon number, asphaltene content and API of oil. A total number of 428 measured data were utilized to train and test of LSSVM algorithm. The average absolute relative deviation (AARD), the coefficient of determination (R²) and root mean square error (RMSE) were determined as 7.7569, 0.98552 and 0.26312 respectively. Based on these statistical parameters and graphical analysis it can be concluded that the predicting algorithm has enough reliability and accuracy in prediction of asphaltene precipitation.     

On the prediction of solubility of alkane in carbon dioxide using the LSSVM algorithm

The increasing global energy demand and declination of oil reservoir in recent years cause the researchers attention focus on the enhancement of oil recovery approaches. One of the extensive applicable methods for enhancement of oil recovery, which has great efficiency and environmental benefits, is carbon dioxide injection. The CO₂ injection has various effects on the reservoir fluid, which causes enhancement of recovery. One of these effects is extraction of lighter components of crude oil, which straightly depends on solubility of hydrocarbons in carbon dioxide. In order to better understand of this parameter, in this study, Least squares support vector machine (LSSVM) algorithm was developed as a novel predictive tool to estimate solubility of alkane in CO₂ as function of carbon number of alkane, carbon dioxide density, pressure, and temperature. The predicting model outputs were compared with the extracted experimental solubility from literature statistically and graphically. The comparison showed the great ability and high accuracy of developed model in prediction of solubility.     

Applying of LSSVM approach as a novel tool for accurate prediction of asphaltene inhibition efficiency

Asphaltene precipitation is one of critical problems for petroleum industries. There are different methods for inhibition of asphaltene precipitation. One of the common and effective methods for inhibition of asphaltene precipitation is utilizing asphaltene inhibitors. In this work, Least squares support vector machine (LSSVM) algorithm was coupled with simplex optimizer to create a novel and accurate tool for estimation of effect of inhibitors on asphaltene precipitation as function of concentration and structure of inhibitors and crude oil properties. To this end a total number of 75 measured data was extracted from the literature for training and testing of predicting model. The average absolute relative deviation (AARD), the coefficient of determination (R²) and root mean square error (RMSE) of total data for prediction algorithm were determined as 1.1479, 0.99406 and 0.61039. According to these parameters and graphical comparisons the LSSVM algorithm has potential to predict asphaltene precipitation in high degree of accuracy.     

An Efficient Approach for the Determination of Oil Production Rate During the Water-flooding Recovery Method

In this work, a mathematical methodology namely, least square support vector machine (LSSVM) is implemented to predict the variation of oil production rate as a function of oil water viscosity ratio and water injection rate for water-flooding. Furthermore, the coupled simulated annealing (CSA) optimization technique is coupled with LSSVM to find the optimal architecture and parameters of the LSSVM. The obtained results demonstrate that the CSA-LSSVM estimations are in a satisfactory agreement with literature-reported data and the previously published correlation. Consequently, the R² and average absolute relative deviation of CSA-LSSVM model in testing phase are reported 0.979 and 8.15, respectively.     

Adsorption kinetics of natural surfactants on carbonate minerals in enhanced oil recovery processes

One of the most promising methods for improving oil recovery from carbonate reservoirs is surfactant flooding in which the trapped oil can be mobilized by alteration in the wettability of rock surfaces and also reduction in the interfacial tension between oil and water. Adsorption of surfactants on carbonate minerals plays a key role in designing this process and may make it less effective for enhancing oil recovery. Natural surfactants have been proposed by many researchers since they have lower cost and also less detrimental environmental effects compared to the industrial surfactants. Well-established predictive models for predicting the adsorption of natural surfactants have some issues which need to be addressed. Therefore, developing an accurate, rapid and simple model is crucial. In this study, a least square support vector machine (LSSVM) optimized with coupled simulated annealing (CSA) algorithm is developed for accurate prediction of natural surfactants kinetic adsorption on carbonate minerals. Obtained results by this model were in a very good agreement with experimental results. Additionally, the results showed that the proposed model has the highest accuracy and performance in comparison to the previous kinetic models. Afterward, the effect of natural surfactants adsorption on the amount of oil recovery and also the quality of the produced oil was investigated via core flooding tests for showing the importance of determining the adsorption of surfactants before any surfactant flooding. Results demonstrated that lower surfactants adsorption yields higher oil recovery factor and oil with higher viscosity.     

The effect of different solvents on the density of undersaturated athabasca bitumen: Application in VAPEX and ES-SAGD

Predicting the density of bitumen after solvent injection is highly required in solvent-based recovery techniques like expanding solvent-steam assisted gravity drainage (ES-SAGD) and vapor extraction (VAPEX) in order to estimate the cumulative oil recovery by these processes. Using experimental procedures for this purpose is so expensive and time-consuming; therefore, it is crucial to propose a rapid and accurate model for predicting the effect of various solvents on the dilution of bitumen. In this study, an adaptive neuro-fuzzy interference system is introduced to estimate the effect of methane, ethane, propane, butane, carbon dioxide, and n-hexane on the density of undersaturated Athabasca bitumen in wide ranges of operating conditions. The obtained results were in an excellent agreement with experimental data with coefficients of determination (R²) of 0.99997 and 0.99948 for training and testing datasets, respectively. Statistical analyses illustrate the superiority of the proposed model in predicting the bitumen density at different conditions.     

The Investigation of SAGD Production Profile in a Naturally Fractured Reservoir (NFR)

Abstract

The steam-assisted gravity drainage (SAGD) process has been found as a promising enhanced oil recovery (EOR) process to recover bitumen and heavy oils. A few studies were done on the SAGD process in naturally fractured reservoirs. The effects of various reservoir variables and operational parameters on production profile were simulated using commercial software. The results showed that three different periods of oil production exist in SAGD process at naturally fractured reservoirs. At first, fractures depletion in the near well region (NWR) affect mainly the initial oil production rate. However, the preheated NWR matrixes cause the oil rate not to decrease suddenly. Then, due to rising of steam into upper layer fractures, oil rate increases and therefore the first flag of the oil rate pulse occurs. At third period, another flag of oil rate pulse with a reduction trend is observed due to fractures depleting and starting of oil drainage from its surrounded matrix blocks. The second and third period occur again as steam penetrates into another block in above well region (AWR). The number of oil rate pulses shifts to left by increasing preheating period and fracture density and also by decreasing well pair length. The number of pulses also increases with reduction in well pair length and fracture density. By increasing well spacing and fracture permeability and also by decreasing production bottom hole pressure and fracture density the pulses vanish.     

A PSO-ANFIS framework for prediction of density of bitumen diluted with solvents

One of the important properties in petroleum engineering calculations in heavy oil reservoirs is the density of bitumen diluted with solvents. It is required in newly developed solvent based enhanced oil recovery methods. Hence, developing accurate models for prediction of this parameter is essential. To tackle this issue, this study presents an accurate model based on adaptive neuro-fuzzy inference system trained by particle swarm optimization (PSO-ANFIS) for estimation of density of bitumen diluted with solvents and hydrocarbon mixtures using experimental data from literature. The accuracy and reliability of results were evaluated by utilizing various statistical and graphical approaches and comparing the predictions of the developed model with literature models. The analysis showed that the PSO-ANFIS model is capable to predict the experimental data with acceptable error and high accuracy. The predictions of the PSO-ANFIS model were also better than the literature models.     

Athabasca oil sands: effect of organic coated solids on bitumen recovery and quality

The Canadian oil sands deposits in northern Alberta contain about 1.3 trillion barrels of crude oil equivalent. The largest of the four major formations is found in the Athabasca region where bitumen is heterogeneously distributed throughout an unconsolidated mineral matrix. About one-tenth of the oil sands in this deposit is economically recoverable by conventional surface mining techniques.The Hot Water Extraction Process (HWEP) is used commercially to recover bitumen from surface mined oil sands ore. The viability of this process relies on the existence of a thin water film around each solid particle in the ore matrix. However, a completely water-wet mineral condition is not generally the case for oil reservoirs, including oil sands deposits. In the latter case, it has been shown that certain solid fractions are associated with significant amounts of toluene insoluble organic matter (TIOM), physically or chemically adsorbed onto particle surfaces. These fractions are generically described as ‘organic rich solids’ (ORS). In bitumen separation processes, the organic matter associated with various ORS fractions represents an impediment to optimum bitumen separation and upgrading. In this sense, these solids are considered to be ‘active’ relative to the ‘inactive’ water wetted quartz particles comprising the bulk of the oil sands ore. Preliminary results indicate that the ORS content of an ore appears to be a better predictor for ore processability than the traditional use of bitumen or fines (−44 μm) contents.Two types of ORS have received particular attention. The first is a coarser fraction, usually less than 44 μm but also occurring as particles greater than 100 μm in diameter. This material typically occurs as aggregates of smaller particles bound together by humic matter and precipitated minerals. During the bitumen separation process, these heavy aggregates carry any associated bitumen into the aqueous tailings, thus reducing overall bitumen recovery. The second important fraction comprises very thin, ultra-fine clay particles with a major dimension of <0.3 μm. These ultra-fine clays, with a surface coating of organic matter, remain with bitumen during the separation process. In bitumen upgrading, these solids may be entrained with volatile overheads and cause problems in downstream operations. This paper summarises the protocols developed to separate and characterise these intractable components from HWEP process streams and discusses their role in determining bitumen recovery and quality.