Prediction Of Asphaltene Precipitation Research Articles

Predicting asphaltene precipitation in CO2 flooding processes using artificial intelligence techniques

The removal of asphaltene depositions from crude oil wellbores and tubing is a very difficult and expensive process. Therefore, a very accurate predictor for the Asphaltene Precipitation (AP) is a very important tool for decision making. In this paper, the AP during the injection of carbon dioxide into crude oil reservoirs is forecasted using an artificial neural network (ANN). For this purpose, experimental data is collected from six different oil fields. 70% of the collected data is utilized for training and to develop the proposed ANN. Trainlm training algorithm was found to be the best ANN architecture for AP forecasting. In order to validate the capability of the designed ANN 30% of the unseen data is utilized for validation. The forecasting results show the Mean Square Error (MSE) of the forecasting is 0.0018 that confirms the high accuracy of the model. In order to evaluate the capability of the proposed methodology, the results of proposed ANN are compared with modified Hirschberg model. Comparison confirmed the proposed ANN has better capability and higher accuracy. In the final stage, the effects of various operating parameters on the AP during gas injection are examined. The results show that the most sensitive parameter is the reservoir temperature. In addition, there is a high correlation between an increase in the AP with an increase in carbon dioxide concentration in liquid phase.

Association parameters for Cubic-Plus-Association equation of state determined by molecular dynamics and Petroleomics: Prediction of asphaltene precipitation

In this study, petroleomics characterization, molecular dynamics (MD) simulations, and a cubic-plus-association (CPA) equation of state (EoS) were conducted to understand how the structural characteristics affect the phase behavior of a fluid. To do this, the asphaltene onset pressure (AOP) was estimated by CPA EoS using association parameters determined from MD simulations. Petroleomics characterization and gas chromatography were conducted to identify appropriate representative molecular structures for describing a live crude oil. The American Petroleum Institute (API) gravity, dead and live crude oil densities, and viscosities at reservoir conditions were calculated from the MD simulation results and were found to differ from the experimental measurements by less than 5%, indicating accurate characterizations for both real dead and live crude oil. A formation damage diagnosis from a compositional perspective was conducted because oxygenated aromatic compounds, which are highly active in aggregation terms, were identified in the aromatic fraction by Fourier-transform ion cyclotron resonance mass spectrometry (FT–ICR MS). The association parameters for the self-associating organics (So: asphaltenes and oxygenated aromatics) and the cross-interaction between these Sos and the maltenes were calculated from the MD simulation results. To the best of our knowledge, this is the first time that association parameters have been obtained from an MD model and successfully implemented in a CPA EoS, allowing the prediction of AOP with an accuracy close to 1% compared to the experimentally determined value for live crude oil. This study provides a promising framework to develop an EoS for crude oils based on their molecular characteristics, reducing the need to perform parametric tuning with experimental data. In addition, this structural characterization proposed could be important to other key topics such as enhanced oil recovery (EOR) and carbon capture use and storage (CCUS), relating the structural composition to phase behavior.

Comparative study of machine learning algorithms in predicting asphaltene precipitation with a novel validation technique

Comparative study of machine learning algorithms in predicting asphaltene precipitation with a novel validation technique

Modelling the effect of the inhibitors on asphaltene precipitation using Flory–Huggins theory

Due to the technical, environmental and economic problems caused by asphaltene precipitation, such as oil production reduction, well shut-ins and the necessity of EOR usage, the prediction of asphaltene precipitation seems to be vital. Considering the larger size of asphaltene molecules compared to the other hydrocarbon, it is reasonable to predict the precipitation using the Flory–Huggins theory. In this study, Flory–Huggins solution theory has been modified regarding the solvent molar volume. The modified model was used to predict the asphaltene precipitation of four oil samples in the absence and presence of the inhibitors. Then, the modeling data given by the Flory–Huggins theory was validated with the experimental data obtained by ASTM D-6560 standard method. The mean error at this modeling was 2–13%, which seems acceptable. The proposed model for the cases where an inhibitor is not involved has higher accuracy. The modified Flory–Huggins theory confirmed that the addition of inhibitors at all concentrations postpones the onset point. The average error of the modified model was found to be 4.5–9.8%, which is in a good range. Also, the model accuracy is less for situations where the asphaltene content of the crude oil is higher. Based on this study, the modification of Flory–Huggins theory, regarding the solvent molar volume leads to a lower error.

A Comparative Study of the Cubic-Plus-Association Equation of State and a Peng-Robinson Equation of State–Based Solid Model for Asphaltene Simulation in the Wellbore

Summary Asphaltene is one of the main flow-assurance concerns in oil production. Its precipitation and further deposition along the flow path (wellbore or reservoir) can cause reduced hydrocarbon flow rates and even a total blockage. These potential damages have caused a growing interest in computationally efficient methods to predict asphaltene precipitation, depending on flow conditions. This paper presents two different approaches: (i) a model from Li and Firoozabadi (2010), using a simplified version of the cubic-plus-association equation of state (CPA EOS), and (ii) a version of a solid model based on the Peng-Robinson (PR) EOS, with no association term. The two approaches are compared in isolated flash calculations and wellbore flow simulations with asphaltene deposition. The comparisons are made using the same fluid parameters for both approaches as much as possible. Results show that, if provided with adequate input data, the solid model can consistently match results from the more complex CPA model quite successfully for several fluid compositions. An attempt is made to explain the “success” of the solid model in reproducing CPA model results. The solid model cannot adjust to fluid-composition changes in a manner similar to that of CPA. Therefore, the solid model seems more suitable for wellbore than reservoir simulation, which tends to involve a higher level of fluid mixing. However, the efficiency of the solid model can reduce the computational time by a factor of 2 in comparison with CPA.

Asphaltene Precipitation Prediction during Bitumen Recovery: Experimental Approach versus Population Balance and Connectionist Models.

Deasphalting bitumen using paraffinic solvent injection is a commonly used technique to reduce both its viscosity and density and ease its flow through pipelines. Common modeling approaches for asphaltene precipitation prediction such as population balance model (PBM) contains complex mathematical relation and require conducting precise experiments to define initial and boundary conditions. Machine learning (ML) approach is considered as a robust, fast, and reliable alternative modeling approach. The main objective of this research work was to model the effect of paraffinic solvent injection on the amount of asphaltene precipitation using ML and PBM approaches. Five hundred and ninety (590) experimental data were collected from the literature for model development. The gathered data was processed using box plot, data scaling, and data splitting. Data pre-processing led to the use of 517 data points for modeling. Then, multilayer perceptron, random forest, decision tree, support vector machine, committee machine intelligent system optimized by annealing, and random search techniques were used for modeling. Precipitant molecular weight, injection rate, API gravity, pressure, C5 asphaltene content, and temperature were determined as the most relevant features for the process. Although the results of the PBM model are precise, the AI/ML model (CMIS) is the preferred model due to its robustness, reliability, and relative accuracy. The committee machine intelligent system is the superior model among the developed smart models with an RMSE of 1.7% for the testing dataset and prediction of asphaltene precipitation during bitumen recovery.

Performance analysis of kinetic models in predicting asphaltene precipitation rate under different conditions of temperature and percentage of precipitant molecular weight

Performance analysis of kinetic models in predicting asphaltene precipitation rate under different conditions of temperature and percentage of precipitant molecular weight

A new vapor-liquid-asphaltene three-phase equilibrium computation algorithm based on the free-asphaltene assumption

Asphaltene precipitation is one of the major flow-assurance problems in petroleum engineering and can occur at different stages of oil production due to pressure/temperature changes or compositional changes. It can cause formation damage in the reservoir and choking of oil tubing in the wellbore, thus leading to the reduction of production rates. Accurately predicting under what conditions asphaltene precipitation is a prerequisite to the effective mitigation of the flow assurance problems related to asphaltene precipitation. In this work, a new three-phase vapor-liquid-asphaltene precipitation algorithm is developed based on Li and Li's work (Li and Li, 2019) using the free-asphaltene assumption, in which the asphaltene phase is considered as a liquid phase that only contains asphaltene component (Nascimento et al., 2018). The algorithm is validated by comparing the calculation results with the asphaltene precipitation data of five different oil samples. The tuned parameters for oil sample 3 and oil sample 4 are then used to predict asphaltene precipitation with different gas injections. The calculation results agree well with the experimental data. It indicates that the three-phase VLA equilibrium calculation algorithm can make reliable predictions of the PT and Px phase diagrams with different gas injections as well as the asphaltene precipitation amount. The calculation results yielded by the free-asphaltene assumption are also compared with the results yielded by the solid assumption and the liquid assumption. The algorithm developed using the free-asphaltene assumption is shown to be able to perform slightly better than the algorithm developed using the solid assumption, and perform as well as the algorithm developed using the liquid assumption.

Smart phase behavior modeling of asphaltene precipitation using advanced computational frameworks: ENN, GMDH, and MPMR

Asphaltene precipitation is the reason behind some of the most destructive issues in the oil industry such as wettability alteration, formation plugging, and reduction of the relative permeability. There are different experimental and modeling methods to study the asphaltene phase behavior. However, those approaches are either time-consuming or costly. In addition, the prediction of asphaltene precipitation is challenging due to the nonlinear dependence. Therefore, in this study, three advanced computational algorithms including group method of data handling (GMDH), emotional neural network (ENN), and minimax probability machine regression (MPMR) are developed and applied to the comprehensive dataset to estimate the amount of asphaltene precipitation as the function of temperature, dilution ratio, and the molecular weight of the n-alkanes. Many different performance metrics are used to evaluate the predictive performance of intelligent algorithms. Obtained results of the modeling reveal that intelligent computational algorithms have a great ability to mimic the nonlinear relationships between the target variable and its influential variables. The results indicate MPMR as the best predictive model with the highest R-squared value on both training and testing datasets with values of R 2 = 0.992 and R 2 = 0.991, respectively. In addition, MPMR is compared and outperformed empirical correlations.

Asphaltene Precipitation from Heavy Oil Diluted with Petroleum Solvents

Heavy oils are sometimes diluted with petroleum solvents for pipeline transportation, and it is necessary to ensure that the selected solvent does not precipitate asphaltenes from the oil. The modified regular solution (MRS) model can predict asphaltene precipitation from heavy oil diluted with pure n-alkanes but has not yet been applied to petroleum solvents. To do so, asphaltene precipitation data were measured for bitumen diluted with petroleum solvents including gas condensates, diesels, kerosene, and naphtha and their blends with n-heptane. The petroleum solvents were characterized into single carbon number (SCN) fractions on the basis of gas chromatography assays. New correlations were proposed for the molecular weight, density, and solubility parameters of the SCN fractions. The MRS model with the characterization and correlations matched asphaltene precipitation data with an average absolute deviation of 0.7 wt % for petroleum solvents with relatively low contents of pure aromatic and cyclic components ( 10 wt %). Hence, the proposed model is currently limited to solvents similar to condensates, diesels, and kerosenes.

Developing a novel colloidal model for predicting asphaltene precipitation from crude oil by alkane dilution

This research aims to propose a thermodynamic model for predicting asphaltene precipitation upon diluting a crude oil with a paraffinic solvent. To this end, a thorough mathematical formulation was carried out to derive a novel micellization model based on the associative properties of asphaltenic compounds. It was assumed that asphaltenes exist in the oil both as monomeric molecules and aggregated cores; with stabilization latter by attachment of resin on its periphery. The aggregation equilibrium was established by taking into account asphaltene's lyophobic tendency, heat of resin adsorption, and interfacial tension between micelle and oil media which is the main driving factor contributing to Gibbs free energy of micellization. In this study, the mean diameter of asphaltene aggregates at different dilution ratios was incorporated into the framework of modeling approach, by measuring particle size distribution of asphaltenes at varying n-heptane fractions. The data density of pure asphaltene for an Iranian crude-oil sample was used in calculations. The proposed model favorably predicted the precipitation trend of the crude oil diluted at varying fractions of n-heptane. This model shows that taking into account the size of asphaltene aggregates can significantly improve the prediction accuracy of the micellization model, compared to the simple model, with normalized root-mean-square error of 9.4 and 28.7, respectively. A detailed sensitivity analysis was performed to gain insight into physical implication of the model parameters.

A new multiphase and dynamic asphaltene deposition tool (MAD-ADEPT) to predict the deposition of asphaltene particles on tubing wall

As expounded, the precipitation and deposition of asphaltene particles in pipelines has been proved to be the most challenging flow assurance problem due to its unknown and complex behaviors. In this work, a new multicomponent, multiphase and dynamic tool was developed to model the aggregation and deposition of asphaltene particles in a bulk medium. The multiphase and dynamic asphaltene deposition tool, shortened as MAD-ADEPT is, in fact, a modified version of the previously developed ADEPT. The new tool was developed to make the asphaltene deposition and aggregation concepts in oil production wells more predictable. To tackle the complexity of the asphaltene problem, a bespoke algorithm was developed to predict asphaltene precipitation, aggregation and deposition. The algorithm was made up of three main modules, i.e., thermodynamic, preprocessor and integrated time step adaptive modules. In the thermodynamic module, the Peng-Robinson equation of state was used and fine-tuned to estimate the fluid properties of multiphase hydrocarbon systems, while the PC-SAFT equation of state was used in order to obtain the asphaltene precipitation phase envelope (APE). In the preprocessor module, the sensitivity of the system to the reduction of wellbore inner diameter (ID) was examined. In the main module which could dynamically model the asphaltene aggregation and deposition, a non-isothermal multiphase mechanistic model, a convective-diffusive-reactive model and a non-convective volume of fluid model were solved simultaneously with an adaptive time step selection method. The new MAD-ADEPT was utilized to simulate a real field case to predict the asphaltene deposition in different time and space intervals along the production tubing. The results obtained from MAD-ADEPT was validated by those scant field data available in the literature. The results confirmed that the new tool can predict the asphaltene deposition reliably at various operating conditions within a production system. Also, the effects of multiphase flow as well as wellbore configurational parameters were studied on the amount of asphaltene deposition thickness on tubing wall. The accuracy of the results for the case studies, fast rate of convergence and less CPU computation time are the clearest advantages of the new asphaltene deposition dynamic tool over the previously developed ones.

Aggregation thermodynamics of asphaltenes: Prediction of asphaltene precipitation in petroleum fluids with NRTL-SAC

Based on the aggregation thermodynamics established for asphaltene precipitation in binary solvents (Wang et al., AIChE J., 2016. 62 (4), 1254–64; Islam et al., Fluid Phase Equilibria, 2018. 473, 255–261), this work extends the aggregation thermodynamics for asphaltene precipitation in blending of heavy oils with light hydrocarbons. In lieu of the commonly accepted pseudocomponents approach for petroleum fluid characterization, the petroleum fluids are characterized in terms of model molecules and their compositions (Chen and Que, U.S. Patent No. 9,934,367 B2, April 3, 2018). Solvent power of the petroleum fluids is further quantified with the conceptual segment-based NRTL-SAC activity coefficient model (Chen and Song, Ind. Eng. Chem. Res., 2004, 43 (26), 8354–62). We show the aggregation thermodynamics with the NRTL-SAC model successfully predicts asphaltene precipitation from crude oils upon blending with normal alkanes.

Evaluation of different thermodynamic models in predicting asphaltene precipitation: A comparative study

One of the major issues in the oil industry is asphaltene precipitation. Modeling asphaltene precipitation is still considered as a complex problem due to various characteristics of different heavy components existing in the crude oil. Thermodynamic models have been found as accurate models for studying asphaltene precipitation in the past few years and a great deal of effort has been devoted to model this process by using different empirical models and equations of state. In this study, the obtained results of asphaltene precipitation from different models based on perturbed-chain statistical associating fluid theory (PC-SAFT), cubic-plus-association (CPA), solid model, Flory-Huggins (FH), and the modified Flory-Huggins (MFH) are compared and their accuracy and reliability are analyzed in detail. For this purpose, twelve crude oil types with different characteristics and asphaltene precipitation behavior are used. Additionally, the performance of the introduced models in predicting asphaltene precipitation during gas injection into the studied oil is investigated. Results demonstrated that PC-SAFT and CPA models have the highest accuracy for both precipitation estimation and behavior trend prediction. Afterward, sensitivity analysis is performed by using Monte-Carlo algorithm for better understanding of the effect of different adjusting parameters, which were used during the tuning process, on each model outputs. Results indicated that cross-association energy between asphaltene and heavy component (HC), self-association energy of asphaltene, and binary interaction coefficient between asphaltene and CO2 are the most sensitive tuning variables for PC-SAFT, CPA, and solid models, respectively. Finally, the CPU times of various models for simulating this process were compared. This comparison showed that the PC-SAFT model has more computational time due to the involved iterative processes for phase equilibrium calculations.

Investigation of Asphaltene Precipitation Using Response Surface Methodology Combined with Artificial Neural Network

The precipitation of asphaltene, one of the components of oil, in reservoirs, transfer lines, and equipment causes many problems. Accordingly, researchers are prompted to determine the factors affecting asphaltene precipitation and methods of avoiding its formation. Predicting precipitation and examining the simultaneous effect of operational variables on asphaltene precipitation are difficult because of the multiplicity, complexity, and nonlinearity of factors affecting asphaltene precipitation and the high cost of experiments. This study combined the use of response surface methodology and the artificial neural network to predict asphaltene precipitation under the mutual effects of various parameters. The values of such parameters were determined to reach the minimum amount of precipitation. We initially selected the appropriate algorithm for predicting asphaltene precipitation from the two neural network algorithms. The outputs of designed experiments in response surface methodology were determined using the optimum algorithm of the neural network. The effects of variables on asphaltene precipitation were then investigated by response surface methodology. According to the results, the minimum precipitation of asphaltene achieved at zero mole percent of injected nitrogen and methane, 10–20 mole percent of injected carbon dioxide, asphaltene content of 0.46, the resin content of 16.8 weight percent, the pressure of 333 psi, and temperature of 180 . Results showed that despite the complexities of asphaltene precipitation, the combination of artificial neural network with response surface methodology can be successfully used to investigate the mutual effect of different variables affecting asphaltene precipitation.

Enhancing asphaltene precipitation modeling by cubic-PR solid model using thermodynamic correlations and averaging techniques

Cubic equation-of-state solid models are one of the most widely used models to predict asphaltene precipitation behavior. Thermodynamic parameters are needed to model precipitation under different pressures and temperatures and are usually obtained through tuning with multi asphaltene onset experiments. For the purpose of enhancing the cubic Peng–Robinson solid model and reducing its dependency on asphaltene experiments, this paper tests the use of aromatics and waxes correlations to obtain these thermodynamic parameters. In addition, weighted averages between both correlations are introduced. The averaging is based on reported saturates, aromatics, resins, asphaltene (SARA) fractions, and wax content. All the methods are tested on four oil samples, with previously published data, covering precipitation and onset experiments. The proposed wax-asphaltene average showed the best match with experimental data, followed by a SARA-weighted average. This new addition enhances the model predictability and agrees with the general molecular structure of asphaltene molecules.

Predicting asphaltene precipitation during titration of diluted crude oil with paraffin using artificial neural network (ANN)

Asphaltene precipitation from crude oil in underground reservoirs and on ground facilities is one of the major problems in a large portion of oil production units around the world. Many scaling equations and intelligent predictive models using the artificial neural network (ANN) are proposed in the literature but none of them can be applied when crude oil is diluted with any types of paraffin. In this study, feed forward artificial neural network is used for prediction of the amount of asphaltene precipitated weight percent of diluted crude oil with paraffin based on titration tests data from published literature. Trial and error method is utilized to optimize the artificial neural network topology in order to enhance its strength of generalization. The results showed that there is good agreement between experimental and predicted values. This predictive model can be applied to estimate the amount of asphaltene precipitated weight percent when the crude oil is diluted with paraffin and to avoid experimental measurement that is time-consuming and requires expensive experimental apparatus as well as complicated interpretation procedure.

A generalized scaling equation to predict asphaltene precipitation during precipitant dilution, natural depletion, water injection and gas injection

Asphaltene precipitation and subsequent deposition on the rock surface during natural depletion and Enhanced Oil Recovery (EOR) methods are two challenging flow assurance issues, causing a sudden decline in oil production. Different precipitation mechanisms and absence of an appropriate generalized characterization parameter have questioned the predictive potential of numerous available thermodynamic models. To overcome the complexity of using available thermodynamic models, different scaling equations were developed to represent a simple yet accurate estimation of the amount of asphaltene precipitation. Nonetheless, most of the proposed correlations are not generalized and are best suited to a specific scenario such as addition of a precipitant or gas injection.In this study, several asphaltene precipitation experiments were conducted on different real oil samples to investigate the amount of precipitated asphaltene due to the compositional changes cause by injection of Methane and Water and also due to pressure variations representing natural depletion mechanism. Besides performing the experiments, the collected literature data for precipitant dilution and gas injection were used to propose a new generalized scaling equation by developing the idea suggested by Kord and Ayatollahi to model all types of asphaltene precipitation phenomena. The results indicated that the new generalized scaling equation can successfully model all types of asphaltene precipitation scenarios. The amount of precipitation can be estimated at different thermodynamic conditions for four precipitation phenomena. The results illustrated that the new generalized scaling equation can successfully model all factors triggering asphaltene precipitation at various stages of production, demonstrating its potential for use as an efficient tool to identify, assess and mitigate risk of asphaltene deposition.

A new model for predicting asphaltene precipitation of diluted crude oil by implementing LSSVM-CSA algorithm

One of the severe problems in all the oil production stages from the pore walls of the reservoir rocks to the wellhead, transfer pipelines, and production units of a large portion of the world’s hydrocarbon reservoirs, is asphaltene precipitation and deposition from crude oil on solid surfaces. In this article, least squares support vector machine optimized by coupled simulated annealing is employed for estimation of the amount of asphaltene precipitated weight percent of diluted crude oil with paraffin based on titration tests data from a recently published article. The results indicated that there is an excellent correlation between predicted and experimental values with an average absolute relative deviation percent, mean square error, and a determination coefficient of 0.0727%, 0.0242, and 0.9972, respectively. The developed predicting model can be applied to estimate the amount of asphaltene precipitated when the crude oil is diluted with paraffin and to eschew experimental titration test that is tedious and time-consuming.

Prediction of Asphaltene Precipitation in Reservoir Model Oils in the Presence of Fe3O4 and NiO Nanoparticles by Cubic Plus Association Equation of State

Nanoparticle-technology was employed to address the challenges of asphaltene precipitation during oil production. Fe3O4 and NiO nanoparticles were synthesized and characterized. Asphaltene was extr...