Journal of Chemical and Petroleum Engineering
Volume:54 Issue: 2, Dec 2020

The present research used novel hybrid computational intelligence (CI) models to predict inorganic indicators of water quality. Two CI models i.e. artificial neural network (ANN) and a hybrid adaptive neuro-fuzzy inference system (ANFIS) trained by genetic algorithm (GA) were used to predict inorganic indicators of water quality including total dissolved solids (TDS), total hardness (TH), total alkalinity (TAlk), and electrical conductivity (σ). The study was conducted on samples collected from water wells of Kermanshah province through analyzing water parameters including pH, temperature (T), and the sum of mill equivalents of cations (SC) and anions (SA). A multilayer perceptron (MLP) structure was used to forecast inorganic indicators of water quality using the ANN approach. A MATLAB code was used for the proposed ANFIS model to adjust and optimize the ANFIS parameters during the training process using GA. The accuracy of the generated models was described using various evaluation techniques such as mean absolute error (MAE), correlation factor (R), and mean relative error percentage (MRE%). The results showed that both methods were suitable for predicting inorganic indicators of water quality. Moreover, the comparison of the two methods showed that the predicted values obtained from the ANFIS/GA model were better than those obtained from the ANN approach.

In this paper, the radiation section of ethylene dichloride (EDC) cracking furnace, considering the chemical reaction, was numerically modelled using computational fluid dynamics (CFD). This study investigated the influence of some parameters such as mass flow rate, the inlet temperature of fluid into the radiation section, and heat flux on the conversion and changes in velocity, pressure, and temperature of the fluid along the coil passes, as well as the outlet stream of the coil. Then, the modelling results were compared with a series of industrial data of an industrial EDC cracking furnace. The results showed that considering the variable heat flux boundary condition is more compatible with the industrial data rather than the constant heat flux boundary condition. Increasing the feed inlet temperature to the furnace, increased the EDC conversion due to the endothermic nature of the thermal cracking reaction. Furthermore, reducing the inlet mass flow rate led to a significant increase in the conversion, temperature, and mass fraction of the products due to an increase in residence time.

The present article investigates the implementation of non-minimal state space (NMSS) representation with proportional-integral-plus (PIP) controller for the carbon dioxide absorption process of Shiraz petrochemical ammonia unit. The PIP controller is a logical extension of conventional PI/PID controllers with additional dynamic feedback and input compensators. PIP controller is used for multivariable control without limitation on the number of controlled variables. A Multi Input - Multi Output (MIMO) square model was extracted from step response test. In this way, input water flow rate to carbon dioxide absorption system, the heat duty of input absorbent cooler to tray (1) of absorption tower and re-boiler heat duty of stripping tower are chosen as manipulated variables (inputs), while carbon dioxide mole fraction in absorption tower vapor product, the water mole fraction in absorption tower liquid product and tray temperature No. 36 of stripping tower are determined as controlled ones (outputs). The system identification is performed with three input and three output variables using step response test. As a result, continuous and discrete time transfer function matrices and suitable NMSS model for PIP controller are reported. Finally, in order to evaluate the PIP control performance, the feed flow rate increases by 2%. The results show the proper performance of designed PIP controller for both disturbance rejection and set point tracking.

Water in oil emulsion is consider one of the major challenges encountered during production of heavy oil or when applying enhanced oil recovery techniques whether thermal or chemical. In this study stability and rheological properties of hot and cold produced heavy oil emulsions formed due to steam injection processes in Kuwaiti reservoirs were investigated thoroughly over a wide range of operation conditions. The effects of temperature, shear rates, and water cuts on the physical and chemical behaviors of the heavy oil emulsions were examined experimentally in detail. The results showed that cold-produced heavy oil emulsion (CP-HO) is more stable than hot produced heavy oil emulsions (HP-HO) because of its high salinity concentrations and low resin/asphaltene (R/A) ratios, and low PH value. Moreover, a new emulsion viscosity correlation was developed using the experimental data. The proposed model was validated against existed models. The results showed that the developed correlation i more applicable than the existed one in predicting the viscosity of heavy oil emulsions with a percentage of deviation almost less than 5 %.

Due to the limited crude oil resources, the role of enhanced oil recovery (EOR) techniques in the production of the oil that has not been extracted during the primary and secondary oil production techniques is crucial. Gas injection is known as an important EOR technology, but one of the main concerns during gas injection is asphaltene precipitation and deposition within reservoir formation. In this study, the effect of temperature (ranges 376-416 K) and concentration of injected gas (N2 (10, 20 and 40, mole percent) and first separator gas (20, 40 and 60, mole percent)) on the onset pressures and amount of asphaltene precipitation in one of the Iranian oil reservoirs were investigated. Two series of experiments were accomplished on live oil by gravimetric method; first: injection of different concentrationsof nitrogen and first separator gas at reservoir temperature and under different pressures (3000-8000 psia) and second: natural depletion at different temperatures. Besides, the experimental data of asphaltene precipitation due to N2, first separator gas, and also CO2 injection were compared together. Finally, the experimental data were modeled with a solid model. The results indicate that the amount of asphaltene precipitation due to N2 injection (0.1-0.2 wt %) is lower than the first separator gas and CO2 injection at the same concentration. Experiments show that in the range of experimental temperatures the asphaltene precipitation changes up to 0.06 wt %. For pressures below the bubble pressure (~ 4700 psi), precipitation changes directly with temperature, and indirect relation is observed for pressures above the bubble point pressure.

Reciprocating piston artificial lift systems are widely adopted especially, for onshore wells. Matching the pump mode to well and reservoir conditions reduces the pumping cost and increases the efficiency of production. Parameters influencing the energy requirement of sucker-rod lifted oil wells are investigated in this study, and new insights are provided for the parametric investigation of design variables required for sizing beam-pumped wells. Two (2) artificial lift simulators are integrated for automated sizing of beam-pumped systems. A sucker-rod artificial lift system is optimally sized for a case study oil well, to obtain the minimum API rating of the pumping unit, sustain the target production rate, and determine the corresponding minimum prime mover required to drive the pump sustainably. Compared to using a single simulator for the case study, the integrated approach reduces the damped and polished rod horsepower by 54.9% and 26.5% respectively, for a corresponding decrease in minimum NEMA D motor size by 38.6%. These key performance indicators demonstrate the benefits of simulator integration in automated sizing of beam pumps.

The accurate description of the phase equilibria of CO2 and n-alkane multicomponent mixtures over a wide range of temperature, pressure, and n-alkane molecular weight, requires the models that are both consistent and mathematically flexible for such highly non-ideal systems. In this study, a predictive correlation was proposed for the vapor-liquid equilibrium data (VLE) of CO2 and n-alkane ternary systems, based on the Peng-Robinson equation of state (PR EOS), coupled to cubic mixing rules (CMRs). The ternary interaction parameters (TIP) correlation is developed using binary VLE data and tested for CO2 + n-alkane+ n-alkane ternary systems. For this purpose, binary VLE data of CO2 + n-alkane and n-alkane + n-alkane systems for n-alkane from C3 to C24, covering a total of about 70 references, used to correlate TIP in the ranges of 0.5-31 MPa and 230-663 K. Two temperature-dependent TIP correlations, based on acentric factor ratio, have been tuned with more than 2000 data points of the CO2 + n-alkane and the n-alkane + n-alkane binary systems with AARD of 3.13% and 6.71%, respectively. The generalized predictive correlation was proposed based on the proper three-body interaction contributions and successfully tested for VLE data of the CO2 + n-alkane + n-alkane ternary systems.

A difficult problem in drilling operation that concerns the very drilling parameters is the cutting transport process. Correct calculation of the cuttings concentration (hole cleaning efficiency) in the wellbore annulus using drilling variables such as the geometry of wellbore, rheology, and density of drilling fluid, and pump rate is very important for optimizing these variables. In this study, a hybrid evolutionary fuzzy system (EFS) using artificial intelligent (AI) techniques is presented for estimation of the cuttings concentration in oil drilling operation using operational drilling parameters. A well-organized genetic learning algorithm that computes fitness values by symbiotic evolution is used for extraction of the Takagi–Sugeno–Kang (TSK) type fuzzy rule-based system for the EFS. A determination coefficient (R2) of 0.877 together with a root mean square error (RMSE) of 1.4 between prediction and measured data for test data verified a very satisfactory model performance. Results confirmed that the estimation accuracy of the proposed EFS is better than other models such as Multiple Linear Regression (MLR), artificial neural network (ANN), and adaptive neuro-fuzzy inference system (ANFIS) for hole cleaning modeling.

Biodiesel is a substitute for fossil fuels which is produced through a transesterification reaction between vegetable oils or animal fats and light alcohols such as methanol or ethanol. In this reaction, along with the production of biodiesel, glycerol as a byproduct and non-reacted alcohol that reduces biodiesel quality is produced. Hence, many studies have been carried out on liquid-liquid equilibrium (LLE) for ternary systems containing biodiesel + glycerol + alcohol. Two phases are formed as 1-rich in biodiesel and 2-rich in glycerol; moreover, alcohol is distributed between these two phases. In this work, based on previous experimental data, the UNIQUAC and NRTL thermodynamic models were used to forecast the composition of the phases. The intermolecular interaction term for each of the models was considered as a linear function of the reverse temperature. In both models, there was no difference between the amount of biodiesel produced from different oils and obtained from the general interaction parameters. Based on the results, the percentage of absolute average deviation for NRTL and UNIQUAC models for biodiesel + glycerol + ethanol system were 1.24% and 2.13%, respectively, and for biodiesel + glycerol + methanol system was 1.13% and 1.71%, respectively.

Experimental study of the effect of gas antisolvent (GAS) system conditions on the particle size distribution of finasteride (FNS) requires a thermodynamic model for the volume expansion process. In this study, the phase behavior of the binary system including carbon dioxide and  Dimethyl sulfoxide, and a ternary system comprising carbon dioxide, dimethyl sulfoxide, and Finasteride was studied. The Peng-Robinson equation of state was employed for the evaluation of the fluid phases and a fugacity expression to represent the solid phase. By developing an accurate predictive model, the GAS operating conditions can be optimized to produce particles with no need for a large number of experiments. First, the critical properties of the FNS were evaluated by the group contribution methods. The method of Marrero and Gani was also selected to predict the normal boiling point, critical temperature, and critical pressure. The correlation of Edmister was chosen for the prediction of the acentric factor. The lowest pressures for the ternary system at 308, 318, 328, and 338 K were 7.49, 8.13, 8.51, and 9.03 MPa, respectively. The precipitation of the dissolved finasteride happened at these operating pressures.

In this study, carbazole-based hypercrosslinked polymer (HCP) adsorbent was synthesized using the knitting method by Friedel-Crafts reaction. The effects of crosslinker to carbazole ratio and synthesis time on the adsorbent structure were investigated to improve CO2/N2 and CO2/H2 selectivity. Crosslinker to carbazole ratio and the synthesis time was considered in the range of 1-4 (mol/mol) and 8-18 (h), respectively. HCP adsorbents were analyzed by energy-dispersive X-ray spectroscopy (EDS), Fourier-transform infrared spectroscopy (FTIR), and Brunauer-Emmett-teller analysis (BET). The adsorption capacity of CO2, N2, and H2 were measured by carbazole-based HCP and it was correlated with the nonlinear form of the Langmuir isotherm model. The achieved BET surface area of adsorbent with the highest amount of synthesis parameters was 922 (m2/g). The ideal adsorbed solution theory (IAST) was utilized to anticipate CO2/N2 and CO2/H2 selectivity at 298 k and 1 bar. CO2/N2 and CO2/H2 selectivity for adsorbent with the maximum amount of synthesis parameters were 8.4 and 4.4, respectively. The high selectivity values of carbazole-based HCPs are due to the presence of nitrogen atoms in the adsorbent structure and a more robust interaction between CO2 molecules and the adsorbent surface.

Prediction of gas reservoir performance in some industrial cases requires costly and time-consuming simulation runs and a strong CPU must be involved in the simulation procedure. Many reservoir parameters conform to a strong aquifer behavior on gas reservoir performance. Effects of parameters, including reservoir permeability, aquifer permeability, initial reservoir pressure, brine water salinity, gas zone thickness, water zone thickness, temperature, tubing diameter, reservoir inclination, the effective intruding angle of the aquifer, and porosity were investigated using Tornado chart, and seven parameters were filtered. Response functions of aquifer productivity index, gas recovery factor, initial maximum gas production, water sweep efficiency, gas production rate, water breakthrough time, and water production were defined statistically, using Eclipse E100 and Box-Behnken design (BBD). According to the formulae generated by the BBD based on simulation runs, reservoir permeability, aquifer permeability, well-head pressure, and gas zone thickness are the most influencing parameters on the gas reservoir performance supported by the strong aquifer. The aquifer was found to be important especially due to its productivity index and sweep water efficiency. Validation of results given by the BBD through simulation runs showed response functions of aquifer productivity index, sweep water efficiency, maximum gas production, and recovery factor are of deviation percentages in the ranges of 10.61%, 6.302%, 3.958%, and 2.04%, respectively.