Phase Splitting Research Articles

Antagonism in the aggregation behaviour of N,N,N′,N′-tetraoctyldiglycolamide in n-dodecane upon adding N,N-dioctylhydroxyacetamide during trivalent metal extraction

The solvent phase composed of tetraoctyl diglycolamide (TODGA) in n-dodecane (n-DD) undergoes the undesirable organic phase splitting during the solvent extraction of trivalent actinides and lanthanides from nitric acid medium. To overcome this limitation a neutral extractant namely the N,N-dioctyl hydroxyacetamide (DOHyA), which was half the molecule of TODGA, has been added to the solvent phase and evaluated for the extraction of Nd(III) from nitric acid medium. Since the reverse micellar aggregates of the extracted species in the organic phase was responsible for organic phase splitting, the aggregation behaviour of the organic phase was probed by dynamic light scattering (DLS) technique. The results have been compared with the individual solvent systems, namely DOHyA/n-DD and TODGA/n-DD. The investigations revealed that the extraction of nitric acid and loading of Nd(III) in organic phase increased upon adding DOHyA, but the size of aggregates formed in the organic phase decreased remarkably with DOHyA addition, which was in contrast to the expected behaviour of increase in aggregate size with increase of extraction in the organic phase. To understand the unexpected behaviour of aggregation, the extracted organic phase was further probed by UV–visible and luminescence spectroscopy. The results further revealed that DOHyA in the binary solution served as a reactive organic phase modifier and induced antagonistic effect on aggregation when mixed with TODGA, which was, in fact an advantage for the extraction of trivalent actinides from high-level nuclear waste.

Continuous Relative Permeability Model for Compositional Simulation

Reservoir simulation using high-fidelity fluid models is typically employed to study subsurface displacement processes that involve complex physics. Such processes are highly nonlinear and occur at the interplay of phase thermodynamics (phase stability and split) and rock/fluid interaction (relative-permeability). Due to its nonlinearity, relative-permeability is an important constitutive of the conservation equations, and has a significant impact on the simulation. Dependence of the phase relative-permeability on fluid compositions, pressure, and temperature is well-documented. In compositional reservoir simulation, however, relative-permeability is typically modeled as a function of phase saturation. Such an approach may lead to serious discontinuities in relative-permeability. To alleviate this issue, several authors proposed models based on phase state indicators (density, parachor, Gibbs Free Energy). However, such techniques cannot represent the complete degrees of freedom that are exhibited by compositional displacements. In this work, we present a relative-permeability model based on a parameterization of the compositional space. The model is independent of the hydrocarbon phase labeling as gas or oil. We show that our proposed model (1) applies regardless of the degrees-of-freedom of the compositional displacement problem, and (2) is guaranteed to yield a continuous relative-permeability function across the entire compositional space. We have implemented this model in our research simulator, and we present test cases using traditional relative-permeability models, as well as, numerical results that compare the nonlinear performance.

The effect of alkyl chain length attached to the diglycolamide and n-paraffin on the aggregation behaviour of diglycolamide and MD simulation of aggregates

The structure of diglycolamide (DGA) plays a vital role in the extraction of trivalent lanthanides and actinides from high-level nuclear waste. The radiolytic degradation of the solvent phase composed of DGAs in n-dodecane produces a range of alkanes differing in chain length and branching, and various other derivatives of DGA in organic phase, known as degradation products. However, the alkyl chain attached to the degraded DGA and n-dodecane degradation products determines the solubility of de-degraded DGA in n-dodecane, as well as the reverse micellar aggregation behavior of the extracted species in organic phase. In the present study, the extraction behavior of nitric acid in 0.2 M alkyl DGAs dissolved n-paraffins was studied. The alkyl chain attached to the DGA was varied from hexyl to dodecyl moiety and that in n-paraffin was varied from n-octane to n-tetradecane to represent the radiolytic degradation products of DGAs and n-dodecane respectively. The nitric acid extraction in all solvent phases (0.2 M alkyl DGAs/n-paraffin) was quite similar. However, the reverse micellar aggregation behavior of the organic phase and the consequent organic phase splitting tendency, known as third phase formation, was significantly different in these solvent phases upon nitric acid extraction. It was observed that the undesirable third phase formation was more in DGAs having lower alkyl chain and in n-paraffins with longer alkyl chain. To understand this behaviour, the nitric acid extracted organic phase was subjected to dynamic light scattering studies and the results revealed that the aggregates size and their distribution in organic phase increased with decreasing the alkyl chain length of DGA and increasing the alkyl chain of n-paraffin, which were found to be responsible for early third phase formation in those systems. The aggregation results obtained from DLS were validated by molecular dynamics simulation studies. It was observed that the size of the clusters obtained from MD simulations was in good agreement with the experimental results. The radial distribution data of the extracted phase showed that the DGAs with shorter alkyl chain and the n-paraffins with longer chain length, if present in the solvent phase were susceptible to early third phase formation.

Discontinuity gravity modes in hybrid stars: Assessing the role of rapid and slow phase conversions

Discontinuity gravity modes may arise in perturbed quark-hadron hybrid stars when a sharp density jump exists in the stellar interior and are a potential fingerprint to infer the existence of quark matter cores in compact objects. When a hybrid star is perturbed, conversion reactions may occur at the quark-hadron interface and may have a key role in global stellar properties such as the dynamic stability and the quasi-normal mode spectrum. In this work we study the role of the conversion rate at the interface. To this end, we first derive the junction conditions that hold at the sharp interface of a non-radially perturbed hybrid star in the case of slow and rapid conversions. Then, we analyse the discontinuity $g$-mode in both cases. For rapid conversions, the discontinuity $g$-mode has zero frequency because a displaced fluid element near the phase splitting surface adjusts almost immediately its composition to its surroundings and gravity cannot provide a buoyancy force. For slow conversions, a $g$-mode exists and its properties are analysed here using modern hadronic and quark equations of state. Moreover, it has been shown recently that in the case of slow conversions an extended branch of stable hybrid configurations arises for which $\partial M/ \partial \epsilon_c <0$. We show that $g$-modes of the standard branch (that is, the one with $\partial M/ \partial \epsilon_c > 0$) have frequencies and damping times in agreement with previous results in the literature. However, $g$-modes of the extended branch have significantly larger frequencies (in the range $1-2 \, \mathrm{kHz}$) and much shorter damping times (few seconds in some cases). We discuss the detectability of $g$-mode GWs with present and planned GW observatories.

FTIR spectroscopic investigations on the aggregation behaviour of N,N,N′,N′-tetraoctyldiglycolamide and N,N-dioctylhydroxyacetamide in n-dodecane during the extraction of Nd(III) from nitric acid medium

Fourier Transform Infra Red (FTIR) spectroscopy is an excellent tool for probing the co-ordination chemistry of the solvent extracted metal complexes in organic phase. The organic phase containing tetraoctyldiglycolamide (TODGA) and N,N-dioctylhydroxyacetamide (DOHyA) ligands dissolved n-dodecane was employed for the extraction of Nd(III) from nitric acid medium. The extracted organic phase was probed by FTIR spectroscopy to understand the coordination chemistry and aggregation behaviour of Nd(III)-ligand complex in the binary solution. The results were compared with those obtained in the individual solvent system. The extraction of Nd(III) and nitric acid in organic phase was accompanied by shifting and broadening of amidic carbonyl stretching bands of TODGA and DOHyA, and the degree of shift or broadening was dependent on the amount of Nd(III) or nitric acid loaded in organic phase. Near complete shift of amidic carbonyl transmittance bands from 1660 cm−1 to 1610 cm−1 was observed prior to the undesirable event known as organic phase splitting. The results showed the possibility of employing the FTIR spectroscopic technique for understanding the extraction and for the prevention of organic phase splitting during the course of solvent extraction.

A self-adaptive deep learning algorithm for accelerating multi-component flash calculation

In this paper, the first self-adaptive deep learning algorithm is proposed in details to accelerate flash calculations, which can quantitatively predict the total number of phases in the mixture and related thermodynamic properties at equilibrium for realistic reservoir fluids with a large number of components under various environmental conditions. A thermodynamically consistent scheme for phase equilibrium calculation is adopted and implemented at specified moles, volume and temperature, and the flash results are used as the ground truth for training and testing the deep neural network. The critical properties of each component are considered as the input features of the neural network and the final output is the total number of phases at equilibrium and the molar compositions in each phase. Two network structures are well designed, one of which transforms the input of various numbers of components in the training and the objective fluid mixture into a unified space before entering the productive neural network. “Ghost components” are defined and introduced to process the data padding work in order to modify the dimension of input flash calculation data to meet the training and testing requirements of the target fluid mixture. Hyperparameters on both two neural networks are carefully tuned in order to ensure the physical correlations underneath the input parameters are preserved properly through the learning process. This combined structure can make our deep learning algorithm to be self-adaptive to the change of input components and dimensions. Furthermore, two Softmax functions are used in the last layer to enforce the constraint that the summation of mole fractions in each phase is equal to 1. An example is presented that the flash calculation results of a 8-component Eagle Ford oil is used as input to estimate the phase equilibrium state of a 14-component Eagle Ford oil. The results are satisfactory with very small estimation errors. The capability of the proposed deep learning algorithm is also verified that simultaneously completes phase stability test and phase splitting calculation. Remarks are concluded at the end to provide some guidance for further research in this direction, especially the potential application of newly developed neural network models.

Vapor-liquid equilibrium calculations at specified composition, density and temperature with the perturbed chain statistical associating fluid theory (PC-SAFT) equation of state

In this study, the PC-SAFT equation of state is used for vapor-liquid equilibrium calculations using as independent variables the mixture composition, density and temperature. The method is based on unconstrained minimisation of the Helmholtz Free energy via a combination of the successive substitution iteration and Newton-Raphson minimisation methods with line-search; the positive definiteness of the Hessian is guaranteed by a modified Cholesky decomposition. The algorithm consists of two stages; initially, the mixture is assumed to be a single-phase and its stability is assessed; in case of being found unstable, a second stage of phase splitting (flash) takes place, in which the pressure of the fluid and compositions of both the liquid and vapor phases are calculated. The reliability of two different methods presented in the existing literature, (i) using mole numbers and (ii) using the logarithm of the equilibrium constants as iterative variables, is evaluated in terms of both iterations and computational time needed to reach convergence, for seven test cases. These include both single and multicomponent Diesel fuel surrogates, known to give incomplete density information when using pressure and temperature as independent variables. Results show that iterating with the logarithm of the equilibrium constants also reproduces this issue, while it requires a smaller number of iterations than using with mole numbers as independent variables. However, the total computational time needed for the latter case is vastly inferior. Pressure and vapor volume fraction fields are discussed for a range of temperatures and densities, apart from the number of iterations needed during the flash calculation stage. A performance comparison is obtained against the Peng-Robinson equation of state, showing similar number of iterations required but a computational time increasing with the number of components. While for a single component PC-SAFT needs around 3 times more CPU time, for 4 components it is 6 times and for a mixture of 8 components it increases up to 14 times. Finally, the method is demonstrated to converge unconditionally for all conditions tested.

Recovery and Purification of Glycerine as By-product from Philippine Coconut Methyl Ester

The study is about the recovery and purification of glycerine as a by-product from coconut methyl ester production in the Philippines. The aqueous layer produced from settling or phase splitting of the methyl ester after the transesterification process was subjected to various treatments like acidification, neutralization, concentration in vacuo in order to get back the crude glycerine. The crude glycerine obtained from the laboratory and scale-up process conformed with the specification set by the British standard for crude glycerine. The recovered glycerine is composed of: 84.92%, glycerine; 8.03%, ash; 4.72%, H20; 2.32%, MONG. Further distillation yielded a refined glycerine that meets with the specification set by USP. The average glycerine content of refined glycerine is 96.86%; ash, 0.06%; water, 1.10%, refractive index @ 20°C, 1.4696, specific gravity at 25°C, 1.296 g.

Development and Testing of a Novel Game Theoretic De-Centralized Traffic Signal Controller

The paper presents a novel de-centralized traffic signal controller, achieved using a Nash bargaining game-theoretic framework, that operates a flexible phasing sequence to adapt to dynamic changes in traffic demand levels. The Nash bargaining algorithm is used to optimize the traffic signal timings at each signalized intersection by modeling each phase as a player in a game, where players cooperate to reach a mutually agreeable outcome. The algorithm was implemented in the INTEGRATION microscopic traffic assignment and simulation software and tested on two sample networks. The proposed control approach was compared to the operation of an optimum fixed-time coordinated plan, an actuated controller, a centralized adaptive phase split controller, a decentralized phase split and cycle length controller, and a fully coordinated adaptive phase split, cycle length, and offset optimization controller to evaluate its performance. Testing was initially conducted on an isolated intersection, showing a 77% reduction in queue length, a a 64% reduction in vehicle delay, and a 17% reduction in vehicle emission levels. In addition, the algorithm was tested on an arterial network producing statistically significant reductions in total delay ranging between 36% and 67% and vehicle emission reductions ranging between 6% and 13%.

An Innovative synthesis of deep learning techniques (DCapsNet &amp; DCOM) for generation electrical renewable energy from wind energy

Renewable energy becomes one of the main resources that help the world to safety the environment from pollution and provide the people of new type of energy; therefore, this paper presents model called multi-objectives renewable energy-generation (MORE-G) for generating electrical energy from the wind. In general, this model consists of five basic phases: in a first phase collecting and preparing the data, so to make it in format suitable for the decision-making stage, this phase split into multi-steps (i.e., handle missing values and normalization dataset), and the second phase focuses on building constraints for each dataset and develops one of the optimization algorithms called cuckoo based on horizontal combination and multi-objective optimization used in third phase to generate the energy. Another model is developed as multi-layer neural network called (DCapsNet) based on linear combination and multi-objective functions used in the fourth phase to generate the energy. Final phase is related to evaluation of both models (DCOM and DCapsNet) to determine the best. The MORE-G is characterized by addressing one of the real problems, saving on material costs (i.e., reducing the need for manpower and reducing dependence on other countries in importing electric power) and upgrading the scope of the ministry of electricity.

Phase-sensitive optical neural recording of cerebellum tissue on a flexible interface

Knowing an increased number of patients suffering from mental disorders, neural signal recording and imaging have become highly prerequisite challenges for providing healing procedures. Despite the fact that novel optical techniques provide highly resolved imaging/recording of large neuron population, most of them suffer from insertion damage, tethering connection, labeling, and photobleaching deficiencies, among which plasmonic ellipsometry is a highly sensitive and label-free platform for detecting neural activity both quantitatively and qualitatively. In this paper, a flexible patterned plasmonic substrate is used as a sensing surface for phase-sensitive neural recording of a cerebellum tissue slice under electrical and chemical stimulations. Although the traditional reflection spectrum cannot represent the changes in neural activity with high precision, phase-sensitive neuroplasmonics can not only reveal the neural activity level but also distinguish different electrical and chemical stimulation types with a considerable phase splitting factor. This study can open up a new insight into label-free and flexible biological sensors with neuroscience applications.

A potential post-perovskite province in D″ beneath the Eastern Pacific: evidence from new analysis of discrepant SKS–SKKS shear-wave splitting

SUMMARY Observations of seismic anisotropy in the lowermost mantle—D″—are abundant. As seismic anisotropy is known to develop as a response to plastic flow in the mantle, constraining lowermost mantle anisotropy allows us to better understand mantle dynamics. Measuring shear-wave splitting in body wave phases which traverse the lowermost mantle is a powerful tool to constrain this anisotropy. Isolating a signal from lowermost mantle anisotropy requires the use of multiple shear-wave phases, such as SKS and SKKS. These phases can also be used to constrain azimuthal anisotropy in D″: the ray paths of SKS and SKKS are nearly coincident in the upper mantle but diverge significantly at the core–mantle boundary. Any significant discrepancy in the shear-wave splitting measured for each phase can be ascribed to anisotropy in D″. We search for statistically significant discrepancies in shear-wave splitting measured for a data set of 420 SKS–SKKS event–station pairs that sample D″ beneath the Eastern Pacific. To ensure robust results, we develop a new multiparameter approach which combines a measure derived from the eigenvalue minimization approach for measuring shear-wave splitting with an existing splitting intensity method. This combined approach allows for easier automation of discrepant shear-wave splitting analysis. Using this approach we identify 30 SKS–SKKS event–station pairs as discrepant. These predominantly sit along a backazimuth range of 260°–290°. From our results we interpret a region of azimuthal anisotropy in D″ beneath the Eastern Pacific, characterized by null SKS splitting, and mean delay time of $1.15 \, \mathrm{ s}$ in SKKS. These measurements corroborate and expand upon previous observations made using SKS–SKKS and S–ScS phases in this region. Our preferred explanation for this anisotropy is the lattice-preferred orientation of post-perovskite. A plausible mechanism for the deformation causing this anisotropy is the impingement of subducted material from the Farallon slab at the core–mantle boundary.

Conceptual study of co-product separation from catalyst-rich recycle streams in thermomorphic multiphase systems by OSN

Homogeneous transition metal catalysts allow for high selectivity at mild reaction conditions. However, very high recoveries of the precious catalyst are decisive for the development of economic processes, especially in case of reactions with an atom economy below 100% since co-products and their potential accumulation have to be considered. To address this issue, an innovative process concept is investigated, which exploits thermomorphic multiphase systems combined with organic solvent nanofiltration for homogeneous catalyst recovery and co-product removal. In this concept, the reaction takes place at elevated temperature in a homogeneous liquid phase, while the reactor effluent is cooled down, inducing a phase split. This allows for the separation of a product-rich non-polar and a catalyst-rich polar phase that is recycled into the reactor. The current article investigates this concept considering a reductive amination reaction as case study based on an experimental membrane screening and a model-based evaluation. DuraMem 150 (T1) proved to be the best membrane for all tested solvents as catalyst ligand rejections higher than 99% and a removal of the co-product water were achieved. The model-based evaluation indicated a final OSN-process of three stages for the solvent methanol and the requirement of a subsequent solvent recovery that is realized by distillation.

Two-Community Noisy Kuramoto Model Suggests Mechanism for Splitting in the Suprachiasmatic Nucleus.

Recent mathematical results for the noisy Kuramoto model on a 2-community network may explain some phenomena observed in the functioning of the suprachiasmatic nucleus (SCN). Specifically, these findings might explain the types of transitions to a state of the SCN in which 2 components are dissociated in phase, for example, in phase splitting. In contrast to previous studies, which required additional time-delayed coupling or large variation in the coupling strengths and other variations in the 2-community model to exhibit the phase-split state, this model requires only the 2-community structure of the SCN to be present. Our model shows that a change in the communication strengths within and between the communities due to external conditions, which changes the excitation-inhibition (E/I) balance of the SCN, may result in the SCN entering an unstable state. With this altered E/I balance, the SCN would try to find a new stable state, which might in some circumstances be the split state. This shows that the 2-community noisy Kuramoto model can help understand the mechanisms of the SCN and explain differences in behavior based on actual E/I balance.

Eco-Driving at Signalized Intersections: A Multiple Signal Optimization Approach

Consecutive traffic signalized intersections can increase vehicle stops, producing vehicle accelerations on arterial roads and potentially increasing vehicle fuel consumption levels. Eco-driving systems are one method to improve vehicle energy efficiency with the help of vehicle connectivity. In this paper, an eco-driving system is developed that computes a fuel-optimized vehicle trajectory while traversing more than one signalized intersection. The system is designed in a modular and scalable fashion allowing it to be implemented in large networks without significantly increasing the computational complexity. The proposed system utilizes signal phasing and timing (SPaT) data that are communicated to connected vehicles (CVs) together with real-time vehicle dynamics to compute fuel-optimum trajectories. The proposed algorithm is incorporated in the INTEGRATION microscopic traffic assignment and simulation software to conduct a comprehensive sensitivity analysis of various variables, including: system market penetration rates (MPRs), demand levels, phase splits, offsets and traffic signal spacings on the system performance. The analysis shows that at 100\% MPR, fuel consumption can be reduced by as high as 13.8\%. Moreover, higher MPRs and shorter phase lengths result in larger fuel savings. Optimum demand levels and traffic signal spacings exist that maximize the effectiveness of the algorithm. Furthermore, the study demonstrates that the algorithm works less effective when the traffic signal offset is closer to its optimal value. Finally, the study highlights the need for further work to enhance the algorithm to deal with over-saturated traffic conditions.

Fine Single Sand Body Characterization of Hulangmao Chang 6 Reservoir in Jing&amp;apos;an Oilfield

The Chang 6 reservoir in Hulangmao area is a lithologic reservoir controlled by Zhidan Jingbian delta depositional system in Northern Shanxi Province. The oil reservoir has entered the middle-high-water-cut stage. The reservoir is characterized by low porosity and low permeability, with strong heterogeneity and great development difficulty. Therefore, it is necessary to further study the distribution of remaining oil through fine single sand body division and correlation，so as to improve the level of the reservoir development. Based on the theory of high-resolution sequence stratigraphy, according to the characteristics of mudstone marker bed and logging combinations, the sedimentary cycles are studied by using the techniques of maximum entropy spectrum and wavelet transform, and the cycle characteristics of different levels are clarified. Chang 6 member is divided into 3 fourth-level cycles, 6 fifth-level cycles (small levels), 13 sixth-order cycles (single level), and 21 seventh-order cycles (single sand body). Based on the stratigraphic correlation analysis model of sand body equal thickness, channel down-cutting, lateral phase transformation and splitting of superimposed sand body, and combined with the core and logging characteristics of the target area, the correlation framework of single sand layer in the whole area is established.

Investigating the strength and trend of seismic anisotropy in the western part of Makran subduction zone and southeast of Iran

The Makran subduction zone, located in southeastern Iran and southern Pakistan, extends approximately 900 km along the Eurasian-Arabian plate boundary, where the Arabian oceanic plate subducts beneath the Eurasian plate. Using broadband seismic data recorded in the western part of Makran in the southeast of Iran, seismic anisotropy was studied to better understand geodynamic processes of this subduction zone. We analyzed shear wave splitting of core-refracted shear-wave phases (SKS, SKKS and PKS) recorded by 17 broadband stations located in the western Makran and southeast of Iran from 542 earthquakes (MW ≥ 6.0) that occurred between 2004 and 2017 at epicentral distances between 90°–145°. Few stations close to the trench line exhibit a trench-oblique trend in the west to a trench-prependicular trend to the east. The magnitude of splitting delay times suggests a mantle source of anisotropy.Our findings show that the majority of the fast axes of seismic anisotropy at farthest stations in the backarc region are oriented in the NW-SE direction with delay times (created by crust and mantle) varying between 0.8 and 1.3 s. The orientations of these fast directions are almost parallel to surface structural features. For stations near the Minab strike-slip fault system (the transition from the Makran subduction zone to the Zagros collision zone), the fast directions are sub-parallel to the strike of this fault system, indicating that simple shear is likely the main source of anisotropy.

Phase split of oil-water two phase flow at double-layer T-junctions pipes

Offshore oil and gas production systems are in urgent need of a more compact, high-efficient oil-water separator. T-junction pipes are a potential solution to the need. Combining dynamic separation principle of T-junctions pipes with shallow pond theory, this paper presents a novel design of double-layer T-junctions pipes to realize high-efficient oil-water separation. The separation system consisted of upper and lower double-layer pipes that are connected by two vertical pipes. There exists an axially floating oil pipe along the top of the upper outer pipe. This design makes the Reynolds number decrease, which is beneficial to the formation of oil-water stratified flow. The slender seam at the bottom of inner pipe reduces the turbulence intensity, thus avoiding remixing oil with water. The influence of mixture velocity, oil content, water superficial velocity, oil superficial velocity, and inlet flow pattern on the phase split of oil-water two phase flow at double-layer T-junctions pipes has been investigated. Experimental results showed that the separation performance was less affected by the inlet oil content than by the inlet mixture velocity. Additionally, the phase split in the separation system was very sensitive to the inlet flow pattern. For separated flow pattern, oil-water can be well separated. But for dispersed flow pattern, the oil-water separation efficiency was low. Besides, the using of floating oil pipe can increase the system separation efficiency.

New Insights and Assessment of Primary Alkanolamine/Sulfolane Biphasic Solutions for Post-combustion CO2 Capture: Absorption, Desorption, Phase Separation, and Technological Process

The sulfolane was, surprisingly, found to be a universal phase splitting solvent in a serial of aqueous primary alkanolamine solution for CO2 removal, the reaction product of carbamate and hydrophilicity of amine can be the key factor for CO2 triggering phase change, also the phase change can have significant effect on CO2 absorption performance. To comprehensively understand the effect of phase change on absorption, stripping and phase separation performance of primary alkanolamine solution for CO2 capture, MEA sulfolane-based solution was chosen as typical biphasic system for investigation. First, the experimental results revealed that phase change can dramatically decrease the absorption rate and lead to less CO2 capacity in an absorption process. Second, the enriched MEA in separated solution can substantially decrease the energy consumption for CO2 stripping, but exorbitant concentration of MEA can significantly reduce the CO2 removal efficiency in a quasi-cycle of absorption and desorption. Third, t...

Practical application of machine learning on fast phase equilibrium calculations in compositional reservoir simulations

To accurately describe the fluid phase behaviour in reservoir simulation, Equation-of-State-based compositional models are usually used. However, phase equilibrium calculations, including stability tests and phase splitting calculations, may require huge computational costs. An improved artificial neural network model is developed based on our previous work to achieve the prediction of phase stability with high accuracy and reduce the computational costs in order of magnitude. This model does not directly tell if a hydrocarbon mixture at given compositions, pressure and temperature is stable or unstable, and it is able to predict the saturation pressures. By comparing the given pressure with the predicted saturation pressure, it is natural to tell the stability of a hydrocarbon mixture. For the phase splitting calculations, another artificial neural network model is developed to provide more reliable initial guesses for commonly used equilibrium methods to reduce their nonlinear iterations. Compared with our previous work, multi-layer neural networks are employed in the models. The improved models have more efficient training processes and more accurate prediction results. Additionally, an adaptive data generation process is introduced to optimize the quality and size of training data; the data transformation and normalization are discussed to reduce the skewness of the data and rescale the data; and a model training and selection strategy is also discussed to efficiently train a high-quality model. The efficiency of the ANN models is first validated by standalone phase equilibrium calculations with a three-component hydrocarbon mixture. Then the ANN models are successfully applied in compositional simulation examples, which demonstrates the practical value of these ANN models on speeding up the phase equilibrium calculations during compositional simulations.