Fuel 115 (2014) 116

Contents lists available at SciVerse ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Development of a thermodynamically consistent, robust and efcient phase equilibrium solver and its validations
Lu Qiu 1, Yue Wang, Qi Jiao, Hu Wang, Rolf D. Reitz
Engine Research Center, University of Wisconsin-Madison, 1500 Engineering Drive, Madison, WI 53705, USA

h i g h l i g h t s
 A robust and efcient equilibrium solver is developed.  The solver is based on rigorous thermodynamics.  Complex phase behaviors of multi-component mixtures are predicted.  Liquidliquid and vaporliquidliquid equilibria are predicted.  Feasible region method is applied to accelerate three-phase ash.

a r t i c l e

i n f o

a b s t r a c t
An applied phase-related equilibrium (APPLE) solver using only the PengRobinson equation of state is developed based on rigorous classical thermodynamics. The solver is theoretically and thermodynamically consistent with the stringent equilibrium criterion. It is mainly composed of phase stability and phase splitting calculations, which will be called routinely in the course of searching for the globally stable equilibrium state. It also makes use of various robust and efcient numerical methods. To demonstrate its performance, the solver is tested against various mixtures, such as oil and gas mixtures, hydrocarbon mixtures and hydrocarbonnitrogen mixtures. Phase diagrams of these mixtures are constructed and veried with available experimental data or other researchers calculations. Results show that the APPLE solver is reliable and fast to solve phase equilibrium problems, including three-phase equilibrium. Finally, its potential applications to droplet evaporation and computational uid dynamics (CFD) calculations are discussed. 2013 Elsevier Ltd. All rights reserved.

Article history: Received 3 May 2013 Received in revised form 25 June 2013 Accepted 25 June 2013 Available online 12 July 2013 Keywords: Phase equilibrium Phase stability Phase splitting Flash calculation

1. Introduction The phase transition/change process is an important physical process that occurs in many engineering applications, such as CO2 injection and CO2 storage in enhanced oil recovery, and liquid phase combustion upon evaporation. Considerations on the dynamics of phase transition, which requires analyses of non-equilibrium thermodynamics such as the nucleation process, are incomplete if the equilibrium state (destination state) is not known. Phase equilibrium, hence, is a starting point for further non-equilibrium analyses and it is of primary importance from the point view of classical thermodynamics. In reality, phase equilibrium thermodynamics-based calculations have been used in a wide range of industrial elds, such as for distillation columns,
Corresponding author. Tel.: +1 608 262 0145; fax: +1 608 262 6707.
E-mail addresses: lqiu4@wisc.edu (L. Qiu), wang46@wisc.edu (Y. Wang), qjiao2@wisc.edu (Q. Jiao), hwang259@wisc.edu (H. Wang), reitz@engr.wisc.edu (R.D. Reitz). 1 Tel.: +1 662 617 9132. 0016-2361/$ - see front matter 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.fuel.2013.06.039

and in reservoir simulation. An equation of state, as an important part of thermodynamics needed to quantify the state properties, is generally used to study non-ideal gas behaviors of mixtures. From the fundamental thermodynamic postulate, the entropy is a continuous and monotonically increasing function of energy [1]. In terms of the Gibbs free energy G for a homogeneous and open system, the second law of thermodynamics limits the possible processes through:

dG VdP SdT

li dni 6 0

Here, T is temperature; P is pressure; V is volume; S is entropy, l is chemical potential; n is mole number and i is the species index. At the equilibrium state, the equality sign in the above equation holds. It is well known that the Gibbs free energy at the equilibrium state is the minimum in the phase space composed of T, P and ~ n. Mathematically, the problem now becomes to search for a global minimum point in multi-dimensional space. This problem is usually tackled by solving for the equality of the chemical potential, which is the partial molar Gibbs free energy. In actual calculations,

L. Qiu et al. / Fuel 115 (2014) 116

fugacity equality is commonly used for identifying the equilibrium state by requiring, for a two-phase equilibrium,

fi1 fi2 ;

i 1; . . . ; N c

Here, j is the index of phase; f is the fugacity and Nc is the number of components. The advantage of this method is that fugacity coefcient can be relative easily calculated from various equations of state, such as the PengRobinson equation of state [2]. However, Eq. (2) is not a sufcient condition for phase equilibrium. Mathematically, the zero derivative location only indicates a local extreme but does not guarantee that it is also the global extreme. Hence, the equality of fugacity in Eq. (2), standing for a local stationary point in the phase space, is not necessarily the global minimum at the same time. In other words, solution of Eq. (2) could represent a false equilibrium state, which only corresponds to some local minima. As a result, fugacity equality is a necessary but not sufcient condition for phase equilibrium. While for simple mixtures, the local extreme could happen to be a global extreme, the situation quickly becomes complex when a multi-component mixture is considered since the phase dimension and the number of local extrema will increase. This is especially the case that occurs quite often at liquidliquid and vaporliquidliquid equilibria [3]. All these noteworthy points are quantitatively illustrated in the work of Baker et al. [4], beneting from the original work of Gibbs [5]. They suggested using Gibbs free energy directly for phase equilibrium to guarantee the true equilibrium state, rather than the traditional way with fugacity. They further proved that the necessary and sufcient condition for a system to be stable, for specied temperature, pressure and species, is that the tangent plane at the species feed composition to the Gibbs free energy surface should at no other point intersect (lie above) the Gibbs free energy, but only the tangent point. This paper is organized as follows. The equation of state model is introduced rst, since it is used for both phases throughout all the calculations. After this, the methodologies adopted for the current development of a robust and efcient applied phase equilibrium solver, named APPLE here, are presented in detail. We then evaluate the performance of the solver for various cases from the literature. Some potential applications of the solver to computational uid dynamics (CFD) calculations are remarked upon. Summary and conclusions are then made. 2. Equation of state The equation of State is a relationship between thermodynamic properties at a specied state. To consider the non-ideality of thermodynamic properties at high pressures, a proper equation of state other than the ideal gas law is needed. While there are no general guidelines for selecting a specic equation of state, the PengRobinson equation of state (PR EOS) [2] was chosen for all the calculations here for several reasons. First, it is a simple form of the cubic equation of state so it is easy to implement for engineering calculations. Second, it has better performance for the prediction of vaporliquid phase equilibrium properties over other cubic equation of states. More importantly, it also has been widely used in the oil and gas industry with great successes. Theoretically, using only one equation of state for both the liquid and gas phases would bring some numerical difculties since more solutions of cubic equations are possible. However, it has the benet of using one unied treatment for both phases and hence it is easier to handle and elegant to implement. In addition, the non-random mixing model, such as UNIFAC, is not used here because it is specially designed for handling the non-ideality of compressed liquids or solutions; and it is mainly used for subcritical conditions where the molecular interactions are strong

and surface tension effects are important. An equation of state model, on the other hand, can be used for a wide range of conditions (from subcritical to supercritical). Besides, there is no need to specify the reference state for equation of state models since ideal gas limit is used exclusively. Finally, enthalpy and some other thermodynamic properties can be determined directly from an equation of state model in a consistent manner. The PR EOS [2] is of the form:

RT

v b v v b bv b

Here, v is the molar volume. a and b are the two parameters determined from:

8 R2 T 2 c > > < a 0:457235 Pc a c b 0:077796 RT Pc > > p 2 : a 1 j1 T r

With

j 0:37464 1:54226x 0:26992x2

5a

Here, Tc, Pc and x are the critical temperature, pressure and acentric factor, respectively. We also adopted a formula that is later expanded [6], which uses, when x is greater than 0.5.

j 0:3796 1:485x 0:1644x2 0:01667x3

5b

In the case of a mixture, the classical Van der Waals mixing rule is used:

8 XX > a xi xj aij > > > > i j < X b xi b i > > > i > > pp : aij 1 di;j ai aj

Here, xi is the mole fraction of species i in the mixture. di,j is the binary interaction parameter between components i and j, and it is generally assumed to be independent of pressure or temperature for a mixture. Temperature-dependent interaction parameters can be used to improved vaporliquid equilibrium predictions with the group contribution method discussed in Ref. [7,8]. This group contribution method is used in the petroleum industry to nd interaction parameters for hydrocarbons (Avaullee et al. [9]), but it is not pursued here. More advanced mixing rules are not pursued in the current work, either. Cubic equations generally have three roots at sub-critical conditions. In all the calculations here, when a compressibility root is to be solved, only the one with minimum Gibbs free energy is chosen. 3. Methodology The equilibrium condition shown in Eq. (1) is written in terms of Gibbs free energy, which has been well applied to isothermal isobaric ash (TPn) calculations. This kind of ash calculation has been investigated by many researchers and there are many welldeveloped theories and numerical methods. There is a growing recognition that robust and efcient phase equilibrium calculations should be composed of two essential parts: phase stability and phase splitting calculations [3,10,11]. 3.1. Phase stability The traditional two-phase ash calculation assumes that a twophase solution exists. If no solution or trivial solution is found, a single phase is then believed to exist. This becomes more and more troublesome when the mixture approaches its thermodynamic

L. Qiu et al. / Fuel 115 (2014) 116

critical point, where most ash techniques are likely to fail due to the similarity between the vapor and liquid phases. Another issue is that sometimes a blind ash search may even yield the wrong solution at a local extreme of Gibbs free energy. This point was also made by Baker et al. [4], as mentioned above. However, their criterion is more descriptive than useful because they did not disclose any numerical methods that can actually be implemented. Since the actual number of phases coexisting at a specied thermodynamic state is not known a priori, Michelsen [12] suggests that a phase stability test be conducted before the ash calculation. Similar to the Gibbs free energy analyses of Baker et al. [4], Michelsen [12] used the Gibbs free energy of mixing to develop a method called the Tangent Plane Distance (TPD). The TPD function is dened as:

tangent hyper-plane of the energy surface at the stationary point and the tangent hyper-plane at the feeding composition. It has been proved that TPD~ x has the same stationary points as TPD~ x and they have the same sign at the stationary points [12]. Nonetheless, TPD~ x is more favorable because it now converts the constrained minimization problem to an unconstrained one (except that X > 0). This treatment bypasses the numerical difculties of the constrained minimization problem such as the boundary constraints. Hence, in the current calculation, Eq. (10) is selected as the objective function for the minimization. To avoid negative mole fractions during the iterations, Michelsen [12] suggested to change the iteration variable as:

ai 2 X i

14

x TPD~

Nc X z x li ~ xi li ~ i 1

z is the feeding composix is the mole fraction array and ~ Here, ~ tion array. Geometrically, the TPD represents the vertical distance from the tangent hyper-plane (to the molar Gibbs free energy surface of ~ z) to the molar Gibbs free energy surface of ~ x. It can be proved that the stability test requires the TPD to be non-negative:

The above conversion bypasses the negative X issue. It should be mentioned that, during the iteration, X is still used when the fugacity coefcient is needed. The rst order derivative of TPD~ x is needed and it is calculated according to

p @ X i lnX i hi TPD X i ln ui ~ @ ai

15

TPD~ x P 0

for any accessible ~ x subject to the species conservation constraints:

xi P 0 and

xi 1

If TPD~ x is always non-negative for all feasible ~ x, then the original mixture is thermodynamically stable and exists as a single phase. On the other hand, if TPD~ x has some negative values, the original mixture is not stable and the phases will separate. Since all the minima are in the region of the physical constraint, TPD will be non-negative if it is non-negative at all the stationary points, where the derivative with respect to all independent variables is zero. Therefore, this method reduces the search space from the whole domain to the local extremes (i.e., local maxima, minima or saddle points) and hence an exhaustive search of all possible f~ xg is not needed. It is remarked that the TPD is strictly zero at saturation point line; Implementation of such criterion to nd the saturation temperature or pressure directly without building the complete phase envelop can be found in Ref. [13]. Even with this treatment, the actual search is still challenging, especially if the dimension of the phase space is large. For a multi-component mixture, geometrically the Gibbs free energy of mixing is a hyper-surface and the tangent plane is a hyper-plane [4]. When an equation of state is used, Eq. (7) can be expressed in terms of fugacity coefcient

Even with all these numerical treatments, convergence is sometimes still quite difcult, and multi-sided initial guesses are further suggested [3,12]. Hoteit and Firoozabadi [14] applied the Broyden FletcherGolfarbShanno (BFGS) algorithm to nd local extremes. This Quasi-Newton type method approximates the Hessian matrix rather than calculating it as in a standard Newtons method, so it is computationally inexpensive. With each initial guess, a local stationary point, if any, is determined. The global minimum is then the minimum of all these candidates. The updated scheme for this method is:

~ g k ; sk ; F k ak1 ak F
With

16

g k F ak1 F ak sk ak1 ak

17

x TPD~
where

Nc X x hi P 0 xi ln xi ln ui ~ i 1

10

z hi ln zi ln ui ~

11

Michelsen [12] formulated a unconstrained minimization problem in terms of a variable X:

TPD ~ x 1
where

Nc X X lnX i hi 1 X i ln ui ~ i 1

12

Here F stands for the gradient of TPD~ x mentioned above. Therefore, a local extreme indicates that F is zero in each direction. ~ is the update matrix simplied from the BFGS algorithm. Details F about the numerical implementation can be found in [14], as well as its application in a reduction method. This method, as will be shown later, is found to perform quite well in the non-linear minimization and it is fast to converge. There is another category of method using global stochastic search methods, such as the Homotopy method [15]. Nichita et al. [16] used a global optimization method called Tunneling to escape local extremes and saddle points. A recent review on Generic and Monte Carlo algorithms can be found in Zhang and Bonilla-Petriciolet [17]. Most of the global optimization methods, however, do not guarantee convergence, either. Even if a global minimum can be found, it can only be achieved after exhaustive and cumbersome searches, so there is always a heavy load on the computation time and resources. Since we are pursuing a fast and robust phase equilibrium solver that will be coupled with computational uid dynamics (CFD) calculations, these stochastic and global searching methods are not pursued. Especially when considering reacting ows, a large amount of computer time is spent on chemistry mechanism, which is already complicated by the stiffness of the reaction system. 3.2. Phase splitting

X i xi expK

13

Xi here can be considered as mole number and K is a constant. Geometrically, K stands for the vertical distance between the

Phase splitting occurs whenever a phase stability test yields a negative TPD value, indicating that the original mixture is non-stable (either stable or metastable). It is traditionally called ash

L. Qiu et al. / Fuel 115 (2014) 116

calculation, which is mainly based on the RachfordRice algorithm [18]. For a two-phase equilibrium, the RachfordRice equation is
Nc X zi ki 1 0 1 kki 1 i 1

18

Here, ki is the phase equilibrium ratio of species i and k is the mole fraction of one phase. The objective function is a monotonically decreasing, but not continuous, function. Together there are Nc poles for k. Hence, caution is needed to window the searching region between two asymptotes and the physical solution is bounded between 0 and 1. At the same time, depending on the initial guess, the trivial solution, showing identical composition in both phases, can also be generated. More seriously, an incorrect solution can be generated by a blind search, especially at liquidliquid and three-phase equilibriums. While implementation of this algorithm is straightforward for traditional two-phase vaporliquid equilibrium, the multi-phase ash is not a trivial task. The multi-phase equilibrium, however, occurs quite often in real situations, such as in carbon dioxide (CO2) utilization for enhanced oil recovery, as well as when oil, gas and water coexist at low temperatures. These situations will be shown in the following calculations. Analogous to Eq. (18), the corresponding three-phase RachfordRice function is [19,20]:

where bi = {1 kij}, b = {bi}, and bi minf1 zi ; minj f1 kij zi gg for i 1; . . . ; N p 1. This feasible region does not contain any poles where the Hessian matrix is ill conditioned, so a minimization scheme with the aid of line search using the Hessian matrix is applicable. In addition, the constraint SC fbjaT i 1; . . . ; N c g i b 6 bi ; creates a much smaller feasible region compared to the traditional constraint LC fbjhi P 0; i 1; . . . ; Nc g developed by Leibovici and Neoschil [24]. For positive ash, another constraint PC fbjbj P 0; j 1; . . . N p g is imposed. Independent of the number of phases for both positive and negative ash, this method guarantees convergence [20]. Generally, it requires only 67 iterations for convergence once the feasible region is determined; making the method is very robust and practical. For three-phase ash calculations, only when close to the bi-critical point will the iterations slow down, which is mainly because the lines from linear constraints are close to parallel with each other. The above-mentioned three widely used methods for three-phase ash, bisection method, direct Newtons method and Okunos feasible region method, have been studied very recently by Li and Firoozabadi [11]. Their calculations show that the direct Newtons method is the most efcient one. 3.3. Summary For the development of the APPLE solver, we adopted the following methodologies. For a TPn ash at specied temperature, pressure and species mole numbers, we apply an equilibrium calculation in a stage-wise manner. Starting from one phase, the stability of the mixture is tested. If it is proven non-stable from the test, another phase is added with phase splitting and a ash calculation is conducted. The above procedures continue until the mixture is stable. Formulated in this way, the true equilibrium solution is guaranteed since a new phase is added if and only if it decreases the mixture Gibbs free energy. This methodology is also mentioned in Li and Firoozabadi [11]. For the phase stability part, the above-mentioned BFGS localized minimization scheme [14] is implemented to benet from its fastness and effectiveness. For the initialization, instead of searching for a global minimum using some stochastic methods, the multi-sided test suggested by Li and Firoozabadi [10] is adopted. For a mixture composed of Nc components, together Nc + 4 tests are needed for a full phase stability test. The total CPU time, however, is actually very cheap even with multi-sided guesses because the Hessian matrix calculation is not needed, which generally requires large amounts of computational effort. While the initialization scheme can locate a maximum of three non-trivial solutions [10], it is sufcient to locate the global minimum for stability testing based on our experience. For the phase splitting part, we solve the RachfordRice equations after the stability test. In a two-phase ash, with the initial guess from the one phase stability, either the bisection or Newtons method is used. The two-phase stability test can be done on either phase, and the phase with the higher molecular weight is selected. For the RachfordRice three-phase ash, both successive substitution and Okunos method are implemented. The latter is used as a backup if the former fails, though this rarely happens in actual calculations. Details on the initialization strategies for the numerical methods can be found in [10,11] and their suggestions were found to be reliable. 4. Results and discussion The tangent plane distance theory for stability testing and the numerical methods for phase splitting summarized in Section 3.3 were implemented in a Fortran program to formulate the applied

 Nc  X zi kij 1
i 1

hi

0;

j 1; . . . ; N p 1

19

where

hi 1

N p 1 X j 1

kij 1bj ;

i 1; . . . ; N c

20

kij

xij ; xiNp

i 1; . . . ; N c ;

j 1; . . . N p 1

21

Here, Np is the number of phases; kij is the phase equilibrium ratio of component i in phase j relative to phase Np; bj is the mole fraction of phase j. Notice that the notation above implicitly assumes that phase Np is the reference phase. Methods for three-phase equilibrium calculations before the TPD method seem not to be widely used and are limited by numerical methods, which lack a systematic thermodynamic treatment. For example, Mehra et al. [21] studied multiphase equilibrium up to four phases using rst-order and second-order methods. Depending on the numerical conditions, switches between these methods are needed. However, the numerical solutions are not guaranteed to be thermodynamically stable states. When two solutions are found, a further comparison on the mixture Gibbs free energy is needed. With the phase stability test introduced, a systematic way of calculating phase equilibrium is possible. Michelsen [12] suggested to check phase stability before the ash calculations and to use the outcomes of stability as an initial guess in the phase splitting calculations. Many calculations [10,11,19] also prove that the phase stability test can supply a very good initial guess for convergence in ash calculations. Nghiem and Li [22] applied the TPD stability test and applied it to a quasiNewton Successive substitution (QNSS) scheme to study multiphase equilibria of reservoir oil and CO2 mixtures. Haugen et al. [23] showed that a combination of bisection and Newton methods is another feasible approach that works quite well for a number of three-phase calculations on reservoir oils. Okuno et al. [20] recently published a robust and efcient multi-phase ash method named the feasible region algorithm. The algorithm is based on minimization of a non-monotonic convex function with Nc linear constraints. The feature of this method is that it limits the solution in a small feasible region constrained by:

aT ib

6 bi

22

L. Qiu et al. / Fuel 115 (2014) 116

phase-related equilibrium (APPLE) solver. In the following, we present tests on a variety of mixtures. Most of them are for reservoir mixtures, as a large amount of available data come from this eld. The phase diagram, which is considered a stability map, is critical to characterize the phase transition processes, and will be the focus in the following sections to evaluate the performance of APPLE solver. Dependence of phase equilibrium state on pressure will also be discussed. Since the PengRobinson equation of state is used, it is important to document clearly the properties and parameters used [25]. The thermodynamic properties used as inputs to the solver, such as critical properties and binary interaction parameters, are all documented in tables for each test. In all the calculations of phase diagrams, unless mentioned separately, the phase boundary is recorded when a new phase appears or an old phase disappears if pressure is changed by 0.001 bar. 4.1. Test problem 1: CH4H2S Calculation of the CH4H2S binary mixture is rst performed to illustrate the usefulness of the stability test for the ash calculations to get the correct solution. The mixture properties are summarized in Table 1. This classical problem has been also tested in Refs. [12,15,16,26]. The tangent plane distance is shown in Fig. 1(a) for the 50:50 (by mole) mixture, it is noticed that it has ve local extremes. The negative sign of the distance indicates that the binary mixture cannot stabilize itself in one phase, neither a single liquid nor a vapor phase. However, if the numerical method becomes stuck at a local extreme at zero TPD, a false one-phase solution will be found. Even if a non-stable phase is detected, if the phase splitting calculation is not robust or the initial guess is not good, the calculation may still converge to another false vaporliquid two-phase solution. This point is best illustrated by considering the Gibbs free energy of mixing, shown in Fig. 1(b). It is noticed that mixture could exhibit both liquidliquid and vaporliquid behaviors, and only the former corresponds to the minimum of Gibbs free energy. At this situation, phase stability is critical in supplying a good initial guess for the subsequent ash calculation. The solution from the current equilibrium solver can read directly from Fig. 1(b), which agrees well with the reference values in Refs. [12,15,16,26]. 4.2. Test problem 2: C1CO2H2SH2O The mixture composed of C1CO2H2SH2O is studied. The thermodynamic data and calculation results are tabulated in Tables 2 and 3, respectively. Convergence was always achieved. The number of calculation iterations increase a little from a two-phase ash to a three-phase ash. In Table 3, T and F stand for stable and non-stable mixture, respectively. The numbers in the table are the mole fractions of each phase. The phase stability results of the three systems in Table 3 are consistent with those in Ref. [10]. 4.3. Test problem 3: acid gas The APPLE solver was applied to study the behaviors of acid gas and its mixing effects with CO2. The inputs needed for the solver are documented in Table 4. This mixture turns out to exhibit very complex behaviors, as shown in Fig. 2(a). Note that since the acid
Table 1 Inputs for binary C1H2S mixture at 190 K and 40.53 bar. Species C1 H2S Composition 0.5 0.5 Tc (K) 190.6 373.2 Pc (bar) 46.0 89.4

gas already has CO2 in it, the data shown in the gure is the total amount of CO2. Depending on the conditions, the mixture can be at one phase, two-phase and three-phase states. For the two-phase equilibrium, it can be at a vaporliquid or a liquidliquid equilibrium. The three-phase region spans a wide range of conditions. At low pressure, the mixture is in the one-phase region. As pressure increases, phase separation occurs and the mixture can enter both two-phase and three-phase regions. To better characterize the complex phase behaviors, especially in the three-phase region, we adopt the saturation line concept, which is an original concept for two-phase mixtures. For an existing mixture, whether one phase or two-phase, when a fresh liquid phase forms across a line, we name this line the dew point line. Similarly, when a fresh vapor phase forms across a line, we name it the bubble point line. Therefore, the meeting point of the bubble point and dew point lines, such as the point at (83%, 22.94 bar, shown by the circle in Fig. 2(a)), does not indicate that it is necessarily the critical point, which indeed will be for mixtures that only show one-phase or two-phase regions. This situation also occurs for the other mixtures. These saturation lines are also explicitly labeled in Fig. 2 and all the subsequent phase diagrams. It is noticed that the single vapor phase only exists at very low pressures and the liquid phase starts to form with pressure increase. Therefore, the phase boundary is mainly characterized by the six branches of the dew point lines. The upper two dew point lines, however, have a large slope so that the mixture is very sensitive to pressure effects. The bubble point line only shows up at higher pressures when the vapor phase disappears. Note the deep penetration of the dew point line into the low-pressure region. Fig. 2(b) shows the disappearance of the vapor phase from the three-phase to the liquidliquid region with 60% CO2 feeding. The fraction and percentage in all the calculations in this paper indicate the mole fraction. It is noticed that the special location (95%, 0.7 bar) (shown in the blow up plot) is the point where the single vapor and vaporliquidliquid regions have the smallest distance, and these two regions do not intersect. It is of great interest to study retrograde behaviors. This special behavior occurs when two saturation point lines are not separated by the critical point, the location where bubble and dew point lines meet. For this mixture, it is noticed that retrograde condensation occurs at the three-phase region along the phase border of the two vaporliquid regions. Due to the curvature effects of the two dew point lines, retrograde condensation occurs at both the left (when CO2 feeding is between 0.261 and 0.469. The starting and ending points were highlighted using yellow circles) and right branches (when CO2 feeding is between 0.832 and 0.857. The starting and ending points were highlighted using green circles). Usually, the maximum amount of liquid phase is located somewhere between the dew points for a given feeding composition. This retrograde condensation can be shown to lie along an isothermal process. As an example, Fig. 3(a) shows retrograde condensation of the CO2-rich liquid phase (L2) at the left branch when the CO2 feeding is 40%. Before reaching the 26.034% maximum liquid amount at 6.0 bar, increase of pressure leads to increase of phase L2 due to compression and condensation. Further increase in pressure does not lead to liquid phase formation. On the contrary, phase L2 starts to decrease and the CO2 lean liquid phase (L1) builds up. The CO2 mole fractions for each phase during this process are plotted in Fig. 3(b). A similar plot is shown in Fig. 3(c) for retrograde condensation of phase L1 at 84% CO2

x ()
0.011 0.097

MW (g/mole) 16 34

ki;C 1 0.083

ki;H2 S 0.083

L. Qiu et al. / Fuel 115 (2014) 116

Fig. 1. Gibbs energy analyses of the binary mixture CH4H2S at 190 K and 40.53 bar. (a) The tangent plane distance (TPD) function. (b) Gibbs free energy of mixing.

Table 2 Inputs for the mixtures of C1CO2H2SH2O. Species C1 CO2 H2S H2O
a b c

Tc (K) 190.6 304.2 373.2 647.3

Pc (bar) 46.0 73.8 89.4 220.5

x () 0.008 0.225 0.1 0.344

MW [g/mole) 16 44 34 18

ki;C 1 0.095 , 0.1005 , 0.13 0.0755b, 0.095c 0.4928

a b c

ki;CO2

ki;H2 S

0.0999b, 0.097c 0.04

C1CO2 mixture. All the cases of C1CO2H2SH2O mixture. C1CO2H2S mixture.

Table 3 Equilibrium solution of the C1CO2H2SH2O mixtures. Mixture C1CO2 T (K) 220.0 P (bar) 60.8 Composition (0.9,0.1) (0.8,0.2) (0.7,0.3) (0.57,0.43) (0.4,0.6) (0.4989, 0.0988, 0.4023) (0.1488, 0.2991, 0.0494, 0.5027) (0.1496, 0.3009, 0.0498, 0.4997) (0.0504, 0.0503, 0.3986, 0.5008) 1P-stab T F F F T F F F F 2P-splitting (0.9701, 0.0299) (0.5695, 0.4305) (0.0488, 0.9512) (0.4323, 0.5677) (0.50404, 0.49596) (0.4936, 0.5064) (0.4956, 0.5044) 2P-stab T T T F T T F (0.4092, 0.0285, 0.5623) 3P-splitting

C1CO2H2S C1CO2H2SH2O

190.16 310.95 380.35 310.95

26.82 76.0 129.3 62.6

(0.4938, 0.0797, 0.4265)

Table 4 Component properties and non-zero interaction parameters for the acid gas. Species CO2 N2 H2S C1 C2 C3 Composition 0.70592 0.07026 0.01966 0.06860 0.10559 0.02967 Tc (K) 304.211 126.2 373.2 190.564 305.322 369.825 Pc (bar) 73.819 33.9 89.4 45.992 48.718 42.462

x (-)
0.225 0.039 0.081 0.01141 0.10574 0.15813

MW (g/mole) 44.0 28.0 34.1 16.0 30.1 44.1

ki;CO2 0.02 0.12 0.125 0.135 0.150

ki;N2 0.02 0.2 0.031 0.042 0.091

ki;H2 S 0.12 0.2 0.1 0.08 0.08

feeding. Notice that the amount of phase L1 is very low, so it is increased by 50 times for a clear illustration. In comparison, it is remarked that the left retrograde region occupies a larger area, indicating that the condensation rate is less at the left branch than at the right one. In other words, careful pressure control is needed to avoid another liquid phase forming at high CO2 concentrations. The corresponding change of CO2 in each phase is shown in Fig. 3(d). In both cases, the vapor phase amount keeps decreasing, as well as the CO2 mole fraction in this phase. In addition, a further comparison between Fig. 2(b) with Fig. 3(a) and (c) shows that in the retrograde region, non-linear behavior due to curvature effects is more obvious.

4.4. Test problem 4: Maljamar reservoir oil and Maljamar separator oil The Maljamar reservoir oil and separator oil, both exhibiting interesting three-phase behaviors, have been well studied in [10,23,27]. The component properties and parameters for these two oils are given in Tables 5 and 6, respectively. The phase diagram of the Maljamar reservoir oil is shown in Fig. 4(a). Two bi-critical points with coordinates (64.6%, 94.02 bar) and (99.6%, 75.00 bar) are shown at the lower right and upper left portions of the phase boundary. The bi-critical point here is dened as the point at which intensive properties of two phases, such as density, compressibility, composition and hence molar weight, in the

L. Qiu et al. / Fuel 115 (2014) 116

Fig. 2. (a) Phase diagram of the acid gas (6 components). Symbols V, L1 and L2 denote the vapor, CO2-lean and CO2-rich liquid phases. Also shown are the bubble and dew point lines. Meeting points between the bubble and dew point lines are not critical points. (b) Evolutions of the three phases during a pressurizing isothermal process at 60% CO2 feeding.

Fig. 3. Two different retrograde behaviors of the acid gas. (a) Retrograde condensation of CO2-rich liquid phase L2 at 40% CO2 feeding. (b) CO2 mole fraction in the three phases during the retrograde condensation process in (a). Note that CO2 fraction in L2 is reduced by 50%. (c) Retrograde condensation of CO2-lean liquid phase L1 at 84% CO2 feeding. Note that phase fraction in L1 is increased by 50 times. (d) CO2 mole fraction in the three phases during the retrograde condensation process in (c). Note that CO2 fraction in L2 is reduced by 50%.

Table 5 Component properties and non-zero interaction parameters for Maljamar reservoir oil. Species CO2 C1 C2 C3 nC4 C57 C810 C1114 C1520 C2128 C29+ Composition 0.0 0.2939 0.1019 0.0835 0.0331 0.1204 0.1581 0.0823 0.0528 0.0276 0.0464 Tc (K) 304.211 190.6 305.4 369.8 425.2 516.667 590.0 668.611 745.778 812.667 914.889 Pc (bar) 73.819 45.4 48.2 41.9 37.5 28.82 23.743 18.589 14.8 11.954 8.523

x ()
0.225 0.008 0.098 0.152 0.193 0.2651 0.3644 0.4987 0.6606 0.8771 1.2789

MW (g/mole) 44.0 16.0 30.1 44.1 58.1 89.9 125.7 174.4 240.3 336.1 536.7

ki;CO2 0.115 0.115 0.115 0.115 0.115 0.115 0.115 0.115 0.115 0.115

ki;C 1 0.115

0.045 0.055 0.055 0.06 0.08 0.28

L. Qiu et al. / Fuel 115 (2014) 116

Table 6 Component properties and non-zero interaction parameters for Maljamar separator oil. Species CO2 C57 C810 C1114 C1520 C2128 C29+ Composition 0.0 0.2354 0.3295 0.1713 0.1099 0.0574 0.0965 Tc (K) 304.211 516.667 590.0 668.611 745.778 812.667 914.889 Pc (bar) 73.819 28.82 23.743 18.589 14.8 11.954 8.523

by green circle) is used to illustrate the details in the contiguous region between phases L1 and V-L1-L2. It is seen from the plot that the single phase region does not intersect with the three-phase region. 4.5. Test problem 5: Oil B

x ()
0.225 0.2651 0.3644 0.4987 0.6606 0.8771 1.2789

MW (g/ mole) 44.0 89.9 125.7 174.4 240.3 336.1 536.7

ki;CO2

0.115 0.115 0.115 0.115 0.115 0.115

three-phase mixture become identical. This also means the equilibrium ratios of species in these two phases are the same. It should be pointed out that these phases, determined from the three-phase splitting calculations, exist in nite amounts that are not negligible. Some other criteria can also be used, such as the tangent plane distance as used by Li and Firoozabadi [10]. Therefore, it is also expected that the bi-critical point is located at the location where the bubble and dew point lines meet, analogously to the critical point for a two-phase mixture. It is noticed that, for a wide range of CO2 feedings up to 65%, the mixture is in vaporliquid state. Another CO2 rich liquid phase comes out at higher concentrations. The three-phase peninsula penetrates all the way up to almost pure CO2. Above the bubble point line, due to the disappearance of the vapor phase, only liquidliquid equilibrium exists. To illustrate the evolutions of the three phases, Fig. 4(b) shows the dynamic change of these phases as a function of pressure with 80% CO2 feeding. Away from the dew point line, the fraction of the vapor phase gradually decreases and reaches zero at the bubble point. The CO2 lean liquid phase amount also decreases but not too much during the pressuring process. Therefore, the overall effect is to increase the CO2 rich liquid phase continuously. The pressure-composition phase diagram of the Maljamar separator oil is presented in Fig. 5(a). There are two distinctive behaviors when compared with the reservoir oil. First, a single liquid phase is found when CO2 feeding is less than 67%. Along the single-phase boundary, a bubble point line exists at low pressures while a dew point line is found at high pressures. Secondly, noticing the different scales in the vertical coordinates, the three-phase region now is stretched to higher CO2 feedings and its area decreases. Thirdly, only one bi-critical point exists at the close boundary between the liquidliquid and vaporliquid regions. The three-phase region ends at this bi-critical point. The phase evolution at 90% CO2 feeding is shown in Fig. 5(b). A similar trend is observed as for the reservoir oil: the vapor phase decreases, L1 phase decreases and L2 phase increases. A blow up plot (indicated

The phase diagram of Oil B was also constructed. The thermodynamic properties used for this oil are documented in Table 7. This mixture was used by Shelton and Yarborough [28] to investigate the mixing effects of CO2 with reservoir oil used in enhanced oil recovery. The general shape of the phase diagram is similar to that of Maljamar separator oil, as presented in Fig. 6(a). There are onephase, two-phase and three-phase regions, too. Interestingly, there is a point where the four regions almost touch at the point (58%, 67.9 bar) and this kind of behavior also appears in some other oils [29]. The three-phase region ends at a bi-critical point at the twophase boundary very close to pure CO2. Below the bubble point line, mixture is in the vaporliquid region for a wide range of CO2 feedings. At higher pressures, the singe phase dominates until reaching the dew point line across which a CO2 rich liquid phase forms when the CO2 feeding is around 73%. Between the liquidliquid and vaporliquid regions, there is a three-phase region acting like a transition region. This kind of phase behavior seems to be general since no sharp phase transition is found, as also can be observed from Figs. 2(a), 4(a) and 5(a). There are no common meeting points between the single phase and the three-phase regions. An enlargement plot (indicated by green circle) around point (70.8%, 86.77 bar) shows the nuances of phase behaviors close to the phase borders. Specically for the three-phase ash, two examples show the feasible region method with this oil. Shown in Fig. 7(a) is the performance of this method at 75% CO2 feeding and 82.4 bar. The inputs, such as the equilibrium ratios are from the two-phase stability test, are tabulated in Table 8, so this is the rst iteration in the three-phase ash. The temperature and pressure corresponds to a point close to the dew point line between the threephase and vaporliquid regions. It is seen that the feasible region method greatly reduces the search space to a limited region and the optimization algorithm converges quickly around six iterations. It is worth mentioning that some constraint lines of become parallel. This is because the mixture is close to the phase boundary. Fig. 7(b) shows results at 95% CO2 feeding at 76.4 bar. The equilibrium ratios are also from the two-phase stability test. This temperature and pressure is located at the upper bubble point line of the three-phase region where the liquidliquid region just terminates. It is easy to notice that the lines of constraints now are more parallel with each other, compared to the case in Fig. 7(a). This again demonstrates the challenges for the liquidliquid equilibrium

Fig. 4. (a) Phase diagram of Maljamar reservoir oil (11 components). Symbols V, L1 and L2 denote the vapor, CO2-lean and CO2-rich liquid phases. Also shown are the bubble and dew point lines. (b) Evolutions of three phases across the three-phase region at 80% CO2 feeding.

L. Qiu et al. / Fuel 115 (2014) 116

Fig. 5. (a) Phase diagram of Maljamar separator oil (7 components). Symbols V, L1 and L2 denote the vapor, CO2-lean and CO2-rich liquid phases. Also shown are the bubble and dew point lines. Meeting points, excepted for the specially noted, between the bubble and dew point lines are not critical points. (b) Evolutions of the three phases across the three-phase region at 90% CO2 feeding.

Table 7 Component properties and non-zero interaction parameters for Oil B. Species CO2 N2 C1 C2 C3 iC4 nC4 iC5 nC5 C6 PC1 PC2 PC3 PC4 PC5 PC6 Composition 0.0011 0.0048 0.1630 0.0403 0.0297 0.0036 0.0329 0.0158 0.0215 0.0332 0.181326 0.161389 0.125314 0.095409 0.057910 0.022752 Tc (K) 304.211 126.2 190.6 305.4 369.8 408.1 425.2 460.4 469.6 506.35 566.55 647.06 719.44 784.93 846.33 919.39 Pc (bar) 73.819 33.5 45.4 48.2 41.9 36.0 37.5 33.4 33.3 33.9 25.3 19.1 14.2 10.5 7.5 4.76

x ()
0.225 0.04 0.008 0.098 0.152 0.176 0.193 0.227 0.251 0.299 0.3884 0.5289 0.6911 0.8782 1.1009 1.4478

MW (g/mole) 44.01 28.01 16.04 30.07 44.1 58.12 58.12 72.15 72.15 84.0 112.8 161.2 223.2 304.4 417.5 636.8

ki;CO2 0.02 0.075 0.08 0.08 0.085 0.085 0.085 0.085 0.095 0.095 0.095 0.095 0.095 0.095 0.095

ki;N2 0.02 0.08 0.07 0.07 0.06 0.06 0.06 0.06 0.05 0.1 0.12 0.12 0.12 0.12 0.12

ki;C 1 0.075 0.08 0.003 0.01 0.018 0.018 0.025 0.026 0.036 0.049 0.073 0.098 0.124 0.149 0.181

Fig. 6. (a) Phase diagram of Oil B (16 components) at 307.6 K. Symbols V, L1 and L2 denote the vapor, CO2-lean and CO2-rich liquid phases. Also shown are the bubble and dew point lines. Meeting points, excepted for the specially noted, between the bubble and dew point lines are not critical points. (b) Evolutions of the three phases across the three-phase region at 80% CO2 feeding.

computation compared to the common vaporliquid equilibrium. Faced with this difculty, the feasible region is more stretched and narrow, and common ash calculation is likely to fail because of a poor initial guess and uncontrolled step size. However, the minimization method combined with a line search in the feasible region quickly found the solution in about ve iterations. 4.6. Test problem 6: Bob Slaughter Block oil and North Ward Estes oil The Bob Slaughter Block oil and North Ward Estes oil were studied by Khan et al. [29] as verication of their uid characterization

scheme. The generated parameters for these two West Texas oil mixtures, including the interaction parameter and the tuned critical constants of methane, are tabulated in Tables 9 and 10, respectively. Since these parameters were already tuned with experimental data, we studied the mixing of impure CO2 (95% CO2 and 5% C1) with the Bob Slaughter Block oil at 313.71 K (105 F). The pressure-composition phase diagram is shown in Fig. 8(a). There is a small three-phase region. A single liquid phase that is lean in CO2 is found to span a wide range of pressures and feeding compositions until reaching the saturation line. At low pressures, the vaporliquid phase dominates. Interestingly, the two liquid

10

L. Qiu et al. / Fuel 115 (2014) 116

Fig. 7. Performance of the feasibility region method in the three-phase ash on Oil B. (a) 75% CO2 feeding and 82.4 bar. (b) 95% CO2 feeding and 76.4 bar. Notice the change of scales from (a) to (b) and the close-to-parallel lines in (b). S limit and P limit stand for the SC and PC constraints in the last paragraph of Section 3.2, respectively.

Table 8 Inputs for the three-phase ash using the feasible region method for the two example cases for Oil B. Species Example case 1 Equilibrium ratio 1 CO2 N2 C1 C2 C3 iC4 nC4 iC5 nC5 C6 PC1 PC2 PC3 PC4 PC5 PC6 0.124763 10 0.311587 10+01 0.208726 10+01 0.854931 10+00 0.527766 10+00 0.380655 10+00 0.327115 10+00 0.232398 10+00 0.207699 10+00 0.135667 10+00 0.601892 1001 0.159753 1001 0.354171 1002 0.676409 1003 0.129447 1003 0.274828 1004
+01

Example case 2 Equilibrium ratio 2 0.123846 10 0.163727 10+01 0.138068 10+01 0.871704 10+00 0.718108 10+00 0.621078 10+00 0.584798 10+00 0.507828 10+00 0.485316 10+00 0.394919 10+00 0.277618 10+00 0.156128 10+00 0.799940 1001 0.374895 1001 0.166739 1001 0.657881 1002
+01

Equilibrium ratio 1 0.132727 10 0.185050 10+01 0.151460 10+01 0.829260 10+00 0.646657 10+00 0.534334 10+00 0.496377 10+00 0.413095 10+00 0.390260 10+00 0.300911 10+00 0.188930 10+00 0.890264 1001 0.367214 1001 0.132442 1001 0.425159 1002 0.100869 1002
+01

Equilibrium ratio 2 0.134109 10+01 0.337815 10+01 0.226808 10+01 0.846343 10+00 0.507804 10+00 0.356389 10+00 0.304867 10+00 0.211426 10+00 0.187989 10+00 0.119582 10+00 0.495008 1001 0.118259 1001 0.228396 1002 0.364778 1003 0.542617 1004 0.721841 1005

Table 9 Component properties and non-zero interaction parameters for Bob Slaughter Block oil. Species CO2 C1 PC1 PC2 Composition 0.0337 0.0861 0.6478 0.2324 Tc (K) 304.2 160.0 529.03 795.33 Pc (bar) 73.77 46.0 27.32 17.31

x ()
0.225 0.008 0.481 1.042

MW (g/mole) 44.01 16.04 98.45 354.2

ki;CO2 0.055 0.081 0.105

The impure CO2 mixing with the North Ward Estes oil is also presented. Its phase diagram is shown in Fig. 8(b). The overall plot looks similar to the Bob Slaughter Block oil plot. However, another bi-critical point appears on the saturation line with the departure of the single-phase mixture. Also, no single liquid phase L2 is observed at the proximity of pure CO2. TPn ash calculations along the left phase boundary do not nd a point where the two phases have identical intensive properties in the pressure range shown; therefore, a two-phase critical point does not exist. 4.7. Test problem 7: reservoir oil of Stenby et al. [30] and North Sea gas condensate The phase behaviors of another two mixtures that was studied by Nichita et al. [31] on developing a delumping procedure were also considered. The rst mixture is the reservoir oil used by Stenby et al. [30]. Its phase diagram is shown in Fig. 9(a) and its properties are tabulated in Table 11. The calculated phase envelope is similar to that of Stenby et al. [30]. It is noteworthy that there exists a small and slim three-phase region at low temperature and pressure conditions for this mixture, which is not reported by Stenby et al. [30]. The North Sea gas condensate mixture is composed of 161 components from experimental analysis. However, Leibovici et al. [32] reduced the original mixture to 27 components, including two pseudo-components for characterizing phase behaviors with available data. Their characterization has been adopted by Nichita et al. [31] and is also used here. Details

Table 10 Component properties and non-zero interaction parameters for North Ward Estes oil. Species CO2 C1 PC1 PC2 PC3 PC4 PC5 Composition 0.0077 0.2025 0.118 0.1484 0.2863 0.149 0.0881 Tc (K) 304.2 190.6 343.64 466.41 603.07 733.79 923.2 Pc (bar) 73.77 46.0 45.05 33.51 24.24 18.03 17.26

x ()
0.225 0.008 0.13 0.244 0.6 0.903 1.229

MW (g/mole) 44.01 16.04 38.4 72.82 135.82 257.75 479.95

ki;co2 0.12 0.12 0.12 0.09 0.09 0.09

phases merge to one single liquid phase at very high CO2 feedings and its borderline is very sharp. This extra liquid phase does not appear in the original pure CO2 mixing with the oil. Only one bicritical point exists at the three-phase boundary at the liquidliquid region side.

L. Qiu et al. / Fuel 115 (2014) 116

11

Fig. 8. (a) Phase diagram of the Bob Slaughter Block Oil (4 components). (b) Phase diagram of the North Ward Estes Oil (7 components). Symbols V, L1 and L2 denote the vapor, CO2-lean and CO2-rich liquid phases. Also shown are the bubble and dew point lines. Meeting points, excepted for the specially noted, between the bubble and dew point lines are not critical points.

Fig. 9. (a) Phase diagram of the 20 components reservoir oil. Note that there is a slim three-phase region at very low temperatures. (b) Results of TPn ash at 344.26 K and 177.65 bar. (c) Phase diagram of the 27 components North Sea gas condensate. (d) Results of TPn ash at 398.15 K and 220.0 bar.

on the component properties can be found in Table 12. The phase diagram of this mixture is plotted in Fig. 9(c). The experimental points are sampled from Nichita et al. [31]. There is only one saturation line and no critical point appears. More data is needed to clarify the low-pressure phase behavior, though the experimental high-pressure, low-temperature behavior is well captured. Two TPn ash calculations are made: one on the reservoir oil at 344.26 K and 177.65 bar and the other is for the gas condensate at 398.15 K and 220 bar, and the results are shown in Fig. 9(b) and (d). The calculation results are identical to those from Nichita et al. [31]. (Note that they swapped the mole fraction of the rst species (N2) and the second one (CO2) for the gas condensate mixture by mistake in their plot.).

4.8. Test problem 8: reservoir oil of Hoffman et al. [33] To construct the global phase diagram, the above calculations call the equilibrium solver for a range of conditions. If only the saturation point at a specied temperature or pressure is of interest, it can be calculated directly by using the tangent plane distance method. That is, (i) the tangent plane distance at the feeding composition must be parallel to that at another composition, and (ii) the distance between these tangent planes must be strictly zero. This method has been applied by Nghiem and Li [13] on the reservoir oil from Hoffman et al. [33]. A comparison with their calculated phase diagram is shown in Fig. 10. The properties data are tabulated in Table 13. The interaction parameter is calculated using

12

L. Qiu et al. / Fuel 115 (2014) 116

Table 11 Component properties and non-zero interaction parameters of Stenby reservoir oil. Species N2 CO2 C1 C2 C3 iC4 nC4 iC5 nC5 C6 C7 C11 C14 C16 C20 C23 C27 C32 C38 C48+ Composition 0.00403 0.01000 0.45396 0.04202 0.00887 0.00561 0.00518 0.00647 0.00294 0.01011 0.13117 0.07127 0.03847 0.06005 0.03352 0.03340 0.02870 0.02179 0.01892 0.01351 Tc (K) 126.20 304.20 190.60 305.40 369.80 408.10 425.20 460.40 469.60 507.40 567.16 633.70 671.11 710.30 752.38 790.47 835.97 887.86 956.19 1090.01 Pc (bar) 33.60 72.90 45.40 48.20 41.90 36.00 37.50 33.40 33.30 29.30 29.01 21.51 18.92 17.18 15.90 15.14 14.54 14.12 13.85 13.80

x ()
0.04 0.23 0.01 0.10 0.15 0.18 0.19 0.23 0.25 0.30 0.52 0.66 0.75 0.85 0.96 1.06 1.17 1.27 1.35 1.24

MW (g/mole) 28.01 44.01 16.04 30.07 44.09 58.12 58.12 72.15 72.15 86.18 111.89 163.22 198.71 239.54 289.22 337.54 399.66 475.59 581.43 797.11

ki;N2

ki;co2

ki;c1

0.02 0.06 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08 0.08

0.12 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15

0.0298 0.0350 0.0442 0.0488 0.0512 0.0544 0.0565 0.0586 0.0609 0.0627 0.0800

Table 12 Component properties and non-zero interaction parameters of North Sea gas condensate. Species N2 CO2 C1 C2 C3 iC4 nC4 iC5 nC5 CC5 PC6 CC6 AC6 PC7 CC7 AC7 PC8 CC8 AC8 PC9 CC9 AC9 PC10 CC10 AC10 CN-1 CN-2 Composition 3.2430 1003 1.9549 1002 7.6153 1001 7.7367 1002 3.6177 1002 5.6670 1003 1.3258 1002 4.4800 1003 6.1000 1003 5.8000 1004 6.0010 1003 4.1480 1003 1.7350 1003 4.2390 1003 5.0560 1003 3.0630 1003 3.2910 1003 2.5400 1003 2.5640 1003 2.6300 1003 1.6400 1003 1.2170 1003 2.4110 1003 4.2700 1004 1.0490 1003 2.6703 1002 3.3380 1003 Tc (K) 126.20 304.20 190.60 305.40 369.80 408.10 425.20 460.26 469.60 511.60 503.79 547.41 562.10 536.44 566.27 591.70 565.05 594.05 619.46 590.64 621.21 644.06 613.72 621.58 670.83 711.04 848.08 Pc (bar) 33.94 73.77 46.00 48.84 42.45 36.48 38.00 33.83 33.74 45.09 30.07 39.90 48.94 27.60 34.69 41.14 25.02 29.74 35.84 23.29 28.39 32.08 21.46 26.25 29.72 18.75 16.33

x ()
0.0400 0.2250 0.0115 0.0908 0.1454 0.1760 0.1928 0.2271 0.2273 0.1923 0.2860 0.2215 0.2100 0.3364 0.2451 0.2566 0.3816 0.2391 0.3228 0.4230 0.2998 0.3725 0.4646 0.4058 0.3642 0.8000 1.3000

MW (g/mole) 28.01 44.01 16.04 30.07 44.10 58.12 58.12 72.15 72.15 70.14 86.18 84.16 78.11 100.21 98.19 92.14 114.23 112.21 106.16 128.25 125.97 120.16 142.28 140.20 133.80 240.00 450.26

ki;N2 0.15 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12 0.12

ki;co2 0.15 0.12 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15

ki;c1

ki;c2

0.03 0.05

0.03

the relationship between the critical specic molar volume data of the pure component according to

" dij 1

2v ci v cj 1=6
1 =3 =3 v ci v1 cj

#h 23

The parameter h = 1.2126. The match with their calculations is quite good except that there is some over prediction around (200 K, 120 bar), which could be due to some minor differences in the model parameters, such as the coefcients used to determine the energy and volume parameters for the PR EOS. Again, it is remarked that the APPLE solver detects a three-phase region penetrating into the two-phase region in a similar slim fashion at low temperature and pressure conditions, which is not shown by

Nghiem and Li [13]. This type of behavior seems to exist for mixtures with very high and opening pressure limits at low temperatures. This appearance of three-phase can only be veried with experimental data and needs more investigations. Also shown in this gure is the stability limit line. The stability limit can be found whenever a saturation point is calculated, but the composition solution is trivial. Only some calculation points are shown since saturation point calculation is not the main topic here. The critical point on stability limit curve is also the point where the bubble and dew point lines meet. Since the critical point is between the cricondentherm and the cricondenbar, retrograde behavior exists for this mixture. The isothermal process at 680 K on the dew point line and the isobaric process at 200 bar are shown in Fig. 10(b) to illustrate this type of behavior. For the isothermal process, a maximum

L. Qiu et al. / Fuel 115 (2014) 116

13

Fig. 10. Comparison on the phase diagram of Hoffman reservoir oil (12 components) calculated with Nghiem and Li. Also shown are the stability limit and critical point. Note that there is a branch of the three-phase region penetrates into the two-phase region at very low temperatures.

Table 13 Component properties of Hoffman reservoir oil. Species C1 C2 C3 iC4 nC4 iC5 nC5 C6 PC1 PC2 PC3 PC4 Composition 0.52 0.0381 0.0237 0.0076 0.0096 0.0069 0.0051 0.0206 0.11972 0.12007 0.08687 0.04174 Tc (K) 190.6 305.4 369.8 408.1 425.2 460.4 469.6 507.5 597.0 696.1 757.2 846.3 Pc (bar) 46.00155 48.83865 42.45517 36.477 37.99688 33.84255 33.74123 32.8901 28.50272 21.49103 15.97895 11.86516

noticed that CH4 mainly exists in the vapor phase, while the liquid phase is occupied mostly by heavy molecules of decane.
x ()
0.0080 0.0098 0.1520 0.1760 0.1930 0.2270 0.2510 0.2637 0.3661 0.5237 0.6658 0.9301 MW (g/mole) 16.043 30.070 44.097 58.124 58.124 72.151 72.151 86.000 120.910 182.310 250.580 361.040

4.10. Test problem 10: binary hydrocarbon and nitrogen mixtures Finally, vaporliquid equilibrium calculations were performed for binary hydrocarbon species with nitrogen. This type of mixture has been widely used in calculations of droplet evaporation [37], non-reacting mixing in high-pressure n-heptane injections, and turbulent mixing modeling [38]. Two sets of mixtures are considered: n-pentane-N2 and n-heptane-N2. Fig. 12(a) and (c) shows comparisons between calculations and the experimental data that are adopted from Refs. [39] and [40]. Their suggested interaction parameters are used. All the other thermodynamic data are summarized in Table 15. The vaporliquid equilibrium behavior of this kind of mixture is rather difcult to model with a cubic equation of state due to the very different molecular structures. It is noticed from the plots that the match with experimental data is good at low pressures but fails at high pressures, especially close to the critical point. Since the only tunable parameter for the equation of state model is the binary interaction parameter, it is expected that using temperaturedependent interaction parameters could improve the accuracy [39,40]. A recent discussion on addressing the difculty of modeling mixtures with nitrogen using the group contribution method can be found in Ref. [41]. It is noteworthy to mention that the critical pressure increases while the critical temperature decreases. In addition, a linear relation between the critical point of a mixture and that of each pure species does not exist. The non-ideal behavior is further shown in Fig. 12(b) and (d) for these two mixtures. The pure species data for the latent heat is taken from the DIPPR database [42]. The latent heat, originally used for pure component is shown here to illustrate the non-ideal behaviors at high pressures. It is dened here as the difference between the partial molar enthalpy of the hydrocarbon in the vapor and liquid phases. In the pure species limit, only the hydrocarbon exists, so the latent heat of pure species is recovered. In both plots, an increase of pressure leads to a substantial decrease of latent heat, which is because the composition in the liquid and vapor phases becomes similar. For a specied pressure, latent heat decreases with temperature increase. This is because the mixture is approaching the critical state. At the thermodynamic critical point, the latent heat is strictly zero. 5. Final remarks

amount of liquid phase is determined at 125 bar corresponding to 26.78% liquid phase and it quickly drops at higher pressures. For the isobaric process, the maximum amount of the vapor phase, 24% occurs at around 475 K.

4.9. Test problem 9: hydrocarbon mixtures Calculations on vaporliquid equilibrium of three hydrocarbon mixtures were also performed. Here, we focused on binary mixtures with methane. The component properties can be found in Table 14. Fig. 11(a) and (b) shows the pressurecomposition diagram of binary CH4 and C2H4 and CH4 and C3H8 mixtures, respectively. Experimental data are from Refs. [34,35] and are shown with symbols. The open symbols are for the vapor phase while the closed ones are for the liquid phase. In all the computations, the equilibrium solver quickly nds the equilibrium composition of two phases. This shows that the APPLE solver is useful in predicting the vaporliquid equilibrium. Equilibrium calculation results for another binary mixture composed of CH4 and nC10H22 are shown in Fig. 11(c) and (d) at higher and lower temperatures, respectively. Experimental data are taken from Ref. [36]. It is

Table 14 Component properties and non-zero interaction parameters for binary CH4C2H4 and CH4nC10H22 mixture. Species C1 C2 C3 nC10 Tc (K) 190.564 282.34 369.83 617.7 Pc (bar) 45.99 50.41 42.48 21.1

x ()
0.0115 0.0862 0.1523 0.4923

MW (g/mole) 16 28 44 142

ki;c1 0.023 0.015 0.043

In internal combustion engine related research, there have been attempts to apply an equation of state to account for real gas

14

L. Qiu et al. / Fuel 115 (2014) 116

Fig. 11. Vaporliquid equilibrium of hydrocarbon mixtures. (a) Vaporliquid equilibrium of CH4C2H4 at three temperatures. (b) Vaporliquid equilibrium of CH4C3H8 at seven temperatures. Note that temperature is in Fahrenheit. (c) Vaporliquid equilibrium of CH4nC10H22 at three higher temperatures. (d)Vaporliquid equilibrium of CH4 nC10H22 at two lower temperatures.

Fig. 12. Phase equilibrium of hydrocarbons with nitrogen. (a) Vaporliquid equilibrium of nC5H12 and nitrogen. (b) Pressure effects on latent heat for nC5H12 and nitrogen. (c) Vaporliquid equilibrium of nC7H16 and nitrogen. (b) Pressure effects on latent heat for nC7H16 and nitrogen.

L. Qiu et al. / Fuel 115 (2014) 116 Table 15 Component properties and non-zero interaction parameters for binary nC5H12N2 and nC7H16N2 mixture. Species nC5 nC7 N2 Tc (K) 469.7 540.3 126.2 Pc (bar) 33.7 27.36 33.9

15

x ()
0.2515 0.3495 0.0377

MW (g/mole) 72.0 100.0 28.0

ki;N2 0.0657 0.0971

phase behaviors, including liquidliquid and two-phase to threephase transitions. Retrograde condensation, which is critical in the CO2 injection process used in enhanced oil recovery, is also well predicted by the solver. Finally, potential applications of the APPLE solver to spray jet simulation, such as droplet evaporation, were discussed. An recent computational uid dynamics simulation that couples the equilibrium solver on jet condensation also proves the robustness of the APPLE solver and its usability in more intensive calculations. Acknowledgements Financial support from the DOE Sandia National Laboratories is acknowledged. One of the authors, Qiu, L., would like to express his gratitude to Dr. Li (now with ExxonMobil) and Prof. Firoozabadi at the Reservoir Engineering Research Institute (RERI) for helpful discussions of phase equilibrium calculations. References
[1] Callen HB. Thermodynamics and an introduction to thermostatics. 2nd ed. Chichester (New York): Wiley; 1985. [2] Peng D-Y, Robinson DB. A new two-constant equation of state. Ind Eng Chem Fund 1976;15:5964. [3] Firoozabadi A. Thermodynamics of hydrocarbon reservoirs. New York: McGraw-Hill; 1999. [4] Baker LE, Pierce AC, Luks KD. Gibbs energy analysis of phase equilibria. SPE J 1982;22:73142. [5] Gibbs JW. A method of geometrical representation of the thermodynamic properties of substances by means of surfaces. Trans Conn Acad 1873;II:382404. [6] Robinson DB, Peng DY. The characterization of the heptanes and heavier fractions for the GPA PengRobinson programs, Gas processors association, Research report RR-28; 1978. [7] Jaubert J-N, Mutelet F. VLE predictions with the PengRobinson equation of state and temperature dependent kij calculated through a group contribution method. Fluid Phase Equilibr 2004;224:285304. [8] Jaubert JN, Privat R, Mutelet F. Predicting the phase equilibria of synthetic petroleum uids with the PPR78 approach. AIChE J 2010;56:322535. [9] Avaullee L, Trassy L, Neau E, Jaubert JN. Thermodynamic modeling for petroleum uids I. Equation of state and group contribution for the estimation of thermodynamic parameters of heavy hydrocarbons. Fluid Phase Equilibr 1997;139:15570. [10] Li ZD, Firoozabadi A. General strategy for stability testing and phase-split calculation in two and three phases. SPE J 2012;17:1096107. [11] Li ZD, Firoozabadi A. Initialization of phase fractions in Rachford-Rice equations for robust and efcient three-phase split calculation. Fluid Phase Equilibr 2012;332:217. [12] Michelsen ML. The isothermal ash problem. Part 1. Stability. Fluid Phase Equilibr 1982;9:119. [13] Nghiem LX, Li Y-K, Heidemann RA. Application of the tangent plane criterion to saturation pressure and temperature computations. Fluid Phase Equilibr 1985;21:3960. [14] Hoteit H, Firoozabadi A. Simple phase stability-testing algorithm in the reduction method. AIChE J 2006;52:290920. [15] Sun AC, Seider WD. Homotopy-continuation method for stability analysis in the global minimization of the Gibbs free-energy. Fluid Phase Equilibr 1995;103:21349. [16] Nichita DV, Gomez S, Luna E. Multiphase equilibria calculation by direct minimization of Gibbs free energy with a global optimization method. Comput Chem Eng 2002;26:170324. [17] Zhang H, Bonilla-Petriciolet RGP. A review on global optimization methods for phase equilibrium modeling and calculations. Open Thermodyn J 2011;5:7192. [18] Rachford Jr HH, Rice JD. Procedure for use of electronic digital computers in calculating ash vaporization hydrocarbon equilibrium. J Pet Technol 1952;4:2. [19] Michelsen ML. The isothermal ash problem. 2. Phase-split calculation. Fluid Phase Equilibr 1982;9:2140. [20] Okuno R, Johns R, Sepehrnoori K. A new algorithm for rachford-rice for multiphase compositional simulation. SPE J 2010:31325. [21] Mehra RK, Heidemann Robert A, Aziz Khalid. Computation of multiphase equilibrium for compositional simulation. Soc Petrol Eng J 1982;22:8. [22] Nghiem LX, Li YK. Computation of multiphase equilibrium phenomena with an equation of state. Fluid Phase Equilibr 1984;17:7795. [23] Haugen KB, Firoozabadi A, Sun L. Efcient and robust three-phase split computations. AIChE J 2011;57:255565. [24] Leibovici CF, Neoschil J. A Solution of rachford-rice equations for multiphase systems. Fluid Phase Equilibr 1995;112:21721. [25] Soave G. Letter to the editor. Fluid Phase Equilibr 1985;21:171.

behavior at high-pressure conditions in the combustion chamber. Researchers were mainly focused on the droplet evaporation process between the liquid phase and the ambient gas phase. This is because state-of-the-art computational uid dynamics (CFD) simulations in direct-injection (DI) diesel engines use an Eulerian gas and Lagrangian (liquid) droplet treatment. Liquid fuel is injected not as a continuous liquid phase, but as discrete droplet of similar size to the nozzle diameter. To match the boundary conditions in droplet evaporation, phase equilibrium is generally applied to provide mass transfer between the two phases [43,44]. However, additional simplications are added beyond the equilibrium assumption. For instance, in developing discrete evaporation model for multi-component evaporation, the phase equilibrium criterion reduces to the relationship between saturation pressure and ambient gas pressure for each component [45]. The interactions between components are fully neglected, so a Lewis mixture is actually assumed. In some cases, the iso-fugacity of each component assumption at the droplet surface is applied [46,47]. However, these calculations are limited to binary mixtures of n-heptane, n-pentane or n-nonane with nitrogen. It is doubtful if the ash calculation procedure is robust enough to handle mixtures of more species. Since liquidliquid equilibrium is not expected for such simple mixtures, the vaporliquid equilibrium always yields the nal state. Therefore, the equilibrium calculation does not need to consider potential phase separations. Nonetheless, as pointed out above, thermodynamically consistent phase equilibria should have much more stringent meaning and hence requires more careful treatments. Another benet of the present approach is that the phase stability test provides a good initial guess to accelerate the ash calculation, which is complicated by the interactions of many different components. The advantage is more obvious, especially when it is used in computational uid dynamic (CFD) simulations such as in simulations of supercritical gas jets [48].

6. Summary and conclusion In summary, an applied phase-related equilibrium solver, named APPLE, that is able to handle complex multicomponent and multiphase equilibrium has been developed and validated. To formulate a thermodynamically consistent equilibrium solver, rigorous analyses of phase equilibrium based on the equilibrium and stability criterion of fundamental thermodynamics were applied. Because the solver is based on minimization of the Gibbs free energy in the TPn ash calculations, it is theoretically correct and thermodynamically consistent. APPLE solver is composed of phase stability test and phase splitting when the stability test shows a negative tangent plane distance. Treated in this way, a new phase is only added when it reduces the Gibbs free energy. Phase stability test results then serve as the starting guess for the splitting calculations. Several efcient numerical methods, such as the BFGS algorithm for solving the tangent plane distance objective function for phase stability, Newton method and feasible region method for efcient multi-phase ash calculations are adopted. Verications are performed on various mixtures including reservoir oils and hydrocarbon mixtures that show intricate

16

L. Qiu et al. / Fuel 115 (2014) 116 [38] Bellan J. Supercritical (and subcritical) uid behavior and modeling: drops, streams, shear and mixing layers, jets and sprays. Prog Energy Combust 2000;26:32966. [39] Silva-Oliver G, Eliosa-Jimnez G, Garca-Snchez F, Avendao-Gmez JR. Highpressure vaporliquid equilibria in the nitrogenn-pentane system. Fluid Phase Equilibr 2006;250:3748. [40] Garcia-Sanchez F, Eliosa-Jimenez G, Silva-Oliver G, Godinez-Silva A. Highpressure (vapor plus liquid) equilibria in the (nitrogen plus n-heptane) system. J Chem Thermodyn 2007;39:893905. [41] Privat R, Jaubert J-N, Mutelet F. Addition of the nitrogen group to the PPR78 model (predictive 1978, peng robinson EOS with temperature-dependent kij calculated through a group contribution method). Ind Eng Chem Res 2008;47:203348. [42] DIPPR Design Institute for Physical Property Data. American Institute of Chemical Engineers; 1996. [43] Turns SR. An introduction to combustion: concepts and applications. New York: McGraw-Hill; 1996. [44] Kuo KK. Principles of combustion. New York: Wiley; 1986. [45] Ra Y, Reitz RD. A vaporization model for discrete multi-component fuel sprays. Int J Multiph Flow 2009;35:10117. [46] Zhang L, Kong S-C. High-pressure vaporization modeling of multi-component petroleumbiofuel mixtures under engine conditions. Combust Flame 2011;158:170517. [47] Trujillo MF, Torres DJ, ORourke PJ. High-pressure multicomponent liquid sprays: departure from ideal behaviour. Int J Engine Res 2004;5:22946. [48] Qiu L, Wang, Y, Wang H, Jiao, Q, Reitz, RD. Simulating condensation in a supercritical gas jet. In: ILASS Americas, 25th annual conference on liquid atomization and spray systems. Pittsburgh (PA); May 2013.

[26] Hua JZ, Brennecke JF, Stadtherr MA. Reliable computation of phase stability using interval analysis: cubic equation of state models. Comput Chem Eng 1998;22:120714. [27] Orr Jr FM, Yu AD, Lien CL. Phase behavior of CO2 and crude oil in lowtemperature reservoirs. Soc Petrol Eng J 1981;21:48092. [28] Shelton JL, Yarborough L. Multiple phase behavior in porous-media during Co2 or rich-gas ooding. J Petrol Technol 1977;29:11718. [29] Khan SA, Pope GA, Sepehrnoori K. Fluid characterization of three-phase CO2/oil mixtures. In: Paper SPE/DOE 24130 presented at the SPE/DOE eighth symposium on enhanced oil recovery. Tulsa (Oklahoma); April 2224, 1992. [30] Stenby EH, Christensen JR, Knudsen K, Leibovici CF. Application of a delumping procedure to compositional reservoir simulations. In: SPE annual technical conference and exhibition, Denver, CO.; 1996. [31] Nichita DV, Broseta D, Leibovici CF. Reservoir uid applications of a pseudocomponent delumping new analytical procedure. J Petrol Sci Eng 2007;59:5972. [32] Leibovici CF. A consistent procedure for the estimation of properties associated to lumped systems. Fluid Phase Equilibr 1993;87:18997. [33] Hoffman AE, Crump JS, Hocott CR. Equilibrium constants for a gas-condensate system. J Pet Technol 1953;5:110. [34] Miller RC, Kidnay AJ, Hiza MJ. Liquid + vapor equilibria in methane + ethene and in methane + ethane from 150.00 to 190.00 K. J Chem Thermodyn 1977;9:16778. [35] Wichterle I, Kobayashi R. Vapor-liquid equilibrium of methane-propane system at low temperatures and high pressures. J Chem Eng Data 1972;17:49. [36] Lin HM, Sebastian HM, Simnick JJ, Chao KC. Gasliquid equilibrium in binarymixtures of methane with N-decane, benzene, and toluene. J Chem Eng Data 1979;24:1469. [37] Zhu GS, Aggarwal SK. Transient supercritical droplet evaporation with emphasis on the effects of equation of state. Int J Heat Mass Trans 2000;43:115771.