Analysis and optimization of two-column cryogenic process for argon recovery from hydrogen-depleted ammonia purge gas
Min Zhao,Yanzhong Li,Shanxiu Sun
DOI: https://doi.org/10.1016/j.cherd.2010.11.014
IF: 4.119
2011-01-01
Chemical Engineering Research and Design
Abstract:Traditionally, high-purity argon recovery from air is considerably difficult owing to the boiling point of argon close to that of oxygen. Recently with the increasing demands for argon, another attractive source of ammonia purge gas has been paid more attention. In this paper with an objective of minimizing energy consumption per argon product, the two-column process for recovering argon from hydrogen-depleted ammonia purge gas is analyzed and optimized in detail on the ASPEN PLUS platform. Firstly, the model of two-column process is set up using the standard unit operation blocks and PENG-ROB property method of ASPEN PLUS, in which validation of PENG-ROB property method is carried out by comparison with a total 623 experimental data from three aspects: vapor–liquid equilibrium, liquid phase density, and enthalpy. It is followed by the thermodynamic and simulation and sensitivity analysis, which on the one hand can reduce the number of decision variables related to optimization problem, and on the other hand can obtain reasonable parameter specification, variables initial values and ranges, thus effectively ensuring the later optimization algorithm converges quickly and accurately. Finally the built-in sequential quadratic programming (SQP) solver of ASPEN PLUS is adopted to solve the minimum energy consumption optimization problem of two-column process. On the processor of 2.66 GHz Intel(R) Core (TM)2 Duo CPU with 4 GB RAM, the whole optimization only takes CPU times 10 s or so to accomplish. The optimal results show that thermal state of feed to demethanizer is a very efficient and valuable means to reduce system energy consumption which at T C05 = 103 K is only 87.4% of that at T C05 = 109 K where T C05 is the temperature of feed to demethanizer directly reflecting its thermal state. The condensing pressure of hydrogen-depleted ammonia purge gas also plays a vital role in reducing system energy consumption which is less at higher condensing pressure, whereas it almost has no influence on the yield and purity of argon recovery. The optimal operating pressure of flash separator used to remove the residual hydrogen in the feed hydrogen-depleted ammonia purge gas is 0.4–0.6 MPa (A); the most economical reflux ratio of argon distillation column is 1.15, and that of demethanizer varies from 0.33 to 0.45 depending on thermal state of feed to demethanizer. Abbreviations API American Petroleum Institute ASPEN PLUS process simulation software PENG-ROB built-in PENG-ROB property method of ASPEN PLUS SRK Soave–Redlich–Kwong equation of state VLE vapor–liquid equilibrium Keywords Hydrogen-depleted ammonia purge gas Argon recovery Cryogenic technology Analysis Optimization Nomenclature a , b , a i , b i constants relating to gas species in Eq. (5) , where subscript i represents component i in a mixture. C p specific heat (J mol −1 K −1 ) F ˙ A R flow rate of argon product in Eqs. (7) and (8) (N m 3 h −1 ) F ˙ D N 01 flow rate of low-pressure nitrogen stream (DN01) (N m 3 h −1 ) F ˙ M N 01 flow rate of refrigerant nitrogen required by middle-pressure compressor (COMP2) (N m 3 h −1 ) F ˙ N 00 flow rate of refrigerant nitrogen adding from outside (N m 3 h −1 ) F ˙ N 04 flow rate of returning cold nitrogen stream (N04) (N m 3 h −1 ) Δ h p isobaric enthalpy difference (kJ kg −1 ) Δ h T Joule–Thomson effect (J mol −1 K −1 ) k ij binary interaction coefficient between component i and j in a mixture of Eqs. (5) and (6) m ˙ i mass flow rate of stream i in Fig. 1 (kg s −1 ) ( m ˙ h ) i enthalpy of stream i in Fig. 1 (kW) p pressure in Eqs. (5) and (6) (MPa) p c , p ci critical pressure in Eq. (6) , where subscript i and c represent component i in a mixture and critical state, respectively (MPa) P i pressure of stream i in Fig. 1 (MPa) P in inlet pressure in Table 1 (MPa) P out outlet pressure in Table 1 (MPa) P TCH 4 Bottom bottom pressure of demethanizer (MPa) P TARBottom bottom pressure of argon column (MPa) Q ˙ C O N D E N 1 heat duty of demethanizer condenser (kW) Q ˙ R E B O I L 2 heat duty of argon column re-boiler (kW) Q ˙ C O M P 1 cooling duty of compressor COMP1 (kW) Q ˙ C O M P 2 cooling duty of compressor COMP2 (kW) Q ˙ l o s s heat loss of the whole process (kW) R gas constant in standard Peng–Robinson cubic equation of state Eq. (5) and Rackette model Eq. (6) (J mol −1 K −1 ) R TCH 4 , R TAR reflux ratios of demethanizer and argon column s 1 systematic deviation s 2 average absolute deviation s 3 root-mean-square deviation T , T c , T ci , T cj , T r temperature in Eq. (5) or (6) , where subscript c , r and i , j represent critical, reduced state, and component i and j in a mixture, respectively (K) T in inlet temperature in Table 1 (K) T out outlet temperature in Table 1 (K) T p C O N D E N 1 s saturation temperature of nitrogen at the pressure of demethanizer condenser (CONDEN1) (K) T p C O N D E N 2 s saturation temperature of nitrogen at the pressure of argon column condenser (CONDEN2) (K) T i temperature of stream i in Fig. 1 (K) Δ T temperature difference (K) V , V m , V m l , V cm , V c , V ci specific volume in Eqs. (5) and (6) , where superscript l represents liquid phase, and subscript i , m , c and represent component i , mixture and critical state, respectively (m 3 kg −1 ) W ˙ C O M P 1 input power of compressor COMP1 (kW) W ˙ C O M P 2 input power of compressor COMP2 (kW) W ˙ E X P isentropic output power of expander EXP (kW) W ˙ t o t a l total power of the whole process in Eq. (7) and (8) (kW) x , x i , x j mole fraction of component in Eqs. (5) and (6) , where subscript i and j represent component i and j in a mixture X vector of decision variable in the objective function of Eq. (7) Z i *, RA , Z m RA parameter in Rackette model Eq. (6) whose default value is the critical compressibility factor Z c Greek letters η m mechanical efficiency of expander η S isentropic efficiency of expander 1 Introduction For many decades, high-purity argon has been obtained from air by using cryogenic distillation approach. However, this conventional separating process is considerably difficult owing to the boiling point of argon being very close to that of oxygen (the difference between them is only 2.9 K). Recently people began to pay attention to other replaceable sources to meet increasing demands for argon. One such attractive source is the purge gas coming from ammonia synthesis plant ( Meisler, 1971; Zhang, 1980; Springmann, 1982; Zhang and Li, 1988; Thorogood et al., 1991; Jian, 1993; Kerry, 2006 ). In ammonia synthesis process, a portion of circulating gas must be purged continuously to maintain inert gas (argon, methane, etc.) concentration below a specified level so that ammonia synthesis efficiency does not decrease a lot. This purged gas is typically composed of 60–63% hydrogen, 20–21% nitrogen, 3–4% argon, 7–15% methane and 2–3% ammonia, which usually is treated in a membrane separator first and divided into two streams: one is permeating gas rich in hydrogen returning to ammonia synthesis system to boost ammonia production; the other is non-permeating gas called hydrogen-depleted ammonia purge gas used to produce argon as a result of containing 6–12% argon mixed with hydrogen, methane and nitrogen (each of their boiling points is far from that of argon) in it. Presently there are two types of technology for recovering argon from this hydrogen-depleted ammonia purge gas: one is cryogenic distillation technology, and the other is pressure swing adsorption (PSA) technology. The former was adopted originally and still has been used widely due to the advantages of high efficiency and high purity, and the latter was developed with the improvement of PSA technology gradually. For example, Krishnamurthy et al. once proposed a new process named HARP which combined PSA technology with cryogenic technology ( Krishnamurthy et al., 1987a, 1987b; Maclean et al., 1987a, 1987b ), and Cho et al. also presented a two-stage PSA method for recovering argon from ammonia purge gas ( Cho et al., 1998 ). However, whatever type they are, these PSA processes have one common shortcoming when compared with cryogenic distillation, i.e., the relatively lower argon purity or yield. The plant with cryogenic distillation approach, made by Costain Oil, Gas & Process Ltd for the AMC plant (at Verdigris in Oklahoma in 1989), was reported to achieve more than 97% recovery of 99.999% pure argon and 94% recovery of 92% purity hydrogen ( Healy, 1992 ), which is almost impossible for those plants using PSA technology. Unfortunately the current cryogenic process for argon recovery from hydrogen-depleted ammonia purge gas is not perfect although great progress has been made whether in process simplification, cost reduction and argon purity or in argon yields since the first plant for recovering both hydrogen and argon from ammonia purge gas was realized in East Germany in the late 1950s ( Sun et al., 1984; Xu, 1984, 1985; Jian, 1988; Tang and Chen, 2002 ). The refrigeration cycle used in the present cryogenic process is a type of heat pump cycle, which is vulnerable to the effects of inlet composition change of hydrogen-depleted ammonia purge gas although having high thermal efficiency. If hydrogen content in hydrogen-depleted ammonia purges gas increases, the whole system cold will become strained, but if methane content in it increases, the whole system cold will be well off. Furthermore, although for many years many authors have been working to reduce energy consumption of the whole system and several innovative cryogenic processes have been proposed, there are some unreasonable parts existing in these ones. For example, we once simulate a reformative process ( Masayuki and Ryuichi, 1985 ) which omitted middle-pressure nitrogen compressor (about 3.0 MPa (A)) by utilizing feed gas in place of middle-pressure nitrogen to heat liquid methane in the re-boiler of demethanizer for rising steam required by demethanizer. It is found that if so, additional liquid nitrogen must be supplied to make up for what produced by middle-pressure nitrogen refrigeration, or else system cold will be insufficient. In addition, the introduction of liquid methane from the bottom of demethanizer into the main heat exchanger (E1) will certainly lead to the enlargement of heat transfer temperature difference Δ T between hot and cold streams in it. In a word, requiring additional liquid nitrogen or other cold sources is bad point of this process, which inevitably leads to the increase of energy consumption and the complexity of process. In 1996, Asami et al. presented a type of coupling column which connects demethanizer with argon column similar to the combination of high-pressure column and low-pressure column in air separation plant ( Asami et al., 1996 ). He said this type of coupling column process can reduce approximately 7% of energy consumption compared with traditional two-column process. However, according to our analysis in Section 3.1 of this paper, this kind of process not only increases the unnecessary energy consumptions but also makes the argon column work in an unexpected vacuum condition. In this paper the model of a two-column cryogenic process for argon recovery from hydrogen-depleted ammonia purge gas is firstly set up using ASPEN PLUS, a commercial simulation software widely used in the chemical processes owing to its convenient interfaces and powerful functions ( Venkataraman et al., 1990; Ongiro et al., 1996; Yan and Rudolph, 2000; Bao et al., 2002; Smejkal and Soos, 2002; Sarvar-Amini et al., 2007; Hao et al., 2008; Doherty et al., 2009; Paviet et al., 2009; Ye et al., 2009 ). For such two-column process model, it is convenient to establish using the standard unit operation blocks (heaters, coolers, heat exchangers, compressors, etc.) and built-in PENG-ROB property method of ASPEN PLUS. Considering property calculation method selection is crucial for correctly simulating and optimizing process, so in this work a total of 623 experimental data on vapor–liquid equilibrium, enthalpy, and liquid phase density are cited to verify the built-in PENG-ROB property method of ASPEN PLUS. The results show that PENG-ROB property method can obtain satisfactory prediction for thermal and caloric properties of fluids involved in two-column process. This is followed by detailed analysis and optimization aiming at minimizing energy consumption per argon product because it is the main factor influencing the entire system economic performance. Generally the whole process's optimization problem is quite complicated and difficult to formulate and converge. So in this work based on the above model on ASPEN PLUS platform, the thermodynamic and simulation and sensitivity analysis is performed firstly to obtain good understanding of two-column cryogenic process (e.g., what is appropriate operating pressure of two distillation columns? what is economical reflux ratio of two distillation columns under ensuring the yield and purity of argon recovery? What are key decision variables affecting system energy consumption among such multiple process parameters? How is the interaction between them?), and then the built-in sequential quadratic programming (SQP) solver of ASPEN PLUS is used to solve two-column process optimization problem. By this optimization procedure, the number of decision variables are reduced, and it is more important that through simulation and sensitivity analysis, the reasonable parameter specification, variables initial values and ranges can be roughly obtained in advance, thus effectively ensuring SQP method as quasi-Newton nonlinear programming algorithm converges quickly and accurately. On the processor of 2.66 GHz Intel(R) Core (TM)2 Duo CPU with 4 GB RAM, it only takes the CPU time about 10 s to accomplish. Its optimal results show excellent agreement with those in the thermodynamic and simulation and sensitivity analysis. 2 Modeling 2.1 ASPEN PLUS flowsheet ASPEN PLUS 11.1 is selected for modeling the steady state of two-column process. Fig. 1 presents the corresponding ASPEN PLUS flowsheet, in which heat exchangers are simulated by HEATX modules, distillation columns by RADFRAC modules, re-boiler and condenser of distillation columns by HEATER modules, compressors by MCOMPR modules, expanders by COMPR modules, flash separators by FLASH2 modules, and throttling valves by VALVE modules. The whole process consists of two interactive subsystems: cryogenic distillation subsystem and refrigeration subsystem. The cryogenic distillation subsystem is actually a combination of three separating processes that are in sequence separation of hydrogen, methane and nitrogen from argon. High pressure hydrogen-depleted ammonia purge gas (C01) consisting of nitrogen, methane, argon and a few hydrogen is firstly cooled and liquefied totally in the main heat exchanger (E1), then through throttling valve (V1) is decompressed to the operating pressure of flash separator (FLASH1) and enters it with the state of vapor–liquid mixture, which results in almost all the residual hydrogen dissolved in the feed hydrogen-depleted ammonia purge gas entering vapor phase stream (H01). By exiting the vapor stream (H01)) from system, hydrogen is successfully removed from ammonia purge gas. It is noted that operating pressure of flash separator (FLASH1) should be chosen appropriately so that the accompanying argon loss is small as far as possible (see Section 3.1 in this paper). After removing hydrogen, liquid fraction (C04) from flash separator (FLASH1) composed of methane, nitrogen and argon enters the main heat exchanger (E1) to be heated to the temperature T C05 , then goes through throttling valve (V2) and enters the first distillation column – demethanizer (TCH4) – in which methane is separated and liquid methane (CH41, 99% methane) is obtained at the bottom of column. Next, the distillate stream (CAR) from column top comprising the remaining two components (i.e., argon and nitrogen) enters the second distillation column – argon column (TAR) – to further separate nitrogen from argon. In argon column (TAR), liquid argon (LAR, 99.999% argon) is obtained at the bottom, and distillate stream (WN01) from the top containing more than 95% nitrogen is used as the medium of refrigeration subsystem. Considering there is a small amount of impurities (e.g., hydrogen) contained in it, if necessary, this part of nitrogen needs pre-treating to get rid of impurities before used as refrigerant so that impurities do not accumulated in the refrigeration subsystem. The refrigeration subsystem of this process is a heat pump nitrogen cycle. Besides the above-mentioned stream (WN01), the remaining nitrogen (N00) needs adding from outside to satisfy system cold requirements. These two parts of nitrogen with atmospheric temperature and pressure are firstly pressurized in compressor (COMP1) to about 0.6 MPa (A) (called low-pressure nitrogen), and then split into three sub-streams (PN0, MN01 and DN01): one (PN0) expands to about atmospheric pressure in turbine (EXP) for obtaining cold required by high-temperature zones of the main heat exchanger (E1), in which it rises to atmospheric temperature and goes back to compressor (COMP1) inlet; another (MN01) continues being pressurized to about 2.7 MPa (A) in compressor (COMP2) (called middle-pressure nitrogen), and then cooled to temperature T MN03 in the main heat exchanger (E1) and enters the re-boiler of demethanizer (REBOIL1), where it is totally condensed into liquid (MN04). Then liquid stream MN04 is decompressed through throttling valve (V3) to about 0.6 MPa (A) into flash separator (FLASH2) in which vapor fraction (VN01) and liquid fraction (LN01) are separated from each other; a third (DN01) being pre-cooled to saturation temperature (about 95.5 K) in the heat exchanger (E3) is mixed with the above vapor fraction (VN01) and then enters the re-boiler of argon column (REBOIL2), where they are completely condensed into liquid nitrogen (LN02). Together with the above liquid fraction (LN01), it provides cold for both distillation columns’ condensers (CONDEN1 and CONDEN2), where they absorb latent heat and vaporize into steam (VN02 and VN03). After being recovered cold in heat exchangers (E2, E3, and E1), they return to compressor (COMP1) inlet finishing one refrigeration cycle. 2.2 Overall mass and energy balance According to the above process description, it is clear that for the cryogenic distillation subsystem, the mass flow rate of feed hydrogen-depleted ammonia purge gas C01, m ˙ C 01 , should be equal to the sum of each separated part's mass flow rate, i.e., (1) m ˙ C 01 = m ˙ L A R + m ˙ C H 4 + m ˙ H 02 + m ˙ W N 03 where m ˙ L A R is the mass flow rate of liquid argon product (LAR, 99.999% argon), m ˙ C H 4 is the mass flow rate of methane byproduct (CH 4 , 99% methane), m ˙ H 02 is the mass flow rate of hydrogen-rich stream (H02) exiting from system, and m ˙ W N 03 is the mass flow rate of remaining nitrogen stream (WN03, more than 95% nitrogen). For refrigeration subsystem, the total refrigerant nitrogen, m ˙ N 01 , is equal to (2) m ˙ N 01 = m ˙ W N 03 + m ˙ N 00 = m ˙ P N 2 + m ˙ N 05 + m ˙ N 07 where m ˙ W N 03 is the mass flow rate of nitrogen coming from cryogenic distillation subsystem, m ˙ N 00 is that of nitrogen added from outside, and m ˙ P N 2 , m ˙ N 05 and m ˙ N 07 are that of three nitrogen branches, respectively. Combining Eqs. (1) and (2) to get rid of m ˙ W N 03 , the overall mass balance of process can be written as, (3) m ˙ C 01 + m ˙ N 00 = m ˙ L A R + m ˙ C H 4 + m ˙ H 02 + m ˙ P N 2 + m ˙ N 05 + m ˙ N 07 Correspondingly, the overall energy balance of process is written as, (4) ( m ˙ h ) C 01 + ( m ˙ h ) N 00 + W ˙ C O M P 1 + W ˙ C O M P 2 + Q ˙ l o s s = ( m ˙ h ) L A R + ( m ˙ h ) C H 4 + ( m ˙ h ) H 02 + ( m ˙ h ) P N 2 + ( m ˙ h ) N 05 + ( m ˙ h ) N 07 + Q ˙ C O M P 1 + Q ˙ C O M P 2 + W ˙ E X P η s η m where ( m ˙ h ) i is the enthalpy of stream i , W ˙ C O M P 1 and W ˙ C O M P 2 are input power of compressor COMP1 and COMP2, respectively, Q ˙ C O M P 1 and Q ˙ C O M P 2 are cooling duty of isothermal compressor COMP1 and COMP2, respectively, W ˙ E X P is the isentropic output power of expander EXP, η S and η m are isentropic and mechanical efficiency of expander EXP, respectively, and Q ˙ l o s s is heat loss of the whole process. 2.3 Property method – the built-in PENG-ROB property method in ASPEN PLUS The accurate calculation of fluid's thermodynamic properties is of indispensable importance for the basic engineering and performance of technical process. In this process model, the built-in PENG-ROB property method of ASPEN PLUS is used to calculate fluids’ thermodynamic properties. To speak concretely, it uses the standard Peng–Robinson cubic equation of state ( Peng and Robinson, 1976 ) to calculate all the thermodynamic properties except for the liquid phase volume that is predicted by the Rackett model ( Rackett, 1970; Spencer and Danner, 1972 ), which is because the cubic equation of state usually has large errors in calculating liquid phase densities. 2.3.1 Standard Peng–Robinson cubic equation of state (5) p = R T V m − b − a V m ( V m + b ) + b ( V m − b ) a = ∑ i ∑ j x i x j a i a j 0.5 1 − k i j b = ∑ i x i b i where p is pressure, R is gas constant, T is temperature, V is specific volume, a and b are constants relating to gas species, x is mole fraction of component, subscript i , j and m represent component i , j and mixture, respectively, and k ij is binary interaction coefficient between component i and j in a mixture, which had better be determined by regressing experimental equilibrium data (for example, VLE data) and whose default value is zero. In ASPEN PLUS databases, there are a lot of such k ij automatically used by PENG-ROB property method. 2.3.2 Rackett model (6) V m l = R T c Z m R A 1 + ( 1 − T r ) 2 / 7 p c T c = ∑ i ∑ j x i x j V c i V c j T c i T c j 0.5 1 − k i j / V c m 2 T c p c = ∑ i x i T c i p c i Z m R A = ∑ i x i Z i * , R A V c m = ∑ i x i V c i T r = T T c where V is specific volume, p is pressure, T is temperature, R is gas constant, x is mole fraction of component, Z i *, RA is a parameter whose default value is the critical compressibility factor Z ci , k ij is binary parameter whose default value is 1 − 8 ( V c i V c j ) 0.5 ( V c i 1 / 3 + V c j 1 / 3 ) , superscript l represents liquid phase, and subscript i , j , m and c represent component i , j , mixture, and critical state, respectively. 2.3.3 Validation of PENG-ROB property method In the two-column process, fluid involves a mixture of nitrogen, argon, methane, and hydrogen, whose temperature is from about 80 K to 313 K and pressure is from about 0.1013 MPa to 10.0 MPa. For cryogenic distillation subsystem, the main property calculation is vapor–liquid equilibrium (VLE) taking place in two distillation columns, that is, in demethanizer, it is the VLE of ternary mixture methane-nitrogen-argon from 85 K to 120 K at ∼0.200 MPa, while in argon column, it is that of binary mixture nitrogen–argon from 80 K to 95 K at ∼0.195 MPa. In addition, mixture enthalpy calculation is also important for mass and energy balance of process, especially for those in the main heat exchanger E1 where the temperature of fluids is changed from 100 K to 313 K at about 10.0 MPa, covering gaseous, supercritical, vapor, vapor–liquid and liquid five kinds of phases. In contrast to cryogenic distillation subsystem, refrigeration subsystem is conventional in fluids’ property calculation, whose refrigerant is pure nitrogen with pressure from about 0.1013 MPa to 3.0 MPa (A), including vapor, vapor–liquid and liquid three phases. Below we will verify the accuracy of PENG-ROB property method in predicting the above mixtures’ thermodynamic properties in such different regions by comparison with experimental data from three aspects: VLE, liquid phase density, and enthalpy. 2.3.3.1 Vapor–liquid equilibrium (VLE) Figs. 2–3 show the deviation of experimental saturated vapor pressures and vapor phase composition from the values calculated by PENG-ROB property method in ASPEN PLUS, in which Fig. 2 is that of binary mixture nitrogen–argon at 122.89 K measured by Jin et al. ( Jin et al., 1993 ) and Fig. 3 is that of ternary mixture nitrogen-argon-methane at 90.67 K, 122.89 K and 123.4 K measured by Sprow et al. ( Sprow and Prausnitz, 1966; Gravelle and Lu, 1973; Jin et al., 1993 ). It can be seen that PENG-ROB property method represents all the experimental vapor pressures with deviations of less than ±5%. A deviation of less than −2% is observed for those measurements of Jin et al. As for vapor mole composition, it can be seen that almost all the experimental data are represented by PENG-ROB property method well within ±(1–2)%, only a few data reaching ±4%. 2.3.3.2 Liquid phase density As we know, the cubic equation of state is generally poor in describing liquid phase density. In ASPEN PLUS's PENG-ROB property method, Rackett model is used to calculate real components’ liquid molar volumes. Fig. 4 compares them with experimental liquid phase density at about 90–140 K for the binary/ternary mixtures composed of nitrogen, argon, and methane. These experimental data ( Massengi and Miller, 1973; Hiza et al., 1977 ) are either saturated liquid densities or those at pressures just above the phase boundary. Clearly PENG-ROB property method represents them with deviation of less than ±2%, which is much superior to those conventional cubic equations of state. 2.3.3.3 Enthalpy in gas, supercritical, vapor, liquid, and vapor–liquid equilibrium regions (1) Liquid phase enthalpy Fig. 5 shows the percentage deviation of experimental liquid phase enthalpy at about 85–145 K from the values calculated by PENG-ROB property method for the binary mixtures composed of nitrogen, argon, or methane. Similar to the above experimental liquid density, these enthalpy measurements ( Lambert and Simon, 1962; Lewis et al., 1975; Lewis and Staveley, 1975; McClure et al., 1976; Mosedale and Wormald, 1977 ) are either saturated liquid enthalpy or those at pressures just above the phase boundary. It can be seen that the PENG-ROB property method represents them with deviations of less than ±2%. In fact most of the measurements are actually within ±0.5%. (2) Vapor, liquid, supercritical and vapor–liquid equilibrium enthalpy Fig. 6 shows the percentage deviation of experimental isobaric enthalpy from the values calculated by PENG-ROB property method for the binary mixtures composed of nitrogen, argon, methane, or hydrogen. These experimental data cover vapor, supercritical, vapor–liquid, and liquid regions with a temperature from 123 K to 180 K and pressure from 1.02 MPa to 10.1 MPa, which are measured by Lewis et al. in 1977 ( Lewis et al., 1977a, 1977b ). The PENG-ROB property method represents them with deviations of ±(3–6)%, being in good agreement with the uncertainty of experimental data. (3) Gas enthalpy For the gaseous enthalpy, Grini et al. ( Grini et al., 1998 ) once issued experimental isobaric enthalpy difference for binary mixture nitrogen-methane at 170–260 K and 0.9–7.0 MPa, and simultaneously compared them with those values of SRK equation of state ( Soave, 1972 ) in which the involved ideal gas enthalpy is calculated with API correlation ( API, 1983 ). Their comparison shows that SRK equation can represent the measurements with reasonable accuracy except in the critical region. To verify whether the PENG-ROB property method can obtain the same results as SRK equation, the corresponding isobaric enthalpy differences for the same inlet and outlet temperatures and pressures using PENG-ROB property method are calculated and listed in Table 1 . In order to facilitate comparison, the experimental data and SRK values by Grini et al. are also listed in Table 1 . It can be seen that although the deviations of PENG-ROB property method (i.e., systematic deviation s 1 = 2.96, average absolute deviation s 2 = 4.09 and the root-mean-square deviation s 3 = 5.77) are a little larger than those of SRK equation (i.e., systematic deviation s 1 = 1.01, average absolute deviation s 2 = 1.40 and the root-mean-square deviation s 3 = 2.05), they are still within acceptable ranges and the results is satisfactory. In short, all the above comparisons demonstrate that the built-in PENG-ROB property method of ASPEN PLUS can obtain satisfactory results in predicting thermal and caloric properties of fluids involved in the two-column process for argon recovery from hydrogen-depleted ammonia purge gas. 2.4 Analysis and optimization 2.4.1 Parameter specification and assumption Before analyze and optimization on the above process model, some process parameters are need specifying beforehand, which includes inlet conditions of feed hydrogen-depleted ammonia purge gas, the efficiency of compressor and expander, the minimum temperature difference of heat exchanger, etc. These parameters are subject to current industry standards and thereby are not as optimized variables to consider. Below lists these parameters and some assumptions used in the analysis and optimization: (a) Composition of feed hydrogen-depleted ammonia purge gas is: 8% hydrogen, 8.96% argon, 35.24% methane and 47.8% nitrogen of mole fraction. The total flow rate is 1000 N m 3 /h; (b) Inlet temperature of hydrogen-depleted ammonia purge gas is 303 K; (c) Inlet pressure of hydrogen-depleted ammonia purge gas is 7.8 MPa (A); (d) Outlet pressure of multi-stage compressor COMP1 and COMP2 are set 0.56 MPa (A) and 2.7 MPa (A), respectively, with outlet temperature of each stage being 313 K; (e) Isentropic and mechanical efficiency of expander are 0.8, 0.98, respectively; (f) The number of demethanizer and argon column trays is 18 and 50, respectively; (g) Pressure drop in the heat exchangers is zero, and their minimal temperature difference is set ∼2 K; (h) Heat loss of the whole system is zero. 2.4.2 Objective Whether a process system is good or not depends largely on its economic performance. As the main indicator reflecting system economical performance, energy consumption is often used as process optimization goals. In this work, we focus on minimizing the energy consumption per argon product, that is, (7) min X . f ( X ) = W ˙ t o t a l F ˙ A R where W ˙ t o t a l is the total energy consumption of process, F ˙ A R is the flow rate of argon product, and X is a vector representing a series of decision variables that affect system energy consumption. 2.4.3 Methodology In the two-column process, there are many parameters influencing system energy consumption, for instance, operating pressure of two distillation columns P TCH4Bottom , P TARBottom , distillation column reflux ratio R TCH4 , R TAR , condensing pressure of hydrogen-depleted ammonia purge gas P C01 , operating pressure of separator FLASH1 P C03 , the thermal state of feed to demethanizer T C05 , etc. In these parameters, some can be roughly determined through process thermodynamic analysis, some can be obtained according to process simulation and sensitivity analysis, and some need to be determined through simultaneously optimization. In order to simplify the optimization problem, more importantly, to ensure optimization algorithm converge fast and reliable, the thermodynamic and simulation and sensitivity analysis is performed first and then the built-in SQP optimization solver of ASPEN PLUS is adopted to solve the minimum energy consumption optimization problem of two-column process, whose decision variables, constraints, and objective function are illustrated as follows: (1) Decision variables Benefit from the thermodynamic and simulation and sensitivity analysis, the optimal values of process parameter P TCH4Bottom , P TARBottom , P C01 , P C03 can be determined in advance and therefore the number of decision variables related to optimization problem can be reduced, which now includes the following parameters: thermal state of feed to demethanizer T C05 , outlet temperature of middle-pressure nitrogen from the main heat exchanger (E1) T MN03 , reflux ratio of two distillation columns R TCH4 and R TAR , flow rate of refrigerant nitrogen adding from outside F ˙ N 00 , flow rate of refrigerant nitrogen required by middle-pressure compressor (COMP2) F ˙ M N 01 , flow rate of low-pressure nitrogen stream (DN01) F ˙ D N 01 , and flow rate of returning cold nitrogen stream (N04) F ˙ N 04 . (2) Constraints First the required argon purity and recovery rate should be ensured, followed by the mass and energy balance should be satisfied by the ASPEN PLUS process model described in Section 2 . In heat exchanger E1, E2, and E3, heat transfer temperature difference ΔT should be larger than the specified minimum temperature difference of ∼2 K; the temperature of streams C03 and C06 should be higher than the triple point of methane (90.68 K) so that their methane component could not freeze. In addition, in order to reduce the required amount of liquid nitrogen and save compressor's power, refrigerant nitrogen in both distillation columns’ condensers should be always in a state of vapor–liquid phase change, i.e., the temperature of stream VN02 and VN03 should be kept at saturation temperature at their respective condenser pressure. Finally the effective varying range of temperature T C05 is given. (3) Objective function Combined with the above (1) and (2), the complete optimization problem with constraints can be described as follows, (8) min . T C 05 , T M N 03 , R T C H 4 , R T A R , F ˙ N 00 , F ˙ M N 01 , F ˙ D N 01 , F ˙ N 04 f ( T C 05 , T M N 03 , R T C H 4 , R T A R , F ˙ N 00 , F ˙ M N 01 , F ˙ D N 01 , F ˙ N 04 ) = W ˙ t o t a l F ˙ A R = ( W ˙ C O M P 1 + W ˙ C O M P 2 − W ˙ E X P η S η m ) F ˙ A R s.t. Purity AR ≥ 99.999% Recovery AR > 90% Δ T E 1 , E 2 , o r E 3 > 2 K T C 03 > T t r i p l e m e t h a n e = 90.68 K T C 06 > T t r i p l e m e t h a n e = 90.68 K T V N 02 = T p C O N D E N 1 s T V N 03 = T p C O N D E N 2 s 103 K ≤ T C 05 ≤ 109 K Mass and energy balance and property calculation described in Section 2 . 3 Results and discussion In this section, the results of the thermodynamic and sensitivity analysis and SQP optimization solver are shown and discussed. 3.1 Thermodynamic and simulation and sensitivity analysis (1) Operating pressure of two distillation columns (demethanizer and argon column), P TCH4Bottom and P TARBottom Demethanizer and argon column are closely related to the process refrigeration subsystem, in which their re-boilers and condensers are acting as hot sources and cold sinks of heat pump refrigeration subsystem, respectively. From refrigeration point of view, the hot middle-pressure nitrogen in demethanizer's re-boiler should be condensed into liquid phase by the cold stream of liquid methane, or else its flow rate required by the re-boiler's heat duty (i.e., distillation requirements) will increase a lot owing to small gas's Cp , thus resulting in significant power increase of compressor. Considering the critical point of pure nitrogen is 126 K, if the temperature difference between hot middle-pressure nitrogen and cold liquid methane (99% methane) is set to 1.5 K, the temperature of liquid methane will be at most 124.5 K (=126–1.5), and correspondingly the bottom pressure of demethanizer will be at most methane saturation pressure of 0.26 MPa (A) according to VLE relationship. From distillation point of view, lower operating pressure is benefit to separation. Taking into account these two factors, the bottom pressure of demethanizer P TCH4Bottom is set to 0.2 MPa (A) in this study. Once P TCH4Bottom is determined, the operating pressure of argon column can be obtained by subtracting resistance of trays and pipes between them, i.e., P TARBottom = 0.198 MPa (A). The following we analyze the shortcomings existing in the coupling column proposed by Asami et al. In order to connect demethanizer with argon column, the condenser of demethanizer and the re-boiler of argon column should be united into one condenser-evaporator, in which their heat duty ( Q ˙ C O N D E N 1 and Q ˙ R E B O I L 2 ) and temperatures ( T CONDEN 1 and T REBOIL 2 ) satisfy the following Eqs. (9) and (10) , respectively, (9) Q ˙ C O N D E N 1 = Q ˙ R E B O I L 2 , (10) Δ T = T C O N D E N 1 − T R E B O I L 1 = 1 ∼ 2 K . Because the boiling point of methane is a lot different from those of argon and nitrogen, Q ˙ C O N D E N 1 is often smaller than Q ˙ R E B O I L 2 . If Eq. (9) is satisfied, the heat duty of demethanizer is bound to increase in vain. Furthermore, the satisfaction of Eq. (10) will lead to argon column working in a vacuum condition, which is because when demethanizer works at the pressure of <0.26 MPa (A), T CONDEN 1 will be <88.4 K. Subtracting heat transfer temperature difference Δ T , T REBOIL 2 will be <86 K which is lower than argon boiling point of 87 K at atmospheric pressure. Obviously this is unreasonable. (2) Condensing pressure of hydrogen-depleted ammonia purge gas, P C01 From the membrane separator, the pressure of hydrogen-depleted ammonia purge gas is 10.0 MPa (A) or so. It is natural for us to consider how to use this energy efficiently. Therefore we design two schemes to examine which one is more energy-saving: one is high pressure hydrogen-depleted ammonia purge gas expanding in expander to get work first and then the obtained low pressure purge gas entering the main heat exchanger (E1) to be liquefied; the other is high pressure hydrogen-depleted ammonia purge gas directly being liquefied in the main heat exchanger (E1). The simulation results that energy consumption of the whole system is about 1.6 kW h/N m 3 argon at P C01 = 7 MPa (A) and over 3 kW h/N m 3 argon at P C01 = 4–5 MPa (A) shows that the higher condensing pressure P C01 , and the lower energy consumption of the whole system. In addition, higher condensing pressure P C01 is beneficial to minimizing heat transfer temperature difference in the main heat exchanger where hydrogen-depleted ammonia purge gas is cooled and liquefied by low pressure cold streams. Owing to specific heat ( Cp ) of low pressure streams varying a little with temperature, the heat transfer temperature difference in the exchanger largely rely on that of high pressure hydrogen-depleted ammonia purge gas. From Fig. 7 a , it can be observed that the Cp curve of high pressure purge gas is generally high in the middle and low on both sides when temperature changed from 100 K to 300 K. With pressure P C01 increasing, the peak of Cp will become small and flush with both sides, and at the same time its location moves towards high-temperature zone, especially when P C01 = 13.0 MPa (A) the Cp curve almost turns into a straight line. This means that in the main heat exchanger (E1), both uniform and small temperature difference Δ T between hot and cold streams is much easier to implement at higher condensing pressure, P C01 . Finally the impact of condensing pressure P C01 on argon recovery is examined. Suppose that T C02 = 100 K, P C03 = 0.5 MPa (A) and condensing pressure P C01 is changed from 2.5 MPa (A) to 13.0 MPa (A), the variation of content of hydrogen, argon, nitrogen and methane in stream (H01) is presented in Fig. 7 c. It is clear that the change of P C01 almost has no impact on the content of argon, methane, and hydrogen in stream (H01) and only a little impact on that of nitrogen in it. Summarizing the above three aspects, the optimal condensing pressure P C01 of hydrogen-depleted ammonia purge gas is set to 7.8 MPa (A) in this study, the same as the inlet pressure of feed hydrogen-depleted ammonia purge gas. (3) Operating pressure of FLASH1, P C03 It is a trade-off that occurs when choosing operating pressure of FLASH1 P C03 . On the one hand removing the residual hydrogen in feed hydrogen-depleted ammonia purge gas as much as possible needs low operating pressure P C03 , but on the other hand too low P C03 will result in much argon loss and methane contained in stream C03 freezing owing to its temperature dropping below 90.68 K (triple point of methane). Through simulation and sensitivity analysis on the ASPEN PLUS platform, it is found that P C03 in 0.4–0.6 MPa (A) is appropriate in which the rate of hydrogen removal and argon recovery are about 95%, >90%, respectively, and simultaneously methane component in mixture fluids can’t freeze (see Fig. 8 ). In this paper, P C03 is set to 0.5 MPa (A). (4) Reflux ratios of two distillation columns (demethanizer and argon column), R TCH4 and R TAR As a separation process, the operating conditions of two distillation columns – demethanizer and argon column – directly influence energy consumption of the whole process, especially their reflux ratios. The smaller the reflux ratio of distillation column is, and the less power is required, but the minimum of reflux ratio is limited by product purity and yield at fixed trays number of column. Under ensuring argon recovery rate (90.4%) and purity (99.999% argon), the simulation and sensitivity analysis results show that the most economical reflux ratio of argon column, R TCH4 , is about 1.15, and that of demethanizer R TAR is from 0.33 to 0.45 depending on thermal state of feed to demethanizer, T C05 . (5) Thermal state of feed to demethanizer, T C05 The thermal state of feed to demethanizer is directly linked to energy consumption of the whole system by changing heat duties of demethanizer's re-boiler and condenser which act as hot source and cold sink of heat pump refrigeration cycle, respectively, while that of feed to argon column has no impact on energy consumption due to it coming from the top of demethanizer and always being in a state of vapor. To seek how thermal state of feed to demethanizer influences system energy consumption, the temperature of stream (C05) T C05 is changed from 103K to 109K. This is because when temperature T C05 > 109 K, that | Q CONDEN1 | > | Q REBOL1 | leads to the amount of liquid nitrogen provided by heat pump being less than what required by both distillation columns so that the whole system does not work normally, while when temperature T C05 < 103 K, methane contained in stream C06 will freeze after stream C05 being decompressed to operating pressure of demethanizer (about 0.2 MPa (A)) through throttling valve (V2). Table 2 lists the variations of a series of relevant parameters vs. temperature T C05 under the premise of the same argon purity (99.999% argon) and recovery rate (90.4%). From it, it can be seen that the lower temperature T C05 , the smaller reflux ratio of demethanizer, the smaller heat duty of demethanizer, the less the amount of middle-pressure nitrogen, and therefore the lower energy consumption which is only 87.4% at T C05 = 103 K of that at T C05 = 109 K. 3.2 ASPEN PLUS's SQP optimization (1) Optimization results The final optimal results shows that the minimal energy consumption per argon product is 1.18 kW h/N m 3 argon at P C01 = 7.8 MPa (A), P C03 = 0.5 MPa (A), P TCH4Bottom = 0.2 MPa (A), P TARBottom = 0.198 MPa (A), T C05 = 103.0005 K, T MN03 = 189 K, R TCH4 = 0.33, R TAR = 1.15, F ˙ N 00 = 382.6748 N m 3 / h , F ˙ M N 01 = 467 N m 3 /h, F ˙ D N 01 = 186 N m 3 /h, and F ˙ N 04 = 199 N m 3 /h, which agrees very well with the preceding simulation and sensitivity analysis results. Hence, both sensitivity analysis and optimization results verify that the lower thermal state of feed to demethanizer T C05 , and the lower energy consumption. The following further analyzes the effects of T C05 on the heat transfer of main heat exchanger (E1). From Fig. 9 , it can be seen that when temperature changes from 100 K to 120 K, heat transfer temperature difference Δ T in the main heat exchanger (E1) is increasing continuously, which indicates surplus cold in this temperature interval and therefore the sooner cold stream (C04) leaves the main heat exchanger (E1) with lower temperature T C05 , the smaller Δ T in it will be, i.e., the less the exergy loss in the main heat exchanger (E1) will be. When temperature is at 120K, the horizontal straight line of cold composite curve in the graph represents phase transition of methane stream CH41 (99% methane) from liquid to vapor. When temperature rises higher than 120 K, liquid methane CH41 has turned into vapor totally and heat transfer temperature difference ΔT begins to decrease fast due to small gas specific heat, i.e., system cold tends to be insufficient in this temperature interval. Note that when T C05 drops from 109 K to 103 K, the pre-cooling temperature T MN03 of middle-pressure nitrogen (in 6th column of Table 2 ) will rise from 122 K to 189 K and reduce the entire cooling loads of system in both quantity and exergy quality, thus alleviating insufficient cold situation. Accordingly the investment of heat transfer equipments of system will be also reduced. However, lower T C05 will result in the increase of ΔT in demethanizer's re-boiler where the ΔT of T C05 = 103 K is changed from about 67 K to 1–2 K owing to middle-pressure nitrogen being cooled from 189 K to 121.4 K (nitrogen condensing point at 2.7 MPa (A)), whereas that of T C05 = 109 K is only 1–2 K with middle-pressure nitrogen being in the state of phase change. (2) CPU times In addition to the reliable optimization results, the optimization efficiency is also important, which not only is an economic problem, but also reflects algorithm feasibility. For the two-column process's minimum energy consumption optimization problem in this work, ASPEN PLUS's SQP optimization solver is quite effective with the help of reasonable initial values and variables ranges obtained from thermodynamic and simulation and sensitivity analysis in advance. On the processor of 2.66 GHz Intel(R) Core (TM)2 Duo CPU with 4 GB RAM, it only takes the CPU time about 10 s to accomplish. 4 Conclusions For the two-column cryogenic process for argon recovery from hydrogen-depleted ammonia purge gas, the ASPEN PLUS process model is established first in this paper, and then the accuracy of property calculation model – the built-in PENG-ROB property method of ASPEN PLUS – is verified by comparison with experimental data from three aspects: VLE, liquid phase density, and enthalpy. Based on the above, the detailed thermodynamic and simulation and sensitivity analysis and optimization using the built-in SQP solver in ASPEN PLUS have been done for minimizing two-column process's energy consumption per argon product, from which the conclusions can be drawn as follows: (1) Condensing pressure of hydrogen-deplete ammonia purge gas, P C01 It plays an important role in reducing the energy consumption of whole system which is less at higher condensing pressure. Besides, the higher condensing pressure P C01 , and the easier for uniform and small ΔT between hot and cold streams in the main heat exchange (E1) to implement. It almost has no impact on the purity and yield of argon recovery. (2) Operating pressure of Flash1, P C03 It has great influence on argon recovery and hydrogen removal. Its choice is a trade-off that on the one hand is to ensure the emission of hydrogen as much as possible, and on the other hand is to ensure the recovery of argon sufficiently. Simulation results show that P C03 = 0.4–0.6 MPa (A) is optimal. (3) Reflux ratio of two distillation columns (demethanizer and argon column) Research shows that the most economical reflux ratio of argon column is 1.15, and that of demethanizer varies from 0.33 to 0.45 depending on the thermal state of feed to demethanizer, T C05 . (4) Thermal state of feed to demethanizer, T C05 In terms of thermodynamic analysis and simulation, temperature T C05 can be changed only from 103 K to 109 K (see Section 3.1 ), but it is the most valuable means of reducing energy consumption at fixed inlet conditions of feed hydrogen-depleted ammonia purge gas. Both sensitivity analysis and optimization results verify that the lower the temperature T C05 , and the lower system energy consumption which at T C05 = 103 K is only 87.4% of that at T C05 = 109 K. In addition, the lower T C05 , the more uniform and smaller ΔT in the main heat exchanger (E1), and thus the less irreversible loss in it. (5) Finally optimization results show that the minimal energy consumption per argon product is 1.18 kW h/N m 3 argon at P C01 = 7.8 MPa (A), P C03 = 0.5 MPa (A), P TCH4Bottom = 0.2 MPa (A), P TARBottom = 0.198 MPa (A), T C05 = 103.0005 K, T MN03 = 189 K, R TCH4 = 0.33, R TAR = 1.15, F ˙ N 00 = 382.6748 N m 3 /h, F ˙ M N 01 = 467 N m 3 /h, F ˙ D N 01 = 186 N m 3 /h, and F ˙ N 04 = 199 N m 3 / h . References API, 1983 API, 1983. Technical Data Book – Petroleum Refining, Washington, DC. Asami et al., 1996 T. Asami M. Tanakat Energy consumption in the recovery of argon from purge gas in ammonia synthesis plant Cryogenics 36 12 1996 997 1003 Bao et al., 2002 J. Bao B.L. Gao Simulation of industrial catalytic-distillation process for production of methyl tert-butyl ether by developing user's model on Aspen plus platform Chemical Engineering Journal 90 3 2002 253 266 Cho et al., 1998 S.-H. Cho K.-T. Chue A two stage PSA for argon and hydrogen recovery from ammonia purge gas Chemical Engineering Communications 163 1 1998 97 109 Doherty et al., 2009 W. Doherty A. Reynolds The effect of air preheating in a biomass CFB gasifier using ASPEN Plus simulation Biomass & Bioenergy 33 9 2009 1158 1167 Gravelle and Lu, 1973 D. Gravelle B.C.Y. Lu Equilibrium between vapor and liquid for nitrogen-argon-methane system Canadian Journal of Chemical Engineering 51 2 1973 227 229 Grini et al., 1998 P.G. Grini G.A. Owren Isobaric enthalpy increment and isenthalpic Joule-Thomson effect measurements on synthetic gas containing binary, or ternary mixtures of methane, ethane, propane, and nitrogen Journal of Chemical Thermodynamics 30 8 1998 1011 1027 Hao et al., 2008 X. Hao M.E. Djatmiko Simulation analysis of a GTL process using ASPEN plus Chemical Engineering & Technology 31 2 2008 188 196 Healy, 1992 M.J. Healy Argon production from ammonia-plant purge gas Gas Separation & Purification 6 4 1992 191 196 Hiza et al., 1977 M.J. Hiza W.M. Haynes Orthobaric liquid densities and excess volumes for binary-mixtures of low molar mass alkanes and nitrogen between 105 and 140 K Journal of Chemical Thermodynamics 9 9 1977 873 896 Jian, 1988 Z.H. Jian Review on separation technology from ammonia purge gas M-Sized Nitrogenous Fertilizer Progress (China) 2 1988 34 43 Jian, 1993 Z.H. Jian The inland and abroad development of separation technology from ammonia purge gas Anhui Chemical Industry (China) 2 1993 15 22 Jin et al., 1993 Z.L. Jin K.Y. Liu Vapor-liquid-equilibrium in binary and ternary mixtures of nitrogen, argon, and methane Journal of Chemical and Engineering Data 38 3 1993 353 355 Kerry, 2006 F.G. Kerry Gas separation and purification Industrial Gas Handbook 2006 CRC Press, Taylor & Francis Group Boca Raton, London, New York p. 539 Krishnamurthy et al., 1987a R. Krishnamurthy S.L. Lerner A new process for argon recovery from hydrogen depleted ammonia plant purge gas Gas Separation & Purification 1 1 1987 16 22 Krishnamurthy et al., 1987b Krishnamurthy, R., Shukla, Y., et al., 1987. High Yield Harp Plant for Argon Recovery from Ammonia Purge Gas, Minneapolis, MN, USA, AIChE, New York, NY, USA. Lambert and Simon, 1962 M. Lambert M. Simon Excess thermodynamic properties of the liquid systems A-CH 4 AND CO-CH 4 Physica 28 1962 1191 1196 Lewis et al., 1977a K.L. Lewis S.E. Mosedale Enthalpies of mixing of methane + argon, methane + nitrogen, and methane + hydrogen in gaseous and 2-phase regions Journal of Chemical Thermodynamics 9 2 1977 121 131 Lewis et al., 1977b K.L. Lewis S.E. Mosedale Enthalpy of mixing of argon + nitrogen at 123, 130, and 140 K at pressures up to 10.1 MPa Journal of Chemical Thermodynamics 9 3 1977 221 229 Lewis et al., 1975 K.L. Lewis G. Saville Excess enthalpies of liquid-oxygen + argon, oxygen + nitrogen, and argon + methane Journal of Chemical Thermodynamics 7 4 1975 389 400 Lewis and Staveley, 1975 K.L. Lewis L.A.K. Staveley Excess enthalpies of liquid-mixtures nitrogen + oxygen, nitrogen + argon, argon + ethane, and methane + carbon tetrafluoride Journal of Chemical Thermodynamics 7 9 1975 855 864 Maclean et al., 1987a Maclean, D.L., Krishnamurthy, R., et al., 1987a. Argon recovery from ammonia plant purge gas utilizing a combination of cryogenic and non-cryogenic separating means. The BOC Group, Inc. (Montvale, NJ), United States. UP 4689062. Maclean et al., 1987b Maclean, D.L., Krishnamurthy, R., et al., 1987b. Argon recovery from hydrogen depleted ammonia plant purge gas utilizing a combination of cryogenic and non-cryogenic separating means. The Boc Group, Inc. (Montvale, NJ), United States. UP 4687498. Masayuki and Ryuichi, 1985 Masayuki, T., Ryuichi, F., 1985. Methods of argon recovery. Japanese. JP 1985-171377. Massengi and Miller, 1973 D.R. Massengi R.C. Miller Temperature-dependence of excess volumes for simple liquid mixtures – N2 + AR, N2 + AR + CH 4 Journal of Chemical Thermodynamics 5 2 1973 207 217 McClure et al., 1976 D.W. McClure K.L. Lewis Excess enthalpies and GIBBS free-energies for nitrogen + methane at temperatures below critical-point of nitrogen Journal of Chemical Thermodynamics 8 8 1976 785 792 Meisler, 1971 Meisler, J.T. (NJ), Van Baush, E.H. (Pearl River, NY), Banikiotes, G.C. (Seaford, NY), 1971. Process for cryogenic purification of hydrogen, United States, Hydrocarbon Research, Inc. (New York, NY). UP 3628340. Mosedale and Wormald, 1977 Mosedale, S.E., Wormald, C.J., 1977. Enthalpy of mixing of argon + methane at 123, 130, 140, and 145 K, at pressures up to 9.12 MPa. Journal of Chemical Thermodynamics 9 (5), 483–490. Ongiro et al., 1996 Ongiro, A., Ugursal, V.I., et al., 1996. Thermodynamic simulation and evaluation of a steam CHP plant using ASPEN Plus. Applied Thermal Engineering 16 (3), 263–271. Paviet et al., 2009 F. Paviet F. Chazarenc Thermo chemical equilibrium modelling of a biomass gasifying process using ASPEN PLUS International Journal of Chemical Reactor Engineering 2009 7 Peng and Robinson, 1976 D.Y. Peng D.B. Robinson A new two-constant equation-of-state Industrial & Engineering Chemistry Fundamentals 15 1976 59 64 Rackett, 1970 H.G. Rackett Equation of state for saturated liquids Journal of Chemical and Engineering Data 15 1970 514 517 Sarvar-Amini et al., 2007 A. Sarvar-Amini R. Sotudeh-Gharebagh Sequential simulation of a fluidized bed membrane reactor for the steam methane reforming using ASPEN PLUS Energy & Fuels 21 2007 3593 3598 Smejkal and Soos, 2002 Q. Smejkal M. Soos Comparison of computer simulation of reactive distillation using ASPEN PLUS and HYSYS software Chemical Engineering and Processing 41 5 2002 413 418 Soave, 1972 G. Soave Equilibrium constants from a modified Redich-Kwong equation of state Chemical Engineering Science 27 1972 1197 1203 Spencer and Danner, 1972 C.F. Spencer R.P. Danner Improved equation for prediction of saturated liquid density Journal of Chemical and Engineering Data 17 1972 236 241 Springmann, 1982 Springmann, H., 1982. Methods for Argon Recovery to Meet Increased Demand on the Argon Market. Cryogenic Processes and Equipment 1982, AlChE Symposium Series, 1983, No. 224, vol. 79, pp. 12–17. Sprow and Prausnitz, 1966 F.B. Sprow J.M. Prausnitz Vapour-liquid equilibria and surface tensions for the nitrogen-argon-methane system at 90.67K Cryogenics 6 6 1966 338 340 Sun et al., 1984 Z.F. Sun Y.W. He The calculation of thermodynamic properties of liquid ammonia mixture Chemical Engineering (China) 5 1984 42 46 Tang and Chen, 2002 L. Tang B.s. Chen Recovery technology of hydrogen from discharge gas in ammonia synthesis Journal of Chemical Industry & Engineering (China) 23 4 2002 16 18 Thorogood et al., 1991 Thorogood, R.M., Tsai, W., et al., 1991. Hybrid membrane – cryogenic generation of argon concurrently with nitrogen. United States, Air Products and Chemicals, Inc. (Allentown, PA). UP 5049174. Venkataraman et al., 1990 S. Venkataraman W.K. Chan Reactive distillation using Aspen Plus Chemical Engineering Progress 86 8 1990 45 54 Xu, 1984 Y.T. Xu Cryogenic distillation of recovery hydrogen, argon and other rare gas from ammonia purge gas Low Temperature and Specialty Gases (China) 1 1984 27 33 Xu, 1985 Y.T. Xu Design of ammonia purge gas's cryogenic separation plant Cryogenic Technology (China) 2 1985 15 20 Yan and Rudolph, 2000 H.M. Yan V. Rudolph Modelling a compartmented fluidised bed coal gasifier process using aspen plus Chemical Engineering Communications 183 2000 1 38 Ye et al., 2009 G.Y. Ye D.L. Xie Modeling of fluidized bed membrane reactors for hydrogen production from steam methane reforming with Aspen Plus International Journal of Hydrogen Energy 34 11 2009 4755 4762 Zhang, 1980 Y.S. Zhang Synthetical utilization of ammonia purge gas Journal of Chemical Fertilizer Industry (China) 6 1980 47 51 Zhang and Li, 1988 Y.S. Zhang Y.X. Li Recovery of hydrogen, methane and rare gases from discharge gas in ammonia synthesis Chemical Engineering Design Communications (China) 14 2 1988 14 22