Bo Lu ,Yuanhui Shen ,Zhongli Tang ,Donghui Zhang,,Gaofei Chen

1 State Key Laboratory of Coal and CBM Co-mining,Jincheng 048000,China

2 The Research Center of Chemical Engineering,School of Chemical Engineering and Technology,Tianjin University,Tianjin 300072,China

3 Key Laboratory of Cryogenics,Technical Institute of Physics and Chemistry,Chinese Academy of Sciences,Beijing 100190,China

Keywords:Coalbed methane enrichment VPSA process Activated carbon Numerical modeling

ABSTRACT The enrichment of low concentration coalbed methane using adsorption process with activated carbon adsorbent was studied in this work.Adsorption isotherms of methane,nitrogen and carbon dioxide on activated carbon were measured by volumetric method,meanwhile a series of breakthrough tests with single component,binary components and three components feed mixture has been performed for exploring dynamic adsorption behaviors.Moreover,a rigorous mathematical model of adsorption bed containing mass,energy,and momentum conservation equation as well as dual-site Langmuir model with the Linear driving force model for gas–solid phase mass transfer has been proposed for numerical modeling and simulation of fixed bed breakthrough process and vacuum pressure swing adsorption process.Furthermore,the lumped mass transfer coefficient of methane,nitrogen and carbon dioxide on activated carbon adsorbent has been determined to be 0.3 s-1,1.0 s-1 and 0.06 s-1 by fitting the breakthrough curves using numerical calculation.Additionally,a six bed VPSA process with twelve step cycle sequence has been proposed and investigated for low concentration coalbed methane enrichment.Results demonstrated that the methane molar fraction in feed mixture ranged from 10%to 50%could be enriched to 32.15% to 88.75% methane in heavy product gas with a methane recovery higher than 83%under the adsorption pressure of 3 bar (1 bar=105 Pa) and desorption pressure of 0.1 bar.Energy consumption of this VPSA process was varied from 0.165 kW﹒h﹒m-3 CH4 to 0.649 kW﹒h﹒m-3 CH4.Finally,a dual-stage VPSA process has been successfully developed to upgrade a low concentration coalbed methane containing 20% methane to a target product gas with methane purity higher than 90%,meanwhile the total methane recovery was up to 98.71% with a total energy consumption of 0.504 kW﹒h﹒m-3 CH4.

1.Introduction

Clean energy and environmental issues have been attracted more and more attention,with the rapid development of modern society [1].CH4has been widely recognized as clean fossil fuel,but it’s also a kind of greenhouse gas with 28 times Global Warming Potential of CO2if it was released into the atmosphere [2].Coalbed methane has been considered as unconventional natural gas associated with coal mining.According to statistics by China National Petroleum Corporation at 2015,the total gas-in-place of China’s CBM resources buried shallower than 2000 m was estimated at 29.8 trillion cubic meters(tcm)and the technically recoverable resource was approximately 12.5 tcm.However,the extraction of CBM was 13.6 billion cubic meters (bcm)and its utilization was lower to 4.8 bcm at the same year in China[3].All that was attributed to characters of underground gas drainage in China,which has resulted low quality coal bed methane with CH4concentration lower than 30%.How to improve the efficiently of lowquality coalbed methane is research hotspot in recent years,since it is not only a kind way to meet the growing demand of energy,but also an effective mean to resolve its’associated environmental problems.

The low concentration coalbed methane mainly consists of air and methane,with a trace of carbon dioxide,water,and hydrogen sulfide.The recovery and further enrichment of low-quality coalbed methane has been considered as an effective way to produce high quality methane that could be sent to natural gas pipeline.Currently,the upgrading and enrichment techniques mainly includes cryogenic distillation,membrane separation,pressure swing adsorption as well as gas hydrate separation [4].Pressure swing adsorption process is more competitive for enrichment of coal bed methane because of the lower power consumption,lower operation cost and higher operational flexibility[5].However,the similar physical properties between CH4and N2have make them more difficult to be separated,because it is a great challenge for the design of adsorbent with higher selectivity and capacity used in CH4/N2adsorption process.Up to now,most adsorbents used for CH4/N2separation can be classified as two categories based on their separation mechanism.Activated carbon(AC) and zeolite belong to equilibrium selective adsorbent,while Carbon molecular sieve,Engelhard titanosilicate(ETS),and Clinoptilolite likely belong to kinetically selective adsorbent [6].

Zhou et al.[7] have evaluated different activated carbon adsorbent with various specific surface and pore volume for CH4/N2separation by breakthrough experiments early.Super-activated carbon with a surface area of 3000 m2﹒g-1and a pore volume of 1.5 cm3﹒g-1could achieve an amazed separation coefficientof 20.13,which was almost 4–5 times of commercial activated carbon.However,there has been no further report related to this super-activated carbon and its application for CH4/N2separation.Olajossy et al.[8] have proposed a three bed VPSA process using AC adsorbent for coalmine methane enrichment with heavy product reflux step.Experimental results demonstrated that 55.2%(vol) CH4in coalmine methane gas could be concentrated to 96%–98% (vol) CH4with a methane recovery of 86%–91%,under adsorption pressure of 300 kPa and desorption pressure of 25 kPa.Subsequently,Yang et al.[9] and Sun et al.[10] have performed a systematical study of CH4/N2separation using AC adsorbent numerically.A gas mixture consisting of 30%CH4-70%N2could produce a methane enriched product gas with CH4purity and recovery higher than 80% and 97%,respectively,using a three bed VPSA process under optimal operating conditions.Furthermore,Saleman et al.[11] and Zhang et al.[12] have developed and constructed a dual-reflux pressure swing apparatus and cycle configuration for low grade methane capture using activated carbon NoritRB3.Their experimental and simulation results indicated that a heavy product with 51.3% CH4can be concentrated from a feed gas containing 2.4% CH4but with a slightly lower methane recovery of 55%.Moreover,their DR-PSA process has been verified to be energetically self-sustainable by numerically analysis.Besides,Liu et al.[13] has also firstly proposed a PSA cycle configuration complemented with CO2displacement for CH4enrichment from low concentration coalbed methane.17.62%(vol)CH4in feed gas can be transferred into a product stream containing more than 90% CH4with a CH4recovery higher than 98% by using self-built activated carbon as adsorbent.Yang et al.[14] and Qu et al.[15]have further confirmed the feasibility of PSA-CO2displacement for the enrichment of ventilation air methane experimentally and numerically.Silicalite is a kind of high ratio Si/Al zeolite that has been reported by Delgado et al.[16,17] for CH4/N2equilibrium selective separation.Numerical results demonstrated that 50%CH4–50% N2feed mixture could be enriched into a methane product with 96.3% CH4purity and 96.7% CH4recovery by using a three bed VPSA cycle configuration containing one rinse step and two equalization steps.

Compared to activated carbon,Carbon Molecular Sieve(CMS)is another kind of carbon material possessed a uniform pore diameter for gas separation by using the difference of kinetic diffusion.Fatehi et al.[18] firstly reported separation of CH4/N2by pressure swing adsorption using a CMS.The methane concentration in product stream could up to 76% for the 60% methane feed and up to 96% for the 92% methane feed with a two-column four-step PSA cycle under optimal operating conditions.Warmuzinski and Sodzawiczny [19] investigated and evaluated effects of operating parameters on two-bed PSA process performances for the separation of CH4and N2with CMS adsorbent,including adsorption pressure,feed concentration,and duration of cycle.Yang et al.[20]investigated different modification method of coal-based carbon molecular sieve for the separation of CH4/N2.Their results indicated that low temperature plasma technology was an alternative modified way for CMS to achieve a high adsorption capacity and equilibrium separation coefficient of 3.3 for CH4/N2,but without consideration of kinetic separation coefficient.Yang et al.[21]have experimentally evaluated three kinds of commercial CMS for the upgrading of coal mine methane consisting of CH4/N2/O2.Their results performed that CMS adsorbents exist an optimal balance between N2/CH4kinetic difference and equilibrium adsorption capacity through careful design with proper existence of microporosity.In addition,Takeda 3KT CMS has been verified as the best adsorbent for upgrading of coal mine methane with their proposed separation factorof 0.414.Zhang et al.[22] have detailed investigated the influence of deposition conditions for the preparation of carbon molecular sieve employed in CH4-N2separation,including deposition temperature,time and flow rate of benzene.Based on the comparison of adsorption capacity and separation coefficient,the deposition temperature of 1023 K,time of 60 min and benzene of flow rate of 4 ml﹒min-1would result in excellent CMS adsorbent with a high N2/CH4kinetic selectivity of 35.26.Their CMS adsorbent could concentrate 27% methane balanced with nitrogen to a methane product stream with purity and recovery of 57.20% and 90.90%,respectively,by using a four bed PVSA process.Li et al.[23] have also performed three bed VPSA process for purification of coal mine methane using CMS adsorbent through experimental and theoretical methods.Feed mixture with methane concentration higher than 50% could be enriched to be 96% methane product gas with a methane recovery of 63% under operating conditions of adsorption pressure of 5 bar and desorption pressure of 0.2 bar.In addition,Bhadra and Farooq [24] have developed a detailed PSA model for comprehensive evaluation of separation performances of three ETS-4 and two CMS adsorbent for nitrogen rejection from feed mixture contained 90%methane–10% nitrogen.Their simulation results performed that Ba400 type ETS-4 adsorbent was the best choice for Pipeline methane production at the view of purity,recovery and productivity.On the contrary,all CMS samples cannot be meet the pipeline specification under the same operation conditions.We assumed that the main reason was due to that CMS adsorbents selected in their studies were design for nitrogen production in industrial air production,in which the pore texture was more suitable for oxygen kinetic adsorption,not for nitrogen kinetic selective adsorption.Here,it is also worth to note that ion-exchanged titanosilicate-4 is an excellent molecular sieving adsorbent discovered by Kuzniki and commercialized by Engelhard Corporation for CH4/N2separation [25,26].Its equilibrium capacity and uptake rates for different species could be precise controlled by the type of exchange ions and the pore texture resulted from dehydration.However,their expensive feedstock and challenging preparation method have resulted in much higher price of ETS-4 adsorbent,which has limited its wide use for natural gas upgrading.On the contrary,Clinoptilolites,naturally occurring zeolites,has been widely used for ion exchange of waste water and drying,while it has attracted more attention for gas separation and purification currently.Jayaraman et al.[27,28] have prepared series of single ion-exchanged clinoptilolites and mixed ion-exchanged Clinoptilolites for the CH4/N2separation.Depend on basic parameters of adsorption equilibrium and kinetics,their process simulation results demonstrated that Mg/Na clinoptilolite and Ce clinoptilolite used in two stage PSA cycle can produce pipeline methane product(methane concentration >96%) with methane recovery more than 90%from 80%CH4–20%N2.Recently,Kennedy et al.[29,30]further investigated ion-exchanged clinoptilolite on adsorption separation of CO2,CH4and N2.Their experimental results indicated that Fe3+exchanged clinoptilolite exhibits a binary CH4/N2separation factors ranged from 3.9 to 7.3,while Ag+exchanged clinoptilolite possesses a binary N2/CH4separation factors ranged from 2.8 to 7.6.

Given the above,although lots of research have been reported for the separation of CH4and N2,but up to now just activated carbon and Engelhard titanosilicate-4 has been successful used for nitrogen rejection from methane contained mixture at industrial scale.Meanwhile,the isotherm parameters and kinetic parameters of component on AC adsorbent are still scarcity.Additionally,reports of VPSA process cycle and configuration for methane enrichment was from a view of lab-scale study,not a practice application.In this study,a micropore coal-based activated carbon has been selected for CH4enrichment by vacuum pressure swing adsorption process from low concentration coalbed methane.Adsorption isotherms of CO2,CH4and N2on activated carbon were measured at 293 K,308 K and 323 K with pressure up to 10 bar.Moreover,a series of breakthrough experiment has been performed with single and binary feed mixture,meanwhile the Linear driving force (LDF) coefficient of component on adsorbent pellet has been estimated from experimental results by using numerical calculation with a rigorous mathematical model.Furthermore,a six bed VPSA process has been proposed for low concentration coalbed methane enrichment and effects of key operating parameters on process performances were evaluated.Finally,a dual-stage VPSA process was further developed to concentrate a coalbed methane to meet the specification of pipeline gas.

2.Experimental

2.1.Characterization of materials

Coal-based activated carbon prepared in author’s Lab,has been selected as methane selective adsorption adsorbent.Nitrogen adsorption and desorption isotherms at 77 K has been performed to characterized the adsorbent measured on Micromeritics ASAP 2020 and presented in Fig.1a.Meanwhile the pore diameter distribution of activated carbon is illustrated in Fig.1b calculated by NLDFT model.As observed from Fig.1,this activated carbon is kind of microporous adsorbent with HK Median pore width of 4.8 ?(1 ?=0.1 nm) and micropore surface area of 669.6 m2﹒g-1.Moreover,an apparent BET surface area of 904.35 m2﹒g-1and total pore volume of 0.5025 cm3﹒g-1of AC adsorbent have been measured.Detailed texture characterizations of AC adsorbent are listed in Table 1.

Table 1 Texture characterization of AC adsorbent

2.2.Adsorption equilibrium measurements

Pure component isotherms of CO2,CH4and N2on activated carbon were measured by using a volumetric apparatus [31,32] to represent thermodynamic equilibrium relationship between adsorbent and adsorbates.The schematic diagram of isotherm measurement apparatus is depicted in Fig.2.It mainly contains adsorption cell,loading cell,bath kettle,vacuum pump and pressure&temperature transducer.The adsorption cell and loading cell were placed in bath kettle to maintain a constant temperature during isotherm measurement,in which the measurement temperature can be ranged from 0 °C to 100 °C with an accuracy of 0.1°C.Moreover,the pressure transducers were deployed for pressure measurement and record during experiments,which can be run at range of 0–20 bar with accuracy of 0.01 bar.Prior to measurement experiment,AC adsorbent was activated firstly in tubular furnace with helium at temperature of 250 °C.Then it has been transferred to adsorption cell and further activated at 100°C under vacuum overnight.After that,isotherm measurement experiments could begin,pure component first introduced into loading cell from gas cylinder to achieve the set pressure,then adsorbate sorted in loading cell could flow into the adsorption cell for the adsorption procedure.Adsorption equilibrium was reached after temperature and pressure of both cells keep same.Finally,equilibrium adsorption amount could be calculated based on the changes of mole numbers of free gas in the two cells before and after adsorption occurring.

2.3.Fixed bed breakthrough experiments

Single component,binary component and three component feed mixture breakthrough experiments were carried out on the fixed bed apparatus,which were used to explore dynamic adsorption behaviors as well as to determine component effective mass transfer coefficient on adsorbent pellet.The schematic diagram is depicted in Fig.3,in which an adsorption bed with a diameter of 2 cm and a length of 40 cm packed with 67.5 g AC adsorbent.The component and flow rate of feed mixture were controlled by mass flow controller,while the analysis of gas mixture sample was completed by gas chromatography.All breakthrough experiments were carried out with feed temperature of 293 K,adsorption pressure of 4 bar as well as environmental temperature of 293 K.While the flow rate were designed and fixed at 250 ml﹒min-1,500 ml﹒min-1,1000 ml﹒min-1for nitrogen,methane and carbon dioxide single component breakthrough experiments,but with various composition of feed mixture (20% adsorbates–80% He,50%adsorbates–50%He,80%adsorbates–20%He)to represent various operating partial pressure for feed mixture breakthrough experiment.In addition,before breakthrough experiments,AC adsorbent in adsorption bed was firstly regenerated under vacuum condition,then the bed was repressurized to adsorption pressure by introducing helium that was assumed to be an inert gas with respect to AC adsorbent.

Fig.1.(a) N2 adsorption and desorption isotherm on AC at 77 K and (b) the curve of pore size distribution of AC.

Fig.2.Schematic diagram of static volumetric apparatus for isotherm measurement.

3.Theoretical Section

3.1.VPSA process description

A six-bed VPSA process depicted in Fig.4 with a twelve steps cycle configuration has been developed with AC adsorbent for low concentration coalbed methane enrichment,which would be investigated numerically.The cycle sequence of VPSA process is listed in Table 2,and the diagram of its cycle configuration is presented in Fig.5.It’s worth to note that three times pressure equalization step have been employed in cycle configuration for energy conservation and enhancing methane concentration retained in adsorption bed by removing weakly adsorbed component.The idle step was introduced to make cycle sequence more reasonable.Detailed description of each step is given below.

Table 2 Cycle sequence of VPSA process

Adsorption step(AD):Low concentration coal bed methane feed mixture was compressed and introduced into adsorption bed,in which strongly adsorbed component(CH4)was selective adsorbed on AC adsorbent.While weakly adsorbed component (N2) was released into the atmosphere from the top of adsorption Bed.The adsorption step should be terminated before the methane break through adsorption bed to keep high methane recovery.

Equalization-depressurization step (ED):Following adsorption step,three successive equalization-depressurization steps were implemented.It should note that pressure equalization step in VPSA cycle configuration is kind of interaction step between high-pressure bed and low-pressure bed.The gas flows from adsorption bed with high pressure to another one with low pressure.Therefore,weakly adsorbed component retained in high pressure adsorption bed could be removed by equalizationdepressurization step,which would result in an enrichment of methane in adsorption bed.

Countercurrent blowdown step (BD):After the equalization depressurization step finished,counter-current blowdown step would be carried out to depressurize of adsorption bed to atmospheric pressure and recover partial heavy product if the pressure in adsorption bed is higher than atmospheric pressure,otherwise the adsorption bed could directly implement next step like vacuum step to recover methane from bed.

Vacuum step (VU):Vacuum pump would be deployed in this step to further depressurize adsorption bed to an appropriate value,meanwhile methane adsorbed on AC adsorbent in adsorption bed would be completely desorbed and collected as the methane product gas.At the same time,AC adsorbent would be almost complete regenerated and could be used to selective adsorption of methane effectively in next cycle.

Fig.3.The schematic diagram of breakthrough experimental setup.

Fig.4.The schematic diagram of six bed VPSA process.

Fig.5.Schematic diagram of cycle configuration and adsorption front of methane in bed.

Equalization-repressurization step (ER):The gas from high pressure adsorption bed just finished adsorption would be accepted by adsorption bed under equalization-repressurization step,which would partially pressurize adsorption bed to a middle pressure.This step or method is more helpful for conserving mechanical energy of high-pressure gas stream.Besides,adsorption front of strongly adsorbed component would move back to the bottom of adsorption bed after this step.Three successive equalization-repressurization steps were carried out with respect to three successive equalization-depressurization steps.

Pressurization step(PR):The adsorption bed was bringing back to adsorption pressure by introducing high purity light product stream that was produced by another bed undergoing adsorption step.Therefore,the adsorption bed would be ready for next cycle sequence.

The composition of low concentration coalbed methane as feed mixture introduced into VPSA process has been simplified.The water vapor and other heavier adsorbed component were assumed to be completely removed by pretreatment step.The compositions of low concentration coalbed methane are consisting of N2and CH4,sometimes including a little portion of CO2.Adsorption dynamic behaviors in adsorption bed were explored by numeric modeling of VPSA process,including process performances,pressure profiles,temperature profiles,gas phase concentration profiles and solid phase concentration profiles in adsorption bed over one cycle after the VPSA process achieve cyclic steady state.The physical properties of adsorption and operating conditions of VPSA process are given in Table 3 in supplementary information.Furthermore,effects of key operating parameters on process performances were investigated and analyzed,including feed flow rate,methane concertation in feed mixture,desorption pressure as well as a little portion of carbon dioxide in feed mixture.Finally,a dual-stage six bed VPSA process has been further constructed and investigated for low concentration coalbed methane enrichment and producing high purity methane product gas to meet the specification of pipeline gas.The cycle sequence and cycle configuration of the second stage VPSA process was same as the cycle sequence and cycle configuration presented in Table 2 and Fig.5,While a light product purge step has been introduced at the end of Vacuum step in the first stage VPSA process cycle sequence to enhance the methane recovery.

3.2.Mathematical model for adsorption bed

Compared to distillation and absorption process,PSA/VPSA process is periodically cyclic and dynamic separation process without true steady state.Moreover,operating schedule step and key operating parameters in VPSA process illustrate stronger couple and interaction.Experimental evaluations of adsorbents and VPSA cycle configuration are time consuming and expensive.On contrary,Numerical modeling and simulation with a rigorous mathematical model is an efficient way for revealing transient behaviors in adsorption bed as well as to accelerate the design and optimization of VPSA process.Here,an axially dispersed plus-flow mathematical model was constructed for adsorption bed that is depicted in Table 4,which containing mass conversation,momentum conversation,energy momentum equation and the dual site Langmuir model for adsorption equilibrium as well as the linear driving force model with a lumped mass transfer coefficient accounting component transfer between gas phase and solid phase.Additionally,some assumptions are adapted to simply model to reduce the computation cost under conditions of reliable results.The assumptions were described below:

Table 3 The physical properties of AC adsorbent and adsorption bed

a.The gas obeys the ideal gas law.

b.There is no radial concentration and temperature gradients.

c.Ergun equation is employed to calculate pressure drop along the bed.

d.The competitive adsorption behaviors are described by Extended Langmuir model.

e.Rigorous energy balance consists of energy balance of gas phase,solid phase and bed wall,respectively.

f.A Linear driving force (LDF) model is used to describe the mass transfer between solid and gas phase.

The resulted mathematical model of adsorption bed,accompanied with necessary initial and boundary conditions have constituted a systems of Partial Differential Equations (PDEs),Ordinary Differential Equations (ODEs),and Algebraic Equations (AEs),which was numerically solved using the gPROMS software.The PDEs were discretized in the axial domain by the second-order Centered finite difference method,then converted into a set of differential and algebraic equations (DAEs).The DAEs system was dynamically solved by intergrading over time in solver DASOLV with a relative error tolerance of 10-5.Additionally,the necessary initial and boundary conditions for adsorption bed in breakthrough process and VPSA process are illustrated in Tables 5 and 6.

4.Results and Discussion

4.1.Adsorption isotherms of CH4/N2/CO2

The isotherm measured pressure was ranged from 0 to 1000 kPa under temperatures of 293 K,308 K and 323 K.As shown in Fig.6,the order of adsorption capacity is CO2>CH4>N2under the same temperature and pressure.The dual-site Langmuir model given in Eq.(1) has been selected to correct experimental data due to its strong applicability on homogeneous and heterogeneous adsorbent,and it describes adsorption behaviors of adsorbates on AC adsorbent very well.The fitting parameters of isotherm model are presented in Table 7 as well as the average value of isosteric heat of adsorption that was calculated by clausius-claypeyron equation.The order of isosteric heat of adsorption is the same of the order adsorption capacity.

Table 4 Mathematical model of adsorption bed

Table 5 Boundary conditions and initial conditions of adsorption bed for breakthrough experiment.

Table 6 Boundary conditions of adsorption bed at each step for VPSA process and its initial conditions

Table 7 Fitting parameters of the Dual-site Langmuir model

4.2.Breakthrough experiment and simulation

All single component breakthrough experimental results are summarized in Fig.7.The left diagrams are breakthrough curve for different species with various feed composition,while the right picture are temperature profiles with a molar fraction of adsorbate in feed mixture fixed at 50%.It can be seen from Fig.7 that a higher component concentration in feed mixture would result in a shorter breakthrough time.This is due to that adsorption capacity of component on AC adsorbent was limited that lead to the adsorption front of component much fast to reach the outlet of adsorption bed,even a higher component partial pressure in adsorption process can be achieved by a higher component concentration in feed mixture.In addition,comparing the shape of breakthrough curve of CO2,CH4and N2,the slope of the nitrogen breakthrough curve is much steeper than that of methane and carbon dioxide,which means that the mass transfer of nitrogen was much faster on AC adsorbent.These phenomena are consistent with results of LDF coefficient extracted from numerical simulation,where the lumped LDF coefficient of N2,CH4and CO2are fitting to approximately 1.0 s-1,0.3 s-1and 0.06 s-1,respectively.Moreover,the same variation trend of temperature profiles can be observed that temperature at the fixed point in adsorption bed increases firstly then decreases over time because of the exothermic character ofadsorption process and heat transfer to environment.The maximum temperature achieved at fixed position in adsorption process follows the order CO2>CH4>N2,this is due to the same order of adsorption amount and adsorption heat of species on AC adsorbent.The temperature profiles obtained from numerical simulation are of well agreement with the experimental results except catching the point of maximum temperature.This may be due to that 1D mathematic model without considering radial variation and an average value of isosteric heat of adsorption has been employed in numerical calculation.

Breakthrough experiments with CH4-N2and CO2-CH4-N2mixture were launched to explore multi-component competitive adsorption thermodynamic,and their breakthrough curves are depicted in Figs.8 and 9,respectively.As shown in Fig.8,the roll-up phenomenon of nitrogen breakthrough curve can be observed due to the preferential adsorption of methane on AC adsorbent over N2,which has resulted in a much higher nitrogen concentration in outlet gas stream than that in feed mixture.Moreover,a higher methane concentration in feed mixture could lead to a much higher enrichment of nitrogen in outlet gas because the displacement effect of methane with respect to nitrogen was more obvious.On the other hand,two roll-up phenomena can be observed in Fig.9 with three component feed mixture,and the roll-up of nitrogen breakthrough curve is more obvious than that of methane.This is because nitrogen would be simultaneously displaced by both methane and carbon dioxide while methane would just be displaced by carbon dioxide.Furthermore,the reasonable agreement between experimental breakthrough curves and simulated breakthrough curves depicted in Figs.8 and 9 indicates that the LDF mass transfer coefficient fitted and extracted from single component breakthrough experiments could be used to predicate multi-component breakthrough curve with an acceptable deviation.Temperature profiles of each breakthrough experiments are also illustrated in Figs.8 and 9,respectively.Two roll-up phenomena and three roll-up phenomena of temperature profiles can be observed in Figs.8 and 9,respectively.The former was attributed to the adsorption of weaker adsorbed component while the later was caused by adsorption of strongly adsorbed component.Vertices in temperature profiles was closely related to concentration of component in feed mixture and its adsorption affinity as well as adsorption heat on AC adsorbent.The vertices of temperature profiles at the 5 cm far from feed end of adsorption bed is much higher than others that was attributed to co-adsorption of all component in feed mixture,in which the adsorption front of each specie has not been separated.Additionally,Comparison between experimental and simulation temperature profiles demonstrated that the developed mathematic model could reveal the experimental temperature profiles very well.

4.3.One stage VPSA process for coalbed methane enrichment

The well agreement between experimental results and simulation results for single component and multi-component breakthrough experiments verified the accuracy and reliability of mathematical model of adsorption bed proposed in this study,which means that it can be further used to predicate adsorption behaviors of adsorption process and to evaluate process performances of different cycle configuration.The six bed VPSA process with twelve steps cycle configuration for coalbed methane enrichment has been explored by numerical modeling and simulation.The resulted pressure profiles,temperature profiles and component profiles of VPSA process for low concentration coalbed methane enrichment are presented in Figs.10 and 11 to analyze the transient behaviors in adsorption bed.As seen from Figs.10and 11,adsorption pressure was kept at 3 bar (1 bar=105Pa) at adsorption step,temperature in adsorption bed increases as adsorption process is exothermic.The temperature wave can be observed in Fig.10b with the adsorption front of methane moves from feed end to product end during adsorption step.The pressure in adsorption bed decreases as three times pressure equalization was carried out,meanwhile adsorption front of methane in adsorption bed move forward to the top of bed.After three times pressure equalization,the methane has been further enriched in adsorption bed by reducing the amount of nitrogen retained in adsorption bed.At the Vacuum step shown in Fig.11a,the methane retained in adsorption bed can be desorbed effectively from adsorbent and recovered as heavy product.At the same time,the temperature in adsorption bed decreases and reaches the minimum value at the end vacuum step because the desorption process is endothermic.In addition,a slightly higher concentration of methane at the top of adsorption bed can be observed at the end of ER3 step because of a slightly breakthrough of methane at the end ED3 step.Nevertheless,this portion of methane has been pushed back to the bottom of adsorption bed by two successive equalizationrepressurization step and pressurization step,meanwhile the temperature increases simultaneously with the adsorption of weakly adsorbed component.

4.3.1.Effects of feed flowrate on process performances

Feed flowrate is one of key operating parameters that is closely related to unit productivity as well as others performance indicator.Effects of Feed flowrate on VPSA process performances for coalbed methane upgrading has been investigated and presented in Fig.12 with a feed flowrate ranged from 0.1 m3﹒min-1to 0.2 m3-﹒min-1.Methane was recovered as extract product since activated carbon is methane preferential adsorption material used in VPSA process.As expected,the purity of methane in product gas increase with increasing feed flowrate shown in Fig.12a,while an opposite trend can be observed for methane recovery.However,there is no obvious increase of methane purity when the feed flowrate is higher than 0.15 m3﹒min-1.It can be assumed that this feed flowrate has resulted a maximum utilization of AC adsorbent in adsorption bed and much higher feed flowrate would result in a significant breakthrough of methane in adsorption bed during adsorption step,which could be judged from the steeper decrease of methane recovery shown in Fig.12a.Hence,the feed flowrate of 0.15 m3﹒min-1is a better choice with other constant operating conditions to produce a methane product gas with methane purity higher and recovery higher than 56% and 94%,respectively.The trend of unit productivity with respect to feed flowrate is more similar to that of methane purity,as depicted in Fig.12b,while the energy consumption of VPSA process performed a minimum value at feed flowrate of 0.125 m3﹒min-1.Much higher feed flowrate with lower methane recovery would result in much higher energy consumption of VPSA process,since energy requirement of compressor for feed mixture could be not effectively utilized.Because low concentration methane existed in feed mixture and much more methane has been lost in exhaust gas.

4.3.2.Effects of feed concentration on process performances

Effects of methane concentration in feed mixture on process performances has been studied to represent coalbed methane withdifferent methane concentration.As shown in Fig.13,methane purity and unit productivity increase with an increase of methane molar fraction in feed mixture.The methane purity in product gas can increase from 32%to 89%as the methane molar fraction in feed mixture ranged from 0.1 to 0.5 by using six bed VPSA process.It’s due to that higher methane molar fraction in feed mixture means higher partial pressure of methane that was beneficial to methane adsorption on AC adsorbent and lead to higher methane purity in product gas.On the other hand,higher methane concentration in feed mixture also means much more methane has been introduced in adsorption bed and recovered in methane product gas at fixed feed flowrate,which has resulted in an increase of unit productivity.Additionally,a maximum value of 93.90% of CH4recovery can be observed at methane molar fraction of 0.2 with a methane purity of 56.04%.The CH4recovery decreases with increasing methane concentration in feed mixture when methane molar fraction is higher than 0.2.This was mainly limited by methane adsorption capacity of AC adsorbent even under high partial pressure of methane in adsorption bed and more methane would travel to exhaust gas.As depicted in Fig.13b,the energy consumption of VPSA process for methane enrichment decreases with increasing of methane concentration in feed mixture,because energy consumption of compressor and vacuum pump was mainly used for methane compression and methane recovered under the situation of higher methane concentration in feed mixture.All that has increased the efficiency of energy utilization.

Fig.6.Adsorption isotherms of CO2,CH4,and N2 on activated carbon at 293 K,308 K,and 323 K.Points-measured data,Solid line-dual site Langmuir model Fitting data.

4.3.3.Effects of desorption pressure on process performances

Desorption pressure is most important operating parameter that is more closely related to desorption of strongly adsorbed component,adsorbent regeneration as well as its effective working capacity.In the studied VPSA cycle configuration,the vacuum step without light product purge has been employed for strongly adsorbed component recovery and adsorbent regeneration because the purity of methane in heavy product gas was expected to be as high as possible.As shown in Fig.14a,higher purity and recovery of methane in product gas could be achieved at lower desorption pressure.As lowering desorption pressure could realize much deeper desorption of methane retained on AC adsorbent to increase methane purity,meanwhile it would also result in higher working capacity of AC adsorbent for both adsorbates,which is more helpful to reducethe loss of methane in exhaust gas as well as to achieve high methane recovery and unit productivity.As depicted in Fig.14b,lower desorption pressure can also result in a lower energy consumption of VPSA process for methane enrichment.The main reason is that higher methane recovery can be achieved at lower desorption pressure that has improved the efficient of energy consumption of compressor for recovered methane per mol.Moreover,the desorption pressure was better set to be 0.1 bar for six bed VPSA process for low concentration coalbed methane enrichment.

Fig.7.Single component breakthrough curve and temperature profiles of CO2,CH4,N2.

4.3.4.Effects of CO2feed concentration on process performances

A portion of CO2was assumed to be contained in feed mixture and its effects on VPSA process performances has been investigated and presented in Fig.15.As the CO2molar fraction varied from zero to 0.08,the methane purity decreases from 56.05%to 49.08%,while the methane recovery increases from 93.90% to 96.40%.It is interesting and weird that methane recovery increases with increasing of CO2concentration in feed mixture,because the methane recovery would decrease if this portion of CO2has been replaced with methane in feed mixture.We assumed that the displacement effectof CO2could lead to an increase of methane concentration and steeper adsorption front,which not enough to break through the adsorption bed.On the other hand,much less methane could retain in adsorption bed as the competitive adsorption between methane and carbon dioxide.Therefore,an increase of methane recovery can be observed within the studied CO2concentration.The decrease of methane purity is due to the dilution effect of collected CO2in heavy product gas.The increasing trend of unit productivity is corresponded to trend of methane recovery since the total amount of methane introduced into adsorption bed per cycle was kept constant.Additionally,an increase of energy consumption for coalbed methane upgrading can be observed with increasing of CO2concentration in feed mixture,which was caused by increasing of energy consumption of vacuum pump due to stronger adsorption affinity between carbon dioxide and AC adsorbent.

Fig.8.Binary feed mixture breakthrough curve and temperature profiles.

4.4.Dual-stage VPSA process for coal bed methane enrichment

The above parameters sensitivity analysis of VPSA process for low concentration coalbed methane enrichment have demonstrated that single train VPSA process was hard to enrich the low concentration coalbed methane (methane <30%) to the methane product gas with a methane purity high enough to be introduced natural gas pipeline.But,a methane product gas with high purity and high recovery could be achieved simultaneously by two stage VPSA process for methane enrichment,which can be deduced from studied results of effects of methane concentration in feed mixture on process performances.Therefore,a dualstage VPSA process has been proposed and investigated in this section for low concentration coalbed methane enrichment.The cycle sequence in the first stage VPSA process is almost same asTable 2,while the cycle sequence in the second stage VPSA process is the same as Table 2.Notably,a purge step was employed in the first stage VPSA process at the end of vacuum step,which has been used to enhance methane recovery as well as to achieve a better regeneration of adsorbent in adsorption bed.In addition,the adsorption pressure and the desorption pressure were set to be 3 bar and 0.1 bar,respectively,for both stage VPSA process.The composition of feed mixture was assigned to 20% methane balanced with nitrogen.

Fig.9.Three component feed mixture breakthrough curve and temperature profiles.

Detailed mass balance of the dual-stage VPSA process is depicted in Fig.16.In order to keep maximum recovery of methane,a recycle stream with methane concentration same as that in feed mixture was recycled from the second stage VPSA and mixed with fresh feed as the feed stream of the first stage VPSA process.The ratio of the amount of recycle stream to that of fresh feed was almost 21%,but it was not required to be compressed because it has been withdrawn as light product steam with adsorption pressure in the second stage VPSA process.As depicted in Fig.16,20% methane in mixed feed mixture could be enriched to a middle product gas with methane purity of 56% and a methane recovery of 98.93% in the first stage VPSA process,while the methane concentration in the exhaust gas is lower to 0.33%.At the second stage VPSA process,the methane product gas with methane purity higher than 90% could be obtained but with a slightly lower methane recovery of 82.48%,which has resulted in 21% recycle stream.However,from an overall perspective of dual-stage VPSA process,20% methane in fresh feed was able to be concentrated to 90.67%methane with a total methane recovery of 98.71%.

The power requirement of compressor and vacuum pump at each stage VPSA process as well as their share is illustrated in Fig.17.The total energy consumption in dual-stage VPSA for producing a methane product gas with methane purity higher than 90% was 40.64 kJ﹒mol-1CH4.The share of energy consumption of the first stage VPSA process was up to 68.72%,Because most of power requirement has been used to compress nitrogen contained in feed mixture and recover nitrogen retained in adsorption bed.Hence,lower adsorption pressure with advanced methane selective adsorbent could decrease significantly the total energy consumption of VPSA process for the enrichment of low concentration methane.In addition,the energy consumption for compressor and vacuum pump at the second stage are 7.96 kJ per mol CH4and 4.75 kJ per mol CH4,respectively,which account to 19.59% and11.69% of total power requirement.

Furthermore,the solid phase and gas phase concentration profile of methane in adsorption bed over one cycle at the first stage VPSA process are presented in Fig.18,after the first stage VPSA process has achieved cyclic steady state.The purge step deployed in cycle configuration is not only helpful for methane deeply desorbed from AC adsorbent,but also helpful to create a much steeper adsorption front of methane in adsorption bed,which has improved the utilization of AC adsorbent and enhanced methane recovery significantly.Unlike methane concentration profiles at first stage adsorption bed,the solid phase and gas phase concentration profiles of methane in the second stage adsorption bed over one cycle are slightly messy and chaotic shown in Fig.19,after the second stage VPSA process reached cyclic steady state.This is because a significant breakthrough of methane was allowed at equalization-depressurization step to increase methane concentration retained in adsorption bed.Moreover,two adsorption fronts of methane can be observed in adsorption bed during adsorption step as one of it has been created by ER3 step depicted in yellow line.

Fig.10.(a) Pressure profiles of six bed VPSA process over one cycle and (b) temperature profiles of adsorption bed over one cycle,under adsorption pressure of 3 bar and desorption pressure of 0.1 bar as well as feed mixture consisting of 0.2 CH4–0.8 N2 with a feed flowrate of 0.15 m3﹒min-1 at feed temperature of 293 K.

Fig.11.(a)CH4 solid phase concentration and(b)CH4 gas phase concentration in adsorption bed at the end of each step over one cycle,under adsorption pressure of 3 bar and desorption pressure of 0.1 bar as well as feed mixture consisting of 0.2 CH4–0.8 N2 with a feed flowrate of 0.15 m3﹒min-1 at feed temperature of 293 K.

5.Conclusions

Adsorption process with activated carbon adsorbent for the enrichment of low concentration coalbed methane has been explored in this study.Adsorption thermodynamics and adsorption kinetics of methane,nitrogen and carbon dioxide on AC adsorbent were measured and calculated by static volumetric isotherm measurement and breakthrough tests.The competitive adsorption behaviors have been observed in the breakthrough experiments with multi-component feed mixture.Moreover,a rigorous mathematical model with the linear driving force model accounting component transfer between gas phase and solid phase for adsorption bed has been developed,and it can reveal the breakthrough curves and temperature curves very well.The lumped mass transfer coefficient of methane,nitrogen and carbon dioxide on AC adsorbent were estimated to be 0.3 s-1,1 s-1,and 0.06 s-1,respectively.Based on the verified mathematical model of adsorption bed,a six bed VPSA process with twelve steps cycle configuration has been proposed for low concentration coalbed methane enrichment.The methane concentration in heavy product gas could up to 32.15% for the 10%methane feed and up to 88.75% for the 50% methane feed with methane recovery higher than 83% under adsorption pressure of 3 bar and desorption pressure of 0.1 bar.Furthermore,effects of key operating conditions on process performances in terms of methane purity,methane recovery,unit productivity and energy consumption,has been detailed investigated.Analysis of energy consumption of VPSA process has demonstrated that lower adsorption pressure of VPSA process with advanced adsorbent was more energy conservation for low concentration coalbed methane upgrading.Finally,a dual-stage VPSA process was further developed for low concentration coalbed methane enrichment with AC adsorbent.A low concentration coalbed methane contained 20% methane could be enriched to 90% methane with total methane recovery higher than 98%,meanwhile the energy consumption of dual-stage VPSA process for coalbed methane enrichment was lower to 0.504 kW﹒h﹒m-3CH4.

Fig.12.Effects of feed flowrate on performance indicators of VPSA process(a)Purity and Recovery and(b)Productivity and energy consumption(feed mixture:20%CH4–80%N2,desorption Pressure:0.1 bar).

Fig.13.Effects of CH4 concentration in feed mixture on performance indicators of VPSA process(a)Purity and Recovery and(b)Productivity and energy consumption(feed flowrate:0.15 m3﹒min-1,desorption pressure:0.1 bar).

Fig.14.Effects of regeneration pressure on performance indicators of VPSA process:(a) Purity and Recovery and (b) Productivity and energy consumption.(Feed flowrate:0.15 m3﹒min-1,Feed mixture:20%CH4-80%N2).

Fig.15.Effects of CO2 concentration in feed mixture on performance indicators of VPSA process:(a)Purity and Recovery and(b)Productivity and energy consumption.(Feed flowrate:0.15 m3﹒min-1,Feed composition:20%CH4,desorption pressure:0.1 bar).

Fig.16.The total mass balance of two stage VPSA process.

Fig.17.Energy consumption of compressor and vacuum pump in each stage for recovering 1 mol methane.

Fig.18.CH4 solid phase concentration (a) and CH4 gas phase concentration (b) in adsorption bed at the end of each step over one cycle at the first stage VPSA process.

Fig.19.(a) CH4 solid phase concentration and (b) CH4 gas phase concentration in adsorption bed at the end of each step over one cycle at the Second stage VPSA process.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work is part of the project supported by the Shanxi Coalbed Methane Joint Research Fund (2015012004).